blob: 30aa76d04d7bd0fabd53cee8a85a1b80b828dc8d [file] [log] [blame]
Willy Tarreau6a06a402007-07-15 20:15:28 +02001 ----------------------
Willy Tarreau8317b282014-04-23 01:49:41 +02002 HAProxy
Willy Tarreau6a06a402007-07-15 20:15:28 +02003 Configuration Manual
4 ----------------------
Willy Tarreau1db55792020-11-05 17:20:35 +01005 version 2.4
Willy Tarreau6a06a402007-07-15 20:15:28 +02006 willy tarreau
Christopher Fauletd83fd762021-10-04 08:46:17 +02007 2021/10/04
Willy Tarreau6a06a402007-07-15 20:15:28 +02008
9
10This document covers the configuration language as implemented in the version
Davor Ocelice9ed2812017-12-25 17:49:28 +010011specified above. It does not provide any hints, examples, or advice. For such
Willy Tarreau0ba27502007-12-24 16:55:16 +010012documentation, please refer to the Reference Manual or the Architecture Manual.
Davor Ocelice9ed2812017-12-25 17:49:28 +010013The summary below is meant to help you find sections by name and navigate
Willy Tarreauc57f0e22009-05-10 13:12:33 +020014through the document.
Willy Tarreau6a06a402007-07-15 20:15:28 +020015
Willy Tarreauc57f0e22009-05-10 13:12:33 +020016Note to documentation contributors :
Jamie Gloudonaaa21002012-08-25 00:18:33 -040017 This document is formatted with 80 columns per line, with even number of
Willy Tarreauc57f0e22009-05-10 13:12:33 +020018 spaces for indentation and without tabs. Please follow these rules strictly
19 so that it remains easily printable everywhere. If a line needs to be
20 printed verbatim and does not fit, please end each line with a backslash
Willy Tarreau62a36c42010-08-17 15:53:10 +020021 ('\') and continue on next line, indented by two characters. It is also
Davor Ocelice9ed2812017-12-25 17:49:28 +010022 sometimes useful to prefix all output lines (logs, console outputs) with 3
23 closing angle brackets ('>>>') in order to emphasize the difference between
24 inputs and outputs when they may be ambiguous. If you add sections,
Willy Tarreau62a36c42010-08-17 15:53:10 +020025 please update the summary below for easier searching.
Willy Tarreauc57f0e22009-05-10 13:12:33 +020026
27
28Summary
29-------
30
311. Quick reminder about HTTP
321.1. The HTTP transaction model
331.2. HTTP request
Davor Ocelice9ed2812017-12-25 17:49:28 +0100341.2.1. The request line
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200351.2.2. The request headers
361.3. HTTP response
Davor Ocelice9ed2812017-12-25 17:49:28 +0100371.3.1. The response line
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200381.3.2. The response headers
39
402. Configuring HAProxy
412.1. Configuration file format
William Lallemandf9873ba2015-05-05 17:37:14 +0200422.2. Quoting and escaping
William Lallemandb2f07452015-05-12 14:27:13 +0200432.3. Environment variables
Willy Tarreau4b103022021-02-12 17:59:10 +0100442.4. Conditional blocks
452.5. Time format
462.6. Examples
Willy Tarreauc57f0e22009-05-10 13:12:33 +020047
483. Global parameters
493.1. Process management and security
503.2. Performance tuning
513.3. Debugging
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +0100523.4. Userlists
Cyril Bontédc4d9032012-04-08 21:57:39 +0200533.5. Peers
Cyril Bonté307ee1e2015-09-28 23:16:06 +0200543.6. Mailers
William Lallemandc9515522019-06-12 16:32:11 +0200553.7. Programs
Christopher Faulet76edc0f2020-01-13 15:52:01 +0100563.8. HTTP-errors
Emeric Brun99c453d2020-05-25 15:01:04 +0200573.9. Rings
William Lallemand0217b7b2020-11-18 10:41:24 +0100583.10. Log forwarding
Willy Tarreauc57f0e22009-05-10 13:12:33 +020059
604. Proxies
614.1. Proxy keywords matrix
624.2. Alphabetically sorted keywords reference
63
Davor Ocelice9ed2812017-12-25 17:49:28 +0100645. Bind and server options
Willy Tarreau086fbf52012-09-24 20:34:51 +0200655.1. Bind options
665.2. Server and default-server options
Baptiste Assmann1fa66662015-04-14 00:28:47 +0200675.3. Server DNS resolution
685.3.1. Global overview
695.3.2. The resolvers section
Willy Tarreauc57f0e22009-05-10 13:12:33 +020070
Julien Pivotto6ccee412019-11-27 15:49:54 +0100716. Cache
726.1. Limitation
736.2. Setup
746.2.1. Cache section
756.2.2. Proxy section
76
Willy Tarreau74ca5042013-06-11 23:12:07 +0200777. Using ACLs and fetching samples
787.1. ACL basics
797.1.1. Matching booleans
807.1.2. Matching integers
817.1.3. Matching strings
827.1.4. Matching regular expressions (regexes)
837.1.5. Matching arbitrary data blocks
847.1.6. Matching IPv4 and IPv6 addresses
857.2. Using ACLs to form conditions
867.3. Fetching samples
Thierry FOURNIER060762e2014-04-23 13:29:15 +0200877.3.1. Converters
887.3.2. Fetching samples from internal states
897.3.3. Fetching samples at Layer 4
907.3.4. Fetching samples at Layer 5
917.3.5. Fetching samples from buffer contents (Layer 6)
927.3.6. Fetching HTTP samples (Layer 7)
Christopher Faulete596d182020-05-05 17:46:34 +0200937.3.7. Fetching samples for developers
Willy Tarreau74ca5042013-06-11 23:12:07 +0200947.4. Pre-defined ACLs
Willy Tarreauc57f0e22009-05-10 13:12:33 +020095
968. Logging
978.1. Log levels
988.2. Log formats
998.2.1. Default log format
1008.2.2. TCP log format
1018.2.3. HTTP log format
William Lallemand48940402012-01-30 16:47:22 +01001028.2.4. Custom log format
Willy Tarreau5f51e1a2012-12-03 18:40:10 +01001038.2.5. Error log format
Willy Tarreauc57f0e22009-05-10 13:12:33 +02001048.3. Advanced logging options
1058.3.1. Disabling logging of external tests
1068.3.2. Logging before waiting for the session to terminate
1078.3.3. Raising log level upon errors
1088.3.4. Disabling logging of successful connections
1098.4. Timing events
1108.5. Session state at disconnection
1118.6. Non-printable characters
1128.7. Capturing HTTP cookies
1138.8. Capturing HTTP headers
1148.9. Examples of logs
115
Christopher Fauletc3fe5332016-04-07 15:30:10 +02001169. Supported filters
1179.1. Trace
1189.2. HTTP compression
Christopher Fauletf7e4e7e2016-10-27 22:29:49 +02001199.3. Stream Processing Offload Engine (SPOE)
Christopher Faulet99a17a22018-12-11 09:18:27 +01001209.4. Cache
Christopher Fauletb30b3102019-09-12 23:03:09 +02001219.5. fcgi-app
Miroslav Zagoracdc32cd92020-12-13 18:32:57 +01001229.6. OpenTracing
Christopher Fauletc3fe5332016-04-07 15:30:10 +0200123
Christopher Fauletb30b3102019-09-12 23:03:09 +020012410. FastCGI applications
12510.1. Setup
12610.1.1. Fcgi-app section
12710.1.2. Proxy section
12810.1.3. Example
12910.2. Default parameters
13010.3. Limitations
131
Emeric Brunce325c42021-04-02 17:05:09 +020013211. Address formats
13311.1. Address family prefixes
13411.2. Socket type prefixes
13511.3. Protocol prefixes
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200136
1371. Quick reminder about HTTP
138----------------------------
139
Davor Ocelice9ed2812017-12-25 17:49:28 +0100140When HAProxy is running in HTTP mode, both the request and the response are
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200141fully analyzed and indexed, thus it becomes possible to build matching criteria
142on almost anything found in the contents.
143
144However, it is important to understand how HTTP requests and responses are
145formed, and how HAProxy decomposes them. It will then become easier to write
146correct rules and to debug existing configurations.
147
148
1491.1. The HTTP transaction model
150-------------------------------
151
152The HTTP protocol is transaction-driven. This means that each request will lead
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +0100153to one and only one response. Traditionally, a TCP connection is established
Davor Ocelice9ed2812017-12-25 17:49:28 +0100154from the client to the server, a request is sent by the client through the
155connection, the server responds, and the connection is closed. A new request
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200156will involve a new connection :
157
158 [CON1] [REQ1] ... [RESP1] [CLO1] [CON2] [REQ2] ... [RESP2] [CLO2] ...
159
160In this mode, called the "HTTP close" mode, there are as many connection
161establishments as there are HTTP transactions. Since the connection is closed
162by the server after the response, the client does not need to know the content
163length.
164
165Due to the transactional nature of the protocol, it was possible to improve it
166to avoid closing a connection between two subsequent transactions. In this mode
167however, it is mandatory that the server indicates the content length for each
168response so that the client does not wait indefinitely. For this, a special
169header is used: "Content-length". This mode is called the "keep-alive" mode :
170
171 [CON] [REQ1] ... [RESP1] [REQ2] ... [RESP2] [CLO] ...
172
173Its advantages are a reduced latency between transactions, and less processing
174power required on the server side. It is generally better than the close mode,
175but not always because the clients often limit their concurrent connections to
Patrick Mezard9ec2ec42010-06-12 17:02:45 +0200176a smaller value.
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200177
Willy Tarreau95c4e142017-11-26 12:18:55 +0100178Another improvement in the communications is the pipelining mode. It still uses
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200179keep-alive, but the client does not wait for the first response to send the
180second request. This is useful for fetching large number of images composing a
181page :
182
183 [CON] [REQ1] [REQ2] ... [RESP1] [RESP2] [CLO] ...
184
185This can obviously have a tremendous benefit on performance because the network
186latency is eliminated between subsequent requests. Many HTTP agents do not
187correctly support pipelining since there is no way to associate a response with
188the corresponding request in HTTP. For this reason, it is mandatory for the
Cyril Bonté78caf842010-03-10 22:41:43 +0100189server to reply in the exact same order as the requests were received.
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200190
Willy Tarreau95c4e142017-11-26 12:18:55 +0100191The next improvement is the multiplexed mode, as implemented in HTTP/2. This
192time, each transaction is assigned a single stream identifier, and all streams
193are multiplexed over an existing connection. Many requests can be sent in
194parallel by the client, and responses can arrive in any order since they also
195carry the stream identifier.
196
Willy Tarreau70dffda2014-01-30 03:07:23 +0100197By default HAProxy operates in keep-alive mode with regards to persistent
198connections: for each connection it processes each request and response, and
199leaves the connection idle on both sides between the end of a response and the
Willy Tarreau95c4e142017-11-26 12:18:55 +0100200start of a new request. When it receives HTTP/2 connections from a client, it
201processes all the requests in parallel and leaves the connection idling,
202waiting for new requests, just as if it was a keep-alive HTTP connection.
Patrick Mezard9ec2ec42010-06-12 17:02:45 +0200203
Christopher Faulet315b39c2018-09-21 16:26:19 +0200204HAProxy supports 4 connection modes :
Willy Tarreau70dffda2014-01-30 03:07:23 +0100205 - keep alive : all requests and responses are processed (default)
206 - tunnel : only the first request and response are processed,
Christopher Faulet6c9bbb22019-03-26 21:37:23 +0100207 everything else is forwarded with no analysis (deprecated).
Willy Tarreau70dffda2014-01-30 03:07:23 +0100208 - server close : the server-facing connection is closed after the response.
Christopher Faulet315b39c2018-09-21 16:26:19 +0200209 - close : the connection is actively closed after end of response.
Willy Tarreau70dffda2014-01-30 03:07:23 +0100210
Willy Tarreau95c4e142017-11-26 12:18:55 +0100211
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200212
2131.2. HTTP request
214-----------------
215
216First, let's consider this HTTP request :
217
218 Line Contents
Willy Tarreaud72758d2010-01-12 10:42:19 +0100219 number
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200220 1 GET /serv/login.php?lang=en&profile=2 HTTP/1.1
221 2 Host: www.mydomain.com
222 3 User-agent: my small browser
223 4 Accept: image/jpeg, image/gif
224 5 Accept: image/png
225
226
2271.2.1. The Request line
228-----------------------
229
230Line 1 is the "request line". It is always composed of 3 fields :
231
232 - a METHOD : GET
233 - a URI : /serv/login.php?lang=en&profile=2
234 - a version tag : HTTP/1.1
235
236All of them are delimited by what the standard calls LWS (linear white spaces),
237which are commonly spaces, but can also be tabs or line feeds/carriage returns
238followed by spaces/tabs. The method itself cannot contain any colon (':') and
239is limited to alphabetic letters. All those various combinations make it
240desirable that HAProxy performs the splitting itself rather than leaving it to
241the user to write a complex or inaccurate regular expression.
242
243The URI itself can have several forms :
244
245 - A "relative URI" :
246
247 /serv/login.php?lang=en&profile=2
248
249 It is a complete URL without the host part. This is generally what is
250 received by servers, reverse proxies and transparent proxies.
251
252 - An "absolute URI", also called a "URL" :
253
254 http://192.168.0.12:8080/serv/login.php?lang=en&profile=2
255
256 It is composed of a "scheme" (the protocol name followed by '://'), a host
257 name or address, optionally a colon (':') followed by a port number, then
258 a relative URI beginning at the first slash ('/') after the address part.
259 This is generally what proxies receive, but a server supporting HTTP/1.1
260 must accept this form too.
261
262 - a star ('*') : this form is only accepted in association with the OPTIONS
263 method and is not relayable. It is used to inquiry a next hop's
264 capabilities.
Willy Tarreaud72758d2010-01-12 10:42:19 +0100265
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200266 - an address:port combination : 192.168.0.12:80
267 This is used with the CONNECT method, which is used to establish TCP
268 tunnels through HTTP proxies, generally for HTTPS, but sometimes for
269 other protocols too.
270
271In a relative URI, two sub-parts are identified. The part before the question
272mark is called the "path". It is typically the relative path to static objects
273on the server. The part after the question mark is called the "query string".
274It is mostly used with GET requests sent to dynamic scripts and is very
275specific to the language, framework or application in use.
276
Willy Tarreau95c4e142017-11-26 12:18:55 +0100277HTTP/2 doesn't convey a version information with the request, so the version is
Davor Ocelice9ed2812017-12-25 17:49:28 +0100278assumed to be the same as the one of the underlying protocol (i.e. "HTTP/2").
Willy Tarreau95c4e142017-11-26 12:18:55 +0100279
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200280
2811.2.2. The request headers
282--------------------------
283
284The headers start at the second line. They are composed of a name at the
285beginning of the line, immediately followed by a colon (':'). Traditionally,
286an LWS is added after the colon but that's not required. Then come the values.
287Multiple identical headers may be folded into one single line, delimiting the
288values with commas, provided that their order is respected. This is commonly
289encountered in the "Cookie:" field. A header may span over multiple lines if
290the subsequent lines begin with an LWS. In the example in 1.2, lines 4 and 5
291define a total of 3 values for the "Accept:" header.
292
Davor Ocelice9ed2812017-12-25 17:49:28 +0100293Contrary to a common misconception, header names are not case-sensitive, and
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200294their values are not either if they refer to other header names (such as the
Willy Tarreau95c4e142017-11-26 12:18:55 +0100295"Connection:" header). In HTTP/2, header names are always sent in lower case,
Willy Tarreau253c2512020-07-07 15:55:23 +0200296as can be seen when running in debug mode. Internally, all header names are
297normalized to lower case so that HTTP/1.x and HTTP/2 use the exact same
298representation, and they are sent as-is on the other side. This explains why an
299HTTP/1.x request typed with camel case is delivered in lower case.
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200300
301The end of the headers is indicated by the first empty line. People often say
302that it's a double line feed, which is not exact, even if a double line feed
303is one valid form of empty line.
304
305Fortunately, HAProxy takes care of all these complex combinations when indexing
306headers, checking values and counting them, so there is no reason to worry
307about the way they could be written, but it is important not to accuse an
308application of being buggy if it does unusual, valid things.
309
310Important note:
Lukas Tribus23953682017-04-28 13:24:30 +0000311 As suggested by RFC7231, HAProxy normalizes headers by replacing line breaks
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200312 in the middle of headers by LWS in order to join multi-line headers. This
313 is necessary for proper analysis and helps less capable HTTP parsers to work
314 correctly and not to be fooled by such complex constructs.
315
316
3171.3. HTTP response
318------------------
319
320An HTTP response looks very much like an HTTP request. Both are called HTTP
321messages. Let's consider this HTTP response :
322
323 Line Contents
Willy Tarreaud72758d2010-01-12 10:42:19 +0100324 number
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200325 1 HTTP/1.1 200 OK
326 2 Content-length: 350
327 3 Content-Type: text/html
328
Willy Tarreau816b9792009-09-15 21:25:21 +0200329As a special case, HTTP supports so called "Informational responses" as status
330codes 1xx. These messages are special in that they don't convey any part of the
331response, they're just used as sort of a signaling message to ask a client to
Willy Tarreau5843d1a2010-02-01 15:13:32 +0100332continue to post its request for instance. In the case of a status 100 response
333the requested information will be carried by the next non-100 response message
334following the informational one. This implies that multiple responses may be
335sent to a single request, and that this only works when keep-alive is enabled
336(1xx messages are HTTP/1.1 only). HAProxy handles these messages and is able to
337correctly forward and skip them, and only process the next non-100 response. As
338such, these messages are neither logged nor transformed, unless explicitly
339state otherwise. Status 101 messages indicate that the protocol is changing
Daniel Corbett9f0843f2021-05-08 10:50:37 -0400340over the same connection and that HAProxy must switch to tunnel mode, just as
Willy Tarreau5843d1a2010-02-01 15:13:32 +0100341if a CONNECT had occurred. Then the Upgrade header would contain additional
342information about the type of protocol the connection is switching to.
Willy Tarreau816b9792009-09-15 21:25:21 +0200343
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200344
Davor Ocelice9ed2812017-12-25 17:49:28 +01003451.3.1. The response line
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200346------------------------
347
348Line 1 is the "response line". It is always composed of 3 fields :
349
350 - a version tag : HTTP/1.1
351 - a status code : 200
352 - a reason : OK
353
354The status code is always 3-digit. The first digit indicates a general status :
Davor Ocelice9ed2812017-12-25 17:49:28 +0100355 - 1xx = informational message to be skipped (e.g. 100, 101)
356 - 2xx = OK, content is following (e.g. 200, 206)
357 - 3xx = OK, no content following (e.g. 302, 304)
358 - 4xx = error caused by the client (e.g. 401, 403, 404)
359 - 5xx = error caused by the server (e.g. 500, 502, 503)
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200360
Lukas Tribus23953682017-04-28 13:24:30 +0000361Please refer to RFC7231 for the detailed meaning of all such codes. The
Willy Tarreaud72758d2010-01-12 10:42:19 +0100362"reason" field is just a hint, but is not parsed by clients. Anything can be
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200363found there, but it's a common practice to respect the well-established
364messages. It can be composed of one or multiple words, such as "OK", "Found",
365or "Authentication Required".
366
Davor Ocelice9ed2812017-12-25 17:49:28 +0100367HAProxy may emit the following status codes by itself :
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200368
369 Code When / reason
370 200 access to stats page, and when replying to monitoring requests
371 301 when performing a redirection, depending on the configured code
372 302 when performing a redirection, depending on the configured code
373 303 when performing a redirection, depending on the configured code
Willy Tarreaub67fdc42013-03-29 19:28:11 +0100374 307 when performing a redirection, depending on the configured code
375 308 when performing a redirection, depending on the configured code
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200376 400 for an invalid or too large request
377 401 when an authentication is required to perform the action (when
378 accessing the stats page)
Christopher Faulet87f1f3d2019-07-18 14:51:20 +0200379 403 when a request is forbidden by a "http-request deny" rule
Florian Tham9205fea2020-01-08 13:35:30 +0100380 404 when the requested resource could not be found
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200381 408 when the request timeout strikes before the request is complete
Florian Tham272e29b2020-01-08 10:19:05 +0100382 410 when the requested resource is no longer available and will not
383 be available again
Daniel Corbett9f0843f2021-05-08 10:50:37 -0400384 500 when HAProxy encounters an unrecoverable internal error, such as a
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200385 memory allocation failure, which should never happen
Daniel Corbett9f0843f2021-05-08 10:50:37 -0400386 501 when HAProxy is unable to satisfy a client request because of an
Christopher Faulete095f312020-12-07 11:22:24 +0100387 unsupported feature
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200388 502 when the server returns an empty, invalid or incomplete response, or
Christopher Faulet87f1f3d2019-07-18 14:51:20 +0200389 when an "http-response deny" rule blocks the response.
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200390 503 when no server was available to handle the request, or in response to
391 monitoring requests which match the "monitor fail" condition
392 504 when the response timeout strikes before the server responds
393
394The error 4xx and 5xx codes above may be customized (see "errorloc" in section
3954.2).
396
397
3981.3.2. The response headers
399---------------------------
400
401Response headers work exactly like request headers, and as such, HAProxy uses
402the same parsing function for both. Please refer to paragraph 1.2.2 for more
403details.
404
405
4062. Configuring HAProxy
407----------------------
408
4092.1. Configuration file format
410------------------------------
Willy Tarreau6a06a402007-07-15 20:15:28 +0200411
412HAProxy's configuration process involves 3 major sources of parameters :
413
414 - the arguments from the command-line, which always take precedence
Willy Tarreau6f1129d2020-11-25 19:58:20 +0100415 - the configuration file(s), whose format is described here
Thayne McCombscdbcca92021-01-07 21:24:41 -0700416 - the running process's environment, in case some environment variables are
Willy Tarreau6f1129d2020-11-25 19:58:20 +0100417 explicitly referenced
Willy Tarreau6a06a402007-07-15 20:15:28 +0200418
Willy Tarreau6f1129d2020-11-25 19:58:20 +0100419The configuration file follows a fairly simple hierarchical format which obey
420a few basic rules:
Willy Tarreau0ba27502007-12-24 16:55:16 +0100421
Willy Tarreau6f1129d2020-11-25 19:58:20 +0100422 1. a configuration file is an ordered sequence of statements
423
424 2. a statement is a single non-empty line before any unprotected "#" (hash)
425
426 3. a line is a series of tokens or "words" delimited by unprotected spaces or
427 tab characters
428
429 4. the first word or sequence of words of a line is one of the keywords or
430 keyword sequences listed in this document
431
432 5. all other words are all arguments of the first one, some being well-known
433 keywords listed in this document, others being values, references to other
434 parts of the configuration, or expressions
435
436 6. certain keywords delimit a section inside which only a subset of keywords
437 are supported
438
439 7. a section ends at the end of a file or on a special keyword starting a new
440 section
441
442This is all that is needed to know to write a simple but reliable configuration
443generator, but this is not enough to reliably parse any configuration nor to
444figure how to deal with certain corner cases.
445
446First, there are a few consequences of the rules above. Rule 6 and 7 imply that
447the keywords used to define a new section are valid everywhere and cannot have
448a different meaning in a specific section. These keywords are always a single
449word (as opposed to a sequence of words), and traditionally the section that
450follows them is designated using the same name. For example when speaking about
451the "global section", it designates the section of configuration that follows
452the "global" keyword. This usage is used a lot in error messages to help locate
453the parts that need to be addressed.
454
455A number of sections create an internal object or configuration space, which
456requires to be distinguished from other ones. In this case they will take an
457extra word which will set the name of this particular section. For some of them
458the section name is mandatory. For example "frontend foo" will create a new
459section of type "frontend" named "foo". Usually a name is specific to its
460section and two sections of different types may use the same name, but this is
461not recommended as it tends to complexify configuration management.
462
463A direct consequence of rule 7 is that when multiple files are read at once,
464each of them must start with a new section, and the end of each file will end
465a section. A file cannot contain sub-sections nor end an existing section and
466start a new one.
467
468Rule 1 mentioned that ordering matters. Indeed, some keywords create directives
469that can be repeated multiple times to create ordered sequences of rules to be
470applied in a certain order. For example "tcp-request" can be used to alternate
471"accept" and "reject" rules on varying criteria. As such, a configuration file
472processor must always preserve a section's ordering when editing a file. The
473ordering of sections usually does not matter except for the global section
474which must be placed before other sections, but it may be repeated if needed.
475In addition, some automatic identifiers may automatically be assigned to some
476of the created objects (e.g. proxies), and by reordering sections, their
477identifiers will change. These ones appear in the statistics for example. As
478such, the configuration below will assign "foo" ID number 1 and "bar" ID number
4792, which will be swapped if the two sections are reversed:
480
481 listen foo
482 bind :80
483
484 listen bar
485 bind :81
486
487Another important point is that according to rules 2 and 3 above, empty lines,
488spaces, tabs, and comments following and unprotected "#" character are not part
489of the configuration as they are just used as delimiters. This implies that the
490following configurations are strictly equivalent:
491
492 global#this is the global section
493 daemon#daemonize
494 frontend foo
495 mode http # or tcp
496
497and:
498
499 global
500 daemon
501
502 # this is the public web frontend
503 frontend foo
504 mode http
505
506The common practice is to align to the left only the keyword that initiates a
507new section, and indent (i.e. prepend a tab character or a few spaces) all
508other keywords so that it's instantly visible that they belong to the same
509section (as done in the second example above). Placing comments before a new
510section helps the reader decide if it's the desired one. Leaving a blank line
511at the end of a section also visually helps spotting the end when editing it.
512
513Tabs are very convenient for indent but they do not copy-paste well. If spaces
514are used instead, it is recommended to avoid placing too many (2 to 4) so that
515editing in field doesn't become a burden with limited editors that do not
516support automatic indent.
517
518In the early days it used to be common to see arguments split at fixed tab
519positions because most keywords would not take more than two arguments. With
520modern versions featuring complex expressions this practice does not stand
521anymore, and is not recommended.
522
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200523
William Lallemandf9873ba2015-05-05 17:37:14 +02005242.2. Quoting and escaping
525-------------------------
526
Willy Tarreau6f1129d2020-11-25 19:58:20 +0100527In modern configurations, some arguments require the use of some characters
528that were previously considered as pure delimiters. In order to make this
529possible, HAProxy supports character escaping by prepending a backslash ('\')
530in front of the character to be escaped, weak quoting within double quotes
531('"') and strong quoting within single quotes ("'").
William Lallemandf9873ba2015-05-05 17:37:14 +0200532
Willy Tarreau6f1129d2020-11-25 19:58:20 +0100533This is pretty similar to what is done in a number of programming languages and
534very close to what is commonly encountered in Bourne shell. The principle is
535the following: while the configuration parser cuts the lines into words, it
536also takes care of quotes and backslashes to decide whether a character is a
537delimiter or is the raw representation of this character within the current
538word. The escape character is then removed, the quotes are removed, and the
539remaining word is used as-is as a keyword or argument for example.
William Lallemandf9873ba2015-05-05 17:37:14 +0200540
Willy Tarreau6f1129d2020-11-25 19:58:20 +0100541If a backslash is needed in a word, it must either be escaped using itself
542(i.e. double backslash) or be strongly quoted.
543
544Escaping outside quotes is achieved by preceding a special character by a
545backslash ('\'):
William Lallemandf9873ba2015-05-05 17:37:14 +0200546
547 \ to mark a space and differentiate it from a delimiter
548 \# to mark a hash and differentiate it from a comment
549 \\ to use a backslash
550 \' to use a single quote and differentiate it from strong quoting
551 \" to use a double quote and differentiate it from weak quoting
552
Willy Tarreau6f1129d2020-11-25 19:58:20 +0100553In addition, a few non-printable characters may be emitted using their usual
554C-language representation:
555
556 \n to insert a line feed (LF, character \x0a or ASCII 10 decimal)
557 \r to insert a carriage return (CR, character \x0d or ASCII 13 decimal)
558 \t to insert a tab (character \x09 or ASCII 9 decimal)
559 \xNN to insert character having ASCII code hex NN (e.g \x0a for LF).
560
561Weak quoting is achieved by surrounding double quotes ("") around the character
562or sequence of characters to protect. Weak quoting prevents the interpretation
563of:
William Lallemandf9873ba2015-05-05 17:37:14 +0200564
Willy Tarreau6f1129d2020-11-25 19:58:20 +0100565 space or tab as a word separator
William Lallemandf9873ba2015-05-05 17:37:14 +0200566 ' single quote as a strong quoting delimiter
567 # hash as a comment start
568
Willy Tarreau6f1129d2020-11-25 19:58:20 +0100569Weak quoting permits the interpretation of environment variables (which are not
570evaluated outside of quotes) by preceding them with a dollar sign ('$'). If a
571dollar character is needed inside double quotes, it must be escaped using a
572backslash.
William Lallemandb2f07452015-05-12 14:27:13 +0200573
Willy Tarreau6f1129d2020-11-25 19:58:20 +0100574Strong quoting is achieved by surrounding single quotes ('') around the
575character or sequence of characters to protect. Inside single quotes, nothing
576is interpreted, it's the efficient way to quote regular expressions.
William Lallemandf9873ba2015-05-05 17:37:14 +0200577
Willy Tarreau6f1129d2020-11-25 19:58:20 +0100578As a result, here is the matrix indicating how special characters can be
579entered in different contexts (unprintable characters are replaced with their
580name within angle brackets). Note that some characters that may only be
581represented escaped have no possible representation inside single quotes,
582hence the '-' there:
William Lallemandf9873ba2015-05-05 17:37:14 +0200583
Willy Tarreau6f1129d2020-11-25 19:58:20 +0100584 Character | Unquoted | Weakly quoted | Strongly quoted
585 -----------+---------------+-----------------------------+-----------------
586 <TAB> | \<TAB>, \x09 | "<TAB>", "\<TAB>", "\x09" | '<TAB>'
587 <LF> | \n, \x0a | "\n", "\x0a" | -
588 <CR> | \r, \x0d | "\r", "\x0d" | -
589 <SPC> | \<SPC>, \x20 | "<SPC>", "\<SPC>", "\x20" | '<SPC>'
590 " | \", \x22 | "\"", "\x22" | '"'
591 # | \#, \x23 | "#", "\#", "\x23" | '#'
592 $ | $, \$, \x24 | "\$", "\x24" | '$'
593 ' | \', \x27 | "'", "\'", "\x27" | -
594 \ | \\, \x5c | "\\", "\x5c" | '\'
William Lallemandf9873ba2015-05-05 17:37:14 +0200595
596 Example:
Willy Tarreau6f1129d2020-11-25 19:58:20 +0100597 # those are all strictly equivalent:
William Lallemandf9873ba2015-05-05 17:37:14 +0200598 log-format %{+Q}o\ %t\ %s\ %{-Q}r
599 log-format "%{+Q}o %t %s %{-Q}r"
600 log-format '%{+Q}o %t %s %{-Q}r'
601 log-format "%{+Q}o %t"' %s %{-Q}r'
602 log-format "%{+Q}o %t"' %s'\ %{-Q}r
603
Willy Tarreau6f1129d2020-11-25 19:58:20 +0100604There is one particular case where a second level of quoting or escaping may be
605necessary. Some keywords take arguments within parenthesis, sometimes delimited
606by commas. These arguments are commonly integers or predefined words, but when
607they are arbitrary strings, it may be required to perform a separate level of
608escaping to disambiguate the characters that belong to the argument from the
609characters that are used to delimit the arguments themselves. A pretty common
610case is the "regsub" converter. It takes a regular expression in argument, and
611if a closing parenthesis is needed inside, this one will require to have its
612own quotes.
613
614The keyword argument parser is exactly the same as the top-level one regarding
Thayne McCombsf87e4f72021-10-04 01:02:58 -0600615quotes, except that the \#, \$, and \xNN escapes are not processed. But what is
Ilya Shipitsin2272d8a2020-12-21 01:22:40 +0500616not always obvious is that the delimiters used inside must first be escaped or
Willy Tarreau6f1129d2020-11-25 19:58:20 +0100617quoted so that they are not resolved at the top level.
618
619Let's take this example making use of the "regsub" converter which takes 3
620arguments, one regular expression, one replacement string and one set of flags:
621
622 # replace all occurrences of "foo" with "blah" in the path:
623 http-request set-path %[path,regsub(foo,blah,g)]
624
625Here no special quoting was necessary. But if now we want to replace either
626"foo" or "bar" with "blah", we'll need the regular expression "(foo|bar)". We
627cannot write:
628
629 http-request set-path %[path,regsub((foo|bar),blah,g)]
630
631because we would like the string to cut like this:
632
633 http-request set-path %[path,regsub((foo|bar),blah,g)]
634 |---------|----|-|
635 arg1 _/ / /
636 arg2 __________/ /
637 arg3 ______________/
638
639but actually what is passed is a string between the opening and closing
640parenthesis then garbage:
641
642 http-request set-path %[path,regsub((foo|bar),blah,g)]
643 |--------|--------|
644 arg1=(foo|bar _/ /
645 trailing garbage _________/
646
647The obvious solution here seems to be that the closing parenthesis needs to be
648quoted, but alone this will not work, because as mentioned above, quotes are
649processed by the top-level parser which will resolve them before processing
650this word:
651
652 http-request set-path %[path,regsub("(foo|bar)",blah,g)]
653 ------------ -------- ----------------------------------
654 word1 word2 word3=%[path,regsub((foo|bar),blah,g)]
655
656So we didn't change anything for the argument parser at the second level which
657still sees a truncated regular expression as the only argument, and garbage at
658the end of the string. By escaping the quotes they will be passed unmodified to
659the second level:
660
661 http-request set-path %[path,regsub(\"(foo|bar)\",blah,g)]
662 ------------ -------- ------------------------------------
663 word1 word2 word3=%[path,regsub("(foo|bar)",blah,g)]
664 |---------||----|-|
665 arg1=(foo|bar) _/ / /
666 arg2=blah ___________/ /
667 arg3=g _______________/
668
Ilya Shipitsin2272d8a2020-12-21 01:22:40 +0500669Another approach consists in using single quotes outside the whole string and
Willy Tarreau6f1129d2020-11-25 19:58:20 +0100670double quotes inside (so that the double quotes are not stripped again):
671
672 http-request set-path '%[path,regsub("(foo|bar)",blah,g)]'
673 ------------ -------- ----------------------------------
674 word1 word2 word3=%[path,regsub("(foo|bar)",blah,g)]
675 |---------||----|-|
676 arg1=(foo|bar) _/ / /
677 arg2 ___________/ /
678 arg3 _______________/
679
680When using regular expressions, it can happen that the dollar ('$') character
681appears in the expression or that a backslash ('\') is used in the replacement
682string. In this case these ones will also be processed inside the double quotes
683thus single quotes are preferred (or double escaping). Example:
684
685 http-request set-path '%[path,regsub("^/(here)(/|$)","my/\1",g)]'
686 ------------ -------- -----------------------------------------
687 word1 word2 word3=%[path,regsub("^/(here)(/|$)","my/\1",g)]
688 |-------------| |-----||-|
689 arg1=(here)(/|$) _/ / /
690 arg2=my/\1 ________________/ /
691 arg3 ______________________/
692
Daniel Corbett9f0843f2021-05-08 10:50:37 -0400693Remember that backslashes are not escape characters within single quotes and
Thayne McCombsf87e4f72021-10-04 01:02:58 -0600694that the whole word above is already protected against them using the single
Willy Tarreau6f1129d2020-11-25 19:58:20 +0100695quotes. Conversely, if double quotes had been used around the whole expression,
696single the dollar character and the backslashes would have been resolved at top
697level, breaking the argument contents at the second level.
698
Thayne McCombsf87e4f72021-10-04 01:02:58 -0600699Unfortunately, since single quotes can't be escaped inside of strong quoting,
700if you need to include single quotes in your argument, you will need to escape
701or quote them twice. There are a few ways to do this:
702
703 http-request set-var(txn.foo) str("\\'foo\\'")
704 http-request set-var(txn.foo) str(\"\'foo\'\")
705 http-request set-var(txn.foo) str(\\\'foo\\\')
706
Willy Tarreau6f1129d2020-11-25 19:58:20 +0100707When in doubt, simply do not use quotes anywhere, and start to place single or
708double quotes around arguments that require a comma or a closing parenthesis,
Thayne McCombsf87e4f72021-10-04 01:02:58 -0600709and think about escaping these quotes using a backslash if the string contains
Willy Tarreau6f1129d2020-11-25 19:58:20 +0100710a dollar or a backslash. Again, this is pretty similar to what is used under
711a Bourne shell when double-escaping a command passed to "eval". For API writers
712the best is probably to place escaped quotes around each and every argument,
713regardless of their contents. Users will probably find that using single quotes
714around the whole expression and double quotes around each argument provides
715more readable configurations.
William Lallemandf9873ba2015-05-05 17:37:14 +0200716
717
William Lallemandb2f07452015-05-12 14:27:13 +02007182.3. Environment variables
719--------------------------
720
721HAProxy's configuration supports environment variables. Those variables are
722interpreted only within double quotes. Variables are expanded during the
723configuration parsing. Variable names must be preceded by a dollar ("$") and
724optionally enclosed with braces ("{}") similarly to what is done in Bourne
725shell. Variable names can contain alphanumerical characters or the character
Amaury Denoyellefa41cb62020-10-01 14:32:35 +0200726underscore ("_") but should not start with a digit. If the variable contains a
727list of several values separated by spaces, it can be expanded as individual
728arguments by enclosing the variable with braces and appending the suffix '[*]'
729before the closing brace.
William Lallemandb2f07452015-05-12 14:27:13 +0200730
731 Example:
732
733 bind "fd@${FD_APP1}"
734
735 log "${LOCAL_SYSLOG}:514" local0 notice # send to local server
736
737 user "$HAPROXY_USER"
738
William Lallemand4d03e432019-06-14 15:35:37 +0200739Some variables are defined by HAProxy, they can be used in the configuration
740file, or could be inherited by a program (See 3.7. Programs):
William Lallemanddaf4cd22018-04-17 16:46:13 +0200741
William Lallemand4d03e432019-06-14 15:35:37 +0200742* HAPROXY_LOCALPEER: defined at the startup of the process which contains the
743 name of the local peer. (See "-L" in the management guide.)
744
745* HAPROXY_CFGFILES: list of the configuration files loaded by HAProxy,
746 separated by semicolons. Can be useful in the case you specified a
747 directory.
748
749* HAPROXY_MWORKER: In master-worker mode, this variable is set to 1.
750
John Roeslerfb2fce12019-07-10 15:45:51 -0500751* HAPROXY_CLI: configured listeners addresses of the stats socket for every
William Lallemand4d03e432019-06-14 15:35:37 +0200752 processes, separated by semicolons.
753
John Roeslerfb2fce12019-07-10 15:45:51 -0500754* HAPROXY_MASTER_CLI: In master-worker mode, listeners addresses of the master
William Lallemand4d03e432019-06-14 15:35:37 +0200755 CLI, separated by semicolons.
756
Willy Tarreaua46f1af2021-05-06 10:25:11 +0200757In addition, some pseudo-variables are internally resolved and may be used as
758regular variables. Pseudo-variables always start with a dot ('.'), and are the
759only ones where the dot is permitted. The current list of pseudo-variables is:
760
761* .FILE: the name of the configuration file currently being parsed.
762
763* .LINE: the line number of the configuration file currently being parsed,
764 starting at one.
765
766* .SECTION: the name of the section currently being parsed, or its type if the
767 section doesn't have a name (e.g. "global"), or an empty string before the
768 first section.
769
770These variables are resolved at the location where they are parsed. For example
771if a ".LINE" variable is used in a "log-format" directive located in a defaults
772section, its line number will be resolved before parsing and compiling the
773"log-format" directive, so this same line number will be reused by subsequent
774proxies.
775
776This way it is possible to emit information to help locate a rule in variables,
777logs, error statuses, health checks, header values, or even to use line numbers
778to name some config objects like servers for example.
779
William Lallemand4d03e432019-06-14 15:35:37 +0200780See also "external-check command" for other variables.
William Lallemandb2f07452015-05-12 14:27:13 +0200781
Willy Tarreau4b103022021-02-12 17:59:10 +0100782
7832.4. Conditional blocks
784-----------------------
785
786It may sometimes be convenient to be able to conditionally enable or disable
787some arbitrary parts of the configuration, for example to enable/disable SSL or
788ciphers, enable or disable some pre-production listeners without modifying the
789configuration, or adjust the configuration's syntax to support two distinct
790versions of HAProxy during a migration.. HAProxy brings a set of nestable
791preprocessor-like directives which allow to integrate or ignore some blocks of
792text. These directives must be placed on their own line and they act on the
793lines that follow them. Two of them support an expression, the other ones only
794switch to an alternate block or end a current level. The 4 following directives
795are defined to form conditional blocks:
796
797 - .if <condition>
798 - .elif <condition>
799 - .else
800 - .endif
801
802The ".if" directive nests a new level, ".elif" stays at the same level, ".else"
803as well, and ".endif" closes a level. Each ".if" must be terminated by a
804matching ".endif". The ".elif" may only be placed after ".if" or ".elif", and
805there is no limit to the number of ".elif" that may be chained. There may be
806only one ".else" per ".if" and it must always be after the ".if" or the last
807".elif" of a block.
808
809Comments may be placed on the same line if needed after a '#', they will be
810ignored. The directives are tokenized like other configuration directives, and
811as such it is possible to use environment variables in conditions.
812
813The conditions are currently limited to:
814
815 - an empty string, always returns "false"
816 - the integer zero ('0'), always returns "false"
817 - a non-nul integer (e.g. '1'), always returns "true".
Willy Tarreau42ed14b2021-05-06 15:55:14 +0200818 - a predicate optionally followed by argument(s) in parenthesis.
819
820The list of currently supported predicates is the following:
821
822 - defined(<name>) : returns true if an environment variable <name>
823 exists, regardless of its contents
824
Willy Tarreau58ca7062021-05-06 16:34:23 +0200825 - feature(<name>) : returns true if feature <name> is listed as present
826 in the features list reported by "haproxy -vv"
827 (which means a <name> appears after a '+')
828
Willy Tarreau6492e872021-05-06 16:10:09 +0200829 - streq(<str1>,<str2>) : returns true only if the two strings are equal
830 - strneq(<str1>,<str2>) : returns true only if the two strings differ
831
Willy Tarreau0b7c78a2021-05-06 16:53:26 +0200832 - version_atleast(<ver>): returns true if the current haproxy version is
833 at least as recent as <ver> otherwise false. The
834 version syntax is the same as shown by "haproxy -v"
835 and missing components are assumed as being zero.
836
837 - version_before(<ver>) : returns true if the current haproxy version is
838 strictly older than <ver> otherwise false. The
839 version syntax is the same as shown by "haproxy -v"
840 and missing components are assumed as being zero.
841
Willy Tarreau42ed14b2021-05-06 15:55:14 +0200842Example:
Willy Tarreau4b103022021-02-12 17:59:10 +0100843
Willy Tarreau42ed14b2021-05-06 15:55:14 +0200844 .if defined(HAPROXY_MWORKER)
845 listen mwcli_px
846 bind :1111
847 ...
848 .endif
Willy Tarreau4b103022021-02-12 17:59:10 +0100849
Willy Tarreau6492e872021-05-06 16:10:09 +0200850 .if strneq("$SSL_ONLY",yes)
851 bind :80
852 .endif
853
854 .if streq("$WITH_SSL",yes)
Willy Tarreau58ca7062021-05-06 16:34:23 +0200855 .if feature(OPENSSL)
Willy Tarreau6492e872021-05-06 16:10:09 +0200856 bind :443 ssl crt ...
Willy Tarreau58ca7062021-05-06 16:34:23 +0200857 .endif
Willy Tarreau6492e872021-05-06 16:10:09 +0200858 .endif
859
Willy Tarreau0b7c78a2021-05-06 16:53:26 +0200860 .if version_atleast(2.4-dev19)
861 profiling.memory on
862 .endif
863
Willy Tarreau7190b982021-05-07 08:59:50 +0200864Four other directives are provided to report some status:
Willy Tarreau4b103022021-02-12 17:59:10 +0100865
Willy Tarreau7190b982021-05-07 08:59:50 +0200866 - .diag "message" : emit this message only when in diagnostic mode (-dD)
Willy Tarreau4b103022021-02-12 17:59:10 +0100867 - .notice "message" : emit this message at level NOTICE
868 - .warning "message" : emit this message at level WARNING
869 - .alert "message" : emit this message at level ALERT
870
871Messages emitted at level WARNING may cause the process to fail to start if the
872"strict-mode" is enabled. Messages emitted at level ALERT will always cause a
873fatal error. These can be used to detect some inappropriate conditions and
874provide advice to the user.
875
876Example:
877
878 .if "${A}"
879 .if "${B}"
880 .notice "A=1, B=1"
881 .elif "${C}"
882 .notice "A=1, B=0, C=1"
883 .elif "${D}"
884 .warning "A=1, B=0, C=0, D=1"
885 .else
886 .alert "A=1, B=0, C=0, D=0"
887 .endif
888 .else
889 .notice "A=0"
890 .endif
891
Willy Tarreau7190b982021-05-07 08:59:50 +0200892 .diag "WTA/2021-05-07: replace 'redirect' with 'return' after switch to 2.4"
893 http-request redirect location /goaway if ABUSE
894
Willy Tarreau4b103022021-02-12 17:59:10 +0100895
8962.5. Time format
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200897----------------
898
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +0100899Some parameters involve values representing time, such as timeouts. These
Willy Tarreau0ba27502007-12-24 16:55:16 +0100900values are generally expressed in milliseconds (unless explicitly stated
901otherwise) but may be expressed in any other unit by suffixing the unit to the
902numeric value. It is important to consider this because it will not be repeated
903for every keyword. Supported units are :
904
905 - us : microseconds. 1 microsecond = 1/1000000 second
906 - ms : milliseconds. 1 millisecond = 1/1000 second. This is the default.
907 - s : seconds. 1s = 1000ms
908 - m : minutes. 1m = 60s = 60000ms
909 - h : hours. 1h = 60m = 3600s = 3600000ms
910 - d : days. 1d = 24h = 1440m = 86400s = 86400000ms
911
912
Willy Tarreau4b103022021-02-12 17:59:10 +01009132.6. Examples
Patrick Mezard35da19c2010-06-12 17:02:47 +0200914-------------
915
916 # Simple configuration for an HTTP proxy listening on port 80 on all
917 # interfaces and forwarding requests to a single backend "servers" with a
918 # single server "server1" listening on 127.0.0.1:8000
919 global
920 daemon
921 maxconn 256
922
923 defaults
924 mode http
925 timeout connect 5000ms
926 timeout client 50000ms
927 timeout server 50000ms
928
929 frontend http-in
930 bind *:80
931 default_backend servers
932
933 backend servers
934 server server1 127.0.0.1:8000 maxconn 32
935
936
937 # The same configuration defined with a single listen block. Shorter but
938 # less expressive, especially in HTTP mode.
939 global
940 daemon
941 maxconn 256
942
943 defaults
944 mode http
945 timeout connect 5000ms
946 timeout client 50000ms
947 timeout server 50000ms
948
949 listen http-in
950 bind *:80
951 server server1 127.0.0.1:8000 maxconn 32
952
953
954Assuming haproxy is in $PATH, test these configurations in a shell with:
955
Willy Tarreauccb289d2010-12-11 20:19:38 +0100956 $ sudo haproxy -f configuration.conf -c
Patrick Mezard35da19c2010-06-12 17:02:47 +0200957
958
Willy Tarreauc57f0e22009-05-10 13:12:33 +02009593. Global parameters
Willy Tarreau6a06a402007-07-15 20:15:28 +0200960--------------------
961
962Parameters in the "global" section are process-wide and often OS-specific. They
963are generally set once for all and do not need being changed once correct. Some
964of them have command-line equivalents.
965
966The following keywords are supported in the "global" section :
967
968 * Process management and security
Emeric Brunc8e8d122012-10-02 18:42:10 +0200969 - ca-base
Willy Tarreau6a06a402007-07-15 20:15:28 +0200970 - chroot
Emeric Brunc8e8d122012-10-02 18:42:10 +0200971 - crt-base
Baptiste Assmann3493d0f2015-10-12 20:21:23 +0200972 - cpu-map
Willy Tarreau6a06a402007-07-15 20:15:28 +0200973 - daemon
Willy Tarreau8a022d52021-04-27 20:29:11 +0200974 - default-path
Baptiste Assmann3493d0f2015-10-12 20:21:23 +0200975 - description
976 - deviceatlas-json-file
977 - deviceatlas-log-level
978 - deviceatlas-separator
979 - deviceatlas-properties-cookie
Amaury Denoyelled2e53cd2021-05-06 16:21:39 +0200980 - expose-experimental-directives
Simon Horman98637e52014-06-20 12:30:16 +0900981 - external-check
Willy Tarreau6a06a402007-07-15 20:15:28 +0200982 - gid
983 - group
Cyril Bonté203ec5a2017-03-23 22:44:13 +0100984 - hard-stop-after
Christopher Faulet98fbe952019-07-22 16:18:24 +0200985 - h1-case-adjust
986 - h1-case-adjust-file
Willy Tarreaud96f1122019-12-03 07:07:36 +0100987 - insecure-fork-wanted
Willy Tarreaua45a8b52019-12-06 16:31:45 +0100988 - insecure-setuid-wanted
Emmanuel Hocdet70df7bf2019-01-04 11:08:20 +0100989 - issuers-chain-path
Amaury Denoyelle0ea2c4f2021-07-09 17:14:30 +0200990 - h2-workaround-bogus-websocket-clients
Dragan Dosen13cd54c2020-06-18 18:24:05 +0200991 - localpeer
Willy Tarreau6a06a402007-07-15 20:15:28 +0200992 - log
Baptiste Assmann3493d0f2015-10-12 20:21:23 +0200993 - log-tag
Joe Williamsdf5b38f2010-12-29 17:05:48 +0100994 - log-send-hostname
Baptiste Assmann3493d0f2015-10-12 20:21:23 +0200995 - lua-load
Thierry Fournier59f11be2020-11-29 00:37:41 +0100996 - lua-load-per-thread
Tim Duesterhusdd74b5f2020-01-12 13:55:40 +0100997 - lua-prepend-path
William Lallemand27edc4b2019-05-07 17:49:33 +0200998 - mworker-max-reloads
Willy Tarreau6a06a402007-07-15 20:15:28 +0200999 - nbproc
Christopher Fauletbe0faa22017-08-29 15:37:10 +02001000 - nbthread
Baptiste Assmann3493d0f2015-10-12 20:21:23 +02001001 - node
Amaury Denoyelle0f50cb92021-03-26 18:50:33 +01001002 - numa-cpu-mapping
Willy Tarreau6a06a402007-07-15 20:15:28 +02001003 - pidfile
Willy Tarreau119e50e2020-05-22 13:53:29 +02001004 - pp2-never-send-local
Willy Tarreau1d549722016-02-16 12:41:57 +01001005 - presetenv
1006 - resetenv
Willy Tarreau6a06a402007-07-15 20:15:28 +02001007 - uid
1008 - ulimit-n
1009 - user
Willy Tarreau636848a2019-04-15 19:38:50 +02001010 - set-dumpable
Willy Tarreau13d2ba22021-03-26 11:38:08 +01001011 - set-var
Willy Tarreau1d549722016-02-16 12:41:57 +01001012 - setenv
Willy Tarreaufbee7132007-10-18 13:53:22 +02001013 - stats
Baptiste Assmann3493d0f2015-10-12 20:21:23 +02001014 - ssl-default-bind-ciphers
Dirkjan Bussink415150f2018-09-14 11:14:21 +02001015 - ssl-default-bind-ciphersuites
Jerome Magninb203ff62020-04-03 15:28:22 +02001016 - ssl-default-bind-curves
Baptiste Assmann3493d0f2015-10-12 20:21:23 +02001017 - ssl-default-bind-options
1018 - ssl-default-server-ciphers
Dirkjan Bussink415150f2018-09-14 11:14:21 +02001019 - ssl-default-server-ciphersuites
Baptiste Assmann3493d0f2015-10-12 20:21:23 +02001020 - ssl-default-server-options
1021 - ssl-dh-param-file
Emeric Brun850efd52014-01-29 12:24:34 +01001022 - ssl-server-verify
Emmanuel Hocdetc3b7e742020-04-22 11:06:19 +02001023 - ssl-skip-self-issued-ca
Willy Tarreauceb24bc2010-11-09 12:46:41 +01001024 - unix-bind
Willy Tarreau1d549722016-02-16 12:41:57 +01001025 - unsetenv
Thomas Holmesdb04f192015-05-18 13:21:39 +01001026 - 51degrees-data-file
1027 - 51degrees-property-name-list
Dragan Dosen93b38d92015-06-29 16:43:25 +02001028 - 51degrees-property-separator
Dragan Dosenae6d39a2015-06-29 16:43:27 +02001029 - 51degrees-cache-size
Willy Tarreaub3cc9f22019-04-19 16:03:32 +02001030 - wurfl-data-file
1031 - wurfl-information-list
1032 - wurfl-information-list-separator
Willy Tarreaub3cc9f22019-04-19 16:03:32 +02001033 - wurfl-cache-size
William Dauchy0fec3ab2019-10-27 20:08:11 +01001034 - strict-limits
Willy Tarreaud72758d2010-01-12 10:42:19 +01001035
Willy Tarreau6a06a402007-07-15 20:15:28 +02001036 * Performance tuning
William Dauchy0a8824f2019-10-27 20:08:09 +01001037 - busy-polling
Willy Tarreau1746eec2014-04-25 10:46:47 +02001038 - max-spread-checks
Willy Tarreau6a06a402007-07-15 20:15:28 +02001039 - maxconn
Willy Tarreau81c25d02011-09-07 15:17:21 +02001040 - maxconnrate
William Lallemandd85f9172012-11-09 17:05:39 +01001041 - maxcomprate
William Lallemand072a2bf2012-11-20 17:01:01 +01001042 - maxcompcpuusage
Willy Tarreauff4f82d2009-02-06 11:28:13 +01001043 - maxpipes
Willy Tarreau93e7c002013-10-07 18:51:07 +02001044 - maxsessrate
Willy Tarreau403edff2012-09-06 11:58:37 +02001045 - maxsslconn
Willy Tarreaue43d5322013-10-07 20:01:52 +02001046 - maxsslrate
Baptiste Assmann3493d0f2015-10-12 20:21:23 +02001047 - maxzlibmem
Willy Tarreau6a06a402007-07-15 20:15:28 +02001048 - noepoll
1049 - nokqueue
Emmanuel Hocdet0ba4f482019-04-08 16:53:32 +00001050 - noevports
Willy Tarreau6a06a402007-07-15 20:15:28 +02001051 - nopoll
Willy Tarreauff4f82d2009-02-06 11:28:13 +01001052 - nosplice
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03001053 - nogetaddrinfo
Lukas Tribusa0bcbdc2016-09-12 21:42:20 +00001054 - noreuseport
Willy Tarreau75c62c22018-11-22 11:02:09 +01001055 - profiling.tasks
Willy Tarreaufe255b72007-10-14 23:09:26 +02001056 - spread-checks
Baptiste Assmann5626f482015-08-23 10:00:10 +02001057 - server-state-base
Baptiste Assmannef1f0fc2015-08-23 10:06:39 +02001058 - server-state-file
Grant Zhang872f9c22017-01-21 01:10:18 +00001059 - ssl-engine
Grant Zhangfa6c7ee2017-01-14 01:42:15 +00001060 - ssl-mode-async
Baptiste Assmann3493d0f2015-10-12 20:21:23 +02001061 - tune.buffers.limit
1062 - tune.buffers.reserve
Willy Tarreau27a674e2009-08-17 07:23:33 +02001063 - tune.bufsize
Willy Tarreau43961d52010-10-04 20:39:20 +02001064 - tune.chksize
William Lallemandf3747832012-11-09 12:33:10 +01001065 - tune.comp.maxlevel
Willy Tarreaubc52bec2020-06-18 08:58:47 +02001066 - tune.fd.edge-triggered
Willy Tarreaufe20e5b2017-07-27 11:42:14 +02001067 - tune.h2.header-table-size
Willy Tarreaue6baec02017-07-27 11:45:11 +02001068 - tune.h2.initial-window-size
Willy Tarreau5242ef82017-07-27 11:47:28 +02001069 - tune.h2.max-concurrent-streams
Willy Tarreau193b8c62012-11-22 00:17:38 +01001070 - tune.http.cookielen
Stéphane Cottin23e9e932017-05-18 08:58:41 +02001071 - tune.http.logurilen
Willy Tarreauac1932d2011-10-24 19:14:41 +02001072 - tune.http.maxhdr
Willy Tarreau76cc6992020-07-01 18:49:24 +02001073 - tune.idle-pool.shared
Willy Tarreau7e312732014-02-12 16:35:14 +01001074 - tune.idletimer
Thierry FOURNIER90da1912015-03-05 11:17:06 +01001075 - tune.lua.forced-yield
Willy Tarreau32f61e22015-03-18 17:54:59 +01001076 - tune.lua.maxmem
Thierry FOURNIER90da1912015-03-05 11:17:06 +01001077 - tune.lua.session-timeout
1078 - tune.lua.task-timeout
Thierry FOURNIER7dd784b2015-10-01 14:49:33 +02001079 - tune.lua.service-timeout
Willy Tarreaua0250ba2008-01-06 11:22:57 +01001080 - tune.maxaccept
1081 - tune.maxpollevents
Willy Tarreau27a674e2009-08-17 07:23:33 +02001082 - tune.maxrewrite
Willy Tarreauf3045d22015-04-29 16:24:50 +02001083 - tune.pattern.cache-size
Willy Tarreaubd9a0a72011-10-23 21:14:29 +02001084 - tune.pipesize
Willy Tarreaua8e2d972020-07-01 18:27:16 +02001085 - tune.pool-high-fd-ratio
1086 - tune.pool-low-fd-ratio
Willy Tarreaue803de22010-01-21 17:43:04 +01001087 - tune.rcvbuf.client
1088 - tune.rcvbuf.server
Willy Tarreaub22fc302015-12-14 12:04:35 +01001089 - tune.recv_enough
Olivier Houchard1599b802018-05-24 18:59:04 +02001090 - tune.runqueue-depth
Willy Tarreaue7723bd2020-06-24 11:11:02 +02001091 - tune.sched.low-latency
Willy Tarreaue803de22010-01-21 17:43:04 +01001092 - tune.sndbuf.client
1093 - tune.sndbuf.server
Willy Tarreau6ec58db2012-11-16 16:32:15 +01001094 - tune.ssl.cachesize
William Lallemand7d42ef52020-07-06 11:41:30 +02001095 - tune.ssl.keylog
Willy Tarreaubfd59462013-02-21 07:46:09 +01001096 - tune.ssl.lifetime
Emeric Brun8dc60392014-05-09 13:52:00 +02001097 - tune.ssl.force-private-cache
Willy Tarreaubfd59462013-02-21 07:46:09 +01001098 - tune.ssl.maxrecord
Remi Gacognef46cd6e2014-06-12 14:58:40 +02001099 - tune.ssl.default-dh-param
Christopher Faulet31af49d2015-06-09 17:29:50 +02001100 - tune.ssl.ssl-ctx-cache-size
Thierry FOURNIER5bf77322017-02-25 12:45:22 +01001101 - tune.ssl.capture-cipherlist-size
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02001102 - tune.vars.global-max-size
Christopher Fauletff2613e2016-11-09 11:36:17 +01001103 - tune.vars.proc-max-size
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02001104 - tune.vars.reqres-max-size
1105 - tune.vars.sess-max-size
1106 - tune.vars.txn-max-size
William Lallemanda509e4c2012-11-07 16:54:34 +01001107 - tune.zlib.memlevel
1108 - tune.zlib.windowsize
Willy Tarreaud72758d2010-01-12 10:42:19 +01001109
Willy Tarreau6a06a402007-07-15 20:15:28 +02001110 * Debugging
Willy Tarreau6a06a402007-07-15 20:15:28 +02001111 - quiet
Willy Tarreau3eb10b82020-04-15 16:42:39 +02001112 - zero-warning
Willy Tarreau6a06a402007-07-15 20:15:28 +02001113
1114
Willy Tarreauc57f0e22009-05-10 13:12:33 +020011153.1. Process management and security
Willy Tarreau6a06a402007-07-15 20:15:28 +02001116------------------------------------
1117
Emeric Brunc8e8d122012-10-02 18:42:10 +02001118ca-base <dir>
1119 Assigns a default directory to fetch SSL CA certificates and CRLs from when a
Emmanuel Hocdet842e94e2019-12-16 16:39:17 +01001120 relative path is used with "ca-file", "ca-verify-file" or "crl-file"
1121 directives. Absolute locations specified in "ca-file", "ca-verify-file" and
1122 "crl-file" prevail and ignore "ca-base".
Emeric Brunc8e8d122012-10-02 18:42:10 +02001123
Willy Tarreau6a06a402007-07-15 20:15:28 +02001124chroot <jail dir>
1125 Changes current directory to <jail dir> and performs a chroot() there before
1126 dropping privileges. This increases the security level in case an unknown
1127 vulnerability would be exploited, since it would make it very hard for the
1128 attacker to exploit the system. This only works when the process is started
1129 with superuser privileges. It is important to ensure that <jail_dir> is both
Davor Ocelice9ed2812017-12-25 17:49:28 +01001130 empty and non-writable to anyone.
Willy Tarreaud72758d2010-01-12 10:42:19 +01001131
Christopher Fauletcb6a9452017-11-22 16:50:41 +01001132cpu-map [auto:]<process-set>[/<thread-set>] <cpu-set>...
1133 On Linux 2.6 and above, it is possible to bind a process or a thread to a
1134 specific CPU set. This means that the process or the thread will never run on
1135 other CPUs. The "cpu-map" directive specifies CPU sets for process or thread
1136 sets. The first argument is a process set, eventually followed by a thread
1137 set. These sets have the format
1138
1139 all | odd | even | number[-[number]]
1140
1141 <number>> must be a number between 1 and 32 or 64, depending on the machine's
Davor Ocelice9ed2812017-12-25 17:49:28 +01001142 word size. Any process IDs above nbproc and any thread IDs above nbthread are
Christopher Fauletcb6a9452017-11-22 16:50:41 +01001143 ignored. It is possible to specify a range with two such number delimited by
1144 a dash ('-'). It also is possible to specify all processes at once using
Christopher Faulet1dcb9cb2017-11-22 10:24:40 +01001145 "all", only odd numbers using "odd" or even numbers using "even", just like
1146 with the "bind-process" directive. The second and forthcoming arguments are
Amaury Denoyelle982fb532021-04-21 18:39:58 +02001147 CPU sets. Each CPU set is either a unique number starting at 0 for the first
1148 CPU or a range with two such numbers delimited by a dash ('-'). Outside of
1149 Linux and BSDs, there may be a limitation on the maximum CPU index to either
1150 31 or 63. Multiple CPU numbers or ranges may be specified, and the processes
1151 or threads will be allowed to bind to all of them. Obviously, multiple
1152 "cpu-map" directives may be specified. Each "cpu-map" directive will replace
1153 the previous ones when they overlap. A thread will be bound on the
1154 intersection of its mapping and the one of the process on which it is
1155 attached. If the intersection is null, no specific binding will be set for
1156 the thread.
Willy Tarreaufc6c0322012-11-16 16:12:27 +01001157
Christopher Fauletff4121f2017-11-22 16:38:49 +01001158 Ranges can be partially defined. The higher bound can be omitted. In such
1159 case, it is replaced by the corresponding maximum value, 32 or 64 depending
1160 on the machine's word size.
1161
Christopher Faulet26028f62017-11-22 15:01:51 +01001162 The prefix "auto:" can be added before the process set to let HAProxy
Christopher Fauletcb6a9452017-11-22 16:50:41 +01001163 automatically bind a process or a thread to a CPU by incrementing
1164 process/thread and CPU sets. To be valid, both sets must have the same
1165 size. No matter the declaration order of the CPU sets, it will be bound from
1166 the lowest to the highest bound. Having a process and a thread range with the
1167 "auto:" prefix is not supported. Only one range is supported, the other one
1168 must be a fixed number.
Christopher Faulet26028f62017-11-22 15:01:51 +01001169
1170 Examples:
Christopher Fauletcb6a9452017-11-22 16:50:41 +01001171 cpu-map 1-4 0-3 # bind processes 1 to 4 on the first 4 CPUs
1172
1173 cpu-map 1/all 0-3 # bind all threads of the first process on the
1174 # first 4 CPUs
1175
1176 cpu-map 1- 0- # will be replaced by "cpu-map 1-64 0-63"
1177 # or "cpu-map 1-32 0-31" depending on the machine's
1178 # word size.
1179
Christopher Faulet26028f62017-11-22 15:01:51 +01001180 # all these lines bind the process 1 to the cpu 0, the process 2 to cpu 1
Christopher Fauletcb6a9452017-11-22 16:50:41 +01001181 # and so on.
Christopher Faulet26028f62017-11-22 15:01:51 +01001182 cpu-map auto:1-4 0-3
1183 cpu-map auto:1-4 0-1 2-3
1184 cpu-map auto:1-4 3 2 1 0
1185
Christopher Fauletcb6a9452017-11-22 16:50:41 +01001186 # all these lines bind the thread 1 to the cpu 0, the thread 2 to cpu 1
1187 # and so on.
1188 cpu-map auto:1/1-4 0-3
1189 cpu-map auto:1/1-4 0-1 2-3
1190 cpu-map auto:1/1-4 3 2 1 0
1191
Davor Ocelice9ed2812017-12-25 17:49:28 +01001192 # bind each process to exactly one CPU using all/odd/even keyword
Christopher Faulet26028f62017-11-22 15:01:51 +01001193 cpu-map auto:all 0-63
1194 cpu-map auto:even 0-31
1195 cpu-map auto:odd 32-63
1196
1197 # invalid cpu-map because process and CPU sets have different sizes.
1198 cpu-map auto:1-4 0 # invalid
1199 cpu-map auto:1 0-3 # invalid
1200
Christopher Fauletcb6a9452017-11-22 16:50:41 +01001201 # invalid cpu-map because automatic binding is used with a process range
1202 # and a thread range.
1203 cpu-map auto:all/all 0 # invalid
1204 cpu-map auto:all/1-4 0 # invalid
1205 cpu-map auto:1-4/all 0 # invalid
1206
Emeric Brunc8e8d122012-10-02 18:42:10 +02001207crt-base <dir>
1208 Assigns a default directory to fetch SSL certificates from when a relative
William Dauchy238ea3b2020-01-11 13:09:12 +01001209 path is used with "crtfile" or "crt" directives. Absolute locations specified
1210 prevail and ignore "crt-base".
Emeric Brunc8e8d122012-10-02 18:42:10 +02001211
Willy Tarreau6a06a402007-07-15 20:15:28 +02001212daemon
1213 Makes the process fork into background. This is the recommended mode of
1214 operation. It is equivalent to the command line "-D" argument. It can be
Lukas Tribusf46bf952017-11-21 12:39:34 +01001215 disabled by the command line "-db" argument. This option is ignored in
1216 systemd mode.
Willy Tarreau6a06a402007-07-15 20:15:28 +02001217
Willy Tarreau8a022d52021-04-27 20:29:11 +02001218default-path { current | config | parent | origin <path> }
Daniel Corbett9f0843f2021-05-08 10:50:37 -04001219 By default HAProxy loads all files designated by a relative path from the
Willy Tarreau8a022d52021-04-27 20:29:11 +02001220 location the process is started in. In some circumstances it might be
1221 desirable to force all relative paths to start from a different location
1222 just as if the process was started from such locations. This is what this
1223 directive is made for. Technically it will perform a temporary chdir() to
1224 the designated location while processing each configuration file, and will
1225 return to the original directory after processing each file. It takes an
1226 argument indicating the policy to use when loading files whose path does
1227 not start with a slash ('/'):
1228 - "current" indicates that all relative files are to be loaded from the
1229 directory the process is started in ; this is the default.
1230
1231 - "config" indicates that all relative files should be loaded from the
1232 directory containing the configuration file. More specifically, if the
1233 configuration file contains a slash ('/'), the longest part up to the
1234 last slash is used as the directory to change to, otherwise the current
1235 directory is used. This mode is convenient to bundle maps, errorfiles,
1236 certificates and Lua scripts together as relocatable packages. When
1237 multiple configuration files are loaded, the directory is updated for
1238 each of them.
1239
1240 - "parent" indicates that all relative files should be loaded from the
1241 parent of the directory containing the configuration file. More
1242 specifically, if the configuration file contains a slash ('/'), ".."
1243 is appended to the longest part up to the last slash is used as the
1244 directory to change to, otherwise the directory is "..". This mode is
1245 convenient to bundle maps, errorfiles, certificates and Lua scripts
1246 together as relocatable packages, but where each part is located in a
1247 different subdirectory (e.g. "config/", "certs/", "maps/", ...).
1248
1249 - "origin" indicates that all relative files should be loaded from the
1250 designated (mandatory) path. This may be used to ease management of
Daniel Corbett9f0843f2021-05-08 10:50:37 -04001251 different HAProxy instances running in parallel on a system, where each
Willy Tarreau8a022d52021-04-27 20:29:11 +02001252 instance uses a different prefix but where the rest of the sections are
1253 made easily relocatable.
1254
1255 Each "default-path" directive instantly replaces any previous one and will
1256 possibly result in switching to a different directory. While this should
1257 always result in the desired behavior, it is really not a good practice to
1258 use multiple default-path directives, and if used, the policy ought to remain
1259 consistent across all configuration files.
1260
1261 Warning: some configuration elements such as maps or certificates are
1262 uniquely identified by their configured path. By using a relocatable layout,
1263 it becomes possible for several of them to end up with the same unique name,
1264 making it difficult to update them at run time, especially when multiple
1265 configuration files are loaded from different directories. It is essential to
1266 observe a strict collision-free file naming scheme before adopting relative
1267 paths. A robust approach could consist in prefixing all files names with
1268 their respective site name, or in doing so at the directory level.
1269
David Carlier8167f302015-06-01 13:50:06 +02001270deviceatlas-json-file <path>
1271 Sets the path of the DeviceAtlas JSON data file to be loaded by the API.
Davor Ocelice9ed2812017-12-25 17:49:28 +01001272 The path must be a valid JSON data file and accessible by HAProxy process.
David Carlier8167f302015-06-01 13:50:06 +02001273
1274deviceatlas-log-level <value>
Davor Ocelice9ed2812017-12-25 17:49:28 +01001275 Sets the level of information returned by the API. This directive is
David Carlier8167f302015-06-01 13:50:06 +02001276 optional and set to 0 by default if not set.
1277
1278deviceatlas-separator <char>
1279 Sets the character separator for the API properties results. This directive
1280 is optional and set to | by default if not set.
1281
Cyril Bonté0306c4a2015-10-26 22:37:38 +01001282deviceatlas-properties-cookie <name>
Cyril Bonté307ee1e2015-09-28 23:16:06 +02001283 Sets the client cookie's name used for the detection if the DeviceAtlas
1284 Client-side component was used during the request. This directive is optional
1285 and set to DAPROPS by default if not set.
David Carlier29b3ca32015-09-25 14:09:21 +01001286
Amaury Denoyelled2e53cd2021-05-06 16:21:39 +02001287expose-experimental-directives
1288 This statement must appear before using directives tagged as experimental or
1289 the config file will be rejected.
1290
Simon Horman98637e52014-06-20 12:30:16 +09001291external-check
Willy Tarreaud96f1122019-12-03 07:07:36 +01001292 Allows the use of an external agent to perform health checks. This is
1293 disabled by default as a security precaution, and even when enabled, checks
Willy Tarreaua45a8b52019-12-06 16:31:45 +01001294 may still fail unless "insecure-fork-wanted" is enabled as well. If the
1295 program launched makes use of a setuid executable (it should really not),
1296 you may also need to set "insecure-setuid-wanted" in the global section.
1297 See "option external-check", and "insecure-fork-wanted", and
1298 "insecure-setuid-wanted".
Simon Horman98637e52014-06-20 12:30:16 +09001299
Willy Tarreau6a06a402007-07-15 20:15:28 +02001300gid <number>
Thayne McCombscdbcca92021-01-07 21:24:41 -07001301 Changes the process's group ID to <number>. It is recommended that the group
Willy Tarreau6a06a402007-07-15 20:15:28 +02001302 ID is dedicated to HAProxy or to a small set of similar daemons. HAProxy must
1303 be started with a user belonging to this group, or with superuser privileges.
Daniel Corbett9f0843f2021-05-08 10:50:37 -04001304 Note that if HAProxy is started from a user having supplementary groups, it
Michael Schererab012dd2013-01-12 18:35:19 +01001305 will only be able to drop these groups if started with superuser privileges.
Willy Tarreau6a06a402007-07-15 20:15:28 +02001306 See also "group" and "uid".
Willy Tarreaud72758d2010-01-12 10:42:19 +01001307
Willy Tarreau11770ce2019-12-03 08:29:22 +01001308group <group name>
1309 Similar to "gid" but uses the GID of group name <group name> from /etc/group.
1310 See also "gid" and "user".
1311
Cyril Bonté203ec5a2017-03-23 22:44:13 +01001312hard-stop-after <time>
1313 Defines the maximum time allowed to perform a clean soft-stop.
1314
1315 Arguments :
1316 <time> is the maximum time (by default in milliseconds) for which the
1317 instance will remain alive when a soft-stop is received via the
1318 SIGUSR1 signal.
1319
1320 This may be used to ensure that the instance will quit even if connections
1321 remain opened during a soft-stop (for example with long timeouts for a proxy
1322 in tcp mode). It applies both in TCP and HTTP mode.
1323
1324 Example:
1325 global
1326 hard-stop-after 30s
1327
Christopher Faulet98fbe952019-07-22 16:18:24 +02001328h1-case-adjust <from> <to>
1329 Defines the case adjustment to apply, when enabled, to the header name
1330 <from>, to change it to <to> before sending it to HTTP/1 clients or
1331 servers. <from> must be in lower case, and <from> and <to> must not differ
1332 except for their case. It may be repeated if several header names need to be
Ilya Shipitsin8525fd92020-02-29 12:34:59 +05001333 adjusted. Duplicate entries are not allowed. If a lot of header names have to
Christopher Faulet98fbe952019-07-22 16:18:24 +02001334 be adjusted, it might be more convenient to use "h1-case-adjust-file".
1335 Please note that no transformation will be applied unless "option
1336 h1-case-adjust-bogus-client" or "option h1-case-adjust-bogus-server" is
1337 specified in a proxy.
1338
1339 There is no standard case for header names because, as stated in RFC7230,
1340 they are case-insensitive. So applications must handle them in a case-
1341 insensitive manner. But some bogus applications violate the standards and
1342 erroneously rely on the cases most commonly used by browsers. This problem
1343 becomes critical with HTTP/2 because all header names must be exchanged in
1344 lower case, and HAProxy follows the same convention. All header names are
1345 sent in lower case to clients and servers, regardless of the HTTP version.
1346
1347 Applications which fail to properly process requests or responses may require
1348 to temporarily use such workarounds to adjust header names sent to them for
1349 the time it takes the application to be fixed. Please note that an
1350 application which requires such workarounds might be vulnerable to content
1351 smuggling attacks and must absolutely be fixed.
1352
1353 Example:
1354 global
1355 h1-case-adjust content-length Content-Length
1356
1357 See "h1-case-adjust-file", "option h1-case-adjust-bogus-client" and
1358 "option h1-case-adjust-bogus-server".
1359
1360h1-case-adjust-file <hdrs-file>
1361 Defines a file containing a list of key/value pairs used to adjust the case
1362 of some header names before sending them to HTTP/1 clients or servers. The
1363 file <hdrs-file> must contain 2 header names per line. The first one must be
1364 in lower case and both must not differ except for their case. Lines which
1365 start with '#' are ignored, just like empty lines. Leading and trailing tabs
1366 and spaces are stripped. Duplicate entries are not allowed. Please note that
1367 no transformation will be applied unless "option h1-case-adjust-bogus-client"
1368 or "option h1-case-adjust-bogus-server" is specified in a proxy.
1369
1370 If this directive is repeated, only the last one will be processed. It is an
1371 alternative to the directive "h1-case-adjust" if a lot of header names need
1372 to be adjusted. Please read the risks associated with using this.
1373
1374 See "h1-case-adjust", "option h1-case-adjust-bogus-client" and
1375 "option h1-case-adjust-bogus-server".
1376
Willy Tarreaud96f1122019-12-03 07:07:36 +01001377insecure-fork-wanted
Daniel Corbett9f0843f2021-05-08 10:50:37 -04001378 By default HAProxy tries hard to prevent any thread and process creation
Willy Tarreaud96f1122019-12-03 07:07:36 +01001379 after it starts. Doing so is particularly important when using Lua files of
1380 uncertain origin, and when experimenting with development versions which may
1381 still contain bugs whose exploitability is uncertain. And generally speaking
1382 it's good hygiene to make sure that no unexpected background activity can be
1383 triggered by traffic. But this prevents external checks from working, and may
1384 break some very specific Lua scripts which actively rely on the ability to
1385 fork. This option is there to disable this protection. Note that it is a bad
Daniel Corbett9f0843f2021-05-08 10:50:37 -04001386 idea to disable it, as a vulnerability in a library or within HAProxy itself
Willy Tarreaud96f1122019-12-03 07:07:36 +01001387 will be easier to exploit once disabled. In addition, forking from Lua or
1388 anywhere else is not reliable as the forked process may randomly embed a lock
1389 set by another thread and never manage to finish an operation. As such it is
1390 highly recommended that this option is never used and that any workload
1391 requiring such a fork be reconsidered and moved to a safer solution (such as
1392 agents instead of external checks). This option supports the "no" prefix to
1393 disable it.
1394
Willy Tarreaua45a8b52019-12-06 16:31:45 +01001395insecure-setuid-wanted
1396 HAProxy doesn't need to call executables at run time (except when using
1397 external checks which are strongly recommended against), and is even expected
1398 to isolate itself into an empty chroot. As such, there basically is no valid
1399 reason to allow a setuid executable to be called without the user being fully
Daniel Corbett9f0843f2021-05-08 10:50:37 -04001400 aware of the risks. In a situation where HAProxy would need to call external
Willy Tarreaua45a8b52019-12-06 16:31:45 +01001401 checks and/or disable chroot, exploiting a vulnerability in a library or in
Daniel Corbett9f0843f2021-05-08 10:50:37 -04001402 HAProxy itself could lead to the execution of an external program. On Linux
Willy Tarreaua45a8b52019-12-06 16:31:45 +01001403 it is possible to lock the process so that any setuid bit present on such an
1404 executable is ignored. This significantly reduces the risk of privilege
Daniel Corbett9f0843f2021-05-08 10:50:37 -04001405 escalation in such a situation. This is what HAProxy does by default. In case
Willy Tarreaua45a8b52019-12-06 16:31:45 +01001406 this causes a problem to an external check (for example one which would need
1407 the "ping" command), then it is possible to disable this protection by
1408 explicitly adding this directive in the global section. If enabled, it is
1409 possible to turn it back off by prefixing it with the "no" keyword.
1410
Emmanuel Hocdet70df7bf2019-01-04 11:08:20 +01001411issuers-chain-path <dir>
1412 Assigns a directory to load certificate chain for issuer completion. All
1413 files must be in PEM format. For certificates loaded with "crt" or "crt-list",
1414 if certificate chain is not included in PEM (also commonly known as
Daniel Corbett9f0843f2021-05-08 10:50:37 -04001415 intermediate certificate), HAProxy will complete chain if the issuer of the
Emmanuel Hocdet70df7bf2019-01-04 11:08:20 +01001416 certificate corresponds to the first certificate of the chain loaded with
1417 "issuers-chain-path".
1418 A "crt" file with PrivateKey+Certificate+IntermediateCA2+IntermediateCA1
1419 could be replaced with PrivateKey+Certificate. HAProxy will complete the
1420 chain if a file with IntermediateCA2+IntermediateCA1 is present in
1421 "issuers-chain-path" directory. All other certificates with the same issuer
1422 will share the chain in memory.
1423
Amaury Denoyelle0ea2c4f2021-07-09 17:14:30 +02001424h2-workaround-bogus-websocket-clients
1425 This disables the announcement of the support for h2 websockets to clients.
1426 This can be use to overcome clients which have issues when implementing the
1427 relatively fresh RFC8441, such as Firefox 88. To allow clients to
1428 automatically downgrade to http/1.1 for the websocket tunnel, specify h2
1429 support on the bind line using "alpn" without an explicit "proto" keyword. If
1430 this statement was previously activated, this can be disabled by prefixing
1431 the keyword with "no'.
1432
Dragan Dosen13cd54c2020-06-18 18:24:05 +02001433localpeer <name>
1434 Sets the local instance's peer name. It will be ignored if the "-L"
1435 command line argument is specified or if used after "peers" section
1436 definitions. In such cases, a warning message will be emitted during
1437 the configuration parsing.
1438
1439 This option will also set the HAPROXY_LOCALPEER environment variable.
1440 See also "-L" in the management guide and "peers" section below.
1441
Jan Wagner3e678602020-12-17 22:22:32 +01001442log <address> [len <length>] [format <format>] [sample <ranges>:<sample_size>]
Frédéric Lécailled690dfa2019-04-25 10:52:17 +02001443 <facility> [max level [min level]]
Cyril Bonté3e954872018-03-20 23:30:27 +01001444 Adds a global syslog server. Several global servers can be defined. They
Davor Ocelice9ed2812017-12-25 17:49:28 +01001445 will receive logs for starts and exits, as well as all logs from proxies
Robert Tsai81ae1952007-12-05 10:47:29 +01001446 configured with "log global".
1447
1448 <address> can be one of:
1449
Willy Tarreau2769aa02007-12-27 18:26:09 +01001450 - An IPv4 address optionally followed by a colon and a UDP port. If
Robert Tsai81ae1952007-12-05 10:47:29 +01001451 no port is specified, 514 is used by default (the standard syslog
1452 port).
1453
David du Colombier24bb5f52011-03-17 10:40:23 +01001454 - An IPv6 address followed by a colon and optionally a UDP port. If
1455 no port is specified, 514 is used by default (the standard syslog
1456 port).
1457
Willy Tarreau5a32ecc2018-11-12 07:34:59 +01001458 - A filesystem path to a datagram UNIX domain socket, keeping in mind
Robert Tsai81ae1952007-12-05 10:47:29 +01001459 considerations for chroot (be sure the path is accessible inside
1460 the chroot) and uid/gid (be sure the path is appropriately
Davor Ocelice9ed2812017-12-25 17:49:28 +01001461 writable).
Robert Tsai81ae1952007-12-05 10:47:29 +01001462
Willy Tarreau5a32ecc2018-11-12 07:34:59 +01001463 - A file descriptor number in the form "fd@<number>", which may point
1464 to a pipe, terminal, or socket. In this case unbuffered logs are used
1465 and one writev() call per log is performed. This is a bit expensive
1466 but acceptable for most workloads. Messages sent this way will not be
1467 truncated but may be dropped, in which case the DroppedLogs counter
1468 will be incremented. The writev() call is atomic even on pipes for
1469 messages up to PIPE_BUF size, which POSIX recommends to be at least
1470 512 and which is 4096 bytes on most modern operating systems. Any
1471 larger message may be interleaved with messages from other processes.
1472 Exceptionally for debugging purposes the file descriptor may also be
Daniel Corbett9f0843f2021-05-08 10:50:37 -04001473 directed to a file, but doing so will significantly slow HAProxy down
Willy Tarreau5a32ecc2018-11-12 07:34:59 +01001474 as non-blocking calls will be ignored. Also there will be no way to
1475 purge nor rotate this file without restarting the process. Note that
1476 the configured syslog format is preserved, so the output is suitable
Willy Tarreauc1b06452018-11-12 11:57:56 +01001477 for use with a TCP syslog server. See also the "short" and "raw"
1478 format below.
Willy Tarreau5a32ecc2018-11-12 07:34:59 +01001479
1480 - "stdout" / "stderr", which are respectively aliases for "fd@1" and
1481 "fd@2", see above.
1482
Willy Tarreauc046d162019-08-30 15:24:59 +02001483 - A ring buffer in the form "ring@<name>", which will correspond to an
1484 in-memory ring buffer accessible over the CLI using the "show events"
1485 command, which will also list existing rings and their sizes. Such
1486 buffers are lost on reload or restart but when used as a complement
1487 this can help troubleshooting by having the logs instantly available.
1488
William Lallemandb2f07452015-05-12 14:27:13 +02001489 You may want to reference some environment variables in the address
1490 parameter, see section 2.3 about environment variables.
Willy Tarreaudad36a32013-03-11 01:20:04 +01001491
Willy Tarreau18324f52014-06-27 18:10:07 +02001492 <length> is an optional maximum line length. Log lines larger than this value
1493 will be truncated before being sent. The reason is that syslog
1494 servers act differently on log line length. All servers support the
1495 default value of 1024, but some servers simply drop larger lines
1496 while others do log them. If a server supports long lines, it may
1497 make sense to set this value here in order to avoid truncating long
1498 lines. Similarly, if a server drops long lines, it is preferable to
1499 truncate them before sending them. Accepted values are 80 to 65535
1500 inclusive. The default value of 1024 is generally fine for all
1501 standard usages. Some specific cases of long captures or
Davor Ocelice9ed2812017-12-25 17:49:28 +01001502 JSON-formatted logs may require larger values. You may also need to
1503 increase "tune.http.logurilen" if your request URIs are truncated.
Willy Tarreau18324f52014-06-27 18:10:07 +02001504
Dragan Dosen7ad31542015-09-28 17:16:47 +02001505 <format> is the log format used when generating syslog messages. It may be
1506 one of the following :
1507
Emeric Brun0237c4e2020-11-27 16:24:34 +01001508 local Analog to rfc3164 syslog message format except that hostname
1509 field is stripped. This is the default.
1510 Note: option "log-send-hostname" switches the default to
1511 rfc3164.
1512
1513 rfc3164 The RFC3164 syslog message format.
Dragan Dosen7ad31542015-09-28 17:16:47 +02001514 (https://tools.ietf.org/html/rfc3164)
1515
1516 rfc5424 The RFC5424 syslog message format.
1517 (https://tools.ietf.org/html/rfc5424)
1518
Emeric Brun54648852020-07-06 15:54:06 +02001519 priority A message containing only a level plus syslog facility between
1520 angle brackets such as '<63>', followed by the text. The PID,
1521 date, time, process name and system name are omitted. This is
1522 designed to be used with a local log server.
1523
Willy Tarreaue8746a02018-11-12 08:45:00 +01001524 short A message containing only a level between angle brackets such as
1525 '<3>', followed by the text. The PID, date, time, process name
1526 and system name are omitted. This is designed to be used with a
1527 local log server. This format is compatible with what the systemd
1528 logger consumes.
1529
Emeric Brun54648852020-07-06 15:54:06 +02001530 timed A message containing only a level between angle brackets such as
1531 '<3>', followed by ISO date and by the text. The PID, process
1532 name and system name are omitted. This is designed to be
1533 used with a local log server.
1534
1535 iso A message containing only the ISO date, followed by the text.
1536 The PID, process name and system name are omitted. This is
1537 designed to be used with a local log server.
1538
Willy Tarreauc1b06452018-11-12 11:57:56 +01001539 raw A message containing only the text. The level, PID, date, time,
1540 process name and system name are omitted. This is designed to be
1541 used in containers or during development, where the severity only
1542 depends on the file descriptor used (stdout/stderr).
1543
Frédéric Lécailled690dfa2019-04-25 10:52:17 +02001544 <ranges> A list of comma-separated ranges to identify the logs to sample.
1545 This is used to balance the load of the logs to send to the log
1546 server. The limits of the ranges cannot be null. They are numbered
1547 from 1. The size or period (in number of logs) of the sample must be
1548 set with <sample_size> parameter.
1549
1550 <sample_size>
1551 The size of the sample in number of logs to consider when balancing
1552 their logging loads. It is used to balance the load of the logs to
1553 send to the syslog server. This size must be greater or equal to the
1554 maximum of the high limits of the ranges.
1555 (see also <ranges> parameter).
1556
Robert Tsai81ae1952007-12-05 10:47:29 +01001557 <facility> must be one of the 24 standard syslog facilities :
Willy Tarreau6a06a402007-07-15 20:15:28 +02001558
Willy Tarreaue8746a02018-11-12 08:45:00 +01001559 kern user mail daemon auth syslog lpr news
1560 uucp cron auth2 ftp ntp audit alert cron2
1561 local0 local1 local2 local3 local4 local5 local6 local7
1562
Willy Tarreauc1b06452018-11-12 11:57:56 +01001563 Note that the facility is ignored for the "short" and "raw"
1564 formats, but still required as a positional field. It is
1565 recommended to use "daemon" in this case to make it clear that
1566 it's only supposed to be used locally.
Willy Tarreau6a06a402007-07-15 20:15:28 +02001567
1568 An optional level can be specified to filter outgoing messages. By default,
Willy Tarreauf7edefa2009-05-10 17:20:05 +02001569 all messages are sent. If a maximum level is specified, only messages with a
1570 severity at least as important as this level will be sent. An optional minimum
1571 level can be specified. If it is set, logs emitted with a more severe level
1572 than this one will be capped to this level. This is used to avoid sending
1573 "emerg" messages on all terminals on some default syslog configurations.
1574 Eight levels are known :
Willy Tarreau6a06a402007-07-15 20:15:28 +02001575
Cyril Bontédc4d9032012-04-08 21:57:39 +02001576 emerg alert crit err warning notice info debug
Willy Tarreau6a06a402007-07-15 20:15:28 +02001577
Joe Williamsdf5b38f2010-12-29 17:05:48 +01001578log-send-hostname [<string>]
1579 Sets the hostname field in the syslog header. If optional "string" parameter
1580 is set the header is set to the string contents, otherwise uses the hostname
1581 of the system. Generally used if one is not relaying logs through an
1582 intermediate syslog server or for simply customizing the hostname printed in
1583 the logs.
1584
Kevinm48936af2010-12-22 16:08:21 +00001585log-tag <string>
1586 Sets the tag field in the syslog header to this string. It defaults to the
1587 program name as launched from the command line, which usually is "haproxy".
1588 Sometimes it can be useful to differentiate between multiple processes
Willy Tarreau094af4e2015-01-07 15:03:42 +01001589 running on the same host. See also the per-proxy "log-tag" directive.
Kevinm48936af2010-12-22 16:08:21 +00001590
Thierry FOURNIER90da1912015-03-05 11:17:06 +01001591lua-load <file>
Thierry Fournier59f11be2020-11-29 00:37:41 +01001592 This global directive loads and executes a Lua file in the shared context
1593 that is visible to all threads. Any variable set in such a context is visible
1594 from any thread. This is the easiest and recommended way to load Lua programs
1595 but it will not scale well if a lot of Lua calls are performed, as only one
1596 thread may be running on the global state at a time. A program loaded this
1597 way will always see 0 in the "core.thread" variable. This directive can be
Thierry FOURNIER90da1912015-03-05 11:17:06 +01001598 used multiple times.
1599
Thierry Fournier59f11be2020-11-29 00:37:41 +01001600lua-load-per-thread <file>
1601 This global directive loads and executes a Lua file into each started thread.
1602 Any global variable has a thread-local visibility so that each thread could
1603 see a different value. As such it is strongly recommended not to use global
1604 variables in programs loaded this way. An independent copy is loaded and
1605 initialized for each thread, everything is done sequentially and in the
1606 thread's numeric order from 1 to nbthread. If some operations need to be
1607 performed only once, the program should check the "core.thread" variable to
1608 figure what thread is being initialized. Programs loaded this way will run
1609 concurrently on all threads and will be highly scalable. This is the
1610 recommended way to load simple functions that register sample-fetches,
1611 converters, actions or services once it is certain the program doesn't depend
1612 on global variables. For the sake of simplicity, the directive is available
1613 even if only one thread is used and even if threads are disabled (in which
1614 case it will be equivalent to lua-load). This directive can be used multiple
1615 times.
1616
Tim Duesterhusdd74b5f2020-01-12 13:55:40 +01001617lua-prepend-path <string> [<type>]
1618 Prepends the given string followed by a semicolon to Lua's package.<type>
1619 variable.
1620 <type> must either be "path" or "cpath". If <type> is not given it defaults
1621 to "path".
1622
1623 Lua's paths are semicolon delimited lists of patterns that specify how the
1624 `require` function attempts to find the source file of a library. Question
1625 marks (?) within a pattern will be replaced by module name. The path is
1626 evaluated left to right. This implies that paths that are prepended later
1627 will be checked earlier.
1628
1629 As an example by specifying the following path:
1630
1631 lua-prepend-path /usr/share/haproxy-lua/?/init.lua
1632 lua-prepend-path /usr/share/haproxy-lua/?.lua
1633
1634 When `require "example"` is being called Lua will first attempt to load the
1635 /usr/share/haproxy-lua/example.lua script, if that does not exist the
1636 /usr/share/haproxy-lua/example/init.lua will be attempted and the default
1637 paths if that does not exist either.
1638
1639 See https://www.lua.org/pil/8.1.html for the details within the Lua
1640 documentation.
1641
William Lallemand4cfede82017-11-24 22:02:34 +01001642master-worker [no-exit-on-failure]
William Lallemande202b1e2017-06-01 17:38:56 +02001643 Master-worker mode. It is equivalent to the command line "-W" argument.
1644 This mode will launch a "master" which will monitor the "workers". Using
1645 this mode, you can reload HAProxy directly by sending a SIGUSR2 signal to
Davor Ocelice9ed2812017-12-25 17:49:28 +01001646 the master. The master-worker mode is compatible either with the foreground
William Lallemande202b1e2017-06-01 17:38:56 +02001647 or daemon mode. It is recommended to use this mode with multiprocess and
1648 systemd.
William Lallemand4cfede82017-11-24 22:02:34 +01001649 By default, if a worker exits with a bad return code, in the case of a
1650 segfault for example, all workers will be killed, and the master will leave.
1651 It is convenient to combine this behavior with Restart=on-failure in a
1652 systemd unit file in order to relaunch the whole process. If you don't want
1653 this behavior, you must use the keyword "no-exit-on-failure".
William Lallemande202b1e2017-06-01 17:38:56 +02001654
William Lallemand4cfede82017-11-24 22:02:34 +01001655 See also "-W" in the management guide.
William Lallemande202b1e2017-06-01 17:38:56 +02001656
William Lallemand27edc4b2019-05-07 17:49:33 +02001657mworker-max-reloads <number>
1658 In master-worker mode, this option limits the number of time a worker can
John Roeslerfb2fce12019-07-10 15:45:51 -05001659 survive to a reload. If the worker did not leave after a reload, once its
William Lallemand27edc4b2019-05-07 17:49:33 +02001660 number of reloads is greater than this number, the worker will receive a
1661 SIGTERM. This option helps to keep under control the number of workers.
1662 See also "show proc" in the Management Guide.
1663
Willy Tarreauf42d7942020-10-20 11:54:49 +02001664nbproc <number> (deprecated)
Willy Tarreau6a06a402007-07-15 20:15:28 +02001665 Creates <number> processes when going daemon. This requires the "daemon"
1666 mode. By default, only one process is created, which is the recommended mode
1667 of operation. For systems limited to small sets of file descriptors per
Willy Tarreau149ab772019-01-26 14:27:06 +01001668 process, it may be needed to fork multiple daemons. When set to a value
1669 larger than 1, threads are automatically disabled. USING MULTIPLE PROCESSES
Willy Tarreauf42d7942020-10-20 11:54:49 +02001670 IS HARDER TO DEBUG AND IS REALLY DISCOURAGED. This directive is deprecated
1671 and scheduled for removal in 2.5. Please use "nbthread" instead. See also
1672 "daemon" and "nbthread".
Willy Tarreau6a06a402007-07-15 20:15:28 +02001673
Christopher Fauletbe0faa22017-08-29 15:37:10 +02001674nbthread <number>
1675 This setting is only available when support for threads was built in. It
Daniel Corbett9f0843f2021-05-08 10:50:37 -04001676 makes HAProxy run on <number> threads. This is exclusive with "nbproc". While
Willy Tarreau26f6ae12019-02-02 12:56:15 +01001677 "nbproc" historically used to be the only way to use multiple processors, it
1678 also involved a number of shortcomings related to the lack of synchronization
1679 between processes (health-checks, peers, stick-tables, stats, ...) which do
1680 not affect threads. As such, any modern configuration is strongly encouraged
Willy Tarreau149ab772019-01-26 14:27:06 +01001681 to migrate away from "nbproc" to "nbthread". "nbthread" also works when
1682 HAProxy is started in foreground. On some platforms supporting CPU affinity,
1683 when nbproc is not used, the default "nbthread" value is automatically set to
1684 the number of CPUs the process is bound to upon startup. This means that the
1685 thread count can easily be adjusted from the calling process using commands
1686 like "taskset" or "cpuset". Otherwise, this value defaults to 1. The default
1687 value is reported in the output of "haproxy -vv". See also "nbproc".
Christopher Fauletbe0faa22017-08-29 15:37:10 +02001688
Amaury Denoyelle0f50cb92021-03-26 18:50:33 +01001689numa-cpu-mapping
Daniel Corbett9f0843f2021-05-08 10:50:37 -04001690 By default, if running on Linux, HAProxy inspects on startup the CPU topology
Amaury Denoyelle0f50cb92021-03-26 18:50:33 +01001691 of the machine. If a multi-socket machine is detected, the affinity is
1692 automatically calculated to run on the CPUs of a single node. This is done in
1693 order to not suffer from the performance penalties caused by the inter-socket
1694 bus latency. However, if the applied binding is non optimal on a particular
1695 architecture, it can be disabled with the statement 'no numa-cpu-mapping'.
1696 This automatic binding is also not applied if a nbthread statement is present
1697 in the configuration, or the affinity of the process is already specified,
1698 for example via the 'cpu-map' directive or the taskset utility.
1699
Willy Tarreau6a06a402007-07-15 20:15:28 +02001700pidfile <pidfile>
MIZUTA Takeshic32f3942020-08-26 13:46:19 +09001701 Writes PIDs of all daemons into file <pidfile> when daemon mode or writes PID
1702 of master process into file <pidfile> when master-worker mode. This option is
1703 equivalent to the "-p" command line argument. The file must be accessible to
1704 the user starting the process. See also "daemon" and "master-worker".
Willy Tarreau6a06a402007-07-15 20:15:28 +02001705
Willy Tarreau119e50e2020-05-22 13:53:29 +02001706pp2-never-send-local
1707 A bug in the PROXY protocol v2 implementation was present in HAProxy up to
1708 version 2.1, causing it to emit a PROXY command instead of a LOCAL command
1709 for health checks. This is particularly minor but confuses some servers'
1710 logs. Sadly, the bug was discovered very late and revealed that some servers
1711 which possibly only tested their PROXY protocol implementation against
1712 HAProxy fail to properly handle the LOCAL command, and permanently remain in
1713 the "down" state when HAProxy checks them. When this happens, it is possible
1714 to enable this global option to revert to the older (bogus) behavior for the
1715 time it takes to contact the affected components' vendors and get them fixed.
1716 This option is disabled by default and acts on all servers having the
1717 "send-proxy-v2" statement.
1718
Willy Tarreau1d549722016-02-16 12:41:57 +01001719presetenv <name> <value>
1720 Sets environment variable <name> to value <value>. If the variable exists, it
1721 is NOT overwritten. The changes immediately take effect so that the next line
1722 in the configuration file sees the new value. See also "setenv", "resetenv",
1723 and "unsetenv".
1724
1725resetenv [<name> ...]
1726 Removes all environment variables except the ones specified in argument. It
1727 allows to use a clean controlled environment before setting new values with
1728 setenv or unsetenv. Please note that some internal functions may make use of
1729 some environment variables, such as time manipulation functions, but also
1730 OpenSSL or even external checks. This must be used with extreme care and only
1731 after complete validation. The changes immediately take effect so that the
1732 next line in the configuration file sees the new environment. See also
1733 "setenv", "presetenv", and "unsetenv".
1734
Christopher Fauletff4121f2017-11-22 16:38:49 +01001735stats bind-process [ all | odd | even | <process_num>[-[process_num>]] ] ...
Willy Tarreau35b7b162012-10-22 23:17:18 +02001736 Limits the stats socket to a certain set of processes numbers. By default the
1737 stats socket is bound to all processes, causing a warning to be emitted when
1738 nbproc is greater than 1 because there is no way to select the target process
1739 when connecting. However, by using this setting, it becomes possible to pin
1740 the stats socket to a specific set of processes, typically the first one. The
1741 warning will automatically be disabled when this setting is used, whatever
Willy Tarreaua9db57e2013-01-18 11:29:29 +01001742 the number of processes used. The maximum process ID depends on the machine's
Christopher Fauletff4121f2017-11-22 16:38:49 +01001743 word size (32 or 64). Ranges can be partially defined. The higher bound can
1744 be omitted. In such case, it is replaced by the corresponding maximum
1745 value. A better option consists in using the "process" setting of the "stats
1746 socket" line to force the process on each line.
Willy Tarreau35b7b162012-10-22 23:17:18 +02001747
Baptiste Assmann5626f482015-08-23 10:00:10 +02001748server-state-base <directory>
1749 Specifies the directory prefix to be prepended in front of all servers state
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02001750 file names which do not start with a '/'. See also "server-state-file",
1751 "load-server-state-from-file" and "server-state-file-name".
Baptiste Assmannef1f0fc2015-08-23 10:06:39 +02001752
1753server-state-file <file>
1754 Specifies the path to the file containing state of servers. If the path starts
1755 with a slash ('/'), it is considered absolute, otherwise it is considered
1756 relative to the directory specified using "server-state-base" (if set) or to
1757 the current directory. Before reloading HAProxy, it is possible to save the
1758 servers' current state using the stats command "show servers state". The
1759 output of this command must be written in the file pointed by <file>. When
1760 starting up, before handling traffic, HAProxy will read, load and apply state
1761 for each server found in the file and available in its current running
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02001762 configuration. See also "server-state-base" and "show servers state",
1763 "load-server-state-from-file" and "server-state-file-name"
Baptiste Assmann5626f482015-08-23 10:00:10 +02001764
Willy Tarreau13d2ba22021-03-26 11:38:08 +01001765set-var <var-name> <expr>
1766 Sets the process-wide variable '<var-name>' to the result of the evaluation
1767 of the sample expression <expr>. The variable '<var-name>' may only be a
1768 process-wide variable (using the 'proc.' prefix). It works exactly like the
1769 'set-var' action in TCP or HTTP rules except that the expression is evaluated
1770 at configuration parsing time and that the variable is instantly set. The
1771 sample fetch functions and converters permitted in the expression are only
1772 those using internal data, typically 'int(value)' or 'str(value)'. It's is
1773 possible to reference previously allocated variables as well. These variables
1774 will then be readable (and modifiable) from the regular rule sets.
1775
1776 Example:
1777 global
1778 set-var proc.current_state str(primary)
1779 set-var proc.prio int(100)
1780 set-var proc.threshold int(200),sub(proc.prio)
1781
Willy Tarreau1d549722016-02-16 12:41:57 +01001782setenv <name> <value>
1783 Sets environment variable <name> to value <value>. If the variable exists, it
1784 is overwritten. The changes immediately take effect so that the next line in
1785 the configuration file sees the new value. See also "presetenv", "resetenv",
1786 and "unsetenv".
1787
Willy Tarreau636848a2019-04-15 19:38:50 +02001788set-dumpable
1789 This option is better left disabled by default and enabled only upon a
William Dauchyec730982019-10-27 20:08:10 +01001790 developer's request. If it has been enabled, it may still be forcibly
1791 disabled by prefixing it with the "no" keyword. It has no impact on
1792 performance nor stability but will try hard to re-enable core dumps that were
1793 possibly disabled by file size limitations (ulimit -f), core size limitations
1794 (ulimit -c), or "dumpability" of a process after changing its UID/GID (such
1795 as /proc/sys/fs/suid_dumpable on Linux). Core dumps might still be limited by
1796 the current directory's permissions (check what directory the file is started
1797 from), the chroot directory's permission (it may be needed to temporarily
1798 disable the chroot directive or to move it to a dedicated writable location),
1799 or any other system-specific constraint. For example, some Linux flavours are
1800 notorious for replacing the default core file with a path to an executable
1801 not even installed on the system (check /proc/sys/kernel/core_pattern). Often,
1802 simply writing "core", "core.%p" or "/var/log/core/core.%p" addresses the
1803 issue. When trying to enable this option waiting for a rare issue to
1804 re-appear, it's often a good idea to first try to obtain such a dump by
Daniel Corbett9f0843f2021-05-08 10:50:37 -04001805 issuing, for example, "kill -11" to the "haproxy" process and verify that it
William Dauchyec730982019-10-27 20:08:10 +01001806 leaves a core where expected when dying.
Willy Tarreau636848a2019-04-15 19:38:50 +02001807
Willy Tarreau610f04b2014-02-13 11:36:41 +01001808ssl-default-bind-ciphers <ciphers>
1809 This setting is only available when support for OpenSSL was built in. It sets
1810 the default string describing the list of cipher algorithms ("cipher suite")
Bertrand Jacquin8cf7c1e2019-02-03 18:35:25 +00001811 that are negotiated during the SSL/TLS handshake up to TLSv1.2 for all
Dirkjan Bussink415150f2018-09-14 11:14:21 +02001812 "bind" lines which do not explicitly define theirs. The format of the string
Bertrand Jacquin4f03ab02019-02-03 18:48:49 +00001813 is defined in "man 1 ciphers" from OpenSSL man pages. For background
1814 information and recommendations see e.g.
1815 (https://wiki.mozilla.org/Security/Server_Side_TLS) and
1816 (https://mozilla.github.io/server-side-tls/ssl-config-generator/). For TLSv1.3
1817 cipher configuration, please check the "ssl-default-bind-ciphersuites" keyword.
1818 Please check the "bind" keyword for more information.
Dirkjan Bussink415150f2018-09-14 11:14:21 +02001819
1820ssl-default-bind-ciphersuites <ciphersuites>
1821 This setting is only available when support for OpenSSL was built in and
1822 OpenSSL 1.1.1 or later was used to build HAProxy. It sets the default string
1823 describing the list of cipher algorithms ("cipher suite") that are negotiated
1824 during the TLSv1.3 handshake for all "bind" lines which do not explicitly define
1825 theirs. The format of the string is defined in
Bertrand Jacquin4f03ab02019-02-03 18:48:49 +00001826 "man 1 ciphers" from OpenSSL man pages under the section "ciphersuites". For
1827 cipher configuration for TLSv1.2 and earlier, please check the
1828 "ssl-default-bind-ciphers" keyword. Please check the "bind" keyword for more
Dirkjan Bussink415150f2018-09-14 11:14:21 +02001829 information.
Willy Tarreau610f04b2014-02-13 11:36:41 +01001830
Jerome Magninb203ff62020-04-03 15:28:22 +02001831ssl-default-bind-curves <curves>
1832 This setting is only available when support for OpenSSL was built in. It sets
1833 the default string describing the list of elliptic curves algorithms ("curve
1834 suite") that are negotiated during the SSL/TLS handshake with ECDHE. The format
1835 of the string is a colon-delimited list of curve name.
1836 Please check the "bind" keyword for more information.
1837
Emeric Brun2c86cbf2014-10-30 15:56:50 +01001838ssl-default-bind-options [<option>]...
1839 This setting is only available when support for OpenSSL was built in. It sets
1840 default ssl-options to force on all "bind" lines. Please check the "bind"
1841 keyword to see available options.
1842
1843 Example:
1844 global
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +02001845 ssl-default-bind-options ssl-min-ver TLSv1.0 no-tls-tickets
Emeric Brun2c86cbf2014-10-30 15:56:50 +01001846
Willy Tarreau610f04b2014-02-13 11:36:41 +01001847ssl-default-server-ciphers <ciphers>
1848 This setting is only available when support for OpenSSL was built in. It
1849 sets the default string describing the list of cipher algorithms that are
Bertrand Jacquin8cf7c1e2019-02-03 18:35:25 +00001850 negotiated during the SSL/TLS handshake up to TLSv1.2 with the server,
Dirkjan Bussink415150f2018-09-14 11:14:21 +02001851 for all "server" lines which do not explicitly define theirs. The format of
Bertrand Jacquin4f03ab02019-02-03 18:48:49 +00001852 the string is defined in "man 1 ciphers" from OpenSSL man pages. For background
1853 information and recommendations see e.g.
1854 (https://wiki.mozilla.org/Security/Server_Side_TLS) and
1855 (https://mozilla.github.io/server-side-tls/ssl-config-generator/).
1856 For TLSv1.3 cipher configuration, please check the
1857 "ssl-default-server-ciphersuites" keyword. Please check the "server" keyword
1858 for more information.
Dirkjan Bussink415150f2018-09-14 11:14:21 +02001859
1860ssl-default-server-ciphersuites <ciphersuites>
1861 This setting is only available when support for OpenSSL was built in and
1862 OpenSSL 1.1.1 or later was used to build HAProxy. It sets the default
1863 string describing the list of cipher algorithms that are negotiated during
1864 the TLSv1.3 handshake with the server, for all "server" lines which do not
1865 explicitly define theirs. The format of the string is defined in
Bertrand Jacquin4f03ab02019-02-03 18:48:49 +00001866 "man 1 ciphers" from OpenSSL man pages under the section "ciphersuites". For
1867 cipher configuration for TLSv1.2 and earlier, please check the
1868 "ssl-default-server-ciphers" keyword. Please check the "server" keyword for
1869 more information.
Willy Tarreau610f04b2014-02-13 11:36:41 +01001870
Emeric Brun2c86cbf2014-10-30 15:56:50 +01001871ssl-default-server-options [<option>]...
1872 This setting is only available when support for OpenSSL was built in. It sets
1873 default ssl-options to force on all "server" lines. Please check the "server"
1874 keyword to see available options.
1875
Remi Gacogne47783ef2015-05-29 15:53:22 +02001876ssl-dh-param-file <file>
1877 This setting is only available when support for OpenSSL was built in. It sets
1878 the default DH parameters that are used during the SSL/TLS handshake when
1879 ephemeral Diffie-Hellman (DHE) key exchange is used, for all "bind" lines
Davor Ocelice9ed2812017-12-25 17:49:28 +01001880 which do not explicitly define theirs. It will be overridden by custom DH
Remi Gacogne47783ef2015-05-29 15:53:22 +02001881 parameters found in a bind certificate file if any. If custom DH parameters
Cyril Bonté307ee1e2015-09-28 23:16:06 +02001882 are not specified either by using ssl-dh-param-file or by setting them
1883 directly in the certificate file, pre-generated DH parameters of the size
1884 specified by tune.ssl.default-dh-param will be used. Custom parameters are
1885 known to be more secure and therefore their use is recommended.
Remi Gacogne47783ef2015-05-29 15:53:22 +02001886 Custom DH parameters may be generated by using the OpenSSL command
1887 "openssl dhparam <size>", where size should be at least 2048, as 1024-bit DH
1888 parameters should not be considered secure anymore.
1889
William Lallemand8e8581e2020-10-20 17:36:46 +02001890ssl-load-extra-del-ext
1891 This setting allows to configure the way HAProxy does the lookup for the
1892 extra SSL files. By default HAProxy adds a new extension to the filename.
William Lallemand089c1382020-10-23 17:35:12 +02001893 (ex: with "foobar.crt" load "foobar.crt.key"). With this option enabled,
William Lallemand8e8581e2020-10-20 17:36:46 +02001894 HAProxy removes the extension before adding the new one (ex: with
William Lallemand089c1382020-10-23 17:35:12 +02001895 "foobar.crt" load "foobar.key").
1896
1897 Your crt file must have a ".crt" extension for this option to work.
William Lallemand8e8581e2020-10-20 17:36:46 +02001898
1899 This option is not compatible with bundle extensions (.ecdsa, .rsa. .dsa)
1900 and won't try to remove them.
1901
1902 This option is disabled by default. See also "ssl-load-extra-files".
1903
William Lallemand4c5adbf2020-02-24 14:23:22 +01001904ssl-load-extra-files <none|all|bundle|sctl|ocsp|issuer|key>*
William Lallemand3af48e72020-02-03 17:15:52 +01001905 This setting alters the way HAProxy will look for unspecified files during
Remi Tricot-Le Breton7c980df2021-05-07 15:28:08 +02001906 the loading of the SSL certificates. This option applies to certificates
1907 associated to "bind" lines as well as "server" lines but some of the extra
1908 files will not have any functional impact for "server" line certificates.
William Lallemand3af48e72020-02-03 17:15:52 +01001909
1910 By default, HAProxy discovers automatically a lot of files not specified in
1911 the configuration, and you may want to disable this behavior if you want to
1912 optimize the startup time.
1913
1914 "none": Only load the files specified in the configuration. Don't try to load
1915 a certificate bundle if the file does not exist. In the case of a directory,
1916 it won't try to bundle the certificates if they have the same basename.
1917
1918 "all": This is the default behavior, it will try to load everything,
William Lallemand4c5adbf2020-02-24 14:23:22 +01001919 bundles, sctl, ocsp, issuer, key.
William Lallemand3af48e72020-02-03 17:15:52 +01001920
1921 "bundle": When a file specified in the configuration does not exist, HAProxy
Remi Tricot-Le Breton7c980df2021-05-07 15:28:08 +02001922 will try to load a "cert bundle". Certificate bundles are only managed on the
1923 frontend side and will not work for backend certificates.
William Lallemandf9ff3ec2020-10-02 17:57:44 +02001924
1925 Starting from HAProxy 2.3, the bundles are not loaded in the same OpenSSL
1926 certificate store, instead it will loads each certificate in a separate
1927 store which is equivalent to declaring multiple "crt". OpenSSL 1.1.1 is
1928 required to achieve this. Which means that bundles are now used only for
1929 backward compatibility and are not mandatory anymore to do an hybrid RSA/ECC
Remi Tricot-Le Breton7c980df2021-05-07 15:28:08 +02001930 bind configuration.
William Lallemandf9ff3ec2020-10-02 17:57:44 +02001931
1932 To associate these PEM files into a "cert bundle" that is recognized by
Daniel Corbett9f0843f2021-05-08 10:50:37 -04001933 HAProxy, they must be named in the following way: All PEM files that are to
William Lallemandf9ff3ec2020-10-02 17:57:44 +02001934 be bundled must have the same base name, with a suffix indicating the key
1935 type. Currently, three suffixes are supported: rsa, dsa and ecdsa. For
1936 example, if www.example.com has two PEM files, an RSA file and an ECDSA
1937 file, they must be named: "example.pem.rsa" and "example.pem.ecdsa". The
1938 first part of the filename is arbitrary; only the suffix matters. To load
Daniel Corbett9f0843f2021-05-08 10:50:37 -04001939 this bundle into HAProxy, specify the base name only:
William Lallemandf9ff3ec2020-10-02 17:57:44 +02001940
1941 Example : bind :8443 ssl crt example.pem
1942
Daniel Corbett9f0843f2021-05-08 10:50:37 -04001943 Note that the suffix is not given to HAProxy; this tells HAProxy to look for
William Lallemandf9ff3ec2020-10-02 17:57:44 +02001944 a cert bundle.
1945
1946 HAProxy will load all PEM files in the bundle as if they were configured
1947 separately in several "crt".
1948
1949 The bundle loading does not have an impact anymore on the directory loading
1950 since files are loading separately.
1951
1952 On the CLI, bundles are seen as separate files, and the bundle extension is
1953 required to commit them.
1954
William Dauchy57dd6f12020-10-06 15:22:37 +02001955 OCSP files (.ocsp), issuer files (.issuer), Certificate Transparency (.sctl)
William Lallemandf9ff3ec2020-10-02 17:57:44 +02001956 as well as private keys (.key) are supported with multi-cert bundling.
William Lallemand3af48e72020-02-03 17:15:52 +01001957
Remi Tricot-Le Breton7c980df2021-05-07 15:28:08 +02001958 "sctl": Try to load "<basename>.sctl" for each crt keyword. If provided for
1959 a backend certificate, it will be loaded but will not have any functional
1960 impact.
William Lallemand3af48e72020-02-03 17:15:52 +01001961
Remi Tricot-Le Breton7c980df2021-05-07 15:28:08 +02001962 "ocsp": Try to load "<basename>.ocsp" for each crt keyword. If provided for
1963 a backend certificate, it will be loaded but will not have any functional
1964 impact.
William Lallemand3af48e72020-02-03 17:15:52 +01001965
1966 "issuer": Try to load "<basename>.issuer" if the issuer of the OCSP file is
Remi Tricot-Le Breton7c980df2021-05-07 15:28:08 +02001967 not provided in the PEM file. If provided for a backend certificate, it will
1968 be loaded but will not have any functional impact.
William Lallemand3af48e72020-02-03 17:15:52 +01001969
William Lallemand4c5adbf2020-02-24 14:23:22 +01001970 "key": If the private key was not provided by the PEM file, try to load a
1971 file "<basename>.key" containing a private key.
1972
William Lallemand3af48e72020-02-03 17:15:52 +01001973 The default behavior is "all".
1974
1975 Example:
1976 ssl-load-extra-files bundle sctl
1977 ssl-load-extra-files sctl ocsp issuer
1978 ssl-load-extra-files none
1979
Remi Tricot-Le Breton7c980df2021-05-07 15:28:08 +02001980 See also: "crt", section 5.1 about bind options and section 5.2 about server
1981 options.
William Lallemand3af48e72020-02-03 17:15:52 +01001982
Emeric Brun850efd52014-01-29 12:24:34 +01001983ssl-server-verify [none|required]
1984 The default behavior for SSL verify on servers side. If specified to 'none',
1985 servers certificates are not verified. The default is 'required' except if
1986 forced using cmdline option '-dV'.
1987
Emmanuel Hocdetc3b7e742020-04-22 11:06:19 +02001988ssl-skip-self-issued-ca
Daniel Corbett67a82712020-07-06 23:01:19 -04001989 Self issued CA, aka x509 root CA, is the anchor for chain validation: as a
Emmanuel Hocdetc3b7e742020-04-22 11:06:19 +02001990 server is useless to send it, client must have it. Standard configuration
1991 need to not include such CA in PEM file. This option allows you to keep such
1992 CA in PEM file without sending it to the client. Use case is to provide
1993 issuer for ocsp without the need for '.issuer' file and be able to share it
1994 with 'issuers-chain-path'. This concerns all certificates without intermediate
1995 certificates. It's useless for BoringSSL, .issuer is ignored because ocsp
William Lallemand9a1d8392020-08-10 17:28:23 +02001996 bits does not need it. Requires at least OpenSSL 1.0.2.
Emmanuel Hocdetc3b7e742020-04-22 11:06:19 +02001997
Willy Tarreauabb175f2012-09-24 12:43:26 +02001998stats socket [<address:port>|<path>] [param*]
1999 Binds a UNIX socket to <path> or a TCPv4/v6 address to <address:port>.
2000 Connections to this socket will return various statistics outputs and even
2001 allow some commands to be issued to change some runtime settings. Please
Willy Tarreau1af20c72017-06-23 16:01:14 +02002002 consult section 9.3 "Unix Socket commands" of Management Guide for more
Kevin Decherf949c7202015-10-13 23:26:44 +02002003 details.
Willy Tarreau6162db22009-10-10 17:13:00 +02002004
Willy Tarreauabb175f2012-09-24 12:43:26 +02002005 All parameters supported by "bind" lines are supported, for instance to
2006 restrict access to some users or their access rights. Please consult
2007 section 5.1 for more information.
Willy Tarreaufbee7132007-10-18 13:53:22 +02002008
2009stats timeout <timeout, in milliseconds>
2010 The default timeout on the stats socket is set to 10 seconds. It is possible
2011 to change this value with "stats timeout". The value must be passed in
Willy Tarreaubefdff12007-12-02 22:27:38 +01002012 milliseconds, or be suffixed by a time unit among { us, ms, s, m, h, d }.
Willy Tarreaufbee7132007-10-18 13:53:22 +02002013
2014stats maxconn <connections>
2015 By default, the stats socket is limited to 10 concurrent connections. It is
2016 possible to change this value with "stats maxconn".
2017
Willy Tarreau6a06a402007-07-15 20:15:28 +02002018uid <number>
Thayne McCombscdbcca92021-01-07 21:24:41 -07002019 Changes the process's user ID to <number>. It is recommended that the user ID
Willy Tarreau6a06a402007-07-15 20:15:28 +02002020 is dedicated to HAProxy or to a small set of similar daemons. HAProxy must
2021 be started with superuser privileges in order to be able to switch to another
2022 one. See also "gid" and "user".
2023
2024ulimit-n <number>
2025 Sets the maximum number of per-process file-descriptors to <number>. By
2026 default, it is automatically computed, so it is recommended not to use this
2027 option.
2028
Willy Tarreauceb24bc2010-11-09 12:46:41 +01002029unix-bind [ prefix <prefix> ] [ mode <mode> ] [ user <user> ] [ uid <uid> ]
2030 [ group <group> ] [ gid <gid> ]
2031
2032 Fixes common settings to UNIX listening sockets declared in "bind" statements.
2033 This is mainly used to simplify declaration of those UNIX sockets and reduce
2034 the risk of errors, since those settings are most commonly required but are
2035 also process-specific. The <prefix> setting can be used to force all socket
2036 path to be relative to that directory. This might be needed to access another
Daniel Corbett9f0843f2021-05-08 10:50:37 -04002037 component's chroot. Note that those paths are resolved before HAProxy chroots
Willy Tarreauceb24bc2010-11-09 12:46:41 +01002038 itself, so they are absolute. The <mode>, <user>, <uid>, <group> and <gid>
2039 all have the same meaning as their homonyms used by the "bind" statement. If
2040 both are specified, the "bind" statement has priority, meaning that the
2041 "unix-bind" settings may be seen as process-wide default settings.
2042
Willy Tarreau1d549722016-02-16 12:41:57 +01002043unsetenv [<name> ...]
2044 Removes environment variables specified in arguments. This can be useful to
2045 hide some sensitive information that are occasionally inherited from the
2046 user's environment during some operations. Variables which did not exist are
2047 silently ignored so that after the operation, it is certain that none of
2048 these variables remain. The changes immediately take effect so that the next
2049 line in the configuration file will not see these variables. See also
2050 "setenv", "presetenv", and "resetenv".
2051
Willy Tarreau6a06a402007-07-15 20:15:28 +02002052user <user name>
2053 Similar to "uid" but uses the UID of user name <user name> from /etc/passwd.
2054 See also "uid" and "group".
2055
Krzysztof Piotr Oledzki48cb2ae2009-10-02 22:51:14 +02002056node <name>
2057 Only letters, digits, hyphen and underscore are allowed, like in DNS names.
2058
2059 This statement is useful in HA configurations where two or more processes or
2060 servers share the same IP address. By setting a different node-name on all
2061 nodes, it becomes easy to immediately spot what server is handling the
2062 traffic.
2063
2064description <text>
2065 Add a text that describes the instance.
2066
2067 Please note that it is required to escape certain characters (# for example)
2068 and this text is inserted into a html page so you should avoid using
2069 "<" and ">" characters.
2070
Thomas Holmesdb04f192015-05-18 13:21:39 +0100207151degrees-data-file <file path>
2072 The path of the 51Degrees data file to provide device detection services. The
Davor Ocelice9ed2812017-12-25 17:49:28 +01002073 file should be unzipped and accessible by HAProxy with relevant permissions.
Thomas Holmesdb04f192015-05-18 13:21:39 +01002074
Daniel Corbett9f0843f2021-05-08 10:50:37 -04002075 Please note that this option is only available when HAProxy has been
Thomas Holmesdb04f192015-05-18 13:21:39 +01002076 compiled with USE_51DEGREES.
2077
Ben Shillitof25e8e52016-12-02 14:25:37 +0000207851degrees-property-name-list [<string> ...]
Thomas Holmesdb04f192015-05-18 13:21:39 +01002079 A list of 51Degrees property names to be load from the dataset. A full list
2080 of names is available on the 51Degrees website:
2081 https://51degrees.com/resources/property-dictionary
2082
Daniel Corbett9f0843f2021-05-08 10:50:37 -04002083 Please note that this option is only available when HAProxy has been
Thomas Holmesdb04f192015-05-18 13:21:39 +01002084 compiled with USE_51DEGREES.
2085
Dragan Dosen93b38d92015-06-29 16:43:25 +0200208651degrees-property-separator <char>
Thomas Holmesdb04f192015-05-18 13:21:39 +01002087 A char that will be appended to every property value in a response header
2088 containing 51Degrees results. If not set that will be set as ','.
2089
Daniel Corbett9f0843f2021-05-08 10:50:37 -04002090 Please note that this option is only available when HAProxy has been
Dragan Dosenae6d39a2015-06-29 16:43:27 +02002091 compiled with USE_51DEGREES.
2092
209351degrees-cache-size <number>
2094 Sets the size of the 51Degrees converter cache to <number> entries. This
2095 is an LRU cache which reminds previous device detections and their results.
2096 By default, this cache is disabled.
2097
Daniel Corbett9f0843f2021-05-08 10:50:37 -04002098 Please note that this option is only available when HAProxy has been
Thomas Holmesdb04f192015-05-18 13:21:39 +01002099 compiled with USE_51DEGREES.
2100
Willy Tarreaub3cc9f22019-04-19 16:03:32 +02002101wurfl-data-file <file path>
2102 The path of the WURFL data file to provide device detection services. The
2103 file should be accessible by HAProxy with relevant permissions.
2104
Daniel Corbett9f0843f2021-05-08 10:50:37 -04002105 Please note that this option is only available when HAProxy has been compiled
Willy Tarreaub3cc9f22019-04-19 16:03:32 +02002106 with USE_WURFL=1.
2107
2108wurfl-information-list [<capability>]*
2109 A space-delimited list of WURFL capabilities, virtual capabilities, property
2110 names we plan to use in injected headers. A full list of capability and
2111 virtual capability names is available on the Scientiamobile website :
2112
2113 https://www.scientiamobile.com/wurflCapability
2114
2115 Valid WURFL properties are:
2116 - wurfl_id Contains the device ID of the matched device.
2117
2118 - wurfl_root_id Contains the device root ID of the matched
2119 device.
2120
2121 - wurfl_isdevroot Tells if the matched device is a root device.
2122 Possible values are "TRUE" or "FALSE".
2123
2124 - wurfl_useragent The original useragent coming with this
2125 particular web request.
2126
2127 - wurfl_api_version Contains a string representing the currently
2128 used Libwurfl API version.
2129
Willy Tarreaub3cc9f22019-04-19 16:03:32 +02002130 - wurfl_info A string containing information on the parsed
2131 wurfl.xml and its full path.
2132
2133 - wurfl_last_load_time Contains the UNIX timestamp of the last time
2134 WURFL has been loaded successfully.
2135
2136 - wurfl_normalized_useragent The normalized useragent.
2137
Daniel Corbett9f0843f2021-05-08 10:50:37 -04002138 Please note that this option is only available when HAProxy has been compiled
Willy Tarreaub3cc9f22019-04-19 16:03:32 +02002139 with USE_WURFL=1.
2140
2141wurfl-information-list-separator <char>
2142 A char that will be used to separate values in a response header containing
2143 WURFL results. If not set that a comma (',') will be used by default.
2144
Daniel Corbett9f0843f2021-05-08 10:50:37 -04002145 Please note that this option is only available when HAProxy has been compiled
Willy Tarreaub3cc9f22019-04-19 16:03:32 +02002146 with USE_WURFL=1.
2147
2148wurfl-patch-file [<file path>]
2149 A list of WURFL patch file paths. Note that patches are loaded during startup
2150 thus before the chroot.
2151
Daniel Corbett9f0843f2021-05-08 10:50:37 -04002152 Please note that this option is only available when HAProxy has been compiled
Willy Tarreaub3cc9f22019-04-19 16:03:32 +02002153 with USE_WURFL=1.
2154
paulborilebad132c2019-04-18 11:57:04 +02002155wurfl-cache-size <size>
2156 Sets the WURFL Useragent cache size. For faster lookups, already processed user
2157 agents are kept in a LRU cache :
Willy Tarreaub3cc9f22019-04-19 16:03:32 +02002158 - "0" : no cache is used.
paulborilebad132c2019-04-18 11:57:04 +02002159 - <size> : size of lru cache in elements.
Willy Tarreaub3cc9f22019-04-19 16:03:32 +02002160
Daniel Corbett9f0843f2021-05-08 10:50:37 -04002161 Please note that this option is only available when HAProxy has been compiled
Willy Tarreaub3cc9f22019-04-19 16:03:32 +02002162 with USE_WURFL=1.
2163
William Dauchy0fec3ab2019-10-27 20:08:11 +01002164strict-limits
Daniel Corbett9f0843f2021-05-08 10:50:37 -04002165 Makes process fail at startup when a setrlimit fails. HAProxy tries to set the
William Dauchya5194602020-03-28 19:29:58 +01002166 best setrlimit according to what has been calculated. If it fails, it will
2167 emit a warning. This option is here to guarantee an explicit failure of
Daniel Corbett9f0843f2021-05-08 10:50:37 -04002168 HAProxy when those limits fail. It is enabled by default. It may still be
William Dauchya5194602020-03-28 19:29:58 +01002169 forcibly disabled by prefixing it with the "no" keyword.
William Dauchy0fec3ab2019-10-27 20:08:11 +01002170
Willy Tarreauc57f0e22009-05-10 13:12:33 +020021713.2. Performance tuning
Willy Tarreau6a06a402007-07-15 20:15:28 +02002172-----------------------
2173
Willy Tarreaubeb859a2018-11-22 18:07:59 +01002174busy-polling
2175 In some situations, especially when dealing with low latency on processors
2176 supporting a variable frequency or when running inside virtual machines, each
2177 time the process waits for an I/O using the poller, the processor goes back
2178 to sleep or is offered to another VM for a long time, and it causes
2179 excessively high latencies. This option provides a solution preventing the
2180 processor from sleeping by always using a null timeout on the pollers. This
2181 results in a significant latency reduction (30 to 100 microseconds observed)
2182 at the expense of a risk to overheat the processor. It may even be used with
2183 threads, in which case improperly bound threads may heavily conflict,
2184 resulting in a worse performance and high values for the CPU stolen fields
2185 in "show info" output, indicating which threads are misconfigured. It is
2186 important not to let the process run on the same processor as the network
2187 interrupts when this option is used. It is also better to avoid using it on
2188 multiple CPU threads sharing the same core. This option is disabled by
2189 default. If it has been enabled, it may still be forcibly disabled by
2190 prefixing it with the "no" keyword. It is ignored by the "select" and
2191 "poll" pollers.
2192
William Dauchy3894d972019-12-28 15:36:02 +01002193 This option is automatically disabled on old processes in the context of
2194 seamless reload; it avoids too much cpu conflicts when multiple processes
2195 stay around for some time waiting for the end of their current connections.
2196
Willy Tarreau1746eec2014-04-25 10:46:47 +02002197max-spread-checks <delay in milliseconds>
Daniel Corbett9f0843f2021-05-08 10:50:37 -04002198 By default, HAProxy tries to spread the start of health checks across the
Willy Tarreau1746eec2014-04-25 10:46:47 +02002199 smallest health check interval of all the servers in a farm. The principle is
2200 to avoid hammering services running on the same server. But when using large
2201 check intervals (10 seconds or more), the last servers in the farm take some
2202 time before starting to be tested, which can be a problem. This parameter is
2203 used to enforce an upper bound on delay between the first and the last check,
2204 even if the servers' check intervals are larger. When servers run with
2205 shorter intervals, their intervals will be respected though.
2206
Willy Tarreau6a06a402007-07-15 20:15:28 +02002207maxconn <number>
2208 Sets the maximum per-process number of concurrent connections to <number>. It
2209 is equivalent to the command-line argument "-n". Proxies will stop accepting
2210 connections when this limit is reached. The "ulimit-n" parameter is
Willy Tarreau8274e102014-06-19 15:31:25 +02002211 automatically adjusted according to this value. See also "ulimit-n". Note:
2212 the "select" poller cannot reliably use more than 1024 file descriptors on
2213 some platforms. If your platform only supports select and reports "select
2214 FAILED" on startup, you need to reduce maxconn until it works (slightly
Willy Tarreaub28f3442019-03-04 08:13:43 +01002215 below 500 in general). If this value is not set, it will automatically be
2216 calculated based on the current file descriptors limit reported by the
2217 "ulimit -n" command, possibly reduced to a lower value if a memory limit
2218 is enforced, based on the buffer size, memory allocated to compression, SSL
2219 cache size, and use or not of SSL and the associated maxsslconn (which can
2220 also be automatic).
Willy Tarreau6a06a402007-07-15 20:15:28 +02002221
Willy Tarreau81c25d02011-09-07 15:17:21 +02002222maxconnrate <number>
2223 Sets the maximum per-process number of connections per second to <number>.
2224 Proxies will stop accepting connections when this limit is reached. It can be
2225 used to limit the global capacity regardless of each frontend capacity. It is
2226 important to note that this can only be used as a service protection measure,
2227 as there will not necessarily be a fair share between frontends when the
2228 limit is reached, so it's a good idea to also limit each frontend to some
2229 value close to its expected share. Also, lowering tune.maxaccept can improve
2230 fairness.
2231
William Lallemandd85f9172012-11-09 17:05:39 +01002232maxcomprate <number>
2233 Sets the maximum per-process input compression rate to <number> kilobytes
Davor Ocelice9ed2812017-12-25 17:49:28 +01002234 per second. For each session, if the maximum is reached, the compression
William Lallemandd85f9172012-11-09 17:05:39 +01002235 level will be decreased during the session. If the maximum is reached at the
2236 beginning of a session, the session will not compress at all. If the maximum
2237 is not reached, the compression level will be increased up to
Davor Ocelice9ed2812017-12-25 17:49:28 +01002238 tune.comp.maxlevel. A value of zero means there is no limit, this is the
William Lallemandd85f9172012-11-09 17:05:39 +01002239 default value.
2240
William Lallemand072a2bf2012-11-20 17:01:01 +01002241maxcompcpuusage <number>
2242 Sets the maximum CPU usage HAProxy can reach before stopping the compression
2243 for new requests or decreasing the compression level of current requests.
2244 It works like 'maxcomprate' but measures CPU usage instead of incoming data
Daniel Corbett9f0843f2021-05-08 10:50:37 -04002245 bandwidth. The value is expressed in percent of the CPU used by HAProxy. In
William Lallemand072a2bf2012-11-20 17:01:01 +01002246 case of multiple processes (nbproc > 1), each process manages its individual
2247 usage. A value of 100 disable the limit. The default value is 100. Setting
2248 a lower value will prevent the compression work from slowing the whole
2249 process down and from introducing high latencies.
2250
Willy Tarreauff4f82d2009-02-06 11:28:13 +01002251maxpipes <number>
2252 Sets the maximum per-process number of pipes to <number>. Currently, pipes
2253 are only used by kernel-based tcp splicing. Since a pipe contains two file
2254 descriptors, the "ulimit-n" value will be increased accordingly. The default
2255 value is maxconn/4, which seems to be more than enough for most heavy usages.
2256 The splice code dynamically allocates and releases pipes, and can fall back
2257 to standard copy, so setting this value too low may only impact performance.
2258
Willy Tarreau93e7c002013-10-07 18:51:07 +02002259maxsessrate <number>
2260 Sets the maximum per-process number of sessions per second to <number>.
2261 Proxies will stop accepting connections when this limit is reached. It can be
2262 used to limit the global capacity regardless of each frontend capacity. It is
2263 important to note that this can only be used as a service protection measure,
2264 as there will not necessarily be a fair share between frontends when the
2265 limit is reached, so it's a good idea to also limit each frontend to some
2266 value close to its expected share. Also, lowering tune.maxaccept can improve
2267 fairness.
2268
Willy Tarreau403edff2012-09-06 11:58:37 +02002269maxsslconn <number>
2270 Sets the maximum per-process number of concurrent SSL connections to
2271 <number>. By default there is no SSL-specific limit, which means that the
2272 global maxconn setting will apply to all connections. Setting this limit
2273 avoids having openssl use too much memory and crash when malloc returns NULL
2274 (since it unfortunately does not reliably check for such conditions). Note
2275 that the limit applies both to incoming and outgoing connections, so one
2276 connection which is deciphered then ciphered accounts for 2 SSL connections.
Willy Tarreaud0256482015-01-15 21:45:22 +01002277 If this value is not set, but a memory limit is enforced, this value will be
2278 automatically computed based on the memory limit, maxconn, the buffer size,
2279 memory allocated to compression, SSL cache size, and use of SSL in either
2280 frontends, backends or both. If neither maxconn nor maxsslconn are specified
Daniel Corbett9f0843f2021-05-08 10:50:37 -04002281 when there is a memory limit, HAProxy will automatically adjust these values
Willy Tarreaud0256482015-01-15 21:45:22 +01002282 so that 100% of the connections can be made over SSL with no risk, and will
2283 consider the sides where it is enabled (frontend, backend, both).
Willy Tarreau403edff2012-09-06 11:58:37 +02002284
Willy Tarreaue43d5322013-10-07 20:01:52 +02002285maxsslrate <number>
2286 Sets the maximum per-process number of SSL sessions per second to <number>.
2287 SSL listeners will stop accepting connections when this limit is reached. It
2288 can be used to limit the global SSL CPU usage regardless of each frontend
2289 capacity. It is important to note that this can only be used as a service
2290 protection measure, as there will not necessarily be a fair share between
2291 frontends when the limit is reached, so it's a good idea to also limit each
2292 frontend to some value close to its expected share. It is also important to
2293 note that the sessions are accounted before they enter the SSL stack and not
2294 after, which also protects the stack against bad handshakes. Also, lowering
2295 tune.maxaccept can improve fairness.
2296
William Lallemand9d5f5482012-11-07 16:12:57 +01002297maxzlibmem <number>
2298 Sets the maximum amount of RAM in megabytes per process usable by the zlib.
2299 When the maximum amount is reached, future sessions will not compress as long
2300 as RAM is unavailable. When sets to 0, there is no limit.
William Lallemande3a7d992012-11-20 11:25:20 +01002301 The default value is 0. The value is available in bytes on the UNIX socket
2302 with "show info" on the line "MaxZlibMemUsage", the memory used by zlib is
2303 "ZlibMemUsage" in bytes.
2304
Willy Tarreau6a06a402007-07-15 20:15:28 +02002305noepoll
2306 Disables the use of the "epoll" event polling system on Linux. It is
2307 equivalent to the command-line argument "-de". The next polling system
Willy Tarreaue9f49e72012-11-11 17:42:00 +01002308 used will generally be "poll". See also "nopoll".
Willy Tarreau6a06a402007-07-15 20:15:28 +02002309
2310nokqueue
2311 Disables the use of the "kqueue" event polling system on BSD. It is
2312 equivalent to the command-line argument "-dk". The next polling system
2313 used will generally be "poll". See also "nopoll".
2314
Emmanuel Hocdet0ba4f482019-04-08 16:53:32 +00002315noevports
2316 Disables the use of the event ports event polling system on SunOS systems
2317 derived from Solaris 10 and later. It is equivalent to the command-line
2318 argument "-dv". The next polling system used will generally be "poll". See
2319 also "nopoll".
2320
Willy Tarreau6a06a402007-07-15 20:15:28 +02002321nopoll
2322 Disables the use of the "poll" event polling system. It is equivalent to the
2323 command-line argument "-dp". The next polling system used will be "select".
Willy Tarreau0ba27502007-12-24 16:55:16 +01002324 It should never be needed to disable "poll" since it's available on all
Emmanuel Hocdet0ba4f482019-04-08 16:53:32 +00002325 platforms supported by HAProxy. See also "nokqueue", "noepoll" and
2326 "noevports".
Willy Tarreau6a06a402007-07-15 20:15:28 +02002327
Willy Tarreauff4f82d2009-02-06 11:28:13 +01002328nosplice
2329 Disables the use of kernel tcp splicing between sockets on Linux. It is
Davor Ocelice9ed2812017-12-25 17:49:28 +01002330 equivalent to the command line argument "-dS". Data will then be copied
Willy Tarreauff4f82d2009-02-06 11:28:13 +01002331 using conventional and more portable recv/send calls. Kernel tcp splicing is
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01002332 limited to some very recent instances of kernel 2.6. Most versions between
Willy Tarreauff4f82d2009-02-06 11:28:13 +01002333 2.6.25 and 2.6.28 are buggy and will forward corrupted data, so they must not
2334 be used. This option makes it easier to globally disable kernel splicing in
2335 case of doubt. See also "option splice-auto", "option splice-request" and
2336 "option splice-response".
2337
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03002338nogetaddrinfo
2339 Disables the use of getaddrinfo(3) for name resolving. It is equivalent to
2340 the command line argument "-dG". Deprecated gethostbyname(3) will be used.
2341
Lukas Tribusa0bcbdc2016-09-12 21:42:20 +00002342noreuseport
2343 Disables the use of SO_REUSEPORT - see socket(7). It is equivalent to the
2344 command line argument "-dR".
2345
Willy Tarreauca3afc22021-05-05 18:33:19 +02002346profiling.memory { on | off }
2347 Enables ('on') or disables ('off') per-function memory profiling. This will
2348 keep usage statistics of malloc/calloc/realloc/free calls anywhere in the
2349 process (including libraries) which will be reported on the CLI using the
2350 "show profiling" command. This is essentially meant to be used when an
2351 abnormal memory usage is observed that cannot be explained by the pools and
2352 other info are required. The performance hit will typically be around 1%,
2353 maybe a bit more on highly threaded machines, so it is normally suitable for
2354 use in production. The same may be achieved at run time on the CLI using the
2355 "set profiling memory" command, please consult the management manual.
2356
Willy Tarreaud2d33482019-04-25 17:09:07 +02002357profiling.tasks { auto | on | off }
2358 Enables ('on') or disables ('off') per-task CPU profiling. When set to 'auto'
2359 the profiling automatically turns on a thread when it starts to suffer from
2360 an average latency of 1000 microseconds or higher as reported in the
2361 "avg_loop_us" activity field, and automatically turns off when the latency
John Roeslerfb2fce12019-07-10 15:45:51 -05002362 returns below 990 microseconds (this value is an average over the last 1024
Willy Tarreaud2d33482019-04-25 17:09:07 +02002363 loops so it does not vary quickly and tends to significantly smooth short
2364 spikes). It may also spontaneously trigger from time to time on overloaded
2365 systems, containers, or virtual machines, or when the system swaps (which
2366 must absolutely never happen on a load balancer).
2367
2368 CPU profiling per task can be very convenient to report where the time is
2369 spent and which requests have what effect on which other request. Enabling
2370 it will typically affect the overall's performance by less than 1%, thus it
2371 is recommended to leave it to the default 'auto' value so that it only
2372 operates when a problem is identified. This feature requires a system
Willy Tarreau75c62c22018-11-22 11:02:09 +01002373 supporting the clock_gettime(2) syscall with clock identifiers
2374 CLOCK_MONOTONIC and CLOCK_THREAD_CPUTIME_ID, otherwise the reported time will
2375 be zero. This option may be changed at run time using "set profiling" on the
2376 CLI.
2377
Willy Tarreaufe255b72007-10-14 23:09:26 +02002378spread-checks <0..50, in percent>
Simon Hormand60d6912013-11-25 10:46:36 +09002379 Sometimes it is desirable to avoid sending agent and health checks to
2380 servers at exact intervals, for instance when many logical servers are
2381 located on the same physical server. With the help of this parameter, it
2382 becomes possible to add some randomness in the check interval between 0
2383 and +/- 50%. A value between 2 and 5 seems to show good results. The
2384 default value remains at 0.
Willy Tarreaufe255b72007-10-14 23:09:26 +02002385
Davor Ocelice9ed2812017-12-25 17:49:28 +01002386ssl-engine <name> [algo <comma-separated list of algorithms>]
Grant Zhang872f9c22017-01-21 01:10:18 +00002387 Sets the OpenSSL engine to <name>. List of valid values for <name> may be
Davor Ocelice9ed2812017-12-25 17:49:28 +01002388 obtained using the command "openssl engine". This statement may be used
Grant Zhang872f9c22017-01-21 01:10:18 +00002389 multiple times, it will simply enable multiple crypto engines. Referencing an
Daniel Corbett9f0843f2021-05-08 10:50:37 -04002390 unsupported engine will prevent HAProxy from starting. Note that many engines
Grant Zhang872f9c22017-01-21 01:10:18 +00002391 will lead to lower HTTPS performance than pure software with recent
2392 processors. The optional command "algo" sets the default algorithms an ENGINE
2393 will supply using the OPENSSL function ENGINE_set_default_string(). A value
Davor Ocelice9ed2812017-12-25 17:49:28 +01002394 of "ALL" uses the engine for all cryptographic operations. If no list of
2395 algo is specified then the value of "ALL" is used. A comma-separated list
Grant Zhang872f9c22017-01-21 01:10:18 +00002396 of different algorithms may be specified, including: RSA, DSA, DH, EC, RAND,
2397 CIPHERS, DIGESTS, PKEY, PKEY_CRYPTO, PKEY_ASN1. This is the same format that
2398 openssl configuration file uses:
2399 https://www.openssl.org/docs/man1.0.2/apps/config.html
2400
Grant Zhangfa6c7ee2017-01-14 01:42:15 +00002401ssl-mode-async
2402 Adds SSL_MODE_ASYNC mode to the SSL context. This enables asynchronous TLS
Emeric Brun3854e012017-05-17 20:42:48 +02002403 I/O operations if asynchronous capable SSL engines are used. The current
Emeric Brunb5e42a82017-06-06 12:35:14 +00002404 implementation supports a maximum of 32 engines. The Openssl ASYNC API
2405 doesn't support moving read/write buffers and is not compliant with
Daniel Corbett9f0843f2021-05-08 10:50:37 -04002406 HAProxy's buffer management. So the asynchronous mode is disabled on
John Roeslerfb2fce12019-07-10 15:45:51 -05002407 read/write operations (it is only enabled during initial and renegotiation
Emeric Brunb5e42a82017-06-06 12:35:14 +00002408 handshakes).
Grant Zhangfa6c7ee2017-01-14 01:42:15 +00002409
Willy Tarreau33cb0652014-12-23 22:52:37 +01002410tune.buffers.limit <number>
2411 Sets a hard limit on the number of buffers which may be allocated per process.
2412 The default value is zero which means unlimited. The minimum non-zero value
2413 will always be greater than "tune.buffers.reserve" and should ideally always
2414 be about twice as large. Forcing this value can be particularly useful to
2415 limit the amount of memory a process may take, while retaining a sane
Davor Ocelice9ed2812017-12-25 17:49:28 +01002416 behavior. When this limit is reached, sessions which need a buffer wait for
Willy Tarreau33cb0652014-12-23 22:52:37 +01002417 another one to be released by another session. Since buffers are dynamically
2418 allocated and released, the waiting time is very short and not perceptible
2419 provided that limits remain reasonable. In fact sometimes reducing the limit
2420 may even increase performance by increasing the CPU cache's efficiency. Tests
2421 have shown good results on average HTTP traffic with a limit to 1/10 of the
2422 expected global maxconn setting, which also significantly reduces memory
2423 usage. The memory savings come from the fact that a number of connections
2424 will not allocate 2*tune.bufsize. It is best not to touch this value unless
Daniel Corbett9f0843f2021-05-08 10:50:37 -04002425 advised to do so by an HAProxy core developer.
Willy Tarreau33cb0652014-12-23 22:52:37 +01002426
Willy Tarreau1058ae72014-12-23 22:40:40 +01002427tune.buffers.reserve <number>
2428 Sets the number of buffers which are pre-allocated and reserved for use only
2429 during memory shortage conditions resulting in failed memory allocations. The
2430 minimum value is 2 and is also the default. There is no reason a user would
Daniel Corbett9f0843f2021-05-08 10:50:37 -04002431 want to change this value, it's mostly aimed at HAProxy core developers.
Willy Tarreau1058ae72014-12-23 22:40:40 +01002432
Willy Tarreau27a674e2009-08-17 07:23:33 +02002433tune.bufsize <number>
2434 Sets the buffer size to this size (in bytes). Lower values allow more
2435 sessions to coexist in the same amount of RAM, and higher values allow some
2436 applications with very large cookies to work. The default value is 16384 and
2437 can be changed at build time. It is strongly recommended not to change this
2438 from the default value, as very low values will break some services such as
2439 statistics, and values larger than default size will increase memory usage,
2440 possibly causing the system to run out of memory. At least the global maxconn
Willy Tarreau45a66cc2017-11-24 11:28:00 +01002441 parameter should be decreased by the same factor as this one is increased. In
2442 addition, use of HTTP/2 mandates that this value must be 16384 or more. If an
Daniel Corbett9f0843f2021-05-08 10:50:37 -04002443 HTTP request is larger than (tune.bufsize - tune.maxrewrite), HAProxy will
Dmitry Sivachenkof6f4f7b2012-10-21 18:10:25 +04002444 return HTTP 400 (Bad Request) error. Similarly if an HTTP response is larger
Daniel Corbett9f0843f2021-05-08 10:50:37 -04002445 than this size, HAProxy will return HTTP 502 (Bad Gateway). Note that the
Willy Tarreauc77d3642018-12-12 06:19:42 +01002446 value set using this parameter will automatically be rounded up to the next
2447 multiple of 8 on 32-bit machines and 16 on 64-bit machines.
Willy Tarreau27a674e2009-08-17 07:23:33 +02002448
Christopher Fauletbb9fb8b2020-11-25 17:20:57 +01002449tune.chksize <number> (deprecated)
2450 This option is deprecated and ignored.
Willy Tarreau43961d52010-10-04 20:39:20 +02002451
William Lallemandf3747832012-11-09 12:33:10 +01002452tune.comp.maxlevel <number>
2453 Sets the maximum compression level. The compression level affects CPU
2454 usage during compression. This value affects CPU usage during compression.
2455 Each session using compression initializes the compression algorithm with
2456 this value. The default value is 1.
2457
Willy Tarreauc299e1e2019-02-27 11:35:12 +01002458tune.fail-alloc
2459 If compiled with DEBUG_FAIL_ALLOC, gives the percentage of chances an
2460 allocation attempt fails. Must be between 0 (no failure) and 100 (no
2461 success). This is useful to debug and make sure memory failures are handled
2462 gracefully.
2463
Willy Tarreaubc52bec2020-06-18 08:58:47 +02002464tune.fd.edge-triggered { on | off } [ EXPERIMENTAL ]
2465 Enables ('on') or disables ('off') the edge-triggered polling mode for FDs
2466 that support it. This is currently only support with epoll. It may noticeably
2467 reduce the number of epoll_ctl() calls and slightly improve performance in
2468 certain scenarios. This is still experimental, it may result in frozen
2469 connections if bugs are still present, and is disabled by default.
2470
Willy Tarreaufe20e5b2017-07-27 11:42:14 +02002471tune.h2.header-table-size <number>
2472 Sets the HTTP/2 dynamic header table size. It defaults to 4096 bytes and
2473 cannot be larger than 65536 bytes. A larger value may help certain clients
2474 send more compact requests, depending on their capabilities. This amount of
2475 memory is consumed for each HTTP/2 connection. It is recommended not to
2476 change it.
2477
Willy Tarreaue6baec02017-07-27 11:45:11 +02002478tune.h2.initial-window-size <number>
2479 Sets the HTTP/2 initial window size, which is the number of bytes the client
Daniel Corbett9f0843f2021-05-08 10:50:37 -04002480 can upload before waiting for an acknowledgment from HAProxy. This setting
Davor Ocelice9ed2812017-12-25 17:49:28 +01002481 only affects payload contents (i.e. the body of POST requests), not headers.
Willy Tarreaue6baec02017-07-27 11:45:11 +02002482 The default value is 65535, which roughly allows up to 5 Mbps of upload
2483 bandwidth per client over a network showing a 100 ms ping time, or 500 Mbps
2484 over a 1-ms local network. It can make sense to increase this value to allow
2485 faster uploads, or to reduce it to increase fairness when dealing with many
2486 clients. It doesn't affect resource usage.
2487
Willy Tarreau5242ef82017-07-27 11:47:28 +02002488tune.h2.max-concurrent-streams <number>
2489 Sets the HTTP/2 maximum number of concurrent streams per connection (ie the
2490 number of outstanding requests on a single connection). The default value is
2491 100. A larger one may slightly improve page load time for complex sites when
2492 visited over high latency networks, but increases the amount of resources a
2493 single client may allocate. A value of zero disables the limit so a single
Daniel Corbett9f0843f2021-05-08 10:50:37 -04002494 client may create as many streams as allocatable by HAProxy. It is highly
Willy Tarreau5242ef82017-07-27 11:47:28 +02002495 recommended not to change this value.
2496
Willy Tarreaua24b35c2019-02-21 13:24:36 +01002497tune.h2.max-frame-size <number>
Daniel Corbett9f0843f2021-05-08 10:50:37 -04002498 Sets the HTTP/2 maximum frame size that HAProxy announces it is willing to
Willy Tarreaua24b35c2019-02-21 13:24:36 +01002499 receive to its peers. The default value is the largest between 16384 and the
Daniel Corbett9f0843f2021-05-08 10:50:37 -04002500 buffer size (tune.bufsize). In any case, HAProxy will not announce support
Willy Tarreaua24b35c2019-02-21 13:24:36 +01002501 for frame sizes larger than buffers. The main purpose of this setting is to
2502 allow to limit the maximum frame size setting when using large buffers. Too
2503 large frame sizes might have performance impact or cause some peers to
2504 misbehave. It is highly recommended not to change this value.
2505
Willy Tarreau193b8c62012-11-22 00:17:38 +01002506tune.http.cookielen <number>
2507 Sets the maximum length of captured cookies. This is the maximum value that
2508 the "capture cookie xxx len yyy" will be allowed to take, and any upper value
2509 will automatically be truncated to this one. It is important not to set too
2510 high a value because all cookie captures still allocate this size whatever
2511 their configured value (they share a same pool). This value is per request
2512 per response, so the memory allocated is twice this value per connection.
2513 When not specified, the limit is set to 63 characters. It is recommended not
2514 to change this value.
2515
Stéphane Cottin23e9e932017-05-18 08:58:41 +02002516tune.http.logurilen <number>
Davor Ocelice9ed2812017-12-25 17:49:28 +01002517 Sets the maximum length of request URI in logs. This prevents truncating long
2518 request URIs with valuable query strings in log lines. This is not related
Stéphane Cottin23e9e932017-05-18 08:58:41 +02002519 to syslog limits. If you increase this limit, you may also increase the
Davor Ocelice9ed2812017-12-25 17:49:28 +01002520 'log ... len yyy' parameter. Your syslog daemon may also need specific
Stéphane Cottin23e9e932017-05-18 08:58:41 +02002521 configuration directives too.
2522 The default value is 1024.
2523
Willy Tarreauac1932d2011-10-24 19:14:41 +02002524tune.http.maxhdr <number>
2525 Sets the maximum number of headers in a request. When a request comes with a
2526 number of headers greater than this value (including the first line), it is
2527 rejected with a "400 Bad Request" status code. Similarly, too large responses
2528 are blocked with "502 Bad Gateway". The default value is 101, which is enough
2529 for all usages, considering that the widely deployed Apache server uses the
2530 same limit. It can be useful to push this limit further to temporarily allow
Christopher Faulet50174f32017-06-21 16:31:35 +02002531 a buggy application to work by the time it gets fixed. The accepted range is
2532 1..32767. Keep in mind that each new header consumes 32bits of memory for
2533 each session, so don't push this limit too high.
Willy Tarreauac1932d2011-10-24 19:14:41 +02002534
Willy Tarreau76cc6992020-07-01 18:49:24 +02002535tune.idle-pool.shared { on | off }
2536 Enables ('on') or disables ('off') sharing of idle connection pools between
2537 threads for a same server. The default is to share them between threads in
2538 order to minimize the number of persistent connections to a server, and to
2539 optimize the connection reuse rate. But to help with debugging or when
2540 suspecting a bug in HAProxy around connection reuse, it can be convenient to
2541 forcefully disable this idle pool sharing between multiple threads, and force
Willy Tarreau0784db82021-02-19 11:45:22 +01002542 this option to "off". The default is on. It is strongly recommended against
2543 disabling this option without setting a conservative value on "pool-low-conn"
2544 for all servers relying on connection reuse to achieve a high performance
2545 level, otherwise connections might be closed very often as the thread count
2546 increases.
Willy Tarreau76cc6992020-07-01 18:49:24 +02002547
Willy Tarreau7e312732014-02-12 16:35:14 +01002548tune.idletimer <timeout>
Daniel Corbett9f0843f2021-05-08 10:50:37 -04002549 Sets the duration after which HAProxy will consider that an empty buffer is
Willy Tarreau7e312732014-02-12 16:35:14 +01002550 probably associated with an idle stream. This is used to optimally adjust
2551 some packet sizes while forwarding large and small data alternatively. The
2552 decision to use splice() or to send large buffers in SSL is modulated by this
2553 parameter. The value is in milliseconds between 0 and 65535. A value of zero
Daniel Corbett9f0843f2021-05-08 10:50:37 -04002554 means that HAProxy will not try to detect idle streams. The default is 1000,
Davor Ocelice9ed2812017-12-25 17:49:28 +01002555 which seems to correctly detect end user pauses (e.g. read a page before
John Roeslerfb2fce12019-07-10 15:45:51 -05002556 clicking). There should be no reason for changing this value. Please check
Willy Tarreau7e312732014-02-12 16:35:14 +01002557 tune.ssl.maxrecord below.
2558
Willy Tarreau7ac908b2019-02-27 12:02:18 +01002559tune.listener.multi-queue { on | off }
2560 Enables ('on') or disables ('off') the listener's multi-queue accept which
2561 spreads the incoming traffic to all threads a "bind" line is allowed to run
2562 on instead of taking them for itself. This provides a smoother traffic
2563 distribution and scales much better, especially in environments where threads
2564 may be unevenly loaded due to external activity (network interrupts colliding
2565 with one thread for example). This option is enabled by default, but it may
2566 be forcefully disabled for troubleshooting or for situations where it is
2567 estimated that the operating system already provides a good enough
2568 distribution and connections are extremely short-lived.
2569
Thierry FOURNIER90da1912015-03-05 11:17:06 +01002570tune.lua.forced-yield <number>
2571 This directive forces the Lua engine to execute a yield each <number> of
Tim Düsterhus4896c442016-11-29 02:15:19 +01002572 instructions executed. This permits interrupting a long script and allows the
Thierry FOURNIER90da1912015-03-05 11:17:06 +01002573 HAProxy scheduler to process other tasks like accepting connections or
2574 forwarding traffic. The default value is 10000 instructions. If HAProxy often
Davor Ocelice9ed2812017-12-25 17:49:28 +01002575 executes some Lua code but more responsiveness is required, this value can be
Thierry FOURNIER90da1912015-03-05 11:17:06 +01002576 lowered. If the Lua code is quite long and its result is absolutely required
2577 to process the data, the <number> can be increased.
2578
Willy Tarreau32f61e22015-03-18 17:54:59 +01002579tune.lua.maxmem
2580 Sets the maximum amount of RAM in megabytes per process usable by Lua. By
2581 default it is zero which means unlimited. It is important to set a limit to
2582 ensure that a bug in a script will not result in the system running out of
2583 memory.
2584
Thierry FOURNIER90da1912015-03-05 11:17:06 +01002585tune.lua.session-timeout <timeout>
2586 This is the execution timeout for the Lua sessions. This is useful for
Thierry FOURNIER7dd784b2015-10-01 14:49:33 +02002587 preventing infinite loops or spending too much time in Lua. This timeout
2588 counts only the pure Lua runtime. If the Lua does a sleep, the sleep is
Davor Ocelice9ed2812017-12-25 17:49:28 +01002589 not taken in account. The default timeout is 4s.
Thierry FOURNIER90da1912015-03-05 11:17:06 +01002590
2591tune.lua.task-timeout <timeout>
2592 Purpose is the same as "tune.lua.session-timeout", but this timeout is
2593 dedicated to the tasks. By default, this timeout isn't set because a task may
2594 remain alive during of the lifetime of HAProxy. For example, a task used to
2595 check servers.
2596
Thierry FOURNIER7dd784b2015-10-01 14:49:33 +02002597tune.lua.service-timeout <timeout>
2598 This is the execution timeout for the Lua services. This is useful for
2599 preventing infinite loops or spending too much time in Lua. This timeout
2600 counts only the pure Lua runtime. If the Lua does a sleep, the sleep is
Davor Ocelice9ed2812017-12-25 17:49:28 +01002601 not taken in account. The default timeout is 4s.
Thierry FOURNIER7dd784b2015-10-01 14:49:33 +02002602
Willy Tarreaua0250ba2008-01-06 11:22:57 +01002603tune.maxaccept <number>
Willy Tarreau16a21472012-11-19 12:39:59 +01002604 Sets the maximum number of consecutive connections a process may accept in a
2605 row before switching to other work. In single process mode, higher numbers
Willy Tarreau66161322021-02-19 15:50:27 +01002606 used to give better performance at high connection rates, though this is not
2607 the case anymore with the multi-queue. This value applies individually to
2608 each listener, so that the number of processes a listener is bound to is
2609 taken into account. This value defaults to 4 which showed best results. If a
2610 significantly higher value was inherited from an ancient config, it might be
2611 worth removing it as it will both increase performance and lower response
2612 time. In multi-process mode, it is divided by twice the number of processes
2613 the listener is bound to. Setting this value to -1 completely disables the
2614 limitation. It should normally not be needed to tweak this value.
Willy Tarreaua0250ba2008-01-06 11:22:57 +01002615
2616tune.maxpollevents <number>
2617 Sets the maximum amount of events that can be processed at once in a call to
2618 the polling system. The default value is adapted to the operating system. It
2619 has been noticed that reducing it below 200 tends to slightly decrease
2620 latency at the expense of network bandwidth, and increasing it above 200
2621 tends to trade latency for slightly increased bandwidth.
2622
Willy Tarreau27a674e2009-08-17 07:23:33 +02002623tune.maxrewrite <number>
2624 Sets the reserved buffer space to this size in bytes. The reserved space is
2625 used for header rewriting or appending. The first reads on sockets will never
2626 fill more than bufsize-maxrewrite. Historically it has defaulted to half of
2627 bufsize, though that does not make much sense since there are rarely large
2628 numbers of headers to add. Setting it too high prevents processing of large
2629 requests or responses. Setting it too low prevents addition of new headers
2630 to already large requests or to POST requests. It is generally wise to set it
2631 to about 1024. It is automatically readjusted to half of bufsize if it is
2632 larger than that. This means you don't have to worry about it when changing
2633 bufsize.
2634
Willy Tarreauf3045d22015-04-29 16:24:50 +02002635tune.pattern.cache-size <number>
2636 Sets the size of the pattern lookup cache to <number> entries. This is an LRU
2637 cache which reminds previous lookups and their results. It is used by ACLs
2638 and maps on slow pattern lookups, namely the ones using the "sub", "reg",
2639 "dir", "dom", "end", "bin" match methods as well as the case-insensitive
2640 strings. It applies to pattern expressions which means that it will be able
2641 to memorize the result of a lookup among all the patterns specified on a
2642 configuration line (including all those loaded from files). It automatically
2643 invalidates entries which are updated using HTTP actions or on the CLI. The
2644 default cache size is set to 10000 entries, which limits its footprint to
Willy Tarreau403bfbb2019-10-23 06:59:31 +02002645 about 5 MB per process/thread on 32-bit systems and 8 MB per process/thread
2646 on 64-bit systems, as caches are thread/process local. There is a very low
Willy Tarreauf3045d22015-04-29 16:24:50 +02002647 risk of collision in this cache, which is in the order of the size of the
2648 cache divided by 2^64. Typically, at 10000 requests per second with the
2649 default cache size of 10000 entries, there's 1% chance that a brute force
2650 attack could cause a single collision after 60 years, or 0.1% after 6 years.
2651 This is considered much lower than the risk of a memory corruption caused by
2652 aging components. If this is not acceptable, the cache can be disabled by
2653 setting this parameter to 0.
2654
Willy Tarreaubd9a0a72011-10-23 21:14:29 +02002655tune.pipesize <number>
2656 Sets the kernel pipe buffer size to this size (in bytes). By default, pipes
2657 are the default size for the system. But sometimes when using TCP splicing,
2658 it can improve performance to increase pipe sizes, especially if it is
2659 suspected that pipes are not filled and that many calls to splice() are
2660 performed. This has an impact on the kernel's memory footprint, so this must
2661 not be changed if impacts are not understood.
2662
Olivier Houchard88698d92019-04-16 19:07:22 +02002663tune.pool-high-fd-ratio <number>
2664 This setting sets the max number of file descriptors (in percentage) used by
Daniel Corbett9f0843f2021-05-08 10:50:37 -04002665 HAProxy globally against the maximum number of file descriptors HAProxy can
Olivier Houchard88698d92019-04-16 19:07:22 +02002666 use before we start killing idle connections when we can't reuse a connection
2667 and we have to create a new one. The default is 25 (one quarter of the file
2668 descriptor will mean that roughly half of the maximum front connections can
2669 keep an idle connection behind, anything beyond this probably doesn't make
John Roeslerfb2fce12019-07-10 15:45:51 -05002670 much sense in the general case when targeting connection reuse).
Olivier Houchard88698d92019-04-16 19:07:22 +02002671
Willy Tarreau83ca3052020-07-01 18:30:16 +02002672tune.pool-low-fd-ratio <number>
2673 This setting sets the max number of file descriptors (in percentage) used by
Daniel Corbett9f0843f2021-05-08 10:50:37 -04002674 HAProxy globally against the maximum number of file descriptors HAProxy can
Willy Tarreau83ca3052020-07-01 18:30:16 +02002675 use before we stop putting connection into the idle pool for reuse. The
2676 default is 20.
2677
Willy Tarreaue803de22010-01-21 17:43:04 +01002678tune.rcvbuf.client <number>
2679tune.rcvbuf.server <number>
2680 Forces the kernel socket receive buffer size on the client or the server side
2681 to the specified value in bytes. This value applies to all TCP/HTTP frontends
2682 and backends. It should normally never be set, and the default size (0) lets
John Roeslerfb2fce12019-07-10 15:45:51 -05002683 the kernel auto-tune this value depending on the amount of available memory.
Davor Ocelice9ed2812017-12-25 17:49:28 +01002684 However it can sometimes help to set it to very low values (e.g. 4096) in
Willy Tarreaue803de22010-01-21 17:43:04 +01002685 order to save kernel memory by preventing it from buffering too large amounts
2686 of received data. Lower values will significantly increase CPU usage though.
2687
Willy Tarreaub22fc302015-12-14 12:04:35 +01002688tune.recv_enough <number>
Davor Ocelice9ed2812017-12-25 17:49:28 +01002689 HAProxy uses some hints to detect that a short read indicates the end of the
Willy Tarreaub22fc302015-12-14 12:04:35 +01002690 socket buffers. One of them is that a read returns more than <recv_enough>
2691 bytes, which defaults to 10136 (7 segments of 1448 each). This default value
2692 may be changed by this setting to better deal with workloads involving lots
2693 of short messages such as telnet or SSH sessions.
2694
Olivier Houchard1599b802018-05-24 18:59:04 +02002695tune.runqueue-depth <number>
John Roeslerfb2fce12019-07-10 15:45:51 -05002696 Sets the maximum amount of task that can be processed at once when running
Willy Tarreau060a7612021-03-10 11:06:26 +01002697 tasks. The default value depends on the number of threads but sits between 35
2698 and 280, which tend to show the highest request rates and lowest latencies.
2699 Increasing it may incur latency when dealing with I/Os, making it too small
2700 can incur extra overhead. Higher thread counts benefit from lower values.
2701 When experimenting with much larger values, it may be useful to also enable
2702 tune.sched.low-latency and possibly tune.fd.edge-triggered to limit the
2703 maximum latency to the lowest possible.
Willy Tarreaue7723bd2020-06-24 11:11:02 +02002704
2705tune.sched.low-latency { on | off }
2706 Enables ('on') or disables ('off') the low-latency task scheduler. By default
Daniel Corbett9f0843f2021-05-08 10:50:37 -04002707 HAProxy processes tasks from several classes one class at a time as this is
Willy Tarreaue7723bd2020-06-24 11:11:02 +02002708 the most efficient. But when running with large values of tune.runqueue-depth
2709 this can have a measurable effect on request or connection latency. When this
2710 low-latency setting is enabled, tasks of lower priority classes will always
2711 be executed before other ones if they exist. This will permit to lower the
2712 maximum latency experienced by new requests or connections in the middle of
2713 massive traffic, at the expense of a higher impact on this large traffic.
2714 For regular usage it is better to leave this off. The default value is off.
Olivier Houchard1599b802018-05-24 18:59:04 +02002715
Willy Tarreaue803de22010-01-21 17:43:04 +01002716tune.sndbuf.client <number>
2717tune.sndbuf.server <number>
2718 Forces the kernel socket send buffer size on the client or the server side to
2719 the specified value in bytes. This value applies to all TCP/HTTP frontends
2720 and backends. It should normally never be set, and the default size (0) lets
John Roeslerfb2fce12019-07-10 15:45:51 -05002721 the kernel auto-tune this value depending on the amount of available memory.
Davor Ocelice9ed2812017-12-25 17:49:28 +01002722 However it can sometimes help to set it to very low values (e.g. 4096) in
Willy Tarreaue803de22010-01-21 17:43:04 +01002723 order to save kernel memory by preventing it from buffering too large amounts
2724 of received data. Lower values will significantly increase CPU usage though.
2725 Another use case is to prevent write timeouts with extremely slow clients due
2726 to the kernel waiting for a large part of the buffer to be read before
Daniel Corbett9f0843f2021-05-08 10:50:37 -04002727 notifying HAProxy again.
Willy Tarreaue803de22010-01-21 17:43:04 +01002728
Willy Tarreau6ec58db2012-11-16 16:32:15 +01002729tune.ssl.cachesize <number>
Emeric Brunaf9619d2012-11-28 18:47:52 +01002730 Sets the size of the global SSL session cache, in a number of blocks. A block
William Dauchy9a4bbfe2021-02-12 15:58:46 +01002731 is large enough to contain an encoded session without peer certificate. An
2732 encoded session with peer certificate is stored in multiple blocks depending
2733 on the size of the peer certificate. A block uses approximately 200 bytes of
2734 memory (based on `sizeof(struct sh_ssl_sess_hdr) + SHSESS_BLOCK_MIN_SIZE`
2735 calculation used for `shctx_init` function). The default value may be forced
2736 at build time, otherwise defaults to 20000. When the cache is full, the most
2737 idle entries are purged and reassigned. Higher values reduce the occurrence
2738 of such a purge, hence the number of CPU-intensive SSL handshakes by ensuring
2739 that all users keep their session as long as possible. All entries are
2740 pre-allocated upon startup and are shared between all processes if "nbproc"
2741 is greater than 1. Setting this value to 0 disables the SSL session cache.
Willy Tarreau6ec58db2012-11-16 16:32:15 +01002742
Emeric Brun8dc60392014-05-09 13:52:00 +02002743tune.ssl.force-private-cache
Lukas Tribus27935782018-10-01 02:00:16 +02002744 This option disables SSL session cache sharing between all processes. It
Emeric Brun8dc60392014-05-09 13:52:00 +02002745 should normally not be used since it will force many renegotiations due to
2746 clients hitting a random process. But it may be required on some operating
2747 systems where none of the SSL cache synchronization method may be used. In
2748 this case, adding a first layer of hash-based load balancing before the SSL
2749 layer might limit the impact of the lack of session sharing.
2750
William Lallemand7d42ef52020-07-06 11:41:30 +02002751tune.ssl.keylog { on | off }
2752 This option activates the logging of the TLS keys. It should be used with
2753 care as it will consume more memory per SSL session and could decrease
2754 performances. This is disabled by default.
2755
2756 These sample fetches should be used to generate the SSLKEYLOGFILE that is
2757 required to decipher traffic with wireshark.
2758
2759 https://developer.mozilla.org/en-US/docs/Mozilla/Projects/NSS/Key_Log_Format
2760
2761 The SSLKEYLOG is a series of lines which are formatted this way:
2762
2763 <Label> <space> <ClientRandom> <space> <Secret>
2764
2765 The ClientRandom is provided by the %[ssl_fc_client_random,hex] sample
2766 fetch, the secret and the Label could be find in the array below. You need
2767 to generate a SSLKEYLOGFILE with all the labels in this array.
2768
2769 The following sample fetches are hexadecimal strings and does not need to be
2770 converted.
2771
2772 SSLKEYLOGFILE Label | Sample fetches for the Secrets
2773 --------------------------------|-----------------------------------------
2774 CLIENT_EARLY_TRAFFIC_SECRET | %[ssl_fc_client_early_traffic_secret]
2775 CLIENT_HANDSHAKE_TRAFFIC_SECRET | %[ssl_fc_client_handshake_traffic_secret]
2776 SERVER_HANDSHAKE_TRAFFIC_SECRET | %[ssl_fc_server_handshake_traffic_secret]
2777 CLIENT_TRAFFIC_SECRET_0 | %[ssl_fc_client_traffic_secret_0]
2778 SERVER_TRAFFIC_SECRET_0 | %[ssl_fc_server_traffic_secret_0]
William Lallemandd742b6c2020-07-07 10:14:56 +02002779 EXPORTER_SECRET | %[ssl_fc_exporter_secret]
2780 EARLY_EXPORTER_SECRET | %[ssl_fc_early_exporter_secret]
William Lallemand7d42ef52020-07-06 11:41:30 +02002781
2782 This is only available with OpenSSL 1.1.1, and useful with TLS1.3 session.
2783
2784 If you want to generate the content of a SSLKEYLOGFILE with TLS < 1.3, you
2785 only need this line:
2786
2787 "CLIENT_RANDOM %[ssl_fc_client_random,hex] %[ssl_fc_session_key,hex]"
2788
Emeric Brun4f65bff2012-11-16 15:11:00 +01002789tune.ssl.lifetime <timeout>
2790 Sets how long a cached SSL session may remain valid. This time is expressed
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03002791 in seconds and defaults to 300 (5 min). It is important to understand that it
Emeric Brun4f65bff2012-11-16 15:11:00 +01002792 does not guarantee that sessions will last that long, because if the cache is
2793 full, the longest idle sessions will be purged despite their configured
2794 lifetime. The real usefulness of this setting is to prevent sessions from
2795 being used for too long.
2796
Willy Tarreaubfd59462013-02-21 07:46:09 +01002797tune.ssl.maxrecord <number>
2798 Sets the maximum amount of bytes passed to SSL_write() at a time. Default
2799 value 0 means there is no limit. Over SSL/TLS, the client can decipher the
2800 data only once it has received a full record. With large records, it means
2801 that clients might have to download up to 16kB of data before starting to
2802 process them. Limiting the value can improve page load times on browsers
2803 located over high latency or low bandwidth networks. It is suggested to find
2804 optimal values which fit into 1 or 2 TCP segments (generally 1448 bytes over
2805 Ethernet with TCP timestamps enabled, or 1460 when timestamps are disabled),
2806 keeping in mind that SSL/TLS add some overhead. Typical values of 1419 and
2807 2859 gave good results during tests. Use "strace -e trace=write" to find the
Davor Ocelice9ed2812017-12-25 17:49:28 +01002808 best value. HAProxy will automatically switch to this setting after an idle
Willy Tarreau7e312732014-02-12 16:35:14 +01002809 stream has been detected (see tune.idletimer above).
Willy Tarreaubfd59462013-02-21 07:46:09 +01002810
Remi Gacognef46cd6e2014-06-12 14:58:40 +02002811tune.ssl.default-dh-param <number>
2812 Sets the maximum size of the Diffie-Hellman parameters used for generating
2813 the ephemeral/temporary Diffie-Hellman key in case of DHE key exchange. The
2814 final size will try to match the size of the server's RSA (or DSA) key (e.g,
2815 a 2048 bits temporary DH key for a 2048 bits RSA key), but will not exceed
Willy Tarreau3ba77d22020-05-08 09:31:18 +02002816 this maximum value. Default value if 2048. Only 1024 or higher values are
Remi Gacognef46cd6e2014-06-12 14:58:40 +02002817 allowed. Higher values will increase the CPU load, and values greater than
2818 1024 bits are not supported by Java 7 and earlier clients. This value is not
Remi Gacogne47783ef2015-05-29 15:53:22 +02002819 used if static Diffie-Hellman parameters are supplied either directly
2820 in the certificate file or by using the ssl-dh-param-file parameter.
Remi Gacognef46cd6e2014-06-12 14:58:40 +02002821
Christopher Faulet31af49d2015-06-09 17:29:50 +02002822tune.ssl.ssl-ctx-cache-size <number>
2823 Sets the size of the cache used to store generated certificates to <number>
2824 entries. This is a LRU cache. Because generating a SSL certificate
2825 dynamically is expensive, they are cached. The default cache size is set to
2826 1000 entries.
2827
Thierry FOURNIER5bf77322017-02-25 12:45:22 +01002828tune.ssl.capture-cipherlist-size <number>
2829 Sets the maximum size of the buffer used for capturing client-hello cipher
2830 list. If the value is 0 (default value) the capture is disabled, otherwise
2831 a buffer is allocated for each SSL/TLS connection.
2832
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02002833tune.vars.global-max-size <size>
Christopher Fauletff2613e2016-11-09 11:36:17 +01002834tune.vars.proc-max-size <size>
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02002835tune.vars.reqres-max-size <size>
2836tune.vars.sess-max-size <size>
2837tune.vars.txn-max-size <size>
Christopher Fauletff2613e2016-11-09 11:36:17 +01002838 These five tunes help to manage the maximum amount of memory used by the
2839 variables system. "global" limits the overall amount of memory available for
2840 all scopes. "proc" limits the memory for the process scope, "sess" limits the
2841 memory for the session scope, "txn" for the transaction scope, and "reqres"
2842 limits the memory for each request or response processing.
2843 Memory accounting is hierarchical, meaning more coarse grained limits include
2844 the finer grained ones: "proc" includes "sess", "sess" includes "txn", and
2845 "txn" includes "reqres".
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02002846
Daniel Schneller0b547052016-03-21 20:46:57 +01002847 For example, when "tune.vars.sess-max-size" is limited to 100,
2848 "tune.vars.txn-max-size" and "tune.vars.reqres-max-size" cannot exceed
2849 100 either. If we create a variable "txn.var" that contains 100 bytes,
2850 all available space is consumed.
2851 Notice that exceeding the limits at runtime will not result in an error
2852 message, but values might be cut off or corrupted. So make sure to accurately
2853 plan for the amount of space needed to store all your variables.
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02002854
William Lallemanda509e4c2012-11-07 16:54:34 +01002855tune.zlib.memlevel <number>
2856 Sets the memLevel parameter in zlib initialization for each session. It
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03002857 defines how much memory should be allocated for the internal compression
William Lallemanda509e4c2012-11-07 16:54:34 +01002858 state. A value of 1 uses minimum memory but is slow and reduces compression
Davor Ocelice9ed2812017-12-25 17:49:28 +01002859 ratio, a value of 9 uses maximum memory for optimal speed. Can be a value
William Lallemanda509e4c2012-11-07 16:54:34 +01002860 between 1 and 9. The default value is 8.
2861
2862tune.zlib.windowsize <number>
2863 Sets the window size (the size of the history buffer) as a parameter of the
2864 zlib initialization for each session. Larger values of this parameter result
Davor Ocelice9ed2812017-12-25 17:49:28 +01002865 in better compression at the expense of memory usage. Can be a value between
2866 8 and 15. The default value is 15.
Willy Tarreau6a06a402007-07-15 20:15:28 +02002867
Willy Tarreauc57f0e22009-05-10 13:12:33 +020028683.3. Debugging
2869--------------
Willy Tarreau6a06a402007-07-15 20:15:28 +02002870
Willy Tarreau6a06a402007-07-15 20:15:28 +02002871quiet
2872 Do not display any message during startup. It is equivalent to the command-
2873 line argument "-q".
2874
Willy Tarreau3eb10b82020-04-15 16:42:39 +02002875zero-warning
Daniel Corbett9f0843f2021-05-08 10:50:37 -04002876 When this option is set, HAProxy will refuse to start if any warning was
Willy Tarreau3eb10b82020-04-15 16:42:39 +02002877 emitted while processing the configuration. It is highly recommended to set
2878 this option on configurations that are not changed often, as it helps detect
2879 subtle mistakes and keep the configuration clean and forward-compatible. Note
2880 that "haproxy -c" will also report errors in such a case. This option is
2881 equivalent to command line argument "-dW".
2882
Emeric Brunf099e792010-09-27 12:05:28 +02002883
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +010028843.4. Userlists
2885--------------
2886It is possible to control access to frontend/backend/listen sections or to
2887http stats by allowing only authenticated and authorized users. To do this,
2888it is required to create at least one userlist and to define users.
2889
2890userlist <listname>
Cyril Bonté78caf842010-03-10 22:41:43 +01002891 Creates new userlist with name <listname>. Many independent userlists can be
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01002892 used to store authentication & authorization data for independent customers.
2893
2894group <groupname> [users <user>,<user>,(...)]
Cyril Bonté78caf842010-03-10 22:41:43 +01002895 Adds group <groupname> to the current userlist. It is also possible to
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01002896 attach users to this group by using a comma separated list of names
2897 proceeded by "users" keyword.
2898
Cyril Bontéf0c60612010-02-06 14:44:47 +01002899user <username> [password|insecure-password <password>]
2900 [groups <group>,<group>,(...)]
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01002901 Adds user <username> to the current userlist. Both secure (encrypted) and
2902 insecure (unencrypted) passwords can be used. Encrypted passwords are
Daniel Schnellerd06f31c2017-11-06 16:51:04 +01002903 evaluated using the crypt(3) function, so depending on the system's
2904 capabilities, different algorithms are supported. For example, modern Glibc
2905 based Linux systems support MD5, SHA-256, SHA-512, and, of course, the
2906 classic DES-based method of encrypting passwords.
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01002907
Daniel Schnellerd06f31c2017-11-06 16:51:04 +01002908 Attention: Be aware that using encrypted passwords might cause significantly
2909 increased CPU usage, depending on the number of requests, and the algorithm
2910 used. For any of the hashed variants, the password for each request must
2911 be processed through the chosen algorithm, before it can be compared to the
2912 value specified in the config file. Most current algorithms are deliberately
2913 designed to be expensive to compute to achieve resistance against brute
2914 force attacks. They do not simply salt/hash the clear text password once,
Daniel Corbett9f0843f2021-05-08 10:50:37 -04002915 but thousands of times. This can quickly become a major factor in HAProxy's
Daniel Schnellerd06f31c2017-11-06 16:51:04 +01002916 overall CPU consumption!
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01002917
2918 Example:
Cyril Bontéf0c60612010-02-06 14:44:47 +01002919 userlist L1
2920 group G1 users tiger,scott
2921 group G2 users xdb,scott
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01002922
Cyril Bontéf0c60612010-02-06 14:44:47 +01002923 user tiger password $6$k6y3o.eP$JlKBx9za9667qe4(...)xHSwRv6J.C0/D7cV91
2924 user scott insecure-password elgato
2925 user xdb insecure-password hello
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01002926
Cyril Bontéf0c60612010-02-06 14:44:47 +01002927 userlist L2
2928 group G1
2929 group G2
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01002930
Cyril Bontéf0c60612010-02-06 14:44:47 +01002931 user tiger password $6$k6y3o.eP$JlKBx(...)xHSwRv6J.C0/D7cV91 groups G1
2932 user scott insecure-password elgato groups G1,G2
2933 user xdb insecure-password hello groups G2
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01002934
2935 Please note that both lists are functionally identical.
Willy Tarreau6a06a402007-07-15 20:15:28 +02002936
Emeric Brunf099e792010-09-27 12:05:28 +02002937
29383.5. Peers
Cyril Bontédc4d9032012-04-08 21:57:39 +02002939----------
Emeric Brun94900952015-06-11 18:25:54 +02002940It is possible to propagate entries of any data-types in stick-tables between
Daniel Corbett9f0843f2021-05-08 10:50:37 -04002941several HAProxy instances over TCP connections in a multi-master fashion. Each
Emeric Brun94900952015-06-11 18:25:54 +02002942instance pushes its local updates and insertions to remote peers. The pushed
2943values overwrite remote ones without aggregation. Interrupted exchanges are
2944automatically detected and recovered from the last known point.
2945In addition, during a soft restart, the old process connects to the new one
2946using such a TCP connection to push all its entries before the new process
2947tries to connect to other peers. That ensures very fast replication during a
2948reload, it typically takes a fraction of a second even for large tables.
2949Note that Server IDs are used to identify servers remotely, so it is important
2950that configurations look similar or at least that the same IDs are forced on
2951each server on all participants.
Emeric Brunf099e792010-09-27 12:05:28 +02002952
2953peers <peersect>
Jamie Gloudon801a0a32012-08-25 00:18:33 -04002954 Creates a new peer list with name <peersect>. It is an independent section,
Emeric Brunf099e792010-09-27 12:05:28 +02002955 which is referenced by one or more stick-tables.
2956
Frédéric Lécaille2f167b32019-01-11 14:13:54 +01002957bind [<address>]:<port_range> [, ...] [param*]
2958 Defines the binding parameters of the local peer of this "peers" section.
2959 Such lines are not supported with "peer" line in the same "peers" section.
2960
Willy Tarreau77e4bd12015-05-01 20:02:17 +02002961disabled
2962 Disables a peers section. It disables both listening and any synchronization
2963 related to this section. This is provided to disable synchronization of stick
2964 tables without having to comment out all "peers" references.
2965
Frédéric Lécaille2f167b32019-01-11 14:13:54 +01002966default-bind [param*]
2967 Defines the binding parameters for the local peer, excepted its address.
2968
2969default-server [param*]
2970 Change default options for a server in a "peers" section.
2971
2972 Arguments:
2973 <param*> is a list of parameters for this server. The "default-server"
2974 keyword accepts an important number of options and has a complete
2975 section dedicated to it. Please refer to section 5 for more
2976 details.
2977
2978
2979 See also: "server" and section 5 about server options
2980
Emeric Brun7214dcf2021-09-29 10:29:52 +02002981enabled
2982 This re-enables a peers section which was previously disabled via the
2983 "disabled" keyword.
Willy Tarreau77e4bd12015-05-01 20:02:17 +02002984
Jan Wagner3e678602020-12-17 22:22:32 +01002985log <address> [len <length>] [format <format>] [sample <ranges>:<sample_size>]
Frédéric Lécailleb6f759b2019-11-05 09:57:45 +01002986 <facility> [<level> [<minlevel>]]
2987 "peers" sections support the same "log" keyword as for the proxies to
2988 log information about the "peers" listener. See "log" option for proxies for
2989 more details.
2990
Frédéric Lécaille2f167b32019-01-11 14:13:54 +01002991peer <peername> <ip>:<port> [param*]
Emeric Brunf099e792010-09-27 12:05:28 +02002992 Defines a peer inside a peers section.
2993 If <peername> is set to the local peer name (by default hostname, or forced
Dragan Dosen13cd54c2020-06-18 18:24:05 +02002994 using "-L" command line option or "localpeer" global configuration setting),
Daniel Corbett9f0843f2021-05-08 10:50:37 -04002995 HAProxy will listen for incoming remote peer connection on <ip>:<port>.
Dragan Dosen13cd54c2020-06-18 18:24:05 +02002996 Otherwise, <ip>:<port> defines where to connect to in order to join the
2997 remote peer, and <peername> is used at the protocol level to identify and
2998 validate the remote peer on the server side.
Emeric Brunf099e792010-09-27 12:05:28 +02002999
3000 During a soft restart, local peer <ip>:<port> is used by the old instance to
3001 connect the new one and initiate a complete replication (teaching process).
3002
3003 It is strongly recommended to have the exact same peers declaration on all
Dragan Dosen13cd54c2020-06-18 18:24:05 +02003004 peers and to only rely on the "-L" command line argument or the "localpeer"
3005 global configuration setting to change the local peer name. This makes it
3006 easier to maintain coherent configuration files across all peers.
Emeric Brunf099e792010-09-27 12:05:28 +02003007
William Lallemandb2f07452015-05-12 14:27:13 +02003008 You may want to reference some environment variables in the address
3009 parameter, see section 2.3 about environment variables.
Willy Tarreaudad36a32013-03-11 01:20:04 +01003010
Frédéric Lécaille2f167b32019-01-11 14:13:54 +01003011 Note: "peer" keyword may transparently be replaced by "server" keyword (see
3012 "server" keyword explanation below).
3013
3014server <peername> [<ip>:<port>] [param*]
Michael Prokop4438c602019-05-24 10:25:45 +02003015 As previously mentioned, "peer" keyword may be replaced by "server" keyword
Frédéric Lécaille2f167b32019-01-11 14:13:54 +01003016 with a support for all "server" parameters found in 5.2 paragraph.
3017 If the underlying peer is local, <ip>:<port> parameters must not be present.
3018 These parameters must be provided on a "bind" line (see "bind" keyword
3019 of this "peers" section).
3020 Some of these parameters are irrelevant for "peers" sections.
3021
3022
Cyril Bontédc4d9032012-04-08 21:57:39 +02003023 Example:
Frédéric Lécaille2f167b32019-01-11 14:13:54 +01003024 # The old way.
Emeric Brunf099e792010-09-27 12:05:28 +02003025 peers mypeers
Willy Tarreauf7b30a92010-12-06 22:59:17 +01003026 peer haproxy1 192.168.0.1:1024
3027 peer haproxy2 192.168.0.2:1024
3028 peer haproxy3 10.2.0.1:1024
Emeric Brunf099e792010-09-27 12:05:28 +02003029
3030 backend mybackend
3031 mode tcp
3032 balance roundrobin
3033 stick-table type ip size 20k peers mypeers
3034 stick on src
3035
Willy Tarreauf7b30a92010-12-06 22:59:17 +01003036 server srv1 192.168.0.30:80
3037 server srv2 192.168.0.31:80
Emeric Brunf099e792010-09-27 12:05:28 +02003038
Frédéric Lécaille2f167b32019-01-11 14:13:54 +01003039 Example:
3040 peers mypeers
3041 bind 127.0.0.11:10001 ssl crt mycerts/pem
3042 default-server ssl verify none
3043 server hostA 127.0.0.10:10000
3044 server hostB #local peer
Emeric Brunf099e792010-09-27 12:05:28 +02003045
Frédéric Lécaille4f5b77c2019-03-18 14:05:58 +01003046
3047table <tablename> type {ip | integer | string [len <length>] | binary [len <length>]}
3048 size <size> [expire <expire>] [nopurge] [store <data_type>]*
3049
3050 Configure a stickiness table for the current section. This line is parsed
3051 exactly the same way as the "stick-table" keyword in others section, except
John Roeslerfb2fce12019-07-10 15:45:51 -05003052 for the "peers" argument which is not required here and with an additional
Frédéric Lécaille4f5b77c2019-03-18 14:05:58 +01003053 mandatory first parameter to designate the stick-table. Contrary to others
3054 sections, there may be several "table" lines in "peers" sections (see also
3055 "stick-table" keyword).
3056
3057 Also be aware of the fact that "peers" sections have their own stick-table
3058 namespaces to avoid collisions between stick-table names identical in
3059 different "peers" section. This is internally handled prepending the "peers"
3060 sections names to the name of the stick-tables followed by a '/' character.
3061 If somewhere else in the configuration file you have to refer to such
3062 stick-tables declared in "peers" sections you must use the prefixed version
3063 of the stick-table name as follows:
3064
3065 peers mypeers
3066 peer A ...
3067 peer B ...
3068 table t1 ...
3069
3070 frontend fe1
3071 tcp-request content track-sc0 src table mypeers/t1
3072
3073 This is also this prefixed version of the stick-table names which must be
3074 used to refer to stick-tables through the CLI.
3075
3076 About "peers" protocol, as only "peers" belonging to the same section may
3077 communicate with each others, there is no need to do such a distinction.
3078 Several "peers" sections may declare stick-tables with the same name.
3079 This is shorter version of the stick-table name which is sent over the network.
3080 There is only a '/' character as prefix to avoid stick-table name collisions between
3081 stick-tables declared as backends and stick-table declared in "peers" sections
3082 as follows in this weird but supported configuration:
3083
3084 peers mypeers
3085 peer A ...
3086 peer B ...
3087 table t1 type string size 10m store gpc0
3088
3089 backend t1
3090 stick-table type string size 10m store gpc0 peers mypeers
3091
Daniel Corbett67a82712020-07-06 23:01:19 -04003092 Here "t1" table declared in "mypeers" section has "mypeers/t1" as global name.
Frédéric Lécaille4f5b77c2019-03-18 14:05:58 +01003093 "t1" table declared as a backend as "t1" as global name. But at peer protocol
3094 level the former table is named "/t1", the latter is again named "t1".
3095
Simon Horman51a1cf62015-02-03 13:00:44 +090030963.6. Mailers
3097------------
3098It is possible to send email alerts when the state of servers changes.
3099If configured email alerts are sent to each mailer that is configured
3100in a mailers section. Email is sent to mailers using SMTP.
3101
Pieter Baauw386a1272015-08-16 15:26:24 +02003102mailers <mailersect>
Simon Horman51a1cf62015-02-03 13:00:44 +09003103 Creates a new mailer list with the name <mailersect>. It is an
3104 independent section which is referenced by one or more proxies.
3105
3106mailer <mailername> <ip>:<port>
3107 Defines a mailer inside a mailers section.
3108
3109 Example:
3110 mailers mymailers
3111 mailer smtp1 192.168.0.1:587
3112 mailer smtp2 192.168.0.2:587
3113
3114 backend mybackend
3115 mode tcp
3116 balance roundrobin
3117
3118 email-alert mailers mymailers
3119 email-alert from test1@horms.org
3120 email-alert to test2@horms.org
3121
3122 server srv1 192.168.0.30:80
3123 server srv2 192.168.0.31:80
3124
Pieter Baauw235fcfc2016-02-13 15:33:40 +01003125timeout mail <time>
3126 Defines the time available for a mail/connection to be made and send to
3127 the mail-server. If not defined the default value is 10 seconds. To allow
3128 for at least two SYN-ACK packets to be send during initial TCP handshake it
3129 is advised to keep this value above 4 seconds.
3130
3131 Example:
3132 mailers mymailers
3133 timeout mail 20s
3134 mailer smtp1 192.168.0.1:587
Simon Horman51a1cf62015-02-03 13:00:44 +09003135
William Lallemandc9515522019-06-12 16:32:11 +020031363.7. Programs
3137-------------
3138In master-worker mode, it is possible to launch external binaries with the
3139master, these processes are called programs. These programs are launched and
3140managed the same way as the workers.
3141
3142During a reload of HAProxy, those processes are dealing with the same
3143sequence as a worker:
3144
3145 - the master is re-executed
3146 - the master sends a SIGUSR1 signal to the program
3147 - if "option start-on-reload" is not disabled, the master launches a new
3148 instance of the program
3149
3150During a stop, or restart, a SIGTERM is sent to the programs.
3151
3152program <name>
3153 This is a new program section, this section will create an instance <name>
3154 which is visible in "show proc" on the master CLI. (See "9.4. Master CLI" in
3155 the management guide).
3156
3157command <command> [arguments*]
3158 Define the command to start with optional arguments. The command is looked
3159 up in the current PATH if it does not include an absolute path. This is a
3160 mandatory option of the program section. Arguments containing spaces must
3161 be enclosed in quotes or double quotes or be prefixed by a backslash.
3162
Andrew Heberle97236962019-07-12 11:50:26 +08003163user <user name>
3164 Changes the executed command user ID to the <user name> from /etc/passwd.
3165 See also "group".
3166
3167group <group name>
3168 Changes the executed command group ID to the <group name> from /etc/group.
3169 See also "user".
3170
William Lallemandc9515522019-06-12 16:32:11 +02003171option start-on-reload
3172no option start-on-reload
3173 Start (or not) a new instance of the program upon a reload of the master.
3174 The default is to start a new instance. This option may only be used in a
3175 program section.
3176
3177
Christopher Faulet76edc0f2020-01-13 15:52:01 +010031783.8. HTTP-errors
3179----------------
3180
3181It is possible to globally declare several groups of HTTP errors, to be
3182imported afterwards in any proxy section. Same group may be referenced at
3183several places and can be fully or partially imported.
3184
3185http-errors <name>
3186 Create a new http-errors group with the name <name>. It is an independent
3187 section that may be referenced by one or more proxies using its name.
3188
3189errorfile <code> <file>
3190 Associate a file contents to an HTTP error code
3191
3192 Arguments :
3193 <code> is the HTTP status code. Currently, HAProxy is capable of
Christopher Faulet612f2ea2020-05-27 09:57:28 +02003194 generating codes 200, 400, 401, 403, 404, 405, 407, 408, 410,
Christopher Faulete095f312020-12-07 11:22:24 +01003195 425, 429, 500, 501, 502, 503, and 504.
Christopher Faulet76edc0f2020-01-13 15:52:01 +01003196
3197 <file> designates a file containing the full HTTP response. It is
3198 recommended to follow the common practice of appending ".http" to
3199 the filename so that people do not confuse the response with HTML
3200 error pages, and to use absolute paths, since files are read
3201 before any chroot is performed.
3202
3203 Please referrers to "errorfile" keyword in section 4 for details.
3204
3205 Example:
3206 http-errors website-1
3207 errorfile 400 /etc/haproxy/errorfiles/site1/400.http
3208 errorfile 404 /etc/haproxy/errorfiles/site1/404.http
3209 errorfile 408 /dev/null # work around Chrome pre-connect bug
3210
3211 http-errors website-2
3212 errorfile 400 /etc/haproxy/errorfiles/site2/400.http
3213 errorfile 404 /etc/haproxy/errorfiles/site2/404.http
3214 errorfile 408 /dev/null # work around Chrome pre-connect bug
3215
Emeric Brun99c453d2020-05-25 15:01:04 +020032163.9. Rings
3217----------
3218
3219It is possible to globally declare ring-buffers, to be used as target for log
3220servers or traces.
3221
3222ring <ringname>
3223 Creates a new ring-buffer with name <ringname>.
3224
3225description <text>
Daniel Corbett67a82712020-07-06 23:01:19 -04003226 The description is an optional description string of the ring. It will
Emeric Brun99c453d2020-05-25 15:01:04 +02003227 appear on CLI. By default, <name> is reused to fill this field.
3228
3229format <format>
3230 Format used to store events into the ring buffer.
3231
3232 Arguments:
3233 <format> is the log format used when generating syslog messages. It may be
3234 one of the following :
3235
3236 iso A message containing only the ISO date, followed by the text.
3237 The PID, process name and system name are omitted. This is
3238 designed to be used with a local log server.
3239
Emeric Brun0237c4e2020-11-27 16:24:34 +01003240 local Analog to rfc3164 syslog message format except that hostname
3241 field is stripped. This is the default.
3242 Note: option "log-send-hostname" switches the default to
3243 rfc3164.
3244
Emeric Brun99c453d2020-05-25 15:01:04 +02003245 raw A message containing only the text. The level, PID, date, time,
3246 process name and system name are omitted. This is designed to be
3247 used in containers or during development, where the severity
3248 only depends on the file descriptor used (stdout/stderr). This
3249 is the default.
3250
Emeric Brun0237c4e2020-11-27 16:24:34 +01003251 rfc3164 The RFC3164 syslog message format.
Emeric Brun99c453d2020-05-25 15:01:04 +02003252 (https://tools.ietf.org/html/rfc3164)
3253
3254 rfc5424 The RFC5424 syslog message format.
3255 (https://tools.ietf.org/html/rfc5424)
3256
3257 short A message containing only a level between angle brackets such as
3258 '<3>', followed by the text. The PID, date, time, process name
3259 and system name are omitted. This is designed to be used with a
3260 local log server. This format is compatible with what the systemd
3261 logger consumes.
3262
Emeric Brun54648852020-07-06 15:54:06 +02003263 priority A message containing only a level plus syslog facility between angle
3264 brackets such as '<63>', followed by the text. The PID, date, time,
3265 process name and system name are omitted. This is designed to be used
3266 with a local log server.
3267
Emeric Brun99c453d2020-05-25 15:01:04 +02003268 timed A message containing only a level between angle brackets such as
3269 '<3>', followed by ISO date and by the text. The PID, process
3270 name and system name are omitted. This is designed to be
3271 used with a local log server.
3272
3273maxlen <length>
3274 The maximum length of an event message stored into the ring,
3275 including formatted header. If an event message is longer than
3276 <length>, it will be truncated to this length.
3277
Emeric Brun494c5052020-05-28 11:13:15 +02003278server <name> <address> [param*]
3279 Used to configure a syslog tcp server to forward messages from ring buffer.
3280 This supports for all "server" parameters found in 5.2 paragraph. Some of
3281 these parameters are irrelevant for "ring" sections. Important point: there
3282 is little reason to add more than one server to a ring, because all servers
3283 will receive the exact same copy of the ring contents, and as such the ring
3284 will progress at the speed of the slowest server. If one server does not
3285 respond, it will prevent old messages from being purged and may block new
3286 messages from being inserted into the ring. The proper way to send messages
3287 to multiple servers is to use one distinct ring per log server, not to
Emeric Brun97556472020-05-30 01:42:45 +02003288 attach multiple servers to the same ring. Note that specific server directive
3289 "log-proto" is used to set the protocol used to send messages.
Emeric Brun494c5052020-05-28 11:13:15 +02003290
Emeric Brun99c453d2020-05-25 15:01:04 +02003291size <size>
3292 This is the optional size in bytes for the ring-buffer. Default value is
3293 set to BUFSIZE.
3294
Emeric Brun494c5052020-05-28 11:13:15 +02003295timeout connect <timeout>
3296 Set the maximum time to wait for a connection attempt to a server to succeed.
3297
3298 Arguments :
3299 <timeout> is the timeout value specified in milliseconds by default, but
3300 can be in any other unit if the number is suffixed by the unit,
3301 as explained at the top of this document.
3302
3303timeout server <timeout>
3304 Set the maximum time for pending data staying into output buffer.
3305
3306 Arguments :
3307 <timeout> is the timeout value specified in milliseconds by default, but
3308 can be in any other unit if the number is suffixed by the unit,
3309 as explained at the top of this document.
3310
Emeric Brun99c453d2020-05-25 15:01:04 +02003311 Example:
3312 global
3313 log ring@myring local7
3314
3315 ring myring
3316 description "My local buffer"
3317 format rfc3164
3318 maxlen 1200
3319 size 32764
Emeric Brun494c5052020-05-28 11:13:15 +02003320 timeout connect 5s
3321 timeout server 10s
Emeric Brun97556472020-05-30 01:42:45 +02003322 server mysyslogsrv 127.0.0.1:6514 log-proto octet-count
Emeric Brun99c453d2020-05-25 15:01:04 +02003323
Emeric Brun12941c82020-07-07 14:19:42 +020033243.10. Log forwarding
3325-------------------
3326
3327It is possible to declare one or multiple log forwarding section,
Daniel Corbett9f0843f2021-05-08 10:50:37 -04003328HAProxy will forward all received log messages to a log servers list.
Emeric Brun12941c82020-07-07 14:19:42 +02003329
3330log-forward <name>
3331 Creates a new log forwarder proxy identified as <name>.
3332
Emeric Bruncbb7bf72020-10-05 14:39:35 +02003333backlog <conns>
3334 Give hints to the system about the approximate listen backlog desired size
3335 on connections accept.
3336
3337bind <addr> [param*]
3338 Used to configure a stream log listener to receive messages to forward.
Emeric Brunda46c1c2020-10-08 08:39:02 +02003339 This supports the "bind" parameters found in 5.1 paragraph including
3340 those about ssl but some statements such as "alpn" may be irrelevant for
3341 syslog protocol over TCP.
3342 Those listeners support both "Octet Counting" and "Non-Transparent-Framing"
Emeric Bruncbb7bf72020-10-05 14:39:35 +02003343 modes as defined in rfc-6587.
3344
Willy Tarreau76aaa7f2020-09-16 15:07:22 +02003345dgram-bind <addr> [param*]
Emeric Bruncbb7bf72020-10-05 14:39:35 +02003346 Used to configure a datagram log listener to receive messages to forward.
3347 Addresses must be in IPv4 or IPv6 form,followed by a port. This supports
3348 for some of the "bind" parameters found in 5.1 paragraph among which
3349 "interface", "namespace" or "transparent", the other ones being
Willy Tarreau26ff5da2020-09-16 15:22:19 +02003350 silently ignored as irrelevant for UDP/syslog case.
Emeric Brun12941c82020-07-07 14:19:42 +02003351
3352log global
Jan Wagner3e678602020-12-17 22:22:32 +01003353log <address> [len <length>] [format <format>] [sample <ranges>:<sample_size>]
Emeric Brun12941c82020-07-07 14:19:42 +02003354 <facility> [<level> [<minlevel>]]
3355 Used to configure target log servers. See more details on proxies
3356 documentation.
Daniel Corbett9f0843f2021-05-08 10:50:37 -04003357 If no format specified, HAProxy tries to keep the incoming log format.
Emeric Brun12941c82020-07-07 14:19:42 +02003358 Configured facility is ignored, except if incoming message does not
3359 present a facility but one is mandatory on the outgoing format.
3360 If there is no timestamp available in the input format, but the field
Daniel Corbett9f0843f2021-05-08 10:50:37 -04003361 exists in output format, HAProxy will use the local date.
Emeric Brun12941c82020-07-07 14:19:42 +02003362
3363 Example:
3364 global
3365 log stderr format iso local7
3366
3367 ring myring
3368 description "My local buffer"
3369 format rfc5424
3370 maxlen 1200
3371 size 32764
3372 timeout connect 5s
3373 timeout server 10s
3374 # syslog tcp server
3375 server mysyslogsrv 127.0.0.1:514 log-proto octet-count
3376
3377 log-forward sylog-loadb
Emeric Bruncbb7bf72020-10-05 14:39:35 +02003378 dgram-bind 127.0.0.1:1514
3379 bind 127.0.0.1:1514
Emeric Brun12941c82020-07-07 14:19:42 +02003380 # all messages on stderr
3381 log global
3382 # all messages on local tcp syslog server
3383 log ring@myring local0
3384 # load balance messages on 4 udp syslog servers
3385 log 127.0.0.1:10001 sample 1:4 local0
3386 log 127.0.0.1:10002 sample 2:4 local0
3387 log 127.0.0.1:10003 sample 3:4 local0
3388 log 127.0.0.1:10004 sample 4:4 local0
Christopher Faulet76edc0f2020-01-13 15:52:01 +01003389
Emeric Bruncbb7bf72020-10-05 14:39:35 +02003390maxconn <conns>
3391 Fix the maximum number of concurrent connections on a log forwarder.
3392 10 is the default.
3393
3394timeout client <timeout>
3395 Set the maximum inactivity time on the client side.
3396
Willy Tarreauc57f0e22009-05-10 13:12:33 +020033974. Proxies
Willy Tarreau6a06a402007-07-15 20:15:28 +02003398----------
Willy Tarreau0ba27502007-12-24 16:55:16 +01003399
Willy Tarreau6a06a402007-07-15 20:15:28 +02003400Proxy configuration can be located in a set of sections :
Willy Tarreau7c0b4d82021-02-12 14:58:08 +01003401 - defaults [<name>] [ from <defaults_name> ]
3402 - frontend <name> [ from <defaults_name> ]
3403 - backend <name> [ from <defaults_name> ]
3404 - listen <name> [ from <defaults_name> ]
Willy Tarreau6a06a402007-07-15 20:15:28 +02003405
3406A "frontend" section describes a set of listening sockets accepting client
3407connections.
3408
3409A "backend" section describes a set of servers to which the proxy will connect
3410to forward incoming connections.
3411
3412A "listen" section defines a complete proxy with its frontend and backend
3413parts combined in one section. It is generally useful for TCP-only traffic.
3414
Willy Tarreau7c0b4d82021-02-12 14:58:08 +01003415A "defaults" section resets all settings to the documented ones and presets new
3416ones for use by subsequent sections. All of "frontend", "backend" and "listen"
3417sections always take their initial settings from a defaults section, by default
3418the latest one that appears before the newly created section. It is possible to
3419explicitly designate a specific "defaults" section to load the initial settings
3420from by indicating its name on the section line after the optional keyword
3421"from". While "defaults" section do not impose a name, this use is encouraged
3422for better readability. It is also the only way to designate a specific section
3423to use instead of the default previous one. Since "defaults" section names are
3424optional, by default a very permissive check is applied on their name and these
3425are even permitted to overlap. However if a "defaults" section is referenced by
3426any other section, its name must comply with the syntax imposed on all proxy
3427names, and this name must be unique among the defaults sections. Please note
3428that regardless of what is currently permitted, it is recommended to avoid
3429duplicate section names in general and to respect the same syntax as for proxy
3430names. This rule might be enforced in a future version.
3431
3432Note that it is even possible for a defaults section to take its initial
3433settings from another one, and as such, inherit settings across multiple levels
3434of defaults sections. This can be convenient to establish certain configuration
3435profiles to carry groups of default settings (e.g. TCP vs HTTP or short vs long
3436timeouts) but can quickly become confusing to follow.
3437
Willy Tarreau0ba27502007-12-24 16:55:16 +01003438All proxy names must be formed from upper and lower case letters, digits,
3439'-' (dash), '_' (underscore) , '.' (dot) and ':' (colon). ACL names are
3440case-sensitive, which means that "www" and "WWW" are two different proxies.
3441
3442Historically, all proxy names could overlap, it just caused troubles in the
3443logs. Since the introduction of content switching, it is mandatory that two
3444proxies with overlapping capabilities (frontend/backend) have different names.
3445However, it is still permitted that a frontend and a backend share the same
3446name, as this configuration seems to be commonly encountered.
3447
3448Right now, two major proxy modes are supported : "tcp", also known as layer 4,
3449and "http", also known as layer 7. In layer 4 mode, HAProxy simply forwards
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01003450bidirectional traffic between two sides. In layer 7 mode, HAProxy analyzes the
Willy Tarreau0ba27502007-12-24 16:55:16 +01003451protocol, and can interact with it by allowing, blocking, switching, adding,
3452modifying, or removing arbitrary contents in requests or responses, based on
3453arbitrary criteria.
3454
Willy Tarreau70dffda2014-01-30 03:07:23 +01003455In HTTP mode, the processing applied to requests and responses flowing over
3456a connection depends in the combination of the frontend's HTTP options and
Julien Pivotto21ad3152019-12-10 13:11:17 +01003457the backend's. HAProxy supports 3 connection modes :
Willy Tarreau70dffda2014-01-30 03:07:23 +01003458
3459 - KAL : keep alive ("option http-keep-alive") which is the default mode : all
3460 requests and responses are processed, and connections remain open but idle
3461 between responses and new requests.
3462
Willy Tarreau70dffda2014-01-30 03:07:23 +01003463 - SCL: server close ("option http-server-close") : the server-facing
3464 connection is closed after the end of the response is received, but the
3465 client-facing connection remains open.
3466
Christopher Faulet315b39c2018-09-21 16:26:19 +02003467 - CLO: close ("option httpclose"): the connection is closed after the end of
3468 the response and "Connection: close" appended in both directions.
Willy Tarreau70dffda2014-01-30 03:07:23 +01003469
3470The effective mode that will be applied to a connection passing through a
3471frontend and a backend can be determined by both proxy modes according to the
3472following matrix, but in short, the modes are symmetric, keep-alive is the
Christopher Faulet315b39c2018-09-21 16:26:19 +02003473weakest option and close is the strongest.
Willy Tarreau70dffda2014-01-30 03:07:23 +01003474
Christopher Faulet315b39c2018-09-21 16:26:19 +02003475 Backend mode
Willy Tarreau70dffda2014-01-30 03:07:23 +01003476
Christopher Faulet315b39c2018-09-21 16:26:19 +02003477 | KAL | SCL | CLO
3478 ----+-----+-----+----
3479 KAL | KAL | SCL | CLO
3480 ----+-----+-----+----
Christopher Faulet315b39c2018-09-21 16:26:19 +02003481 mode SCL | SCL | SCL | CLO
3482 ----+-----+-----+----
3483 CLO | CLO | CLO | CLO
Willy Tarreau70dffda2014-01-30 03:07:23 +01003484
Christopher Faulet4d37e532021-03-26 14:44:00 +01003485It is possible to chain a TCP frontend to an HTTP backend. It is pointless if
Daniel Corbett9f0843f2021-05-08 10:50:37 -04003486only HTTP traffic is handled. But it may be used to handle several protocols
3487within the same frontend. In this case, the client's connection is first handled
Christopher Faulet4d37e532021-03-26 14:44:00 +01003488as a raw tcp connection before being upgraded to HTTP. Before the upgrade, the
Daniel Corbett9f0843f2021-05-08 10:50:37 -04003489content processings are performend on raw data. Once upgraded, data is parsed
Christopher Faulet4d37e532021-03-26 14:44:00 +01003490and stored using an internal representation called HTX and it is no longer
3491possible to rely on raw representation. There is no way to go back.
Willy Tarreau0ba27502007-12-24 16:55:16 +01003492
Christopher Faulet4d37e532021-03-26 14:44:00 +01003493There are two kind of upgrades, in-place upgrades and destructive upgrades. The
Daniel Corbett9f0843f2021-05-08 10:50:37 -04003494first ones involves a TCP to HTTP/1 upgrade. In HTTP/1, the request
Christopher Faulet4d37e532021-03-26 14:44:00 +01003495processings are serialized, thus the applicative stream can be preserved. The
Daniel Corbett9f0843f2021-05-08 10:50:37 -04003496second one involves a TCP to HTTP/2 upgrade. Because it is a multiplexed
Christopher Faulet4d37e532021-03-26 14:44:00 +01003497protocol, the applicative stream cannot be associated to any HTTP/2 stream and
3498is destroyed. New applicative streams are then created when HAProxy receives
3499new HTTP/2 streams at the lower level, in the H2 multiplexer. It is important
Daniel Corbett9f0843f2021-05-08 10:50:37 -04003500to understand this difference because that drastically changes the way to
Christopher Faulet4d37e532021-03-26 14:44:00 +01003501process data. When an HTTP/1 upgrade is performed, the content processings
3502already performed on raw data are neither lost nor reexecuted while for an
3503HTTP/2 upgrade, applicative streams are distinct and all frontend rules are
3504evaluated systematically on each one. And as said, the first stream, the TCP
3505one, is destroyed, but only after the frontend rules were evaluated.
3506
3507There is another importnat point to understand when HTTP processings are
3508performed from a TCP proxy. While HAProxy is able to parse HTTP/1 in-fly from
3509tcp-request content rules, it is not possible for HTTP/2. Only the HTTP/2
3510preface can be parsed. This is a huge limitation regarding the HTTP content
3511analysis in TCP. Concretely it is only possible to know if received data are
3512HTTP. For instance, it is not possible to choose a backend based on the Host
3513header value while it is trivial in HTTP/1. Hopefully, there is a solution to
3514mitigate this drawback.
3515
Daniel Corbett9f0843f2021-05-08 10:50:37 -04003516There are two ways to perform an HTTP upgrade. The first one, the historical
Christopher Faulet4d37e532021-03-26 14:44:00 +01003517method, is to select an HTTP backend. The upgrade happens when the backend is
3518set. Thus, for in-place upgrades, only the backend configuration is considered
3519in the HTTP data processing. For destructive upgrades, the applicative stream
3520is destroyed, thus its processing is stopped. With this method, possibilities
3521to choose a backend with an HTTP/2 connection are really limited, as mentioned
3522above, and a bit useless because the stream is destroyed. The second method is
3523to upgrade during the tcp-request content rules evaluation, thanks to the
3524"switch-mode http" action. In this case, the upgrade is performed in the
3525frontend context and it is possible to define HTTP directives in this
3526frontend. For in-place upgrades, it offers all the power of the HTTP analysis
3527as soon as possible. It is not that far from an HTTP frontend. For destructive
3528upgrades, it does not change anything except it is useless to choose a backend
3529on limited information. It is of course the recommended method. Thus, testing
3530the request protocol from the tcp-request content rules to perform an HTTP
3531upgrade is enough. All the remaining HTTP manipulation may be moved to the
3532frontend http-request ruleset. But keep in mind that tcp-request content rules
3533remains evaluated on each streams, that can't be changed.
Willy Tarreau70dffda2014-01-30 03:07:23 +01003534
Willy Tarreauc57f0e22009-05-10 13:12:33 +020035354.1. Proxy keywords matrix
3536--------------------------
Willy Tarreau0ba27502007-12-24 16:55:16 +01003537
Willy Tarreauc57f0e22009-05-10 13:12:33 +02003538The following list of keywords is supported. Most of them may only be used in a
3539limited set of section types. Some of them are marked as "deprecated" because
3540they are inherited from an old syntax which may be confusing or functionally
3541limited, and there are new recommended keywords to replace them. Keywords
Davor Ocelice9ed2812017-12-25 17:49:28 +01003542marked with "(*)" can be optionally inverted using the "no" prefix, e.g. "no
Willy Tarreauc57f0e22009-05-10 13:12:33 +02003543option contstats". This makes sense when the option has been enabled by default
Willy Tarreau3842f002009-06-14 11:39:52 +02003544and must be disabled for a specific instance. Such options may also be prefixed
3545with "default" in order to restore default settings regardless of what has been
3546specified in a previous "defaults" section.
Willy Tarreau0ba27502007-12-24 16:55:16 +01003547
Willy Tarreau6a06a402007-07-15 20:15:28 +02003548
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01003549 keyword defaults frontend listen backend
3550------------------------------------+----------+----------+---------+---------
3551acl - X X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01003552backlog X X X -
3553balance X - X X
3554bind - X X -
3555bind-process X X X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01003556capture cookie - X X -
3557capture request header - X X -
3558capture response header - X X -
MIZUTA Takeshib24bc0d2020-07-09 11:13:20 +09003559clitcpka-cnt X X X -
3560clitcpka-idle X X X -
3561clitcpka-intvl X X X -
William Lallemand82fe75c2012-10-23 10:25:10 +02003562compression X X X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01003563cookie X - X X
Thierry FOURNIERa0a1b752015-05-26 17:44:32 +02003564declare capture - X X -
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01003565default-server X - X X
3566default_backend X X X -
3567description - X X X
3568disabled X X X X
3569dispatch - - X X
Simon Horman51a1cf62015-02-03 13:00:44 +09003570email-alert from X X X X
Simon Horman64e34162015-02-06 11:11:57 +09003571email-alert level X X X X
Simon Horman51a1cf62015-02-03 13:00:44 +09003572email-alert mailers X X X X
3573email-alert myhostname X X X X
3574email-alert to X X X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01003575enabled X X X X
3576errorfile X X X X
Christopher Faulet76edc0f2020-01-13 15:52:01 +01003577errorfiles X X X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01003578errorloc X X X X
3579errorloc302 X X X X
3580-- keyword -------------------------- defaults - frontend - listen -- backend -
3581errorloc303 X X X X
Cyril Bonté4288c5a2018-03-12 22:02:59 +01003582force-persist - - X X
Christopher Fauletc3fe5332016-04-07 15:30:10 +02003583filter - X X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01003584fullconn X - X X
3585grace X X X X
3586hash-type X - X X
Christopher Faulet6d0c3df2020-01-22 09:26:35 +01003587http-after-response - X X X
Christopher Faulet4f5c2e22020-04-23 15:22:33 +02003588http-check comment X - X X
Christopher Faulete5870d82020-04-15 11:32:03 +02003589http-check connect X - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01003590http-check disable-on-404 X - X X
Christopher Faulete5870d82020-04-15 11:32:03 +02003591http-check expect X - X X
Peter Gervai8912ae62020-06-11 18:26:36 +02003592http-check send X - X X
Willy Tarreau7ab6aff2010-10-12 06:30:16 +02003593http-check send-state X - X X
Christopher Faulete5870d82020-04-15 11:32:03 +02003594http-check set-var X - X X
3595http-check unset-var X - X X
Christopher Faulet3b967c12020-05-15 15:47:44 +02003596http-error X X X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01003597http-request - X X X
Willy Tarreaue365c0b2013-06-11 16:06:12 +02003598http-response - X X X
Willy Tarreau30631952015-08-06 15:05:24 +02003599http-reuse X - X X
Baptiste Assmann2c42ef52013-10-09 21:57:02 +02003600http-send-name-header - - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01003601id - X X X
Cyril Bonté4288c5a2018-03-12 22:02:59 +01003602ignore-persist - - X X
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02003603load-server-state-from-file X - X X
William Lallemand0f99e342011-10-12 17:50:54 +02003604log (*) X X X X
Willy Tarreaufb4e7ea2015-01-07 14:55:17 +01003605log-format X X X -
Dragan Dosen7ad31542015-09-28 17:16:47 +02003606log-format-sd X X X -
Willy Tarreau094af4e2015-01-07 15:03:42 +01003607log-tag X X X X
Willy Tarreauc35362a2014-04-25 13:58:37 +02003608max-keep-alive-queue X - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01003609maxconn X X X -
3610mode X X X X
3611monitor fail - X X -
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01003612monitor-uri X X X -
3613option abortonclose (*) X - X X
3614option accept-invalid-http-request (*) X X X -
3615option accept-invalid-http-response (*) X - X X
3616option allbackups (*) X - X X
3617option checkcache (*) X - X X
3618option clitcpka (*) X X X -
3619option contstats (*) X X X -
Christopher Faulet89aed322020-06-02 17:33:56 +02003620option disable-h2-upgrade (*) X X X -
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01003621option dontlog-normal (*) X X X -
3622option dontlognull (*) X X X -
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01003623-- keyword -------------------------- defaults - frontend - listen -- backend -
3624option forwardfor X X X X
Christopher Faulet98fbe952019-07-22 16:18:24 +02003625option h1-case-adjust-bogus-client (*) X X X -
3626option h1-case-adjust-bogus-server (*) X - X X
Willy Tarreau9fbe18e2015-05-01 22:42:08 +02003627option http-buffer-request (*) X X X X
Willy Tarreau82649f92015-05-01 22:40:51 +02003628option http-ignore-probes (*) X X X -
Willy Tarreau16bfb022010-01-16 19:48:41 +01003629option http-keep-alive (*) X X X X
Willy Tarreau96e31212011-05-30 18:10:30 +02003630option http-no-delay (*) X X X X
Christopher Faulet98db9762018-09-21 10:25:19 +02003631option http-pretend-keepalive (*) X - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01003632option http-server-close (*) X X X X
3633option http-use-proxy-header (*) X X X -
3634option httpchk X - X X
3635option httpclose (*) X X X X
Freddy Spierenburge88b7732019-03-25 14:35:17 +01003636option httplog X X X -
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01003637option http_proxy (*) X X X X
Jamie Gloudon801a0a32012-08-25 00:18:33 -04003638option independent-streams (*) X X X X
Gabor Lekenyb4c81e42010-09-29 18:17:05 +02003639option ldap-check X - X X
Simon Horman98637e52014-06-20 12:30:16 +09003640option external-check X - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01003641option log-health-checks (*) X - X X
3642option log-separate-errors (*) X X X -
3643option logasap (*) X X X -
3644option mysql-check X - X X
3645option nolinger (*) X X X X
3646option originalto X X X X
3647option persist (*) X - X X
Baptiste Assmann809e22a2015-10-12 20:22:55 +02003648option pgsql-check X - X X
3649option prefer-last-server (*) X - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01003650option redispatch (*) X - X X
Hervé COMMOWICKec032d62011-08-05 16:23:48 +02003651option redis-check X - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01003652option smtpchk X - X X
3653option socket-stats (*) X X X -
3654option splice-auto (*) X X X X
3655option splice-request (*) X X X X
3656option splice-response (*) X X X X
Christopher Fauletba7bc162016-11-07 21:07:38 +01003657option spop-check - - - X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01003658option srvtcpka (*) X - X X
3659option ssl-hello-chk X - X X
3660-- keyword -------------------------- defaults - frontend - listen -- backend -
Willy Tarreaued179852013-12-16 01:07:00 +01003661option tcp-check X - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01003662option tcp-smart-accept (*) X X X -
3663option tcp-smart-connect (*) X - X X
3664option tcpka X X X X
3665option tcplog X X X X
3666option transparent (*) X - X X
Simon Horman98637e52014-06-20 12:30:16 +09003667external-check command X - X X
3668external-check path X - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01003669persist rdp-cookie X - X X
3670rate-limit sessions X X X -
3671redirect - X X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01003672-- keyword -------------------------- defaults - frontend - listen -- backend -
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01003673retries X - X X
Olivier Houcharda254a372019-04-05 15:30:12 +02003674retry-on X - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01003675server - - X X
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02003676server-state-file-name X - X X
Frédéric Lécaillecb4502e2017-04-20 13:36:25 +02003677server-template - - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01003678source X - X X
MIZUTA Takeshib24bc0d2020-07-09 11:13:20 +09003679srvtcpka-cnt X - X X
3680srvtcpka-idle X - X X
3681srvtcpka-intvl X - X X
Baptiste Assmann5a549212015-10-12 20:30:24 +02003682stats admin - X X X
3683stats auth X X X X
3684stats enable X X X X
3685stats hide-version X X X X
3686stats http-request - X X X
3687stats realm X X X X
3688stats refresh X X X X
3689stats scope X X X X
3690stats show-desc X X X X
3691stats show-legends X X X X
3692stats show-node X X X X
3693stats uri X X X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01003694-- keyword -------------------------- defaults - frontend - listen -- backend -
3695stick match - - X X
3696stick on - - X X
3697stick store-request - - X X
Willy Tarreaud8dc99f2011-07-01 11:33:25 +02003698stick store-response - - X X
Adam Spiers68af3c12017-04-06 16:31:39 +01003699stick-table - X X X
Christopher Faulet4f5c2e22020-04-23 15:22:33 +02003700tcp-check comment X - X X
Christopher Faulet404f9192020-04-09 23:13:54 +02003701tcp-check connect X - X X
3702tcp-check expect X - X X
3703tcp-check send X - X X
Christopher Fauletb50b3e62020-05-05 18:43:43 +02003704tcp-check send-lf X - X X
Christopher Faulet404f9192020-04-09 23:13:54 +02003705tcp-check send-binary X - X X
Christopher Fauletb50b3e62020-05-05 18:43:43 +02003706tcp-check send-binary-lf X - X X
Christopher Faulet404f9192020-04-09 23:13:54 +02003707tcp-check set-var X - X X
3708tcp-check unset-var X - X X
Willy Tarreaue9656522010-08-17 15:40:09 +02003709tcp-request connection - X X -
3710tcp-request content - X X X
Willy Tarreaua56235c2010-09-14 11:31:36 +02003711tcp-request inspect-delay - X X X
Willy Tarreau4f614292016-10-21 17:49:36 +02003712tcp-request session - X X -
Emeric Brun0a3b67f2010-09-24 15:34:53 +02003713tcp-response content - - X X
3714tcp-response inspect-delay - - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01003715timeout check X - X X
3716timeout client X X X -
Willy Tarreau05cdd962014-05-10 14:30:07 +02003717timeout client-fin X X X -
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01003718timeout connect X - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01003719timeout http-keep-alive X X X X
3720timeout http-request X X X X
3721timeout queue X - X X
3722timeout server X - X X
Willy Tarreau05cdd962014-05-10 14:30:07 +02003723timeout server-fin X - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01003724timeout tarpit X X X X
Willy Tarreauce887fd2012-05-12 12:50:00 +02003725timeout tunnel X - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01003726transparent (deprecated) X - X X
William Lallemanda73203e2012-03-12 12:48:57 +01003727unique-id-format X X X -
3728unique-id-header X X X -
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01003729use_backend - X X -
Christopher Fauletb30b3102019-09-12 23:03:09 +02003730use-fcgi-app - - X X
Willy Tarreau4a5cade2012-04-05 21:09:48 +02003731use-server - - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01003732------------------------------------+----------+----------+---------+---------
3733 keyword defaults frontend listen backend
Willy Tarreau6a06a402007-07-15 20:15:28 +02003734
Willy Tarreau0ba27502007-12-24 16:55:16 +01003735
Willy Tarreauc57f0e22009-05-10 13:12:33 +020037364.2. Alphabetically sorted keywords reference
3737---------------------------------------------
Willy Tarreau0ba27502007-12-24 16:55:16 +01003738
3739This section provides a description of each keyword and its usage.
3740
3741
3742acl <aclname> <criterion> [flags] [operator] <value> ...
3743 Declare or complete an access list.
3744 May be used in sections : defaults | frontend | listen | backend
3745 no | yes | yes | yes
3746 Example:
3747 acl invalid_src src 0.0.0.0/7 224.0.0.0/3
3748 acl invalid_src src_port 0:1023
3749 acl local_dst hdr(host) -i localhost
3750
Willy Tarreauc57f0e22009-05-10 13:12:33 +02003751 See section 7 about ACL usage.
Willy Tarreau0ba27502007-12-24 16:55:16 +01003752
3753
Willy Tarreauc73ce2b2008-01-06 10:55:10 +01003754backlog <conns>
3755 Give hints to the system about the approximate listen backlog desired size
3756 May be used in sections : defaults | frontend | listen | backend
3757 yes | yes | yes | no
3758 Arguments :
3759 <conns> is the number of pending connections. Depending on the operating
3760 system, it may represent the number of already acknowledged
Cyril Bontédc4d9032012-04-08 21:57:39 +02003761 connections, of non-acknowledged ones, or both.
Willy Tarreauc73ce2b2008-01-06 10:55:10 +01003762
3763 In order to protect against SYN flood attacks, one solution is to increase
3764 the system's SYN backlog size. Depending on the system, sometimes it is just
3765 tunable via a system parameter, sometimes it is not adjustable at all, and
3766 sometimes the system relies on hints given by the application at the time of
3767 the listen() syscall. By default, HAProxy passes the frontend's maxconn value
3768 to the listen() syscall. On systems which can make use of this value, it can
3769 sometimes be useful to be able to specify a different value, hence this
3770 backlog parameter.
3771
3772 On Linux 2.4, the parameter is ignored by the system. On Linux 2.6, it is
3773 used as a hint and the system accepts up to the smallest greater power of
3774 two, and never more than some limits (usually 32768).
3775
3776 See also : "maxconn" and the target operating system's tuning guide.
3777
3778
Willy Tarreau0ba27502007-12-24 16:55:16 +01003779balance <algorithm> [ <arguments> ]
Willy Tarreau226071e2014-04-10 11:55:45 +02003780balance url_param <param> [check_post]
Willy Tarreau0ba27502007-12-24 16:55:16 +01003781 Define the load balancing algorithm to be used in a backend.
3782 May be used in sections : defaults | frontend | listen | backend
3783 yes | no | yes | yes
3784 Arguments :
3785 <algorithm> is the algorithm used to select a server when doing load
3786 balancing. This only applies when no persistence information
3787 is available, or when a connection is redispatched to another
3788 server. <algorithm> may be one of the following :
3789
3790 roundrobin Each server is used in turns, according to their weights.
3791 This is the smoothest and fairest algorithm when the server's
3792 processing time remains equally distributed. This algorithm
3793 is dynamic, which means that server weights may be adjusted
Willy Tarreau9757a382009-10-03 12:56:50 +02003794 on the fly for slow starts for instance. It is limited by
Godbacha34bdc02013-07-22 07:44:53 +08003795 design to 4095 active servers per backend. Note that in some
Willy Tarreau9757a382009-10-03 12:56:50 +02003796 large farms, when a server becomes up after having been down
3797 for a very short time, it may sometimes take a few hundreds
3798 requests for it to be re-integrated into the farm and start
3799 receiving traffic. This is normal, though very rare. It is
3800 indicated here in case you would have the chance to observe
3801 it, so that you don't worry.
3802
3803 static-rr Each server is used in turns, according to their weights.
3804 This algorithm is as similar to roundrobin except that it is
3805 static, which means that changing a server's weight on the
3806 fly will have no effect. On the other hand, it has no design
3807 limitation on the number of servers, and when a server goes
3808 up, it is always immediately reintroduced into the farm, once
3809 the full map is recomputed. It also uses slightly less CPU to
3810 run (around -1%).
Willy Tarreau0ba27502007-12-24 16:55:16 +01003811
Willy Tarreau2d2a7f82008-03-17 12:07:56 +01003812 leastconn The server with the lowest number of connections receives the
3813 connection. Round-robin is performed within groups of servers
3814 of the same load to ensure that all servers will be used. Use
3815 of this algorithm is recommended where very long sessions are
3816 expected, such as LDAP, SQL, TSE, etc... but is not very well
3817 suited for protocols using short sessions such as HTTP. This
3818 algorithm is dynamic, which means that server weights may be
Willy Tarreau8c855f62020-10-22 17:41:45 +02003819 adjusted on the fly for slow starts for instance. It will
3820 also consider the number of queued connections in addition to
3821 the established ones in order to minimize queuing.
Willy Tarreau2d2a7f82008-03-17 12:07:56 +01003822
Willy Tarreauf09c6602012-02-13 17:12:08 +01003823 first The first server with available connection slots receives the
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03003824 connection. The servers are chosen from the lowest numeric
Willy Tarreauf09c6602012-02-13 17:12:08 +01003825 identifier to the highest (see server parameter "id"), which
3826 defaults to the server's position in the farm. Once a server
Willy Tarreau64559c52012-04-07 09:08:45 +02003827 reaches its maxconn value, the next server is used. It does
Willy Tarreauf09c6602012-02-13 17:12:08 +01003828 not make sense to use this algorithm without setting maxconn.
3829 The purpose of this algorithm is to always use the smallest
3830 number of servers so that extra servers can be powered off
3831 during non-intensive hours. This algorithm ignores the server
3832 weight, and brings more benefit to long session such as RDP
Willy Tarreau64559c52012-04-07 09:08:45 +02003833 or IMAP than HTTP, though it can be useful there too. In
3834 order to use this algorithm efficiently, it is recommended
3835 that a cloud controller regularly checks server usage to turn
3836 them off when unused, and regularly checks backend queue to
3837 turn new servers on when the queue inflates. Alternatively,
3838 using "http-check send-state" may inform servers on the load.
Willy Tarreauf09c6602012-02-13 17:12:08 +01003839
Willy Tarreau0ba27502007-12-24 16:55:16 +01003840 source The source IP address is hashed and divided by the total
3841 weight of the running servers to designate which server will
3842 receive the request. This ensures that the same client IP
3843 address will always reach the same server as long as no
3844 server goes down or up. If the hash result changes due to the
3845 number of running servers changing, many clients will be
3846 directed to a different server. This algorithm is generally
3847 used in TCP mode where no cookie may be inserted. It may also
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01003848 be used on the Internet to provide a best-effort stickiness
Willy Tarreau0ba27502007-12-24 16:55:16 +01003849 to clients which refuse session cookies. This algorithm is
Willy Tarreau6b2e11b2009-10-01 07:52:15 +02003850 static by default, which means that changing a server's
3851 weight on the fly will have no effect, but this can be
3852 changed using "hash-type".
Willy Tarreau0ba27502007-12-24 16:55:16 +01003853
Oskar Stolc8dc41842012-05-19 10:19:54 +01003854 uri This algorithm hashes either the left part of the URI (before
3855 the question mark) or the whole URI (if the "whole" parameter
3856 is present) and divides the hash value by the total weight of
3857 the running servers. The result designates which server will
3858 receive the request. This ensures that the same URI will
3859 always be directed to the same server as long as no server
3860 goes up or down. This is used with proxy caches and
3861 anti-virus proxies in order to maximize the cache hit rate.
3862 Note that this algorithm may only be used in an HTTP backend.
3863 This algorithm is static by default, which means that
3864 changing a server's weight on the fly will have no effect,
3865 but this can be changed using "hash-type".
Willy Tarreau0ba27502007-12-24 16:55:16 +01003866
Oskar Stolc8dc41842012-05-19 10:19:54 +01003867 This algorithm supports two optional parameters "len" and
Marek Majkowski9c30fc12008-04-27 23:25:55 +02003868 "depth", both followed by a positive integer number. These
3869 options may be helpful when it is needed to balance servers
3870 based on the beginning of the URI only. The "len" parameter
3871 indicates that the algorithm should only consider that many
3872 characters at the beginning of the URI to compute the hash.
3873 Note that having "len" set to 1 rarely makes sense since most
3874 URIs start with a leading "/".
3875
3876 The "depth" parameter indicates the maximum directory depth
3877 to be used to compute the hash. One level is counted for each
3878 slash in the request. If both parameters are specified, the
3879 evaluation stops when either is reached.
3880
Willy Tarreau57a37412020-09-23 08:56:29 +02003881 A "path-only" parameter indicates that the hashing key starts
3882 at the first '/' of the path. This can be used to ignore the
3883 authority part of absolute URIs, and to make sure that HTTP/1
3884 and HTTP/2 URIs will provide the same hash.
3885
Willy Tarreau0ba27502007-12-24 16:55:16 +01003886 url_param The URL parameter specified in argument will be looked up in
matt.farnsworth@nokia.com1c2ab962008-04-14 20:47:37 +02003887 the query string of each HTTP GET request.
3888
3889 If the modifier "check_post" is used, then an HTTP POST
Cyril Bontédc4d9032012-04-08 21:57:39 +02003890 request entity will be searched for the parameter argument,
3891 when it is not found in a query string after a question mark
Willy Tarreau226071e2014-04-10 11:55:45 +02003892 ('?') in the URL. The message body will only start to be
3893 analyzed once either the advertised amount of data has been
3894 received or the request buffer is full. In the unlikely event
3895 that chunked encoding is used, only the first chunk is
Cyril Bontédc4d9032012-04-08 21:57:39 +02003896 scanned. Parameter values separated by a chunk boundary, may
Willy Tarreau226071e2014-04-10 11:55:45 +02003897 be randomly balanced if at all. This keyword used to support
3898 an optional <max_wait> parameter which is now ignored.
matt.farnsworth@nokia.com1c2ab962008-04-14 20:47:37 +02003899
3900 If the parameter is found followed by an equal sign ('=') and
3901 a value, then the value is hashed and divided by the total
3902 weight of the running servers. The result designates which
3903 server will receive the request.
3904
3905 This is used to track user identifiers in requests and ensure
3906 that a same user ID will always be sent to the same server as
3907 long as no server goes up or down. If no value is found or if
3908 the parameter is not found, then a round robin algorithm is
3909 applied. Note that this algorithm may only be used in an HTTP
Willy Tarreau6b2e11b2009-10-01 07:52:15 +02003910 backend. This algorithm is static by default, which means
3911 that changing a server's weight on the fly will have no
3912 effect, but this can be changed using "hash-type".
Willy Tarreau0ba27502007-12-24 16:55:16 +01003913
Cyril Bontédc4d9032012-04-08 21:57:39 +02003914 hdr(<name>) The HTTP header <name> will be looked up in each HTTP
3915 request. Just as with the equivalent ACL 'hdr()' function,
3916 the header name in parenthesis is not case sensitive. If the
3917 header is absent or if it does not contain any value, the
3918 roundrobin algorithm is applied instead.
Benoitaffb4812009-03-25 13:02:10 +01003919
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01003920 An optional 'use_domain_only' parameter is available, for
Benoitaffb4812009-03-25 13:02:10 +01003921 reducing the hash algorithm to the main domain part with some
3922 specific headers such as 'Host'. For instance, in the Host
3923 value "haproxy.1wt.eu", only "1wt" will be considered.
3924
Willy Tarreau6b2e11b2009-10-01 07:52:15 +02003925 This algorithm is static by default, which means that
3926 changing a server's weight on the fly will have no effect,
3927 but this can be changed using "hash-type".
3928
Willy Tarreau21c741a2019-01-14 18:14:27 +01003929 random
3930 random(<draws>)
3931 A random number will be used as the key for the consistent
Willy Tarreau760e81d2018-05-03 07:20:40 +02003932 hashing function. This means that the servers' weights are
3933 respected, dynamic weight changes immediately take effect, as
3934 well as new server additions. Random load balancing can be
3935 useful with large farms or when servers are frequently added
Willy Tarreau21c741a2019-01-14 18:14:27 +01003936 or removed as it may avoid the hammering effect that could
3937 result from roundrobin or leastconn in this situation. The
3938 hash-balance-factor directive can be used to further improve
3939 fairness of the load balancing, especially in situations
3940 where servers show highly variable response times. When an
3941 argument <draws> is present, it must be an integer value one
3942 or greater, indicating the number of draws before selecting
3943 the least loaded of these servers. It was indeed demonstrated
3944 that picking the least loaded of two servers is enough to
3945 significantly improve the fairness of the algorithm, by
3946 always avoiding to pick the most loaded server within a farm
3947 and getting rid of any bias that could be induced by the
3948 unfair distribution of the consistent list. Higher values N
3949 will take away N-1 of the highest loaded servers at the
3950 expense of performance. With very high values, the algorithm
3951 will converge towards the leastconn's result but much slower.
3952 The default value is 2, which generally shows very good
3953 distribution and performance. This algorithm is also known as
3954 the Power of Two Random Choices and is described here :
3955 http://www.eecs.harvard.edu/~michaelm/postscripts/handbook2001.pdf
Willy Tarreau760e81d2018-05-03 07:20:40 +02003956
Emeric Brun736aa232009-06-30 17:56:00 +02003957 rdp-cookie
Hervé COMMOWICKa3eb39c2011-08-05 18:48:51 +02003958 rdp-cookie(<name>)
Emeric Brun736aa232009-06-30 17:56:00 +02003959 The RDP cookie <name> (or "mstshash" if omitted) will be
3960 looked up and hashed for each incoming TCP request. Just as
3961 with the equivalent ACL 'req_rdp_cookie()' function, the name
3962 is not case-sensitive. This mechanism is useful as a degraded
3963 persistence mode, as it makes it possible to always send the
3964 same user (or the same session ID) to the same server. If the
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01003965 cookie is not found, the normal roundrobin algorithm is
Emeric Brun736aa232009-06-30 17:56:00 +02003966 used instead.
3967
3968 Note that for this to work, the frontend must ensure that an
3969 RDP cookie is already present in the request buffer. For this
3970 you must use 'tcp-request content accept' rule combined with
3971 a 'req_rdp_cookie_cnt' ACL.
3972
Willy Tarreau6b2e11b2009-10-01 07:52:15 +02003973 This algorithm is static by default, which means that
3974 changing a server's weight on the fly will have no effect,
3975 but this can be changed using "hash-type".
3976
Cyril Bontédc4d9032012-04-08 21:57:39 +02003977 See also the rdp_cookie pattern fetch function.
Simon Hormanab814e02011-06-24 14:50:20 +09003978
Willy Tarreau0ba27502007-12-24 16:55:16 +01003979 <arguments> is an optional list of arguments which may be needed by some
Marek Majkowski9c30fc12008-04-27 23:25:55 +02003980 algorithms. Right now, only "url_param" and "uri" support an
3981 optional argument.
matt.farnsworth@nokia.com1c2ab962008-04-14 20:47:37 +02003982
Willy Tarreau3cd9af22009-03-15 14:06:41 +01003983 The load balancing algorithm of a backend is set to roundrobin when no other
3984 algorithm, mode nor option have been set. The algorithm may only be set once
3985 for each backend.
Willy Tarreau0ba27502007-12-24 16:55:16 +01003986
Lukas Tribus80512b12018-10-27 20:07:40 +02003987 With authentication schemes that require the same connection like NTLM, URI
John Roeslerfb2fce12019-07-10 15:45:51 -05003988 based algorithms must not be used, as they would cause subsequent requests
Lukas Tribus80512b12018-10-27 20:07:40 +02003989 to be routed to different backend servers, breaking the invalid assumptions
3990 NTLM relies on.
3991
Willy Tarreau0ba27502007-12-24 16:55:16 +01003992 Examples :
3993 balance roundrobin
3994 balance url_param userid
matt.farnsworth@nokia.com1c2ab962008-04-14 20:47:37 +02003995 balance url_param session_id check_post 64
Benoitaffb4812009-03-25 13:02:10 +01003996 balance hdr(User-Agent)
3997 balance hdr(host)
3998 balance hdr(Host) use_domain_only
matt.farnsworth@nokia.com1c2ab962008-04-14 20:47:37 +02003999
4000 Note: the following caveats and limitations on using the "check_post"
4001 extension with "url_param" must be considered :
4002
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01004003 - all POST requests are eligible for consideration, because there is no way
matt.farnsworth@nokia.com1c2ab962008-04-14 20:47:37 +02004004 to determine if the parameters will be found in the body or entity which
4005 may contain binary data. Therefore another method may be required to
4006 restrict consideration of POST requests that have no URL parameters in
Christopher Faulet87f1f3d2019-07-18 14:51:20 +02004007 the body. (see acl http_end)
matt.farnsworth@nokia.com1c2ab962008-04-14 20:47:37 +02004008
4009 - using a <max_wait> value larger than the request buffer size does not
4010 make sense and is useless. The buffer size is set at build time, and
4011 defaults to 16 kB.
4012
4013 - Content-Encoding is not supported, the parameter search will probably
4014 fail; and load balancing will fall back to Round Robin.
4015
4016 - Expect: 100-continue is not supported, load balancing will fall back to
4017 Round Robin.
4018
Lukas Tribus23953682017-04-28 13:24:30 +00004019 - Transfer-Encoding (RFC7230 3.3.1) is only supported in the first chunk.
matt.farnsworth@nokia.com1c2ab962008-04-14 20:47:37 +02004020 If the entire parameter value is not present in the first chunk, the
4021 selection of server is undefined (actually, defined by how little
4022 actually appeared in the first chunk).
4023
4024 - This feature does not support generation of a 100, 411 or 501 response.
4025
4026 - In some cases, requesting "check_post" MAY attempt to scan the entire
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01004027 contents of a message body. Scanning normally terminates when linear
matt.farnsworth@nokia.com1c2ab962008-04-14 20:47:37 +02004028 white space or control characters are found, indicating the end of what
4029 might be a URL parameter list. This is probably not a concern with SGML
4030 type message bodies.
Willy Tarreau0ba27502007-12-24 16:55:16 +01004031
Willy Tarreau294d0f02015-08-10 19:40:12 +02004032 See also : "dispatch", "cookie", "transparent", "hash-type" and "http_proxy".
Willy Tarreau0ba27502007-12-24 16:55:16 +01004033
4034
Willy Tarreaub6205fd2012-09-24 12:27:33 +02004035bind [<address>]:<port_range> [, ...] [param*]
4036bind /<path> [, ...] [param*]
Willy Tarreau0ba27502007-12-24 16:55:16 +01004037 Define one or several listening addresses and/or ports in a frontend.
4038 May be used in sections : defaults | frontend | listen | backend
4039 no | yes | yes | no
4040 Arguments :
Willy Tarreaub1e52e82008-01-13 14:49:51 +01004041 <address> is optional and can be a host name, an IPv4 address, an IPv6
4042 address, or '*'. It designates the address the frontend will
4043 listen on. If unset, all IPv4 addresses of the system will be
4044 listened on. The same will apply for '*' or the system's
David du Colombier9c938da2011-03-17 10:40:27 +01004045 special address "0.0.0.0". The IPv6 equivalent is '::'.
Willy Tarreau24709282013-03-10 21:32:12 +01004046 Optionally, an address family prefix may be used before the
4047 address to force the family regardless of the address format,
4048 which can be useful to specify a path to a unix socket with
4049 no slash ('/'). Currently supported prefixes are :
4050 - 'ipv4@' -> address is always IPv4
4051 - 'ipv6@' -> address is always IPv6
Emeric Brun3835c0d2020-07-07 09:46:09 +02004052 - 'udp@' -> address is resolved as IPv4 or IPv6 and
Emeric Brun12941c82020-07-07 14:19:42 +02004053 protocol UDP is used. Currently those listeners are
4054 supported only in log-forward sections.
Emeric Brun3835c0d2020-07-07 09:46:09 +02004055 - 'udp4@' -> address is always IPv4 and protocol UDP
Emeric Brun12941c82020-07-07 14:19:42 +02004056 is used. Currently those listeners are supported
4057 only in log-forward sections.
Emeric Brun3835c0d2020-07-07 09:46:09 +02004058 - 'udp6@' -> address is always IPv6 and protocol UDP
Emeric Brun12941c82020-07-07 14:19:42 +02004059 is used. Currently those listeners are supported
4060 only in log-forward sections.
Willy Tarreau24709282013-03-10 21:32:12 +01004061 - 'unix@' -> address is a path to a local unix socket
Willy Tarreau70f72e02014-07-08 00:37:50 +02004062 - 'abns@' -> address is in abstract namespace (Linux only).
4063 Note: since abstract sockets are not "rebindable", they
4064 do not cope well with multi-process mode during
4065 soft-restart, so it is better to avoid them if
4066 nbproc is greater than 1. The effect is that if the
4067 new process fails to start, only one of the old ones
4068 will be able to rebind to the socket.
Willy Tarreau40aa0702013-03-10 23:51:38 +01004069 - 'fd@<n>' -> use file descriptor <n> inherited from the
4070 parent. The fd must be bound and may or may not already
4071 be listening.
William Lallemand2fe7dd02018-09-11 16:51:29 +02004072 - 'sockpair@<n>'-> like fd@ but you must use the fd of a
4073 connected unix socket or of a socketpair. The bind waits
4074 to receive a FD over the unix socket and uses it as if it
4075 was the FD of an accept(). Should be used carefully.
William Lallemandb2f07452015-05-12 14:27:13 +02004076 You may want to reference some environment variables in the
4077 address parameter, see section 2.3 about environment
4078 variables.
Willy Tarreaub1e52e82008-01-13 14:49:51 +01004079
Willy Tarreauc5011ca2010-03-22 11:53:56 +01004080 <port_range> is either a unique TCP port, or a port range for which the
4081 proxy will accept connections for the IP address specified
Willy Tarreauceb24bc2010-11-09 12:46:41 +01004082 above. The port is mandatory for TCP listeners. Note that in
4083 the case of an IPv6 address, the port is always the number
4084 after the last colon (':'). A range can either be :
Willy Tarreauc5011ca2010-03-22 11:53:56 +01004085 - a numerical port (ex: '80')
4086 - a dash-delimited ports range explicitly stating the lower
4087 and upper bounds (ex: '2000-2100') which are included in
4088 the range.
4089
4090 Particular care must be taken against port ranges, because
4091 every <address:port> couple consumes one socket (= a file
4092 descriptor), so it's easy to consume lots of descriptors
4093 with a simple range, and to run out of sockets. Also, each
4094 <address:port> couple must be used only once among all
4095 instances running on a same system. Please note that binding
4096 to ports lower than 1024 generally require particular
Jamie Gloudon801a0a32012-08-25 00:18:33 -04004097 privileges to start the program, which are independent of
Willy Tarreauc5011ca2010-03-22 11:53:56 +01004098 the 'uid' parameter.
Willy Tarreau0ba27502007-12-24 16:55:16 +01004099
Willy Tarreauceb24bc2010-11-09 12:46:41 +01004100 <path> is a UNIX socket path beginning with a slash ('/'). This is
Davor Ocelice9ed2812017-12-25 17:49:28 +01004101 alternative to the TCP listening port. HAProxy will then
Willy Tarreauceb24bc2010-11-09 12:46:41 +01004102 receive UNIX connections on the socket located at this place.
4103 The path must begin with a slash and by default is absolute.
4104 It can be relative to the prefix defined by "unix-bind" in
4105 the global section. Note that the total length of the prefix
4106 followed by the socket path cannot exceed some system limits
4107 for UNIX sockets, which commonly are set to 107 characters.
4108
Willy Tarreaub6205fd2012-09-24 12:27:33 +02004109 <param*> is a list of parameters common to all sockets declared on the
4110 same line. These numerous parameters depend on OS and build
4111 options and have a complete section dedicated to them. Please
4112 refer to section 5 to for more details.
Willy Tarreaua0ee1d02012-09-10 09:01:23 +02004113
Willy Tarreau0ba27502007-12-24 16:55:16 +01004114 It is possible to specify a list of address:port combinations delimited by
4115 commas. The frontend will then listen on all of these addresses. There is no
4116 fixed limit to the number of addresses and ports which can be listened on in
4117 a frontend, as well as there is no limit to the number of "bind" statements
4118 in a frontend.
4119
4120 Example :
4121 listen http_proxy
4122 bind :80,:443
4123 bind 10.0.0.1:10080,10.0.0.1:10443
Willy Tarreauceb24bc2010-11-09 12:46:41 +01004124 bind /var/run/ssl-frontend.sock user root mode 600 accept-proxy
Willy Tarreau0ba27502007-12-24 16:55:16 +01004125
Willy Tarreaua0ee1d02012-09-10 09:01:23 +02004126 listen http_https_proxy
4127 bind :80
Cyril Bonté0d44fc62012-10-09 22:45:33 +02004128 bind :443 ssl crt /etc/haproxy/site.pem
Willy Tarreaua0ee1d02012-09-10 09:01:23 +02004129
Willy Tarreau24709282013-03-10 21:32:12 +01004130 listen http_https_proxy_explicit
4131 bind ipv6@:80
4132 bind ipv4@public_ssl:443 ssl crt /etc/haproxy/site.pem
4133 bind unix@ssl-frontend.sock user root mode 600 accept-proxy
4134
Willy Tarreaudad36a32013-03-11 01:20:04 +01004135 listen external_bind_app1
William Lallemandb2f07452015-05-12 14:27:13 +02004136 bind "fd@${FD_APP1}"
Willy Tarreaudad36a32013-03-11 01:20:04 +01004137
Willy Tarreau55dcaf62015-09-27 15:03:15 +02004138 Note: regarding Linux's abstract namespace sockets, HAProxy uses the whole
4139 sun_path length is used for the address length. Some other programs
4140 such as socat use the string length only by default. Pass the option
4141 ",unix-tightsocklen=0" to any abstract socket definition in socat to
4142 make it compatible with HAProxy's.
4143
Willy Tarreauceb24bc2010-11-09 12:46:41 +01004144 See also : "source", "option forwardfor", "unix-bind" and the PROXY protocol
Willy Tarreaub6205fd2012-09-24 12:27:33 +02004145 documentation, and section 5 about bind options.
Willy Tarreau0ba27502007-12-24 16:55:16 +01004146
4147
Christopher Fauletff4121f2017-11-22 16:38:49 +01004148bind-process [ all | odd | even | <process_num>[-[<process_num>]] ] ...
Willy Tarreau0b9c02c2009-02-04 22:05:05 +01004149 Limit visibility of an instance to a certain set of processes numbers.
4150 May be used in sections : defaults | frontend | listen | backend
4151 yes | yes | yes | yes
4152 Arguments :
4153 all All process will see this instance. This is the default. It
4154 may be used to override a default value.
4155
Willy Tarreaua9db57e2013-01-18 11:29:29 +01004156 odd This instance will be enabled on processes 1,3,5,...63. This
Willy Tarreau0b9c02c2009-02-04 22:05:05 +01004157 option may be combined with other numbers.
4158
Willy Tarreaua9db57e2013-01-18 11:29:29 +01004159 even This instance will be enabled on processes 2,4,6,...64. This
Willy Tarreau0b9c02c2009-02-04 22:05:05 +01004160 option may be combined with other numbers. Do not use it
4161 with less than 2 processes otherwise some instances might be
4162 missing from all processes.
4163
Christopher Fauletff4121f2017-11-22 16:38:49 +01004164 process_num The instance will be enabled on this process number or range,
Willy Tarreaua9db57e2013-01-18 11:29:29 +01004165 whose values must all be between 1 and 32 or 64 depending on
Christopher Fauletff4121f2017-11-22 16:38:49 +01004166 the machine's word size. Ranges can be partially defined. The
4167 higher bound can be omitted. In such case, it is replaced by
4168 the corresponding maximum value. If a proxy is bound to
4169 process numbers greater than the configured global.nbproc, it
4170 will either be forced to process #1 if a single process was
Willy Tarreau102df612014-05-07 23:56:38 +02004171 specified, or to all processes otherwise.
Willy Tarreau0b9c02c2009-02-04 22:05:05 +01004172
4173 This keyword limits binding of certain instances to certain processes. This
4174 is useful in order not to have too many processes listening to the same
4175 ports. For instance, on a dual-core machine, it might make sense to set
4176 'nbproc 2' in the global section, then distributes the listeners among 'odd'
4177 and 'even' instances.
4178
Willy Tarreaua9db57e2013-01-18 11:29:29 +01004179 At the moment, it is not possible to reference more than 32 or 64 processes
4180 using this keyword, but this should be more than enough for most setups.
4181 Please note that 'all' really means all processes regardless of the machine's
4182 word size, and is not limited to the first 32 or 64.
Willy Tarreau0b9c02c2009-02-04 22:05:05 +01004183
Willy Tarreau6ae1ba62014-05-07 19:01:58 +02004184 Each "bind" line may further be limited to a subset of the proxy's processes,
4185 please consult the "process" bind keyword in section 5.1.
4186
Willy Tarreaub369a042014-09-16 13:21:03 +02004187 When a frontend has no explicit "bind-process" line, it tries to bind to all
4188 the processes referenced by its "bind" lines. That means that frontends can
4189 easily adapt to their listeners' processes.
4190
Willy Tarreau0b9c02c2009-02-04 22:05:05 +01004191 If some backends are referenced by frontends bound to other processes, the
4192 backend automatically inherits the frontend's processes.
4193
4194 Example :
4195 listen app_ip1
4196 bind 10.0.0.1:80
Willy Tarreaubfcd3112010-10-23 11:22:08 +02004197 bind-process odd
Willy Tarreau0b9c02c2009-02-04 22:05:05 +01004198
4199 listen app_ip2
4200 bind 10.0.0.2:80
Willy Tarreaubfcd3112010-10-23 11:22:08 +02004201 bind-process even
Willy Tarreau0b9c02c2009-02-04 22:05:05 +01004202
4203 listen management
4204 bind 10.0.0.3:80
Willy Tarreaubfcd3112010-10-23 11:22:08 +02004205 bind-process 1 2 3 4
Willy Tarreau0b9c02c2009-02-04 22:05:05 +01004206
Willy Tarreau110ecc12012-11-15 17:50:01 +01004207 listen management
4208 bind 10.0.0.4:80
4209 bind-process 1-4
4210
Willy Tarreau6ae1ba62014-05-07 19:01:58 +02004211 See also : "nbproc" in global section, and "process" in section 5.1.
Willy Tarreau0b9c02c2009-02-04 22:05:05 +01004212
4213
Willy Tarreau0ba27502007-12-24 16:55:16 +01004214capture cookie <name> len <length>
4215 Capture and log a cookie in the request and in the response.
4216 May be used in sections : defaults | frontend | listen | backend
4217 no | yes | yes | no
4218 Arguments :
4219 <name> is the beginning of the name of the cookie to capture. In order
4220 to match the exact name, simply suffix the name with an equal
4221 sign ('='). The full name will appear in the logs, which is
4222 useful with application servers which adjust both the cookie name
Davor Ocelice9ed2812017-12-25 17:49:28 +01004223 and value (e.g. ASPSESSIONXXX).
Willy Tarreau0ba27502007-12-24 16:55:16 +01004224
4225 <length> is the maximum number of characters to report in the logs, which
4226 include the cookie name, the equal sign and the value, all in the
4227 standard "name=value" form. The string will be truncated on the
4228 right if it exceeds <length>.
4229
4230 Only the first cookie is captured. Both the "cookie" request headers and the
4231 "set-cookie" response headers are monitored. This is particularly useful to
4232 check for application bugs causing session crossing or stealing between
4233 users, because generally the user's cookies can only change on a login page.
4234
4235 When the cookie was not presented by the client, the associated log column
4236 will report "-". When a request does not cause a cookie to be assigned by the
4237 server, a "-" is reported in the response column.
4238
4239 The capture is performed in the frontend only because it is necessary that
4240 the log format does not change for a given frontend depending on the
4241 backends. This may change in the future. Note that there can be only one
Willy Tarreau193b8c62012-11-22 00:17:38 +01004242 "capture cookie" statement in a frontend. The maximum capture length is set
4243 by the global "tune.http.cookielen" setting and defaults to 63 characters. It
4244 is not possible to specify a capture in a "defaults" section.
Willy Tarreau0ba27502007-12-24 16:55:16 +01004245
4246 Example:
4247 capture cookie ASPSESSION len 32
4248
4249 See also : "capture request header", "capture response header" as well as
Willy Tarreauc57f0e22009-05-10 13:12:33 +02004250 section 8 about logging.
Willy Tarreau0ba27502007-12-24 16:55:16 +01004251
4252
4253capture request header <name> len <length>
Willy Tarreau4460d032012-11-21 23:37:37 +01004254 Capture and log the last occurrence of the specified request header.
Willy Tarreau0ba27502007-12-24 16:55:16 +01004255 May be used in sections : defaults | frontend | listen | backend
4256 no | yes | yes | no
4257 Arguments :
4258 <name> is the name of the header to capture. The header names are not
Willy Tarreaud2a4aa22008-01-31 15:28:22 +01004259 case-sensitive, but it is a common practice to write them as they
Willy Tarreau0ba27502007-12-24 16:55:16 +01004260 appear in the requests, with the first letter of each word in
4261 upper case. The header name will not appear in the logs, only the
4262 value is reported, but the position in the logs is respected.
4263
4264 <length> is the maximum number of characters to extract from the value and
4265 report in the logs. The string will be truncated on the right if
4266 it exceeds <length>.
4267
Willy Tarreau4460d032012-11-21 23:37:37 +01004268 The complete value of the last occurrence of the header is captured. The
Willy Tarreau0ba27502007-12-24 16:55:16 +01004269 value will be added to the logs between braces ('{}'). If multiple headers
4270 are captured, they will be delimited by a vertical bar ('|') and will appear
Willy Tarreaucc6c8912009-02-22 10:53:55 +01004271 in the same order they were declared in the configuration. Non-existent
4272 headers will be logged just as an empty string. Common uses for request
4273 header captures include the "Host" field in virtual hosting environments, the
4274 "Content-length" when uploads are supported, "User-agent" to quickly
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01004275 differentiate between real users and robots, and "X-Forwarded-For" in proxied
Willy Tarreaucc6c8912009-02-22 10:53:55 +01004276 environments to find where the request came from.
4277
4278 Note that when capturing headers such as "User-agent", some spaces may be
4279 logged, making the log analysis more difficult. Thus be careful about what
4280 you log if you know your log parser is not smart enough to rely on the
4281 braces.
Willy Tarreau0ba27502007-12-24 16:55:16 +01004282
Willy Tarreau0900abb2012-11-22 00:21:46 +01004283 There is no limit to the number of captured request headers nor to their
4284 length, though it is wise to keep them low to limit memory usage per session.
4285 In order to keep log format consistent for a same frontend, header captures
4286 can only be declared in a frontend. It is not possible to specify a capture
4287 in a "defaults" section.
Willy Tarreau0ba27502007-12-24 16:55:16 +01004288
4289 Example:
4290 capture request header Host len 15
4291 capture request header X-Forwarded-For len 15
Cyril Bontéd1b0f7c2015-10-26 22:37:39 +01004292 capture request header Referer len 15
Willy Tarreau0ba27502007-12-24 16:55:16 +01004293
Willy Tarreauc57f0e22009-05-10 13:12:33 +02004294 See also : "capture cookie", "capture response header" as well as section 8
Willy Tarreau0ba27502007-12-24 16:55:16 +01004295 about logging.
4296
4297
4298capture response header <name> len <length>
Willy Tarreau4460d032012-11-21 23:37:37 +01004299 Capture and log the last occurrence of the specified response header.
Willy Tarreau0ba27502007-12-24 16:55:16 +01004300 May be used in sections : defaults | frontend | listen | backend
4301 no | yes | yes | no
4302 Arguments :
4303 <name> is the name of the header to capture. The header names are not
Willy Tarreaud2a4aa22008-01-31 15:28:22 +01004304 case-sensitive, but it is a common practice to write them as they
Willy Tarreau0ba27502007-12-24 16:55:16 +01004305 appear in the response, with the first letter of each word in
4306 upper case. The header name will not appear in the logs, only the
4307 value is reported, but the position in the logs is respected.
4308
4309 <length> is the maximum number of characters to extract from the value and
4310 report in the logs. The string will be truncated on the right if
4311 it exceeds <length>.
4312
Willy Tarreau4460d032012-11-21 23:37:37 +01004313 The complete value of the last occurrence of the header is captured. The
Willy Tarreau0ba27502007-12-24 16:55:16 +01004314 result will be added to the logs between braces ('{}') after the captured
4315 request headers. If multiple headers are captured, they will be delimited by
4316 a vertical bar ('|') and will appear in the same order they were declared in
Willy Tarreaucc6c8912009-02-22 10:53:55 +01004317 the configuration. Non-existent headers will be logged just as an empty
4318 string. Common uses for response header captures include the "Content-length"
4319 header which indicates how many bytes are expected to be returned, the
4320 "Location" header to track redirections.
Willy Tarreau0ba27502007-12-24 16:55:16 +01004321
Willy Tarreau0900abb2012-11-22 00:21:46 +01004322 There is no limit to the number of captured response headers nor to their
4323 length, though it is wise to keep them low to limit memory usage per session.
4324 In order to keep log format consistent for a same frontend, header captures
4325 can only be declared in a frontend. It is not possible to specify a capture
4326 in a "defaults" section.
Willy Tarreau0ba27502007-12-24 16:55:16 +01004327
4328 Example:
4329 capture response header Content-length len 9
4330 capture response header Location len 15
4331
Willy Tarreauc57f0e22009-05-10 13:12:33 +02004332 See also : "capture cookie", "capture request header" as well as section 8
Willy Tarreau0ba27502007-12-24 16:55:16 +01004333 about logging.
4334
4335
MIZUTA Takeshib24bc0d2020-07-09 11:13:20 +09004336clitcpka-cnt <count>
4337 Sets the maximum number of keepalive probes TCP should send before dropping
4338 the connection on the client side.
4339 May be used in sections : defaults | frontend | listen | backend
4340 yes | yes | yes | no
4341 Arguments :
4342 <count> is the maximum number of keepalive probes.
4343
4344 This keyword corresponds to the socket option TCP_KEEPCNT. If this keyword
4345 is not specified, system-wide TCP parameter (tcp_keepalive_probes) is used.
Willy Tarreau52543212020-07-09 05:58:51 +02004346 The availability of this setting depends on the operating system. It is
4347 known to work on Linux.
MIZUTA Takeshib24bc0d2020-07-09 11:13:20 +09004348
4349 See also : "option clitcpka", "clitcpka-idle", "clitcpka-intvl".
4350
4351
4352clitcpka-idle <timeout>
4353 Sets the time the connection needs to remain idle before TCP starts sending
4354 keepalive probes, if enabled the sending of TCP keepalive packets on the
4355 client side.
4356 May be used in sections : defaults | frontend | listen | backend
4357 yes | yes | yes | no
4358 Arguments :
4359 <timeout> is the time the connection needs to remain idle before TCP starts
4360 sending keepalive probes. It is specified in seconds by default,
4361 but can be in any other unit if the number is suffixed by the
4362 unit, as explained at the top of this document.
4363
4364 This keyword corresponds to the socket option TCP_KEEPIDLE. If this keyword
4365 is not specified, system-wide TCP parameter (tcp_keepalive_time) is used.
Willy Tarreau52543212020-07-09 05:58:51 +02004366 The availability of this setting depends on the operating system. It is
4367 known to work on Linux.
MIZUTA Takeshib24bc0d2020-07-09 11:13:20 +09004368
4369 See also : "option clitcpka", "clitcpka-cnt", "clitcpka-intvl".
4370
4371
4372clitcpka-intvl <timeout>
4373 Sets the time between individual keepalive probes on the client side.
4374 May be used in sections : defaults | frontend | listen | backend
4375 yes | yes | yes | no
4376 Arguments :
4377 <timeout> is the time between individual keepalive probes. It is specified
4378 in seconds by default, but can be in any other unit if the number
4379 is suffixed by the unit, as explained at the top of this
4380 document.
4381
4382 This keyword corresponds to the socket option TCP_KEEPINTVL. If this keyword
4383 is not specified, system-wide TCP parameter (tcp_keepalive_intvl) is used.
Willy Tarreau52543212020-07-09 05:58:51 +02004384 The availability of this setting depends on the operating system. It is
4385 known to work on Linux.
MIZUTA Takeshib24bc0d2020-07-09 11:13:20 +09004386
4387 See also : "option clitcpka", "clitcpka-cnt", "clitcpka-idle".
4388
4389
Cyril Bonté316a8cf2012-11-11 13:38:27 +01004390compression algo <algorithm> ...
4391compression type <mime type> ...
Willy Tarreau70737d12012-10-27 00:34:28 +02004392compression offload
William Lallemand82fe75c2012-10-23 10:25:10 +02004393 Enable HTTP compression.
4394 May be used in sections : defaults | frontend | listen | backend
4395 yes | yes | yes | yes
4396 Arguments :
Cyril Bonté316a8cf2012-11-11 13:38:27 +01004397 algo is followed by the list of supported compression algorithms.
4398 type is followed by the list of MIME types that will be compressed.
Daniel Corbett9f0843f2021-05-08 10:50:37 -04004399 offload makes HAProxy work as a compression offloader only (see notes).
Cyril Bonté316a8cf2012-11-11 13:38:27 +01004400
4401 The currently supported algorithms are :
Willy Tarreauc91840a2015-03-28 17:00:39 +01004402 identity this is mostly for debugging, and it was useful for developing
4403 the compression feature. Identity does not apply any change on
4404 data.
Cyril Bonté316a8cf2012-11-11 13:38:27 +01004405
Willy Tarreauc91840a2015-03-28 17:00:39 +01004406 gzip applies gzip compression. This setting is only available when
Baptiste Assmannf085d632015-12-21 17:57:32 +01004407 support for zlib or libslz was built in.
Willy Tarreauc91840a2015-03-28 17:00:39 +01004408
4409 deflate same as "gzip", but with deflate algorithm and zlib format.
4410 Note that this algorithm has ambiguous support on many
4411 browsers and no support at all from recent ones. It is
4412 strongly recommended not to use it for anything else than
4413 experimentation. This setting is only available when support
Baptiste Assmannf085d632015-12-21 17:57:32 +01004414 for zlib or libslz was built in.
Cyril Bonté316a8cf2012-11-11 13:38:27 +01004415
Willy Tarreauc91840a2015-03-28 17:00:39 +01004416 raw-deflate same as "deflate" without the zlib wrapper, and used as an
4417 alternative when the browser wants "deflate". All major
4418 browsers understand it and despite violating the standards,
4419 it is known to work better than "deflate", at least on MSIE
4420 and some versions of Safari. Do not use it in conjunction
4421 with "deflate", use either one or the other since both react
4422 to the same Accept-Encoding token. This setting is only
Baptiste Assmannf085d632015-12-21 17:57:32 +01004423 available when support for zlib or libslz was built in.
Cyril Bonté316a8cf2012-11-11 13:38:27 +01004424
Dmitry Sivachenko87c208b2012-11-22 20:03:26 +04004425 Compression will be activated depending on the Accept-Encoding request
Cyril Bonté316a8cf2012-11-11 13:38:27 +01004426 header. With identity, it does not take care of that header.
Dmitry Sivachenkoc9f3b452012-11-28 17:47:11 +04004427 If backend servers support HTTP compression, these directives
Daniel Corbett9f0843f2021-05-08 10:50:37 -04004428 will be no-op: HAProxy will see the compressed response and will not
Dmitry Sivachenkoc9f3b452012-11-28 17:47:11 +04004429 compress again. If backend servers do not support HTTP compression and
Daniel Corbett9f0843f2021-05-08 10:50:37 -04004430 there is Accept-Encoding header in request, HAProxy will compress the
Dmitry Sivachenkoc9f3b452012-11-28 17:47:11 +04004431 matching response.
Willy Tarreau70737d12012-10-27 00:34:28 +02004432
Daniel Corbett9f0843f2021-05-08 10:50:37 -04004433 The "offload" setting makes HAProxy remove the Accept-Encoding header to
Willy Tarreau70737d12012-10-27 00:34:28 +02004434 prevent backend servers from compressing responses. It is strongly
4435 recommended not to do this because this means that all the compression work
Daniel Corbett9f0843f2021-05-08 10:50:37 -04004436 will be done on the single point where HAProxy is located. However in some
4437 deployment scenarios, HAProxy may be installed in front of a buggy gateway
Dmitry Sivachenkoc9f3b452012-11-28 17:47:11 +04004438 with broken HTTP compression implementation which can't be turned off.
Daniel Corbett9f0843f2021-05-08 10:50:37 -04004439 In that case HAProxy can be used to prevent that gateway from emitting
Dmitry Sivachenkoc9f3b452012-11-28 17:47:11 +04004440 invalid payloads. In this case, simply removing the header in the
4441 configuration does not work because it applies before the header is parsed,
Daniel Corbett9f0843f2021-05-08 10:50:37 -04004442 so that prevents HAProxy from compressing. The "offload" setting should
Willy Tarreauffea9fd2014-07-12 16:37:02 +02004443 then be used for such scenarios. Note: for now, the "offload" setting is
4444 ignored when set in a defaults section.
William Lallemand82fe75c2012-10-23 10:25:10 +02004445
William Lallemand05097442012-11-20 12:14:28 +01004446 Compression is disabled when:
Baptiste Assmann650d53d2013-01-05 15:44:44 +01004447 * the request does not advertise a supported compression algorithm in the
4448 "Accept-Encoding" header
Julien Pivottoff80c822021-03-29 12:41:40 +02004449 * the response message is not HTTP/1.1 or above
Tim Duesterhusbb48c9a2019-01-30 23:46:04 +01004450 * HTTP status code is not one of 200, 201, 202, or 203
Baptiste Assmann650d53d2013-01-05 15:44:44 +01004451 * response contain neither a "Content-Length" header nor a
4452 "Transfer-Encoding" whose last value is "chunked"
4453 * response contains a "Content-Type" header whose first value starts with
4454 "multipart"
4455 * the response contains the "no-transform" value in the "Cache-control"
4456 header
4457 * User-Agent matches "Mozilla/4" unless it is MSIE 6 with XP SP2, or MSIE 7
4458 and later
4459 * The response contains a "Content-Encoding" header, indicating that the
4460 response is already compressed (see compression offload)
Tim Duesterhusbb48c9a2019-01-30 23:46:04 +01004461 * The response contains an invalid "ETag" header or multiple ETag headers
William Lallemand05097442012-11-20 12:14:28 +01004462
Tim Duesterhusb229f012019-01-29 16:38:56 +01004463 Note: The compression does not emit the Warning header.
William Lallemand05097442012-11-20 12:14:28 +01004464
William Lallemand82fe75c2012-10-23 10:25:10 +02004465 Examples :
4466 compression algo gzip
4467 compression type text/html text/plain
Willy Tarreau0ba27502007-12-24 16:55:16 +01004468
Christopher Fauletc3fe5332016-04-07 15:30:10 +02004469
Willy Tarreau55165fe2009-05-10 12:02:55 +02004470cookie <name> [ rewrite | insert | prefix ] [ indirect ] [ nocache ]
Willy Tarreau4992dd22012-05-31 21:02:17 +02004471 [ postonly ] [ preserve ] [ httponly ] [ secure ]
4472 [ domain <domain> ]* [ maxidle <idle> ] [ maxlife <life> ]
Christopher Faulet2f533902020-01-21 11:06:48 +01004473 [ dynamic ] [ attr <value> ]*
Willy Tarreau0ba27502007-12-24 16:55:16 +01004474 Enable cookie-based persistence in a backend.
4475 May be used in sections : defaults | frontend | listen | backend
4476 yes | no | yes | yes
4477 Arguments :
4478 <name> is the name of the cookie which will be monitored, modified or
4479 inserted in order to bring persistence. This cookie is sent to
4480 the client via a "Set-Cookie" header in the response, and is
4481 brought back by the client in a "Cookie" header in all requests.
4482 Special care should be taken to choose a name which does not
4483 conflict with any likely application cookie. Also, if the same
Davor Ocelice9ed2812017-12-25 17:49:28 +01004484 backends are subject to be used by the same clients (e.g.
Willy Tarreau0ba27502007-12-24 16:55:16 +01004485 HTTP/HTTPS), care should be taken to use different cookie names
4486 between all backends if persistence between them is not desired.
4487
4488 rewrite This keyword indicates that the cookie will be provided by the
Daniel Corbett9f0843f2021-05-08 10:50:37 -04004489 server and that HAProxy will have to modify its value to set the
Willy Tarreau0ba27502007-12-24 16:55:16 +01004490 server's identifier in it. This mode is handy when the management
4491 of complex combinations of "Set-cookie" and "Cache-control"
4492 headers is left to the application. The application can then
4493 decide whether or not it is appropriate to emit a persistence
Lukas Tribusf01a9cd2016-02-03 18:09:37 +01004494 cookie. Since all responses should be monitored, this mode
4495 doesn't work in HTTP tunnel mode. Unless the application
Davor Ocelice9ed2812017-12-25 17:49:28 +01004496 behavior is very complex and/or broken, it is advised not to
Lukas Tribusf01a9cd2016-02-03 18:09:37 +01004497 start with this mode for new deployments. This keyword is
4498 incompatible with "insert" and "prefix".
Willy Tarreau0ba27502007-12-24 16:55:16 +01004499
4500 insert This keyword indicates that the persistence cookie will have to
Daniel Corbett9f0843f2021-05-08 10:50:37 -04004501 be inserted by HAProxy in server responses if the client did not
Willy Tarreauba4c5be2010-10-23 12:46:42 +02004502
Willy Tarreaua79094d2010-08-31 22:54:15 +02004503 already have a cookie that would have permitted it to access this
Willy Tarreauba4c5be2010-10-23 12:46:42 +02004504 server. When used without the "preserve" option, if the server
Michael Prokop4438c602019-05-24 10:25:45 +02004505 emits a cookie with the same name, it will be removed before
Davor Ocelice9ed2812017-12-25 17:49:28 +01004506 processing. For this reason, this mode can be used to upgrade
Willy Tarreauba4c5be2010-10-23 12:46:42 +02004507 existing configurations running in the "rewrite" mode. The cookie
4508 will only be a session cookie and will not be stored on the
4509 client's disk. By default, unless the "indirect" option is added,
4510 the server will see the cookies emitted by the client. Due to
4511 caching effects, it is generally wise to add the "nocache" or
4512 "postonly" keywords (see below). The "insert" keyword is not
4513 compatible with "rewrite" and "prefix".
Willy Tarreau0ba27502007-12-24 16:55:16 +01004514
4515 prefix This keyword indicates that instead of relying on a dedicated
4516 cookie for the persistence, an existing one will be completed.
4517 This may be needed in some specific environments where the client
4518 does not support more than one single cookie and the application
4519 already needs it. In this case, whenever the server sets a cookie
4520 named <name>, it will be prefixed with the server's identifier
4521 and a delimiter. The prefix will be removed from all client
4522 requests so that the server still finds the cookie it emitted.
4523 Since all requests and responses are subject to being modified,
Lukas Tribusf01a9cd2016-02-03 18:09:37 +01004524 this mode doesn't work with tunnel mode. The "prefix" keyword is
Willy Tarreau37229df2011-10-17 12:24:55 +02004525 not compatible with "rewrite" and "insert". Note: it is highly
4526 recommended not to use "indirect" with "prefix", otherwise server
4527 cookie updates would not be sent to clients.
Willy Tarreau0ba27502007-12-24 16:55:16 +01004528
Willy Tarreaua79094d2010-08-31 22:54:15 +02004529 indirect When this option is specified, no cookie will be emitted to a
4530 client which already has a valid one for the server which has
4531 processed the request. If the server sets such a cookie itself,
Willy Tarreauba4c5be2010-10-23 12:46:42 +02004532 it will be removed, unless the "preserve" option is also set. In
4533 "insert" mode, this will additionally remove cookies from the
4534 requests transmitted to the server, making the persistence
4535 mechanism totally transparent from an application point of view.
Willy Tarreau37229df2011-10-17 12:24:55 +02004536 Note: it is highly recommended not to use "indirect" with
4537 "prefix", otherwise server cookie updates would not be sent to
4538 clients.
Willy Tarreau0ba27502007-12-24 16:55:16 +01004539
4540 nocache This option is recommended in conjunction with the insert mode
4541 when there is a cache between the client and HAProxy, as it
4542 ensures that a cacheable response will be tagged non-cacheable if
4543 a cookie needs to be inserted. This is important because if all
4544 persistence cookies are added on a cacheable home page for
4545 instance, then all customers will then fetch the page from an
4546 outer cache and will all share the same persistence cookie,
4547 leading to one server receiving much more traffic than others.
4548 See also the "insert" and "postonly" options.
4549
4550 postonly This option ensures that cookie insertion will only be performed
4551 on responses to POST requests. It is an alternative to the
4552 "nocache" option, because POST responses are not cacheable, so
4553 this ensures that the persistence cookie will never get cached.
4554 Since most sites do not need any sort of persistence before the
4555 first POST which generally is a login request, this is a very
4556 efficient method to optimize caching without risking to find a
4557 persistence cookie in the cache.
4558 See also the "insert" and "nocache" options.
4559
Willy Tarreauba4c5be2010-10-23 12:46:42 +02004560 preserve This option may only be used with "insert" and/or "indirect". It
4561 allows the server to emit the persistence cookie itself. In this
Daniel Corbett9f0843f2021-05-08 10:50:37 -04004562 case, if a cookie is found in the response, HAProxy will leave it
Willy Tarreauba4c5be2010-10-23 12:46:42 +02004563 untouched. This is useful in order to end persistence after a
4564 logout request for instance. For this, the server just has to
Davor Ocelice9ed2812017-12-25 17:49:28 +01004565 emit a cookie with an invalid value (e.g. empty) or with a date in
Willy Tarreauba4c5be2010-10-23 12:46:42 +02004566 the past. By combining this mechanism with the "disable-on-404"
4567 check option, it is possible to perform a completely graceful
4568 shutdown because users will definitely leave the server after
4569 they logout.
4570
Daniel Corbett9f0843f2021-05-08 10:50:37 -04004571 httponly This option tells HAProxy to add an "HttpOnly" cookie attribute
Willy Tarreau4992dd22012-05-31 21:02:17 +02004572 when a cookie is inserted. This attribute is used so that a
4573 user agent doesn't share the cookie with non-HTTP components.
4574 Please check RFC6265 for more information on this attribute.
4575
Daniel Corbett9f0843f2021-05-08 10:50:37 -04004576 secure This option tells HAProxy to add a "Secure" cookie attribute when
Willy Tarreau4992dd22012-05-31 21:02:17 +02004577 a cookie is inserted. This attribute is used so that a user agent
4578 never emits this cookie over non-secure channels, which means
4579 that a cookie learned with this flag will be presented only over
4580 SSL/TLS connections. Please check RFC6265 for more information on
4581 this attribute.
4582
Krzysztof Piotr Oledzkiefe3b6f2008-05-23 23:49:32 +02004583 domain This option allows to specify the domain at which a cookie is
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01004584 inserted. It requires exactly one parameter: a valid domain
Willy Tarreau68a897b2009-12-03 23:28:34 +01004585 name. If the domain begins with a dot, the browser is allowed to
4586 use it for any host ending with that name. It is also possible to
4587 specify several domain names by invoking this option multiple
4588 times. Some browsers might have small limits on the number of
4589 domains, so be careful when doing that. For the record, sending
4590 10 domains to MSIE 6 or Firefox 2 works as expected.
Krzysztof Piotr Oledzkiefe3b6f2008-05-23 23:49:32 +02004591
Willy Tarreau996a92c2010-10-13 19:30:47 +02004592 maxidle This option allows inserted cookies to be ignored after some idle
4593 time. It only works with insert-mode cookies. When a cookie is
4594 sent to the client, the date this cookie was emitted is sent too.
4595 Upon further presentations of this cookie, if the date is older
4596 than the delay indicated by the parameter (in seconds), it will
4597 be ignored. Otherwise, it will be refreshed if needed when the
4598 response is sent to the client. This is particularly useful to
4599 prevent users who never close their browsers from remaining for
Davor Ocelice9ed2812017-12-25 17:49:28 +01004600 too long on the same server (e.g. after a farm size change). When
Willy Tarreau996a92c2010-10-13 19:30:47 +02004601 this option is set and a cookie has no date, it is always
4602 accepted, but gets refreshed in the response. This maintains the
4603 ability for admins to access their sites. Cookies that have a
4604 date in the future further than 24 hours are ignored. Doing so
4605 lets admins fix timezone issues without risking kicking users off
4606 the site.
4607
4608 maxlife This option allows inserted cookies to be ignored after some life
4609 time, whether they're in use or not. It only works with insert
4610 mode cookies. When a cookie is first sent to the client, the date
4611 this cookie was emitted is sent too. Upon further presentations
4612 of this cookie, if the date is older than the delay indicated by
4613 the parameter (in seconds), it will be ignored. If the cookie in
4614 the request has no date, it is accepted and a date will be set.
4615 Cookies that have a date in the future further than 24 hours are
4616 ignored. Doing so lets admins fix timezone issues without risking
4617 kicking users off the site. Contrary to maxidle, this value is
4618 not refreshed, only the first visit date counts. Both maxidle and
4619 maxlife may be used at the time. This is particularly useful to
4620 prevent users who never close their browsers from remaining for
Davor Ocelice9ed2812017-12-25 17:49:28 +01004621 too long on the same server (e.g. after a farm size change). This
Willy Tarreau996a92c2010-10-13 19:30:47 +02004622 is stronger than the maxidle method in that it forces a
4623 redispatch after some absolute delay.
4624
Olivier Houchard4e694042017-03-14 20:01:29 +01004625 dynamic Activate dynamic cookies. When used, a session cookie is
4626 dynamically created for each server, based on the IP and port
4627 of the server, and a secret key, specified in the
4628 "dynamic-cookie-key" backend directive.
4629 The cookie will be regenerated each time the IP address change,
4630 and is only generated for IPv4/IPv6.
4631
Daniel Corbett9f0843f2021-05-08 10:50:37 -04004632 attr This option tells HAProxy to add an extra attribute when a
Christopher Faulet2f533902020-01-21 11:06:48 +01004633 cookie is inserted. The attribute value can contain any
4634 characters except control ones or ";". This option may be
4635 repeated.
4636
Willy Tarreau0ba27502007-12-24 16:55:16 +01004637 There can be only one persistence cookie per HTTP backend, and it can be
4638 declared in a defaults section. The value of the cookie will be the value
4639 indicated after the "cookie" keyword in a "server" statement. If no cookie
4640 is declared for a given server, the cookie is not set.
Willy Tarreau6a06a402007-07-15 20:15:28 +02004641
Willy Tarreau0ba27502007-12-24 16:55:16 +01004642 Examples :
4643 cookie JSESSIONID prefix
4644 cookie SRV insert indirect nocache
4645 cookie SRV insert postonly indirect
Willy Tarreau996a92c2010-10-13 19:30:47 +02004646 cookie SRV insert indirect nocache maxidle 30m maxlife 8h
Willy Tarreau0ba27502007-12-24 16:55:16 +01004647
Willy Tarreau294d0f02015-08-10 19:40:12 +02004648 See also : "balance source", "capture cookie", "server" and "ignore-persist".
Willy Tarreau0ba27502007-12-24 16:55:16 +01004649
Willy Tarreau983e01e2010-01-11 18:42:06 +01004650
Thierry FOURNIERa0a1b752015-05-26 17:44:32 +02004651declare capture [ request | response ] len <length>
4652 Declares a capture slot.
4653 May be used in sections : defaults | frontend | listen | backend
4654 no | yes | yes | no
4655 Arguments:
4656 <length> is the length allowed for the capture.
4657
4658 This declaration is only available in the frontend or listen section, but the
4659 reserved slot can be used in the backends. The "request" keyword allocates a
4660 capture slot for use in the request, and "response" allocates a capture slot
4661 for use in the response.
4662
4663 See also: "capture-req", "capture-res" (sample converters),
Baptiste Assmann5ac425c2015-10-21 23:13:46 +02004664 "capture.req.hdr", "capture.res.hdr" (sample fetches),
Thierry FOURNIERa0a1b752015-05-26 17:44:32 +02004665 "http-request capture" and "http-response capture".
4666
4667
Krzysztof Piotr Oledzkic6df0662010-01-05 16:38:49 +01004668default-server [param*]
4669 Change default options for a server in a backend
4670 May be used in sections : defaults | frontend | listen | backend
4671 yes | no | yes | yes
4672 Arguments:
Willy Tarreau983e01e2010-01-11 18:42:06 +01004673 <param*> is a list of parameters for this server. The "default-server"
4674 keyword accepts an important number of options and has a complete
4675 section dedicated to it. Please refer to section 5 for more
4676 details.
Krzysztof Piotr Oledzkic6df0662010-01-05 16:38:49 +01004677
Willy Tarreau983e01e2010-01-11 18:42:06 +01004678 Example :
Krzysztof Piotr Oledzkic6df0662010-01-05 16:38:49 +01004679 default-server inter 1000 weight 13
4680
4681 See also: "server" and section 5 about server options
Willy Tarreau0ba27502007-12-24 16:55:16 +01004682
Willy Tarreau983e01e2010-01-11 18:42:06 +01004683
Willy Tarreau0ba27502007-12-24 16:55:16 +01004684default_backend <backend>
4685 Specify the backend to use when no "use_backend" rule has been matched.
4686 May be used in sections : defaults | frontend | listen | backend
4687 yes | yes | yes | no
4688 Arguments :
4689 <backend> is the name of the backend to use.
4690
4691 When doing content-switching between frontend and backends using the
4692 "use_backend" keyword, it is often useful to indicate which backend will be
4693 used when no rule has matched. It generally is the dynamic backend which
4694 will catch all undetermined requests.
4695
Willy Tarreau0ba27502007-12-24 16:55:16 +01004696 Example :
4697
4698 use_backend dynamic if url_dyn
4699 use_backend static if url_css url_img extension_img
4700 default_backend dynamic
4701
Willy Tarreau98d04852015-05-26 12:18:29 +02004702 See also : "use_backend"
Willy Tarreau2769aa02007-12-27 18:26:09 +01004703
Willy Tarreau0ba27502007-12-24 16:55:16 +01004704
Baptiste Assmann27f51342013-10-09 06:51:49 +02004705description <string>
4706 Describe a listen, frontend or backend.
4707 May be used in sections : defaults | frontend | listen | backend
4708 no | yes | yes | yes
4709 Arguments : string
4710
4711 Allows to add a sentence to describe the related object in the HAProxy HTML
4712 stats page. The description will be printed on the right of the object name
4713 it describes.
4714 No need to backslash spaces in the <string> arguments.
4715
4716
Willy Tarreau0ba27502007-12-24 16:55:16 +01004717disabled
4718 Disable a proxy, frontend or backend.
4719 May be used in sections : defaults | frontend | listen | backend
4720 yes | yes | yes | yes
4721 Arguments : none
4722
4723 The "disabled" keyword is used to disable an instance, mainly in order to
4724 liberate a listening port or to temporarily disable a service. The instance
4725 will still be created and its configuration will be checked, but it will be
4726 created in the "stopped" state and will appear as such in the statistics. It
4727 will not receive any traffic nor will it send any health-checks or logs. It
4728 is possible to disable many instances at once by adding the "disabled"
4729 keyword in a "defaults" section.
4730
4731 See also : "enabled"
4732
4733
Willy Tarreau5ce94572010-06-07 14:35:41 +02004734dispatch <address>:<port>
4735 Set a default server address
4736 May be used in sections : defaults | frontend | listen | backend
4737 no | no | yes | yes
Cyril Bonté108cf6e2012-04-21 23:30:29 +02004738 Arguments :
Willy Tarreau5ce94572010-06-07 14:35:41 +02004739
4740 <address> is the IPv4 address of the default server. Alternatively, a
4741 resolvable hostname is supported, but this name will be resolved
4742 during start-up.
4743
4744 <ports> is a mandatory port specification. All connections will be sent
4745 to this port, and it is not permitted to use port offsets as is
4746 possible with normal servers.
4747
Willy Tarreau787aed52011-04-15 06:45:37 +02004748 The "dispatch" keyword designates a default server for use when no other
Willy Tarreau5ce94572010-06-07 14:35:41 +02004749 server can take the connection. In the past it was used to forward non
4750 persistent connections to an auxiliary load balancer. Due to its simple
4751 syntax, it has also been used for simple TCP relays. It is recommended not to
4752 use it for more clarity, and to use the "server" directive instead.
4753
4754 See also : "server"
4755
Olivier Houchard4e694042017-03-14 20:01:29 +01004756
4757dynamic-cookie-key <string>
4758 Set the dynamic cookie secret key for a backend.
4759 May be used in sections : defaults | frontend | listen | backend
4760 yes | no | yes | yes
4761 Arguments : The secret key to be used.
4762
4763 When dynamic cookies are enabled (see the "dynamic" directive for cookie),
Davor Ocelice9ed2812017-12-25 17:49:28 +01004764 a dynamic cookie is created for each server (unless one is explicitly
Olivier Houchard4e694042017-03-14 20:01:29 +01004765 specified on the "server" line), using a hash of the IP address of the
4766 server, the TCP port, and the secret key.
Davor Ocelice9ed2812017-12-25 17:49:28 +01004767 That way, we can ensure session persistence across multiple load-balancers,
Olivier Houchard4e694042017-03-14 20:01:29 +01004768 even if servers are dynamically added or removed.
Willy Tarreau5ce94572010-06-07 14:35:41 +02004769
Willy Tarreau0ba27502007-12-24 16:55:16 +01004770enabled
4771 Enable a proxy, frontend or backend.
4772 May be used in sections : defaults | frontend | listen | backend
4773 yes | yes | yes | yes
4774 Arguments : none
4775
4776 The "enabled" keyword is used to explicitly enable an instance, when the
4777 defaults has been set to "disabled". This is very rarely used.
4778
4779 See also : "disabled"
4780
4781
4782errorfile <code> <file>
4783 Return a file contents instead of errors generated by HAProxy
4784 May be used in sections : defaults | frontend | listen | backend
4785 yes | yes | yes | yes
4786 Arguments :
4787 <code> is the HTTP status code. Currently, HAProxy is capable of
Christopher Faulet612f2ea2020-05-27 09:57:28 +02004788 generating codes 200, 400, 401, 403, 404, 405, 407, 408, 410,
Christopher Faulete095f312020-12-07 11:22:24 +01004789 413, 425, 429, 500, 501, 502, 503, and 504.
Willy Tarreau0ba27502007-12-24 16:55:16 +01004790
4791 <file> designates a file containing the full HTTP response. It is
Willy Tarreaud2a4aa22008-01-31 15:28:22 +01004792 recommended to follow the common practice of appending ".http" to
Willy Tarreau0ba27502007-12-24 16:55:16 +01004793 the filename so that people do not confuse the response with HTML
Willy Tarreau59140a22009-02-22 12:02:30 +01004794 error pages, and to use absolute paths, since files are read
4795 before any chroot is performed.
Willy Tarreau0ba27502007-12-24 16:55:16 +01004796
4797 It is important to understand that this keyword is not meant to rewrite
4798 errors returned by the server, but errors detected and returned by HAProxy.
4799 This is why the list of supported errors is limited to a small set.
4800
Willy Tarreauae94d4d2011-05-11 16:28:49 +02004801 Code 200 is emitted in response to requests matching a "monitor-uri" rule.
4802
Christopher Faulet70170672020-05-18 17:42:48 +02004803 The files are parsed when HAProxy starts and must be valid according to the
4804 HTTP specification. They should not exceed the configured buffer size
4805 (BUFSIZE), which generally is 16 kB, otherwise an internal error will be
4806 returned. It is also wise not to put any reference to local contents
4807 (e.g. images) in order to avoid loops between the client and HAProxy when all
4808 servers are down, causing an error to be returned instead of an
4809 image. Finally, The response cannot exceed (tune.bufsize - tune.maxrewrite)
4810 so that "http-after-response" rules still have room to operate (see
4811 "tune.maxrewrite").
Willy Tarreau59140a22009-02-22 12:02:30 +01004812
Willy Tarreau0ba27502007-12-24 16:55:16 +01004813 The files are read at the same time as the configuration and kept in memory.
4814 For this reason, the errors continue to be returned even when the process is
4815 chrooted, and no file change is considered while the process is running. A
Willy Tarreauc27debf2008-01-06 08:57:02 +01004816 simple method for developing those files consists in associating them to the
Willy Tarreau0ba27502007-12-24 16:55:16 +01004817 403 status code and interrogating a blocked URL.
4818
Christopher Faulet3b967c12020-05-15 15:47:44 +02004819 See also : "http-error", "errorloc", "errorloc302", "errorloc303"
Willy Tarreau0ba27502007-12-24 16:55:16 +01004820
Willy Tarreau59140a22009-02-22 12:02:30 +01004821 Example :
4822 errorfile 400 /etc/haproxy/errorfiles/400badreq.http
Willy Tarreau989222a2016-01-15 10:26:26 +01004823 errorfile 408 /dev/null # work around Chrome pre-connect bug
Willy Tarreau59140a22009-02-22 12:02:30 +01004824 errorfile 403 /etc/haproxy/errorfiles/403forbid.http
4825 errorfile 503 /etc/haproxy/errorfiles/503sorry.http
4826
Willy Tarreau2769aa02007-12-27 18:26:09 +01004827
Christopher Faulet76edc0f2020-01-13 15:52:01 +01004828errorfiles <name> [<code> ...]
4829 Import, fully or partially, the error files defined in the <name> http-errors
4830 section.
4831 May be used in sections : defaults | frontend | listen | backend
4832 yes | yes | yes | yes
4833 Arguments :
4834 <name> is the name of an existing http-errors section.
4835
4836 <code> is a HTTP status code. Several status code may be listed.
Christopher Faulet612f2ea2020-05-27 09:57:28 +02004837 Currently, HAProxy is capable of generating codes 200, 400, 401,
Christopher Faulete095f312020-12-07 11:22:24 +01004838 403, 404, 405, 407, 408, 410, 413, 425, 429, 500, 501, 502, 503,
4839 and 504.
Christopher Faulet76edc0f2020-01-13 15:52:01 +01004840
4841 Errors defined in the http-errors section with the name <name> are imported
4842 in the current proxy. If no status code is specified, all error files of the
4843 http-errors section are imported. Otherwise, only error files associated to
4844 the listed status code are imported. Those error files override the already
4845 defined custom errors for the proxy. And they may be overridden by following
Daniel Corbett67a82712020-07-06 23:01:19 -04004846 ones. Functionally, it is exactly the same as declaring all error files by
Christopher Faulet76edc0f2020-01-13 15:52:01 +01004847 hand using "errorfile" directives.
4848
Christopher Faulet3b967c12020-05-15 15:47:44 +02004849 See also : "http-error", "errorfile", "errorloc", "errorloc302" ,
4850 "errorloc303" and section 3.8 about http-errors.
Christopher Faulet76edc0f2020-01-13 15:52:01 +01004851
4852 Example :
4853 errorfiles generic
4854 errorfiles site-1 403 404
4855
4856
Willy Tarreau2769aa02007-12-27 18:26:09 +01004857errorloc <code> <url>
4858errorloc302 <code> <url>
4859 Return an HTTP redirection to a URL instead of errors generated by HAProxy
4860 May be used in sections : defaults | frontend | listen | backend
4861 yes | yes | yes | yes
4862 Arguments :
4863 <code> is the HTTP status code. Currently, HAProxy is capable of
Christopher Faulet612f2ea2020-05-27 09:57:28 +02004864 generating codes 200, 400, 401, 403, 404, 405, 407, 408, 410,
Christopher Faulete095f312020-12-07 11:22:24 +01004865 413, 425, 429, 500, 501, 502, 503, and 504.
Willy Tarreau2769aa02007-12-27 18:26:09 +01004866
4867 <url> it is the exact contents of the "Location" header. It may contain
4868 either a relative URI to an error page hosted on the same site,
4869 or an absolute URI designating an error page on another site.
4870 Special care should be given to relative URIs to avoid redirect
Davor Ocelice9ed2812017-12-25 17:49:28 +01004871 loops if the URI itself may generate the same error (e.g. 500).
Willy Tarreau2769aa02007-12-27 18:26:09 +01004872
4873 It is important to understand that this keyword is not meant to rewrite
4874 errors returned by the server, but errors detected and returned by HAProxy.
4875 This is why the list of supported errors is limited to a small set.
4876
Willy Tarreauae94d4d2011-05-11 16:28:49 +02004877 Code 200 is emitted in response to requests matching a "monitor-uri" rule.
4878
Willy Tarreau2769aa02007-12-27 18:26:09 +01004879 Note that both keyword return the HTTP 302 status code, which tells the
4880 client to fetch the designated URL using the same HTTP method. This can be
4881 quite problematic in case of non-GET methods such as POST, because the URL
4882 sent to the client might not be allowed for something other than GET. To
Willy Tarreau989222a2016-01-15 10:26:26 +01004883 work around this problem, please use "errorloc303" which send the HTTP 303
Willy Tarreau2769aa02007-12-27 18:26:09 +01004884 status code, indicating to the client that the URL must be fetched with a GET
4885 request.
4886
Christopher Faulet3b967c12020-05-15 15:47:44 +02004887 See also : "http-error", "errorfile", "errorloc303"
Willy Tarreau2769aa02007-12-27 18:26:09 +01004888
4889
4890errorloc303 <code> <url>
4891 Return an HTTP redirection to a URL instead of errors generated by HAProxy
4892 May be used in sections : defaults | frontend | listen | backend
4893 yes | yes | yes | yes
4894 Arguments :
4895 <code> is the HTTP status code. Currently, HAProxy is capable of
Christopher Faulet612f2ea2020-05-27 09:57:28 +02004896 generating codes 200, 400, 401, 403, 404, 405, 407, 408, 410,
Christopher Faulete095f312020-12-07 11:22:24 +01004897 413, 425, 429, 500, 501, 502, 503, and 504.
Willy Tarreau2769aa02007-12-27 18:26:09 +01004898
4899 <url> it is the exact contents of the "Location" header. It may contain
4900 either a relative URI to an error page hosted on the same site,
4901 or an absolute URI designating an error page on another site.
4902 Special care should be given to relative URIs to avoid redirect
Davor Ocelice9ed2812017-12-25 17:49:28 +01004903 loops if the URI itself may generate the same error (e.g. 500).
Willy Tarreau2769aa02007-12-27 18:26:09 +01004904
4905 It is important to understand that this keyword is not meant to rewrite
4906 errors returned by the server, but errors detected and returned by HAProxy.
4907 This is why the list of supported errors is limited to a small set.
4908
Willy Tarreauae94d4d2011-05-11 16:28:49 +02004909 Code 200 is emitted in response to requests matching a "monitor-uri" rule.
4910
Willy Tarreau2769aa02007-12-27 18:26:09 +01004911 Note that both keyword return the HTTP 303 status code, which tells the
4912 client to fetch the designated URL using the same HTTP GET method. This
4913 solves the usual problems associated with "errorloc" and the 302 code. It is
4914 possible that some very old browsers designed before HTTP/1.1 do not support
Willy Tarreaud2a4aa22008-01-31 15:28:22 +01004915 it, but no such problem has been reported till now.
Willy Tarreau2769aa02007-12-27 18:26:09 +01004916
Christopher Faulet3b967c12020-05-15 15:47:44 +02004917 See also : "http-error", "errorfile", "errorloc", "errorloc302"
Willy Tarreau2769aa02007-12-27 18:26:09 +01004918
4919
Simon Horman51a1cf62015-02-03 13:00:44 +09004920email-alert from <emailaddr>
4921 Declare the from email address to be used in both the envelope and header
Davor Ocelice9ed2812017-12-25 17:49:28 +01004922 of email alerts. This is the address that email alerts are sent from.
Simon Horman51a1cf62015-02-03 13:00:44 +09004923 May be used in sections: defaults | frontend | listen | backend
4924 yes | yes | yes | yes
4925
4926 Arguments :
4927
4928 <emailaddr> is the from email address to use when sending email alerts
4929
4930 Also requires "email-alert mailers" and "email-alert to" to be set
4931 and if so sending email alerts is enabled for the proxy.
4932
Simon Horman64e34162015-02-06 11:11:57 +09004933 See also : "email-alert level", "email-alert mailers",
Cyril Bonté307ee1e2015-09-28 23:16:06 +02004934 "email-alert myhostname", "email-alert to", section 3.6 about
4935 mailers.
Simon Horman64e34162015-02-06 11:11:57 +09004936
4937
4938email-alert level <level>
4939 Declare the maximum log level of messages for which email alerts will be
4940 sent. This acts as a filter on the sending of email alerts.
4941 May be used in sections: defaults | frontend | listen | backend
4942 yes | yes | yes | yes
4943
4944 Arguments :
4945
4946 <level> One of the 8 syslog levels:
4947 emerg alert crit err warning notice info debug
4948 The above syslog levels are ordered from lowest to highest.
4949
4950 By default level is alert
4951
4952 Also requires "email-alert from", "email-alert mailers" and
4953 "email-alert to" to be set and if so sending email alerts is enabled
4954 for the proxy.
4955
Simon Horman1421e212015-04-30 13:10:35 +09004956 Alerts are sent when :
4957
4958 * An un-paused server is marked as down and <level> is alert or lower
4959 * A paused server is marked as down and <level> is notice or lower
4960 * A server is marked as up or enters the drain state and <level>
4961 is notice or lower
4962 * "option log-health-checks" is enabled, <level> is info or lower,
4963 and a health check status update occurs
4964
Simon Horman64e34162015-02-06 11:11:57 +09004965 See also : "email-alert from", "email-alert mailers",
4966 "email-alert myhostname", "email-alert to",
Simon Horman51a1cf62015-02-03 13:00:44 +09004967 section 3.6 about mailers.
4968
4969
4970email-alert mailers <mailersect>
4971 Declare the mailers to be used when sending email alerts
4972 May be used in sections: defaults | frontend | listen | backend
4973 yes | yes | yes | yes
4974
4975 Arguments :
4976
4977 <mailersect> is the name of the mailers section to send email alerts.
4978
4979 Also requires "email-alert from" and "email-alert to" to be set
4980 and if so sending email alerts is enabled for the proxy.
4981
Simon Horman64e34162015-02-06 11:11:57 +09004982 See also : "email-alert from", "email-alert level", "email-alert myhostname",
4983 "email-alert to", section 3.6 about mailers.
Simon Horman51a1cf62015-02-03 13:00:44 +09004984
4985
4986email-alert myhostname <hostname>
4987 Declare the to hostname address to be used when communicating with
4988 mailers.
4989 May be used in sections: defaults | frontend | listen | backend
4990 yes | yes | yes | yes
4991
4992 Arguments :
4993
Baptiste Assmann738bad92015-12-21 15:27:53 +01004994 <hostname> is the hostname to use when communicating with mailers
Simon Horman51a1cf62015-02-03 13:00:44 +09004995
4996 By default the systems hostname is used.
4997
4998 Also requires "email-alert from", "email-alert mailers" and
4999 "email-alert to" to be set and if so sending email alerts is enabled
5000 for the proxy.
5001
Simon Horman64e34162015-02-06 11:11:57 +09005002 See also : "email-alert from", "email-alert level", "email-alert mailers",
5003 "email-alert to", section 3.6 about mailers.
Simon Horman51a1cf62015-02-03 13:00:44 +09005004
5005
5006email-alert to <emailaddr>
Davor Ocelice9ed2812017-12-25 17:49:28 +01005007 Declare both the recipient address in the envelope and to address in the
Simon Horman51a1cf62015-02-03 13:00:44 +09005008 header of email alerts. This is the address that email alerts are sent to.
5009 May be used in sections: defaults | frontend | listen | backend
5010 yes | yes | yes | yes
5011
5012 Arguments :
5013
5014 <emailaddr> is the to email address to use when sending email alerts
5015
5016 Also requires "email-alert mailers" and "email-alert to" to be set
5017 and if so sending email alerts is enabled for the proxy.
5018
Simon Horman64e34162015-02-06 11:11:57 +09005019 See also : "email-alert from", "email-alert level", "email-alert mailers",
Simon Horman51a1cf62015-02-03 13:00:44 +09005020 "email-alert myhostname", section 3.6 about mailers.
5021
5022
Willy Tarreau4de91492010-01-22 19:10:05 +01005023force-persist { if | unless } <condition>
5024 Declare a condition to force persistence on down servers
5025 May be used in sections: defaults | frontend | listen | backend
Cyril Bonté4288c5a2018-03-12 22:02:59 +01005026 no | no | yes | yes
Willy Tarreau4de91492010-01-22 19:10:05 +01005027
5028 By default, requests are not dispatched to down servers. It is possible to
5029 force this using "option persist", but it is unconditional and redispatches
5030 to a valid server if "option redispatch" is set. That leaves with very little
5031 possibilities to force some requests to reach a server which is artificially
5032 marked down for maintenance operations.
5033
5034 The "force-persist" statement allows one to declare various ACL-based
5035 conditions which, when met, will cause a request to ignore the down status of
5036 a server and still try to connect to it. That makes it possible to start a
5037 server, still replying an error to the health checks, and run a specially
5038 configured browser to test the service. Among the handy methods, one could
5039 use a specific source IP address, or a specific cookie. The cookie also has
5040 the advantage that it can easily be added/removed on the browser from a test
5041 page. Once the service is validated, it is then possible to open the service
5042 to the world by returning a valid response to health checks.
5043
5044 The forced persistence is enabled when an "if" condition is met, or unless an
5045 "unless" condition is met. The final redispatch is always disabled when this
5046 is used.
5047
Cyril Bonté0d4bf012010-04-25 23:21:46 +02005048 See also : "option redispatch", "ignore-persist", "persist",
Cyril Bontéa8e7bbc2010-04-25 22:29:29 +02005049 and section 7 about ACL usage.
Willy Tarreau4de91492010-01-22 19:10:05 +01005050
Christopher Fauletc3fe5332016-04-07 15:30:10 +02005051
5052filter <name> [param*]
5053 Add the filter <name> in the filter list attached to the proxy.
5054 May be used in sections : defaults | frontend | listen | backend
5055 no | yes | yes | yes
5056 Arguments :
5057 <name> is the name of the filter. Officially supported filters are
5058 referenced in section 9.
5059
Tim Düsterhus4896c442016-11-29 02:15:19 +01005060 <param*> is a list of parameters accepted by the filter <name>. The
Christopher Fauletc3fe5332016-04-07 15:30:10 +02005061 parsing of these parameters are the responsibility of the
Tim Düsterhus4896c442016-11-29 02:15:19 +01005062 filter. Please refer to the documentation of the corresponding
5063 filter (section 9) for all details on the supported parameters.
Christopher Fauletc3fe5332016-04-07 15:30:10 +02005064
5065 Multiple occurrences of the filter line can be used for the same proxy. The
5066 same filter can be referenced many times if needed.
5067
5068 Example:
5069 listen
5070 bind *:80
5071
5072 filter trace name BEFORE-HTTP-COMP
5073 filter compression
5074 filter trace name AFTER-HTTP-COMP
5075
5076 compression algo gzip
5077 compression offload
5078
5079 server srv1 192.168.0.1:80
5080
5081 See also : section 9.
5082
Willy Tarreau4de91492010-01-22 19:10:05 +01005083
Willy Tarreau2769aa02007-12-27 18:26:09 +01005084fullconn <conns>
5085 Specify at what backend load the servers will reach their maxconn
5086 May be used in sections : defaults | frontend | listen | backend
5087 yes | no | yes | yes
5088 Arguments :
5089 <conns> is the number of connections on the backend which will make the
5090 servers use the maximal number of connections.
5091
Willy Tarreau198a7442008-01-17 12:05:32 +01005092 When a server has a "maxconn" parameter specified, it means that its number
Willy Tarreau2769aa02007-12-27 18:26:09 +01005093 of concurrent connections will never go higher. Additionally, if it has a
Willy Tarreau198a7442008-01-17 12:05:32 +01005094 "minconn" parameter, it indicates a dynamic limit following the backend's
Willy Tarreau2769aa02007-12-27 18:26:09 +01005095 load. The server will then always accept at least <minconn> connections,
5096 never more than <maxconn>, and the limit will be on the ramp between both
5097 values when the backend has less than <conns> concurrent connections. This
5098 makes it possible to limit the load on the servers during normal loads, but
5099 push it further for important loads without overloading the servers during
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01005100 exceptional loads.
Willy Tarreau2769aa02007-12-27 18:26:09 +01005101
Daniel Corbett9f0843f2021-05-08 10:50:37 -04005102 Since it's hard to get this value right, HAProxy automatically sets it to
Willy Tarreaufbb78422011-06-05 15:38:35 +02005103 10% of the sum of the maxconns of all frontends that may branch to this
Bertrand Jacquin702d44f2013-11-19 11:43:06 +01005104 backend (based on "use_backend" and "default_backend" rules). That way it's
5105 safe to leave it unset. However, "use_backend" involving dynamic names are
5106 not counted since there is no way to know if they could match or not.
Willy Tarreaufbb78422011-06-05 15:38:35 +02005107
Willy Tarreau2769aa02007-12-27 18:26:09 +01005108 Example :
5109 # The servers will accept between 100 and 1000 concurrent connections each
5110 # and the maximum of 1000 will be reached when the backend reaches 10000
5111 # connections.
5112 backend dynamic
5113 fullconn 10000
5114 server srv1 dyn1:80 minconn 100 maxconn 1000
5115 server srv2 dyn2:80 minconn 100 maxconn 1000
5116
5117 See also : "maxconn", "server"
5118
5119
Willy Tarreauab0a5192020-10-09 19:07:01 +02005120grace <time> (deprecated)
Willy Tarreau2769aa02007-12-27 18:26:09 +01005121 Maintain a proxy operational for some time after a soft stop
5122 May be used in sections : defaults | frontend | listen | backend
Cyril Bonté99ed3272010-01-24 23:29:44 +01005123 yes | yes | yes | yes
Willy Tarreau2769aa02007-12-27 18:26:09 +01005124 Arguments :
5125 <time> is the time (by default in milliseconds) for which the instance
5126 will remain operational with the frontend sockets still listening
5127 when a soft-stop is received via the SIGUSR1 signal.
5128
5129 This may be used to ensure that the services disappear in a certain order.
5130 This was designed so that frontends which are dedicated to monitoring by an
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01005131 external equipment fail immediately while other ones remain up for the time
Willy Tarreau2769aa02007-12-27 18:26:09 +01005132 needed by the equipment to detect the failure.
5133
5134 Note that currently, there is very little benefit in using this parameter,
5135 and it may in fact complicate the soft-reconfiguration process more than
5136 simplify it.
5137
Willy Tarreau0ba27502007-12-24 16:55:16 +01005138
Andrew Rodland17be45e2016-10-25 17:04:12 -04005139hash-balance-factor <factor>
5140 Specify the balancing factor for bounded-load consistent hashing
5141 May be used in sections : defaults | frontend | listen | backend
5142 yes | no | no | yes
5143 Arguments :
5144 <factor> is the control for the maximum number of concurrent requests to
5145 send to a server, expressed as a percentage of the average number
Frédéric Lécaille93d33162019-03-06 09:35:59 +01005146 of concurrent requests across all of the active servers.
Andrew Rodland17be45e2016-10-25 17:04:12 -04005147
5148 Specifying a "hash-balance-factor" for a server with "hash-type consistent"
5149 enables an algorithm that prevents any one server from getting too many
5150 requests at once, even if some hash buckets receive many more requests than
5151 others. Setting <factor> to 0 (the default) disables the feature. Otherwise,
5152 <factor> is a percentage greater than 100. For example, if <factor> is 150,
5153 then no server will be allowed to have a load more than 1.5 times the average.
5154 If server weights are used, they will be respected.
5155
5156 If the first-choice server is disqualified, the algorithm will choose another
5157 server based on the request hash, until a server with additional capacity is
5158 found. A higher <factor> allows more imbalance between the servers, while a
5159 lower <factor> means that more servers will be checked on average, affecting
5160 performance. Reasonable values are from 125 to 200.
5161
Willy Tarreau760e81d2018-05-03 07:20:40 +02005162 This setting is also used by "balance random" which internally relies on the
5163 consistent hashing mechanism.
5164
Andrew Rodland17be45e2016-10-25 17:04:12 -04005165 See also : "balance" and "hash-type".
5166
5167
Bhaskar Maddalab6c0ac92013-11-05 11:54:02 -05005168hash-type <method> <function> <modifier>
Willy Tarreau6b2e11b2009-10-01 07:52:15 +02005169 Specify a method to use for mapping hashes to servers
5170 May be used in sections : defaults | frontend | listen | backend
5171 yes | no | yes | yes
5172 Arguments :
Bhaskar98634f02013-10-29 23:30:51 -04005173 <method> is the method used to select a server from the hash computed by
5174 the <function> :
Willy Tarreau6b2e11b2009-10-01 07:52:15 +02005175
Bhaskar98634f02013-10-29 23:30:51 -04005176 map-based the hash table is a static array containing all alive servers.
5177 The hashes will be very smooth, will consider weights, but
5178 will be static in that weight changes while a server is up
5179 will be ignored. This means that there will be no slow start.
5180 Also, since a server is selected by its position in the array,
5181 most mappings are changed when the server count changes. This
5182 means that when a server goes up or down, or when a server is
5183 added to a farm, most connections will be redistributed to
5184 different servers. This can be inconvenient with caches for
5185 instance.
Willy Tarreau798a39c2010-11-24 15:04:29 +01005186
Bhaskar98634f02013-10-29 23:30:51 -04005187 consistent the hash table is a tree filled with many occurrences of each
5188 server. The hash key is looked up in the tree and the closest
5189 server is chosen. This hash is dynamic, it supports changing
5190 weights while the servers are up, so it is compatible with the
5191 slow start feature. It has the advantage that when a server
5192 goes up or down, only its associations are moved. When a
5193 server is added to the farm, only a few part of the mappings
5194 are redistributed, making it an ideal method for caches.
5195 However, due to its principle, the distribution will never be
5196 very smooth and it may sometimes be necessary to adjust a
5197 server's weight or its ID to get a more balanced distribution.
5198 In order to get the same distribution on multiple load
5199 balancers, it is important that all servers have the exact
Bhaskar Maddalab6c0ac92013-11-05 11:54:02 -05005200 same IDs. Note: consistent hash uses sdbm and avalanche if no
5201 hash function is specified.
Bhaskar98634f02013-10-29 23:30:51 -04005202
5203 <function> is the hash function to be used :
5204
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03005205 sdbm this function was created initially for sdbm (a public-domain
Bhaskar98634f02013-10-29 23:30:51 -04005206 reimplementation of ndbm) database library. It was found to do
5207 well in scrambling bits, causing better distribution of the keys
5208 and fewer splits. It also happens to be a good general hashing
Bhaskar Maddalab6c0ac92013-11-05 11:54:02 -05005209 function with good distribution, unless the total server weight
5210 is a multiple of 64, in which case applying the avalanche
5211 modifier may help.
Bhaskar98634f02013-10-29 23:30:51 -04005212
5213 djb2 this function was first proposed by Dan Bernstein many years ago
5214 on comp.lang.c. Studies have shown that for certain workload this
Bhaskar Maddalab6c0ac92013-11-05 11:54:02 -05005215 function provides a better distribution than sdbm. It generally
5216 works well with text-based inputs though it can perform extremely
5217 poorly with numeric-only input or when the total server weight is
5218 a multiple of 33, unless the avalanche modifier is also used.
5219
Daniel Corbett9f0843f2021-05-08 10:50:37 -04005220 wt6 this function was designed for HAProxy while testing other
Willy Tarreaua0f42712013-11-14 14:30:35 +01005221 functions in the past. It is not as smooth as the other ones, but
5222 is much less sensible to the input data set or to the number of
5223 servers. It can make sense as an alternative to sdbm+avalanche or
5224 djb2+avalanche for consistent hashing or when hashing on numeric
5225 data such as a source IP address or a visitor identifier in a URL
5226 parameter.
5227
Willy Tarreau324f07f2015-01-20 19:44:50 +01005228 crc32 this is the most common CRC32 implementation as used in Ethernet,
5229 gzip, PNG, etc. It is slower than the other ones but may provide
5230 a better distribution or less predictable results especially when
5231 used on strings.
5232
Bhaskar Maddalab6c0ac92013-11-05 11:54:02 -05005233 <modifier> indicates an optional method applied after hashing the key :
5234
5235 avalanche This directive indicates that the result from the hash
5236 function above should not be used in its raw form but that
5237 a 4-byte full avalanche hash must be applied first. The
5238 purpose of this step is to mix the resulting bits from the
5239 previous hash in order to avoid any undesired effect when
5240 the input contains some limited values or when the number of
5241 servers is a multiple of one of the hash's components (64
5242 for SDBM, 33 for DJB2). Enabling avalanche tends to make the
5243 result less predictable, but it's also not as smooth as when
5244 using the original function. Some testing might be needed
5245 with some workloads. This hash is one of the many proposed
5246 by Bob Jenkins.
Willy Tarreau6b2e11b2009-10-01 07:52:15 +02005247
Bhaskar98634f02013-10-29 23:30:51 -04005248 The default hash type is "map-based" and is recommended for most usages. The
5249 default function is "sdbm", the selection of a function should be based on
5250 the range of the values being hashed.
Willy Tarreau6b2e11b2009-10-01 07:52:15 +02005251
Andrew Rodland17be45e2016-10-25 17:04:12 -04005252 See also : "balance", "hash-balance-factor", "server"
Willy Tarreau6b2e11b2009-10-01 07:52:15 +02005253
5254
Christopher Faulet6d0c3df2020-01-22 09:26:35 +01005255http-after-response <action> <options...> [ { if | unless } <condition> ]
5256 Access control for all Layer 7 responses (server, applet/service and internal
5257 ones).
5258
5259 May be used in sections: defaults | frontend | listen | backend
5260 no | yes | yes | yes
5261
5262 The http-after-response statement defines a set of rules which apply to layer
5263 7 processing. The rules are evaluated in their declaration order when they
5264 are met in a frontend, listen or backend section. Any rule may optionally be
5265 followed by an ACL-based condition, in which case it will only be evaluated
5266 if the condition is true. Since these rules apply on responses, the backend
5267 rules are applied first, followed by the frontend's rules.
5268
5269 Unlike http-response rules, these ones are applied on all responses, the
5270 server ones but also to all responses generated by HAProxy. These rules are
5271 evaluated at the end of the responses analysis, before the data forwarding.
5272
5273 The first keyword is the rule's action. The supported actions are described
5274 below.
5275
5276 There is no limit to the number of http-after-response statements per
5277 instance.
5278
Christopher Fauletd5ac6de2020-12-02 08:40:14 +01005279 Note: Errors emitted in early stage of the request parsing are handled by the
5280 multiplexer at a lower level, before any http analysis. Thus no
5281 http-after-response ruleset is evaluated on these errors.
5282
Christopher Faulet6d0c3df2020-01-22 09:26:35 +01005283 Example:
5284 http-after-response set-header Strict-Transport-Security "max-age=31536000"
5285 http-after-response set-header Cache-Control "no-store,no-cache,private"
5286 http-after-response set-header Pragma "no-cache"
5287
5288http-after-response add-header <name> <fmt> [ { if | unless } <condition> ]
5289
5290 This appends an HTTP header field whose name is specified in <name> and whose
5291 value is defined by <fmt> which follows the log-format rules (see Custom Log
5292 Format in section 8.2.4). This may be used to send a cookie to a client for
5293 example, or to pass some internal information.
5294 This rule is not final, so it is possible to add other similar rules.
5295 Note that header addition is performed immediately, so one rule might reuse
5296 the resulting header from a previous rule.
5297
5298http-after-response allow [ { if | unless } <condition> ]
5299
5300 This stops the evaluation of the rules and lets the response pass the check.
5301 No further "http-after-response" rules are evaluated.
5302
Maciej Zdebebdd4c52020-11-20 13:58:48 +00005303http-after-response del-header <name> [ -m <meth> ] [ { if | unless } <condition> ]
Christopher Faulet6d0c3df2020-01-22 09:26:35 +01005304
Maciej Zdebebdd4c52020-11-20 13:58:48 +00005305 This removes all HTTP header fields whose name is specified in <name>. <meth>
5306 is the matching method, applied on the header name. Supported matching methods
5307 are "str" (exact match), "beg" (prefix match), "end" (suffix match), "sub"
5308 (substring match) and "reg" (regex match). If not specified, exact matching
5309 method is used.
Christopher Faulet6d0c3df2020-01-22 09:26:35 +01005310
5311http-after-response replace-header <name> <regex-match> <replace-fmt>
5312 [ { if | unless } <condition> ]
5313
5314 This works like "http-response replace-header".
5315
5316 Example:
5317 http-after-response replace-header Set-Cookie (C=[^;]*);(.*) \1;ip=%bi;\2
5318
5319 # applied to:
5320 Set-Cookie: C=1; expires=Tue, 14-Jun-2016 01:40:45 GMT
5321
5322 # outputs:
5323 Set-Cookie: C=1;ip=192.168.1.20; expires=Tue, 14-Jun-2016 01:40:45 GMT
5324
5325 # assuming the backend IP is 192.168.1.20.
5326
5327http-after-response replace-value <name> <regex-match> <replace-fmt>
5328 [ { if | unless } <condition> ]
5329
5330 This works like "http-response replace-value".
5331
5332 Example:
5333 http-after-response replace-value Cache-control ^public$ private
5334
5335 # applied to:
5336 Cache-Control: max-age=3600, public
5337
5338 # outputs:
5339 Cache-Control: max-age=3600, private
5340
5341http-after-response set-header <name> <fmt> [ { if | unless } <condition> ]
5342
5343 This does the same as "add-header" except that the header name is first
5344 removed if it existed. This is useful when passing security information to
5345 the server, where the header must not be manipulated by external users.
5346
5347http-after-response set-status <status> [reason <str>]
5348 [ { if | unless } <condition> ]
5349
5350 This replaces the response status code with <status> which must be an integer
5351 between 100 and 999. Optionally, a custom reason text can be provided defined
5352 by <str>, or the default reason for the specified code will be used as a
5353 fallback.
5354
5355 Example:
5356 # return "431 Request Header Fields Too Large"
5357 http-response set-status 431
5358 # return "503 Slow Down", custom reason
5359 http-response set-status 503 reason "Slow Down"
5360
5361http-after-response set-var(<var-name>) <expr> [ { if | unless } <condition> ]
5362
5363 This is used to set the contents of a variable. The variable is declared
5364 inline.
5365
5366 Arguments:
5367 <var-name> The name of the variable starts with an indication about its
5368 scope. The scopes allowed are:
5369 "proc" : the variable is shared with the whole process
5370 "sess" : the variable is shared with the whole session
5371 "txn" : the variable is shared with the transaction
5372 (request and response)
5373 "req" : the variable is shared only during request
5374 processing
5375 "res" : the variable is shared only during response
5376 processing
5377 This prefix is followed by a name. The separator is a '.'.
5378 The name may only contain characters 'a-z', 'A-Z', '0-9', '.'
5379 and '_'.
5380
5381 <expr> Is a standard HAProxy expression formed by a sample-fetch
5382 followed by some converters.
5383
5384 Example:
5385 http-after-response set-var(sess.last_redir) res.hdr(location)
5386
5387http-after-response strict-mode { on | off }
5388
5389 This enables or disables the strict rewriting mode for following rules. It
5390 does not affect rules declared before it and it is only applicable on rules
5391 performing a rewrite on the responses. When the strict mode is enabled, any
5392 rewrite failure triggers an internal error. Otherwise, such errors are
5393 silently ignored. The purpose of the strict rewriting mode is to make some
Ilya Shipitsin8525fd92020-02-29 12:34:59 +05005394 rewrites optional while others must be performed to continue the response
Christopher Faulet6d0c3df2020-01-22 09:26:35 +01005395 processing.
5396
5397 By default, the strict rewriting mode is enabled. Its value is also reset
5398 when a ruleset evaluation ends. So, for instance, if you change the mode on
Daniel Corbett67a82712020-07-06 23:01:19 -04005399 the backend, the default mode is restored when HAProxy starts the frontend
Christopher Faulet6d0c3df2020-01-22 09:26:35 +01005400 rules evaluation.
5401
5402http-after-response unset-var(<var-name>) [ { if | unless } <condition> ]
5403
5404 This is used to unset a variable. See "http-after-response set-var" for
5405 details about <var-name>.
5406
5407 Example:
5408 http-after-response unset-var(sess.last_redir)
5409
Christopher Faulet4f5c2e22020-04-23 15:22:33 +02005410
5411http-check comment <string>
5412 Defines a comment for the following the http-check rule, reported in logs if
5413 it fails.
5414 May be used in sections : defaults | frontend | listen | backend
5415 yes | no | yes | yes
5416
Christopher Faulet4f5c2e22020-04-23 15:22:33 +02005417 Arguments :
5418 <string> is the comment message to add in logs if the following http-check
5419 rule fails.
5420
Christopher Fauletc52ea4d2020-04-23 15:43:35 +02005421 It only works for connect, send and expect rules. It is useful to make
5422 user-friendly error reporting.
5423
Daniel Corbett67a82712020-07-06 23:01:19 -04005424 See also : "option httpchk", "http-check connect", "http-check send" and
Christopher Faulet4f5c2e22020-04-23 15:22:33 +02005425 "http-check expect".
5426
5427
Christopher Fauletc52ea4d2020-04-23 15:43:35 +02005428http-check connect [default] [port <expr>] [addr <ip>] [send-proxy]
5429 [via-socks4] [ssl] [sni <sni>] [alpn <alpn>] [linger]
Christopher Fauletedc6ed92020-04-23 16:27:59 +02005430 [proto <name>] [comment <msg>]
Christopher Faulete5870d82020-04-15 11:32:03 +02005431 Opens a new connection to perform an HTTP health check
5432 May be used in sections : defaults | frontend | listen | backend
5433 yes | no | yes | yes
5434
Christopher Fauletc52ea4d2020-04-23 15:43:35 +02005435 Arguments :
Christopher Faulet4f5c2e22020-04-23 15:22:33 +02005436 comment <msg> defines a message to report if the rule evaluation fails.
5437
Christopher Faulete5870d82020-04-15 11:32:03 +02005438 default Use default options of the server line to do the health
Daniel Corbett67a82712020-07-06 23:01:19 -04005439 checks. The server options are used only if not redefined.
Christopher Faulete5870d82020-04-15 11:32:03 +02005440
5441 port <expr> if not set, check port or server port is used.
5442 It tells HAProxy where to open the connection to.
5443 <port> must be a valid TCP port source integer, from 1 to
5444 65535 or an sample-fetch expression.
5445
5446 addr <ip> defines the IP address to do the health check.
5447
5448 send-proxy send a PROXY protocol string
5449
5450 via-socks4 enables outgoing health checks using upstream socks4 proxy.
5451
5452 ssl opens a ciphered connection
5453
5454 sni <sni> specifies the SNI to use to do health checks over SSL.
5455
5456 alpn <alpn> defines which protocols to advertise with ALPN. The protocol
5457 list consists in a comma-delimited list of protocol names,
5458 for instance: "h2,http/1.1". If it is not set, the server ALPN
5459 is used.
5460
Christopher Fauletedc6ed92020-04-23 16:27:59 +02005461 proto <name> forces the multiplexer's protocol to use for this connection.
5462 It must be an HTTP mux protocol and it must be usable on the
5463 backend side. The list of available protocols is reported in
5464 haproxy -vv.
5465
Christopher Faulete5870d82020-04-15 11:32:03 +02005466 linger cleanly close the connection instead of using a single RST.
5467
Christopher Fauletc52ea4d2020-04-23 15:43:35 +02005468 Just like tcp-check health checks, it is possible to configure the connection
5469 to use to perform HTTP health check. This directive should also be used to
5470 describe a scenario involving several request/response exchanges, possibly on
5471 different ports or with different servers.
5472
5473 When there are no TCP port configured on the server line neither server port
5474 directive, then the first step of the http-check sequence must be to specify
5475 the port with a "http-check connect".
5476
5477 In an http-check ruleset a 'connect' is required, it is also mandatory to start
5478 the ruleset with a 'connect' rule. Purpose is to ensure admin know what they
5479 do.
5480
5481 When a connect must start the ruleset, if may still be preceded by set-var,
5482 unset-var or comment rules.
5483
5484 Examples :
Christopher Faulete5870d82020-04-15 11:32:03 +02005485 # check HTTP and HTTPs services on a server.
5486 # first open port 80 thanks to server line port directive, then
5487 # tcp-check opens port 443, ciphered and run a request on it:
5488 option httpchk
5489
5490 http-check connect
Christopher Fauleta5c14ef2020-04-29 14:19:13 +02005491 http-check send meth GET uri / ver HTTP/1.1 hdr host haproxy.1wt.eu
Christopher Faulet8021a5f2020-04-24 13:53:12 +02005492 http-check expect status 200-399
Christopher Faulete5870d82020-04-15 11:32:03 +02005493 http-check connect port 443 ssl sni haproxy.1wt.eu
Christopher Fauleta5c14ef2020-04-29 14:19:13 +02005494 http-check send meth GET uri / ver HTTP/1.1 hdr host haproxy.1wt.eu
Christopher Faulet8021a5f2020-04-24 13:53:12 +02005495 http-check expect status 200-399
Christopher Faulete5870d82020-04-15 11:32:03 +02005496
5497 server www 10.0.0.1 check port 80
5498
5499 See also : "option httpchk", "http-check send", "http-check expect"
Christopher Faulet6d0c3df2020-01-22 09:26:35 +01005500
Christopher Fauletc52ea4d2020-04-23 15:43:35 +02005501
Willy Tarreau0ba27502007-12-24 16:55:16 +01005502http-check disable-on-404
5503 Enable a maintenance mode upon HTTP/404 response to health-checks
5504 May be used in sections : defaults | frontend | listen | backend
Willy Tarreau2769aa02007-12-27 18:26:09 +01005505 yes | no | yes | yes
Willy Tarreau0ba27502007-12-24 16:55:16 +01005506 Arguments : none
5507
5508 When this option is set, a server which returns an HTTP code 404 will be
5509 excluded from further load-balancing, but will still receive persistent
5510 connections. This provides a very convenient method for Web administrators
5511 to perform a graceful shutdown of their servers. It is also important to note
5512 that a server which is detected as failed while it was in this mode will not
5513 generate an alert, just a notice. If the server responds 2xx or 3xx again, it
5514 will immediately be reinserted into the farm. The status on the stats page
5515 reports "NOLB" for a server in this mode. It is important to note that this
Willy Tarreaubd741542010-03-16 18:46:54 +01005516 option only works in conjunction with the "httpchk" option. If this option
5517 is used with "http-check expect", then it has precedence over it so that 404
Christopher Fauletfa8b89a2020-11-20 18:54:13 +01005518 responses will still be considered as soft-stop. Note also that a stopped
5519 server will stay stopped even if it replies 404s. This option is only
5520 evaluated for running servers.
Willy Tarreaubd741542010-03-16 18:46:54 +01005521
Christopher Fauletc52ea4d2020-04-23 15:43:35 +02005522 See also : "option httpchk" and "http-check expect".
Willy Tarreaubd741542010-03-16 18:46:54 +01005523
5524
Christopher Faulet4f5c2e22020-04-23 15:22:33 +02005525http-check expect [min-recv <int>] [comment <msg>]
Christopher Faulete5870d82020-04-15 11:32:03 +02005526 [ok-status <st>] [error-status <st>] [tout-status <st>]
5527 [on-success <fmt>] [on-error <fmt>] [status-code <expr>]
5528 [!] <match> <pattern>
Jamie Gloudonaaa21002012-08-25 00:18:33 -04005529 Make HTTP health checks consider response contents or specific status codes
Willy Tarreaubd741542010-03-16 18:46:54 +01005530 May be used in sections : defaults | frontend | listen | backend
Willy Tarreau1ee51a62011-08-19 20:04:17 +02005531 yes | no | yes | yes
Christopher Faulete5870d82020-04-15 11:32:03 +02005532
Willy Tarreaubd741542010-03-16 18:46:54 +01005533 Arguments :
Christopher Faulet4f5c2e22020-04-23 15:22:33 +02005534 comment <msg> defines a message to report if the rule evaluation fails.
5535
Christopher Faulete5870d82020-04-15 11:32:03 +02005536 min-recv is optional and can define the minimum amount of data required to
5537 evaluate the current expect rule. If the number of received bytes
5538 is under this limit, the check will wait for more data. This
5539 option can be used to resolve some ambiguous matching rules or to
5540 avoid executing costly regex matches on content known to be still
5541 incomplete. If an exact string is used, the minimum between the
5542 string length and this parameter is used. This parameter is
5543 ignored if it is set to -1. If the expect rule does not match,
5544 the check will wait for more data. If set to 0, the evaluation
5545 result is always conclusive.
5546
5547 ok-status <st> is optional and can be used to set the check status if
5548 the expect rule is successfully evaluated and if it is
5549 the last rule in the tcp-check ruleset. "L7OK", "L7OKC",
Christopher Fauletd888f0f2020-05-07 07:40:17 +02005550 "L6OK" and "L4OK" are supported :
5551 - L7OK : check passed on layer 7
Christopher Faulet83662b52020-11-20 17:47:47 +01005552 - L7OKC : check conditionally passed on layer 7, set
5553 server to NOLB state.
Christopher Fauletd888f0f2020-05-07 07:40:17 +02005554 - L6OK : check passed on layer 6
5555 - L4OK : check passed on layer 4
5556 By default "L7OK" is used.
Christopher Faulete5870d82020-04-15 11:32:03 +02005557
5558 error-status <st> is optional and can be used to set the check status if
5559 an error occurred during the expect rule evaluation.
Christopher Faulet83662b52020-11-20 17:47:47 +01005560 "L7OKC", "L7RSP", "L7STS", "L6RSP" and "L4CON" are
5561 supported :
5562 - L7OKC : check conditionally passed on layer 7, set
5563 server to NOLB state.
Christopher Fauletd888f0f2020-05-07 07:40:17 +02005564 - L7RSP : layer 7 invalid response - protocol error
5565 - L7STS : layer 7 response error, for example HTTP 5xx
5566 - L6RSP : layer 6 invalid response - protocol error
5567 - L4CON : layer 1-4 connection problem
5568 By default "L7RSP" is used.
Christopher Faulete5870d82020-04-15 11:32:03 +02005569
5570 tout-status <st> is optional and can be used to set the check status if
5571 a timeout occurred during the expect rule evaluation.
Christopher Fauletd888f0f2020-05-07 07:40:17 +02005572 "L7TOUT", "L6TOUT", and "L4TOUT" are supported :
5573 - L7TOUT : layer 7 (HTTP/SMTP) timeout
5574 - L6TOUT : layer 6 (SSL) timeout
5575 - L4TOUT : layer 1-4 timeout
Christopher Faulete5870d82020-04-15 11:32:03 +02005576 By default "L7TOUT" is used.
5577
5578 on-success <fmt> is optional and can be used to customize the
5579 informational message reported in logs if the expect
5580 rule is successfully evaluated and if it is the last rule
5581 in the tcp-check ruleset. <fmt> is a log-format string.
5582
5583 on-error <fmt> is optional and can be used to customize the
5584 informational message reported in logs if an error
5585 occurred during the expect rule evaluation. <fmt> is a
5586 log-format string.
5587
Willy Tarreaubd741542010-03-16 18:46:54 +01005588 <match> is a keyword indicating how to look for a specific pattern in the
Christopher Fauletb5594262020-05-05 20:23:13 +02005589 response. The keyword may be one of "status", "rstatus", "hdr",
5590 "fhdr", "string", or "rstring". The keyword may be preceded by an
Willy Tarreaubd741542010-03-16 18:46:54 +01005591 exclamation mark ("!") to negate the match. Spaces are allowed
5592 between the exclamation mark and the keyword. See below for more
5593 details on the supported keywords.
5594
Christopher Faulet39708192020-05-05 10:47:36 +02005595 <pattern> is the pattern to look for. It may be a string, a regular
5596 expression or a more complex pattern with several arguments. If
5597 the string pattern contains spaces, they must be escaped with the
5598 usual backslash ('\').
Willy Tarreaubd741542010-03-16 18:46:54 +01005599
5600 By default, "option httpchk" considers that response statuses 2xx and 3xx
5601 are valid, and that others are invalid. When "http-check expect" is used,
5602 it defines what is considered valid or invalid. Only one "http-check"
5603 statement is supported in a backend. If a server fails to respond or times
5604 out, the check obviously fails. The available matches are :
5605
Christopher Faulet8021a5f2020-04-24 13:53:12 +02005606 status <codes> : test the status codes found parsing <codes> string. it
5607 must be a comma-separated list of status codes or range
5608 codes. A health check response will be considered as
5609 valid if the response's status code matches any status
5610 code or is inside any range of the list. If the "status"
5611 keyword is prefixed with "!", then the response will be
5612 considered invalid if the status code matches.
Willy Tarreaubd741542010-03-16 18:46:54 +01005613
5614 rstatus <regex> : test a regular expression for the HTTP status code.
Jamie Gloudonaaa21002012-08-25 00:18:33 -04005615 A health check response will be considered valid if the
Willy Tarreaubd741542010-03-16 18:46:54 +01005616 response's status code matches the expression. If the
5617 "rstatus" keyword is prefixed with "!", then the response
5618 will be considered invalid if the status code matches.
5619 This is mostly used to check for multiple codes.
5620
Christopher Fauletb5594262020-05-05 20:23:13 +02005621 hdr { name | name-lf } [ -m <meth> ] <name>
5622 [ { value | value-lf } [ -m <meth> ] <value> :
Christopher Faulet39708192020-05-05 10:47:36 +02005623 test the specified header pattern on the HTTP response
5624 headers. The name pattern is mandatory but the value
5625 pattern is optional. If not specified, only the header
5626 presence is verified. <meth> is the matching method,
5627 applied on the header name or the header value. Supported
5628 matching methods are "str" (exact match), "beg" (prefix
5629 match), "end" (suffix match), "sub" (substring match) or
5630 "reg" (regex match). If not specified, exact matching
Christopher Fauletb5594262020-05-05 20:23:13 +02005631 method is used. If the "name-lf" parameter is used,
5632 <name> is evaluated as a log-format string. If "value-lf"
5633 parameter is used, <value> is evaluated as a log-format
5634 string. These parameters cannot be used with the regex
5635 matching method. Finally, the header value is considered
5636 as comma-separated list. Note that matchings are case
5637 insensitive on the header names.
5638
5639 fhdr { name | name-lf } [ -m <meth> ] <name>
5640 [ { value | value-lf } [ -m <meth> ] <value> :
5641 test the specified full header pattern on the HTTP
5642 response headers. It does exactly the same than "hdr"
5643 keyword, except the full header value is tested, commas
5644 are not considered as delimiters.
Christopher Faulet39708192020-05-05 10:47:36 +02005645
Willy Tarreaubd741542010-03-16 18:46:54 +01005646 string <string> : test the exact string match in the HTTP response body.
Jamie Gloudonaaa21002012-08-25 00:18:33 -04005647 A health check response will be considered valid if the
Willy Tarreaubd741542010-03-16 18:46:54 +01005648 response's body contains this exact string. If the
5649 "string" keyword is prefixed with "!", then the response
5650 will be considered invalid if the body contains this
5651 string. This can be used to look for a mandatory word at
5652 the end of a dynamic page, or to detect a failure when a
Davor Ocelice9ed2812017-12-25 17:49:28 +01005653 specific error appears on the check page (e.g. a stack
Willy Tarreaubd741542010-03-16 18:46:54 +01005654 trace).
5655
5656 rstring <regex> : test a regular expression on the HTTP response body.
Jamie Gloudonaaa21002012-08-25 00:18:33 -04005657 A health check response will be considered valid if the
Willy Tarreaubd741542010-03-16 18:46:54 +01005658 response's body matches this expression. If the "rstring"
5659 keyword is prefixed with "!", then the response will be
5660 considered invalid if the body matches the expression.
5661 This can be used to look for a mandatory word at the end
5662 of a dynamic page, or to detect a failure when a specific
Davor Ocelice9ed2812017-12-25 17:49:28 +01005663 error appears on the check page (e.g. a stack trace).
Willy Tarreaubd741542010-03-16 18:46:54 +01005664
Christopher Fauletaaab0832020-05-05 15:54:22 +02005665 string-lf <fmt> : test a log-format string match in the HTTP response body.
5666 A health check response will be considered valid if the
5667 response's body contains the string resulting of the
5668 evaluation of <fmt>, which follows the log-format rules.
5669 If prefixed with "!", then the response will be
5670 considered invalid if the body contains the string.
5671
Willy Tarreaubd741542010-03-16 18:46:54 +01005672 It is important to note that the responses will be limited to a certain size
Christopher Fauletbb9fb8b2020-11-25 17:20:57 +01005673 defined by the global "tune.bufsize" option, which defaults to 16384 bytes.
Willy Tarreaubd741542010-03-16 18:46:54 +01005674 Thus, too large responses may not contain the mandatory pattern when using
5675 "string" or "rstring". If a large response is absolutely required, it is
5676 possible to change the default max size by setting the global variable.
5677 However, it is worth keeping in mind that parsing very large responses can
5678 waste some CPU cycles, especially when regular expressions are used, and that
5679 it is always better to focus the checks on smaller resources.
5680
Christopher Faulete5870d82020-04-15 11:32:03 +02005681 In an http-check ruleset, the last expect rule may be implicit. If no expect
5682 rule is specified after the last "http-check send", an implicit expect rule
5683 is defined to match on 2xx or 3xx status codes. It means this rule is also
5684 defined if there is no "http-check" rule at all, when only "option httpchk"
5685 is set.
Cyril Bonté32602d22015-01-30 00:07:07 +01005686
Willy Tarreaubd741542010-03-16 18:46:54 +01005687 Last, if "http-check expect" is combined with "http-check disable-on-404",
5688 then this last one has precedence when the server responds with 404.
5689
5690 Examples :
5691 # only accept status 200 as valid
Christopher Faulet8021a5f2020-04-24 13:53:12 +02005692 http-check expect status 200,201,300-310
Willy Tarreaubd741542010-03-16 18:46:54 +01005693
Christopher Faulet39708192020-05-05 10:47:36 +02005694 # be sure a sessid coookie is set
5695 http-check expect header name "set-cookie" value -m beg "sessid="
5696
Willy Tarreaubd741542010-03-16 18:46:54 +01005697 # consider SQL errors as errors
Willy Tarreau8f2a1e72011-01-06 16:36:10 +01005698 http-check expect ! string SQL\ Error
Willy Tarreaubd741542010-03-16 18:46:54 +01005699
5700 # consider status 5xx only as errors
Willy Tarreau8f2a1e72011-01-06 16:36:10 +01005701 http-check expect ! rstatus ^5
Willy Tarreaubd741542010-03-16 18:46:54 +01005702
5703 # check that we have a correct hexadecimal tag before /html
Jarno Huuskonene5ae7022017-04-03 14:36:21 +03005704 http-check expect rstring <!--tag:[0-9a-f]*--></html>
Willy Tarreau0ba27502007-12-24 16:55:16 +01005705
Christopher Faulete5870d82020-04-15 11:32:03 +02005706 See also : "option httpchk", "http-check connect", "http-check disable-on-404"
Christopher Fauletc52ea4d2020-04-23 15:43:35 +02005707 and "http-check send".
Willy Tarreau2769aa02007-12-27 18:26:09 +01005708
5709
Christopher Faulet7c95f5f2020-05-06 15:06:34 +02005710http-check send [meth <method>] [{ uri <uri> | uri-lf <fmt> }>] [ver <version>]
Christopher Faulet574e7bd2020-05-06 15:38:58 +02005711 [hdr <name> <fmt>]* [{ body <string> | body-lf <fmt> }]
5712 [comment <msg>]
Christopher Faulet8acb1282020-04-09 08:44:06 +02005713 Add a possible list of headers and/or a body to the request sent during HTTP
5714 health checks.
5715 May be used in sections : defaults | frontend | listen | backend
5716 yes | no | yes | yes
5717 Arguments :
Christopher Faulet4f5c2e22020-04-23 15:22:33 +02005718 comment <msg> defines a message to report if the rule evaluation fails.
5719
Christopher Faulete5870d82020-04-15 11:32:03 +02005720 meth <method> is the optional HTTP method used with the requests. When not
5721 set, the "OPTIONS" method is used, as it generally requires
5722 low server processing and is easy to filter out from the
5723 logs. Any method may be used, though it is not recommended
5724 to invent non-standard ones.
5725
Christopher Faulet7c95f5f2020-05-06 15:06:34 +02005726 uri <uri> is optional and set the URI referenced in the HTTP requests
5727 to the string <uri>. It defaults to "/" which is accessible
5728 by default on almost any server, but may be changed to any
5729 other URI. Query strings are permitted.
5730
5731 uri-lf <fmt> is optional and set the URI referenced in the HTTP requests
5732 using the log-format string <fmt>. It defaults to "/" which
5733 is accessible by default on almost any server, but may be
5734 changed to any other URI. Query strings are permitted.
Christopher Faulet8acb1282020-04-09 08:44:06 +02005735
Christopher Faulet907701b2020-04-28 09:37:00 +02005736 ver <version> is the optional HTTP version string. It defaults to
Christopher Faulete5870d82020-04-15 11:32:03 +02005737 "HTTP/1.0" but some servers might behave incorrectly in HTTP
Daniel Corbett67a82712020-07-06 23:01:19 -04005738 1.0, so turning it to HTTP/1.1 may sometimes help. Note that
Christopher Faulete5870d82020-04-15 11:32:03 +02005739 the Host field is mandatory in HTTP/1.1, use "hdr" argument
5740 to add it.
5741
5742 hdr <name> <fmt> adds the HTTP header field whose name is specified in
5743 <name> and whose value is defined by <fmt>, which follows
5744 to the log-format rules.
5745
5746 body <string> add the body defined by <string> to the request sent during
5747 HTTP health checks. If defined, the "Content-Length" header
5748 is thus automatically added to the request.
Christopher Faulet8acb1282020-04-09 08:44:06 +02005749
Christopher Faulet574e7bd2020-05-06 15:38:58 +02005750 body-lf <fmt> add the body defined by the log-format string <fmt> to the
5751 request sent during HTTP health checks. If defined, the
5752 "Content-Length" header is thus automatically added to the
5753 request.
5754
Christopher Faulet8acb1282020-04-09 08:44:06 +02005755 In addition to the request line defined by the "option httpchk" directive,
5756 this one is the valid way to add some headers and optionally a body to the
5757 request sent during HTTP health checks. If a body is defined, the associate
Christopher Faulet9df910c2020-04-29 14:20:47 +02005758 "Content-Length" header is automatically added. Thus, this header or
5759 "Transfer-encoding" header should not be present in the request provided by
5760 "http-check send". If so, it will be ignored. The old trick consisting to add
5761 headers after the version string on the "option httpchk" line is now
Amaury Denoyelle6d975f02020-12-22 14:08:52 +01005762 deprecated.
Christopher Faulet8acb1282020-04-09 08:44:06 +02005763
Christopher Faulete5870d82020-04-15 11:32:03 +02005764 Also "http-check send" doesn't support HTTP keep-alive. Keep in mind that it
Amaury Denoyelle6d975f02020-12-22 14:08:52 +01005765 will automatically append a "Connection: close" header, unless a Connection
5766 header has already already been configured via a hdr entry.
Christopher Faulet9df910c2020-04-29 14:20:47 +02005767
5768 Note that the Host header and the request authority, when both defined, are
5769 automatically synchronized. It means when the HTTP request is sent, when a
5770 Host is inserted in the request, the request authority is accordingly
5771 updated. Thus, don't be surprised if the Host header value overwrites the
5772 configured request authority.
5773
5774 Note also for now, no Host header is automatically added in HTTP/1.1 or above
5775 requests. You should add it explicitly.
Christopher Faulete5870d82020-04-15 11:32:03 +02005776
Christopher Fauletc52ea4d2020-04-23 15:43:35 +02005777 See also : "option httpchk", "http-check send-state" and "http-check expect".
Christopher Faulet8acb1282020-04-09 08:44:06 +02005778
5779
Willy Tarreauef781042010-01-27 11:53:01 +01005780http-check send-state
5781 Enable emission of a state header with HTTP health checks
5782 May be used in sections : defaults | frontend | listen | backend
5783 yes | no | yes | yes
5784 Arguments : none
5785
Daniel Corbett9f0843f2021-05-08 10:50:37 -04005786 When this option is set, HAProxy will systematically send a special header
Willy Tarreauef781042010-01-27 11:53:01 +01005787 "X-Haproxy-Server-State" with a list of parameters indicating to each server
Daniel Corbett9f0843f2021-05-08 10:50:37 -04005788 how they are seen by HAProxy. This can be used for instance when a server is
5789 manipulated without access to HAProxy and the operator needs to know whether
5790 HAProxy still sees it up or not, or if the server is the last one in a farm.
Willy Tarreauef781042010-01-27 11:53:01 +01005791
5792 The header is composed of fields delimited by semi-colons, the first of which
5793 is a word ("UP", "DOWN", "NOLB"), possibly followed by a number of valid
5794 checks on the total number before transition, just as appears in the stats
5795 interface. Next headers are in the form "<variable>=<value>", indicating in
5796 no specific order some values available in the stats interface :
Joseph Lynch514061c2015-01-15 17:52:59 -08005797 - a variable "address", containing the address of the backend server.
5798 This corresponds to the <address> field in the server declaration. For
5799 unix domain sockets, it will read "unix".
5800
5801 - a variable "port", containing the port of the backend server. This
5802 corresponds to the <port> field in the server declaration. For unix
5803 domain sockets, it will read "unix".
5804
Willy Tarreauef781042010-01-27 11:53:01 +01005805 - a variable "name", containing the name of the backend followed by a slash
5806 ("/") then the name of the server. This can be used when a server is
5807 checked in multiple backends.
5808
Daniel Corbett9f0843f2021-05-08 10:50:37 -04005809 - a variable "node" containing the name of the HAProxy node, as set in the
Willy Tarreauef781042010-01-27 11:53:01 +01005810 global "node" variable, otherwise the system's hostname if unspecified.
5811
5812 - a variable "weight" indicating the weight of the server, a slash ("/")
5813 and the total weight of the farm (just counting usable servers). This
5814 helps to know if other servers are available to handle the load when this
5815 one fails.
5816
5817 - a variable "scur" indicating the current number of concurrent connections
5818 on the server, followed by a slash ("/") then the total number of
5819 connections on all servers of the same backend.
5820
5821 - a variable "qcur" indicating the current number of requests in the
5822 server's queue.
5823
5824 Example of a header received by the application server :
5825 >>> X-Haproxy-Server-State: UP 2/3; name=bck/srv2; node=lb1; weight=1/2; \
5826 scur=13/22; qcur=0
5827
Christopher Fauletc52ea4d2020-04-23 15:43:35 +02005828 See also : "option httpchk", "http-check disable-on-404" and
5829 "http-check send".
Willy Tarreauef781042010-01-27 11:53:01 +01005830
Christopher Faulete5870d82020-04-15 11:32:03 +02005831
5832http-check set-var(<var-name>) <expr>
Christopher Faulete5870d82020-04-15 11:32:03 +02005833 This operation sets the content of a variable. The variable is declared inline.
Christopher Faulete5870d82020-04-15 11:32:03 +02005834 May be used in sections: defaults | frontend | listen | backend
5835 yes | no | yes | yes
5836
Christopher Fauletc52ea4d2020-04-23 15:43:35 +02005837 Arguments :
Christopher Faulete5870d82020-04-15 11:32:03 +02005838 <var-name> The name of the variable starts with an indication about its
5839 scope. The scopes allowed for http-check are:
5840 "proc" : the variable is shared with the whole process.
5841 "sess" : the variable is shared with the tcp-check session.
5842 "check": the variable is declared for the lifetime of the tcp-check.
5843 This prefix is followed by a name. The separator is a '.'.
5844 The name may only contain characters 'a-z', 'A-Z', '0-9', '.',
5845 and '-'.
5846
5847 <expr> Is a sample-fetch expression potentially followed by converters.
5848
Christopher Fauletc52ea4d2020-04-23 15:43:35 +02005849 Examples :
5850 http-check set-var(check.port) int(1234)
Christopher Faulete5870d82020-04-15 11:32:03 +02005851
5852
5853http-check unset-var(<var-name>)
Christopher Faulete5870d82020-04-15 11:32:03 +02005854 Free a reference to a variable within its scope.
Christopher Faulete5870d82020-04-15 11:32:03 +02005855 May be used in sections: defaults | frontend | listen | backend
5856 yes | no | yes | yes
5857
Christopher Fauletc52ea4d2020-04-23 15:43:35 +02005858 Arguments :
Christopher Faulete5870d82020-04-15 11:32:03 +02005859 <var-name> The name of the variable starts with an indication about its
5860 scope. The scopes allowed for http-check are:
5861 "proc" : the variable is shared with the whole process.
5862 "sess" : the variable is shared with the tcp-check session.
5863 "check": the variable is declared for the lifetime of the tcp-check.
5864 This prefix is followed by a name. The separator is a '.'.
5865 The name may only contain characters 'a-z', 'A-Z', '0-9', '.',
5866 and '-'.
5867
Christopher Fauletc52ea4d2020-04-23 15:43:35 +02005868 Examples :
5869 http-check unset-var(check.port)
Christopher Faulete5870d82020-04-15 11:32:03 +02005870
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02005871
Christopher Faulet3b967c12020-05-15 15:47:44 +02005872http-error status <code> [content-type <type>]
5873 [ { default-errorfiles | errorfile <file> | errorfiles <name> |
5874 file <file> | lf-file <file> | string <str> | lf-string <fmt> } ]
5875 [ hdr <name> <fmt> ]*
5876 Defines a custom error message to use instead of errors generated by HAProxy.
5877 May be used in sections : defaults | frontend | listen | backend
5878 yes | yes | yes | yes
5879 Arguments :
Ilya Shipitsin11057a32020-06-21 21:18:27 +05005880 status <code> is the HTTP status code. It must be specified.
Christopher Faulet3b967c12020-05-15 15:47:44 +02005881 Currently, HAProxy is capable of generating codes
Anthonin Bonnefoy85048f82020-06-22 09:17:01 +02005882 200, 400, 401, 403, 404, 405, 407, 408, 410, 413, 425,
Christopher Faulete095f312020-12-07 11:22:24 +01005883 429, 500, 501, 502, 503, and 504.
Christopher Faulet3b967c12020-05-15 15:47:44 +02005884
5885 content-type <type> is the response content type, for instance
5886 "text/plain". This parameter is ignored and should be
5887 omitted when an errorfile is configured or when the
5888 payload is empty. Otherwise, it must be defined.
5889
5890 default-errorfiles Reset the previously defined error message for current
5891 proxy for the status <code>. If used on a backend, the
5892 frontend error message is used, if defined. If used on
5893 a frontend, the default error message is used.
5894
5895 errorfile <file> designates a file containing the full HTTP response.
5896 It is recommended to follow the common practice of
5897 appending ".http" to the filename so that people do
5898 not confuse the response with HTML error pages, and to
5899 use absolute paths, since files are read before any
5900 chroot is performed.
5901
5902 errorfiles <name> designates the http-errors section to use to import
5903 the error message with the status code <code>. If no
5904 such message is found, the proxy's error messages are
5905 considered.
5906
5907 file <file> specifies the file to use as response payload. If the
5908 file is not empty, its content-type must be set as
5909 argument to "content-type", otherwise, any
5910 "content-type" argument is ignored. <file> is
5911 considered as a raw string.
5912
5913 string <str> specifies the raw string to use as response payload.
5914 The content-type must always be set as argument to
5915 "content-type".
5916
5917 lf-file <file> specifies the file to use as response payload. If the
5918 file is not empty, its content-type must be set as
5919 argument to "content-type", otherwise, any
5920 "content-type" argument is ignored. <file> is
5921 evaluated as a log-format string.
5922
5923 lf-string <str> specifies the log-format string to use as response
5924 payload. The content-type must always be set as
5925 argument to "content-type".
5926
5927 hdr <name> <fmt> adds to the response the HTTP header field whose name
5928 is specified in <name> and whose value is defined by
5929 <fmt>, which follows to the log-format rules.
5930 This parameter is ignored if an errorfile is used.
5931
5932 This directive may be used instead of "errorfile", to define a custom error
5933 message. As "errorfile" directive, it is used for errors detected and
5934 returned by HAProxy. If an errorfile is defined, it is parsed when HAProxy
5935 starts and must be valid according to the HTTP standards. The generated
5936 response must not exceed the configured buffer size (BUFFSIZE), otherwise an
5937 internal error will be returned. Finally, if you consider to use some
5938 http-after-response rules to rewrite these errors, the reserved buffer space
5939 should be available (see "tune.maxrewrite").
5940
5941 The files are read at the same time as the configuration and kept in memory.
5942 For this reason, the errors continue to be returned even when the process is
5943 chrooted, and no file change is considered while the process is running.
5944
Christopher Fauletd5ac6de2020-12-02 08:40:14 +01005945 Note: 400/408/500 errors emitted in early stage of the request parsing are
5946 handled by the multiplexer at a lower level. No custom formatting is
5947 supported at this level. Thus only static error messages, defined with
5948 "errorfile" directive, are supported. However, this limitation only
5949 exists during the request headers parsing or between two transactions.
5950
Christopher Faulet3b967c12020-05-15 15:47:44 +02005951 See also : "errorfile", "errorfiles", "errorloc", "errorloc302",
5952 "errorloc303" and section 3.8 about http-errors.
5953
5954
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02005955http-request <action> [options...] [ { if | unless } <condition> ]
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01005956 Access control for Layer 7 requests
5957
5958 May be used in sections: defaults | frontend | listen | backend
5959 no | yes | yes | yes
5960
Willy Tarreau20b0de52012-12-24 15:45:22 +01005961 The http-request statement defines a set of rules which apply to layer 7
5962 processing. The rules are evaluated in their declaration order when they are
5963 met in a frontend, listen or backend section. Any rule may optionally be
5964 followed by an ACL-based condition, in which case it will only be evaluated
5965 if the condition is true.
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01005966
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02005967 The first keyword is the rule's action. The supported actions are described
5968 below.
Willy Tarreau20b0de52012-12-24 15:45:22 +01005969
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02005970 There is no limit to the number of http-request statements per instance.
Willy Tarreau20b0de52012-12-24 15:45:22 +01005971
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02005972 Example:
5973 acl nagios src 192.168.129.3
5974 acl local_net src 192.168.0.0/16
5975 acl auth_ok http_auth(L1)
Willy Tarreau20b0de52012-12-24 15:45:22 +01005976
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02005977 http-request allow if nagios
5978 http-request allow if local_net auth_ok
5979 http-request auth realm Gimme if local_net auth_ok
5980 http-request deny
Willy Tarreau81499eb2012-12-27 12:19:02 +01005981
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02005982 Example:
5983 acl key req.hdr(X-Add-Acl-Key) -m found
5984 acl add path /addacl
5985 acl del path /delacl
Willy Tarreau20b0de52012-12-24 15:45:22 +01005986
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02005987 acl myhost hdr(Host) -f myhost.lst
Willy Tarreau20b0de52012-12-24 15:45:22 +01005988
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02005989 http-request add-acl(myhost.lst) %[req.hdr(X-Add-Acl-Key)] if key add
5990 http-request del-acl(myhost.lst) %[req.hdr(X-Add-Acl-Key)] if key del
Thierry FOURNIERdad3d1d2014-04-22 18:07:25 +02005991
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02005992 Example:
5993 acl value req.hdr(X-Value) -m found
5994 acl setmap path /setmap
5995 acl delmap path /delmap
Sasha Pachev218f0642014-06-16 12:05:59 -06005996
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02005997 use_backend bk_appli if { hdr(Host),map_str(map.lst) -m found }
Sasha Pachev218f0642014-06-16 12:05:59 -06005998
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02005999 http-request set-map(map.lst) %[src] %[req.hdr(X-Value)] if setmap value
6000 http-request del-map(map.lst) %[src] if delmap
Sasha Pachev218f0642014-06-16 12:05:59 -06006001
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006002 See also : "stats http-request", section 3.4 about userlists and section 7
6003 about ACL usage.
Sasha Pachev218f0642014-06-16 12:05:59 -06006004
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006005http-request add-acl(<file-name>) <key fmt> [ { if | unless } <condition> ]
Sasha Pachev218f0642014-06-16 12:05:59 -06006006
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006007 This is used to add a new entry into an ACL. The ACL must be loaded from a
6008 file (even a dummy empty file). The file name of the ACL to be updated is
6009 passed between parentheses. It takes one argument: <key fmt>, which follows
6010 log-format rules, to collect content of the new entry. It performs a lookup
6011 in the ACL before insertion, to avoid duplicated (or more) values. This
6012 lookup is done by a linear search and can be expensive with large lists!
6013 It is the equivalent of the "add acl" command from the stats socket, but can
6014 be triggered by an HTTP request.
Sasha Pachev218f0642014-06-16 12:05:59 -06006015
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006016http-request add-header <name> <fmt> [ { if | unless } <condition> ]
Sasha Pachev218f0642014-06-16 12:05:59 -06006017
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006018 This appends an HTTP header field whose name is specified in <name> and
6019 whose value is defined by <fmt> which follows the log-format rules (see
6020 Custom Log Format in section 8.2.4). This is particularly useful to pass
6021 connection-specific information to the server (e.g. the client's SSL
6022 certificate), or to combine several headers into one. This rule is not
6023 final, so it is possible to add other similar rules. Note that header
6024 addition is performed immediately, so one rule might reuse the resulting
6025 header from a previous rule.
Sasha Pachev218f0642014-06-16 12:05:59 -06006026
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006027http-request allow [ { if | unless } <condition> ]
Sasha Pachev218f0642014-06-16 12:05:59 -06006028
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006029 This stops the evaluation of the rules and lets the request pass the check.
6030 No further "http-request" rules are evaluated.
Sasha Pachev218f0642014-06-16 12:05:59 -06006031
Sasha Pachev218f0642014-06-16 12:05:59 -06006032
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006033http-request auth [realm <realm>] [ { if | unless } <condition> ]
Sasha Pachev218f0642014-06-16 12:05:59 -06006034
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006035 This stops the evaluation of the rules and immediately responds with an
6036 HTTP 401 or 407 error code to invite the user to present a valid user name
6037 and password. No further "http-request" rules are evaluated. An optional
6038 "realm" parameter is supported, it sets the authentication realm that is
6039 returned with the response (typically the application's name).
Sasha Pachev218f0642014-06-16 12:05:59 -06006040
Christopher Faulet612f2ea2020-05-27 09:57:28 +02006041 The corresponding proxy's error message is used. It may be customized using
6042 an "errorfile" or an "http-error" directive. For 401 responses, all
6043 occurrences of the WWW-Authenticate header are removed and replaced by a new
6044 one with a basic authentication challenge for realm "<realm>". For 407
6045 responses, the same is done on the Proxy-Authenticate header. If the error
6046 message must not be altered, consider to use "http-request return" rule
6047 instead.
6048
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006049 Example:
6050 acl auth_ok http_auth_group(L1) G1
6051 http-request auth unless auth_ok
Sasha Pachev218f0642014-06-16 12:05:59 -06006052
Jarno Huuskonen251a6b72019-01-04 14:05:02 +02006053http-request cache-use <name> [ { if | unless } <condition> ]
Sasha Pachev218f0642014-06-16 12:05:59 -06006054
Christopher Faulet87f1f3d2019-07-18 14:51:20 +02006055 See section 6.2 about cache setup.
Willy Tarreaua0dc23f2015-01-22 20:46:11 +01006056
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006057http-request capture <sample> [ len <length> | id <id> ]
6058 [ { if | unless } <condition> ]
Willy Tarreaua0dc23f2015-01-22 20:46:11 +01006059
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006060 This captures sample expression <sample> from the request buffer, and
6061 converts it to a string of at most <len> characters. The resulting string is
6062 stored into the next request "capture" slot, so it will possibly appear next
6063 to some captured HTTP headers. It will then automatically appear in the logs,
6064 and it will be possible to extract it using sample fetch rules to feed it
6065 into headers or anything. The length should be limited given that this size
6066 will be allocated for each capture during the whole session life.
6067 Please check section 7.3 (Fetching samples) and "capture request header" for
6068 more information.
Willy Tarreaua0dc23f2015-01-22 20:46:11 +01006069
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006070 If the keyword "id" is used instead of "len", the action tries to store the
6071 captured string in a previously declared capture slot. This is useful to run
6072 captures in backends. The slot id can be declared by a previous directive
Baptiste Assmann19a69b32020-01-16 14:34:22 +01006073 "http-request capture" or with the "declare capture" keyword.
6074
6075 When using this action in a backend, double check that the relevant
6076 frontend(s) have the required capture slots otherwise, this rule will be
6077 ignored at run time. This can't be detected at configuration parsing time
6078 due to HAProxy's ability to dynamically resolve backend name at runtime.
Willy Tarreaua0dc23f2015-01-22 20:46:11 +01006079
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006080http-request del-acl(<file-name>) <key fmt> [ { if | unless } <condition> ]
Willy Tarreaua0dc23f2015-01-22 20:46:11 +01006081
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006082 This is used to delete an entry from an ACL. The ACL must be loaded from a
6083 file (even a dummy empty file). The file name of the ACL to be updated is
6084 passed between parentheses. It takes one argument: <key fmt>, which follows
6085 log-format rules, to collect content of the entry to delete.
6086 It is the equivalent of the "del acl" command from the stats socket, but can
6087 be triggered by an HTTP request.
Willy Tarreaua0dc23f2015-01-22 20:46:11 +01006088
Maciej Zdebebdd4c52020-11-20 13:58:48 +00006089http-request del-header <name> [ -m <meth> ] [ { if | unless } <condition> ]
Willy Tarreauf4c43c12013-06-11 17:01:13 +02006090
Maciej Zdebebdd4c52020-11-20 13:58:48 +00006091 This removes all HTTP header fields whose name is specified in <name>. <meth>
6092 is the matching method, applied on the header name. Supported matching methods
6093 are "str" (exact match), "beg" (prefix match), "end" (suffix match), "sub"
6094 (substring match) and "reg" (regex match). If not specified, exact matching
6095 method is used.
Willy Tarreau9a355ec2013-06-11 17:45:46 +02006096
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006097http-request del-map(<file-name>) <key fmt> [ { if | unless } <condition> ]
Willy Tarreau42cf39e2013-06-11 18:51:32 +02006098
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006099 This is used to delete an entry from a MAP. The MAP must be loaded from a
6100 file (even a dummy empty file). The file name of the MAP to be updated is
6101 passed between parentheses. It takes one argument: <key fmt>, which follows
6102 log-format rules, to collect content of the entry to delete.
6103 It takes one argument: "file name" It is the equivalent of the "del map"
6104 command from the stats socket, but can be triggered by an HTTP request.
Willy Tarreau51347ed2013-06-11 19:34:13 +02006105
Christopher Faulet5cb513a2020-05-13 17:56:56 +02006106http-request deny [deny_status <status>] [ { if | unless } <condition> ]
6107http-request deny [ { status | deny_status } <code>] [content-type <type>]
6108 [ { default-errorfiles | errorfile <file> | errorfiles <name> |
6109 file <file> | lf-file <file> | string <str> | lf-string <fmt> } ]
6110 [ hdr <name> <fmt> ]*
6111 [ { if | unless } <condition> ]
Patrick Hemmer268a7072018-05-11 12:52:31 -04006112
Christopher Faulet5cb513a2020-05-13 17:56:56 +02006113 This stops the evaluation of the rules and immediately rejects the request.
6114 By default an HTTP 403 error is returned. But the response may be customized
6115 using same syntax than "http-request return" rules. Thus, see "http-request
Ilya Shipitsin11057a32020-06-21 21:18:27 +05006116 return" for details. For compatibility purpose, when no argument is defined,
Christopher Faulet5cb513a2020-05-13 17:56:56 +02006117 or only "deny_status", the argument "default-errorfiles" is implied. It means
6118 "http-request deny [deny_status <status>]" is an alias of
6119 "http-request deny [status <status>] default-errorfiles".
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006120 No further "http-request" rules are evaluated.
Christopher Faulet5cb513a2020-05-13 17:56:56 +02006121 See also "http-request return".
Patrick Hemmer268a7072018-05-11 12:52:31 -04006122
Olivier Houchard602bf7d2019-05-10 13:59:15 +02006123http-request disable-l7-retry [ { if | unless } <condition> ]
6124 This disables any attempt to retry the request if it fails for any other
6125 reason than a connection failure. This can be useful for example to make
6126 sure POST requests aren't retried on failure.
6127
Baptiste Assmann333939c2019-01-21 08:34:50 +01006128http-request do-resolve(<var>,<resolvers>,[ipv4,ipv6]) <expr> :
6129
6130 This action performs a DNS resolution of the output of <expr> and stores
6131 the result in the variable <var>. It uses the DNS resolvers section
6132 pointed by <resolvers>.
6133 It is possible to choose a resolution preference using the optional
6134 arguments 'ipv4' or 'ipv6'.
6135 When performing the DNS resolution, the client side connection is on
6136 pause waiting till the end of the resolution.
6137 If an IP address can be found, it is stored into <var>. If any kind of
6138 error occurs, then <var> is not set.
6139 One can use this action to discover a server IP address at run time and
6140 based on information found in the request (IE a Host header).
6141 If this action is used to find the server's IP address (using the
6142 "set-dst" action), then the server IP address in the backend must be set
6143 to 0.0.0.0.
6144
6145 Example:
6146 resolvers mydns
6147 nameserver local 127.0.0.53:53
6148 nameserver google 8.8.8.8:53
6149 timeout retry 1s
6150 hold valid 10s
6151 hold nx 3s
6152 hold other 3s
6153 hold obsolete 0s
6154 accepted_payload_size 8192
6155
6156 frontend fe
6157 bind 10.42.0.1:80
6158 http-request do-resolve(txn.myip,mydns,ipv4) hdr(Host),lower
6159 http-request capture var(txn.myip) len 40
6160
6161 # return 503 when the variable is not set,
6162 # which mean DNS resolution error
6163 use_backend b_503 unless { var(txn.myip) -m found }
6164
6165 default_backend be
6166
6167 backend b_503
6168 # dummy backend used to return 503.
6169 # one can use the errorfile directive to send a nice
6170 # 503 error page to end users
6171
6172 backend be
6173 # rule to prevent HAProxy from reconnecting to services
6174 # on the local network (forged DNS name used to scan the network)
6175 http-request deny if { var(txn.myip) -m ip 127.0.0.0/8 10.0.0.0/8 }
6176 http-request set-dst var(txn.myip)
6177 server clear 0.0.0.0:0
6178
6179 NOTE: Don't forget to set the "protection" rules to ensure HAProxy won't
6180 be used to scan the network or worst won't loop over itself...
6181
Frédéric Lécaille06f5b642018-11-12 11:01:10 +01006182http-request early-hint <name> <fmt> [ { if | unless } <condition> ]
6183
6184 This is used to build an HTTP 103 Early Hints response prior to any other one.
6185 This appends an HTTP header field to this response whose name is specified in
6186 <name> and whose value is defined by <fmt> which follows the log-format rules
6187 (see Custom Log Format in section 8.2.4). This is particularly useful to pass
Frédéric Lécaille3aac1062018-11-13 09:42:13 +01006188 to the client some Link headers to preload resources required to render the
6189 HTML documents.
Frédéric Lécaille06f5b642018-11-12 11:01:10 +01006190
6191 See RFC 8297 for more information.
6192
Tim Duesterhusd371e992021-04-15 21:45:58 +02006193http-request normalize-uri <normalizer> [ { if | unless } <condition> ]
Tim Duesterhusdec1c362021-05-10 17:28:26 +02006194http-request normalize-uri fragment-encode [ { if | unless } <condition> ]
Tim Duesterhusc9e05ab2021-05-10 17:28:25 +02006195http-request normalize-uri fragment-strip [ { if | unless } <condition> ]
Tim Duesterhus5be6ab22021-04-17 11:21:10 +02006196http-request normalize-uri path-merge-slashes [ { if | unless } <condition> ]
Maximilian Maderff3bb8b2021-04-21 00:22:50 +02006197http-request normalize-uri path-strip-dot [ { if | unless } <condition> ]
Tim Duesterhus5be6ab22021-04-17 11:21:10 +02006198http-request normalize-uri path-strip-dotdot [ full ] [ { if | unless } <condition> ]
Tim Duesterhus2e4a18e2021-04-21 21:20:36 +02006199http-request normalize-uri percent-decode-unreserved [ strict ] [ { if | unless } <condition> ]
Tim Duesterhus5be6ab22021-04-17 11:21:10 +02006200http-request normalize-uri percent-to-uppercase [ strict ] [ { if | unless } <condition> ]
6201http-request normalize-uri query-sort-by-name [ { if | unless } <condition> ]
Tim Duesterhusd371e992021-04-15 21:45:58 +02006202
Tim Duesterhusb918a4a2021-04-16 23:52:29 +02006203 Performs normalization of the request's URI.
6204
Tim Duesterhus2963fd32021-04-17 00:24:56 +02006205 URI normalization in HAProxy 2.4 is currently available as an experimental
Amaury Denoyellea9e639a2021-05-06 15:50:12 +02006206 technical preview. As such, it requires the global directive
6207 'expose-experimental-directives' first to be able to invoke it. You should be
6208 prepared that the behavior of normalizers might change to fix possible
6209 issues, possibly breaking proper request processing in your infrastructure.
Tim Duesterhus2963fd32021-04-17 00:24:56 +02006210
Tim Duesterhusb918a4a2021-04-16 23:52:29 +02006211 Each normalizer handles a single type of normalization to allow for a
6212 fine-grained selection of the level of normalization that is appropriate for
6213 the supported backend.
6214
6215 As an example the "path-strip-dotdot" normalizer might be useful for a static
6216 fileserver that directly maps the requested URI to the path within the local
6217 filesystem. However it might break routing of an API that expects a specific
6218 number of segments in the path.
6219
Tim Duesterhus2e4a18e2021-04-21 21:20:36 +02006220 It is important to note that some normalizers might result in unsafe
6221 transformations for broken URIs. It might also be possible that a combination
6222 of normalizers that are safe by themselves results in unsafe transformations
6223 when improperly combined.
6224
6225 As an example the "percent-decode-unreserved" normalizer might result in
6226 unexpected results when a broken URI includes bare percent characters. One
6227 such a broken URI is "/%%36%36" which would be decoded to "/%66" which in
6228 turn is equivalent to "/f". By specifying the "strict" option requests to
6229 such a broken URI would safely be rejected.
6230
Tim Duesterhusb918a4a2021-04-16 23:52:29 +02006231 The following normalizers are available:
Tim Duesterhusd371e992021-04-15 21:45:58 +02006232
Tim Duesterhusdec1c362021-05-10 17:28:26 +02006233 - fragment-encode: Encodes "#" as "%23".
6234
6235 The "fragment-strip" normalizer should be preferred, unless it is known
6236 that broken clients do not correctly encode '#' within the path component.
6237
6238 Example:
6239 - /#foo -> /%23foo
6240
Tim Duesterhusc9e05ab2021-05-10 17:28:25 +02006241 - fragment-strip: Removes the URI's "fragment" component.
6242
6243 According to RFC 3986#3.5 the "fragment" component of an URI should not
6244 be sent, but handled by the User Agent after retrieving a resource.
6245
6246 This normalizer should be applied first to ensure that the fragment is
6247 not interpreted as part of the request's path component.
6248
6249 Example:
6250 - /#foo -> /
6251
Tim Duesterhusd6d33de2021-04-21 21:20:35 +02006252 - path-strip-dot: Removes "/./" segments within the "path" component
6253 (RFC 3986#6.2.2.3).
Maximilian Maderff3bb8b2021-04-21 00:22:50 +02006254
Tim Duesterhus2e4a18e2021-04-21 21:20:36 +02006255 Segments including percent encoded dots ("%2E") will not be detected. Use
6256 the "percent-decode-unreserved" normalizer first if this is undesired.
6257
Tim Duesterhus7a95f412021-04-21 21:20:33 +02006258 Example:
6259 - /. -> /
6260 - /./bar/ -> /bar/
6261 - /a/./a -> /a/a
6262 - /.well-known/ -> /.well-known/ (no change)
Maximilian Maderff3bb8b2021-04-21 00:22:50 +02006263
Tim Duesterhusd6d33de2021-04-21 21:20:35 +02006264 - path-strip-dotdot: Normalizes "/../" segments within the "path" component
6265 (RFC 3986#6.2.2.3).
6266
Tim Duesterhus5be6ab22021-04-17 11:21:10 +02006267 This merges segments that attempt to access the parent directory with
Tim Duesterhus2e4a18e2021-04-21 21:20:36 +02006268 their preceding segment.
6269
6270 Empty segments do not receive special treatment. Use the "merge-slashes"
6271 normalizer first if this is undesired.
6272
6273 Segments including percent encoded dots ("%2E") will not be detected. Use
6274 the "percent-decode-unreserved" normalizer first if this is undesired.
Tim Duesterhus9982fc22021-04-15 21:45:59 +02006275
6276 Example:
6277 - /foo/../ -> /
6278 - /foo/../bar/ -> /bar/
6279 - /foo/bar/../ -> /foo/
6280 - /../bar/ -> /../bar/
Tim Duesterhus560e1a62021-04-15 21:46:00 +02006281 - /bar/../../ -> /../
Tim Duesterhus9982fc22021-04-15 21:45:59 +02006282 - /foo//../ -> /foo/
Tim Duesterhus2e4a18e2021-04-21 21:20:36 +02006283 - /foo/%2E%2E/ -> /foo/%2E%2E/
Tim Duesterhus9982fc22021-04-15 21:45:59 +02006284
Tim Duesterhus560e1a62021-04-15 21:46:00 +02006285 If the "full" option is specified then "../" at the beginning will be
6286 removed as well:
6287
6288 Example:
6289 - /../bar/ -> /bar/
6290 - /bar/../../ -> /
6291
Tim Duesterhus5be6ab22021-04-17 11:21:10 +02006292 - path-merge-slashes: Merges adjacent slashes within the "path" component
6293 into a single slash.
Tim Duesterhusd371e992021-04-15 21:45:58 +02006294
6295 Example:
6296 - // -> /
6297 - /foo//bar -> /foo/bar
6298
Tim Duesterhus2e4a18e2021-04-21 21:20:36 +02006299 - percent-decode-unreserved: Decodes unreserved percent encoded characters to
6300 their representation as a regular character (RFC 3986#6.2.2.2).
6301
6302 The set of unreserved characters includes all letters, all digits, "-",
6303 ".", "_", and "~".
6304
6305 Example:
6306 - /%61dmin -> /admin
6307 - /foo%3Fbar=baz -> /foo%3Fbar=baz (no change)
6308 - /%%36%36 -> /%66 (unsafe)
6309 - /%ZZ -> /%ZZ
6310
6311 If the "strict" option is specified then invalid sequences will result
6312 in a HTTP 400 Bad Request being returned.
6313
6314 Example:
6315 - /%%36%36 -> HTTP 400
6316 - /%ZZ -> HTTP 400
6317
Tim Duesterhus5be6ab22021-04-17 11:21:10 +02006318 - percent-to-uppercase: Uppercases letters within percent-encoded sequences
Tim Duesterhusc315efd2021-04-21 21:20:34 +02006319 (RFC 3986#6.2.2.1).
Tim Duesterhusa4071932021-04-15 21:46:02 +02006320
6321 Example:
6322 - /%6f -> /%6F
6323 - /%zz -> /%zz
6324
6325 If the "strict" option is specified then invalid sequences will result
6326 in a HTTP 400 Bad Request being returned.
6327
6328 Example:
6329 - /%zz -> HTTP 400
6330
Tim Duesterhus5be6ab22021-04-17 11:21:10 +02006331 - query-sort-by-name: Sorts the query string parameters by parameter name.
Tim Duesterhusd7b89be2021-04-15 21:46:01 +02006332 Parameters are assumed to be delimited by '&'. Shorter names sort before
6333 longer names and identical parameter names maintain their relative order.
6334
6335 Example:
6336 - /?c=3&a=1&b=2 -> /?a=1&b=2&c=3
6337 - /?aaa=3&a=1&aa=2 -> /?a=1&aa=2&aaa=3
6338 - /?a=3&b=4&a=1&b=5&a=2 -> /?a=3&a=1&a=2&b=4&b=5
6339
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006340http-request redirect <rule> [ { if | unless } <condition> ]
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02006341
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006342 This performs an HTTP redirection based on a redirect rule. This is exactly
6343 the same as the "redirect" statement except that it inserts a redirect rule
6344 which can be processed in the middle of other "http-request" rules and that
6345 these rules use the "log-format" strings. See the "redirect" keyword for the
6346 rule's syntax.
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02006347
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006348http-request reject [ { if | unless } <condition> ]
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02006349
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006350 This stops the evaluation of the rules and immediately closes the connection
6351 without sending any response. It acts similarly to the
6352 "tcp-request content reject" rules. It can be useful to force an immediate
6353 connection closure on HTTP/2 connections.
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02006354
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006355http-request replace-header <name> <match-regex> <replace-fmt>
6356 [ { if | unless } <condition> ]
Willy Tarreaua9083d02015-05-08 15:27:59 +02006357
Ilya Shipitsin8525fd92020-02-29 12:34:59 +05006358 This matches the value of all occurrences of header field <name> against
Tim Duesterhus2252beb2019-10-29 00:05:13 +01006359 <match-regex>. Matching is performed case-sensitively. Matching values are
6360 completely replaced by <replace-fmt>. Format characters are allowed in
6361 <replace-fmt> and work like <fmt> arguments in "http-request add-header".
6362 Standard back-references using the backslash ('\') followed by a number are
6363 supported.
Thierry FOURNIER82bf70d2015-05-26 17:58:29 +02006364
Tim Duesterhus2252beb2019-10-29 00:05:13 +01006365 This action acts on whole header lines, regardless of the number of values
6366 they may contain. Thus it is well-suited to process headers naturally
6367 containing commas in their value, such as If-Modified-Since. Headers that
6368 contain a comma-separated list of values, such as Accept, should be processed
6369 using "http-request replace-value".
William Lallemand86d0df02017-11-24 21:36:45 +01006370
Tim Duesterhus2252beb2019-10-29 00:05:13 +01006371 Example:
6372 http-request replace-header Cookie foo=([^;]*);(.*) foo=\1;ip=%bi;\2
6373
6374 # applied to:
6375 Cookie: foo=foobar; expires=Tue, 14-Jun-2016 01:40:45 GMT;
6376
6377 # outputs:
6378 Cookie: foo=foobar;ip=192.168.1.20; expires=Tue, 14-Jun-2016 01:40:45 GMT;
6379
6380 # assuming the backend IP is 192.168.1.20
Willy Tarreau09448f72014-06-25 18:12:15 +02006381
Tim Duesterhus2252beb2019-10-29 00:05:13 +01006382 http-request replace-header User-Agent curl foo
6383
6384 # applied to:
6385 User-Agent: curl/7.47.0
Willy Tarreau09448f72014-06-25 18:12:15 +02006386
Tim Duesterhus2252beb2019-10-29 00:05:13 +01006387 # outputs:
6388 User-Agent: foo
Willy Tarreau09448f72014-06-25 18:12:15 +02006389
Willy Tarreau262c3f12019-12-17 06:52:51 +01006390http-request replace-path <match-regex> <replace-fmt>
6391 [ { if | unless } <condition> ]
6392
6393 This works like "replace-header" except that it works on the request's path
6394 component instead of a header. The path component starts at the first '/'
Christopher Faulet82c83322020-09-02 14:16:59 +02006395 after an optional scheme+authority and ends before the question mark. Thus,
6396 the replacement does not modify the scheme, the authority and the
6397 query-string.
Willy Tarreau262c3f12019-12-17 06:52:51 +01006398
6399 It is worth noting that regular expressions may be more expensive to evaluate
6400 than certain ACLs, so rare replacements may benefit from a condition to avoid
6401 performing the evaluation at all if it does not match.
6402
6403 Example:
6404 # prefix /foo : turn /bar?q=1 into /foo/bar?q=1 :
6405 http-request replace-path (.*) /foo\1
6406
Willy Tarreau262c3f12019-12-17 06:52:51 +01006407 # strip /foo : turn /foo/bar?q=1 into /bar?q=1
6408 http-request replace-path /foo/(.*) /\1
6409 # or more efficient if only some requests match :
6410 http-request replace-path /foo/(.*) /\1 if { url_beg /foo/ }
6411
Christopher Faulet312294f2020-09-02 17:17:44 +02006412http-request replace-pathq <match-regex> <replace-fmt>
6413 [ { if | unless } <condition> ]
6414
6415 This does the same as "http-request replace-path" except that the path
6416 contains the query-string if any is present. Thus, the path and the
6417 query-string are replaced.
6418
6419 Example:
6420 # suffix /foo : turn /bar?q=1 into /bar/foo?q=1 :
6421 http-request replace-pathq ([^?]*)(\?(.*))? \1/foo\2
6422
Willy Tarreau33810222019-06-12 17:44:02 +02006423http-request replace-uri <match-regex> <replace-fmt>
6424 [ { if | unless } <condition> ]
6425
Tim Duesterhus2252beb2019-10-29 00:05:13 +01006426 This works like "replace-header" except that it works on the request's URI part
6427 instead of a header. The URI part may contain an optional scheme, authority or
6428 query string. These are considered to be part of the value that is matched
6429 against.
6430
6431 It is worth noting that regular expressions may be more expensive to evaluate
6432 than certain ACLs, so rare replacements may benefit from a condition to avoid
6433 performing the evaluation at all if it does not match.
Willy Tarreau33810222019-06-12 17:44:02 +02006434
Willy Tarreau62b59132019-12-17 06:51:20 +01006435 IMPORTANT NOTE: historically in HTTP/1.x, the vast majority of requests sent
6436 by browsers use the "origin form", which differs from the "absolute form" in
6437 that they do not contain a scheme nor authority in the URI portion. Mostly
6438 only requests sent to proxies, those forged by hand and some emitted by
6439 certain applications use the absolute form. As such, "replace-uri" usually
6440 works fine most of the time in HTTP/1.x with rules starting with a "/". But
6441 with HTTP/2, clients are encouraged to send absolute URIs only, which look
6442 like the ones HTTP/1 clients use to talk to proxies. Such partial replace-uri
6443 rules may then fail in HTTP/2 when they work in HTTP/1. Either the rules need
Willy Tarreau262c3f12019-12-17 06:52:51 +01006444 to be adapted to optionally match a scheme and authority, or replace-path
6445 should be used.
Willy Tarreau33810222019-06-12 17:44:02 +02006446
Willy Tarreau62b59132019-12-17 06:51:20 +01006447 Example:
6448 # rewrite all "http" absolute requests to "https":
6449 http-request replace-uri ^http://(.*) https://\1
Willy Tarreau33810222019-06-12 17:44:02 +02006450
Willy Tarreau62b59132019-12-17 06:51:20 +01006451 # prefix /foo : turn /bar?q=1 into /foo/bar?q=1 :
6452 http-request replace-uri ([^/:]*://[^/]*)?(.*) \1/foo\2
Willy Tarreau33810222019-06-12 17:44:02 +02006453
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006454http-request replace-value <name> <match-regex> <replace-fmt>
6455 [ { if | unless } <condition> ]
Willy Tarreau09448f72014-06-25 18:12:15 +02006456
Tim Duesterhus2252beb2019-10-29 00:05:13 +01006457 This works like "replace-header" except that it matches the regex against
6458 every comma-delimited value of the header field <name> instead of the
6459 entire header. This is suited for all headers which are allowed to carry
6460 more than one value. An example could be the Accept header.
Willy Tarreau09448f72014-06-25 18:12:15 +02006461
Tim Duesterhus2252beb2019-10-29 00:05:13 +01006462 Example:
6463 http-request replace-value X-Forwarded-For ^192\.168\.(.*)$ 172.16.\1
Thierry FOURNIER236657b2015-08-19 08:25:14 +02006464
Tim Duesterhus2252beb2019-10-29 00:05:13 +01006465 # applied to:
6466 X-Forwarded-For: 192.168.10.1, 192.168.13.24, 10.0.0.37
Thierry FOURNIERe0627bd2015-08-04 08:20:33 +02006467
Tim Duesterhus2252beb2019-10-29 00:05:13 +01006468 # outputs:
6469 X-Forwarded-For: 172.16.10.1, 172.16.13.24, 10.0.0.37
Frédéric Lécaille6778b272018-01-29 15:22:53 +01006470
Christopher Faulet24231ab2020-01-24 17:44:23 +01006471http-request return [status <code>] [content-type <type>]
6472 [ { default-errorfiles | errorfile <file> | errorfiles <name> |
6473 file <file> | lf-file <file> | string <str> | lf-string <fmt> } ]
Christopher Faulet4a2c1422020-01-31 17:36:01 +01006474 [ hdr <name> <fmt> ]*
Christopher Faulet24231ab2020-01-24 17:44:23 +01006475 [ { if | unless } <condition> ]
6476
Ilya Shipitsin8525fd92020-02-29 12:34:59 +05006477 This stops the evaluation of the rules and immediately returns a response. The
Christopher Faulet24231ab2020-01-24 17:44:23 +01006478 default status code used for the response is 200. It can be optionally
6479 specified as an arguments to "status". The response content-type may also be
Daniel Corbett67a82712020-07-06 23:01:19 -04006480 specified as an argument to "content-type". Finally the response itself may
Sébastien Grossab877122020-10-08 10:06:03 +02006481 be defined. It can be a full HTTP response specifying the errorfile to use,
Ilya Shipitsin8525fd92020-02-29 12:34:59 +05006482 or the response payload specifying the file or the string to use. These rules
Christopher Faulet24231ab2020-01-24 17:44:23 +01006483 are followed to create the response :
6484
6485 * If neither the errorfile nor the payload to use is defined, a dummy
6486 response is returned. Only the "status" argument is considered. It can be
6487 any code in the range [200, 599]. The "content-type" argument, if any, is
6488 ignored.
6489
6490 * If "default-errorfiles" argument is set, the proxy's errorfiles are
6491 considered. If the "status" argument is defined, it must be one of the
Daniel Corbett9f0843f2021-05-08 10:50:37 -04006492 status code handled by HAProxy (200, 400, 403, 404, 405, 408, 410, 413,
Christopher Faulete095f312020-12-07 11:22:24 +01006493 425, 429, 500, 501, 502, 503, and 504). The "content-type" argument, if
6494 any, is ignored.
Christopher Faulet24231ab2020-01-24 17:44:23 +01006495
6496 * If a specific errorfile is defined, with an "errorfile" argument, the
6497 corresponding file, containing a full HTTP response, is returned. Only the
6498 "status" argument is considered. It must be one of the status code handled
Daniel Corbett9f0843f2021-05-08 10:50:37 -04006499 by HAProxy (200, 400, 403, 404, 405, 408, 410, 413, 425, 429, 500, 501,
Christopher Faulete095f312020-12-07 11:22:24 +01006500 502, 503, and 504). The "content-type" argument, if any, is ignored.
Christopher Faulet24231ab2020-01-24 17:44:23 +01006501
6502 * If an http-errors section is defined, with an "errorfiles" argument, the
6503 corresponding file in the specified http-errors section, containing a full
6504 HTTP response, is returned. Only the "status" argument is considered. It
Daniel Corbett9f0843f2021-05-08 10:50:37 -04006505 must be one of the status code handled by HAProxy (200, 400, 403, 404, 405,
Christopher Faulete095f312020-12-07 11:22:24 +01006506 408, 410, 413, 425, 429, 500, 501, 502, 503, and 504). The "content-type"
Anthonin Bonnefoy85048f82020-06-22 09:17:01 +02006507 argument, if any, is ignored.
Christopher Faulet24231ab2020-01-24 17:44:23 +01006508
6509 * If a "file" or a "lf-file" argument is specified, the file's content is
6510 used as the response payload. If the file is not empty, its content-type
6511 must be set as argument to "content-type". Otherwise, any "content-type"
6512 argument is ignored. With a "lf-file" argument, the file's content is
6513 evaluated as a log-format string. With a "file" argument, it is considered
6514 as a raw content.
6515
6516 * If a "string" or "lf-string" argument is specified, the defined string is
6517 used as the response payload. The content-type must always be set as
6518 argument to "content-type". With a "lf-string" argument, the string is
6519 evaluated as a log-format string. With a "string" argument, it is
6520 considered as a raw string.
6521
Sébastien Grossab877122020-10-08 10:06:03 +02006522 When the response is not based on an errorfile, it is possible to append HTTP
Christopher Faulet4a2c1422020-01-31 17:36:01 +01006523 header fields to the response using "hdr" arguments. Otherwise, all "hdr"
6524 arguments are ignored. For each one, the header name is specified in <name>
6525 and its value is defined by <fmt> which follows the log-format rules.
6526
Christopher Faulet24231ab2020-01-24 17:44:23 +01006527 Note that the generated response must be smaller than a buffer. And to avoid
6528 any warning, when an errorfile or a raw file is loaded, the buffer space
Sébastien Grossab877122020-10-08 10:06:03 +02006529 reserved for the headers rewriting should also be free.
Christopher Faulet24231ab2020-01-24 17:44:23 +01006530
6531 No further "http-request" rules are evaluated.
6532
6533 Example:
Daniel Corbett67a82712020-07-06 23:01:19 -04006534 http-request return errorfile /etc/haproxy/errorfiles/200.http \
Christopher Faulet24231ab2020-01-24 17:44:23 +01006535 if { path /ping }
6536
6537 http-request return content-type image/x-icon file /var/www/favicon.ico \
6538 if { path /favicon.ico }
6539
6540 http-request return status 403 content-type text/plain \
6541 lf-string "Access denied. IP %[src] is blacklisted." \
6542 if { src -f /etc/haproxy/blacklist.lst }
6543
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006544http-request sc-inc-gpc0(<sc-id>) [ { if | unless } <condition> ]
6545http-request sc-inc-gpc1(<sc-id>) [ { if | unless } <condition> ]
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02006546
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006547 This actions increments the GPC0 or GPC1 counter according with the sticky
6548 counter designated by <sc-id>. If an error occurs, this action silently fails
6549 and the actions evaluation continues.
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02006550
Cédric Dufour0d7712d2019-11-06 18:38:53 +01006551http-request sc-set-gpt0(<sc-id>) { <int> | <expr> }
6552 [ { if | unless } <condition> ]
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02006553
Cédric Dufour0d7712d2019-11-06 18:38:53 +01006554 This action sets the 32-bit unsigned GPT0 tag according to the sticky counter
6555 designated by <sc-id> and the value of <int>/<expr>. The expected result is a
6556 boolean. If an error occurs, this action silently fails and the actions
6557 evaluation continues.
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02006558
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006559http-request set-dst <expr> [ { if | unless } <condition> ]
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02006560
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006561 This is used to set the destination IP address to the value of specified
6562 expression. Useful when a proxy in front of HAProxy rewrites destination IP,
6563 but provides the correct IP in a HTTP header; or you want to mask the IP for
6564 privacy. If you want to connect to the new address/port, use '0.0.0.0:0' as a
6565 server address in the backend.
Christopher Faulet85d79c92016-11-09 16:54:56 +01006566
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006567 Arguments:
6568 <expr> Is a standard HAProxy expression formed by a sample-fetch followed
6569 by some converters.
Christopher Faulet85d79c92016-11-09 16:54:56 +01006570
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006571 Example:
6572 http-request set-dst hdr(x-dst)
6573 http-request set-dst dst,ipmask(24)
Christopher Faulet85d79c92016-11-09 16:54:56 +01006574
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006575 When possible, set-dst preserves the original destination port as long as the
6576 address family allows it, otherwise the destination port is set to 0.
Adis Nezirovic2fbcafc2015-07-06 15:44:30 +02006577
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006578http-request set-dst-port <expr> [ { if | unless } <condition> ]
Adis Nezirovic2fbcafc2015-07-06 15:44:30 +02006579
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006580 This is used to set the destination port address to the value of specified
6581 expression. If you want to connect to the new address/port, use '0.0.0.0:0'
6582 as a server address in the backend.
Adis Nezirovic2fbcafc2015-07-06 15:44:30 +02006583
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006584 Arguments:
6585 <expr> Is a standard HAProxy expression formed by a sample-fetch
6586 followed by some converters.
Adis Nezirovic2fbcafc2015-07-06 15:44:30 +02006587
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006588 Example:
6589 http-request set-dst-port hdr(x-port)
6590 http-request set-dst-port int(4000)
Adis Nezirovic2fbcafc2015-07-06 15:44:30 +02006591
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006592 When possible, set-dst-port preserves the original destination address as
6593 long as the address family supports a port, otherwise it forces the
6594 destination address to IPv4 "0.0.0.0" before rewriting the port.
William Lallemand44be6402016-05-25 01:51:35 +02006595
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006596http-request set-header <name> <fmt> [ { if | unless } <condition> ]
William Lallemand44be6402016-05-25 01:51:35 +02006597
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006598 This does the same as "http-request add-header" except that the header name
6599 is first removed if it existed. This is useful when passing security
6600 information to the server, where the header must not be manipulated by
6601 external users. Note that the new value is computed before the removal so it
6602 is possible to concatenate a value to an existing header.
William Lallemand44be6402016-05-25 01:51:35 +02006603
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006604 Example:
6605 http-request set-header X-Haproxy-Current-Date %T
6606 http-request set-header X-SSL %[ssl_fc]
6607 http-request set-header X-SSL-Session_ID %[ssl_fc_session_id,hex]
6608 http-request set-header X-SSL-Client-Verify %[ssl_c_verify]
6609 http-request set-header X-SSL-Client-DN %{+Q}[ssl_c_s_dn]
6610 http-request set-header X-SSL-Client-CN %{+Q}[ssl_c_s_dn(cn)]
6611 http-request set-header X-SSL-Issuer %{+Q}[ssl_c_i_dn]
6612 http-request set-header X-SSL-Client-NotBefore %{+Q}[ssl_c_notbefore]
6613 http-request set-header X-SSL-Client-NotAfter %{+Q}[ssl_c_notafter]
William Lallemand44be6402016-05-25 01:51:35 +02006614
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006615http-request set-log-level <level> [ { if | unless } <condition> ]
William Lallemand44be6402016-05-25 01:51:35 +02006616
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006617 This is used to change the log level of the current request when a certain
6618 condition is met. Valid levels are the 8 syslog levels (see the "log"
6619 keyword) plus the special level "silent" which disables logging for this
6620 request. This rule is not final so the last matching rule wins. This rule
6621 can be useful to disable health checks coming from another equipment.
William Lallemand13e9b0c2016-05-25 02:34:07 +02006622
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006623http-request set-map(<file-name>) <key fmt> <value fmt>
6624 [ { if | unless } <condition> ]
William Lallemand13e9b0c2016-05-25 02:34:07 +02006625
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006626 This is used to add a new entry into a MAP. The MAP must be loaded from a
6627 file (even a dummy empty file). The file name of the MAP to be updated is
6628 passed between parentheses. It takes 2 arguments: <key fmt>, which follows
6629 log-format rules, used to collect MAP key, and <value fmt>, which follows
6630 log-format rules, used to collect content for the new entry.
6631 It performs a lookup in the MAP before insertion, to avoid duplicated (or
6632 more) values. This lookup is done by a linear search and can be expensive
6633 with large lists! It is the equivalent of the "set map" command from the
6634 stats socket, but can be triggered by an HTTP request.
William Lallemand13e9b0c2016-05-25 02:34:07 +02006635
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006636http-request set-mark <mark> [ { if | unless } <condition> ]
William Lallemand13e9b0c2016-05-25 02:34:07 +02006637
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006638 This is used to set the Netfilter MARK on all packets sent to the client to
6639 the value passed in <mark> on platforms which support it. This value is an
6640 unsigned 32 bit value which can be matched by netfilter and by the routing
6641 table. It can be expressed both in decimal or hexadecimal format (prefixed by
6642 "0x"). This can be useful to force certain packets to take a different route
6643 (for example a cheaper network path for bulk downloads). This works on Linux
6644 kernels 2.6.32 and above and requires admin privileges.
Willy Tarreau00005ce2016-10-21 15:07:45 +02006645
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006646http-request set-method <fmt> [ { if | unless } <condition> ]
William Lallemand13e9b0c2016-05-25 02:34:07 +02006647
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006648 This rewrites the request method with the result of the evaluation of format
6649 string <fmt>. There should be very few valid reasons for having to do so as
6650 this is more likely to break something than to fix it.
William Lallemand13e9b0c2016-05-25 02:34:07 +02006651
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006652http-request set-nice <nice> [ { if | unless } <condition> ]
William Lallemand13e9b0c2016-05-25 02:34:07 +02006653
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006654 This sets the "nice" factor of the current request being processed. It only
6655 has effect against the other requests being processed at the same time.
6656 The default value is 0, unless altered by the "nice" setting on the "bind"
6657 line. The accepted range is -1024..1024. The higher the value, the nicest
6658 the request will be. Lower values will make the request more important than
6659 other ones. This can be useful to improve the speed of some requests, or
6660 lower the priority of non-important requests. Using this setting without
6661 prior experimentation can cause some major slowdown.
William Lallemand13e9b0c2016-05-25 02:34:07 +02006662
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006663http-request set-path <fmt> [ { if | unless } <condition> ]
Willy Tarreau00005ce2016-10-21 15:07:45 +02006664
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006665 This rewrites the request path with the result of the evaluation of format
6666 string <fmt>. The query string, if any, is left intact. If a scheme and
6667 authority is found before the path, they are left intact as well. If the
6668 request doesn't have a path ("*"), this one is replaced with the format.
6669 This can be used to prepend a directory component in front of a path for
6670 example. See also "http-request set-query" and "http-request set-uri".
Willy Tarreau2d392c22015-08-24 01:43:45 +02006671
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006672 Example :
6673 # prepend the host name before the path
6674 http-request set-path /%[hdr(host)]%[path]
Christopher Faulet76c09ef2017-09-21 11:03:52 +02006675
Christopher Faulet312294f2020-09-02 17:17:44 +02006676http-request set-pathq <fmt> [ { if | unless } <condition> ]
6677
6678 This does the same as "http-request set-path" except that the query-string is
6679 also rewritten. It may be used to remove the query-string, including the
6680 question mark (it is not possible using "http-request set-query").
6681
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006682http-request set-priority-class <expr> [ { if | unless } <condition> ]
Olivier Houchardccaa7de2017-10-02 11:51:03 +02006683
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006684 This is used to set the queue priority class of the current request.
6685 The value must be a sample expression which converts to an integer in the
6686 range -2047..2047. Results outside this range will be truncated.
6687 The priority class determines the order in which queued requests are
6688 processed. Lower values have higher priority.
Christopher Faulet76c09ef2017-09-21 11:03:52 +02006689
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006690http-request set-priority-offset <expr> [ { if | unless } <condition> ]
Christopher Faulet76c09ef2017-09-21 11:03:52 +02006691
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006692 This is used to set the queue priority timestamp offset of the current
6693 request. The value must be a sample expression which converts to an integer
6694 in the range -524287..524287. Results outside this range will be truncated.
6695 When a request is queued, it is ordered first by the priority class, then by
6696 the current timestamp adjusted by the given offset in milliseconds. Lower
6697 values have higher priority.
6698 Note that the resulting timestamp is is only tracked with enough precision
6699 for 524,287ms (8m44s287ms). If the request is queued long enough to where the
6700 adjusted timestamp exceeds this value, it will be misidentified as highest
6701 priority. Thus it is important to set "timeout queue" to a value, where when
6702 combined with the offset, does not exceed this limit.
Christopher Faulet76c09ef2017-09-21 11:03:52 +02006703
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006704http-request set-query <fmt> [ { if | unless } <condition> ]
Willy Tarreau20b0de52012-12-24 15:45:22 +01006705
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006706 This rewrites the request's query string which appears after the first
6707 question mark ("?") with the result of the evaluation of format string <fmt>.
6708 The part prior to the question mark is left intact. If the request doesn't
6709 contain a question mark and the new value is not empty, then one is added at
6710 the end of the URI, followed by the new value. If a question mark was
6711 present, it will never be removed even if the value is empty. This can be
6712 used to add or remove parameters from the query string.
Ruoshan Huangeb5a3632015-12-08 21:00:23 +08006713
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006714 See also "http-request set-query" and "http-request set-uri".
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01006715
6716 Example:
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006717 # replace "%3D" with "=" in the query string
6718 http-request set-query %[query,regsub(%3D,=,g)]
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01006719
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006720http-request set-src <expr> [ { if | unless } <condition> ]
6721 This is used to set the source IP address to the value of specified
6722 expression. Useful when a proxy in front of HAProxy rewrites source IP, but
6723 provides the correct IP in a HTTP header; or you want to mask source IP for
Olivier Doucet56e31202020-04-21 09:32:56 +02006724 privacy. All subsequent calls to "src" fetch will return this value
6725 (see example).
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006726
6727 Arguments :
6728 <expr> Is a standard HAProxy expression formed by a sample-fetch followed
6729 by some converters.
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01006730
Olivier Doucet56e31202020-04-21 09:32:56 +02006731 See also "option forwardfor".
6732
Cyril Bonté78caf842010-03-10 22:41:43 +01006733 Example:
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006734 http-request set-src hdr(x-forwarded-for)
6735 http-request set-src src,ipmask(24)
6736
Olivier Doucet56e31202020-04-21 09:32:56 +02006737 # After the masking this will track connections
6738 # based on the IP address with the last byte zeroed out.
6739 http-request track-sc0 src
6740
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006741 When possible, set-src preserves the original source port as long as the
6742 address family allows it, otherwise the source port is set to 0.
6743
6744http-request set-src-port <expr> [ { if | unless } <condition> ]
6745
6746 This is used to set the source port address to the value of specified
6747 expression.
6748
6749 Arguments:
6750 <expr> Is a standard HAProxy expression formed by a sample-fetch followed
6751 by some converters.
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01006752
Willy Tarreau20b0de52012-12-24 15:45:22 +01006753 Example:
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006754 http-request set-src-port hdr(x-port)
6755 http-request set-src-port int(4000)
6756
6757 When possible, set-src-port preserves the original source address as long as
6758 the address family supports a port, otherwise it forces the source address to
6759 IPv4 "0.0.0.0" before rewriting the port.
6760
Alex59c53352021-04-27 12:57:07 +02006761http-request set-timeout { server | tunnel } { <timeout> | <expr> }
Amaury Denoyelle8d228232020-12-10 13:43:54 +01006762 [ { if | unless } <condition> ]
6763
6764 This action overrides the specified "server" or "tunnel" timeout for the
6765 current stream only. The timeout can be specified in millisecond or with any
6766 other unit if the number is suffixed by the unit as explained at the top of
6767 this document. It is also possible to write an expression which must returns
6768 a number interpreted as a timeout in millisecond.
6769
6770 Note that the server/tunnel timeouts are only relevant on the backend side
6771 and thus this rule is only available for the proxies with backend
6772 capabilities. Also the timeout value must be non-null to obtain the expected
6773 results.
6774
6775 Example:
Alex59c53352021-04-27 12:57:07 +02006776 http-request set-timeout tunnel 5s
6777 http-request set-timeout server req.hdr(host),map_int(host.lst)
Amaury Denoyelle8d228232020-12-10 13:43:54 +01006778
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006779http-request set-tos <tos> [ { if | unless } <condition> ]
6780
6781 This is used to set the TOS or DSCP field value of packets sent to the client
6782 to the value passed in <tos> on platforms which support this. This value
6783 represents the whole 8 bits of the IP TOS field, and can be expressed both in
6784 decimal or hexadecimal format (prefixed by "0x"). Note that only the 6 higher
6785 bits are used in DSCP or TOS, and the two lower bits are always 0. This can
6786 be used to adjust some routing behavior on border routers based on some
6787 information from the request.
6788
6789 See RFC 2474, 2597, 3260 and 4594 for more information.
6790
6791http-request set-uri <fmt> [ { if | unless } <condition> ]
6792
6793 This rewrites the request URI with the result of the evaluation of format
6794 string <fmt>. The scheme, authority, path and query string are all replaced
6795 at once. This can be used to rewrite hosts in front of proxies, or to
6796 perform complex modifications to the URI such as moving parts between the
6797 path and the query string.
6798 See also "http-request set-path" and "http-request set-query".
6799
6800http-request set-var(<var-name>) <expr> [ { if | unless } <condition> ]
6801
6802 This is used to set the contents of a variable. The variable is declared
6803 inline.
6804
6805 Arguments:
6806 <var-name> The name of the variable starts with an indication about its
6807 scope. The scopes allowed are:
6808 "proc" : the variable is shared with the whole process
6809 "sess" : the variable is shared with the whole session
6810 "txn" : the variable is shared with the transaction
6811 (request and response)
6812 "req" : the variable is shared only during request
6813 processing
6814 "res" : the variable is shared only during response
6815 processing
6816 This prefix is followed by a name. The separator is a '.'.
6817 The name may only contain characters 'a-z', 'A-Z', '0-9'
6818 and '_'.
6819
6820 <expr> Is a standard HAProxy expression formed by a sample-fetch
6821 followed by some converters.
Willy Tarreau20b0de52012-12-24 15:45:22 +01006822
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02006823 Example:
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006824 http-request set-var(req.my_var) req.fhdr(user-agent),lower
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02006825
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006826http-request send-spoe-group <engine-name> <group-name>
6827 [ { if | unless } <condition> ]
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02006828
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006829 This action is used to trigger sending of a group of SPOE messages. To do so,
6830 the SPOE engine used to send messages must be defined, as well as the SPOE
6831 group to send. Of course, the SPOE engine must refer to an existing SPOE
6832 filter. If not engine name is provided on the SPOE filter line, the SPOE
6833 agent name must be used.
6834
6835 Arguments:
6836 <engine-name> The SPOE engine name.
6837
6838 <group-name> The SPOE group name as specified in the engine
6839 configuration.
6840
6841http-request silent-drop [ { if | unless } <condition> ]
6842
6843 This stops the evaluation of the rules and makes the client-facing connection
6844 suddenly disappear using a system-dependent way that tries to prevent the
6845 client from being notified. The effect it then that the client still sees an
6846 established connection while there's none on HAProxy. The purpose is to
6847 achieve a comparable effect to "tarpit" except that it doesn't use any local
6848 resource at all on the machine running HAProxy. It can resist much higher
6849 loads than "tarpit", and slow down stronger attackers. It is important to
6850 understand the impact of using this mechanism. All stateful equipment placed
6851 between the client and HAProxy (firewalls, proxies, load balancers) will also
6852 keep the established connection for a long time and may suffer from this
6853 action.
6854 On modern Linux systems running with enough privileges, the TCP_REPAIR socket
6855 option is used to block the emission of a TCP reset. On other systems, the
6856 socket's TTL is reduced to 1 so that the TCP reset doesn't pass the first
6857 router, though it's still delivered to local networks. Do not use it unless
6858 you fully understand how it works.
6859
Christopher Faulet46f95542019-12-20 10:07:22 +01006860http-request strict-mode { on | off }
6861
6862 This enables or disables the strict rewriting mode for following rules. It
6863 does not affect rules declared before it and it is only applicable on rules
6864 performing a rewrite on the requests. When the strict mode is enabled, any
6865 rewrite failure triggers an internal error. Otherwise, such errors are
6866 silently ignored. The purpose of the strict rewriting mode is to make some
Ilya Shipitsin8525fd92020-02-29 12:34:59 +05006867 rewrites optional while others must be performed to continue the request
Christopher Faulet46f95542019-12-20 10:07:22 +01006868 processing.
6869
Christopher Faulet1aea50e2020-01-17 16:03:53 +01006870 By default, the strict rewriting mode is enabled. Its value is also reset
Christopher Faulet46f95542019-12-20 10:07:22 +01006871 when a ruleset evaluation ends. So, for instance, if you change the mode on
6872 the frontend, the default mode is restored when HAProxy starts the backend
6873 rules evaluation.
6874
Christopher Faulet5cb513a2020-05-13 17:56:56 +02006875http-request tarpit [deny_status <status>] [ { if | unless } <condition> ]
6876http-request tarpit [ { status | deny_status } <code>] [content-type <type>]
6877 [ { default-errorfiles | errorfile <file> | errorfiles <name> |
6878 file <file> | lf-file <file> | string <str> | lf-string <fmt> } ]
6879 [ hdr <name> <fmt> ]*
6880 [ { if | unless } <condition> ]
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006881
6882 This stops the evaluation of the rules and immediately blocks the request
6883 without responding for a delay specified by "timeout tarpit" or
6884 "timeout connect" if the former is not set. After that delay, if the client
Christopher Faulet5cb513a2020-05-13 17:56:56 +02006885 is still connected, a response is returned so that the client does not
6886 suspect it has been tarpitted. Logs will report the flags "PT". The goal of
6887 the tarpit rule is to slow down robots during an attack when they're limited
6888 on the number of concurrent requests. It can be very efficient against very
6889 dumb robots, and will significantly reduce the load on firewalls compared to
6890 a "deny" rule. But when facing "correctly" developed robots, it can make
Daniel Corbett9f0843f2021-05-08 10:50:37 -04006891 things worse by forcing HAProxy and the front firewall to support insane
Christopher Faulet5cb513a2020-05-13 17:56:56 +02006892 number of concurrent connections. By default an HTTP error 500 is returned.
6893 But the response may be customized using same syntax than
6894 "http-request return" rules. Thus, see "http-request return" for details.
Ilya Shipitsin11057a32020-06-21 21:18:27 +05006895 For compatibility purpose, when no argument is defined, or only "deny_status",
Christopher Faulet5cb513a2020-05-13 17:56:56 +02006896 the argument "default-errorfiles" is implied. It means
6897 "http-request tarpit [deny_status <status>]" is an alias of
6898 "http-request tarpit [status <status>] default-errorfiles".
6899 No further "http-request" rules are evaluated.
6900 See also "http-request return" and "http-request silent-drop".
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006901
6902http-request track-sc0 <key> [table <table>] [ { if | unless } <condition> ]
6903http-request track-sc1 <key> [table <table>] [ { if | unless } <condition> ]
6904http-request track-sc2 <key> [table <table>] [ { if | unless } <condition> ]
6905
6906 This enables tracking of sticky counters from current request. These rules do
6907 not stop evaluation and do not change default action. The number of counters
6908 that may be simultaneously tracked by the same connection is set in
6909 MAX_SESS_STKCTR at build time (reported in haproxy -vv) which defaults to 3,
Matteo Contrini1857b8c2020-10-16 17:35:54 +02006910 so the track-sc number is between 0 and (MAX_SESS_STKCTR-1). The first
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006911 "track-sc0" rule executed enables tracking of the counters of the specified
6912 table as the first set. The first "track-sc1" rule executed enables tracking
6913 of the counters of the specified table as the second set. The first
6914 "track-sc2" rule executed enables tracking of the counters of the specified
6915 table as the third set. It is a recommended practice to use the first set of
6916 counters for the per-frontend counters and the second set for the per-backend
6917 ones. But this is just a guideline, all may be used everywhere.
6918
6919 Arguments :
6920 <key> is mandatory, and is a sample expression rule as described in
6921 section 7.3. It describes what elements of the incoming request or
6922 connection will be analyzed, extracted, combined, and used to
6923 select which table entry to update the counters.
6924
6925 <table> is an optional table to be used instead of the default one, which
6926 is the stick-table declared in the current proxy. All the counters
6927 for the matches and updates for the key will then be performed in
6928 that table until the session ends.
6929
6930 Once a "track-sc*" rule is executed, the key is looked up in the table and if
6931 it is not found, an entry is allocated for it. Then a pointer to that entry
6932 is kept during all the session's life, and this entry's counters are updated
6933 as often as possible, every time the session's counters are updated, and also
6934 systematically when the session ends. Counters are only updated for events
6935 that happen after the tracking has been started. As an exception, connection
6936 counters and request counters are systematically updated so that they reflect
6937 useful information.
6938
6939 If the entry tracks concurrent connection counters, one connection is counted
6940 for as long as the entry is tracked, and the entry will not expire during
6941 that time. Tracking counters also provides a performance advantage over just
6942 checking the keys, because only one table lookup is performed for all ACL
6943 checks that make use of it.
6944
6945http-request unset-var(<var-name>) [ { if | unless } <condition> ]
6946
6947 This is used to unset a variable. See above for details about <var-name>.
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02006948
6949 Example:
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006950 http-request unset-var(req.my_var)
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02006951
Christopher Faulet579d83b2019-11-22 15:34:17 +01006952http-request use-service <service-name> [ { if | unless } <condition> ]
6953
6954 This directive executes the configured HTTP service to reply to the request
6955 and stops the evaluation of the rules. An HTTP service may choose to reply by
6956 sending any valid HTTP response or it may immediately close the connection
6957 without sending any response. Outside natives services, for instance the
6958 Prometheus exporter, it is possible to write your own services in Lua. No
6959 further "http-request" rules are evaluated.
6960
6961 Arguments :
6962 <service-name> is mandatory. It is the service to call
6963
6964 Example:
6965 http-request use-service prometheus-exporter if { path /metrics }
6966
Christopher Faulet021a8e42021-03-29 10:46:38 +02006967http-request wait-for-body time <time> [ at-least <bytes> ]
6968 [ { if | unless } <condition> ]
6969
6970 This will delay the processing of the request waiting for the payload for at
6971 most <time> milliseconds. if "at-least" argument is specified, HAProxy stops
6972 to wait the payload when the first <bytes> bytes are received. 0 means no
6973 limit, it is the default value. Regardless the "at-least" argument value,
6974 HAProxy stops to wait if the whole payload is received or if the request
6975 buffer is full. This action may be used as a replacement to "option
6976 http-buffer-request".
6977
6978 Arguments :
6979
6980 <time> is mandatory. It is the maximum time to wait for the body. It
6981 follows the HAProxy time format and is expressed in milliseconds.
6982
6983 <bytes> is optional. It is the minimum payload size to receive to stop to
Ilya Shipitsinb2be9a12021-04-24 13:25:42 +05006984 wait. It follows the HAProxy size format and is expressed in
Christopher Faulet021a8e42021-03-29 10:46:38 +02006985 bytes.
6986
6987 Example:
6988 http-request wait-for-body time 1s at-least 1k if METH_POST
6989
6990 See also : "option http-buffer-request"
6991
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006992http-request wait-for-handshake [ { if | unless } <condition> ]
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02006993
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006994 This will delay the processing of the request until the SSL handshake
6995 happened. This is mostly useful to delay processing early data until we're
6996 sure they are valid.
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02006997
Willy Tarreauef781042010-01-27 11:53:01 +01006998
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02006999http-response <action> <options...> [ { if | unless } <condition> ]
Willy Tarreaue365c0b2013-06-11 16:06:12 +02007000 Access control for Layer 7 responses
7001
7002 May be used in sections: defaults | frontend | listen | backend
7003 no | yes | yes | yes
7004
7005 The http-response statement defines a set of rules which apply to layer 7
7006 processing. The rules are evaluated in their declaration order when they are
7007 met in a frontend, listen or backend section. Any rule may optionally be
7008 followed by an ACL-based condition, in which case it will only be evaluated
7009 if the condition is true. Since these rules apply on responses, the backend
7010 rules are applied first, followed by the frontend's rules.
7011
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02007012 The first keyword is the rule's action. The supported actions are described
7013 below.
Willy Tarreaue365c0b2013-06-11 16:06:12 +02007014
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02007015 There is no limit to the number of http-response statements per instance.
Willy Tarreaue365c0b2013-06-11 16:06:12 +02007016
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02007017 Example:
7018 acl key_acl res.hdr(X-Acl-Key) -m found
Thierry FOURNIERdad3d1d2014-04-22 18:07:25 +02007019
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02007020 acl myhost hdr(Host) -f myhost.lst
Sasha Pachev218f0642014-06-16 12:05:59 -06007021
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02007022 http-response add-acl(myhost.lst) %[res.hdr(X-Acl-Key)] if key_acl
7023 http-response del-acl(myhost.lst) %[res.hdr(X-Acl-Key)] if key_acl
Sasha Pachev218f0642014-06-16 12:05:59 -06007024
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02007025 Example:
7026 acl value res.hdr(X-Value) -m found
Sasha Pachev218f0642014-06-16 12:05:59 -06007027
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02007028 use_backend bk_appli if { hdr(Host),map_str(map.lst) -m found }
Sasha Pachev218f0642014-06-16 12:05:59 -06007029
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02007030 http-response set-map(map.lst) %[src] %[res.hdr(X-Value)] if value
7031 http-response del-map(map.lst) %[src] if ! value
Sasha Pachev218f0642014-06-16 12:05:59 -06007032
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02007033 See also : "http-request", section 3.4 about userlists and section 7 about
7034 ACL usage.
Sasha Pachev218f0642014-06-16 12:05:59 -06007035
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02007036http-response add-acl(<file-name>) <key fmt> [ { if | unless } <condition> ]
Sasha Pachev218f0642014-06-16 12:05:59 -06007037
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02007038 This is used to add a new entry into an ACL. The ACL must be loaded from a
7039 file (even a dummy empty file). The file name of the ACL to be updated is
7040 passed between parentheses. It takes one argument: <key fmt>, which follows
7041 log-format rules, to collect content of the new entry. It performs a lookup
7042 in the ACL before insertion, to avoid duplicated (or more) values.
7043 This lookup is done by a linear search and can be expensive with large lists!
7044 It is the equivalent of the "add acl" command from the stats socket, but can
7045 be triggered by an HTTP response.
Sasha Pachev218f0642014-06-16 12:05:59 -06007046
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02007047http-response add-header <name> <fmt> [ { if | unless } <condition> ]
Sasha Pachev218f0642014-06-16 12:05:59 -06007048
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02007049 This appends an HTTP header field whose name is specified in <name> and whose
7050 value is defined by <fmt> which follows the log-format rules (see Custom Log
7051 Format in section 8.2.4). This may be used to send a cookie to a client for
7052 example, or to pass some internal information.
7053 This rule is not final, so it is possible to add other similar rules.
7054 Note that header addition is performed immediately, so one rule might reuse
7055 the resulting header from a previous rule.
Sasha Pachev218f0642014-06-16 12:05:59 -06007056
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02007057http-response allow [ { if | unless } <condition> ]
Sasha Pachev218f0642014-06-16 12:05:59 -06007058
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02007059 This stops the evaluation of the rules and lets the response pass the check.
7060 No further "http-response" rules are evaluated for the current section.
Sasha Pachev218f0642014-06-16 12:05:59 -06007061
Jarno Huuskonen251a6b72019-01-04 14:05:02 +02007062http-response cache-store <name> [ { if | unless } <condition> ]
Sasha Pachev218f0642014-06-16 12:05:59 -06007063
Christopher Faulet87f1f3d2019-07-18 14:51:20 +02007064 See section 6.2 about cache setup.
Sasha Pachev218f0642014-06-16 12:05:59 -06007065
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02007066http-response capture <sample> id <id> [ { if | unless } <condition> ]
Sasha Pachev218f0642014-06-16 12:05:59 -06007067
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02007068 This captures sample expression <sample> from the response buffer, and
7069 converts it to a string. The resulting string is stored into the next request
7070 "capture" slot, so it will possibly appear next to some captured HTTP
7071 headers. It will then automatically appear in the logs, and it will be
7072 possible to extract it using sample fetch rules to feed it into headers or
7073 anything. Please check section 7.3 (Fetching samples) and
7074 "capture response header" for more information.
Thierry FOURNIER35d70ef2015-08-26 16:21:56 +02007075
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02007076 The keyword "id" is the id of the capture slot which is used for storing the
7077 string. The capture slot must be defined in an associated frontend.
7078 This is useful to run captures in backends. The slot id can be declared by a
7079 previous directive "http-response capture" or with the "declare capture"
7080 keyword.
Baptiste Assmann19a69b32020-01-16 14:34:22 +01007081
7082 When using this action in a backend, double check that the relevant
7083 frontend(s) have the required capture slots otherwise, this rule will be
7084 ignored at run time. This can't be detected at configuration parsing time
7085 due to HAProxy's ability to dynamically resolve backend name at runtime.
Thierry FOURNIER35d70ef2015-08-26 16:21:56 +02007086
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02007087http-response del-acl(<file-name>) <key fmt> [ { if | unless } <condition> ]
Thierry FOURNIER35d70ef2015-08-26 16:21:56 +02007088
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02007089 This is used to delete an entry from an ACL. The ACL must be loaded from a
7090 file (even a dummy empty file). The file name of the ACL to be updated is
7091 passed between parentheses. It takes one argument: <key fmt>, which follows
7092 log-format rules, to collect content of the entry to delete.
7093 It is the equivalent of the "del acl" command from the stats socket, but can
7094 be triggered by an HTTP response.
Willy Tarreauf4c43c12013-06-11 17:01:13 +02007095
Maciej Zdebebdd4c52020-11-20 13:58:48 +00007096http-response del-header <name> [ -m <meth> ] [ { if | unless } <condition> ]
Willy Tarreau9a355ec2013-06-11 17:45:46 +02007097
Maciej Zdebebdd4c52020-11-20 13:58:48 +00007098 This removes all HTTP header fields whose name is specified in <name>. <meth>
7099 is the matching method, applied on the header name. Supported matching methods
7100 are "str" (exact match), "beg" (prefix match), "end" (suffix match), "sub"
7101 (substring match) and "reg" (regex match). If not specified, exact matching
7102 method is used.
Willy Tarreau42cf39e2013-06-11 18:51:32 +02007103
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02007104http-response del-map(<file-name>) <key fmt> [ { if | unless } <condition> ]
Willy Tarreau51347ed2013-06-11 19:34:13 +02007105
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02007106 This is used to delete an entry from a MAP. The MAP must be loaded from a
7107 file (even a dummy empty file). The file name of the MAP to be updated is
7108 passed between parentheses. It takes one argument: <key fmt>, which follows
7109 log-format rules, to collect content of the entry to delete.
7110 It takes one argument: "file name" It is the equivalent of the "del map"
7111 command from the stats socket, but can be triggered by an HTTP response.
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02007112
Christopher Faulet5cb513a2020-05-13 17:56:56 +02007113http-response deny [deny_status <status>] [ { if | unless } <condition> ]
7114http-response deny [ { status | deny_status } <code>] [content-type <type>]
7115 [ { default-errorfiles | errorfile <file> | errorfiles <name> |
7116 file <file> | lf-file <file> | string <str> | lf-string <fmt> } ]
7117 [ hdr <name> <fmt> ]*
7118 [ { if | unless } <condition> ]
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02007119
Christopher Faulet5cb513a2020-05-13 17:56:56 +02007120 This stops the evaluation of the rules and immediately rejects the response.
7121 By default an HTTP 502 error is returned. But the response may be customized
7122 using same syntax than "http-response return" rules. Thus, see
Ilya Shipitsin11057a32020-06-21 21:18:27 +05007123 "http-response return" for details. For compatibility purpose, when no
Christopher Faulet5cb513a2020-05-13 17:56:56 +02007124 argument is defined, or only "deny_status", the argument "default-errorfiles"
7125 is implied. It means "http-response deny [deny_status <status>]" is an alias
7126 of "http-response deny [status <status>] default-errorfiles".
Christopher Faulet040c8cd2020-01-13 16:43:45 +01007127 No further "http-response" rules are evaluated.
Christopher Faulet5cb513a2020-05-13 17:56:56 +02007128 See also "http-response return".
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02007129
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02007130http-response redirect <rule> [ { if | unless } <condition> ]
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02007131
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02007132 This performs an HTTP redirection based on a redirect rule.
7133 This supports a format string similarly to "http-request redirect" rules,
7134 with the exception that only the "location" type of redirect is possible on
7135 the response. See the "redirect" keyword for the rule's syntax. When a
7136 redirect rule is applied during a response, connections to the server are
7137 closed so that no data can be forwarded from the server to the client.
Thierry FOURNIERe80fada2015-05-26 18:06:31 +02007138
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02007139http-response replace-header <name> <regex-match> <replace-fmt>
7140 [ { if | unless } <condition> ]
Thierry FOURNIERe80fada2015-05-26 18:06:31 +02007141
Tim Duesterhus2252beb2019-10-29 00:05:13 +01007142 This works like "http-request replace-header" except that it works on the
7143 server's response instead of the client's request.
William Lallemand86d0df02017-11-24 21:36:45 +01007144
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02007145 Example:
7146 http-response replace-header Set-Cookie (C=[^;]*);(.*) \1;ip=%bi;\2
Willy Tarreau51d861a2015-05-22 17:30:48 +02007147
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02007148 # applied to:
7149 Set-Cookie: C=1; expires=Tue, 14-Jun-2016 01:40:45 GMT
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02007150
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02007151 # outputs:
7152 Set-Cookie: C=1;ip=192.168.1.20; expires=Tue, 14-Jun-2016 01:40:45 GMT
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02007153
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02007154 # assuming the backend IP is 192.168.1.20.
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02007155
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02007156http-response replace-value <name> <regex-match> <replace-fmt>
7157 [ { if | unless } <condition> ]
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02007158
Tim Duesterhus6bd909b2020-01-17 15:53:18 +01007159 This works like "http-request replace-value" except that it works on the
Tim Duesterhus2252beb2019-10-29 00:05:13 +01007160 server's response instead of the client's request.
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02007161
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02007162 Example:
7163 http-response replace-value Cache-control ^public$ private
Christopher Faulet85d79c92016-11-09 16:54:56 +01007164
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02007165 # applied to:
7166 Cache-Control: max-age=3600, public
Christopher Faulet85d79c92016-11-09 16:54:56 +01007167
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02007168 # outputs:
7169 Cache-Control: max-age=3600, private
Christopher Faulet85d79c92016-11-09 16:54:56 +01007170
Christopher Faulet24231ab2020-01-24 17:44:23 +01007171http-response return [status <code>] [content-type <type>]
7172 [ { default-errorfiles | errorfile <file> | errorfiles <name> |
7173 file <file> | lf-file <file> | string <str> | lf-string <fmt> } ]
Christopher Faulet4a2c1422020-01-31 17:36:01 +01007174 [ hdr <name> <value> ]*
Christopher Faulet24231ab2020-01-24 17:44:23 +01007175 [ { if | unless } <condition> ]
7176
Ilya Shipitsin8525fd92020-02-29 12:34:59 +05007177 This stops the evaluation of the rules and immediately returns a response. The
Christopher Faulet24231ab2020-01-24 17:44:23 +01007178 default status code used for the response is 200. It can be optionally
7179 specified as an arguments to "status". The response content-type may also be
Daniel Corbett67a82712020-07-06 23:01:19 -04007180 specified as an argument to "content-type". Finally the response itself may
Christopher Faulet24231ab2020-01-24 17:44:23 +01007181 be defined. If can be a full HTTP response specifying the errorfile to use,
Ilya Shipitsin8525fd92020-02-29 12:34:59 +05007182 or the response payload specifying the file or the string to use. These rules
Christopher Faulet24231ab2020-01-24 17:44:23 +01007183 are followed to create the response :
7184
7185 * If neither the errorfile nor the payload to use is defined, a dummy
7186 response is returned. Only the "status" argument is considered. It can be
7187 any code in the range [200, 599]. The "content-type" argument, if any, is
7188 ignored.
7189
7190 * If "default-errorfiles" argument is set, the proxy's errorfiles are
7191 considered. If the "status" argument is defined, it must be one of the
Daniel Corbett9f0843f2021-05-08 10:50:37 -04007192 status code handled by HAProxy (200, 400, 403, 404, 405, 408, 410, 413,
Christopher Faulete095f312020-12-07 11:22:24 +01007193 425, 429, 500, 501, 502, 503, and 504). The "content-type" argument, if
7194 any, is ignored.
Christopher Faulet24231ab2020-01-24 17:44:23 +01007195
7196 * If a specific errorfile is defined, with an "errorfile" argument, the
7197 corresponding file, containing a full HTTP response, is returned. Only the
7198 "status" argument is considered. It must be one of the status code handled
Daniel Corbett9f0843f2021-05-08 10:50:37 -04007199 by HAProxy (200, 400, 403, 404, 405, 408, 410, 413, 425, 429, 500, 501,
Christopher Faulete095f312020-12-07 11:22:24 +01007200 502, 503, and 504). The "content-type" argument, if any, is ignored.
Christopher Faulet24231ab2020-01-24 17:44:23 +01007201
7202 * If an http-errors section is defined, with an "errorfiles" argument, the
7203 corresponding file in the specified http-errors section, containing a full
7204 HTTP response, is returned. Only the "status" argument is considered. It
Daniel Corbett9f0843f2021-05-08 10:50:37 -04007205 must be one of the status code handled by HAProxy (200, 400, 403, 404, 405,
Christopher Faulete095f312020-12-07 11:22:24 +01007206 408, 410, 413, 425, 429, 500, 501, 502, 503, and 504). The "content-type"
Anthonin Bonnefoy85048f82020-06-22 09:17:01 +02007207 argument, if any, is ignored.
Christopher Faulet24231ab2020-01-24 17:44:23 +01007208
7209 * If a "file" or a "lf-file" argument is specified, the file's content is
7210 used as the response payload. If the file is not empty, its content-type
7211 must be set as argument to "content-type". Otherwise, any "content-type"
7212 argument is ignored. With a "lf-file" argument, the file's content is
7213 evaluated as a log-format string. With a "file" argument, it is considered
7214 as a raw content.
7215
7216 * If a "string" or "lf-string" argument is specified, the defined string is
7217 used as the response payload. The content-type must always be set as
7218 argument to "content-type". With a "lf-string" argument, the string is
7219 evaluated as a log-format string. With a "string" argument, it is
7220 considered as a raw string.
7221
Christopher Faulet4a2c1422020-01-31 17:36:01 +01007222 When the response is not based an errorfile, it is possible to appends HTTP
7223 header fields to the response using "hdr" arguments. Otherwise, all "hdr"
7224 arguments are ignored. For each one, the header name is specified in <name>
7225 and its value is defined by <fmt> which follows the log-format rules.
7226
Christopher Faulet24231ab2020-01-24 17:44:23 +01007227 Note that the generated response must be smaller than a buffer. And to avoid
7228 any warning, when an errorfile or a raw file is loaded, the buffer space
Ilya Shipitsin11057a32020-06-21 21:18:27 +05007229 reserved to the headers rewriting should also be free.
Christopher Faulet24231ab2020-01-24 17:44:23 +01007230
7231 No further "http-response" rules are evaluated.
7232
7233 Example:
Daniel Corbett67a82712020-07-06 23:01:19 -04007234 http-response return errorfile /etc/haproxy/errorfiles/200.http \
Christopher Faulet24231ab2020-01-24 17:44:23 +01007235 if { status eq 404 }
7236
7237 http-response return content-type text/plain \
7238 string "This is the end !" \
7239 if { status eq 500 }
7240
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02007241http-response sc-inc-gpc0(<sc-id>) [ { if | unless } <condition> ]
7242http-response sc-inc-gpc1(<sc-id>) [ { if | unless } <condition> ]
Ruoshan Huange4edc6b2016-07-14 15:07:45 +08007243
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02007244 This action increments the GPC0 or GPC1 counter according with the sticky
7245 counter designated by <sc-id>. If an error occurs, this action silently fails
7246 and the actions evaluation continues.
Thierry FOURNIER236657b2015-08-19 08:25:14 +02007247
Cédric Dufour0d7712d2019-11-06 18:38:53 +01007248http-response sc-set-gpt0(<sc-id>) { <int> | <expr> }
7249 [ { if | unless } <condition> ]
Thierry FOURNIERe0627bd2015-08-04 08:20:33 +02007250
Cédric Dufour0d7712d2019-11-06 18:38:53 +01007251 This action sets the 32-bit unsigned GPT0 tag according to the sticky counter
7252 designated by <sc-id> and the value of <int>/<expr>. The expected result is a
7253 boolean. If an error occurs, this action silently fails and the actions
7254 evaluation continues.
Frédéric Lécaille6778b272018-01-29 15:22:53 +01007255
Christopher Faulet68fc3a12021-08-12 09:32:07 +02007256http-response send-spoe-group <engine-name> <group-name>
7257 [ { if | unless } <condition> ]
Willy Tarreau2d392c22015-08-24 01:43:45 +02007258
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02007259 This action is used to trigger sending of a group of SPOE messages. To do so,
7260 the SPOE engine used to send messages must be defined, as well as the SPOE
7261 group to send. Of course, the SPOE engine must refer to an existing SPOE
7262 filter. If not engine name is provided on the SPOE filter line, the SPOE
7263 agent name must be used.
Christopher Faulet76c09ef2017-09-21 11:03:52 +02007264
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02007265 Arguments:
7266 <engine-name> The SPOE engine name.
Christopher Faulet76c09ef2017-09-21 11:03:52 +02007267
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02007268 <group-name> The SPOE group name as specified in the engine
7269 configuration.
Christopher Faulet76c09ef2017-09-21 11:03:52 +02007270
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02007271http-response set-header <name> <fmt> [ { if | unless } <condition> ]
Willy Tarreaue365c0b2013-06-11 16:06:12 +02007272
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02007273 This does the same as "add-header" except that the header name is first
7274 removed if it existed. This is useful when passing security information to
7275 the server, where the header must not be manipulated by external users.
Willy Tarreaue365c0b2013-06-11 16:06:12 +02007276
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02007277http-response set-log-level <level> [ { if | unless } <condition> ]
7278
7279 This is used to change the log level of the current request when a certain
7280 condition is met. Valid levels are the 8 syslog levels (see the "log"
7281 keyword) plus the special level "silent" which disables logging for this
7282 request. This rule is not final so the last matching rule wins. This rule can
7283 be useful to disable health checks coming from another equipment.
7284
7285http-response set-map(<file-name>) <key fmt> <value fmt>
7286
7287 This is used to add a new entry into a MAP. The MAP must be loaded from a
7288 file (even a dummy empty file). The file name of the MAP to be updated is
7289 passed between parentheses. It takes 2 arguments: <key fmt>, which follows
7290 log-format rules, used to collect MAP key, and <value fmt>, which follows
7291 log-format rules, used to collect content for the new entry. It performs a
7292 lookup in the MAP before insertion, to avoid duplicated (or more) values.
7293 This lookup is done by a linear search and can be expensive with large lists!
7294 It is the equivalent of the "set map" command from the stats socket, but can
7295 be triggered by an HTTP response.
7296
7297http-response set-mark <mark> [ { if | unless } <condition> ]
7298
7299 This is used to set the Netfilter MARK on all packets sent to the client to
7300 the value passed in <mark> on platforms which support it. This value is an
7301 unsigned 32 bit value which can be matched by netfilter and by the routing
7302 table. It can be expressed both in decimal or hexadecimal format (prefixed
7303 by "0x"). This can be useful to force certain packets to take a different
7304 route (for example a cheaper network path for bulk downloads). This works on
7305 Linux kernels 2.6.32 and above and requires admin privileges.
7306
7307http-response set-nice <nice> [ { if | unless } <condition> ]
7308
7309 This sets the "nice" factor of the current request being processed.
7310 It only has effect against the other requests being processed at the same
7311 time. The default value is 0, unless altered by the "nice" setting on the
7312 "bind" line. The accepted range is -1024..1024. The higher the value, the
7313 nicest the request will be. Lower values will make the request more important
7314 than other ones. This can be useful to improve the speed of some requests, or
7315 lower the priority of non-important requests. Using this setting without
7316 prior experimentation can cause some major slowdown.
7317
7318http-response set-status <status> [reason <str>]
7319 [ { if | unless } <condition> ]
7320
7321 This replaces the response status code with <status> which must be an integer
7322 between 100 and 999. Optionally, a custom reason text can be provided defined
7323 by <str>, or the default reason for the specified code will be used as a
7324 fallback.
Ruoshan Huangeb5a3632015-12-08 21:00:23 +08007325
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02007326 Example:
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02007327 # return "431 Request Header Fields Too Large"
7328 http-response set-status 431
7329 # return "503 Slow Down", custom reason
7330 http-response set-status 503 reason "Slow Down".
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02007331
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02007332http-response set-tos <tos> [ { if | unless } <condition> ]
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02007333
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02007334 This is used to set the TOS or DSCP field value of packets sent to the client
7335 to the value passed in <tos> on platforms which support this.
7336 This value represents the whole 8 bits of the IP TOS field, and can be
7337 expressed both in decimal or hexadecimal format (prefixed by "0x"). Note that
7338 only the 6 higher bits are used in DSCP or TOS, and the two lower bits are
7339 always 0. This can be used to adjust some routing behavior on border routers
7340 based on some information from the request.
7341
7342 See RFC 2474, 2597, 3260 and 4594 for more information.
7343
7344http-response set-var(<var-name>) <expr> [ { if | unless } <condition> ]
7345
7346 This is used to set the contents of a variable. The variable is declared
7347 inline.
7348
7349 Arguments:
7350 <var-name> The name of the variable starts with an indication about its
7351 scope. The scopes allowed are:
7352 "proc" : the variable is shared with the whole process
7353 "sess" : the variable is shared with the whole session
7354 "txn" : the variable is shared with the transaction
7355 (request and response)
7356 "req" : the variable is shared only during request
7357 processing
7358 "res" : the variable is shared only during response
7359 processing
7360 This prefix is followed by a name. The separator is a '.'.
7361 The name may only contain characters 'a-z', 'A-Z', '0-9', '.'
7362 and '_'.
7363
7364 <expr> Is a standard HAProxy expression formed by a sample-fetch
7365 followed by some converters.
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02007366
7367 Example:
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02007368 http-response set-var(sess.last_redir) res.hdr(location)
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02007369
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02007370http-response silent-drop [ { if | unless } <condition> ]
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02007371
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02007372 This stops the evaluation of the rules and makes the client-facing connection
7373 suddenly disappear using a system-dependent way that tries to prevent the
7374 client from being notified. The effect it then that the client still sees an
7375 established connection while there's none on HAProxy. The purpose is to
7376 achieve a comparable effect to "tarpit" except that it doesn't use any local
7377 resource at all on the machine running HAProxy. It can resist much higher
7378 loads than "tarpit", and slow down stronger attackers. It is important to
7379 understand the impact of using this mechanism. All stateful equipment placed
7380 between the client and HAProxy (firewalls, proxies, load balancers) will also
7381 keep the established connection for a long time and may suffer from this
7382 action.
7383 On modern Linux systems running with enough privileges, the TCP_REPAIR socket
7384 option is used to block the emission of a TCP reset. On other systems, the
7385 socket's TTL is reduced to 1 so that the TCP reset doesn't pass the first
7386 router, though it's still delivered to local networks. Do not use it unless
7387 you fully understand how it works.
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02007388
Christopher Faulet46f95542019-12-20 10:07:22 +01007389http-response strict-mode { on | off }
7390
7391 This enables or disables the strict rewriting mode for following rules. It
7392 does not affect rules declared before it and it is only applicable on rules
7393 performing a rewrite on the responses. When the strict mode is enabled, any
7394 rewrite failure triggers an internal error. Otherwise, such errors are
7395 silently ignored. The purpose of the strict rewriting mode is to make some
Ilya Shipitsin8525fd92020-02-29 12:34:59 +05007396 rewrites optional while others must be performed to continue the response
Christopher Faulet46f95542019-12-20 10:07:22 +01007397 processing.
7398
Christopher Faulet1aea50e2020-01-17 16:03:53 +01007399 By default, the strict rewriting mode is enabled. Its value is also reset
Christopher Faulet46f95542019-12-20 10:07:22 +01007400 when a ruleset evaluation ends. So, for instance, if you change the mode on
Daniel Corbett67a82712020-07-06 23:01:19 -04007401 the backend, the default mode is restored when HAProxy starts the frontend
Christopher Faulet46f95542019-12-20 10:07:22 +01007402 rules evaluation.
7403
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02007404http-response track-sc0 <key> [table <table>] [ { if | unless } <condition> ]
7405http-response track-sc1 <key> [table <table>] [ { if | unless } <condition> ]
7406http-response track-sc2 <key> [table <table>] [ { if | unless } <condition> ]
Willy Tarreaue365c0b2013-06-11 16:06:12 +02007407
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02007408 This enables tracking of sticky counters from current response. Please refer
7409 to "http-request track-sc" for a complete description. The only difference
7410 from "http-request track-sc" is the <key> sample expression can only make use
7411 of samples in response (e.g. res.*, status etc.) and samples below Layer 6
7412 (e.g. SSL-related samples, see section 7.3.4). If the sample is not
Daniel Corbett9f0843f2021-05-08 10:50:37 -04007413 supported, HAProxy will fail and warn while parsing the config.
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02007414
7415http-response unset-var(<var-name>) [ { if | unless } <condition> ]
7416
7417 This is used to unset a variable. See "http-response set-var" for details
7418 about <var-name>.
7419
7420 Example:
7421 http-response unset-var(sess.last_redir)
7422
Christopher Faulet021a8e42021-03-29 10:46:38 +02007423http-response wait-for-body time <time> [ at-least <bytes> ]
7424 [ { if | unless } <condition> ]
7425
7426 This will delay the processing of the response waiting for the payload for at
7427 most <time> milliseconds. if "at-least" argument is specified, HAProxy stops
7428 to wait the payload when the first <bytes> bytes are received. 0 means no
7429 limit, it is the default value. Regardless the "at-least" argument value,
7430 HAProxy stops to wait if the whole payload is received or if the response
7431 buffer is full.
7432
7433 Arguments :
7434
7435 <time> is mandatory. It is the maximum time to wait for the body. It
7436 follows the HAProxy time format and is expressed in milliseconds.
7437
7438 <bytes> is optional. It is the minimum payload size to receive to stop to
Ilya Shipitsinb2be9a12021-04-24 13:25:42 +05007439 wait. It follows the HAProxy size format and is expressed in
Christopher Faulet021a8e42021-03-29 10:46:38 +02007440 bytes.
7441
7442 Example:
7443 http-response wait-for-body time 1s at-least 10k
Baptiste Assmann5ecb77f2013-10-06 23:24:13 +02007444
Willy Tarreau30631952015-08-06 15:05:24 +02007445http-reuse { never | safe | aggressive | always }
7446 Declare how idle HTTP connections may be shared between requests
7447
7448 May be used in sections: defaults | frontend | listen | backend
7449 yes | no | yes | yes
7450
Daniel Corbett9f0843f2021-05-08 10:50:37 -04007451 By default, a connection established between HAProxy and the backend server
Olivier Houchard86006a52018-12-14 19:37:49 +01007452 which is considered safe for reuse is moved back to the server's idle
7453 connections pool so that any other request can make use of it. This is the
7454 "safe" strategy below.
Willy Tarreau30631952015-08-06 15:05:24 +02007455
7456 The argument indicates the desired connection reuse strategy :
7457
Olivier Houchard86006a52018-12-14 19:37:49 +01007458 - "never" : idle connections are never shared between sessions. This mode
7459 may be enforced to cancel a different strategy inherited from
7460 a defaults section or for troubleshooting. For example, if an
7461 old bogus application considers that multiple requests over
7462 the same connection come from the same client and it is not
7463 possible to fix the application, it may be desirable to
7464 disable connection sharing in a single backend. An example of
Daniel Corbett9f0843f2021-05-08 10:50:37 -04007465 such an application could be an old HAProxy using cookie
Olivier Houchard86006a52018-12-14 19:37:49 +01007466 insertion in tunnel mode and not checking any request past the
7467 first one.
Willy Tarreau30631952015-08-06 15:05:24 +02007468
Olivier Houchard86006a52018-12-14 19:37:49 +01007469 - "safe" : this is the default and the recommended strategy. The first
7470 request of a session is always sent over its own connection,
7471 and only subsequent requests may be dispatched over other
7472 existing connections. This ensures that in case the server
7473 closes the connection when the request is being sent, the
7474 browser can decide to silently retry it. Since it is exactly
7475 equivalent to regular keep-alive, there should be no side
Amaury Denoyelle27179652020-10-14 18:17:12 +02007476 effects. There is also a special handling for the connections
7477 using protocols subject to Head-of-line blocking (backend with
7478 h2 or fcgi). In this case, when at least one stream is
7479 processed, the used connection is reserved to handle streams
7480 of the same session. When no more streams are processed, the
7481 connection is released and can be reused.
Willy Tarreau30631952015-08-06 15:05:24 +02007482
7483 - "aggressive" : this mode may be useful in webservices environments where
7484 all servers are not necessarily known and where it would be
7485 appreciable to deliver most first requests over existing
7486 connections. In this case, first requests are only delivered
7487 over existing connections that have been reused at least once,
7488 proving that the server correctly supports connection reuse.
7489 It should only be used when it's sure that the client can
7490 retry a failed request once in a while and where the benefit
Michael Prokop4438c602019-05-24 10:25:45 +02007491 of aggressive connection reuse significantly outweighs the
Willy Tarreau30631952015-08-06 15:05:24 +02007492 downsides of rare connection failures.
7493
7494 - "always" : this mode is only recommended when the path to the server is
7495 known for never breaking existing connections quickly after
7496 releasing them. It allows the first request of a session to be
7497 sent to an existing connection. This can provide a significant
7498 performance increase over the "safe" strategy when the backend
7499 is a cache farm, since such components tend to show a
Davor Ocelice9ed2812017-12-25 17:49:28 +01007500 consistent behavior and will benefit from the connection
Willy Tarreau30631952015-08-06 15:05:24 +02007501 sharing. It is recommended that the "http-keep-alive" timeout
7502 remains low in this mode so that no dead connections remain
7503 usable. In most cases, this will lead to the same performance
7504 gains as "aggressive" but with more risks. It should only be
7505 used when it improves the situation over "aggressive".
7506
7507 When http connection sharing is enabled, a great care is taken to respect the
Amaury Denoyelled773a4e2021-01-29 15:18:49 +01007508 connection properties and compatibility. Indeed, some properties are specific
7509 and it is not possibly to reuse it blindly. Those are the SSL SNI, source
7510 and destination address and proxy protocol block. A connection is reused only
7511 if it shares the same set of properties with the request.
Willy Tarreau30631952015-08-06 15:05:24 +02007512
Amaury Denoyelled773a4e2021-01-29 15:18:49 +01007513 Also note that connections with certain bogus authentication schemes (relying
7514 on the connection) like NTLM are marked private and never shared.
Willy Tarreau30631952015-08-06 15:05:24 +02007515
Lukas Tribuse8adfeb2019-11-06 11:50:25 +01007516 A connection pool is involved and configurable with "pool-max-conn".
Willy Tarreau30631952015-08-06 15:05:24 +02007517
7518 Note: connection reuse improves the accuracy of the "server maxconn" setting,
7519 because almost no new connection will be established while idle connections
7520 remain available. This is particularly true with the "always" strategy.
7521
7522 See also : "option http-keep-alive", "server maxconn"
7523
7524
Mark Lamourinec2247f02012-01-04 13:02:01 -05007525http-send-name-header [<header>]
7526 Add the server name to a request. Use the header string given by <header>
Mark Lamourinec2247f02012-01-04 13:02:01 -05007527 May be used in sections: defaults | frontend | listen | backend
7528 yes | no | yes | yes
Mark Lamourinec2247f02012-01-04 13:02:01 -05007529 Arguments :
Mark Lamourinec2247f02012-01-04 13:02:01 -05007530 <header> The header string to use to send the server name
7531
Willy Tarreau81bef7e2019-10-07 14:58:02 +02007532 The "http-send-name-header" statement causes the header field named <header>
7533 to be set to the name of the target server at the moment the request is about
7534 to be sent on the wire. Any existing occurrences of this header are removed.
7535 Upon retries and redispatches, the header field is updated to always reflect
7536 the server being attempted to connect to. Given that this header is modified
7537 very late in the connection setup, it may have unexpected effects on already
7538 modified headers. For example using it with transport-level header such as
7539 connection, content-length, transfer-encoding and so on will likely result in
7540 invalid requests being sent to the server. Additionally it has been reported
7541 that this directive is currently being used as a way to overwrite the Host
7542 header field in outgoing requests; while this trick has been known to work
7543 as a side effect of the feature for some time, it is not officially supported
7544 and might possibly not work anymore in a future version depending on the
7545 technical difficulties this feature induces. A long-term solution instead
7546 consists in fixing the application which required this trick so that it binds
7547 to the correct host name.
Mark Lamourinec2247f02012-01-04 13:02:01 -05007548
7549 See also : "server"
7550
Krzysztof Piotr Oledzkif58a9622008-02-23 01:19:10 +01007551id <value>
Willy Tarreau53fb4ae2009-10-04 23:04:08 +02007552 Set a persistent ID to a proxy.
7553 May be used in sections : defaults | frontend | listen | backend
7554 no | yes | yes | yes
7555 Arguments : none
7556
7557 Set a persistent ID for the proxy. This ID must be unique and positive.
7558 An unused ID will automatically be assigned if unset. The first assigned
7559 value will be 1. This ID is currently only returned in statistics.
Krzysztof Piotr Oledzkif58a9622008-02-23 01:19:10 +01007560
7561
Cyril Bonté0d4bf012010-04-25 23:21:46 +02007562ignore-persist { if | unless } <condition>
7563 Declare a condition to ignore persistence
7564 May be used in sections: defaults | frontend | listen | backend
Cyril Bonté4288c5a2018-03-12 22:02:59 +01007565 no | no | yes | yes
Cyril Bonté0d4bf012010-04-25 23:21:46 +02007566
7567 By default, when cookie persistence is enabled, every requests containing
7568 the cookie are unconditionally persistent (assuming the target server is up
7569 and running).
7570
7571 The "ignore-persist" statement allows one to declare various ACL-based
7572 conditions which, when met, will cause a request to ignore persistence.
7573 This is sometimes useful to load balance requests for static files, which
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03007574 often don't require persistence. This can also be used to fully disable
Cyril Bonté0d4bf012010-04-25 23:21:46 +02007575 persistence for a specific User-Agent (for example, some web crawler bots).
7576
Cyril Bonté0d4bf012010-04-25 23:21:46 +02007577 The persistence is ignored when an "if" condition is met, or unless an
7578 "unless" condition is met.
7579
Jarno Huuskonene5ae7022017-04-03 14:36:21 +03007580 Example:
7581 acl url_static path_beg /static /images /img /css
7582 acl url_static path_end .gif .png .jpg .css .js
7583 ignore-persist if url_static
7584
Cyril Bonté0d4bf012010-04-25 23:21:46 +02007585 See also : "force-persist", "cookie", and section 7 about ACL usage.
7586
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02007587load-server-state-from-file { global | local | none }
7588 Allow seamless reload of HAProxy
7589 May be used in sections: defaults | frontend | listen | backend
7590 yes | no | yes | yes
7591
7592 This directive points HAProxy to a file where server state from previous
7593 running process has been saved. That way, when starting up, before handling
7594 traffic, the new process can apply old states to servers exactly has if no
Davor Ocelice9ed2812017-12-25 17:49:28 +01007595 reload occurred. The purpose of the "load-server-state-from-file" directive is
Daniel Corbett9f0843f2021-05-08 10:50:37 -04007596 to tell HAProxy which file to use. For now, only 2 arguments to either prevent
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02007597 loading state or load states from a file containing all backends and servers.
7598 The state file can be generated by running the command "show servers state"
7599 over the stats socket and redirect output.
7600
Davor Ocelice9ed2812017-12-25 17:49:28 +01007601 The format of the file is versioned and is very specific. To understand it,
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02007602 please read the documentation of the "show servers state" command (chapter
Willy Tarreau1af20c72017-06-23 16:01:14 +02007603 9.3 of Management Guide).
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02007604
7605 Arguments:
7606 global load the content of the file pointed by the global directive
7607 named "server-state-file".
7608
7609 local load the content of the file pointed by the directive
7610 "server-state-file-name" if set. If not set, then the backend
7611 name is used as a file name.
7612
7613 none don't load any stat for this backend
7614
7615 Notes:
Willy Tarreaue5a60682016-11-09 14:54:53 +01007616 - server's IP address is preserved across reloads by default, but the
7617 order can be changed thanks to the server's "init-addr" setting. This
7618 means that an IP address change performed on the CLI at run time will
Davor Ocelice9ed2812017-12-25 17:49:28 +01007619 be preserved, and that any change to the local resolver (e.g. /etc/hosts)
Willy Tarreaue5a60682016-11-09 14:54:53 +01007620 will possibly not have any effect if the state file is in use.
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02007621
7622 - server's weight is applied from previous running process unless it has
7623 has changed between previous and new configuration files.
7624
Olivier Doucetaa1ea8a2016-08-05 17:15:20 +02007625 Example: Minimal configuration
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02007626
Olivier Doucetaa1ea8a2016-08-05 17:15:20 +02007627 global
7628 stats socket /tmp/socket
7629 server-state-file /tmp/server_state
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02007630
Olivier Doucetaa1ea8a2016-08-05 17:15:20 +02007631 defaults
7632 load-server-state-from-file global
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02007633
Olivier Doucetaa1ea8a2016-08-05 17:15:20 +02007634 backend bk
7635 server s1 127.0.0.1:22 check weight 11
7636 server s2 127.0.0.1:22 check weight 12
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02007637
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02007638
7639 Then one can run :
7640
7641 socat /tmp/socket - <<< "show servers state" > /tmp/server_state
7642
7643 Content of the file /tmp/server_state would be like this:
7644
7645 1
7646 # <field names skipped for the doc example>
7647 1 bk 1 s1 127.0.0.1 2 0 11 11 4 6 3 4 6 0 0
7648 1 bk 2 s2 127.0.0.1 2 0 12 12 4 6 3 4 6 0 0
7649
Olivier Doucetaa1ea8a2016-08-05 17:15:20 +02007650 Example: Minimal configuration
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02007651
7652 global
7653 stats socket /tmp/socket
7654 server-state-base /etc/haproxy/states
7655
7656 defaults
7657 load-server-state-from-file local
7658
7659 backend bk
7660 server s1 127.0.0.1:22 check weight 11
7661 server s2 127.0.0.1:22 check weight 12
7662
Olivier Doucetaa1ea8a2016-08-05 17:15:20 +02007663
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02007664 Then one can run :
7665
7666 socat /tmp/socket - <<< "show servers state bk" > /etc/haproxy/states/bk
7667
7668 Content of the file /etc/haproxy/states/bk would be like this:
7669
7670 1
7671 # <field names skipped for the doc example>
7672 1 bk 1 s1 127.0.0.1 2 0 11 11 4 6 3 4 6 0 0
7673 1 bk 2 s2 127.0.0.1 2 0 12 12 4 6 3 4 6 0 0
7674
7675 See also: "server-state-file", "server-state-file-name", and
7676 "show servers state"
7677
Cyril Bonté0d4bf012010-04-25 23:21:46 +02007678
Willy Tarreau2769aa02007-12-27 18:26:09 +01007679log global
Jan Wagner3e678602020-12-17 22:22:32 +01007680log <address> [len <length>] [format <format>] [sample <ranges>:<sample_size>]
Frédéric Lécailled690dfa2019-04-25 10:52:17 +02007681 <facility> [<level> [<minlevel>]]
William Lallemand0f99e342011-10-12 17:50:54 +02007682no log
Willy Tarreau2769aa02007-12-27 18:26:09 +01007683 Enable per-instance logging of events and traffic.
7684 May be used in sections : defaults | frontend | listen | backend
7685 yes | yes | yes | yes
William Lallemand0f99e342011-10-12 17:50:54 +02007686
7687 Prefix :
7688 no should be used when the logger list must be flushed. For example,
7689 if you don't want to inherit from the default logger list. This
7690 prefix does not allow arguments.
7691
Willy Tarreau2769aa02007-12-27 18:26:09 +01007692 Arguments :
7693 global should be used when the instance's logging parameters are the
7694 same as the global ones. This is the most common usage. "global"
7695 replaces <address>, <facility> and <level> with those of the log
7696 entries found in the "global" section. Only one "log global"
7697 statement may be used per instance, and this form takes no other
7698 parameter.
7699
7700 <address> indicates where to send the logs. It takes the same format as
7701 for the "global" section's logs, and can be one of :
7702
7703 - An IPv4 address optionally followed by a colon (':') and a UDP
7704 port. If no port is specified, 514 is used by default (the
7705 standard syslog port).
7706
David du Colombier24bb5f52011-03-17 10:40:23 +01007707 - An IPv6 address followed by a colon (':') and optionally a UDP
7708 port. If no port is specified, 514 is used by default (the
7709 standard syslog port).
7710
Willy Tarreau2769aa02007-12-27 18:26:09 +01007711 - A filesystem path to a UNIX domain socket, keeping in mind
7712 considerations for chroot (be sure the path is accessible
7713 inside the chroot) and uid/gid (be sure the path is
Davor Ocelice9ed2812017-12-25 17:49:28 +01007714 appropriately writable).
Willy Tarreau2769aa02007-12-27 18:26:09 +01007715
Willy Tarreau5a32ecc2018-11-12 07:34:59 +01007716 - A file descriptor number in the form "fd@<number>", which may
7717 point to a pipe, terminal, or socket. In this case unbuffered
7718 logs are used and one writev() call per log is performed. This
7719 is a bit expensive but acceptable for most workloads. Messages
7720 sent this way will not be truncated but may be dropped, in
7721 which case the DroppedLogs counter will be incremented. The
7722 writev() call is atomic even on pipes for messages up to
7723 PIPE_BUF size, which POSIX recommends to be at least 512 and
7724 which is 4096 bytes on most modern operating systems. Any
7725 larger message may be interleaved with messages from other
7726 processes. Exceptionally for debugging purposes the file
7727 descriptor may also be directed to a file, but doing so will
Daniel Corbett9f0843f2021-05-08 10:50:37 -04007728 significantly slow HAProxy down as non-blocking calls will be
Willy Tarreau5a32ecc2018-11-12 07:34:59 +01007729 ignored. Also there will be no way to purge nor rotate this
7730 file without restarting the process. Note that the configured
7731 syslog format is preserved, so the output is suitable for use
Willy Tarreauc1b06452018-11-12 11:57:56 +01007732 with a TCP syslog server. See also the "short" and "raw"
7733 formats below.
Willy Tarreau5a32ecc2018-11-12 07:34:59 +01007734
7735 - "stdout" / "stderr", which are respectively aliases for "fd@1"
7736 and "fd@2", see above.
7737
Willy Tarreauc046d162019-08-30 15:24:59 +02007738 - A ring buffer in the form "ring@<name>", which will correspond
7739 to an in-memory ring buffer accessible over the CLI using the
7740 "show events" command, which will also list existing rings and
7741 their sizes. Such buffers are lost on reload or restart but
7742 when used as a complement this can help troubleshooting by
7743 having the logs instantly available.
7744
Emeric Brun94aab062021-04-02 10:41:36 +02007745 - An explicit stream address prefix such as "tcp@","tcp6@",
7746 "tcp4@" or "uxst@" will allocate an implicit ring buffer with
7747 a stream forward server targeting the given address.
7748
Willy Tarreau5a32ecc2018-11-12 07:34:59 +01007749 You may want to reference some environment variables in the
7750 address parameter, see section 2.3 about environment variables.
Willy Tarreaudad36a32013-03-11 01:20:04 +01007751
Willy Tarreau18324f52014-06-27 18:10:07 +02007752 <length> is an optional maximum line length. Log lines larger than this
7753 value will be truncated before being sent. The reason is that
7754 syslog servers act differently on log line length. All servers
7755 support the default value of 1024, but some servers simply drop
7756 larger lines while others do log them. If a server supports long
7757 lines, it may make sense to set this value here in order to avoid
7758 truncating long lines. Similarly, if a server drops long lines,
7759 it is preferable to truncate them before sending them. Accepted
7760 values are 80 to 65535 inclusive. The default value of 1024 is
7761 generally fine for all standard usages. Some specific cases of
Davor Ocelice9ed2812017-12-25 17:49:28 +01007762 long captures or JSON-formatted logs may require larger values.
Willy Tarreau18324f52014-06-27 18:10:07 +02007763
Frédéric Lécailled690dfa2019-04-25 10:52:17 +02007764 <ranges> A list of comma-separated ranges to identify the logs to sample.
7765 This is used to balance the load of the logs to send to the log
7766 server. The limits of the ranges cannot be null. They are numbered
7767 from 1. The size or period (in number of logs) of the sample must
7768 be set with <sample_size> parameter.
7769
7770 <sample_size>
7771 The size of the sample in number of logs to consider when balancing
7772 their logging loads. It is used to balance the load of the logs to
7773 send to the syslog server. This size must be greater or equal to the
7774 maximum of the high limits of the ranges.
7775 (see also <ranges> parameter).
7776
Willy Tarreauadb345d2018-11-12 07:56:13 +01007777 <format> is the log format used when generating syslog messages. It may be
7778 one of the following :
7779
Emeric Brun0237c4e2020-11-27 16:24:34 +01007780 local Analog to rfc3164 syslog message format except that hostname
7781 field is stripped. This is the default.
7782 Note: option "log-send-hostname" switches the default to
7783 rfc3164.
7784
7785 rfc3164 The RFC3164 syslog message format.
Willy Tarreauadb345d2018-11-12 07:56:13 +01007786 (https://tools.ietf.org/html/rfc3164)
7787
7788 rfc5424 The RFC5424 syslog message format.
7789 (https://tools.ietf.org/html/rfc5424)
7790
Emeric Brun54648852020-07-06 15:54:06 +02007791 priority A message containing only a level plus syslog facility between
7792 angle brackets such as '<63>', followed by the text. The PID,
7793 date, time, process name and system name are omitted. This is
7794 designed to be used with a local log server.
7795
Willy Tarreaue8746a02018-11-12 08:45:00 +01007796 short A message containing only a level between angle brackets such as
7797 '<3>', followed by the text. The PID, date, time, process name
7798 and system name are omitted. This is designed to be used with a
7799 local log server. This format is compatible with what the
7800 systemd logger consumes.
7801
Emeric Brun54648852020-07-06 15:54:06 +02007802 timed A message containing only a level between angle brackets such as
7803 '<3>', followed by ISO date and by the text. The PID, process
7804 name and system name are omitted. This is designed to be
7805 used with a local log server.
7806
7807 iso A message containing only the ISO date, followed by the text.
7808 The PID, process name and system name are omitted. This is
7809 designed to be used with a local log server.
7810
Willy Tarreauc1b06452018-11-12 11:57:56 +01007811 raw A message containing only the text. The level, PID, date, time,
7812 process name and system name are omitted. This is designed to
7813 be used in containers or during development, where the severity
7814 only depends on the file descriptor used (stdout/stderr).
7815
Willy Tarreau2769aa02007-12-27 18:26:09 +01007816 <facility> must be one of the 24 standard syslog facilities :
7817
Willy Tarreaue8746a02018-11-12 08:45:00 +01007818 kern user mail daemon auth syslog lpr news
7819 uucp cron auth2 ftp ntp audit alert cron2
7820 local0 local1 local2 local3 local4 local5 local6 local7
7821
Willy Tarreauc1b06452018-11-12 11:57:56 +01007822 Note that the facility is ignored for the "short" and "raw"
7823 formats, but still required as a positional field. It is
7824 recommended to use "daemon" in this case to make it clear that
7825 it's only supposed to be used locally.
Willy Tarreau2769aa02007-12-27 18:26:09 +01007826
7827 <level> is optional and can be specified to filter outgoing messages. By
7828 default, all messages are sent. If a level is specified, only
7829 messages with a severity at least as important as this level
Willy Tarreauf7edefa2009-05-10 17:20:05 +02007830 will be sent. An optional minimum level can be specified. If it
7831 is set, logs emitted with a more severe level than this one will
7832 be capped to this level. This is used to avoid sending "emerg"
7833 messages on all terminals on some default syslog configurations.
7834 Eight levels are known :
Willy Tarreau2769aa02007-12-27 18:26:09 +01007835
7836 emerg alert crit err warning notice info debug
7837
William Lallemand0f99e342011-10-12 17:50:54 +02007838 It is important to keep in mind that it is the frontend which decides what to
7839 log from a connection, and that in case of content switching, the log entries
7840 from the backend will be ignored. Connections are logged at level "info".
Willy Tarreaucc6c8912009-02-22 10:53:55 +01007841
7842 However, backend log declaration define how and where servers status changes
7843 will be logged. Level "notice" will be used to indicate a server going up,
7844 "warning" will be used for termination signals and definitive service
7845 termination, and "alert" will be used for when a server goes down.
7846
7847 Note : According to RFC3164, messages are truncated to 1024 bytes before
7848 being emitted.
Willy Tarreau2769aa02007-12-27 18:26:09 +01007849
7850 Example :
7851 log global
Willy Tarreauc1b06452018-11-12 11:57:56 +01007852 log stdout format short daemon # send log to systemd
7853 log stdout format raw daemon # send everything to stdout
7854 log stderr format raw daemon notice # send important events to stderr
Willy Tarreauf7edefa2009-05-10 17:20:05 +02007855 log 127.0.0.1:514 local0 notice # only send important events
Emeric Brun94aab062021-04-02 10:41:36 +02007856 log tcp@127.0.0.1:514 local0 notice notice # same but limit output
7857 # level and send in tcp
William Lallemandb2f07452015-05-12 14:27:13 +02007858 log "${LOCAL_SYSLOG}:514" local0 notice # send to local server
Willy Tarreaudad36a32013-03-11 01:20:04 +01007859
Willy Tarreau2769aa02007-12-27 18:26:09 +01007860
William Lallemand48940402012-01-30 16:47:22 +01007861log-format <string>
Willy Tarreaufb4e7ea2015-01-07 14:55:17 +01007862 Specifies the log format string to use for traffic logs
7863 May be used in sections: defaults | frontend | listen | backend
7864 yes | yes | yes | no
William Lallemand48940402012-01-30 16:47:22 +01007865
Willy Tarreaufb4e7ea2015-01-07 14:55:17 +01007866 This directive specifies the log format string that will be used for all logs
7867 resulting from traffic passing through the frontend using this line. If the
7868 directive is used in a defaults section, all subsequent frontends will use
7869 the same log format. Please see section 8.2.4 which covers the log format
7870 string in depth.
William Lallemand48940402012-01-30 16:47:22 +01007871
Guillaume de Lafond29f45602017-03-31 19:52:15 +02007872 "log-format" directive overrides previous "option tcplog", "log-format" and
7873 "option httplog" directives.
7874
Dragan Dosen7ad31542015-09-28 17:16:47 +02007875log-format-sd <string>
7876 Specifies the RFC5424 structured-data log format string
7877 May be used in sections: defaults | frontend | listen | backend
7878 yes | yes | yes | no
7879
7880 This directive specifies the RFC5424 structured-data log format string that
7881 will be used for all logs resulting from traffic passing through the frontend
7882 using this line. If the directive is used in a defaults section, all
7883 subsequent frontends will use the same log format. Please see section 8.2.4
7884 which covers the log format string in depth.
7885
7886 See https://tools.ietf.org/html/rfc5424#section-6.3 for more information
7887 about the RFC5424 structured-data part.
7888
7889 Note : This log format string will be used only for loggers that have set
7890 log format to "rfc5424".
7891
7892 Example :
7893 log-format-sd [exampleSDID@1234\ bytes=\"%B\"\ status=\"%ST\"]
7894
7895
Willy Tarreau094af4e2015-01-07 15:03:42 +01007896log-tag <string>
7897 Specifies the log tag to use for all outgoing logs
7898 May be used in sections: defaults | frontend | listen | backend
7899 yes | yes | yes | yes
7900
7901 Sets the tag field in the syslog header to this string. It defaults to the
7902 log-tag set in the global section, otherwise the program name as launched
Daniel Corbett9f0843f2021-05-08 10:50:37 -04007903 from the command line, which usually is "HAProxy". Sometimes it can be useful
Willy Tarreau094af4e2015-01-07 15:03:42 +01007904 to differentiate between multiple processes running on the same host, or to
7905 differentiate customer instances running in the same process. In the backend,
7906 logs about servers up/down will use this tag. As a hint, it can be convenient
7907 to set a log-tag related to a hosted customer in a defaults section then put
7908 all the frontends and backends for that customer, then start another customer
7909 in a new defaults section. See also the global "log-tag" directive.
Willy Tarreau2769aa02007-12-27 18:26:09 +01007910
Willy Tarreauc35362a2014-04-25 13:58:37 +02007911max-keep-alive-queue <value>
7912 Set the maximum server queue size for maintaining keep-alive connections
7913 May be used in sections: defaults | frontend | listen | backend
7914 yes | no | yes | yes
7915
7916 HTTP keep-alive tries to reuse the same server connection whenever possible,
7917 but sometimes it can be counter-productive, for example if a server has a lot
7918 of connections while other ones are idle. This is especially true for static
7919 servers.
7920
7921 The purpose of this setting is to set a threshold on the number of queued
Daniel Corbett9f0843f2021-05-08 10:50:37 -04007922 connections at which HAProxy stops trying to reuse the same server and prefers
Willy Tarreauc35362a2014-04-25 13:58:37 +02007923 to find another one. The default value, -1, means there is no limit. A value
7924 of zero means that keep-alive requests will never be queued. For very close
7925 servers which can be reached with a low latency and which are not sensible to
Davor Ocelice9ed2812017-12-25 17:49:28 +01007926 breaking keep-alive, a low value is recommended (e.g. local static server can
Willy Tarreauc35362a2014-04-25 13:58:37 +02007927 use a value of 10 or less). For remote servers suffering from a high latency,
7928 higher values might be needed to cover for the latency and/or the cost of
7929 picking a different server.
7930
7931 Note that this has no impact on responses which are maintained to the same
7932 server consecutively to a 401 response. They will still go to the same server
7933 even if they have to be queued.
7934
7935 See also : "option http-server-close", "option prefer-last-server", server
7936 "maxconn" and cookie persistence.
7937
Olivier Houcharda4d4fdf2018-12-14 19:27:06 +01007938max-session-srv-conns <nb>
7939 Set the maximum number of outgoing connections we can keep idling for a given
7940 client session. The default is 5 (it precisely equals MAX_SRV_LIST which is
7941 defined at build time).
Willy Tarreauc35362a2014-04-25 13:58:37 +02007942
Willy Tarreau2769aa02007-12-27 18:26:09 +01007943maxconn <conns>
7944 Fix the maximum number of concurrent connections on a frontend
7945 May be used in sections : defaults | frontend | listen | backend
7946 yes | yes | yes | no
7947 Arguments :
7948 <conns> is the maximum number of concurrent connections the frontend will
7949 accept to serve. Excess connections will be queued by the system
7950 in the socket's listen queue and will be served once a connection
7951 closes.
7952
7953 If the system supports it, it can be useful on big sites to raise this limit
Daniel Corbett9f0843f2021-05-08 10:50:37 -04007954 very high so that HAProxy manages connection queues, instead of leaving the
Willy Tarreau2769aa02007-12-27 18:26:09 +01007955 clients with unanswered connection attempts. This value should not exceed the
7956 global maxconn. Also, keep in mind that a connection contains two buffers
Baptiste Assmann79fb45d2016-03-06 23:34:31 +01007957 of tune.bufsize (16kB by default) each, as well as some other data resulting
7958 in about 33 kB of RAM being consumed per established connection. That means
7959 that a medium system equipped with 1GB of RAM can withstand around
7960 20000-25000 concurrent connections if properly tuned.
Willy Tarreau2769aa02007-12-27 18:26:09 +01007961
7962 Also, when <conns> is set to large values, it is possible that the servers
7963 are not sized to accept such loads, and for this reason it is generally wise
7964 to assign them some reasonable connection limits.
7965
Willy Tarreauc8d5b952019-02-27 17:25:52 +01007966 When this value is set to zero, which is the default, the global "maxconn"
7967 value is used.
Vincent Bernat6341be52012-06-27 17:18:30 +02007968
Willy Tarreau2769aa02007-12-27 18:26:09 +01007969 See also : "server", global section's "maxconn", "fullconn"
7970
7971
Willy Tarreau77e0dae2020-10-14 15:44:27 +02007972mode { tcp|http }
Willy Tarreau2769aa02007-12-27 18:26:09 +01007973 Set the running mode or protocol of the instance
7974 May be used in sections : defaults | frontend | listen | backend
7975 yes | yes | yes | yes
7976 Arguments :
7977 tcp The instance will work in pure TCP mode. A full-duplex connection
7978 will be established between clients and servers, and no layer 7
7979 examination will be performed. This is the default mode. It
7980 should be used for SSL, SSH, SMTP, ...
7981
7982 http The instance will work in HTTP mode. The client request will be
7983 analyzed in depth before connecting to any server. Any request
7984 which is not RFC-compliant will be rejected. Layer 7 filtering,
7985 processing and switching will be possible. This is the mode which
7986 brings HAProxy most of its value.
7987
Cyril Bonté108cf6e2012-04-21 23:30:29 +02007988 When doing content switching, it is mandatory that the frontend and the
7989 backend are in the same mode (generally HTTP), otherwise the configuration
7990 will be refused.
Willy Tarreau2769aa02007-12-27 18:26:09 +01007991
Cyril Bonté108cf6e2012-04-21 23:30:29 +02007992 Example :
Willy Tarreau2769aa02007-12-27 18:26:09 +01007993 defaults http_instances
7994 mode http
7995
Willy Tarreau0ba27502007-12-24 16:55:16 +01007996
Cyril Bontéf0c60612010-02-06 14:44:47 +01007997monitor fail { if | unless } <condition>
Willy Tarreau2769aa02007-12-27 18:26:09 +01007998 Add a condition to report a failure to a monitor HTTP request.
Willy Tarreau0ba27502007-12-24 16:55:16 +01007999 May be used in sections : defaults | frontend | listen | backend
8000 no | yes | yes | no
Willy Tarreau0ba27502007-12-24 16:55:16 +01008001 Arguments :
8002 if <cond> the monitor request will fail if the condition is satisfied,
8003 and will succeed otherwise. The condition should describe a
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01008004 combined test which must induce a failure if all conditions
Willy Tarreau0ba27502007-12-24 16:55:16 +01008005 are met, for instance a low number of servers both in a
8006 backend and its backup.
8007
8008 unless <cond> the monitor request will succeed only if the condition is
8009 satisfied, and will fail otherwise. Such a condition may be
8010 based on a test on the presence of a minimum number of active
8011 servers in a list of backends.
8012
8013 This statement adds a condition which can force the response to a monitor
8014 request to report a failure. By default, when an external component queries
8015 the URI dedicated to monitoring, a 200 response is returned. When one of the
Daniel Corbett9f0843f2021-05-08 10:50:37 -04008016 conditions above is met, HAProxy will return 503 instead of 200. This is
Willy Tarreau0ba27502007-12-24 16:55:16 +01008017 very useful to report a site failure to an external component which may base
8018 routing advertisements between multiple sites on the availability reported by
Daniel Corbett9f0843f2021-05-08 10:50:37 -04008019 HAProxy. In this case, one would rely on an ACL involving the "nbsrv"
Willy Tarreauae94d4d2011-05-11 16:28:49 +02008020 criterion. Note that "monitor fail" only works in HTTP mode. Both status
8021 messages may be tweaked using "errorfile" or "errorloc" if needed.
Willy Tarreau0ba27502007-12-24 16:55:16 +01008022
8023 Example:
8024 frontend www
Willy Tarreau2769aa02007-12-27 18:26:09 +01008025 mode http
Willy Tarreau0ba27502007-12-24 16:55:16 +01008026 acl site_dead nbsrv(dynamic) lt 2
8027 acl site_dead nbsrv(static) lt 2
8028 monitor-uri /site_alive
8029 monitor fail if site_dead
8030
Willy Tarreau9e9919d2020-10-14 15:55:23 +02008031 See also : "monitor-uri", "errorfile", "errorloc"
Willy Tarreau2769aa02007-12-27 18:26:09 +01008032
8033
Willy Tarreau2769aa02007-12-27 18:26:09 +01008034monitor-uri <uri>
8035 Intercept a URI used by external components' monitor requests
8036 May be used in sections : defaults | frontend | listen | backend
8037 yes | yes | yes | no
8038 Arguments :
8039 <uri> is the exact URI which we want to intercept to return HAProxy's
8040 health status instead of forwarding the request.
8041
8042 When an HTTP request referencing <uri> will be received on a frontend,
8043 HAProxy will not forward it nor log it, but instead will return either
8044 "HTTP/1.0 200 OK" or "HTTP/1.0 503 Service unavailable", depending on failure
8045 conditions defined with "monitor fail". This is normally enough for any
8046 front-end HTTP probe to detect that the service is UP and running without
8047 forwarding the request to a backend server. Note that the HTTP method, the
8048 version and all headers are ignored, but the request must at least be valid
8049 at the HTTP level. This keyword may only be used with an HTTP-mode frontend.
8050
Willy Tarreau721d8e02017-12-01 18:25:08 +01008051 Monitor requests are processed very early, just after the request is parsed
Christopher Faulet87f1f3d2019-07-18 14:51:20 +02008052 and even before any "http-request". The only rulesets applied before are the
8053 tcp-request ones. They cannot be logged either, and it is the intended
8054 purpose. They are only used to report HAProxy's health to an upper component,
8055 nothing more. However, it is possible to add any number of conditions using
8056 "monitor fail" and ACLs so that the result can be adjusted to whatever check
8057 can be imagined (most often the number of available servers in a backend).
Willy Tarreau2769aa02007-12-27 18:26:09 +01008058
Christopher Faulet6072beb2020-02-18 15:34:58 +01008059 Note: if <uri> starts by a slash ('/'), the matching is performed against the
8060 request's path instead of the request's uri. It is a workaround to let
8061 the HTTP/2 requests match the monitor-uri. Indeed, in HTTP/2, clients
8062 are encouraged to send absolute URIs only.
8063
Willy Tarreau2769aa02007-12-27 18:26:09 +01008064 Example :
Daniel Corbett9f0843f2021-05-08 10:50:37 -04008065 # Use /haproxy_test to report HAProxy's status
Willy Tarreau2769aa02007-12-27 18:26:09 +01008066 frontend www
8067 mode http
8068 monitor-uri /haproxy_test
8069
Willy Tarreau9e9919d2020-10-14 15:55:23 +02008070 See also : "monitor fail"
Willy Tarreau2769aa02007-12-27 18:26:09 +01008071
Willy Tarreau0ba27502007-12-24 16:55:16 +01008072
Willy Tarreaubf1f8162007-12-28 17:42:56 +01008073option abortonclose
8074no option abortonclose
8075 Enable or disable early dropping of aborted requests pending in queues.
8076 May be used in sections : defaults | frontend | listen | backend
8077 yes | no | yes | yes
8078 Arguments : none
8079
8080 In presence of very high loads, the servers will take some time to respond.
8081 The per-instance connection queue will inflate, and the response time will
8082 increase respective to the size of the queue times the average per-session
8083 response time. When clients will wait for more than a few seconds, they will
Willy Tarreau198a7442008-01-17 12:05:32 +01008084 often hit the "STOP" button on their browser, leaving a useless request in
Willy Tarreaubf1f8162007-12-28 17:42:56 +01008085 the queue, and slowing down other users, and the servers as well, because the
8086 request will eventually be served, then aborted at the first error
8087 encountered while delivering the response.
8088
8089 As there is no way to distinguish between a full STOP and a simple output
8090 close on the client side, HTTP agents should be conservative and consider
8091 that the client might only have closed its output channel while waiting for
8092 the response. However, this introduces risks of congestion when lots of users
8093 do the same, and is completely useless nowadays because probably no client at
8094 all will close the session while waiting for the response. Some HTTP agents
Davor Ocelice9ed2812017-12-25 17:49:28 +01008095 support this behavior (Squid, Apache, HAProxy), and others do not (TUX, most
Willy Tarreaubf1f8162007-12-28 17:42:56 +01008096 hardware-based load balancers). So the probability for a closed input channel
Willy Tarreau198a7442008-01-17 12:05:32 +01008097 to represent a user hitting the "STOP" button is close to 100%, and the risk
Willy Tarreaubf1f8162007-12-28 17:42:56 +01008098 of being the single component to break rare but valid traffic is extremely
8099 low, which adds to the temptation to be able to abort a session early while
8100 still not served and not pollute the servers.
8101
Davor Ocelice9ed2812017-12-25 17:49:28 +01008102 In HAProxy, the user can choose the desired behavior using the option
8103 "abortonclose". By default (without the option) the behavior is HTTP
Willy Tarreaubf1f8162007-12-28 17:42:56 +01008104 compliant and aborted requests will be served. But when the option is
8105 specified, a session with an incoming channel closed will be aborted while
8106 it is still possible, either pending in the queue for a connection slot, or
8107 during the connection establishment if the server has not yet acknowledged
8108 the connection request. This considerably reduces the queue size and the load
8109 on saturated servers when users are tempted to click on STOP, which in turn
Willy Tarreaud72758d2010-01-12 10:42:19 +01008110 reduces the response time for other users.
Willy Tarreaubf1f8162007-12-28 17:42:56 +01008111
8112 If this option has been enabled in a "defaults" section, it can be disabled
8113 in a specific instance by prepending the "no" keyword before it.
8114
8115 See also : "timeout queue" and server's "maxconn" and "maxqueue" parameters
8116
8117
Willy Tarreau4076a152009-04-02 15:18:36 +02008118option accept-invalid-http-request
8119no option accept-invalid-http-request
8120 Enable or disable relaxing of HTTP request parsing
8121 May be used in sections : defaults | frontend | listen | backend
8122 yes | yes | yes | no
8123 Arguments : none
8124
Willy Tarreau91852eb2015-05-01 13:26:00 +02008125 By default, HAProxy complies with RFC7230 in terms of message parsing. This
Willy Tarreau4076a152009-04-02 15:18:36 +02008126 means that invalid characters in header names are not permitted and cause an
Davor Ocelice9ed2812017-12-25 17:49:28 +01008127 error to be returned to the client. This is the desired behavior as such
Willy Tarreau4076a152009-04-02 15:18:36 +02008128 forbidden characters are essentially used to build attacks exploiting server
8129 weaknesses, and bypass security filtering. Sometimes, a buggy browser or
8130 server will emit invalid header names for whatever reason (configuration,
8131 implementation) and the issue will not be immediately fixed. In such a case,
8132 it is possible to relax HAProxy's header name parser to accept any character
Willy Tarreau422246e2012-01-07 23:54:13 +01008133 even if that does not make sense, by specifying this option. Similarly, the
8134 list of characters allowed to appear in a URI is well defined by RFC3986, and
8135 chars 0-31, 32 (space), 34 ('"'), 60 ('<'), 62 ('>'), 92 ('\'), 94 ('^'), 96
8136 ('`'), 123 ('{'), 124 ('|'), 125 ('}'), 127 (delete) and anything above are
Davor Ocelice9ed2812017-12-25 17:49:28 +01008137 not allowed at all. HAProxy always blocks a number of them (0..32, 127). The
Willy Tarreau91852eb2015-05-01 13:26:00 +02008138 remaining ones are blocked by default unless this option is enabled. This
Willy Tarreau13317662015-05-01 13:47:08 +02008139 option also relaxes the test on the HTTP version, it allows HTTP/0.9 requests
8140 to pass through (no version specified) and multiple digits for both the major
8141 and the minor version.
Willy Tarreau4076a152009-04-02 15:18:36 +02008142
8143 This option should never be enabled by default as it hides application bugs
8144 and open security breaches. It should only be deployed after a problem has
8145 been confirmed.
8146
8147 When this option is enabled, erroneous header names will still be accepted in
8148 requests, but the complete request will be captured in order to permit later
Willy Tarreau422246e2012-01-07 23:54:13 +01008149 analysis using the "show errors" request on the UNIX stats socket. Similarly,
8150 requests containing invalid chars in the URI part will be logged. Doing this
Willy Tarreau4076a152009-04-02 15:18:36 +02008151 also helps confirming that the issue has been solved.
8152
8153 If this option has been enabled in a "defaults" section, it can be disabled
8154 in a specific instance by prepending the "no" keyword before it.
8155
8156 See also : "option accept-invalid-http-response" and "show errors" on the
8157 stats socket.
8158
8159
8160option accept-invalid-http-response
8161no option accept-invalid-http-response
8162 Enable or disable relaxing of HTTP response parsing
8163 May be used in sections : defaults | frontend | listen | backend
8164 yes | no | yes | yes
8165 Arguments : none
8166
Willy Tarreau91852eb2015-05-01 13:26:00 +02008167 By default, HAProxy complies with RFC7230 in terms of message parsing. This
Willy Tarreau4076a152009-04-02 15:18:36 +02008168 means that invalid characters in header names are not permitted and cause an
Davor Ocelice9ed2812017-12-25 17:49:28 +01008169 error to be returned to the client. This is the desired behavior as such
Willy Tarreau4076a152009-04-02 15:18:36 +02008170 forbidden characters are essentially used to build attacks exploiting server
8171 weaknesses, and bypass security filtering. Sometimes, a buggy browser or
8172 server will emit invalid header names for whatever reason (configuration,
8173 implementation) and the issue will not be immediately fixed. In such a case,
8174 it is possible to relax HAProxy's header name parser to accept any character
Willy Tarreau91852eb2015-05-01 13:26:00 +02008175 even if that does not make sense, by specifying this option. This option also
8176 relaxes the test on the HTTP version format, it allows multiple digits for
8177 both the major and the minor version.
Willy Tarreau4076a152009-04-02 15:18:36 +02008178
8179 This option should never be enabled by default as it hides application bugs
8180 and open security breaches. It should only be deployed after a problem has
8181 been confirmed.
8182
8183 When this option is enabled, erroneous header names will still be accepted in
8184 responses, but the complete response will be captured in order to permit
8185 later analysis using the "show errors" request on the UNIX stats socket.
8186 Doing this also helps confirming that the issue has been solved.
8187
8188 If this option has been enabled in a "defaults" section, it can be disabled
8189 in a specific instance by prepending the "no" keyword before it.
8190
8191 See also : "option accept-invalid-http-request" and "show errors" on the
8192 stats socket.
8193
8194
Willy Tarreaubf1f8162007-12-28 17:42:56 +01008195option allbackups
8196no option allbackups
8197 Use either all backup servers at a time or only the first one
8198 May be used in sections : defaults | frontend | listen | backend
8199 yes | no | yes | yes
8200 Arguments : none
8201
8202 By default, the first operational backup server gets all traffic when normal
8203 servers are all down. Sometimes, it may be preferred to use multiple backups
8204 at once, because one will not be enough. When "option allbackups" is enabled,
8205 the load balancing will be performed among all backup servers when all normal
8206 ones are unavailable. The same load balancing algorithm will be used and the
8207 servers' weights will be respected. Thus, there will not be any priority
8208 order between the backup servers anymore.
8209
8210 This option is mostly used with static server farms dedicated to return a
8211 "sorry" page when an application is completely offline.
8212
8213 If this option has been enabled in a "defaults" section, it can be disabled
8214 in a specific instance by prepending the "no" keyword before it.
8215
8216
8217option checkcache
8218no option checkcache
Godbach7056a352013-12-11 20:01:07 +08008219 Analyze all server responses and block responses with cacheable cookies
Willy Tarreaubf1f8162007-12-28 17:42:56 +01008220 May be used in sections : defaults | frontend | listen | backend
8221 yes | no | yes | yes
8222 Arguments : none
8223
8224 Some high-level frameworks set application cookies everywhere and do not
8225 always let enough control to the developer to manage how the responses should
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01008226 be cached. When a session cookie is returned on a cacheable object, there is a
Willy Tarreaubf1f8162007-12-28 17:42:56 +01008227 high risk of session crossing or stealing between users traversing the same
8228 caches. In some situations, it is better to block the response than to let
Willy Tarreau3c92c5f2011-08-28 09:45:47 +02008229 some sensitive session information go in the wild.
Willy Tarreaubf1f8162007-12-28 17:42:56 +01008230
8231 The option "checkcache" enables deep inspection of all server responses for
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01008232 strict compliance with HTTP specification in terms of cacheability. It
Willy Tarreau198a7442008-01-17 12:05:32 +01008233 carefully checks "Cache-control", "Pragma" and "Set-cookie" headers in server
Willy Tarreaubf1f8162007-12-28 17:42:56 +01008234 response to check if there's a risk of caching a cookie on a client-side
8235 proxy. When this option is enabled, the only responses which can be delivered
Willy Tarreau198a7442008-01-17 12:05:32 +01008236 to the client are :
Davor Ocelice9ed2812017-12-25 17:49:28 +01008237 - all those without "Set-Cookie" header;
Willy Tarreauc55ddce2017-12-21 11:41:38 +01008238 - all those with a return code other than 200, 203, 204, 206, 300, 301,
8239 404, 405, 410, 414, 501, provided that the server has not set a
Davor Ocelice9ed2812017-12-25 17:49:28 +01008240 "Cache-control: public" header field;
Willy Tarreau24ea0bc2017-12-21 11:32:55 +01008241 - all those that result from a request using a method other than GET, HEAD,
8242 OPTIONS, TRACE, provided that the server has not set a 'Cache-Control:
Davor Ocelice9ed2812017-12-25 17:49:28 +01008243 public' header field;
Willy Tarreaubf1f8162007-12-28 17:42:56 +01008244 - those with a 'Pragma: no-cache' header
8245 - those with a 'Cache-control: private' header
8246 - those with a 'Cache-control: no-store' header
8247 - those with a 'Cache-control: max-age=0' header
8248 - those with a 'Cache-control: s-maxage=0' header
8249 - those with a 'Cache-control: no-cache' header
8250 - those with a 'Cache-control: no-cache="set-cookie"' header
8251 - those with a 'Cache-control: no-cache="set-cookie,' header
8252 (allowing other fields after set-cookie)
8253
8254 If a response doesn't respect these requirements, then it will be blocked
Christopher Faulet87f1f3d2019-07-18 14:51:20 +02008255 just as if it was from an "http-response deny" rule, with an "HTTP 502 bad
8256 gateway". The session state shows "PH--" meaning that the proxy blocked the
8257 response during headers processing. Additionally, an alert will be sent in
8258 the logs so that admins are informed that there's something to be fixed.
Willy Tarreaubf1f8162007-12-28 17:42:56 +01008259
8260 Due to the high impact on the application, the application should be tested
8261 in depth with the option enabled before going to production. It is also a
Willy Tarreaud2a4aa22008-01-31 15:28:22 +01008262 good practice to always activate it during tests, even if it is not used in
Davor Ocelice9ed2812017-12-25 17:49:28 +01008263 production, as it will report potentially dangerous application behaviors.
Willy Tarreaubf1f8162007-12-28 17:42:56 +01008264
8265 If this option has been enabled in a "defaults" section, it can be disabled
8266 in a specific instance by prepending the "no" keyword before it.
8267
8268
8269option clitcpka
8270no option clitcpka
8271 Enable or disable the sending of TCP keepalive packets on the client side
8272 May be used in sections : defaults | frontend | listen | backend
8273 yes | yes | yes | no
8274 Arguments : none
8275
8276 When there is a firewall or any session-aware component between a client and
8277 a server, and when the protocol involves very long sessions with long idle
Davor Ocelice9ed2812017-12-25 17:49:28 +01008278 periods (e.g. remote desktops), there is a risk that one of the intermediate
Willy Tarreaubf1f8162007-12-28 17:42:56 +01008279 components decides to expire a session which has remained idle for too long.
8280
8281 Enabling socket-level TCP keep-alives makes the system regularly send packets
8282 to the other end of the connection, leaving it active. The delay between
8283 keep-alive probes is controlled by the system only and depends both on the
8284 operating system and its tuning parameters.
8285
8286 It is important to understand that keep-alive packets are neither emitted nor
8287 received at the application level. It is only the network stacks which sees
8288 them. For this reason, even if one side of the proxy already uses keep-alives
8289 to maintain its connection alive, those keep-alive packets will not be
8290 forwarded to the other side of the proxy.
8291
8292 Please note that this has nothing to do with HTTP keep-alive.
8293
8294 Using option "clitcpka" enables the emission of TCP keep-alive probes on the
8295 client side of a connection, which should help when session expirations are
8296 noticed between HAProxy and a client.
8297
8298 If this option has been enabled in a "defaults" section, it can be disabled
8299 in a specific instance by prepending the "no" keyword before it.
8300
8301 See also : "option srvtcpka", "option tcpka"
8302
8303
Willy Tarreau0ba27502007-12-24 16:55:16 +01008304option contstats
8305 Enable continuous traffic statistics updates
8306 May be used in sections : defaults | frontend | listen | backend
8307 yes | yes | yes | no
8308 Arguments : none
8309
8310 By default, counters used for statistics calculation are incremented
8311 only when a session finishes. It works quite well when serving small
8312 objects, but with big ones (for example large images or archives) or
Daniel Corbett9f0843f2021-05-08 10:50:37 -04008313 with A/V streaming, a graph generated from HAProxy counters looks like
Willy Tarreaudef0d222016-11-08 22:03:00 +01008314 a hedgehog. With this option enabled counters get incremented frequently
8315 along the session, typically every 5 seconds, which is often enough to
8316 produce clean graphs. Recounting touches a hotpath directly so it is not
8317 not enabled by default, as it can cause a lot of wakeups for very large
8318 session counts and cause a small performance drop.
Willy Tarreau0ba27502007-12-24 16:55:16 +01008319
Christopher Faulet89aed322020-06-02 17:33:56 +02008320option disable-h2-upgrade
8321no option disable-h2-upgrade
8322 Enable or disable the implicit HTTP/2 upgrade from an HTTP/1.x client
8323 connection.
8324 May be used in sections : defaults | frontend | listen | backend
8325 yes | yes | yes | no
8326 Arguments : none
8327
8328 By default, HAProxy is able to implicitly upgrade an HTTP/1.x client
8329 connection to an HTTP/2 connection if the first request it receives from a
8330 given HTTP connection matches the HTTP/2 connection preface (i.e. the string
8331 "PRI * HTTP/2.0\r\n\r\nSM\r\n\r\n"). This way, it is possible to support
Christopher Faulet982e17d2021-03-26 14:44:18 +01008332 HTTP/1.x and HTTP/2 clients on a non-SSL connections. This option must be
8333 used to disable the implicit upgrade. Note this implicit upgrade is only
8334 supported for HTTP proxies, thus this option too. Note also it is possible to
8335 force the HTTP/2 on clear connections by specifying "proto h2" on the bind
8336 line. Finally, this option is applied on all bind lines. To disable implicit
8337 HTTP/2 upgrades for a specific bind line, it is possible to use "proto h1".
Christopher Faulet89aed322020-06-02 17:33:56 +02008338
8339 If this option has been enabled in a "defaults" section, it can be disabled
8340 in a specific instance by prepending the "no" keyword before it.
Willy Tarreau0ba27502007-12-24 16:55:16 +01008341
Willy Tarreauc9bd0cc2009-05-10 11:57:02 +02008342option dontlog-normal
8343no option dontlog-normal
8344 Enable or disable logging of normal, successful connections
8345 May be used in sections : defaults | frontend | listen | backend
8346 yes | yes | yes | no
8347 Arguments : none
8348
8349 There are large sites dealing with several thousand connections per second
8350 and for which logging is a major pain. Some of them are even forced to turn
8351 logs off and cannot debug production issues. Setting this option ensures that
8352 normal connections, those which experience no error, no timeout, no retry nor
8353 redispatch, will not be logged. This leaves disk space for anomalies. In HTTP
8354 mode, the response status code is checked and return codes 5xx will still be
8355 logged.
8356
8357 It is strongly discouraged to use this option as most of the time, the key to
8358 complex issues is in the normal logs which will not be logged here. If you
8359 need to separate logs, see the "log-separate-errors" option instead.
8360
Willy Tarreauc57f0e22009-05-10 13:12:33 +02008361 See also : "log", "dontlognull", "log-separate-errors" and section 8 about
Willy Tarreauc9bd0cc2009-05-10 11:57:02 +02008362 logging.
8363
8364
Willy Tarreaubf1f8162007-12-28 17:42:56 +01008365option dontlognull
8366no option dontlognull
8367 Enable or disable logging of null connections
8368 May be used in sections : defaults | frontend | listen | backend
8369 yes | yes | yes | no
8370 Arguments : none
8371
8372 In certain environments, there are components which will regularly connect to
8373 various systems to ensure that they are still alive. It can be the case from
8374 another load balancer as well as from monitoring systems. By default, even a
8375 simple port probe or scan will produce a log. If those connections pollute
8376 the logs too much, it is possible to enable option "dontlognull" to indicate
8377 that a connection on which no data has been transferred will not be logged,
Willy Tarreau0f228a02015-05-01 15:37:53 +02008378 which typically corresponds to those probes. Note that errors will still be
8379 returned to the client and accounted for in the stats. If this is not what is
8380 desired, option http-ignore-probes can be used instead.
Willy Tarreaubf1f8162007-12-28 17:42:56 +01008381
8382 It is generally recommended not to use this option in uncontrolled
Davor Ocelice9ed2812017-12-25 17:49:28 +01008383 environments (e.g. internet), otherwise scans and other malicious activities
Willy Tarreaubf1f8162007-12-28 17:42:56 +01008384 would not be logged.
8385
8386 If this option has been enabled in a "defaults" section, it can be disabled
8387 in a specific instance by prepending the "no" keyword before it.
8388
Willy Tarreau9e9919d2020-10-14 15:55:23 +02008389 See also : "log", "http-ignore-probes", "monitor-uri", and
Willy Tarreau0f228a02015-05-01 15:37:53 +02008390 section 8 about logging.
Willy Tarreaubf1f8162007-12-28 17:42:56 +01008391
Willy Tarreaubf1f8162007-12-28 17:42:56 +01008392
Willy Tarreau87cf5142011-08-19 22:57:24 +02008393option forwardfor [ except <network> ] [ header <name> ] [ if-none ]
Willy Tarreauc27debf2008-01-06 08:57:02 +01008394 Enable insertion of the X-Forwarded-For header to requests sent to servers
8395 May be used in sections : defaults | frontend | listen | backend
8396 yes | yes | yes | yes
8397 Arguments :
8398 <network> is an optional argument used to disable this option for sources
8399 matching <network>
Ross Westaf72a1d2008-08-03 10:51:45 +02008400 <name> an optional argument to specify a different "X-Forwarded-For"
Willy Tarreaud72758d2010-01-12 10:42:19 +01008401 header name.
Willy Tarreauc27debf2008-01-06 08:57:02 +01008402
8403 Since HAProxy works in reverse-proxy mode, the servers see its IP address as
8404 their client address. This is sometimes annoying when the client's IP address
8405 is expected in server logs. To solve this problem, the well-known HTTP header
8406 "X-Forwarded-For" may be added by HAProxy to all requests sent to the server.
8407 This header contains a value representing the client's IP address. Since this
8408 header is always appended at the end of the existing header list, the server
8409 must be configured to always use the last occurrence of this header only. See
Ross Westaf72a1d2008-08-03 10:51:45 +02008410 the server's manual to find how to enable use of this standard header. Note
8411 that only the last occurrence of the header must be used, since it is really
8412 possible that the client has already brought one.
8413
Willy Tarreaud72758d2010-01-12 10:42:19 +01008414 The keyword "header" may be used to supply a different header name to replace
Ross Westaf72a1d2008-08-03 10:51:45 +02008415 the default "X-Forwarded-For". This can be useful where you might already
Davor Ocelice9ed2812017-12-25 17:49:28 +01008416 have a "X-Forwarded-For" header from a different application (e.g. stunnel),
Willy Tarreaud72758d2010-01-12 10:42:19 +01008417 and you need preserve it. Also if your backend server doesn't use the
Davor Ocelice9ed2812017-12-25 17:49:28 +01008418 "X-Forwarded-For" header and requires different one (e.g. Zeus Web Servers
Ross Westaf72a1d2008-08-03 10:51:45 +02008419 require "X-Cluster-Client-IP").
Willy Tarreauc27debf2008-01-06 08:57:02 +01008420
8421 Sometimes, a same HAProxy instance may be shared between a direct client
8422 access and a reverse-proxy access (for instance when an SSL reverse-proxy is
8423 used to decrypt HTTPS traffic). It is possible to disable the addition of the
8424 header for a known source address or network by adding the "except" keyword
8425 followed by the network address. In this case, any source IP matching the
8426 network will not cause an addition of this header. Most common uses are with
Christopher Faulet5d1def62021-02-26 09:19:15 +01008427 private networks or 127.0.0.1. IPv4 and IPv6 are both supported.
Willy Tarreauc27debf2008-01-06 08:57:02 +01008428
Willy Tarreau87cf5142011-08-19 22:57:24 +02008429 Alternatively, the keyword "if-none" states that the header will only be
8430 added if it is not present. This should only be used in perfectly trusted
Daniel Corbett9f0843f2021-05-08 10:50:37 -04008431 environment, as this might cause a security issue if headers reaching HAProxy
Willy Tarreau87cf5142011-08-19 22:57:24 +02008432 are under the control of the end-user.
8433
Willy Tarreauc27debf2008-01-06 08:57:02 +01008434 This option may be specified either in the frontend or in the backend. If at
Ross Westaf72a1d2008-08-03 10:51:45 +02008435 least one of them uses it, the header will be added. Note that the backend's
8436 setting of the header subargument takes precedence over the frontend's if
Willy Tarreau87cf5142011-08-19 22:57:24 +02008437 both are defined. In the case of the "if-none" argument, if at least one of
8438 the frontend or the backend does not specify it, it wants the addition to be
8439 mandatory, so it wins.
Willy Tarreauc27debf2008-01-06 08:57:02 +01008440
Olivier Doucetaa1ea8a2016-08-05 17:15:20 +02008441 Example :
Willy Tarreauc27debf2008-01-06 08:57:02 +01008442 # Public HTTP address also used by stunnel on the same machine
8443 frontend www
8444 mode http
8445 option forwardfor except 127.0.0.1 # stunnel already adds the header
8446
Ross Westaf72a1d2008-08-03 10:51:45 +02008447 # Those servers want the IP Address in X-Client
8448 backend www
8449 mode http
8450 option forwardfor header X-Client
8451
Willy Tarreau87cf5142011-08-19 22:57:24 +02008452 See also : "option httpclose", "option http-server-close",
Christopher Faulet315b39c2018-09-21 16:26:19 +02008453 "option http-keep-alive"
Willy Tarreauc27debf2008-01-06 08:57:02 +01008454
Willy Tarreau8a8e1d92010-04-05 16:15:16 +02008455
Christopher Faulet98fbe952019-07-22 16:18:24 +02008456option h1-case-adjust-bogus-client
8457no option h1-case-adjust-bogus-client
8458 Enable or disable the case adjustment of HTTP/1 headers sent to bogus clients
8459 May be used in sections : defaults | frontend | listen | backend
8460 yes | yes | yes | no
8461 Arguments : none
8462
8463 There is no standard case for header names because, as stated in RFC7230,
8464 they are case-insensitive. So applications must handle them in a case-
8465 insensitive manner. But some bogus applications violate the standards and
8466 erroneously rely on the cases most commonly used by browsers. This problem
8467 becomes critical with HTTP/2 because all header names must be exchanged in
8468 lower case, and HAProxy follows the same convention. All header names are
8469 sent in lower case to clients and servers, regardless of the HTTP version.
8470
8471 When HAProxy receives an HTTP/1 response, its header names are converted to
8472 lower case and manipulated and sent this way to the clients. If a client is
8473 known to violate the HTTP standards and to fail to process a response coming
8474 from HAProxy, it is possible to transform the lower case header names to a
8475 different format when the response is formatted and sent to the client, by
8476 enabling this option and specifying the list of headers to be reformatted
8477 using the global directives "h1-case-adjust" or "h1-case-adjust-file". This
8478 must only be a temporary workaround for the time it takes the client to be
8479 fixed, because clients which require such workarounds might be vulnerable to
8480 content smuggling attacks and must absolutely be fixed.
8481
8482 Please note that this option will not affect standards-compliant clients.
8483
8484 If this option has been enabled in a "defaults" section, it can be disabled
8485 in a specific instance by prepending the "no" keyword before it.
8486
8487 See also: "option h1-case-adjust-bogus-server", "h1-case-adjust",
8488 "h1-case-adjust-file".
8489
8490
8491option h1-case-adjust-bogus-server
8492no option h1-case-adjust-bogus-server
8493 Enable or disable the case adjustment of HTTP/1 headers sent to bogus servers
8494 May be used in sections : defaults | frontend | listen | backend
8495 yes | no | yes | yes
8496 Arguments : none
8497
8498 There is no standard case for header names because, as stated in RFC7230,
8499 they are case-insensitive. So applications must handle them in a case-
8500 insensitive manner. But some bogus applications violate the standards and
8501 erroneously rely on the cases most commonly used by browsers. This problem
8502 becomes critical with HTTP/2 because all header names must be exchanged in
8503 lower case, and HAProxy follows the same convention. All header names are
8504 sent in lower case to clients and servers, regardless of the HTTP version.
8505
8506 When HAProxy receives an HTTP/1 request, its header names are converted to
8507 lower case and manipulated and sent this way to the servers. If a server is
8508 known to violate the HTTP standards and to fail to process a request coming
8509 from HAProxy, it is possible to transform the lower case header names to a
8510 different format when the request is formatted and sent to the server, by
8511 enabling this option and specifying the list of headers to be reformatted
8512 using the global directives "h1-case-adjust" or "h1-case-adjust-file". This
8513 must only be a temporary workaround for the time it takes the server to be
8514 fixed, because servers which require such workarounds might be vulnerable to
8515 content smuggling attacks and must absolutely be fixed.
8516
8517 Please note that this option will not affect standards-compliant servers.
8518
8519 If this option has been enabled in a "defaults" section, it can be disabled
8520 in a specific instance by prepending the "no" keyword before it.
8521
8522 See also: "option h1-case-adjust-bogus-client", "h1-case-adjust",
8523 "h1-case-adjust-file".
8524
8525
Willy Tarreau9fbe18e2015-05-01 22:42:08 +02008526option http-buffer-request
8527no option http-buffer-request
8528 Enable or disable waiting for whole HTTP request body before proceeding
8529 May be used in sections : defaults | frontend | listen | backend
8530 yes | yes | yes | yes
8531 Arguments : none
8532
8533 It is sometimes desirable to wait for the body of an HTTP request before
8534 taking a decision. This is what is being done by "balance url_param" for
8535 example. The first use case is to buffer requests from slow clients before
8536 connecting to the server. Another use case consists in taking the routing
8537 decision based on the request body's contents. This option placed in a
8538 frontend or backend forces the HTTP processing to wait until either the whole
Christopher Faulet6db8a2e2019-11-19 16:27:25 +01008539 body is received or the request buffer is full. It can have undesired side
8540 effects with some applications abusing HTTP by expecting unbuffered
8541 transmissions between the frontend and the backend, so this should definitely
8542 not be used by default.
Willy Tarreau9fbe18e2015-05-01 22:42:08 +02008543
Christopher Faulet021a8e42021-03-29 10:46:38 +02008544 See also : "option http-no-delay", "timeout http-request",
8545 "http-request wait-for-body"
Willy Tarreau9fbe18e2015-05-01 22:42:08 +02008546
8547
Willy Tarreau0f228a02015-05-01 15:37:53 +02008548option http-ignore-probes
8549no option http-ignore-probes
8550 Enable or disable logging of null connections and request timeouts
8551 May be used in sections : defaults | frontend | listen | backend
8552 yes | yes | yes | no
8553 Arguments : none
8554
8555 Recently some browsers started to implement a "pre-connect" feature
8556 consisting in speculatively connecting to some recently visited web sites
8557 just in case the user would like to visit them. This results in many
8558 connections being established to web sites, which end up in 408 Request
8559 Timeout if the timeout strikes first, or 400 Bad Request when the browser
8560 decides to close them first. These ones pollute the log and feed the error
8561 counters. There was already "option dontlognull" but it's insufficient in
8562 this case. Instead, this option does the following things :
8563 - prevent any 400/408 message from being sent to the client if nothing
Davor Ocelice9ed2812017-12-25 17:49:28 +01008564 was received over a connection before it was closed;
8565 - prevent any log from being emitted in this situation;
Willy Tarreau0f228a02015-05-01 15:37:53 +02008566 - prevent any error counter from being incremented
8567
8568 That way the empty connection is silently ignored. Note that it is better
8569 not to use this unless it is clear that it is needed, because it will hide
8570 real problems. The most common reason for not receiving a request and seeing
8571 a 408 is due to an MTU inconsistency between the client and an intermediary
8572 element such as a VPN, which blocks too large packets. These issues are
8573 generally seen with POST requests as well as GET with large cookies. The logs
8574 are often the only way to detect them.
8575
8576 If this option has been enabled in a "defaults" section, it can be disabled
8577 in a specific instance by prepending the "no" keyword before it.
8578
8579 See also : "log", "dontlognull", "errorfile", and section 8 about logging.
8580
8581
Willy Tarreau16bfb022010-01-16 19:48:41 +01008582option http-keep-alive
8583no option http-keep-alive
8584 Enable or disable HTTP keep-alive from client to server
8585 May be used in sections : defaults | frontend | listen | backend
8586 yes | yes | yes | yes
8587 Arguments : none
8588
Willy Tarreau70dffda2014-01-30 03:07:23 +01008589 By default HAProxy operates in keep-alive mode with regards to persistent
8590 connections: for each connection it processes each request and response, and
Christopher Faulet315b39c2018-09-21 16:26:19 +02008591 leaves the connection idle on both sides between the end of a response and
8592 the start of a new request. This mode may be changed by several options such
Christopher Faulet159e6672019-07-16 15:09:52 +02008593 as "option http-server-close" or "option httpclose". This option allows to
8594 set back the keep-alive mode, which can be useful when another mode was used
8595 in a defaults section.
Willy Tarreau70dffda2014-01-30 03:07:23 +01008596
8597 Setting "option http-keep-alive" enables HTTP keep-alive mode on the client-
8598 and server- sides. This provides the lowest latency on the client side (slow
Willy Tarreau16bfb022010-01-16 19:48:41 +01008599 network) and the fastest session reuse on the server side at the expense
8600 of maintaining idle connections to the servers. In general, it is possible
8601 with this option to achieve approximately twice the request rate that the
8602 "http-server-close" option achieves on small objects. There are mainly two
8603 situations where this option may be useful :
8604
8605 - when the server is non-HTTP compliant and authenticates the connection
Davor Ocelice9ed2812017-12-25 17:49:28 +01008606 instead of requests (e.g. NTLM authentication)
Willy Tarreau16bfb022010-01-16 19:48:41 +01008607
8608 - when the cost of establishing the connection to the server is significant
8609 compared to the cost of retrieving the associated object from the server.
8610
8611 This last case can happen when the server is a fast static server of cache.
8612 In this case, the server will need to be properly tuned to support high enough
8613 connection counts because connections will last until the client sends another
8614 request.
8615
8616 If the client request has to go to another backend or another server due to
8617 content switching or the load balancing algorithm, the idle connection will
Willy Tarreau9420b122013-12-15 18:58:25 +01008618 immediately be closed and a new one re-opened. Option "prefer-last-server" is
8619 available to try optimize server selection so that if the server currently
8620 attached to an idle connection is usable, it will be used.
Willy Tarreau16bfb022010-01-16 19:48:41 +01008621
Willy Tarreau16bfb022010-01-16 19:48:41 +01008622 At the moment, logs will not indicate whether requests came from the same
8623 session or not. The accept date reported in the logs corresponds to the end
8624 of the previous request, and the request time corresponds to the time spent
8625 waiting for a new request. The keep-alive request time is still bound to the
8626 timeout defined by "timeout http-keep-alive" or "timeout http-request" if
8627 not set.
8628
Christopher Faulet159e6672019-07-16 15:09:52 +02008629 This option disables and replaces any previous "option httpclose" or "option
8630 http-server-close". When backend and frontend options differ, all of these 4
8631 options have precedence over "option http-keep-alive".
Willy Tarreau16bfb022010-01-16 19:48:41 +01008632
Christopher Faulet315b39c2018-09-21 16:26:19 +02008633 See also : "option httpclose",, "option http-server-close",
Willy Tarreau9420b122013-12-15 18:58:25 +01008634 "option prefer-last-server", "option http-pretend-keepalive",
Frédéric Lécaille93d33162019-03-06 09:35:59 +01008635 and "1.1. The HTTP transaction model".
Willy Tarreau16bfb022010-01-16 19:48:41 +01008636
8637
Willy Tarreau96e31212011-05-30 18:10:30 +02008638option http-no-delay
8639no option http-no-delay
8640 Instruct the system to favor low interactive delays over performance in HTTP
8641 May be used in sections : defaults | frontend | listen | backend
8642 yes | yes | yes | yes
8643 Arguments : none
8644
8645 In HTTP, each payload is unidirectional and has no notion of interactivity.
8646 Any agent is expected to queue data somewhat for a reasonably low delay.
8647 There are some very rare server-to-server applications that abuse the HTTP
8648 protocol and expect the payload phase to be highly interactive, with many
8649 interleaved data chunks in both directions within a single request. This is
8650 absolutely not supported by the HTTP specification and will not work across
8651 most proxies or servers. When such applications attempt to do this through
Daniel Corbett9f0843f2021-05-08 10:50:37 -04008652 HAProxy, it works but they will experience high delays due to the network
Willy Tarreau96e31212011-05-30 18:10:30 +02008653 optimizations which favor performance by instructing the system to wait for
8654 enough data to be available in order to only send full packets. Typical
8655 delays are around 200 ms per round trip. Note that this only happens with
8656 abnormal uses. Normal uses such as CONNECT requests nor WebSockets are not
8657 affected.
8658
8659 When "option http-no-delay" is present in either the frontend or the backend
8660 used by a connection, all such optimizations will be disabled in order to
8661 make the exchanges as fast as possible. Of course this offers no guarantee on
8662 the functionality, as it may break at any other place. But if it works via
8663 HAProxy, it will work as fast as possible. This option should never be used
8664 by default, and should never be used at all unless such a buggy application
8665 is discovered. The impact of using this option is an increase of bandwidth
8666 usage and CPU usage, which may significantly lower performance in high
8667 latency environments.
8668
Willy Tarreau9fbe18e2015-05-01 22:42:08 +02008669 See also : "option http-buffer-request"
8670
Willy Tarreau96e31212011-05-30 18:10:30 +02008671
Willy Tarreau8a8e1d92010-04-05 16:15:16 +02008672option http-pretend-keepalive
8673no option http-pretend-keepalive
Daniel Corbett9f0843f2021-05-08 10:50:37 -04008674 Define whether HAProxy will announce keepalive to the server or not
Willy Tarreau8a8e1d92010-04-05 16:15:16 +02008675 May be used in sections : defaults | frontend | listen | backend
Christopher Faulet98db9762018-09-21 10:25:19 +02008676 yes | no | yes | yes
Willy Tarreau8a8e1d92010-04-05 16:15:16 +02008677 Arguments : none
8678
Daniel Corbett9f0843f2021-05-08 10:50:37 -04008679 When running with "option http-server-close" or "option httpclose", HAProxy
Willy Tarreau8a8e1d92010-04-05 16:15:16 +02008680 adds a "Connection: close" header to the request forwarded to the server.
8681 Unfortunately, when some servers see this header, they automatically refrain
8682 from using the chunked encoding for responses of unknown length, while this
Daniel Corbett9f0843f2021-05-08 10:50:37 -04008683 is totally unrelated. The immediate effect is that this prevents HAProxy from
Willy Tarreau8a8e1d92010-04-05 16:15:16 +02008684 maintaining the client connection alive. A second effect is that a client or
8685 a cache could receive an incomplete response without being aware of it, and
8686 consider the response complete.
8687
Daniel Corbett9f0843f2021-05-08 10:50:37 -04008688 By setting "option http-pretend-keepalive", HAProxy will make the server
Willy Tarreau8a8e1d92010-04-05 16:15:16 +02008689 believe it will keep the connection alive. The server will then not fall back
Daniel Corbett9f0843f2021-05-08 10:50:37 -04008690 to the abnormal undesired above. When HAProxy gets the whole response, it
Willy Tarreau8a8e1d92010-04-05 16:15:16 +02008691 will close the connection with the server just as it would do with the
Christopher Faulet315b39c2018-09-21 16:26:19 +02008692 "option httpclose". That way the client gets a normal response and the
Willy Tarreau8a8e1d92010-04-05 16:15:16 +02008693 connection is correctly closed on the server side.
8694
8695 It is recommended not to enable this option by default, because most servers
8696 will more efficiently close the connection themselves after the last packet,
8697 and release its buffers slightly earlier. Also, the added packet on the
8698 network could slightly reduce the overall peak performance. However it is
Daniel Corbett9f0843f2021-05-08 10:50:37 -04008699 worth noting that when this option is enabled, HAProxy will have slightly
8700 less work to do. So if HAProxy is the bottleneck on the whole architecture,
Willy Tarreau8a8e1d92010-04-05 16:15:16 +02008701 enabling this option might save a few CPU cycles.
8702
Christopher Faulet98db9762018-09-21 10:25:19 +02008703 This option may be set in backend and listen sections. Using it in a frontend
8704 section will be ignored and a warning will be reported during startup. It is
8705 a backend related option, so there is no real reason to set it on a
8706 frontend. This option may be combined with "option httpclose", which will
8707 cause keepalive to be announced to the server and close to be announced to
8708 the client. This practice is discouraged though.
Willy Tarreau8a8e1d92010-04-05 16:15:16 +02008709
8710 If this option has been enabled in a "defaults" section, it can be disabled
8711 in a specific instance by prepending the "no" keyword before it.
8712
Christopher Faulet315b39c2018-09-21 16:26:19 +02008713 See also : "option httpclose", "option http-server-close", and
Willy Tarreau16bfb022010-01-16 19:48:41 +01008714 "option http-keep-alive"
Willy Tarreau8a8e1d92010-04-05 16:15:16 +02008715
Willy Tarreauc27debf2008-01-06 08:57:02 +01008716
Willy Tarreaub608feb2010-01-02 22:47:18 +01008717option http-server-close
8718no option http-server-close
8719 Enable or disable HTTP connection closing on the server side
8720 May be used in sections : defaults | frontend | listen | backend
8721 yes | yes | yes | yes
8722 Arguments : none
8723
Willy Tarreau70dffda2014-01-30 03:07:23 +01008724 By default HAProxy operates in keep-alive mode with regards to persistent
8725 connections: for each connection it processes each request and response, and
8726 leaves the connection idle on both sides between the end of a response and
8727 the start of a new request. This mode may be changed by several options such
Christopher Faulet159e6672019-07-16 15:09:52 +02008728 as "option http-server-close" or "option httpclose". Setting "option
8729 http-server-close" enables HTTP connection-close mode on the server side
8730 while keeping the ability to support HTTP keep-alive and pipelining on the
8731 client side. This provides the lowest latency on the client side (slow
8732 network) and the fastest session reuse on the server side to save server
8733 resources, similarly to "option httpclose". It also permits non-keepalive
8734 capable servers to be served in keep-alive mode to the clients if they
8735 conform to the requirements of RFC7230. Please note that some servers do not
8736 always conform to those requirements when they see "Connection: close" in the
8737 request. The effect will be that keep-alive will never be used. A workaround
8738 consists in enabling "option http-pretend-keepalive".
Willy Tarreaub608feb2010-01-02 22:47:18 +01008739
8740 At the moment, logs will not indicate whether requests came from the same
8741 session or not. The accept date reported in the logs corresponds to the end
8742 of the previous request, and the request time corresponds to the time spent
8743 waiting for a new request. The keep-alive request time is still bound to the
Willy Tarreaub16a5742010-01-10 14:46:16 +01008744 timeout defined by "timeout http-keep-alive" or "timeout http-request" if
8745 not set.
Willy Tarreaub608feb2010-01-02 22:47:18 +01008746
8747 This option may be set both in a frontend and in a backend. It is enabled if
8748 at least one of the frontend or backend holding a connection has it enabled.
Christopher Faulet159e6672019-07-16 15:09:52 +02008749 It disables and replaces any previous "option httpclose" or "option
8750 http-keep-alive". Please check section 4 ("Proxies") to see how this option
8751 combines with others when frontend and backend options differ.
Willy Tarreaub608feb2010-01-02 22:47:18 +01008752
8753 If this option has been enabled in a "defaults" section, it can be disabled
8754 in a specific instance by prepending the "no" keyword before it.
8755
Christopher Faulet315b39c2018-09-21 16:26:19 +02008756 See also : "option httpclose", "option http-pretend-keepalive",
8757 "option http-keep-alive", and "1.1. The HTTP transaction model".
Willy Tarreaub608feb2010-01-02 22:47:18 +01008758
Willy Tarreau88d349d2010-01-25 12:15:43 +01008759option http-use-proxy-header
Cyril Bontéf0c60612010-02-06 14:44:47 +01008760no option http-use-proxy-header
Willy Tarreau88d349d2010-01-25 12:15:43 +01008761 Make use of non-standard Proxy-Connection header instead of Connection
8762 May be used in sections : defaults | frontend | listen | backend
8763 yes | yes | yes | no
8764 Arguments : none
8765
Lukas Tribus23953682017-04-28 13:24:30 +00008766 While RFC7230 explicitly states that HTTP/1.1 agents must use the
Willy Tarreau88d349d2010-01-25 12:15:43 +01008767 Connection header to indicate their wish of persistent or non-persistent
8768 connections, both browsers and proxies ignore this header for proxied
8769 connections and make use of the undocumented, non-standard Proxy-Connection
8770 header instead. The issue begins when trying to put a load balancer between
8771 browsers and such proxies, because there will be a difference between what
Daniel Corbett9f0843f2021-05-08 10:50:37 -04008772 HAProxy understands and what the client and the proxy agree on.
Willy Tarreau88d349d2010-01-25 12:15:43 +01008773
Daniel Corbett9f0843f2021-05-08 10:50:37 -04008774 By setting this option in a frontend, HAProxy can automatically switch to use
Willy Tarreau88d349d2010-01-25 12:15:43 +01008775 that non-standard header if it sees proxied requests. A proxied request is
Lukas Tribusf01a9cd2016-02-03 18:09:37 +01008776 defined here as one where the URI begins with neither a '/' nor a '*'. This
8777 is incompatible with the HTTP tunnel mode. Note that this option can only be
8778 specified in a frontend and will affect the request along its whole life.
Willy Tarreau88d349d2010-01-25 12:15:43 +01008779
Willy Tarreau844a7e72010-01-31 21:46:18 +01008780 Also, when this option is set, a request which requires authentication will
8781 automatically switch to use proxy authentication headers if it is itself a
8782 proxied request. That makes it possible to check or enforce authentication in
8783 front of an existing proxy.
8784
Willy Tarreau88d349d2010-01-25 12:15:43 +01008785 This option should normally never be used, except in front of a proxy.
8786
Christopher Faulet315b39c2018-09-21 16:26:19 +02008787 See also : "option httpclose", and "option http-server-close".
Willy Tarreau88d349d2010-01-25 12:15:43 +01008788
Willy Tarreaud63335a2010-02-26 12:56:52 +01008789option httpchk
8790option httpchk <uri>
8791option httpchk <method> <uri>
8792option httpchk <method> <uri> <version>
Christopher Fauletc52ea4d2020-04-23 15:43:35 +02008793 Enables HTTP protocol to check on the servers health
Willy Tarreaud63335a2010-02-26 12:56:52 +01008794 May be used in sections : defaults | frontend | listen | backend
8795 yes | no | yes | yes
8796 Arguments :
8797 <method> is the optional HTTP method used with the requests. When not set,
8798 the "OPTIONS" method is used, as it generally requires low server
8799 processing and is easy to filter out from the logs. Any method
8800 may be used, though it is not recommended to invent non-standard
8801 ones.
8802
8803 <uri> is the URI referenced in the HTTP requests. It defaults to " / "
8804 which is accessible by default on almost any server, but may be
8805 changed to any other URI. Query strings are permitted.
8806
8807 <version> is the optional HTTP version string. It defaults to "HTTP/1.0"
8808 but some servers might behave incorrectly in HTTP 1.0, so turning
8809 it to HTTP/1.1 may sometimes help. Note that the Host field is
Christopher Faulet8acb1282020-04-09 08:44:06 +02008810 mandatory in HTTP/1.1, use "http-check send" directive to add it.
Willy Tarreaud63335a2010-02-26 12:56:52 +01008811
8812 By default, server health checks only consist in trying to establish a TCP
8813 connection. When "option httpchk" is specified, a complete HTTP request is
8814 sent once the TCP connection is established, and responses 2xx and 3xx are
8815 considered valid, while all other ones indicate a server failure, including
8816 the lack of any response.
8817
Christopher Faulete5870d82020-04-15 11:32:03 +02008818 Combined with "http-check" directives, it is possible to customize the
8819 request sent during the HTTP health checks or the matching rules on the
8820 response. It is also possible to configure a send/expect sequence, just like
8821 with the directive "tcp-check" for TCP health checks.
8822
8823 The server configuration is used by default to open connections to perform
8824 HTTP health checks. By it is also possible to overwrite server parameters
8825 using "http-check connect" rules.
Willy Tarreaud63335a2010-02-26 12:56:52 +01008826
Christopher Faulete5870d82020-04-15 11:32:03 +02008827 "httpchk" option does not necessarily require an HTTP backend, it also works
8828 with plain TCP backends. This is particularly useful to check simple scripts
Christopher Faulet14cd3162020-04-16 14:50:06 +02008829 bound to some dedicated ports using the inetd daemon. However, it will always
Daniel Corbett67a82712020-07-06 23:01:19 -04008830 internally relies on an HTX multiplexer. Thus, it means the request
Christopher Faulet14cd3162020-04-16 14:50:06 +02008831 formatting and the response parsing will be strict.
Willy Tarreaud63335a2010-02-26 12:56:52 +01008832
Christopher Faulet8acb1282020-04-09 08:44:06 +02008833 Note : For a while, there was no way to add headers or body in the request
8834 used for HTTP health checks. So a workaround was to hide it at the end
8835 of the version string with a "\r\n" after the version. It is now
8836 deprecated. The directive "http-check send" must be used instead.
8837
Willy Tarreaud63335a2010-02-26 12:56:52 +01008838 Examples :
Christopher Fauletc52ea4d2020-04-23 15:43:35 +02008839 # Relay HTTPS traffic to Apache instance and check service availability
8840 # using HTTP request "OPTIONS * HTTP/1.1" on port 80.
8841 backend https_relay
8842 mode tcp
8843 option httpchk OPTIONS * HTTP/1.1
8844 http-check send hdr Host www
8845 server apache1 192.168.1.1:443 check port 80
Willy Tarreaud63335a2010-02-26 12:56:52 +01008846
Simon Hormanafc47ee2013-11-25 10:46:35 +09008847 See also : "option ssl-hello-chk", "option smtpchk", "option mysql-check",
8848 "option pgsql-check", "http-check" and the "check", "port" and
8849 "inter" server options.
Willy Tarreaud63335a2010-02-26 12:56:52 +01008850
8851
Willy Tarreauc27debf2008-01-06 08:57:02 +01008852option httpclose
8853no option httpclose
Christopher Faulet315b39c2018-09-21 16:26:19 +02008854 Enable or disable HTTP connection closing
Willy Tarreauc27debf2008-01-06 08:57:02 +01008855 May be used in sections : defaults | frontend | listen | backend
8856 yes | yes | yes | yes
8857 Arguments : none
8858
Willy Tarreau70dffda2014-01-30 03:07:23 +01008859 By default HAProxy operates in keep-alive mode with regards to persistent
8860 connections: for each connection it processes each request and response, and
8861 leaves the connection idle on both sides between the end of a response and
8862 the start of a new request. This mode may be changed by several options such
Christopher Faulet159e6672019-07-16 15:09:52 +02008863 as "option http-server-close" or "option httpclose".
Willy Tarreau70dffda2014-01-30 03:07:23 +01008864
Christopher Faulet315b39c2018-09-21 16:26:19 +02008865 If "option httpclose" is set, HAProxy will close connections with the server
8866 and the client as soon as the request and the response are received. It will
John Roeslerfb2fce12019-07-10 15:45:51 -05008867 also check if a "Connection: close" header is already set in each direction,
Christopher Faulet315b39c2018-09-21 16:26:19 +02008868 and will add one if missing. Any "Connection" header different from "close"
8869 will also be removed.
Willy Tarreauc27debf2008-01-06 08:57:02 +01008870
Christopher Faulet315b39c2018-09-21 16:26:19 +02008871 This option may also be combined with "option http-pretend-keepalive", which
8872 will disable sending of the "Connection: close" header, but will still cause
8873 the connection to be closed once the whole response is received.
Willy Tarreauc27debf2008-01-06 08:57:02 +01008874
8875 This option may be set both in a frontend and in a backend. It is enabled if
8876 at least one of the frontend or backend holding a connection has it enabled.
Christopher Faulet159e6672019-07-16 15:09:52 +02008877 It disables and replaces any previous "option http-server-close" or "option
8878 http-keep-alive". Please check section 4 ("Proxies") to see how this option
8879 combines with others when frontend and backend options differ.
Willy Tarreauc27debf2008-01-06 08:57:02 +01008880
8881 If this option has been enabled in a "defaults" section, it can be disabled
8882 in a specific instance by prepending the "no" keyword before it.
8883
Christopher Faulet315b39c2018-09-21 16:26:19 +02008884 See also : "option http-server-close" and "1.1. The HTTP transaction model".
Willy Tarreauc27debf2008-01-06 08:57:02 +01008885
8886
Emeric Brun3a058f32009-06-30 18:26:00 +02008887option httplog [ clf ]
Willy Tarreauc27debf2008-01-06 08:57:02 +01008888 Enable logging of HTTP request, session state and timers
8889 May be used in sections : defaults | frontend | listen | backend
Tim Duesterhus9ad9f352018-02-05 20:52:27 +01008890 yes | yes | yes | no
Emeric Brun3a058f32009-06-30 18:26:00 +02008891 Arguments :
8892 clf if the "clf" argument is added, then the output format will be
8893 the CLF format instead of HAProxy's default HTTP format. You can
8894 use this when you need to feed HAProxy's logs through a specific
Davor Ocelice9ed2812017-12-25 17:49:28 +01008895 log analyzer which only support the CLF format and which is not
Emeric Brun3a058f32009-06-30 18:26:00 +02008896 extensible.
Willy Tarreauc27debf2008-01-06 08:57:02 +01008897
8898 By default, the log output format is very poor, as it only contains the
8899 source and destination addresses, and the instance name. By specifying
8900 "option httplog", each log line turns into a much richer format including,
8901 but not limited to, the HTTP request, the connection timers, the session
8902 status, the connections numbers, the captured headers and cookies, the
8903 frontend, backend and server name, and of course the source address and
8904 ports.
8905
PiBa-NLbd556bf2014-12-11 21:31:54 +01008906 Specifying only "option httplog" will automatically clear the 'clf' mode
8907 if it was set by default.
Emeric Brun3a058f32009-06-30 18:26:00 +02008908
Guillaume de Lafond29f45602017-03-31 19:52:15 +02008909 "option httplog" overrides any previous "log-format" directive.
8910
Willy Tarreauc57f0e22009-05-10 13:12:33 +02008911 See also : section 8 about logging.
Willy Tarreauc27debf2008-01-06 08:57:02 +01008912
Willy Tarreau55165fe2009-05-10 12:02:55 +02008913
8914option http_proxy
8915no option http_proxy
8916 Enable or disable plain HTTP proxy mode
8917 May be used in sections : defaults | frontend | listen | backend
8918 yes | yes | yes | yes
8919 Arguments : none
8920
8921 It sometimes happens that people need a pure HTTP proxy which understands
8922 basic proxy requests without caching nor any fancy feature. In this case,
8923 it may be worth setting up an HAProxy instance with the "option http_proxy"
8924 set. In this mode, no server is declared, and the connection is forwarded to
8925 the IP address and port found in the URL after the "http://" scheme.
8926
8927 No host address resolution is performed, so this only works when pure IP
8928 addresses are passed. Since this option's usage perimeter is rather limited,
Lukas Tribusf01a9cd2016-02-03 18:09:37 +01008929 it will probably be used only by experts who know they need exactly it. This
8930 is incompatible with the HTTP tunnel mode.
Willy Tarreau55165fe2009-05-10 12:02:55 +02008931
8932 If this option has been enabled in a "defaults" section, it can be disabled
8933 in a specific instance by prepending the "no" keyword before it.
8934
8935 Example :
8936 # this backend understands HTTP proxy requests and forwards them directly.
8937 backend direct_forward
8938 option httpclose
8939 option http_proxy
8940
8941 See also : "option httpclose"
8942
Willy Tarreau211ad242009-10-03 21:45:07 +02008943
Jamie Gloudon801a0a32012-08-25 00:18:33 -04008944option independent-streams
8945no option independent-streams
8946 Enable or disable independent timeout processing for both directions
Willy Tarreauf27b5ea2009-10-03 22:01:18 +02008947 May be used in sections : defaults | frontend | listen | backend
8948 yes | yes | yes | yes
8949 Arguments : none
8950
8951 By default, when data is sent over a socket, both the write timeout and the
8952 read timeout for that socket are refreshed, because we consider that there is
8953 activity on that socket, and we have no other means of guessing if we should
8954 receive data or not.
8955
Davor Ocelice9ed2812017-12-25 17:49:28 +01008956 While this default behavior is desirable for almost all applications, there
Willy Tarreauf27b5ea2009-10-03 22:01:18 +02008957 exists a situation where it is desirable to disable it, and only refresh the
8958 read timeout if there are incoming data. This happens on sessions with large
8959 timeouts and low amounts of exchanged data such as telnet session. If the
8960 server suddenly disappears, the output data accumulates in the system's
8961 socket buffers, both timeouts are correctly refreshed, and there is no way
8962 to know the server does not receive them, so we don't timeout. However, when
8963 the underlying protocol always echoes sent data, it would be enough by itself
8964 to detect the issue using the read timeout. Note that this problem does not
8965 happen with more verbose protocols because data won't accumulate long in the
8966 socket buffers.
8967
8968 When this option is set on the frontend, it will disable read timeout updates
8969 on data sent to the client. There probably is little use of this case. When
8970 the option is set on the backend, it will disable read timeout updates on
8971 data sent to the server. Doing so will typically break large HTTP posts from
8972 slow lines, so use it with caution.
8973
Willy Tarreauce887fd2012-05-12 12:50:00 +02008974 See also : "timeout client", "timeout server" and "timeout tunnel"
Willy Tarreauf27b5ea2009-10-03 22:01:18 +02008975
8976
Gabor Lekenyb4c81e42010-09-29 18:17:05 +02008977option ldap-check
8978 Use LDAPv3 health checks for server testing
8979 May be used in sections : defaults | frontend | listen | backend
8980 yes | no | yes | yes
8981 Arguments : none
8982
8983 It is possible to test that the server correctly talks LDAPv3 instead of just
8984 testing that it accepts the TCP connection. When this option is set, an
8985 LDAPv3 anonymous simple bind message is sent to the server, and the response
8986 is analyzed to find an LDAPv3 bind response message.
8987
8988 The server is considered valid only when the LDAP response contains success
8989 resultCode (http://tools.ietf.org/html/rfc4511#section-4.1.9).
8990
8991 Logging of bind requests is server dependent see your documentation how to
8992 configure it.
8993
8994 Example :
8995 option ldap-check
8996
8997 See also : "option httpchk"
8998
8999
Simon Horman98637e52014-06-20 12:30:16 +09009000option external-check
9001 Use external processes for server health checks
9002 May be used in sections : defaults | frontend | listen | backend
9003 yes | no | yes | yes
9004
9005 It is possible to test the health of a server using an external command.
9006 This is achieved by running the executable set using "external-check
9007 command".
9008
9009 Requires the "external-check" global to be set.
9010
9011 See also : "external-check", "external-check command", "external-check path"
9012
9013
Willy Tarreau211ad242009-10-03 21:45:07 +02009014option log-health-checks
9015no option log-health-checks
Willy Tarreaubef1b322014-05-13 21:01:39 +02009016 Enable or disable logging of health checks status updates
Willy Tarreau211ad242009-10-03 21:45:07 +02009017 May be used in sections : defaults | frontend | listen | backend
9018 yes | no | yes | yes
9019 Arguments : none
9020
Willy Tarreaubef1b322014-05-13 21:01:39 +02009021 By default, failed health check are logged if server is UP and successful
9022 health checks are logged if server is DOWN, so the amount of additional
9023 information is limited.
Willy Tarreau211ad242009-10-03 21:45:07 +02009024
Willy Tarreaubef1b322014-05-13 21:01:39 +02009025 When this option is enabled, any change of the health check status or to
9026 the server's health will be logged, so that it becomes possible to know
9027 that a server was failing occasional checks before crashing, or exactly when
9028 it failed to respond a valid HTTP status, then when the port started to
9029 reject connections, then when the server stopped responding at all.
9030
Davor Ocelice9ed2812017-12-25 17:49:28 +01009031 Note that status changes not caused by health checks (e.g. enable/disable on
Willy Tarreaubef1b322014-05-13 21:01:39 +02009032 the CLI) are intentionally not logged by this option.
Willy Tarreau211ad242009-10-03 21:45:07 +02009033
Willy Tarreaubef1b322014-05-13 21:01:39 +02009034 See also: "option httpchk", "option ldap-check", "option mysql-check",
9035 "option pgsql-check", "option redis-check", "option smtpchk",
9036 "option tcp-check", "log" and section 8 about logging.
Willy Tarreau211ad242009-10-03 21:45:07 +02009037
Willy Tarreauc9bd0cc2009-05-10 11:57:02 +02009038
9039option log-separate-errors
9040no option log-separate-errors
9041 Change log level for non-completely successful connections
9042 May be used in sections : defaults | frontend | listen | backend
9043 yes | yes | yes | no
9044 Arguments : none
9045
Daniel Corbett9f0843f2021-05-08 10:50:37 -04009046 Sometimes looking for errors in logs is not easy. This option makes HAProxy
Willy Tarreauc9bd0cc2009-05-10 11:57:02 +02009047 raise the level of logs containing potentially interesting information such
9048 as errors, timeouts, retries, redispatches, or HTTP status codes 5xx. The
9049 level changes from "info" to "err". This makes it possible to log them
9050 separately to a different file with most syslog daemons. Be careful not to
9051 remove them from the original file, otherwise you would lose ordering which
9052 provides very important information.
9053
9054 Using this option, large sites dealing with several thousand connections per
9055 second may log normal traffic to a rotating buffer and only archive smaller
9056 error logs.
9057
Willy Tarreauc57f0e22009-05-10 13:12:33 +02009058 See also : "log", "dontlognull", "dontlog-normal" and section 8 about
Willy Tarreauc9bd0cc2009-05-10 11:57:02 +02009059 logging.
9060
Willy Tarreauc27debf2008-01-06 08:57:02 +01009061
9062option logasap
9063no option logasap
Jerome Magnin95fb57b2020-04-23 19:01:17 +02009064 Enable or disable early logging.
Willy Tarreauc27debf2008-01-06 08:57:02 +01009065 May be used in sections : defaults | frontend | listen | backend
9066 yes | yes | yes | no
9067 Arguments : none
9068
Jerome Magnin95fb57b2020-04-23 19:01:17 +02009069 By default, logs are emitted when all the log format variables and sample
9070 fetches used in the definition of the log-format string return a value, or
9071 when the session is terminated. This allows the built in log-format strings
9072 to account for the transfer time, or the number of bytes in log messages.
9073
9074 When handling long lived connections such as large file transfers or RDP,
9075 it may take a while for the request or connection to appear in the logs.
9076 Using "option logasap", the log message is created as soon as the server
9077 connection is established in mode tcp, or as soon as the server sends the
9078 complete headers in mode http. Missing information in the logs will be the
Ilya Shipitsin4329a9a2020-05-05 21:17:10 +05009079 total number of bytes which will only indicate the amount of data transferred
Jerome Magnin95fb57b2020-04-23 19:01:17 +02009080 before the message was created and the total time which will not take the
9081 remainder of the connection life or transfer time into account. For the case
9082 of HTTP, it is good practice to capture the Content-Length response header
9083 so that the logs at least indicate how many bytes are expected to be
Ilya Shipitsin4329a9a2020-05-05 21:17:10 +05009084 transferred.
Willy Tarreauc27debf2008-01-06 08:57:02 +01009085
Willy Tarreaucc6c8912009-02-22 10:53:55 +01009086 Examples :
9087 listen http_proxy 0.0.0.0:80
9088 mode http
9089 option httplog
9090 option logasap
9091 log 192.168.2.200 local3
9092
9093 >>> Feb 6 12:14:14 localhost \
9094 haproxy[14389]: 10.0.1.2:33317 [06/Feb/2009:12:14:14.655] http-in \
9095 static/srv1 9/10/7/14/+30 200 +243 - - ---- 3/1/1/1/0 1/0 \
9096 "GET /image.iso HTTP/1.0"
9097
Willy Tarreauc57f0e22009-05-10 13:12:33 +02009098 See also : "option httplog", "capture response header", and section 8 about
Willy Tarreauc27debf2008-01-06 08:57:02 +01009099 logging.
9100
9101
Christopher Faulet62f79fe2020-05-18 18:13:03 +02009102option mysql-check [ user <username> [ { post-41 | pre-41 } ] ]
Hervé COMMOWICK8776f1b2010-10-18 15:58:36 +02009103 Use MySQL health checks for server testing
Hervé COMMOWICK698ae002010-01-12 09:25:13 +01009104 May be used in sections : defaults | frontend | listen | backend
9105 yes | no | yes | yes
Hervé COMMOWICK8776f1b2010-10-18 15:58:36 +02009106 Arguments :
Cyril Bonté108cf6e2012-04-21 23:30:29 +02009107 <username> This is the username which will be used when connecting to MySQL
9108 server.
Christopher Faulet62f79fe2020-05-18 18:13:03 +02009109 post-41 Send post v4.1 client compatible checks (the default)
9110 pre-41 Send pre v4.1 client compatible checks
Hervé COMMOWICK8776f1b2010-10-18 15:58:36 +02009111
9112 If you specify a username, the check consists of sending two MySQL packet,
9113 one Client Authentication packet, and one QUIT packet, to correctly close
Davor Ocelice9ed2812017-12-25 17:49:28 +01009114 MySQL session. We then parse the MySQL Handshake Initialization packet and/or
Hervé COMMOWICK8776f1b2010-10-18 15:58:36 +02009115 Error packet. It is a basic but useful test which does not produce error nor
Daniel Black9dc310d2021-07-01 12:09:32 +10009116 aborted connect on the server. However, it requires an unlocked authorised
9117 user without a password. To create a basic limited user in MySQL with
9118 optional resource limits:
Hervé COMMOWICK8776f1b2010-10-18 15:58:36 +02009119
Daniel Black9dc310d2021-07-01 12:09:32 +10009120 CREATE USER '<username>'@'<ip_of_haproxy|network_of_haproxy/netmask>'
9121 /*!50701 WITH MAX_QUERIES_PER_HOUR 1 MAX_UPDATES_PER_HOUR 0 */
9122 /*M!100201 MAX_STATEMENT_TIME 0.0001 */;
Hervé COMMOWICK8776f1b2010-10-18 15:58:36 +02009123
9124 If you don't specify a username (it is deprecated and not recommended), the
Davor Ocelice9ed2812017-12-25 17:49:28 +01009125 check only consists in parsing the Mysql Handshake Initialization packet or
Hervé COMMOWICK8776f1b2010-10-18 15:58:36 +02009126 Error packet, we don't send anything in this mode. It was reported that it
9127 can generate lockout if check is too frequent and/or if there is not enough
9128 traffic. In fact, you need in this case to check MySQL "max_connect_errors"
9129 value as if a connection is established successfully within fewer than MySQL
9130 "max_connect_errors" attempts after a previous connection was interrupted,
9131 the error count for the host is cleared to zero. If HAProxy's server get
9132 blocked, the "FLUSH HOSTS" statement is the only way to unblock it.
9133
9134 Remember that this does not check database presence nor database consistency.
9135 To do this, you can use an external check with xinetd for example.
Hervé COMMOWICK698ae002010-01-12 09:25:13 +01009136
Hervé COMMOWICK212f7782011-06-10 14:05:59 +02009137 The check requires MySQL >=3.22, for older version, please use TCP check.
Hervé COMMOWICK698ae002010-01-12 09:25:13 +01009138
9139 Most often, an incoming MySQL server needs to see the client's IP address for
9140 various purposes, including IP privilege matching and connection logging.
9141 When possible, it is often wise to masquerade the client's IP address when
9142 connecting to the server using the "usesrc" argument of the "source" keyword,
Willy Tarreau29fbe512015-08-20 19:35:14 +02009143 which requires the transparent proxy feature to be compiled in, and the MySQL
Daniel Corbett9f0843f2021-05-08 10:50:37 -04009144 server to route the client via the machine hosting HAProxy.
Hervé COMMOWICK698ae002010-01-12 09:25:13 +01009145
9146 See also: "option httpchk"
9147
9148
Willy Tarreaua453bdd2008-01-08 19:50:52 +01009149option nolinger
9150no option nolinger
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01009151 Enable or disable immediate session resource cleaning after close
Willy Tarreaua453bdd2008-01-08 19:50:52 +01009152 May be used in sections: defaults | frontend | listen | backend
9153 yes | yes | yes | yes
Willy Tarreaueabeafa2008-01-16 16:17:06 +01009154 Arguments : none
Willy Tarreaua453bdd2008-01-08 19:50:52 +01009155
Davor Ocelice9ed2812017-12-25 17:49:28 +01009156 When clients or servers abort connections in a dirty way (e.g. they are
Willy Tarreaua453bdd2008-01-08 19:50:52 +01009157 physically disconnected), the session timeouts triggers and the session is
9158 closed. But it will remain in FIN_WAIT1 state for some time in the system,
9159 using some resources and possibly limiting the ability to establish newer
9160 connections.
9161
9162 When this happens, it is possible to activate "option nolinger" which forces
9163 the system to immediately remove any socket's pending data on close. Thus,
Willy Tarreau4a321032020-10-16 04:55:19 +02009164 a TCP RST is emitted, any pending data are truncated, and the session is
9165 instantly purged from the system's tables. The generally visible effect for
9166 a client is that responses are truncated if the close happens with a last
9167 block of data (e.g. on a redirect or error response). On the server side,
9168 it may help release the source ports immediately when forwarding a client
9169 aborts in tunnels. In both cases, TCP resets are emitted and given that
9170 the session is instantly destroyed, there will be no retransmit. On a lossy
9171 network this can increase problems, especially when there is a firewall on
9172 the lossy side, because the firewall might see and process the reset (hence
9173 purge its session) and block any further traffic for this session,, including
9174 retransmits from the other side. So if the other side doesn't receive it,
9175 it will never receive any RST again, and the firewall might log many blocked
9176 packets.
Willy Tarreaua453bdd2008-01-08 19:50:52 +01009177
Willy Tarreau4a321032020-10-16 04:55:19 +02009178 For all these reasons, it is strongly recommended NOT to use this option,
9179 unless absolutely needed as a last resort. In most situations, using the
9180 "client-fin" or "server-fin" timeouts achieves similar results with a more
9181 reliable behavior. On Linux it's also possible to use the "tcp-ut" bind or
9182 server setting.
Willy Tarreaua453bdd2008-01-08 19:50:52 +01009183
9184 This option may be used both on frontends and backends, depending on the side
9185 where it is required. Use it on the frontend for clients, and on the backend
Willy Tarreau4a321032020-10-16 04:55:19 +02009186 for servers. While this option is technically supported in "defaults"
Ilya Shipitsin2272d8a2020-12-21 01:22:40 +05009187 sections, it must really not be used there as it risks to accidentally
Willy Tarreau4a321032020-10-16 04:55:19 +02009188 propagate to sections that must no use it and to cause problems there.
Willy Tarreaua453bdd2008-01-08 19:50:52 +01009189
9190 If this option has been enabled in a "defaults" section, it can be disabled
9191 in a specific instance by prepending the "no" keyword before it.
9192
Willy Tarreau4a321032020-10-16 04:55:19 +02009193 See also: "timeout client-fin", "timeout server-fin", "tcp-ut" bind or server
9194 keywords.
Willy Tarreaua453bdd2008-01-08 19:50:52 +01009195
Willy Tarreau55165fe2009-05-10 12:02:55 +02009196option originalto [ except <network> ] [ header <name> ]
9197 Enable insertion of the X-Original-To header to requests sent to servers
9198 May be used in sections : defaults | frontend | listen | backend
9199 yes | yes | yes | yes
9200 Arguments :
9201 <network> is an optional argument used to disable this option for sources
9202 matching <network>
9203 <name> an optional argument to specify a different "X-Original-To"
9204 header name.
9205
9206 Since HAProxy can work in transparent mode, every request from a client can
9207 be redirected to the proxy and HAProxy itself can proxy every request to a
9208 complex SQUID environment and the destination host from SO_ORIGINAL_DST will
9209 be lost. This is annoying when you want access rules based on destination ip
9210 addresses. To solve this problem, a new HTTP header "X-Original-To" may be
9211 added by HAProxy to all requests sent to the server. This header contains a
9212 value representing the original destination IP address. Since this must be
9213 configured to always use the last occurrence of this header only. Note that
9214 only the last occurrence of the header must be used, since it is really
9215 possible that the client has already brought one.
9216
9217 The keyword "header" may be used to supply a different header name to replace
9218 the default "X-Original-To". This can be useful where you might already
9219 have a "X-Original-To" header from a different application, and you need
9220 preserve it. Also if your backend server doesn't use the "X-Original-To"
9221 header and requires different one.
9222
9223 Sometimes, a same HAProxy instance may be shared between a direct client
9224 access and a reverse-proxy access (for instance when an SSL reverse-proxy is
9225 used to decrypt HTTPS traffic). It is possible to disable the addition of the
Amaury Denoyellef8b42922021-03-04 18:41:14 +01009226 header for a known destination address or network by adding the "except"
9227 keyword followed by the network address. In this case, any destination IP
9228 matching the network will not cause an addition of this header. Most common
9229 uses are with private networks or 127.0.0.1. IPv4 and IPv6 are both
9230 supported.
Willy Tarreau55165fe2009-05-10 12:02:55 +02009231
9232 This option may be specified either in the frontend or in the backend. If at
9233 least one of them uses it, the header will be added. Note that the backend's
9234 setting of the header subargument takes precedence over the frontend's if
9235 both are defined.
9236
Willy Tarreau55165fe2009-05-10 12:02:55 +02009237 Examples :
9238 # Original Destination address
9239 frontend www
9240 mode http
9241 option originalto except 127.0.0.1
9242
9243 # Those servers want the IP Address in X-Client-Dst
9244 backend www
9245 mode http
9246 option originalto header X-Client-Dst
9247
Christopher Faulet315b39c2018-09-21 16:26:19 +02009248 See also : "option httpclose", "option http-server-close".
Willy Tarreau55165fe2009-05-10 12:02:55 +02009249
9250
Willy Tarreaua453bdd2008-01-08 19:50:52 +01009251option persist
9252no option persist
9253 Enable or disable forced persistence on down servers
9254 May be used in sections: defaults | frontend | listen | backend
9255 yes | no | yes | yes
Willy Tarreaueabeafa2008-01-16 16:17:06 +01009256 Arguments : none
Willy Tarreaua453bdd2008-01-08 19:50:52 +01009257
9258 When an HTTP request reaches a backend with a cookie which references a dead
9259 server, by default it is redispatched to another server. It is possible to
9260 force the request to be sent to the dead server first using "option persist"
9261 if absolutely needed. A common use case is when servers are under extreme
9262 load and spend their time flapping. In this case, the users would still be
9263 directed to the server they opened the session on, in the hope they would be
9264 correctly served. It is recommended to use "option redispatch" in conjunction
9265 with this option so that in the event it would not be possible to connect to
9266 the server at all (server definitely dead), the client would finally be
9267 redirected to another valid server.
9268
9269 If this option has been enabled in a "defaults" section, it can be disabled
9270 in a specific instance by prepending the "no" keyword before it.
9271
Willy Tarreau4de91492010-01-22 19:10:05 +01009272 See also : "option redispatch", "retries", "force-persist"
Willy Tarreaua453bdd2008-01-08 19:50:52 +01009273
9274
Willy Tarreau0c122822013-12-15 18:49:01 +01009275option pgsql-check [ user <username> ]
9276 Use PostgreSQL health checks for server testing
9277 May be used in sections : defaults | frontend | listen | backend
9278 yes | no | yes | yes
9279 Arguments :
9280 <username> This is the username which will be used when connecting to
9281 PostgreSQL server.
9282
9283 The check sends a PostgreSQL StartupMessage and waits for either
9284 Authentication request or ErrorResponse message. It is a basic but useful
9285 test which does not produce error nor aborted connect on the server.
9286 This check is identical with the "mysql-check".
9287
9288 See also: "option httpchk"
9289
9290
Willy Tarreau9420b122013-12-15 18:58:25 +01009291option prefer-last-server
9292no option prefer-last-server
9293 Allow multiple load balanced requests to remain on the same server
9294 May be used in sections: defaults | frontend | listen | backend
9295 yes | no | yes | yes
9296 Arguments : none
9297
9298 When the load balancing algorithm in use is not deterministic, and a previous
Daniel Corbett9f0843f2021-05-08 10:50:37 -04009299 request was sent to a server to which HAProxy still holds a connection, it is
Willy Tarreau9420b122013-12-15 18:58:25 +01009300 sometimes desirable that subsequent requests on a same session go to the same
9301 server as much as possible. Note that this is different from persistence, as
Daniel Corbett9f0843f2021-05-08 10:50:37 -04009302 we only indicate a preference which HAProxy tries to apply without any form
Willy Tarreau9420b122013-12-15 18:58:25 +01009303 of warranty. The real use is for keep-alive connections sent to servers. When
Daniel Corbett9f0843f2021-05-08 10:50:37 -04009304 this option is used, HAProxy will try to reuse the same connection that is
Willy Tarreau9420b122013-12-15 18:58:25 +01009305 attached to the server instead of rebalancing to another server, causing a
9306 close of the connection. This can make sense for static file servers. It does
Willy Tarreau068621e2013-12-23 15:11:25 +01009307 not make much sense to use this in combination with hashing algorithms. Note,
Daniel Corbett9f0843f2021-05-08 10:50:37 -04009308 HAProxy already automatically tries to stick to a server which sends a 401 or
Lukas Tribus80512b12018-10-27 20:07:40 +02009309 to a proxy which sends a 407 (authentication required), when the load
9310 balancing algorithm is not deterministic. This is mandatory for use with the
9311 broken NTLM authentication challenge, and significantly helps in
Willy Tarreau068621e2013-12-23 15:11:25 +01009312 troubleshooting some faulty applications. Option prefer-last-server might be
9313 desirable in these environments as well, to avoid redistributing the traffic
9314 after every other response.
Willy Tarreau9420b122013-12-15 18:58:25 +01009315
9316 If this option has been enabled in a "defaults" section, it can be disabled
9317 in a specific instance by prepending the "no" keyword before it.
9318
9319 See also: "option http-keep-alive"
9320
9321
Krzysztof Piotr Oledzki25b501a2008-01-06 16:36:16 +01009322option redispatch
Joseph Lynch726ab712015-05-11 23:25:34 -07009323option redispatch <interval>
Krzysztof Piotr Oledzki25b501a2008-01-06 16:36:16 +01009324no option redispatch
9325 Enable or disable session redistribution in case of connection failure
9326 May be used in sections: defaults | frontend | listen | backend
9327 yes | no | yes | yes
Joseph Lynch726ab712015-05-11 23:25:34 -07009328 Arguments :
9329 <interval> The optional integer value that controls how often redispatches
9330 occur when retrying connections. Positive value P indicates a
9331 redispatch is desired on every Pth retry, and negative value
Davor Ocelice9ed2812017-12-25 17:49:28 +01009332 N indicate a redispatch is desired on the Nth retry prior to the
Joseph Lynch726ab712015-05-11 23:25:34 -07009333 last retry. For example, the default of -1 preserves the
Davor Ocelice9ed2812017-12-25 17:49:28 +01009334 historical behavior of redispatching on the last retry, a
Joseph Lynch726ab712015-05-11 23:25:34 -07009335 positive value of 1 would indicate a redispatch on every retry,
9336 and a positive value of 3 would indicate a redispatch on every
9337 third retry. You can disable redispatches with a value of 0.
9338
Krzysztof Piotr Oledzki25b501a2008-01-06 16:36:16 +01009339
9340 In HTTP mode, if a server designated by a cookie is down, clients may
9341 definitely stick to it because they cannot flush the cookie, so they will not
9342 be able to access the service anymore.
9343
Willy Tarreau59884a62019-01-02 14:48:31 +01009344 Specifying "option redispatch" will allow the proxy to break cookie or
9345 consistent hash based persistence and redistribute them to a working server.
Krzysztof Piotr Oledzki25b501a2008-01-06 16:36:16 +01009346
Olivier Carrère6e6f59b2020-04-15 11:30:18 +02009347 Active servers are selected from a subset of the list of available
9348 servers. Active servers that are not down or in maintenance (i.e., whose
9349 health is not checked or that have been checked as "up"), are selected in the
9350 following order:
9351
9352 1. Any active, non-backup server, if any, or,
9353
9354 2. If the "allbackups" option is not set, the first backup server in the
9355 list, or
9356
9357 3. If the "allbackups" option is set, any backup server.
9358
9359 When a retry occurs, HAProxy tries to select another server than the last
9360 one. The new server is selected from the current list of servers.
9361
9362 Sometimes, if the list is updated between retries (e.g., if numerous retries
9363 occur and last longer than the time needed to check that a server is down,
9364 remove it from the list and fall back on the list of backup servers),
9365 connections may be redirected to a backup server, though.
9366
Joseph Lynch726ab712015-05-11 23:25:34 -07009367 It also allows to retry connections to another server in case of multiple
Krzysztof Piotr Oledzki25b501a2008-01-06 16:36:16 +01009368 connection failures. Of course, it requires having "retries" set to a nonzero
9369 value.
Willy Tarreaud72758d2010-01-12 10:42:19 +01009370
Krzysztof Piotr Oledzki25b501a2008-01-06 16:36:16 +01009371 If this option has been enabled in a "defaults" section, it can be disabled
9372 in a specific instance by prepending the "no" keyword before it.
9373
Christopher Faulet87f1f3d2019-07-18 14:51:20 +02009374 See also : "retries", "force-persist"
Krzysztof Piotr Oledzki25b501a2008-01-06 16:36:16 +01009375
Willy Tarreaua453bdd2008-01-08 19:50:52 +01009376
Hervé COMMOWICKec032d62011-08-05 16:23:48 +02009377option redis-check
9378 Use redis health checks for server testing
9379 May be used in sections : defaults | frontend | listen | backend
9380 yes | no | yes | yes
9381 Arguments : none
9382
9383 It is possible to test that the server correctly talks REDIS protocol instead
9384 of just testing that it accepts the TCP connection. When this option is set,
9385 a PING redis command is sent to the server, and the response is analyzed to
9386 find the "+PONG" response message.
9387
9388 Example :
9389 option redis-check
9390
Jarno Huuskonene5ae7022017-04-03 14:36:21 +03009391 See also : "option httpchk", "option tcp-check", "tcp-check expect"
Hervé COMMOWICKec032d62011-08-05 16:23:48 +02009392
9393
Willy Tarreaua453bdd2008-01-08 19:50:52 +01009394option smtpchk
9395option smtpchk <hello> <domain>
9396 Use SMTP health checks for server testing
9397 May be used in sections : defaults | frontend | listen | backend
9398 yes | no | yes | yes
Willy Tarreaud72758d2010-01-12 10:42:19 +01009399 Arguments :
Willy Tarreaua453bdd2008-01-08 19:50:52 +01009400 <hello> is an optional argument. It is the "hello" command to use. It can
Lukas Tribus27935782018-10-01 02:00:16 +02009401 be either "HELO" (for SMTP) or "EHLO" (for ESMTP). All other
Willy Tarreaua453bdd2008-01-08 19:50:52 +01009402 values will be turned into the default command ("HELO").
9403
9404 <domain> is the domain name to present to the server. It may only be
9405 specified (and is mandatory) if the hello command has been
9406 specified. By default, "localhost" is used.
9407
9408 When "option smtpchk" is set, the health checks will consist in TCP
9409 connections followed by an SMTP command. By default, this command is
9410 "HELO localhost". The server's return code is analyzed and only return codes
9411 starting with a "2" will be considered as valid. All other responses,
9412 including a lack of response will constitute an error and will indicate a
9413 dead server.
9414
9415 This test is meant to be used with SMTP servers or relays. Depending on the
9416 request, it is possible that some servers do not log each connection attempt,
Davor Ocelice9ed2812017-12-25 17:49:28 +01009417 so you may want to experiment to improve the behavior. Using telnet on port
Willy Tarreaua453bdd2008-01-08 19:50:52 +01009418 25 is often easier than adjusting the configuration.
9419
9420 Most often, an incoming SMTP server needs to see the client's IP address for
9421 various purposes, including spam filtering, anti-spoofing and logging. When
9422 possible, it is often wise to masquerade the client's IP address when
9423 connecting to the server using the "usesrc" argument of the "source" keyword,
Willy Tarreau29fbe512015-08-20 19:35:14 +02009424 which requires the transparent proxy feature to be compiled in.
Willy Tarreaua453bdd2008-01-08 19:50:52 +01009425
9426 Example :
9427 option smtpchk HELO mydomain.org
9428
9429 See also : "option httpchk", "source"
9430
Krzysztof Piotr Oledzki25b501a2008-01-06 16:36:16 +01009431
Krzysztof Piotr Oledzkiaeebf9b2009-10-04 15:43:17 +02009432option socket-stats
9433no option socket-stats
9434
9435 Enable or disable collecting & providing separate statistics for each socket.
9436 May be used in sections : defaults | frontend | listen | backend
9437 yes | yes | yes | no
9438
9439 Arguments : none
9440
9441
Willy Tarreauff4f82d2009-02-06 11:28:13 +01009442option splice-auto
9443no option splice-auto
9444 Enable or disable automatic kernel acceleration on sockets in both directions
9445 May be used in sections : defaults | frontend | listen | backend
9446 yes | yes | yes | yes
9447 Arguments : none
9448
Daniel Corbett9f0843f2021-05-08 10:50:37 -04009449 When this option is enabled either on a frontend or on a backend, HAProxy
Willy Tarreauff4f82d2009-02-06 11:28:13 +01009450 will automatically evaluate the opportunity to use kernel tcp splicing to
Davor Ocelice9ed2812017-12-25 17:49:28 +01009451 forward data between the client and the server, in either direction. HAProxy
Willy Tarreauff4f82d2009-02-06 11:28:13 +01009452 uses heuristics to estimate if kernel splicing might improve performance or
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01009453 not. Both directions are handled independently. Note that the heuristics used
Willy Tarreauff4f82d2009-02-06 11:28:13 +01009454 are not much aggressive in order to limit excessive use of splicing. This
9455 option requires splicing to be enabled at compile time, and may be globally
9456 disabled with the global option "nosplice". Since splice uses pipes, using it
9457 requires that there are enough spare pipes.
9458
9459 Important note: kernel-based TCP splicing is a Linux-specific feature which
9460 first appeared in kernel 2.6.25. It offers kernel-based acceleration to
9461 transfer data between sockets without copying these data to user-space, thus
9462 providing noticeable performance gains and CPU cycles savings. Since many
9463 early implementations are buggy, corrupt data and/or are inefficient, this
9464 feature is not enabled by default, and it should be used with extreme care.
9465 While it is not possible to detect the correctness of an implementation,
9466 2.6.29 is the first version offering a properly working implementation. In
9467 case of doubt, splicing may be globally disabled using the global "nosplice"
9468 keyword.
9469
9470 Example :
9471 option splice-auto
9472
9473 If this option has been enabled in a "defaults" section, it can be disabled
9474 in a specific instance by prepending the "no" keyword before it.
9475
9476 See also : "option splice-request", "option splice-response", and global
9477 options "nosplice" and "maxpipes"
9478
9479
9480option splice-request
9481no option splice-request
9482 Enable or disable automatic kernel acceleration on sockets for requests
9483 May be used in sections : defaults | frontend | listen | backend
9484 yes | yes | yes | yes
9485 Arguments : none
9486
Daniel Corbett9f0843f2021-05-08 10:50:37 -04009487 When this option is enabled either on a frontend or on a backend, HAProxy
Jamie Gloudonaaa21002012-08-25 00:18:33 -04009488 will use kernel tcp splicing whenever possible to forward data going from
Willy Tarreauff4f82d2009-02-06 11:28:13 +01009489 the client to the server. It might still use the recv/send scheme if there
9490 are no spare pipes left. This option requires splicing to be enabled at
9491 compile time, and may be globally disabled with the global option "nosplice".
9492 Since splice uses pipes, using it requires that there are enough spare pipes.
9493
9494 Important note: see "option splice-auto" for usage limitations.
9495
9496 Example :
9497 option splice-request
9498
9499 If this option has been enabled in a "defaults" section, it can be disabled
9500 in a specific instance by prepending the "no" keyword before it.
9501
9502 See also : "option splice-auto", "option splice-response", and global options
9503 "nosplice" and "maxpipes"
9504
9505
9506option splice-response
9507no option splice-response
9508 Enable or disable automatic kernel acceleration on sockets for responses
9509 May be used in sections : defaults | frontend | listen | backend
9510 yes | yes | yes | yes
9511 Arguments : none
9512
Daniel Corbett9f0843f2021-05-08 10:50:37 -04009513 When this option is enabled either on a frontend or on a backend, HAProxy
Jamie Gloudonaaa21002012-08-25 00:18:33 -04009514 will use kernel tcp splicing whenever possible to forward data going from
Willy Tarreauff4f82d2009-02-06 11:28:13 +01009515 the server to the client. It might still use the recv/send scheme if there
9516 are no spare pipes left. This option requires splicing to be enabled at
9517 compile time, and may be globally disabled with the global option "nosplice".
9518 Since splice uses pipes, using it requires that there are enough spare pipes.
9519
9520 Important note: see "option splice-auto" for usage limitations.
9521
9522 Example :
9523 option splice-response
9524
9525 If this option has been enabled in a "defaults" section, it can be disabled
9526 in a specific instance by prepending the "no" keyword before it.
9527
9528 See also : "option splice-auto", "option splice-request", and global options
9529 "nosplice" and "maxpipes"
9530
9531
Christopher Fauletba7bc162016-11-07 21:07:38 +01009532option spop-check
9533 Use SPOP health checks for server testing
9534 May be used in sections : defaults | frontend | listen | backend
9535 no | no | no | yes
9536 Arguments : none
9537
9538 It is possible to test that the server correctly talks SPOP protocol instead
9539 of just testing that it accepts the TCP connection. When this option is set,
9540 a HELLO handshake is performed between HAProxy and the server, and the
9541 response is analyzed to check no error is reported.
9542
9543 Example :
9544 option spop-check
9545
9546 See also : "option httpchk"
9547
9548
Willy Tarreaubf1f8162007-12-28 17:42:56 +01009549option srvtcpka
9550no option srvtcpka
9551 Enable or disable the sending of TCP keepalive packets on the server side
9552 May be used in sections : defaults | frontend | listen | backend
9553 yes | no | yes | yes
9554 Arguments : none
9555
9556 When there is a firewall or any session-aware component between a client and
9557 a server, and when the protocol involves very long sessions with long idle
Davor Ocelice9ed2812017-12-25 17:49:28 +01009558 periods (e.g. remote desktops), there is a risk that one of the intermediate
Willy Tarreaubf1f8162007-12-28 17:42:56 +01009559 components decides to expire a session which has remained idle for too long.
9560
9561 Enabling socket-level TCP keep-alives makes the system regularly send packets
9562 to the other end of the connection, leaving it active. The delay between
9563 keep-alive probes is controlled by the system only and depends both on the
9564 operating system and its tuning parameters.
9565
9566 It is important to understand that keep-alive packets are neither emitted nor
9567 received at the application level. It is only the network stacks which sees
9568 them. For this reason, even if one side of the proxy already uses keep-alives
9569 to maintain its connection alive, those keep-alive packets will not be
9570 forwarded to the other side of the proxy.
9571
9572 Please note that this has nothing to do with HTTP keep-alive.
9573
9574 Using option "srvtcpka" enables the emission of TCP keep-alive probes on the
9575 server side of a connection, which should help when session expirations are
9576 noticed between HAProxy and a server.
9577
9578 If this option has been enabled in a "defaults" section, it can be disabled
9579 in a specific instance by prepending the "no" keyword before it.
9580
9581 See also : "option clitcpka", "option tcpka"
9582
9583
Willy Tarreaua453bdd2008-01-08 19:50:52 +01009584option ssl-hello-chk
9585 Use SSLv3 client hello health checks for server testing
9586 May be used in sections : defaults | frontend | listen | backend
9587 yes | no | yes | yes
9588 Arguments : none
9589
9590 When some SSL-based protocols are relayed in TCP mode through HAProxy, it is
9591 possible to test that the server correctly talks SSL instead of just testing
9592 that it accepts the TCP connection. When "option ssl-hello-chk" is set, pure
9593 SSLv3 client hello messages are sent once the connection is established to
9594 the server, and the response is analyzed to find an SSL server hello message.
9595 The server is considered valid only when the response contains this server
9596 hello message.
9597
9598 All servers tested till there correctly reply to SSLv3 client hello messages,
9599 and most servers tested do not even log the requests containing only hello
9600 messages, which is appreciable.
9601
Daniel Corbett9f0843f2021-05-08 10:50:37 -04009602 Note that this check works even when SSL support was not built into HAProxy
Willy Tarreau763a95b2012-10-04 23:15:39 +02009603 because it forges the SSL message. When SSL support is available, it is best
9604 to use native SSL health checks instead of this one.
Willy Tarreaua453bdd2008-01-08 19:50:52 +01009605
Willy Tarreau763a95b2012-10-04 23:15:39 +02009606 See also: "option httpchk", "check-ssl"
9607
Willy Tarreaua453bdd2008-01-08 19:50:52 +01009608
Willy Tarreaued179852013-12-16 01:07:00 +01009609option tcp-check
9610 Perform health checks using tcp-check send/expect sequences
9611 May be used in sections: defaults | frontend | listen | backend
9612 yes | no | yes | yes
9613
9614 This health check method is intended to be combined with "tcp-check" command
9615 lists in order to support send/expect types of health check sequences.
9616
9617 TCP checks currently support 4 modes of operations :
9618 - no "tcp-check" directive : the health check only consists in a connection
9619 attempt, which remains the default mode.
9620
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03009621 - "tcp-check send" or "tcp-check send-binary" only is mentioned : this is
Willy Tarreaued179852013-12-16 01:07:00 +01009622 used to send a string along with a connection opening. With some
9623 protocols, it helps sending a "QUIT" message for example that prevents
9624 the server from logging a connection error for each health check. The
9625 check result will still be based on the ability to open the connection
9626 only.
9627
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03009628 - "tcp-check expect" only is mentioned : this is used to test a banner.
Daniel Corbett9f0843f2021-05-08 10:50:37 -04009629 The connection is opened and HAProxy waits for the server to present some
Willy Tarreaued179852013-12-16 01:07:00 +01009630 contents which must validate some rules. The check result will be based
9631 on the matching between the contents and the rules. This is suited for
9632 POP, IMAP, SMTP, FTP, SSH, TELNET.
9633
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03009634 - both "tcp-check send" and "tcp-check expect" are mentioned : this is
Davor Ocelice9ed2812017-12-25 17:49:28 +01009635 used to test a hello-type protocol. HAProxy sends a message, the server
9636 responds and its response is analyzed. the check result will be based on
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03009637 the matching between the response contents and the rules. This is often
Willy Tarreaued179852013-12-16 01:07:00 +01009638 suited for protocols which require a binding or a request/response model.
9639 LDAP, MySQL, Redis and SSL are example of such protocols, though they
9640 already all have their dedicated checks with a deeper understanding of
9641 the respective protocols.
9642 In this mode, many questions may be sent and many answers may be
Davor Ocelice9ed2812017-12-25 17:49:28 +01009643 analyzed.
Willy Tarreaued179852013-12-16 01:07:00 +01009644
Christopher Fauletc52ea4d2020-04-23 15:43:35 +02009645 A fifth mode can be used to insert comments in different steps of the script.
Baptiste Assmannd60a9e52015-04-25 16:27:23 +02009646
Christopher Fauletc52ea4d2020-04-23 15:43:35 +02009647 For each tcp-check rule you create, you can add a "comment" directive,
9648 followed by a string. This string will be reported in the log and stderr in
9649 debug mode. It is useful to make user-friendly error reporting. The
9650 "comment" is of course optional.
Baptiste Assmannd60a9e52015-04-25 16:27:23 +02009651
Christopher Fauletc52ea4d2020-04-23 15:43:35 +02009652 During the execution of a health check, a variable scope is made available to
9653 store data samples, using the "tcp-check set-var" operation. Freeing those
9654 variable is possible using "tcp-check unset-var".
Gaetan Rivet0c39ecc2020-02-24 17:34:11 +01009655
Baptiste Assmannd60a9e52015-04-25 16:27:23 +02009656
Willy Tarreaued179852013-12-16 01:07:00 +01009657 Examples :
Davor Ocelice9ed2812017-12-25 17:49:28 +01009658 # perform a POP check (analyze only server's banner)
Willy Tarreaued179852013-12-16 01:07:00 +01009659 option tcp-check
Baptiste Assmannd60a9e52015-04-25 16:27:23 +02009660 tcp-check expect string +OK\ POP3\ ready comment POP\ protocol
Willy Tarreaued179852013-12-16 01:07:00 +01009661
Davor Ocelice9ed2812017-12-25 17:49:28 +01009662 # perform an IMAP check (analyze only server's banner)
Willy Tarreaued179852013-12-16 01:07:00 +01009663 option tcp-check
Baptiste Assmannd60a9e52015-04-25 16:27:23 +02009664 tcp-check expect string *\ OK\ IMAP4\ ready comment IMAP\ protocol
Willy Tarreaued179852013-12-16 01:07:00 +01009665
9666 # look for the redis master server after ensuring it speaks well
9667 # redis protocol, then it exits properly.
Davor Ocelice9ed2812017-12-25 17:49:28 +01009668 # (send a command then analyze the response 3 times)
Willy Tarreaued179852013-12-16 01:07:00 +01009669 option tcp-check
Baptiste Assmannd60a9e52015-04-25 16:27:23 +02009670 tcp-check comment PING\ phase
Willy Tarreaued179852013-12-16 01:07:00 +01009671 tcp-check send PING\r\n
Baptiste Assmanna3322992015-08-04 10:12:18 +02009672 tcp-check expect string +PONG
Baptiste Assmannd60a9e52015-04-25 16:27:23 +02009673 tcp-check comment role\ check
Willy Tarreaued179852013-12-16 01:07:00 +01009674 tcp-check send info\ replication\r\n
9675 tcp-check expect string role:master
Baptiste Assmannd60a9e52015-04-25 16:27:23 +02009676 tcp-check comment QUIT\ phase
Willy Tarreaued179852013-12-16 01:07:00 +01009677 tcp-check send QUIT\r\n
9678 tcp-check expect string +OK
9679
Davor Ocelice9ed2812017-12-25 17:49:28 +01009680 forge a HTTP request, then analyze the response
Willy Tarreaued179852013-12-16 01:07:00 +01009681 (send many headers before analyzing)
9682 option tcp-check
Baptiste Assmannd60a9e52015-04-25 16:27:23 +02009683 tcp-check comment forge\ and\ send\ HTTP\ request
Willy Tarreaued179852013-12-16 01:07:00 +01009684 tcp-check send HEAD\ /\ HTTP/1.1\r\n
9685 tcp-check send Host:\ www.mydomain.com\r\n
9686 tcp-check send User-Agent:\ HAProxy\ tcpcheck\r\n
9687 tcp-check send \r\n
Baptiste Assmannd60a9e52015-04-25 16:27:23 +02009688 tcp-check expect rstring HTTP/1\..\ (2..|3..) comment check\ HTTP\ response
Willy Tarreaued179852013-12-16 01:07:00 +01009689
9690
Christopher Fauletc52ea4d2020-04-23 15:43:35 +02009691 See also : "tcp-check connect", "tcp-check expect" and "tcp-check send".
Willy Tarreaued179852013-12-16 01:07:00 +01009692
9693
Willy Tarreau9ea05a72009-06-14 12:07:01 +02009694option tcp-smart-accept
9695no option tcp-smart-accept
9696 Enable or disable the saving of one ACK packet during the accept sequence
9697 May be used in sections : defaults | frontend | listen | backend
9698 yes | yes | yes | no
9699 Arguments : none
9700
9701 When an HTTP connection request comes in, the system acknowledges it on
9702 behalf of HAProxy, then the client immediately sends its request, and the
9703 system acknowledges it too while it is notifying HAProxy about the new
9704 connection. HAProxy then reads the request and responds. This means that we
9705 have one TCP ACK sent by the system for nothing, because the request could
9706 very well be acknowledged by HAProxy when it sends its response.
9707
9708 For this reason, in HTTP mode, HAProxy automatically asks the system to avoid
9709 sending this useless ACK on platforms which support it (currently at least
9710 Linux). It must not cause any problem, because the system will send it anyway
9711 after 40 ms if the response takes more time than expected to come.
9712
9713 During complex network debugging sessions, it may be desirable to disable
9714 this optimization because delayed ACKs can make troubleshooting more complex
9715 when trying to identify where packets are delayed. It is then possible to
Davor Ocelice9ed2812017-12-25 17:49:28 +01009716 fall back to normal behavior by specifying "no option tcp-smart-accept".
Willy Tarreau9ea05a72009-06-14 12:07:01 +02009717
9718 It is also possible to force it for non-HTTP proxies by simply specifying
9719 "option tcp-smart-accept". For instance, it can make sense with some services
9720 such as SMTP where the server speaks first.
9721
9722 It is recommended to avoid forcing this option in a defaults section. In case
9723 of doubt, consider setting it back to automatic values by prepending the
9724 "default" keyword before it, or disabling it using the "no" keyword.
9725
Willy Tarreaud88edf22009-06-14 15:48:17 +02009726 See also : "option tcp-smart-connect"
9727
9728
9729option tcp-smart-connect
9730no option tcp-smart-connect
9731 Enable or disable the saving of one ACK packet during the connect sequence
9732 May be used in sections : defaults | frontend | listen | backend
9733 yes | no | yes | yes
9734 Arguments : none
9735
9736 On certain systems (at least Linux), HAProxy can ask the kernel not to
9737 immediately send an empty ACK upon a connection request, but to directly
9738 send the buffer request instead. This saves one packet on the network and
9739 thus boosts performance. It can also be useful for some servers, because they
9740 immediately get the request along with the incoming connection.
9741
9742 This feature is enabled when "option tcp-smart-connect" is set in a backend.
9743 It is not enabled by default because it makes network troubleshooting more
9744 complex.
9745
9746 It only makes sense to enable it with protocols where the client speaks first
9747 such as HTTP. In other situations, if there is no data to send in place of
9748 the ACK, a normal ACK is sent.
9749
9750 If this option has been enabled in a "defaults" section, it can be disabled
9751 in a specific instance by prepending the "no" keyword before it.
9752
9753 See also : "option tcp-smart-accept"
9754
Willy Tarreau9ea05a72009-06-14 12:07:01 +02009755
Willy Tarreaubf1f8162007-12-28 17:42:56 +01009756option tcpka
9757 Enable or disable the sending of TCP keepalive packets on both sides
9758 May be used in sections : defaults | frontend | listen | backend
9759 yes | yes | yes | yes
9760 Arguments : none
9761
9762 When there is a firewall or any session-aware component between a client and
9763 a server, and when the protocol involves very long sessions with long idle
Davor Ocelice9ed2812017-12-25 17:49:28 +01009764 periods (e.g. remote desktops), there is a risk that one of the intermediate
Willy Tarreaubf1f8162007-12-28 17:42:56 +01009765 components decides to expire a session which has remained idle for too long.
9766
9767 Enabling socket-level TCP keep-alives makes the system regularly send packets
9768 to the other end of the connection, leaving it active. The delay between
9769 keep-alive probes is controlled by the system only and depends both on the
9770 operating system and its tuning parameters.
9771
9772 It is important to understand that keep-alive packets are neither emitted nor
9773 received at the application level. It is only the network stacks which sees
9774 them. For this reason, even if one side of the proxy already uses keep-alives
9775 to maintain its connection alive, those keep-alive packets will not be
9776 forwarded to the other side of the proxy.
9777
9778 Please note that this has nothing to do with HTTP keep-alive.
9779
9780 Using option "tcpka" enables the emission of TCP keep-alive probes on both
9781 the client and server sides of a connection. Note that this is meaningful
9782 only in "defaults" or "listen" sections. If this option is used in a
9783 frontend, only the client side will get keep-alives, and if this option is
9784 used in a backend, only the server side will get keep-alives. For this
9785 reason, it is strongly recommended to explicitly use "option clitcpka" and
9786 "option srvtcpka" when the configuration is split between frontends and
9787 backends.
9788
9789 See also : "option clitcpka", "option srvtcpka"
9790
Willy Tarreau844e3c52008-01-11 16:28:18 +01009791
9792option tcplog
9793 Enable advanced logging of TCP connections with session state and timers
9794 May be used in sections : defaults | frontend | listen | backend
Tim Duesterhus9ad9f352018-02-05 20:52:27 +01009795 yes | yes | yes | no
Willy Tarreau844e3c52008-01-11 16:28:18 +01009796 Arguments : none
9797
9798 By default, the log output format is very poor, as it only contains the
9799 source and destination addresses, and the instance name. By specifying
9800 "option tcplog", each log line turns into a much richer format including, but
9801 not limited to, the connection timers, the session status, the connections
9802 numbers, the frontend, backend and server name, and of course the source
9803 address and ports. This option is useful for pure TCP proxies in order to
9804 find which of the client or server disconnects or times out. For normal HTTP
9805 proxies, it's better to use "option httplog" which is even more complete.
9806
Guillaume de Lafond29f45602017-03-31 19:52:15 +02009807 "option tcplog" overrides any previous "log-format" directive.
9808
Willy Tarreauc57f0e22009-05-10 13:12:33 +02009809 See also : "option httplog", and section 8 about logging.
Willy Tarreau844e3c52008-01-11 16:28:18 +01009810
9811
Willy Tarreau844e3c52008-01-11 16:28:18 +01009812option transparent
9813no option transparent
9814 Enable client-side transparent proxying
9815 May be used in sections : defaults | frontend | listen | backend
Willy Tarreau4b1f8592008-12-23 23:13:55 +01009816 yes | no | yes | yes
Willy Tarreau844e3c52008-01-11 16:28:18 +01009817 Arguments : none
9818
9819 This option was introduced in order to provide layer 7 persistence to layer 3
9820 load balancers. The idea is to use the OS's ability to redirect an incoming
9821 connection for a remote address to a local process (here HAProxy), and let
9822 this process know what address was initially requested. When this option is
9823 used, sessions without cookies will be forwarded to the original destination
9824 IP address of the incoming request (which should match that of another
9825 equipment), while requests with cookies will still be forwarded to the
9826 appropriate server.
9827
9828 Note that contrary to a common belief, this option does NOT make HAProxy
9829 present the client's IP to the server when establishing the connection.
9830
Willy Tarreaua1146052011-03-01 09:51:54 +01009831 See also: the "usesrc" argument of the "source" keyword, and the
Willy Tarreaueabeafa2008-01-16 16:17:06 +01009832 "transparent" option of the "bind" keyword.
Willy Tarreau844e3c52008-01-11 16:28:18 +01009833
Willy Tarreaubf1f8162007-12-28 17:42:56 +01009834
Simon Horman98637e52014-06-20 12:30:16 +09009835external-check command <command>
9836 Executable to run when performing an external-check
9837 May be used in sections : defaults | frontend | listen | backend
9838 yes | no | yes | yes
9839
9840 Arguments :
9841 <command> is the external command to run
9842
Simon Horman98637e52014-06-20 12:30:16 +09009843 The arguments passed to the to the command are:
9844
Cyril Bonté777be862014-12-02 21:21:35 +01009845 <proxy_address> <proxy_port> <server_address> <server_port>
Simon Horman98637e52014-06-20 12:30:16 +09009846
Cyril Bonté777be862014-12-02 21:21:35 +01009847 The <proxy_address> and <proxy_port> are derived from the first listener
9848 that is either IPv4, IPv6 or a UNIX socket. In the case of a UNIX socket
9849 listener the proxy_address will be the path of the socket and the
9850 <proxy_port> will be the string "NOT_USED". In a backend section, it's not
9851 possible to determine a listener, and both <proxy_address> and <proxy_port>
9852 will have the string value "NOT_USED".
Simon Horman98637e52014-06-20 12:30:16 +09009853
Cyril Bonté72cda2a2014-12-27 22:28:39 +01009854 Some values are also provided through environment variables.
9855
9856 Environment variables :
9857 HAPROXY_PROXY_ADDR The first bind address if available (or empty if not
9858 applicable, for example in a "backend" section).
9859
9860 HAPROXY_PROXY_ID The backend id.
9861
9862 HAPROXY_PROXY_NAME The backend name.
9863
9864 HAPROXY_PROXY_PORT The first bind port if available (or empty if not
9865 applicable, for example in a "backend" section or
9866 for a UNIX socket).
9867
9868 HAPROXY_SERVER_ADDR The server address.
9869
9870 HAPROXY_SERVER_CURCONN The current number of connections on the server.
9871
9872 HAPROXY_SERVER_ID The server id.
9873
9874 HAPROXY_SERVER_MAXCONN The server max connections.
9875
9876 HAPROXY_SERVER_NAME The server name.
9877
9878 HAPROXY_SERVER_PORT The server port if available (or empty for a UNIX
9879 socket).
9880
9881 PATH The PATH environment variable used when executing
9882 the command may be set using "external-check path".
9883
William Lallemand4d03e432019-06-14 15:35:37 +02009884 See also "2.3. Environment variables" for other variables.
9885
Simon Horman98637e52014-06-20 12:30:16 +09009886 If the command executed and exits with a zero status then the check is
9887 considered to have passed, otherwise the check is considered to have
9888 failed.
9889
9890 Example :
9891 external-check command /bin/true
9892
9893 See also : "external-check", "option external-check", "external-check path"
9894
9895
9896external-check path <path>
9897 The value of the PATH environment variable used when running an external-check
9898 May be used in sections : defaults | frontend | listen | backend
9899 yes | no | yes | yes
9900
9901 Arguments :
9902 <path> is the path used when executing external command to run
9903
9904 The default path is "".
9905
9906 Example :
9907 external-check path "/usr/bin:/bin"
9908
9909 See also : "external-check", "option external-check",
9910 "external-check command"
9911
9912
Emeric Brun647caf12009-06-30 17:57:00 +02009913persist rdp-cookie
Hervé COMMOWICKa3eb39c2011-08-05 18:48:51 +02009914persist rdp-cookie(<name>)
Emeric Brun647caf12009-06-30 17:57:00 +02009915 Enable RDP cookie-based persistence
9916 May be used in sections : defaults | frontend | listen | backend
9917 yes | no | yes | yes
9918 Arguments :
9919 <name> is the optional name of the RDP cookie to check. If omitted, the
Willy Tarreau61e28f22010-05-16 22:31:05 +02009920 default cookie name "msts" will be used. There currently is no
9921 valid reason to change this name.
Emeric Brun647caf12009-06-30 17:57:00 +02009922
9923 This statement enables persistence based on an RDP cookie. The RDP cookie
9924 contains all information required to find the server in the list of known
Davor Ocelice9ed2812017-12-25 17:49:28 +01009925 servers. So when this option is set in the backend, the request is analyzed
Emeric Brun647caf12009-06-30 17:57:00 +02009926 and if an RDP cookie is found, it is decoded. If it matches a known server
9927 which is still UP (or if "option persist" is set), then the connection is
9928 forwarded to this server.
9929
9930 Note that this only makes sense in a TCP backend, but for this to work, the
9931 frontend must have waited long enough to ensure that an RDP cookie is present
9932 in the request buffer. This is the same requirement as with the "rdp-cookie"
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01009933 load-balancing method. Thus it is highly recommended to put all statements in
Emeric Brun647caf12009-06-30 17:57:00 +02009934 a single "listen" section.
9935
Willy Tarreau61e28f22010-05-16 22:31:05 +02009936 Also, it is important to understand that the terminal server will emit this
9937 RDP cookie only if it is configured for "token redirection mode", which means
9938 that the "IP address redirection" option is disabled.
9939
Emeric Brun647caf12009-06-30 17:57:00 +02009940 Example :
9941 listen tse-farm
9942 bind :3389
9943 # wait up to 5s for an RDP cookie in the request
9944 tcp-request inspect-delay 5s
9945 tcp-request content accept if RDP_COOKIE
9946 # apply RDP cookie persistence
9947 persist rdp-cookie
9948 # if server is unknown, let's balance on the same cookie.
Cyril Bontédc4d9032012-04-08 21:57:39 +02009949 # alternatively, "balance leastconn" may be useful too.
Emeric Brun647caf12009-06-30 17:57:00 +02009950 balance rdp-cookie
9951 server srv1 1.1.1.1:3389
9952 server srv2 1.1.1.2:3389
9953
Simon Hormanab814e02011-06-24 14:50:20 +09009954 See also : "balance rdp-cookie", "tcp-request", the "req_rdp_cookie" ACL and
9955 the rdp_cookie pattern fetch function.
Emeric Brun647caf12009-06-30 17:57:00 +02009956
9957
Willy Tarreau3a7d2072009-03-05 23:48:25 +01009958rate-limit sessions <rate>
9959 Set a limit on the number of new sessions accepted per second on a frontend
9960 May be used in sections : defaults | frontend | listen | backend
9961 yes | yes | yes | no
9962 Arguments :
9963 <rate> The <rate> parameter is an integer designating the maximum number
9964 of new sessions per second to accept on the frontend.
9965
9966 When the frontend reaches the specified number of new sessions per second, it
9967 stops accepting new connections until the rate drops below the limit again.
9968 During this time, the pending sessions will be kept in the socket's backlog
Daniel Corbett9f0843f2021-05-08 10:50:37 -04009969 (in system buffers) and HAProxy will not even be aware that sessions are
Willy Tarreau3a7d2072009-03-05 23:48:25 +01009970 pending. When applying very low limit on a highly loaded service, it may make
9971 sense to increase the socket's backlog using the "backlog" keyword.
9972
9973 This feature is particularly efficient at blocking connection-based attacks
9974 or service abuse on fragile servers. Since the session rate is measured every
9975 millisecond, it is extremely accurate. Also, the limit applies immediately,
9976 no delay is needed at all to detect the threshold.
9977
9978 Example : limit the connection rate on SMTP to 10 per second max
9979 listen smtp
9980 mode tcp
9981 bind :25
9982 rate-limit sessions 10
Panagiotis Panagiotopoulos7282d8e2016-02-11 16:37:15 +02009983 server smtp1 127.0.0.1:1025
Willy Tarreau3a7d2072009-03-05 23:48:25 +01009984
Willy Tarreaua17c2d92011-07-25 08:16:20 +02009985 Note : when the maximum rate is reached, the frontend's status is not changed
9986 but its sockets appear as "WAITING" in the statistics if the
9987 "socket-stats" option is enabled.
Willy Tarreau3a7d2072009-03-05 23:48:25 +01009988
9989 See also : the "backlog" keyword and the "fe_sess_rate" ACL criterion.
9990
9991
Willy Tarreau2e1dca82012-09-12 08:43:15 +02009992redirect location <loc> [code <code>] <option> [{if | unless} <condition>]
9993redirect prefix <pfx> [code <code>] <option> [{if | unless} <condition>]
9994redirect scheme <sch> [code <code>] <option> [{if | unless} <condition>]
Willy Tarreaub463dfb2008-06-07 23:08:56 +02009995 Return an HTTP redirection if/unless a condition is matched
9996 May be used in sections : defaults | frontend | listen | backend
9997 no | yes | yes | yes
9998
9999 If/unless the condition is matched, the HTTP request will lead to a redirect
Willy Tarreauf285f542010-01-03 20:03:03 +010010000 response. If no condition is specified, the redirect applies unconditionally.
Willy Tarreaub463dfb2008-06-07 23:08:56 +020010001
Willy Tarreau0140f252008-11-19 21:07:09 +010010002 Arguments :
Willy Tarreau2e1dca82012-09-12 08:43:15 +020010003 <loc> With "redirect location", the exact value in <loc> is placed into
Thierry FOURNIERd18cd0f2013-11-29 12:15:45 +010010004 the HTTP "Location" header. When used in an "http-request" rule,
10005 <loc> value follows the log-format rules and can include some
10006 dynamic values (see Custom Log Format in section 8.2.4).
Willy Tarreau2e1dca82012-09-12 08:43:15 +020010007
10008 <pfx> With "redirect prefix", the "Location" header is built from the
10009 concatenation of <pfx> and the complete URI path, including the
10010 query string, unless the "drop-query" option is specified (see
10011 below). As a special case, if <pfx> equals exactly "/", then
10012 nothing is inserted before the original URI. It allows one to
Thierry FOURNIERd18cd0f2013-11-29 12:15:45 +010010013 redirect to the same URL (for instance, to insert a cookie). When
10014 used in an "http-request" rule, <pfx> value follows the log-format
10015 rules and can include some dynamic values (see Custom Log Format
10016 in section 8.2.4).
Willy Tarreau2e1dca82012-09-12 08:43:15 +020010017
10018 <sch> With "redirect scheme", then the "Location" header is built by
10019 concatenating <sch> with "://" then the first occurrence of the
10020 "Host" header, and then the URI path, including the query string
10021 unless the "drop-query" option is specified (see below). If no
10022 path is found or if the path is "*", then "/" is used instead. If
10023 no "Host" header is found, then an empty host component will be
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030010024 returned, which most recent browsers interpret as redirecting to
Willy Tarreau2e1dca82012-09-12 08:43:15 +020010025 the same host. This directive is mostly used to redirect HTTP to
Thierry FOURNIERd18cd0f2013-11-29 12:15:45 +010010026 HTTPS. When used in an "http-request" rule, <sch> value follows
10027 the log-format rules and can include some dynamic values (see
10028 Custom Log Format in section 8.2.4).
Willy Tarreau0140f252008-11-19 21:07:09 +010010029
10030 <code> The code is optional. It indicates which type of HTTP redirection
Willy Tarreaub67fdc42013-03-29 19:28:11 +010010031 is desired. Only codes 301, 302, 303, 307 and 308 are supported,
10032 with 302 used by default if no code is specified. 301 means
10033 "Moved permanently", and a browser may cache the Location. 302
Baptiste Assmannea849c02015-08-03 11:42:50 +020010034 means "Moved temporarily" and means that the browser should not
Willy Tarreaub67fdc42013-03-29 19:28:11 +010010035 cache the redirection. 303 is equivalent to 302 except that the
10036 browser will fetch the location with a GET method. 307 is just
10037 like 302 but makes it clear that the same method must be reused.
10038 Likewise, 308 replaces 301 if the same method must be used.
Willy Tarreau0140f252008-11-19 21:07:09 +010010039
10040 <option> There are several options which can be specified to adjust the
Davor Ocelice9ed2812017-12-25 17:49:28 +010010041 expected behavior of a redirection :
Willy Tarreau0140f252008-11-19 21:07:09 +010010042
10043 - "drop-query"
10044 When this keyword is used in a prefix-based redirection, then the
10045 location will be set without any possible query-string, which is useful
10046 for directing users to a non-secure page for instance. It has no effect
10047 with a location-type redirect.
10048
Willy Tarreau81e3b4f2010-01-10 00:42:19 +010010049 - "append-slash"
10050 This keyword may be used in conjunction with "drop-query" to redirect
10051 users who use a URL not ending with a '/' to the same one with the '/'.
10052 It can be useful to ensure that search engines will only see one URL.
10053 For this, a return code 301 is preferred.
10054
Willy Tarreau0140f252008-11-19 21:07:09 +010010055 - "set-cookie NAME[=value]"
10056 A "Set-Cookie" header will be added with NAME (and optionally "=value")
10057 to the response. This is sometimes used to indicate that a user has
10058 been seen, for instance to protect against some types of DoS. No other
10059 cookie option is added, so the cookie will be a session cookie. Note
10060 that for a browser, a sole cookie name without an equal sign is
10061 different from a cookie with an equal sign.
10062
10063 - "clear-cookie NAME[=]"
10064 A "Set-Cookie" header will be added with NAME (and optionally "="), but
10065 with the "Max-Age" attribute set to zero. This will tell the browser to
10066 delete this cookie. It is useful for instance on logout pages. It is
10067 important to note that clearing the cookie "NAME" will not remove a
10068 cookie set with "NAME=value". You have to clear the cookie "NAME=" for
10069 that, because the browser makes the difference.
Willy Tarreaub463dfb2008-06-07 23:08:56 +020010070
10071 Example: move the login URL only to HTTPS.
10072 acl clear dst_port 80
10073 acl secure dst_port 8080
10074 acl login_page url_beg /login
Willy Tarreau0140f252008-11-19 21:07:09 +010010075 acl logout url_beg /logout
Willy Tarreau79da4692008-11-19 20:03:04 +010010076 acl uid_given url_reg /login?userid=[^&]+
Willy Tarreau0140f252008-11-19 21:07:09 +010010077 acl cookie_set hdr_sub(cookie) SEEN=1
10078
10079 redirect prefix https://mysite.com set-cookie SEEN=1 if !cookie_set
Willy Tarreau79da4692008-11-19 20:03:04 +010010080 redirect prefix https://mysite.com if login_page !secure
10081 redirect prefix http://mysite.com drop-query if login_page !uid_given
10082 redirect location http://mysite.com/ if !login_page secure
Willy Tarreau0140f252008-11-19 21:07:09 +010010083 redirect location / clear-cookie USERID= if logout
Willy Tarreaub463dfb2008-06-07 23:08:56 +020010084
Willy Tarreau81e3b4f2010-01-10 00:42:19 +010010085 Example: send redirects for request for articles without a '/'.
10086 acl missing_slash path_reg ^/article/[^/]*$
10087 redirect code 301 prefix / drop-query append-slash if missing_slash
10088
Daniel Corbett9f0843f2021-05-08 10:50:37 -040010089 Example: redirect all HTTP traffic to HTTPS when SSL is handled by HAProxy.
David BERARDe7153042012-11-03 00:11:31 +010010090 redirect scheme https if !{ ssl_fc }
Willy Tarreau2e1dca82012-09-12 08:43:15 +020010091
Thierry FOURNIERd18cd0f2013-11-29 12:15:45 +010010092 Example: append 'www.' prefix in front of all hosts not having it
Coen Rosdorff596659b2016-04-11 11:33:49 +020010093 http-request redirect code 301 location \
10094 http://www.%[hdr(host)]%[capture.req.uri] \
10095 unless { hdr_beg(host) -i www }
Thierry FOURNIERd18cd0f2013-11-29 12:15:45 +010010096
Willy Tarreauc57f0e22009-05-10 13:12:33 +020010097 See section 7 about ACL usage.
Willy Tarreaub463dfb2008-06-07 23:08:56 +020010098
Willy Tarreau303c0352008-01-17 19:01:39 +010010099
Willy Tarreaue5c5ce92008-06-20 17:27:19 +020010100retries <value>
10101 Set the number of retries to perform on a server after a connection failure
10102 May be used in sections: defaults | frontend | listen | backend
10103 yes | no | yes | yes
10104 Arguments :
10105 <value> is the number of times a connection attempt should be retried on
10106 a server when a connection either is refused or times out. The
10107 default value is 3.
10108
10109 It is important to understand that this value applies to the number of
10110 connection attempts, not full requests. When a connection has effectively
10111 been established to a server, there will be no more retry.
10112
10113 In order to avoid immediate reconnections to a server which is restarting,
Joseph Lynch726ab712015-05-11 23:25:34 -070010114 a turn-around timer of min("timeout connect", one second) is applied before
10115 a retry occurs.
Willy Tarreaue5c5ce92008-06-20 17:27:19 +020010116
10117 When "option redispatch" is set, the last retry may be performed on another
10118 server even if a cookie references a different server.
10119
10120 See also : "option redispatch"
10121
10122
Olivier Houcharda254a372019-04-05 15:30:12 +020010123retry-on [list of keywords]
Jerome Magnin5ce3c142020-05-13 20:09:57 +020010124 Specify when to attempt to automatically retry a failed request.
10125 This setting is only valid when "mode" is set to http and is silently ignored
10126 otherwise.
Olivier Houcharda254a372019-04-05 15:30:12 +020010127 May be used in sections: defaults | frontend | listen | backend
10128 yes | no | yes | yes
10129 Arguments :
10130 <keywords> is a list of keywords or HTTP status codes, each representing a
10131 type of failure event on which an attempt to retry the request
10132 is desired. Please read the notes at the bottom before changing
10133 this setting. The following keywords are supported :
10134
10135 none never retry
10136
10137 conn-failure retry when the connection or the SSL handshake failed
10138 and the request could not be sent. This is the default.
10139
10140 empty-response retry when the server connection was closed after part
10141 of the request was sent, and nothing was received from
10142 the server. This type of failure may be caused by the
10143 request timeout on the server side, poor network
10144 condition, or a server crash or restart while
10145 processing the request.
10146
Olivier Houcharde3249a92019-05-03 23:01:47 +020010147 junk-response retry when the server returned something not looking
10148 like a complete HTTP response. This includes partial
10149 responses headers as well as non-HTTP contents. It
10150 usually is a bad idea to retry on such events, which
10151 may be caused a configuration issue (wrong server port)
10152 or by the request being harmful to the server (buffer
10153 overflow attack for example).
10154
Olivier Houcharda254a372019-04-05 15:30:12 +020010155 response-timeout the server timeout stroke while waiting for the server
10156 to respond to the request. This may be caused by poor
10157 network condition, the reuse of an idle connection
10158 which has expired on the path, or by the request being
10159 extremely expensive to process. It generally is a bad
10160 idea to retry on such events on servers dealing with
10161 heavy database processing (full scans, etc) as it may
10162 amplify denial of service attacks.
10163
Olivier Houchard865d8392019-05-03 22:46:27 +020010164 0rtt-rejected retry requests which were sent over early data and were
10165 rejected by the server. These requests are generally
10166 considered to be safe to retry.
10167
Julien Pivotto2de240a2020-11-12 11:14:05 +010010168 <status> any HTTP status code among "401" (Unauthorized), "403"
10169 (Forbidden), "404" (Not Found), "408" (Request Timeout),
10170 "425" (Too Early), "500" (Server Error), "501" (Not
10171 Implemented), "502" (Bad Gateway), "503" (Service
10172 Unavailable), "504" (Gateway Timeout).
Olivier Houcharda254a372019-04-05 15:30:12 +020010173
Olivier Houchardddf0e032019-05-10 18:05:40 +020010174 all-retryable-errors
10175 retry request for any error that are considered
10176 retryable. This currently activates "conn-failure",
10177 "empty-response", "junk-response", "response-timeout",
10178 "0rtt-rejected", "500", "502", "503", and "504".
10179
Olivier Houcharda254a372019-04-05 15:30:12 +020010180 Using this directive replaces any previous settings with the new ones; it is
10181 not cumulative.
10182
10183 Please note that using anything other than "none" and "conn-failure" requires
10184 to allocate a buffer and copy the whole request into it, so it has memory and
10185 performance impacts. Requests not fitting in a single buffer will never be
10186 retried (see the global tune.bufsize setting).
10187
10188 You have to make sure the application has a replay protection mechanism built
10189 in such as a unique transaction IDs passed in requests, or that replaying the
10190 same request has no consequence, or it is very dangerous to use any retry-on
10191 value beside "conn-failure" and "none". Static file servers and caches are
10192 generally considered safe against any type of retry. Using a status code can
10193 be useful to quickly leave a server showing an abnormal behavior (out of
10194 memory, file system issues, etc), but in this case it may be a good idea to
10195 immediately redispatch the connection to another server (please see "option
10196 redispatch" for this). Last, it is important to understand that most causes
10197 of failures are the requests themselves and that retrying a request causing a
10198 server to misbehave will often make the situation even worse for this server,
10199 or for the whole service in case of redispatch.
10200
10201 Unless you know exactly how the application deals with replayed requests, you
10202 should not use this directive.
10203
10204 The default is "conn-failure".
10205
10206 See also: "retries", "option redispatch", "tune.bufsize"
10207
David du Colombier486df472011-03-17 10:40:26 +010010208server <name> <address>[:[port]] [param*]
Willy Tarreaueabeafa2008-01-16 16:17:06 +010010209 Declare a server in a backend
10210 May be used in sections : defaults | frontend | listen | backend
10211 no | no | yes | yes
10212 Arguments :
10213 <name> is the internal name assigned to this server. This name will
Davor Ocelice9ed2812017-12-25 17:49:28 +010010214 appear in logs and alerts. If "http-send-name-header" is
Mark Lamourinec2247f02012-01-04 13:02:01 -050010215 set, it will be added to the request header sent to the server.
Willy Tarreaueabeafa2008-01-16 16:17:06 +010010216
David du Colombier486df472011-03-17 10:40:26 +010010217 <address> is the IPv4 or IPv6 address of the server. Alternatively, a
10218 resolvable hostname is supported, but this name will be resolved
10219 during start-up. Address "0.0.0.0" or "*" has a special meaning.
10220 It indicates that the connection will be forwarded to the same IP
Willy Tarreaud669a4f2010-07-13 14:49:50 +020010221 address as the one from the client connection. This is useful in
10222 transparent proxy architectures where the client's connection is
Daniel Corbett9f0843f2021-05-08 10:50:37 -040010223 intercepted and HAProxy must forward to the original destination
Willy Tarreaud669a4f2010-07-13 14:49:50 +020010224 address. This is more or less what the "transparent" keyword does
10225 except that with a server it's possible to limit concurrency and
Willy Tarreau24709282013-03-10 21:32:12 +010010226 to report statistics. Optionally, an address family prefix may be
10227 used before the address to force the family regardless of the
10228 address format, which can be useful to specify a path to a unix
10229 socket with no slash ('/'). Currently supported prefixes are :
10230 - 'ipv4@' -> address is always IPv4
10231 - 'ipv6@' -> address is always IPv6
10232 - 'unix@' -> address is a path to a local unix socket
Willy Tarreauccfccef2014-05-10 01:49:15 +020010233 - 'abns@' -> address is in abstract namespace (Linux only)
William Lallemand2fe7dd02018-09-11 16:51:29 +020010234 - 'sockpair@' -> address is the FD of a connected unix
10235 socket or of a socketpair. During a connection, the
10236 backend creates a pair of connected sockets, and passes
10237 one of them over the FD. The bind part will use the
10238 received socket as the client FD. Should be used
10239 carefully.
William Lallemandb2f07452015-05-12 14:27:13 +020010240 You may want to reference some environment variables in the
10241 address parameter, see section 2.3 about environment
Willy Tarreau6a031d12016-11-07 19:42:35 +010010242 variables. The "init-addr" setting can be used to modify the way
10243 IP addresses should be resolved upon startup.
Willy Tarreaueabeafa2008-01-16 16:17:06 +010010244
Willy Tarreaub6205fd2012-09-24 12:27:33 +020010245 <port> is an optional port specification. If set, all connections will
Willy Tarreaueabeafa2008-01-16 16:17:06 +010010246 be sent to this port. If unset, the same port the client
10247 connected to will be used. The port may also be prefixed by a "+"
10248 or a "-". In this case, the server's port will be determined by
10249 adding this value to the client's port.
10250
10251 <param*> is a list of parameters for this server. The "server" keywords
10252 accepts an important number of options and has a complete section
Willy Tarreauc57f0e22009-05-10 13:12:33 +020010253 dedicated to it. Please refer to section 5 for more details.
Willy Tarreaueabeafa2008-01-16 16:17:06 +010010254
10255 Examples :
10256 server first 10.1.1.1:1080 cookie first check inter 1000
10257 server second 10.1.1.2:1080 cookie second check inter 1000
Willy Tarreau24709282013-03-10 21:32:12 +010010258 server transp ipv4@
William Lallemandb2f07452015-05-12 14:27:13 +020010259 server backup "${SRV_BACKUP}:1080" backup
10260 server www1_dc1 "${LAN_DC1}.101:80"
10261 server www1_dc2 "${LAN_DC2}.101:80"
Willy Tarreaueabeafa2008-01-16 16:17:06 +010010262
Willy Tarreau55dcaf62015-09-27 15:03:15 +020010263 Note: regarding Linux's abstract namespace sockets, HAProxy uses the whole
10264 sun_path length is used for the address length. Some other programs
10265 such as socat use the string length only by default. Pass the option
10266 ",unix-tightsocklen=0" to any abstract socket definition in socat to
10267 make it compatible with HAProxy's.
10268
Mark Lamourinec2247f02012-01-04 13:02:01 -050010269 See also: "default-server", "http-send-name-header" and section 5 about
10270 server options
Willy Tarreaueabeafa2008-01-16 16:17:06 +010010271
Christopher Faulet583b6de2021-02-12 09:27:10 +010010272server-state-file-name [ { use-backend-name | <file> } ]
Baptiste Assmann01c6cc32015-08-23 11:45:29 +020010273 Set the server state file to read, load and apply to servers available in
Christopher Faulet583b6de2021-02-12 09:27:10 +010010274 this backend.
10275 May be used in sections: defaults | frontend | listen | backend
10276 no | no | yes | yes
10277
10278 It only applies when the directive "load-server-state-from-file" is set to
10279 "local". When <file> is not provided, if "use-backend-name" is used or if
10280 this directive is not set, then backend name is used. If <file> starts with a
10281 slash '/', then it is considered as an absolute path. Otherwise, <file> is
10282 concatenated to the global directive "server-state-base".
Baptiste Assmann01c6cc32015-08-23 11:45:29 +020010283
10284 Example: the minimal configuration below would make HAProxy look for the
10285 state server file '/etc/haproxy/states/bk':
10286
10287 global
10288 server-state-file-base /etc/haproxy/states
10289
Willy Tarreauc9c6cdb2020-03-05 16:03:58 +010010290 backend bk
Baptiste Assmann01c6cc32015-08-23 11:45:29 +020010291 load-server-state-from-file
10292
Christopher Faulet583b6de2021-02-12 09:27:10 +010010293 See also: "server-state-base", "load-server-state-from-file", and
Baptiste Assmann01c6cc32015-08-23 11:45:29 +020010294 "show servers state"
Willy Tarreaueabeafa2008-01-16 16:17:06 +010010295
Frédéric Lécaillecb4502e2017-04-20 13:36:25 +020010296server-template <prefix> <num | range> <fqdn>[:<port>] [params*]
10297 Set a template to initialize servers with shared parameters.
10298 The names of these servers are built from <prefix> and <num | range> parameters.
10299 May be used in sections : defaults | frontend | listen | backend
10300 no | no | yes | yes
10301
10302 Arguments:
10303 <prefix> A prefix for the server names to be built.
10304
10305 <num | range>
10306 If <num> is provided, this template initializes <num> servers
10307 with 1 up to <num> as server name suffixes. A range of numbers
10308 <num_low>-<num_high> may also be used to use <num_low> up to
10309 <num_high> as server name suffixes.
10310
10311 <fqdn> A FQDN for all the servers this template initializes.
10312
10313 <port> Same meaning as "server" <port> argument (see "server" keyword).
10314
10315 <params*>
10316 Remaining server parameters among all those supported by "server"
10317 keyword.
10318
10319 Examples:
10320 # Initializes 3 servers with srv1, srv2 and srv3 as names,
10321 # google.com as FQDN, and health-check enabled.
10322 server-template srv 1-3 google.com:80 check
10323
10324 # or
10325 server-template srv 3 google.com:80 check
10326
10327 # would be equivalent to:
10328 server srv1 google.com:80 check
10329 server srv2 google.com:80 check
10330 server srv3 google.com:80 check
10331
10332
10333
Willy Tarreaueabeafa2008-01-16 16:17:06 +010010334source <addr>[:<port>] [usesrc { <addr2>[:<port2>] | client | clientip } ]
Willy Tarreaubce70882009-09-07 11:51:47 +020010335source <addr>[:<port>] [usesrc { <addr2>[:<port2>] | hdr_ip(<hdr>[,<occ>]) } ]
Willy Tarreaud53f96b2009-02-04 18:46:54 +010010336source <addr>[:<port>] [interface <name>]
Willy Tarreaueabeafa2008-01-16 16:17:06 +010010337 Set the source address for outgoing connections
10338 May be used in sections : defaults | frontend | listen | backend
10339 yes | no | yes | yes
10340 Arguments :
10341 <addr> is the IPv4 address HAProxy will bind to before connecting to a
10342 server. This address is also used as a source for health checks.
Willy Tarreau24709282013-03-10 21:32:12 +010010343
Willy Tarreaueabeafa2008-01-16 16:17:06 +010010344 The default value of 0.0.0.0 means that the system will select
Willy Tarreau24709282013-03-10 21:32:12 +010010345 the most appropriate address to reach its destination. Optionally
10346 an address family prefix may be used before the address to force
10347 the family regardless of the address format, which can be useful
10348 to specify a path to a unix socket with no slash ('/'). Currently
10349 supported prefixes are :
10350 - 'ipv4@' -> address is always IPv4
10351 - 'ipv6@' -> address is always IPv6
10352 - 'unix@' -> address is a path to a local unix socket
Willy Tarreauccfccef2014-05-10 01:49:15 +020010353 - 'abns@' -> address is in abstract namespace (Linux only)
Cyril Bonté307ee1e2015-09-28 23:16:06 +020010354 You may want to reference some environment variables in the
10355 address parameter, see section 2.3 about environment variables.
Willy Tarreaueabeafa2008-01-16 16:17:06 +010010356
10357 <port> is an optional port. It is normally not needed but may be useful
10358 in some very specific contexts. The default value of zero means
Willy Tarreauc6f4ce82009-06-10 11:09:37 +020010359 the system will select a free port. Note that port ranges are not
10360 supported in the backend. If you want to force port ranges, you
10361 have to specify them on each "server" line.
Willy Tarreaueabeafa2008-01-16 16:17:06 +010010362
10363 <addr2> is the IP address to present to the server when connections are
10364 forwarded in full transparent proxy mode. This is currently only
10365 supported on some patched Linux kernels. When this address is
10366 specified, clients connecting to the server will be presented
10367 with this address, while health checks will still use the address
10368 <addr>.
10369
10370 <port2> is the optional port to present to the server when connections
10371 are forwarded in full transparent proxy mode (see <addr2> above).
10372 The default value of zero means the system will select a free
10373 port.
10374
Willy Tarreaubce70882009-09-07 11:51:47 +020010375 <hdr> is the name of a HTTP header in which to fetch the IP to bind to.
10376 This is the name of a comma-separated header list which can
10377 contain multiple IP addresses. By default, the last occurrence is
10378 used. This is designed to work with the X-Forwarded-For header
Baptiste Assmannea3e73b2013-02-02 23:47:49 +010010379 and to automatically bind to the client's IP address as seen
Willy Tarreaubce70882009-09-07 11:51:47 +020010380 by previous proxy, typically Stunnel. In order to use another
10381 occurrence from the last one, please see the <occ> parameter
10382 below. When the header (or occurrence) is not found, no binding
10383 is performed so that the proxy's default IP address is used. Also
10384 keep in mind that the header name is case insensitive, as for any
10385 HTTP header.
10386
10387 <occ> is the occurrence number of a value to be used in a multi-value
10388 header. This is to be used in conjunction with "hdr_ip(<hdr>)",
Jamie Gloudonaaa21002012-08-25 00:18:33 -040010389 in order to specify which occurrence to use for the source IP
Willy Tarreaubce70882009-09-07 11:51:47 +020010390 address. Positive values indicate a position from the first
10391 occurrence, 1 being the first one. Negative values indicate
10392 positions relative to the last one, -1 being the last one. This
10393 is helpful for situations where an X-Forwarded-For header is set
10394 at the entry point of an infrastructure and must be used several
10395 proxy layers away. When this value is not specified, -1 is
10396 assumed. Passing a zero here disables the feature.
10397
Willy Tarreaud53f96b2009-02-04 18:46:54 +010010398 <name> is an optional interface name to which to bind to for outgoing
10399 traffic. On systems supporting this features (currently, only
10400 Linux), this allows one to bind all traffic to the server to
10401 this interface even if it is not the one the system would select
10402 based on routing tables. This should be used with extreme care.
10403 Note that using this option requires root privileges.
10404
Willy Tarreaueabeafa2008-01-16 16:17:06 +010010405 The "source" keyword is useful in complex environments where a specific
10406 address only is allowed to connect to the servers. It may be needed when a
10407 private address must be used through a public gateway for instance, and it is
10408 known that the system cannot determine the adequate source address by itself.
10409
10410 An extension which is available on certain patched Linux kernels may be used
10411 through the "usesrc" optional keyword. It makes it possible to connect to the
10412 servers with an IP address which does not belong to the system itself. This
10413 is called "full transparent proxy mode". For this to work, the destination
10414 servers have to route their traffic back to this address through the machine
10415 running HAProxy, and IP forwarding must generally be enabled on this machine.
10416
10417 In this "full transparent proxy" mode, it is possible to force a specific IP
10418 address to be presented to the servers. This is not much used in fact. A more
10419 common use is to tell HAProxy to present the client's IP address. For this,
10420 there are two methods :
10421
10422 - present the client's IP and port addresses. This is the most transparent
10423 mode, but it can cause problems when IP connection tracking is enabled on
10424 the machine, because a same connection may be seen twice with different
10425 states. However, this solution presents the huge advantage of not
10426 limiting the system to the 64k outgoing address+port couples, because all
10427 of the client ranges may be used.
10428
10429 - present only the client's IP address and select a spare port. This
10430 solution is still quite elegant but slightly less transparent (downstream
10431 firewalls logs will not match upstream's). It also presents the downside
10432 of limiting the number of concurrent connections to the usual 64k ports.
10433 However, since the upstream and downstream ports are different, local IP
10434 connection tracking on the machine will not be upset by the reuse of the
10435 same session.
10436
Willy Tarreaueabeafa2008-01-16 16:17:06 +010010437 This option sets the default source for all servers in the backend. It may
10438 also be specified in a "defaults" section. Finer source address specification
10439 is possible at the server level using the "source" server option. Refer to
Willy Tarreauc57f0e22009-05-10 13:12:33 +020010440 section 5 for more information.
Willy Tarreaueabeafa2008-01-16 16:17:06 +010010441
Baptiste Assmann91bd3372015-07-17 21:59:42 +020010442 In order to work, "usesrc" requires root privileges.
10443
Willy Tarreaueabeafa2008-01-16 16:17:06 +010010444 Examples :
10445 backend private
10446 # Connect to the servers using our 192.168.1.200 source address
10447 source 192.168.1.200
10448
10449 backend transparent_ssl1
10450 # Connect to the SSL farm from the client's source address
10451 source 192.168.1.200 usesrc clientip
10452
10453 backend transparent_ssl2
10454 # Connect to the SSL farm from the client's source address and port
10455 # not recommended if IP conntrack is present on the local machine.
10456 source 192.168.1.200 usesrc client
10457
10458 backend transparent_ssl3
10459 # Connect to the SSL farm from the client's source address. It
10460 # is more conntrack-friendly.
10461 source 192.168.1.200 usesrc clientip
10462
10463 backend transparent_smtp
10464 # Connect to the SMTP farm from the client's source address/port
10465 # with Tproxy version 4.
10466 source 0.0.0.0 usesrc clientip
10467
Willy Tarreaubce70882009-09-07 11:51:47 +020010468 backend transparent_http
10469 # Connect to the servers using the client's IP as seen by previous
10470 # proxy.
10471 source 0.0.0.0 usesrc hdr_ip(x-forwarded-for,-1)
10472
Willy Tarreauc57f0e22009-05-10 13:12:33 +020010473 See also : the "source" server option in section 5, the Tproxy patches for
Willy Tarreaueabeafa2008-01-16 16:17:06 +010010474 the Linux kernel on www.balabit.com, the "bind" keyword.
10475
Willy Tarreau844e3c52008-01-11 16:28:18 +010010476
MIZUTA Takeshib24bc0d2020-07-09 11:13:20 +090010477srvtcpka-cnt <count>
10478 Sets the maximum number of keepalive probes TCP should send before dropping
10479 the connection on the server side.
10480 May be used in sections : defaults | frontend | listen | backend
10481 yes | no | yes | yes
10482 Arguments :
10483 <count> is the maximum number of keepalive probes.
10484
10485 This keyword corresponds to the socket option TCP_KEEPCNT. If this keyword
10486 is not specified, system-wide TCP parameter (tcp_keepalive_probes) is used.
Willy Tarreau52543212020-07-09 05:58:51 +020010487 The availability of this setting depends on the operating system. It is
10488 known to work on Linux.
MIZUTA Takeshib24bc0d2020-07-09 11:13:20 +090010489
10490 See also : "option srvtcpka", "srvtcpka-idle", "srvtcpka-intvl".
10491
10492
10493srvtcpka-idle <timeout>
10494 Sets the time the connection needs to remain idle before TCP starts sending
10495 keepalive probes, if enabled the sending of TCP keepalive packets on the
10496 server side.
10497 May be used in sections : defaults | frontend | listen | backend
10498 yes | no | yes | yes
10499 Arguments :
10500 <timeout> is the time the connection needs to remain idle before TCP starts
10501 sending keepalive probes. It is specified in seconds by default,
10502 but can be in any other unit if the number is suffixed by the
10503 unit, as explained at the top of this document.
10504
10505 This keyword corresponds to the socket option TCP_KEEPIDLE. If this keyword
10506 is not specified, system-wide TCP parameter (tcp_keepalive_time) is used.
Willy Tarreau52543212020-07-09 05:58:51 +020010507 The availability of this setting depends on the operating system. It is
10508 known to work on Linux.
MIZUTA Takeshib24bc0d2020-07-09 11:13:20 +090010509
10510 See also : "option srvtcpka", "srvtcpka-cnt", "srvtcpka-intvl".
10511
10512
10513srvtcpka-intvl <timeout>
10514 Sets the time between individual keepalive probes on the server side.
10515 May be used in sections : defaults | frontend | listen | backend
10516 yes | no | yes | yes
10517 Arguments :
10518 <timeout> is the time between individual keepalive probes. It is specified
10519 in seconds by default, but can be in any other unit if the number
10520 is suffixed by the unit, as explained at the top of this
10521 document.
10522
10523 This keyword corresponds to the socket option TCP_KEEPINTVL. If this keyword
10524 is not specified, system-wide TCP parameter (tcp_keepalive_intvl) is used.
Willy Tarreau52543212020-07-09 05:58:51 +020010525 The availability of this setting depends on the operating system. It is
10526 known to work on Linux.
MIZUTA Takeshib24bc0d2020-07-09 11:13:20 +090010527
10528 See also : "option srvtcpka", "srvtcpka-cnt", "srvtcpka-idle".
10529
10530
Cyril Bonté66c327d2010-10-12 00:14:37 +020010531stats admin { if | unless } <cond>
10532 Enable statistics admin level if/unless a condition is matched
10533 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaued2119c2014-04-24 22:10:39 +020010534 no | yes | yes | yes
Cyril Bonté66c327d2010-10-12 00:14:37 +020010535
10536 This statement enables the statistics admin level if/unless a condition is
10537 matched.
10538
10539 The admin level allows to enable/disable servers from the web interface. By
10540 default, statistics page is read-only for security reasons.
10541
Cyril Bonté02ff8ef2010-12-14 22:48:49 +010010542 Note : Consider not using this feature in multi-process mode (nbproc > 1)
10543 unless you know what you do : memory is not shared between the
Davor Ocelice9ed2812017-12-25 17:49:28 +010010544 processes, which can result in random behaviors.
Cyril Bonté02ff8ef2010-12-14 22:48:49 +010010545
Cyril Bonté23b39d92011-02-10 22:54:44 +010010546 Currently, the POST request is limited to the buffer size minus the reserved
10547 buffer space, which means that if the list of servers is too long, the
10548 request won't be processed. It is recommended to alter few servers at a
10549 time.
Cyril Bonté66c327d2010-10-12 00:14:37 +020010550
10551 Example :
10552 # statistics admin level only for localhost
10553 backend stats_localhost
10554 stats enable
10555 stats admin if LOCALHOST
10556
10557 Example :
10558 # statistics admin level always enabled because of the authentication
10559 backend stats_auth
10560 stats enable
10561 stats auth admin:AdMiN123
10562 stats admin if TRUE
10563
10564 Example :
10565 # statistics admin level depends on the authenticated user
10566 userlist stats-auth
10567 group admin users admin
10568 user admin insecure-password AdMiN123
10569 group readonly users haproxy
10570 user haproxy insecure-password haproxy
10571
10572 backend stats_auth
10573 stats enable
10574 acl AUTH http_auth(stats-auth)
10575 acl AUTH_ADMIN http_auth_group(stats-auth) admin
10576 stats http-request auth unless AUTH
10577 stats admin if AUTH_ADMIN
10578
Cyril Bonté02ff8ef2010-12-14 22:48:49 +010010579 See also : "stats enable", "stats auth", "stats http-request", "nbproc",
10580 "bind-process", section 3.4 about userlists and section 7 about
10581 ACL usage.
Cyril Bonté66c327d2010-10-12 00:14:37 +020010582
10583
Willy Tarreaueabeafa2008-01-16 16:17:06 +010010584stats auth <user>:<passwd>
10585 Enable statistics with authentication and grant access to an account
10586 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaued2119c2014-04-24 22:10:39 +020010587 yes | yes | yes | yes
Willy Tarreaueabeafa2008-01-16 16:17:06 +010010588 Arguments :
10589 <user> is a user name to grant access to
10590
10591 <passwd> is the cleartext password associated to this user
10592
10593 This statement enables statistics with default settings, and restricts access
10594 to declared users only. It may be repeated as many times as necessary to
10595 allow as many users as desired. When a user tries to access the statistics
10596 without a valid account, a "401 Forbidden" response will be returned so that
10597 the browser asks the user to provide a valid user and password. The real
10598 which will be returned to the browser is configurable using "stats realm".
10599
10600 Since the authentication method is HTTP Basic Authentication, the passwords
10601 circulate in cleartext on the network. Thus, it was decided that the
10602 configuration file would also use cleartext passwords to remind the users
Willy Tarreau3c92c5f2011-08-28 09:45:47 +020010603 that those ones should not be sensitive and not shared with any other account.
Willy Tarreaueabeafa2008-01-16 16:17:06 +010010604
10605 It is also possible to reduce the scope of the proxies which appear in the
10606 report using "stats scope".
10607
10608 Though this statement alone is enough to enable statistics reporting, it is
10609 recommended to set all other settings in order to avoid relying on default
10610 unobvious parameters.
10611
10612 Example :
10613 # public access (limited to this backend only)
10614 backend public_www
10615 server srv1 192.168.0.1:80
10616 stats enable
10617 stats hide-version
10618 stats scope .
10619 stats uri /admin?stats
Davor Ocelice9ed2812017-12-25 17:49:28 +010010620 stats realm HAProxy\ Statistics
Willy Tarreaueabeafa2008-01-16 16:17:06 +010010621 stats auth admin1:AdMiN123
10622 stats auth admin2:AdMiN321
10623
10624 # internal monitoring access (unlimited)
10625 backend private_monitoring
10626 stats enable
10627 stats uri /admin?stats
10628 stats refresh 5s
10629
10630 See also : "stats enable", "stats realm", "stats scope", "stats uri"
10631
10632
10633stats enable
10634 Enable statistics reporting with default settings
10635 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaued2119c2014-04-24 22:10:39 +020010636 yes | yes | yes | yes
Willy Tarreaueabeafa2008-01-16 16:17:06 +010010637 Arguments : none
10638
10639 This statement enables statistics reporting with default settings defined
10640 at build time. Unless stated otherwise, these settings are used :
10641 - stats uri : /haproxy?stats
10642 - stats realm : "HAProxy Statistics"
10643 - stats auth : no authentication
10644 - stats scope : no restriction
10645
10646 Though this statement alone is enough to enable statistics reporting, it is
10647 recommended to set all other settings in order to avoid relying on default
10648 unobvious parameters.
10649
10650 Example :
10651 # public access (limited to this backend only)
10652 backend public_www
10653 server srv1 192.168.0.1:80
10654 stats enable
10655 stats hide-version
10656 stats scope .
10657 stats uri /admin?stats
Davor Ocelice9ed2812017-12-25 17:49:28 +010010658 stats realm HAProxy\ Statistics
Willy Tarreaueabeafa2008-01-16 16:17:06 +010010659 stats auth admin1:AdMiN123
10660 stats auth admin2:AdMiN321
10661
10662 # internal monitoring access (unlimited)
10663 backend private_monitoring
10664 stats enable
10665 stats uri /admin?stats
10666 stats refresh 5s
10667
10668 See also : "stats auth", "stats realm", "stats uri"
10669
10670
Willy Tarreaud63335a2010-02-26 12:56:52 +010010671stats hide-version
10672 Enable statistics and hide HAProxy version reporting
Willy Tarreau1d45b7c2009-08-16 10:29:18 +020010673 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaued2119c2014-04-24 22:10:39 +020010674 yes | yes | yes | yes
Willy Tarreaud63335a2010-02-26 12:56:52 +010010675 Arguments : none
Willy Tarreau1d45b7c2009-08-16 10:29:18 +020010676
Willy Tarreaud63335a2010-02-26 12:56:52 +010010677 By default, the stats page reports some useful status information along with
10678 the statistics. Among them is HAProxy's version. However, it is generally
10679 considered dangerous to report precise version to anyone, as it can help them
10680 target known weaknesses with specific attacks. The "stats hide-version"
10681 statement removes the version from the statistics report. This is recommended
10682 for public sites or any site with a weak login/password.
Willy Tarreau1d45b7c2009-08-16 10:29:18 +020010683
Krzysztof Piotr Oledzki48cb2ae2009-10-02 22:51:14 +020010684 Though this statement alone is enough to enable statistics reporting, it is
10685 recommended to set all other settings in order to avoid relying on default
10686 unobvious parameters.
10687
Willy Tarreaud63335a2010-02-26 12:56:52 +010010688 Example :
10689 # public access (limited to this backend only)
10690 backend public_www
10691 server srv1 192.168.0.1:80
Krzysztof Piotr Oledzki48cb2ae2009-10-02 22:51:14 +020010692 stats enable
Willy Tarreaud63335a2010-02-26 12:56:52 +010010693 stats hide-version
10694 stats scope .
10695 stats uri /admin?stats
Davor Ocelice9ed2812017-12-25 17:49:28 +010010696 stats realm HAProxy\ Statistics
Willy Tarreaud63335a2010-02-26 12:56:52 +010010697 stats auth admin1:AdMiN123
10698 stats auth admin2:AdMiN321
Willy Tarreau1d45b7c2009-08-16 10:29:18 +020010699
Willy Tarreau1d45b7c2009-08-16 10:29:18 +020010700 # internal monitoring access (unlimited)
10701 backend private_monitoring
10702 stats enable
Willy Tarreaud63335a2010-02-26 12:56:52 +010010703 stats uri /admin?stats
10704 stats refresh 5s
Krzysztof Piotr Oledzki15514c22010-01-04 16:03:09 +010010705
Willy Tarreaud63335a2010-02-26 12:56:52 +010010706 See also : "stats auth", "stats enable", "stats realm", "stats uri"
Willy Tarreau1d45b7c2009-08-16 10:29:18 +020010707
Willy Tarreau983e01e2010-01-11 18:42:06 +010010708
Cyril Bonté2be1b3f2010-09-30 23:46:30 +020010709stats http-request { allow | deny | auth [realm <realm>] }
10710 [ { if | unless } <condition> ]
10711 Access control for statistics
10712
10713 May be used in sections: defaults | frontend | listen | backend
10714 no | no | yes | yes
10715
10716 As "http-request", these set of options allow to fine control access to
10717 statistics. Each option may be followed by if/unless and acl.
10718 First option with matched condition (or option without condition) is final.
10719 For "deny" a 403 error will be returned, for "allow" normal processing is
10720 performed, for "auth" a 401/407 error code is returned so the client
10721 should be asked to enter a username and password.
10722
10723 There is no fixed limit to the number of http-request statements per
10724 instance.
10725
10726 See also : "http-request", section 3.4 about userlists and section 7
10727 about ACL usage.
10728
10729
Willy Tarreaueabeafa2008-01-16 16:17:06 +010010730stats realm <realm>
10731 Enable statistics and set authentication realm
10732 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaued2119c2014-04-24 22:10:39 +020010733 yes | yes | yes | yes
Willy Tarreaueabeafa2008-01-16 16:17:06 +010010734 Arguments :
10735 <realm> is the name of the HTTP Basic Authentication realm reported to
10736 the browser. The browser uses it to display it in the pop-up
10737 inviting the user to enter a valid username and password.
10738
10739 The realm is read as a single word, so any spaces in it should be escaped
10740 using a backslash ('\').
10741
10742 This statement is useful only in conjunction with "stats auth" since it is
10743 only related to authentication.
10744
10745 Though this statement alone is enough to enable statistics reporting, it is
10746 recommended to set all other settings in order to avoid relying on default
10747 unobvious parameters.
10748
10749 Example :
10750 # public access (limited to this backend only)
10751 backend public_www
10752 server srv1 192.168.0.1:80
10753 stats enable
10754 stats hide-version
10755 stats scope .
10756 stats uri /admin?stats
Davor Ocelice9ed2812017-12-25 17:49:28 +010010757 stats realm HAProxy\ Statistics
Willy Tarreaueabeafa2008-01-16 16:17:06 +010010758 stats auth admin1:AdMiN123
10759 stats auth admin2:AdMiN321
10760
10761 # internal monitoring access (unlimited)
10762 backend private_monitoring
10763 stats enable
10764 stats uri /admin?stats
10765 stats refresh 5s
10766
10767 See also : "stats auth", "stats enable", "stats uri"
10768
10769
10770stats refresh <delay>
10771 Enable statistics with automatic refresh
10772 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaued2119c2014-04-24 22:10:39 +020010773 yes | yes | yes | yes
Willy Tarreaueabeafa2008-01-16 16:17:06 +010010774 Arguments :
10775 <delay> is the suggested refresh delay, specified in seconds, which will
10776 be returned to the browser consulting the report page. While the
10777 browser is free to apply any delay, it will generally respect it
10778 and refresh the page this every seconds. The refresh interval may
10779 be specified in any other non-default time unit, by suffixing the
10780 unit after the value, as explained at the top of this document.
10781
10782 This statement is useful on monitoring displays with a permanent page
10783 reporting the load balancer's activity. When set, the HTML report page will
10784 include a link "refresh"/"stop refresh" so that the user can select whether
Jackie Tapia749f74c2020-07-22 18:59:40 -050010785 they want automatic refresh of the page or not.
Willy Tarreaueabeafa2008-01-16 16:17:06 +010010786
10787 Though this statement alone is enough to enable statistics reporting, it is
10788 recommended to set all other settings in order to avoid relying on default
10789 unobvious parameters.
10790
10791 Example :
10792 # public access (limited to this backend only)
10793 backend public_www
10794 server srv1 192.168.0.1:80
10795 stats enable
10796 stats hide-version
10797 stats scope .
10798 stats uri /admin?stats
Davor Ocelice9ed2812017-12-25 17:49:28 +010010799 stats realm HAProxy\ Statistics
Willy Tarreaueabeafa2008-01-16 16:17:06 +010010800 stats auth admin1:AdMiN123
10801 stats auth admin2:AdMiN321
10802
10803 # internal monitoring access (unlimited)
10804 backend private_monitoring
10805 stats enable
10806 stats uri /admin?stats
10807 stats refresh 5s
10808
10809 See also : "stats auth", "stats enable", "stats realm", "stats uri"
10810
10811
10812stats scope { <name> | "." }
10813 Enable statistics and limit access scope
10814 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaued2119c2014-04-24 22:10:39 +020010815 yes | yes | yes | yes
Willy Tarreaueabeafa2008-01-16 16:17:06 +010010816 Arguments :
10817 <name> is the name of a listen, frontend or backend section to be
10818 reported. The special name "." (a single dot) designates the
10819 section in which the statement appears.
10820
10821 When this statement is specified, only the sections enumerated with this
10822 statement will appear in the report. All other ones will be hidden. This
10823 statement may appear as many times as needed if multiple sections need to be
10824 reported. Please note that the name checking is performed as simple string
10825 comparisons, and that it is never checked that a give section name really
10826 exists.
10827
10828 Though this statement alone is enough to enable statistics reporting, it is
10829 recommended to set all other settings in order to avoid relying on default
10830 unobvious parameters.
10831
10832 Example :
10833 # public access (limited to this backend only)
10834 backend public_www
10835 server srv1 192.168.0.1:80
10836 stats enable
10837 stats hide-version
10838 stats scope .
10839 stats uri /admin?stats
Davor Ocelice9ed2812017-12-25 17:49:28 +010010840 stats realm HAProxy\ Statistics
Willy Tarreaueabeafa2008-01-16 16:17:06 +010010841 stats auth admin1:AdMiN123
10842 stats auth admin2:AdMiN321
10843
10844 # internal monitoring access (unlimited)
10845 backend private_monitoring
10846 stats enable
10847 stats uri /admin?stats
10848 stats refresh 5s
10849
10850 See also : "stats auth", "stats enable", "stats realm", "stats uri"
10851
Willy Tarreaud63335a2010-02-26 12:56:52 +010010852
Willy Tarreauc9705a12010-07-27 20:05:50 +020010853stats show-desc [ <desc> ]
Willy Tarreaud63335a2010-02-26 12:56:52 +010010854 Enable reporting of a description on the statistics page.
10855 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaued2119c2014-04-24 22:10:39 +020010856 yes | yes | yes | yes
Willy Tarreaud63335a2010-02-26 12:56:52 +010010857
Willy Tarreauc9705a12010-07-27 20:05:50 +020010858 <desc> is an optional description to be reported. If unspecified, the
Willy Tarreaud63335a2010-02-26 12:56:52 +010010859 description from global section is automatically used instead.
10860
10861 This statement is useful for users that offer shared services to their
10862 customers, where node or description should be different for each customer.
10863
10864 Though this statement alone is enough to enable statistics reporting, it is
10865 recommended to set all other settings in order to avoid relying on default
Davor Ocelice9ed2812017-12-25 17:49:28 +010010866 unobvious parameters. By default description is not shown.
Willy Tarreaud63335a2010-02-26 12:56:52 +010010867
10868 Example :
10869 # internal monitoring access (unlimited)
10870 backend private_monitoring
10871 stats enable
10872 stats show-desc Master node for Europe, Asia, Africa
10873 stats uri /admin?stats
10874 stats refresh 5s
10875
10876 See also: "show-node", "stats enable", "stats uri" and "description" in
10877 global section.
10878
10879
10880stats show-legends
Willy Tarreaued2119c2014-04-24 22:10:39 +020010881 Enable reporting additional information on the statistics page
10882 May be used in sections : defaults | frontend | listen | backend
10883 yes | yes | yes | yes
10884 Arguments : none
10885
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030010886 Enable reporting additional information on the statistics page :
Willy Tarreaud63335a2010-02-26 12:56:52 +010010887 - cap: capabilities (proxy)
10888 - mode: one of tcp, http or health (proxy)
10889 - id: SNMP ID (proxy, socket, server)
10890 - IP (socket, server)
10891 - cookie (backend, server)
10892
10893 Though this statement alone is enough to enable statistics reporting, it is
10894 recommended to set all other settings in order to avoid relying on default
Davor Ocelice9ed2812017-12-25 17:49:28 +010010895 unobvious parameters. Default behavior is not to show this information.
Willy Tarreaud63335a2010-02-26 12:56:52 +010010896
10897 See also: "stats enable", "stats uri".
10898
10899
Amaury Denoyelle0b70a8a2020-10-05 11:49:45 +020010900stats show-modules
10901 Enable display of extra statistics module on the statistics page
10902 May be used in sections : defaults | frontend | listen | backend
10903 yes | yes | yes | yes
10904 Arguments : none
10905
10906 New columns are added at the end of the line containing the extra statistics
10907 values as a tooltip.
10908
10909 Though this statement alone is enough to enable statistics reporting, it is
10910 recommended to set all other settings in order to avoid relying on default
10911 unobvious parameters. Default behavior is not to show this information.
10912
10913 See also: "stats enable", "stats uri".
10914
10915
Willy Tarreaud63335a2010-02-26 12:56:52 +010010916stats show-node [ <name> ]
10917 Enable reporting of a host name on the statistics page.
10918 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaued2119c2014-04-24 22:10:39 +020010919 yes | yes | yes | yes
Willy Tarreaud63335a2010-02-26 12:56:52 +010010920 Arguments:
10921 <name> is an optional name to be reported. If unspecified, the
10922 node name from global section is automatically used instead.
10923
10924 This statement is useful for users that offer shared services to their
10925 customers, where node or description might be different on a stats page
Davor Ocelice9ed2812017-12-25 17:49:28 +010010926 provided for each customer. Default behavior is not to show host name.
Willy Tarreaud63335a2010-02-26 12:56:52 +010010927
10928 Though this statement alone is enough to enable statistics reporting, it is
10929 recommended to set all other settings in order to avoid relying on default
10930 unobvious parameters.
10931
10932 Example:
10933 # internal monitoring access (unlimited)
10934 backend private_monitoring
10935 stats enable
10936 stats show-node Europe-1
10937 stats uri /admin?stats
10938 stats refresh 5s
10939
10940 See also: "show-desc", "stats enable", "stats uri", and "node" in global
10941 section.
10942
Willy Tarreaueabeafa2008-01-16 16:17:06 +010010943
10944stats uri <prefix>
10945 Enable statistics and define the URI prefix to access them
10946 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaued2119c2014-04-24 22:10:39 +020010947 yes | yes | yes | yes
Willy Tarreaueabeafa2008-01-16 16:17:06 +010010948 Arguments :
10949 <prefix> is the prefix of any URI which will be redirected to stats. This
10950 prefix may contain a question mark ('?') to indicate part of a
10951 query string.
10952
10953 The statistics URI is intercepted on the relayed traffic, so it appears as a
10954 page within the normal application. It is strongly advised to ensure that the
10955 selected URI will never appear in the application, otherwise it will never be
10956 possible to reach it in the application.
10957
Daniel Corbett9f0843f2021-05-08 10:50:37 -040010958 The default URI compiled in HAProxy is "/haproxy?stats", but this may be
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +010010959 changed at build time, so it's better to always explicitly specify it here.
Willy Tarreaueabeafa2008-01-16 16:17:06 +010010960 It is generally a good idea to include a question mark in the URI so that
10961 intermediate proxies refrain from caching the results. Also, since any string
10962 beginning with the prefix will be accepted as a stats request, the question
10963 mark helps ensuring that no valid URI will begin with the same words.
10964
10965 It is sometimes very convenient to use "/" as the URI prefix, and put that
10966 statement in a "listen" instance of its own. That makes it easy to dedicate
10967 an address or a port to statistics only.
10968
10969 Though this statement alone is enough to enable statistics reporting, it is
10970 recommended to set all other settings in order to avoid relying on default
10971 unobvious parameters.
10972
10973 Example :
10974 # public access (limited to this backend only)
10975 backend public_www
10976 server srv1 192.168.0.1:80
10977 stats enable
10978 stats hide-version
10979 stats scope .
10980 stats uri /admin?stats
Davor Ocelice9ed2812017-12-25 17:49:28 +010010981 stats realm HAProxy\ Statistics
Willy Tarreaueabeafa2008-01-16 16:17:06 +010010982 stats auth admin1:AdMiN123
10983 stats auth admin2:AdMiN321
10984
10985 # internal monitoring access (unlimited)
10986 backend private_monitoring
10987 stats enable
10988 stats uri /admin?stats
10989 stats refresh 5s
10990
10991 See also : "stats auth", "stats enable", "stats realm"
10992
10993
Willy Tarreaud63335a2010-02-26 12:56:52 +010010994stick match <pattern> [table <table>] [{if | unless} <cond>]
10995 Define a request pattern matching condition to stick a user to a server
Willy Tarreaueabeafa2008-01-16 16:17:06 +010010996 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaud63335a2010-02-26 12:56:52 +010010997 no | no | yes | yes
Willy Tarreaub937b7e2010-01-12 15:27:54 +010010998
10999 Arguments :
Willy Tarreaube722a22014-06-13 16:31:59 +020011000 <pattern> is a sample expression rule as described in section 7.3. It
Willy Tarreaub937b7e2010-01-12 15:27:54 +010011001 describes what elements of the incoming request or connection
Davor Ocelice9ed2812017-12-25 17:49:28 +010011002 will be analyzed in the hope to find a matching entry in a
Willy Tarreaub937b7e2010-01-12 15:27:54 +010011003 stickiness table. This rule is mandatory.
11004
11005 <table> is an optional stickiness table name. If unspecified, the same
11006 backend's table is used. A stickiness table is declared using
11007 the "stick-table" statement.
11008
11009 <cond> is an optional matching condition. It makes it possible to match
11010 on a certain criterion only when other conditions are met (or
11011 not met). For instance, it could be used to match on a source IP
11012 address except when a request passes through a known proxy, in
11013 which case we'd match on a header containing that IP address.
11014
11015 Some protocols or applications require complex stickiness rules and cannot
11016 always simply rely on cookies nor hashing. The "stick match" statement
11017 describes a rule to extract the stickiness criterion from an incoming request
11018 or connection. See section 7 for a complete list of possible patterns and
11019 transformation rules.
11020
11021 The table has to be declared using the "stick-table" statement. It must be of
11022 a type compatible with the pattern. By default it is the one which is present
11023 in the same backend. It is possible to share a table with other backends by
11024 referencing it using the "table" keyword. If another table is referenced,
11025 the server's ID inside the backends are used. By default, all server IDs
11026 start at 1 in each backend, so the server ordering is enough. But in case of
11027 doubt, it is highly recommended to force server IDs using their "id" setting.
11028
11029 It is possible to restrict the conditions where a "stick match" statement
11030 will apply, using "if" or "unless" followed by a condition. See section 7 for
11031 ACL based conditions.
11032
11033 There is no limit on the number of "stick match" statements. The first that
11034 applies and matches will cause the request to be directed to the same server
11035 as was used for the request which created the entry. That way, multiple
11036 matches can be used as fallbacks.
11037
11038 The stick rules are checked after the persistence cookies, so they will not
11039 affect stickiness if a cookie has already been used to select a server. That
11040 way, it becomes very easy to insert cookies and match on IP addresses in
11041 order to maintain stickiness between HTTP and HTTPS.
11042
Cyril Bonté02ff8ef2010-12-14 22:48:49 +010011043 Note : Consider not using this feature in multi-process mode (nbproc > 1)
11044 unless you know what you do : memory is not shared between the
Davor Ocelice9ed2812017-12-25 17:49:28 +010011045 processes, which can result in random behaviors.
Cyril Bonté02ff8ef2010-12-14 22:48:49 +010011046
Willy Tarreaub937b7e2010-01-12 15:27:54 +010011047 Example :
11048 # forward SMTP users to the same server they just used for POP in the
11049 # last 30 minutes
11050 backend pop
11051 mode tcp
11052 balance roundrobin
11053 stick store-request src
11054 stick-table type ip size 200k expire 30m
11055 server s1 192.168.1.1:110
11056 server s2 192.168.1.1:110
11057
11058 backend smtp
11059 mode tcp
11060 balance roundrobin
11061 stick match src table pop
11062 server s1 192.168.1.1:25
11063 server s2 192.168.1.1:25
11064
Cyril Bonté02ff8ef2010-12-14 22:48:49 +010011065 See also : "stick-table", "stick on", "nbproc", "bind-process" and section 7
Willy Tarreaube722a22014-06-13 16:31:59 +020011066 about ACLs and samples fetching.
Willy Tarreaub937b7e2010-01-12 15:27:54 +010011067
11068
11069stick on <pattern> [table <table>] [{if | unless} <condition>]
11070 Define a request pattern to associate a user to a server
11071 May be used in sections : defaults | frontend | listen | backend
11072 no | no | yes | yes
11073
11074 Note : This form is exactly equivalent to "stick match" followed by
11075 "stick store-request", all with the same arguments. Please refer
11076 to both keywords for details. It is only provided as a convenience
11077 for writing more maintainable configurations.
11078
Cyril Bonté02ff8ef2010-12-14 22:48:49 +010011079 Note : Consider not using this feature in multi-process mode (nbproc > 1)
11080 unless you know what you do : memory is not shared between the
Davor Ocelice9ed2812017-12-25 17:49:28 +010011081 processes, which can result in random behaviors.
Cyril Bonté02ff8ef2010-12-14 22:48:49 +010011082
Willy Tarreaub937b7e2010-01-12 15:27:54 +010011083 Examples :
11084 # The following form ...
Willy Tarreauec579d82010-02-26 19:15:04 +010011085 stick on src table pop if !localhost
Willy Tarreaub937b7e2010-01-12 15:27:54 +010011086
11087 # ...is strictly equivalent to this one :
11088 stick match src table pop if !localhost
11089 stick store-request src table pop if !localhost
11090
11091
11092 # Use cookie persistence for HTTP, and stick on source address for HTTPS as
11093 # well as HTTP without cookie. Share the same table between both accesses.
11094 backend http
11095 mode http
11096 balance roundrobin
11097 stick on src table https
11098 cookie SRV insert indirect nocache
11099 server s1 192.168.1.1:80 cookie s1
11100 server s2 192.168.1.1:80 cookie s2
11101
11102 backend https
11103 mode tcp
11104 balance roundrobin
11105 stick-table type ip size 200k expire 30m
11106 stick on src
11107 server s1 192.168.1.1:443
11108 server s2 192.168.1.1:443
11109
Cyril Bonté02ff8ef2010-12-14 22:48:49 +010011110 See also : "stick match", "stick store-request", "nbproc" and "bind-process".
Willy Tarreaub937b7e2010-01-12 15:27:54 +010011111
11112
11113stick store-request <pattern> [table <table>] [{if | unless} <condition>]
11114 Define a request pattern used to create an entry in a stickiness table
11115 May be used in sections : defaults | frontend | listen | backend
11116 no | no | yes | yes
11117
11118 Arguments :
Willy Tarreaube722a22014-06-13 16:31:59 +020011119 <pattern> is a sample expression rule as described in section 7.3. It
Willy Tarreaub937b7e2010-01-12 15:27:54 +010011120 describes what elements of the incoming request or connection
Davor Ocelice9ed2812017-12-25 17:49:28 +010011121 will be analyzed, extracted and stored in the table once a
Willy Tarreaub937b7e2010-01-12 15:27:54 +010011122 server is selected.
11123
11124 <table> is an optional stickiness table name. If unspecified, the same
11125 backend's table is used. A stickiness table is declared using
11126 the "stick-table" statement.
11127
11128 <cond> is an optional storage condition. It makes it possible to store
11129 certain criteria only when some conditions are met (or not met).
11130 For instance, it could be used to store the source IP address
11131 except when the request passes through a known proxy, in which
11132 case we'd store a converted form of a header containing that IP
11133 address.
11134
11135 Some protocols or applications require complex stickiness rules and cannot
11136 always simply rely on cookies nor hashing. The "stick store-request" statement
11137 describes a rule to decide what to extract from the request and when to do
11138 it, in order to store it into a stickiness table for further requests to
11139 match it using the "stick match" statement. Obviously the extracted part must
11140 make sense and have a chance to be matched in a further request. Storing a
11141 client's IP address for instance often makes sense. Storing an ID found in a
11142 URL parameter also makes sense. Storing a source port will almost never make
11143 any sense because it will be randomly matched. See section 7 for a complete
11144 list of possible patterns and transformation rules.
11145
11146 The table has to be declared using the "stick-table" statement. It must be of
11147 a type compatible with the pattern. By default it is the one which is present
11148 in the same backend. It is possible to share a table with other backends by
11149 referencing it using the "table" keyword. If another table is referenced,
11150 the server's ID inside the backends are used. By default, all server IDs
11151 start at 1 in each backend, so the server ordering is enough. But in case of
11152 doubt, it is highly recommended to force server IDs using their "id" setting.
11153
11154 It is possible to restrict the conditions where a "stick store-request"
11155 statement will apply, using "if" or "unless" followed by a condition. This
11156 condition will be evaluated while parsing the request, so any criteria can be
11157 used. See section 7 for ACL based conditions.
11158
11159 There is no limit on the number of "stick store-request" statements, but
11160 there is a limit of 8 simultaneous stores per request or response. This
11161 makes it possible to store up to 8 criteria, all extracted from either the
11162 request or the response, regardless of the number of rules. Only the 8 first
11163 ones which match will be kept. Using this, it is possible to feed multiple
11164 tables at once in the hope to increase the chance to recognize a user on
Willy Tarreau9667a802013-12-09 12:52:13 +010011165 another protocol or access method. Using multiple store-request rules with
11166 the same table is possible and may be used to find the best criterion to rely
11167 on, by arranging the rules by decreasing preference order. Only the first
11168 extracted criterion for a given table will be stored. All subsequent store-
11169 request rules referencing the same table will be skipped and their ACLs will
11170 not be evaluated.
Willy Tarreaub937b7e2010-01-12 15:27:54 +010011171
11172 The "store-request" rules are evaluated once the server connection has been
11173 established, so that the table will contain the real server that processed
11174 the request.
11175
Cyril Bonté02ff8ef2010-12-14 22:48:49 +010011176 Note : Consider not using this feature in multi-process mode (nbproc > 1)
11177 unless you know what you do : memory is not shared between the
Davor Ocelice9ed2812017-12-25 17:49:28 +010011178 processes, which can result in random behaviors.
Cyril Bonté02ff8ef2010-12-14 22:48:49 +010011179
Willy Tarreaub937b7e2010-01-12 15:27:54 +010011180 Example :
11181 # forward SMTP users to the same server they just used for POP in the
11182 # last 30 minutes
11183 backend pop
11184 mode tcp
11185 balance roundrobin
11186 stick store-request src
11187 stick-table type ip size 200k expire 30m
11188 server s1 192.168.1.1:110
11189 server s2 192.168.1.1:110
11190
11191 backend smtp
11192 mode tcp
11193 balance roundrobin
11194 stick match src table pop
11195 server s1 192.168.1.1:25
11196 server s2 192.168.1.1:25
11197
Cyril Bonté02ff8ef2010-12-14 22:48:49 +010011198 See also : "stick-table", "stick on", "nbproc", "bind-process" and section 7
Willy Tarreaube722a22014-06-13 16:31:59 +020011199 about ACLs and sample fetching.
Willy Tarreaub937b7e2010-01-12 15:27:54 +010011200
11201
Emeric Brun7c6b82e2010-09-24 16:34:28 +020011202stick-table type {ip | integer | string [len <length>] | binary [len <length>]}
Thayne McCombs92149f92020-11-20 01:28:26 -070011203 size <size> [expire <expire>] [nopurge] [peers <peersect>] [srvkey <srvkey>]
Emeric Brunf099e792010-09-27 12:05:28 +020011204 [store <data_type>]*
Godbach64cef792013-12-04 16:08:22 +080011205 Configure the stickiness table for the current section
Willy Tarreaub937b7e2010-01-12 15:27:54 +010011206 May be used in sections : defaults | frontend | listen | backend
Willy Tarreauc00cdc22010-06-06 16:48:26 +020011207 no | yes | yes | yes
Willy Tarreaub937b7e2010-01-12 15:27:54 +010011208
11209 Arguments :
11210 ip a table declared with "type ip" will only store IPv4 addresses.
11211 This form is very compact (about 50 bytes per entry) and allows
11212 very fast entry lookup and stores with almost no overhead. This
11213 is mainly used to store client source IP addresses.
11214
David du Colombier9a6d3c92011-03-17 10:40:24 +010011215 ipv6 a table declared with "type ipv6" will only store IPv6 addresses.
11216 This form is very compact (about 60 bytes per entry) and allows
11217 very fast entry lookup and stores with almost no overhead. This
11218 is mainly used to store client source IP addresses.
11219
Willy Tarreaub937b7e2010-01-12 15:27:54 +010011220 integer a table declared with "type integer" will store 32bit integers
11221 which can represent a client identifier found in a request for
11222 instance.
11223
11224 string a table declared with "type string" will store substrings of up
11225 to <len> characters. If the string provided by the pattern
11226 extractor is larger than <len>, it will be truncated before
11227 being stored. During matching, at most <len> characters will be
11228 compared between the string in the table and the extracted
11229 pattern. When not specified, the string is automatically limited
Emeric Brun7c6b82e2010-09-24 16:34:28 +020011230 to 32 characters.
11231
11232 binary a table declared with "type binary" will store binary blocks
11233 of <len> bytes. If the block provided by the pattern
11234 extractor is larger than <len>, it will be truncated before
Willy Tarreaube722a22014-06-13 16:31:59 +020011235 being stored. If the block provided by the sample expression
Emeric Brun7c6b82e2010-09-24 16:34:28 +020011236 is shorter than <len>, it will be padded by 0. When not
11237 specified, the block is automatically limited to 32 bytes.
Willy Tarreaub937b7e2010-01-12 15:27:54 +010011238
11239 <length> is the maximum number of characters that will be stored in a
Emeric Brun7c6b82e2010-09-24 16:34:28 +020011240 "string" type table (See type "string" above). Or the number
11241 of bytes of the block in "binary" type table. Be careful when
Willy Tarreaub937b7e2010-01-12 15:27:54 +010011242 changing this parameter as memory usage will proportionally
11243 increase.
11244
11245 <size> is the maximum number of entries that can fit in the table. This
Cyril Bonté78caf842010-03-10 22:41:43 +010011246 value directly impacts memory usage. Count approximately
11247 50 bytes per entry, plus the size of a string if any. The size
11248 supports suffixes "k", "m", "g" for 2^10, 2^20 and 2^30 factors.
Willy Tarreaub937b7e2010-01-12 15:27:54 +010011249
11250 [nopurge] indicates that we refuse to purge older entries when the table
Daniel Corbett9f0843f2021-05-08 10:50:37 -040011251 is full. When not specified and the table is full when HAProxy
Willy Tarreaub937b7e2010-01-12 15:27:54 +010011252 wants to store an entry in it, it will flush a few of the oldest
11253 entries in order to release some space for the new ones. This is
Davor Ocelice9ed2812017-12-25 17:49:28 +010011254 most often the desired behavior. In some specific cases, it
Willy Tarreaub937b7e2010-01-12 15:27:54 +010011255 be desirable to refuse new entries instead of purging the older
11256 ones. That may be the case when the amount of data to store is
11257 far above the hardware limits and we prefer not to offer access
11258 to new clients than to reject the ones already connected. When
11259 using this parameter, be sure to properly set the "expire"
11260 parameter (see below).
11261
Emeric Brunf099e792010-09-27 12:05:28 +020011262 <peersect> is the name of the peers section to use for replication. Entries
11263 which associate keys to server IDs are kept synchronized with
11264 the remote peers declared in this section. All entries are also
11265 automatically learned from the local peer (old process) during a
11266 soft restart.
11267
Willy Tarreau1abc6732015-05-01 19:21:02 +020011268 NOTE : each peers section may be referenced only by tables
11269 belonging to the same unique process.
Cyril Bonté02ff8ef2010-12-14 22:48:49 +010011270
Willy Tarreaub937b7e2010-01-12 15:27:54 +010011271 <expire> defines the maximum duration of an entry in the table since it
11272 was last created, refreshed or matched. The expiration delay is
11273 defined using the standard time format, similarly as the various
11274 timeouts. The maximum duration is slightly above 24 days. See
Willy Tarreau4b103022021-02-12 17:59:10 +010011275 section 2.5 for more information. If this delay is not specified,
Cyril Bontédc4d9032012-04-08 21:57:39 +020011276 the session won't automatically expire, but older entries will
Willy Tarreaub937b7e2010-01-12 15:27:54 +010011277 be removed once full. Be sure not to use the "nopurge" parameter
11278 if not expiration delay is specified.
11279
Thayne McCombs92149f92020-11-20 01:28:26 -070011280 <srvkey> specifies how each server is identified for the purposes of the
11281 stick table. The valid values are "name" and "addr". If "name" is
11282 given, then <name> argument for the server (may be generated by
11283 a template). If "addr" is given, then the server is identified
11284 by its current network address, including the port. "addr" is
11285 especially useful if you are using service discovery to generate
11286 the addresses for servers with peered stick-tables and want
11287 to consistently use the same host across peers for a stickiness
11288 token.
11289
Willy Tarreau08d5f982010-06-06 13:34:54 +020011290 <data_type> is used to store additional information in the stick-table. This
11291 may be used by ACLs in order to control various criteria related
11292 to the activity of the client matching the stick-table. For each
11293 item specified here, the size of each entry will be inflated so
Willy Tarreauc9705a12010-07-27 20:05:50 +020011294 that the additional data can fit. Several data types may be
11295 stored with an entry. Multiple data types may be specified after
11296 the "store" keyword, as a comma-separated list. Alternatively,
11297 it is possible to repeat the "store" keyword followed by one or
11298 several data types. Except for the "server_id" type which is
11299 automatically detected and enabled, all data types must be
11300 explicitly declared to be stored. If an ACL references a data
11301 type which is not stored, the ACL will simply not match. Some
11302 data types require an argument which must be passed just after
11303 the type between parenthesis. See below for the supported data
11304 types and their arguments.
11305
11306 The data types that can be stored with an entry are the following :
11307 - server_id : this is an integer which holds the numeric ID of the server a
11308 request was assigned to. It is used by the "stick match", "stick store",
11309 and "stick on" rules. It is automatically enabled when referenced.
11310
11311 - gpc0 : first General Purpose Counter. It is a positive 32-bit integer
11312 integer which may be used for anything. Most of the time it will be used
11313 to put a special tag on some entries, for instance to note that a
Davor Ocelice9ed2812017-12-25 17:49:28 +010011314 specific behavior was detected and must be known for future matches.
Willy Tarreauc9705a12010-07-27 20:05:50 +020011315
Willy Tarreauba2ffd12013-05-29 15:54:14 +020011316 - gpc0_rate(<period>) : increment rate of the first General Purpose Counter
11317 over a period. It is a positive 32-bit integer integer which may be used
11318 for anything. Just like <gpc0>, it counts events, but instead of keeping
Davor Ocelice9ed2812017-12-25 17:49:28 +010011319 a cumulative number, it maintains the rate at which the counter is
Willy Tarreauba2ffd12013-05-29 15:54:14 +020011320 incremented. Most of the time it will be used to measure the frequency of
Davor Ocelice9ed2812017-12-25 17:49:28 +010011321 occurrence of certain events (e.g. requests to a specific URL).
Willy Tarreauba2ffd12013-05-29 15:54:14 +020011322
Frédéric Lécaille6778b272018-01-29 15:22:53 +010011323 - gpc1 : second General Purpose Counter. It is a positive 32-bit integer
11324 integer which may be used for anything. Most of the time it will be used
11325 to put a special tag on some entries, for instance to note that a
11326 specific behavior was detected and must be known for future matches.
11327
11328 - gpc1_rate(<period>) : increment rate of the second General Purpose Counter
11329 over a period. It is a positive 32-bit integer integer which may be used
11330 for anything. Just like <gpc1>, it counts events, but instead of keeping
11331 a cumulative number, it maintains the rate at which the counter is
11332 incremented. Most of the time it will be used to measure the frequency of
11333 occurrence of certain events (e.g. requests to a specific URL).
11334
Emeric Bruna5d15312021-07-01 18:34:48 +020011335 - gpt0 : first General Purpose Tag. It is a positive 32-bit integer
11336 integer which may be used for anything. Most of the time it will be used
11337 to put a special tag on some entries, for instance to note that a
11338 specific behavior was detected and must be known for future matches
11339
Willy Tarreauc9705a12010-07-27 20:05:50 +020011340 - conn_cnt : Connection Count. It is a positive 32-bit integer which counts
11341 the absolute number of connections received from clients which matched
11342 this entry. It does not mean the connections were accepted, just that
11343 they were received.
11344
11345 - conn_cur : Current Connections. It is a positive 32-bit integer which
11346 stores the concurrent connection counts for the entry. It is incremented
11347 once an incoming connection matches the entry, and decremented once the
11348 connection leaves. That way it is possible to know at any time the exact
11349 number of concurrent connections for an entry.
11350
11351 - conn_rate(<period>) : frequency counter (takes 12 bytes). It takes an
11352 integer parameter <period> which indicates in milliseconds the length
11353 of the period over which the average is measured. It reports the average
11354 incoming connection rate over that period, in connections per period. The
11355 result is an integer which can be matched using ACLs.
11356
11357 - sess_cnt : Session Count. It is a positive 32-bit integer which counts
11358 the absolute number of sessions received from clients which matched this
11359 entry. A session is a connection that was accepted by the layer 4 rules.
11360
11361 - sess_rate(<period>) : frequency counter (takes 12 bytes). It takes an
11362 integer parameter <period> which indicates in milliseconds the length
11363 of the period over which the average is measured. It reports the average
11364 incoming session rate over that period, in sessions per period. The
11365 result is an integer which can be matched using ACLs.
11366
11367 - http_req_cnt : HTTP request Count. It is a positive 32-bit integer which
11368 counts the absolute number of HTTP requests received from clients which
11369 matched this entry. It does not matter whether they are valid requests or
11370 not. Note that this is different from sessions when keep-alive is used on
11371 the client side.
11372
11373 - http_req_rate(<period>) : frequency counter (takes 12 bytes). It takes an
11374 integer parameter <period> which indicates in milliseconds the length
11375 of the period over which the average is measured. It reports the average
11376 HTTP request rate over that period, in requests per period. The result is
11377 an integer which can be matched using ACLs. It does not matter whether
11378 they are valid requests or not. Note that this is different from sessions
11379 when keep-alive is used on the client side.
11380
11381 - http_err_cnt : HTTP Error Count. It is a positive 32-bit integer which
11382 counts the absolute number of HTTP requests errors induced by clients
11383 which matched this entry. Errors are counted on invalid and truncated
11384 requests, as well as on denied or tarpitted requests, and on failed
11385 authentications. If the server responds with 4xx, then the request is
11386 also counted as an error since it's an error triggered by the client
Davor Ocelice9ed2812017-12-25 17:49:28 +010011387 (e.g. vulnerability scan).
Willy Tarreauc9705a12010-07-27 20:05:50 +020011388
11389 - http_err_rate(<period>) : frequency counter (takes 12 bytes). It takes an
11390 integer parameter <period> which indicates in milliseconds the length
11391 of the period over which the average is measured. It reports the average
11392 HTTP request error rate over that period, in requests per period (see
11393 http_err_cnt above for what is accounted as an error). The result is an
11394 integer which can be matched using ACLs.
11395
Willy Tarreau826f3ab2021-02-10 12:07:15 +010011396 - http_fail_cnt : HTTP Failure Count. It is a positive 32-bit integer which
11397 counts the absolute number of HTTP response failures induced by servers
11398 which matched this entry. Errors are counted on invalid and truncated
11399 responses, as well as any 5xx response other than 501 or 505. It aims at
11400 being used combined with path or URI to detect service failures.
11401
11402 - http_fail_rate(<period>) : frequency counter (takes 12 bytes). It takes
11403 an integer parameter <period> which indicates in milliseconds the length
11404 of the period over which the average is measured. It reports the average
11405 HTTP response failure rate over that period, in requests per period (see
11406 http_fail_cnt above for what is accounted as a failure). The result is an
11407 integer which can be matched using ACLs.
11408
Willy Tarreauc9705a12010-07-27 20:05:50 +020011409 - bytes_in_cnt : client to server byte count. It is a positive 64-bit
Davor Ocelice9ed2812017-12-25 17:49:28 +010011410 integer which counts the cumulative number of bytes received from clients
Willy Tarreauc9705a12010-07-27 20:05:50 +020011411 which matched this entry. Headers are included in the count. This may be
11412 used to limit abuse of upload features on photo or video servers.
11413
11414 - bytes_in_rate(<period>) : frequency counter (takes 12 bytes). It takes an
11415 integer parameter <period> which indicates in milliseconds the length
11416 of the period over which the average is measured. It reports the average
11417 incoming bytes rate over that period, in bytes per period. It may be used
11418 to detect users which upload too much and too fast. Warning: with large
11419 uploads, it is possible that the amount of uploaded data will be counted
11420 once upon termination, thus causing spikes in the average transfer speed
11421 instead of having a smooth one. This may partially be smoothed with
11422 "option contstats" though this is not perfect yet. Use of byte_in_cnt is
11423 recommended for better fairness.
11424
11425 - bytes_out_cnt : server to client byte count. It is a positive 64-bit
Davor Ocelice9ed2812017-12-25 17:49:28 +010011426 integer which counts the cumulative number of bytes sent to clients which
Willy Tarreauc9705a12010-07-27 20:05:50 +020011427 matched this entry. Headers are included in the count. This may be used
11428 to limit abuse of bots sucking the whole site.
11429
11430 - bytes_out_rate(<period>) : frequency counter (takes 12 bytes). It takes
11431 an integer parameter <period> which indicates in milliseconds the length
11432 of the period over which the average is measured. It reports the average
11433 outgoing bytes rate over that period, in bytes per period. It may be used
11434 to detect users which download too much and too fast. Warning: with large
11435 transfers, it is possible that the amount of transferred data will be
11436 counted once upon termination, thus causing spikes in the average
11437 transfer speed instead of having a smooth one. This may partially be
11438 smoothed with "option contstats" though this is not perfect yet. Use of
11439 byte_out_cnt is recommended for better fairness.
Willy Tarreau08d5f982010-06-06 13:34:54 +020011440
Willy Tarreauc00cdc22010-06-06 16:48:26 +020011441 There is only one stick-table per proxy. At the moment of writing this doc,
11442 it does not seem useful to have multiple tables per proxy. If this happens
Willy Tarreaub937b7e2010-01-12 15:27:54 +010011443 to be required, simply create a dummy backend with a stick-table in it and
11444 reference it.
11445
11446 It is important to understand that stickiness based on learning information
11447 has some limitations, including the fact that all learned associations are
Baptiste Assmann123ff042016-03-06 23:29:28 +010011448 lost upon restart unless peers are properly configured to transfer such
11449 information upon restart (recommended). In general it can be good as a
11450 complement but not always as an exclusive stickiness.
Willy Tarreaub937b7e2010-01-12 15:27:54 +010011451
Willy Tarreauc9705a12010-07-27 20:05:50 +020011452 Last, memory requirements may be important when storing many data types.
11453 Indeed, storing all indicators above at once in each entry requires 116 bytes
11454 per entry, or 116 MB for a 1-million entries table. This is definitely not
11455 something that can be ignored.
11456
11457 Example:
11458 # Keep track of counters of up to 1 million IP addresses over 5 minutes
11459 # and store a general purpose counter and the average connection rate
11460 # computed over a sliding window of 30 seconds.
11461 stick-table type ip size 1m expire 5m store gpc0,conn_rate(30s)
11462
Willy Tarreau4b103022021-02-12 17:59:10 +010011463 See also : "stick match", "stick on", "stick store-request", section 2.5
David du Colombiera13d1b92011-03-17 10:40:22 +010011464 about time format and section 7 about ACLs.
Willy Tarreaub937b7e2010-01-12 15:27:54 +010011465
11466
Emeric Brun6a1cefa2010-09-24 18:15:17 +020011467stick store-response <pattern> [table <table>] [{if | unless} <condition>]
Baptiste Assmann2f2d2ec2016-03-06 23:27:24 +010011468 Define a response pattern used to create an entry in a stickiness table
Emeric Brun6a1cefa2010-09-24 18:15:17 +020011469 May be used in sections : defaults | frontend | listen | backend
11470 no | no | yes | yes
11471
11472 Arguments :
Willy Tarreaube722a22014-06-13 16:31:59 +020011473 <pattern> is a sample expression rule as described in section 7.3. It
Emeric Brun6a1cefa2010-09-24 18:15:17 +020011474 describes what elements of the response or connection will
Davor Ocelice9ed2812017-12-25 17:49:28 +010011475 be analyzed, extracted and stored in the table once a
Emeric Brun6a1cefa2010-09-24 18:15:17 +020011476 server is selected.
11477
11478 <table> is an optional stickiness table name. If unspecified, the same
11479 backend's table is used. A stickiness table is declared using
11480 the "stick-table" statement.
11481
11482 <cond> is an optional storage condition. It makes it possible to store
11483 certain criteria only when some conditions are met (or not met).
11484 For instance, it could be used to store the SSL session ID only
11485 when the response is a SSL server hello.
11486
11487 Some protocols or applications require complex stickiness rules and cannot
11488 always simply rely on cookies nor hashing. The "stick store-response"
11489 statement describes a rule to decide what to extract from the response and
11490 when to do it, in order to store it into a stickiness table for further
11491 requests to match it using the "stick match" statement. Obviously the
11492 extracted part must make sense and have a chance to be matched in a further
Cyril Bonté108cf6e2012-04-21 23:30:29 +020011493 request. Storing an ID found in a header of a response makes sense.
Emeric Brun6a1cefa2010-09-24 18:15:17 +020011494 See section 7 for a complete list of possible patterns and transformation
11495 rules.
11496
11497 The table has to be declared using the "stick-table" statement. It must be of
11498 a type compatible with the pattern. By default it is the one which is present
11499 in the same backend. It is possible to share a table with other backends by
11500 referencing it using the "table" keyword. If another table is referenced,
11501 the server's ID inside the backends are used. By default, all server IDs
11502 start at 1 in each backend, so the server ordering is enough. But in case of
11503 doubt, it is highly recommended to force server IDs using their "id" setting.
11504
11505 It is possible to restrict the conditions where a "stick store-response"
11506 statement will apply, using "if" or "unless" followed by a condition. This
11507 condition will be evaluated while parsing the response, so any criteria can
11508 be used. See section 7 for ACL based conditions.
11509
11510 There is no limit on the number of "stick store-response" statements, but
11511 there is a limit of 8 simultaneous stores per request or response. This
11512 makes it possible to store up to 8 criteria, all extracted from either the
11513 request or the response, regardless of the number of rules. Only the 8 first
11514 ones which match will be kept. Using this, it is possible to feed multiple
11515 tables at once in the hope to increase the chance to recognize a user on
Willy Tarreau9667a802013-12-09 12:52:13 +010011516 another protocol or access method. Using multiple store-response rules with
11517 the same table is possible and may be used to find the best criterion to rely
11518 on, by arranging the rules by decreasing preference order. Only the first
11519 extracted criterion for a given table will be stored. All subsequent store-
11520 response rules referencing the same table will be skipped and their ACLs will
11521 not be evaluated. However, even if a store-request rule references a table, a
11522 store-response rule may also use the same table. This means that each table
11523 may learn exactly one element from the request and one element from the
11524 response at once.
Emeric Brun6a1cefa2010-09-24 18:15:17 +020011525
11526 The table will contain the real server that processed the request.
11527
11528 Example :
11529 # Learn SSL session ID from both request and response and create affinity.
11530 backend https
11531 mode tcp
11532 balance roundrobin
Cyril Bontédc4d9032012-04-08 21:57:39 +020011533 # maximum SSL session ID length is 32 bytes.
Emeric Brun6a1cefa2010-09-24 18:15:17 +020011534 stick-table type binary len 32 size 30k expire 30m
Cyril Bonté108cf6e2012-04-21 23:30:29 +020011535
Emeric Brun6a1cefa2010-09-24 18:15:17 +020011536 acl clienthello req_ssl_hello_type 1
11537 acl serverhello rep_ssl_hello_type 2
11538
11539 # use tcp content accepts to detects ssl client and server hello.
11540 tcp-request inspect-delay 5s
11541 tcp-request content accept if clienthello
11542
11543 # no timeout on response inspect delay by default.
11544 tcp-response content accept if serverhello
Cyril Bonté108cf6e2012-04-21 23:30:29 +020011545
Emeric Brun6a1cefa2010-09-24 18:15:17 +020011546 # SSL session ID (SSLID) may be present on a client or server hello.
11547 # Its length is coded on 1 byte at offset 43 and its value starts
11548 # at offset 44.
11549
11550 # Match and learn on request if client hello.
11551 stick on payload_lv(43,1) if clienthello
11552
11553 # Learn on response if server hello.
11554 stick store-response payload_lv(43,1) if serverhello
Cyril Bontédc4d9032012-04-08 21:57:39 +020011555
Emeric Brun6a1cefa2010-09-24 18:15:17 +020011556 server s1 192.168.1.1:443
11557 server s2 192.168.1.1:443
11558
11559 See also : "stick-table", "stick on", and section 7 about ACLs and pattern
11560 extraction.
11561
11562
Christopher Faulet4f5c2e22020-04-23 15:22:33 +020011563tcp-check comment <string>
11564 Defines a comment for the following the tcp-check rule, reported in logs if
11565 it fails.
11566 May be used in sections : defaults | frontend | listen | backend
11567 yes | no | yes | yes
11568
Christopher Faulet4f5c2e22020-04-23 15:22:33 +020011569 Arguments :
11570 <string> is the comment message to add in logs if the following tcp-check
11571 rule fails.
11572
Christopher Fauletc52ea4d2020-04-23 15:43:35 +020011573 It only works for connect, send and expect rules. It is useful to make
11574 user-friendly error reporting.
11575
Christopher Faulet4f5c2e22020-04-23 15:22:33 +020011576 See also : "option tcp-check", "tcp-check connect", "tcp-check send" and
11577 "tcp-check expect".
11578
11579
Christopher Fauletc52ea4d2020-04-23 15:43:35 +020011580tcp-check connect [default] [port <expr>] [addr <ip>] [send-proxy] [via-socks4]
11581 [ssl] [sni <sni>] [alpn <alpn>] [linger]
Christopher Fauletedc6ed92020-04-23 16:27:59 +020011582 [proto <name>] [comment <msg>]
Willy Tarreau938c7fe2014-04-25 14:21:39 +020011583 Opens a new connection
11584 May be used in sections: defaults | frontend | listen | backend
Christopher Faulet404f9192020-04-09 23:13:54 +020011585 yes | no | yes | yes
Willy Tarreau938c7fe2014-04-25 14:21:39 +020011586
Christopher Fauletc52ea4d2020-04-23 15:43:35 +020011587 Arguments :
Christopher Faulet4f5c2e22020-04-23 15:22:33 +020011588 comment <msg> defines a message to report if the rule evaluation fails.
11589
Christopher Faulet4dce5922020-03-30 13:54:42 +020011590 default Use default options of the server line to do the health
Daniel Corbett67a82712020-07-06 23:01:19 -040011591 checks. The server options are used only if not redefined.
Christopher Faulet4dce5922020-03-30 13:54:42 +020011592
Christopher Fauletb7d30092020-03-30 15:19:03 +020011593 port <expr> if not set, check port or server port is used.
Christopher Faulet5c288742020-03-31 08:15:58 +020011594 It tells HAProxy where to open the connection to.
11595 <port> must be a valid TCP port source integer, from 1 to
Christopher Fauletb7d30092020-03-30 15:19:03 +020011596 65535 or an sample-fetch expression.
Christopher Faulet5c288742020-03-31 08:15:58 +020011597
11598 addr <ip> defines the IP address to do the health check.
Willy Tarreau938c7fe2014-04-25 14:21:39 +020011599
11600 send-proxy send a PROXY protocol string
11601
Christopher Faulet085426a2020-03-30 13:07:02 +020011602 via-socks4 enables outgoing health checks using upstream socks4 proxy.
11603
Willy Tarreau938c7fe2014-04-25 14:21:39 +020011604 ssl opens a ciphered connection
11605
Christopher Faulet79b31d42020-03-30 13:00:05 +020011606 sni <sni> specifies the SNI to use to do health checks over SSL.
11607
Christopher Faulet98572322020-03-30 13:16:44 +020011608 alpn <alpn> defines which protocols to advertise with ALPN. The protocol
11609 list consists in a comma-delimited list of protocol names,
11610 for instance: "http/1.1,http/1.0" (without quotes).
11611 If it is not set, the server ALPN is used.
11612
Christopher Fauletedc6ed92020-04-23 16:27:59 +020011613 proto <name> forces the multiplexer's protocol to use for this connection.
11614 It must be a TCP mux protocol and it must be usable on the
11615 backend side. The list of available protocols is reported in
11616 haproxy -vv.
11617
Christopher Faulet5c288742020-03-31 08:15:58 +020011618 linger cleanly close the connection instead of using a single RST.
Gaetan Rivetf8ba6772020-02-07 15:37:17 +010011619
Christopher Fauletc52ea4d2020-04-23 15:43:35 +020011620 When an application lies on more than a single TCP port or when HAProxy
11621 load-balance many services in a single backend, it makes sense to probe all
11622 the services individually before considering a server as operational.
11623
11624 When there are no TCP port configured on the server line neither server port
11625 directive, then the 'tcp-check connect port <port>' must be the first step
11626 of the sequence.
11627
11628 In a tcp-check ruleset a 'connect' is required, it is also mandatory to start
11629 the ruleset with a 'connect' rule. Purpose is to ensure admin know what they
11630 do.
11631
11632 When a connect must start the ruleset, if may still be preceded by set-var,
11633 unset-var or comment rules.
11634
11635 Examples :
Willy Tarreau938c7fe2014-04-25 14:21:39 +020011636 # check HTTP and HTTPs services on a server.
11637 # first open port 80 thanks to server line port directive, then
11638 # tcp-check opens port 443, ciphered and run a request on it:
11639 option tcp-check
11640 tcp-check connect
11641 tcp-check send GET\ /\ HTTP/1.0\r\n
11642 tcp-check send Host:\ haproxy.1wt.eu\r\n
11643 tcp-check send \r\n
11644 tcp-check expect rstring (2..|3..)
11645 tcp-check connect port 443 ssl
11646 tcp-check send GET\ /\ HTTP/1.0\r\n
11647 tcp-check send Host:\ haproxy.1wt.eu\r\n
11648 tcp-check send \r\n
11649 tcp-check expect rstring (2..|3..)
11650 server www 10.0.0.1 check port 80
11651
11652 # check both POP and IMAP from a single server:
11653 option tcp-check
Gaetan Rivetf8ba6772020-02-07 15:37:17 +010011654 tcp-check connect port 110 linger
Willy Tarreau938c7fe2014-04-25 14:21:39 +020011655 tcp-check expect string +OK\ POP3\ ready
11656 tcp-check connect port 143
11657 tcp-check expect string *\ OK\ IMAP4\ ready
11658 server mail 10.0.0.1 check
11659
11660 See also : "option tcp-check", "tcp-check send", "tcp-check expect"
11661
11662
Christopher Faulet4f5c2e22020-04-23 15:22:33 +020011663tcp-check expect [min-recv <int>] [comment <msg>]
Christopher Fauletec07e382020-04-07 14:56:26 +020011664 [ok-status <st>] [error-status <st>] [tout-status <st>]
Christopher Faulet98cc57c2020-04-01 20:52:31 +020011665 [on-success <fmt>] [on-error <fmt>] [status-code <expr>]
Christopher Fauletcf80f2f2020-04-01 11:04:52 +020011666 [!] <match> <pattern>
Davor Ocelice9ed2812017-12-25 17:49:28 +010011667 Specify data to be collected and analyzed during a generic health check
Willy Tarreau938c7fe2014-04-25 14:21:39 +020011668 May be used in sections: defaults | frontend | listen | backend
Christopher Faulet404f9192020-04-09 23:13:54 +020011669 yes | no | yes | yes
Willy Tarreau938c7fe2014-04-25 14:21:39 +020011670
11671 Arguments :
Christopher Faulet4f5c2e22020-04-23 15:22:33 +020011672 comment <msg> defines a message to report if the rule evaluation fails.
11673
Gaetan Rivet1afd8262020-02-07 15:37:17 +010011674 min-recv is optional and can define the minimum amount of data required to
11675 evaluate the current expect rule. If the number of received bytes
11676 is under this limit, the check will wait for more data. This
11677 option can be used to resolve some ambiguous matching rules or to
11678 avoid executing costly regex matches on content known to be still
11679 incomplete. If an exact string (string or binary) is used, the
11680 minimum between the string length and this parameter is used.
11681 This parameter is ignored if it is set to -1. If the expect rule
11682 does not match, the check will wait for more data. If set to 0,
11683 the evaluation result is always conclusive.
11684
Willy Tarreau938c7fe2014-04-25 14:21:39 +020011685 <match> is a keyword indicating how to look for a specific pattern in the
Gaetan Rivetefab6c62020-02-07 15:37:17 +010011686 response. The keyword may be one of "string", "rstring", "binary" or
11687 "rbinary".
Willy Tarreau938c7fe2014-04-25 14:21:39 +020011688 The keyword may be preceded by an exclamation mark ("!") to negate
11689 the match. Spaces are allowed between the exclamation mark and the
11690 keyword. See below for more details on the supported keywords.
11691
Christopher Fauletec07e382020-04-07 14:56:26 +020011692 ok-status <st> is optional and can be used to set the check status if
11693 the expect rule is successfully evaluated and if it is
11694 the last rule in the tcp-check ruleset. "L7OK", "L7OKC",
Christopher Fauletd888f0f2020-05-07 07:40:17 +020011695 "L6OK" and "L4OK" are supported :
11696 - L7OK : check passed on layer 7
Christopher Faulet83662b52020-11-20 17:47:47 +010011697 - L7OKC : check conditionally passed on layer 7, set
11698 server to NOLB state.
Christopher Fauletd888f0f2020-05-07 07:40:17 +020011699 - L6OK : check passed on layer 6
11700 - L4OK : check passed on layer 4
Christopher Fauletec07e382020-04-07 14:56:26 +020011701 By default "L7OK" is used.
11702
Christopher Fauletcf80f2f2020-04-01 11:04:52 +020011703 error-status <st> is optional and can be used to set the check status if
11704 an error occurred during the expect rule evaluation.
Christopher Faulet83662b52020-11-20 17:47:47 +010011705 "L7OKC", "L7RSP", "L7STS", "L6RSP" and "L4CON" are
11706 supported :
11707 - L7OKC : check conditionally passed on layer 7, set
11708 server to NOLB state.
Christopher Fauletd888f0f2020-05-07 07:40:17 +020011709 - L7RSP : layer 7 invalid response - protocol error
11710 - L7STS : layer 7 response error, for example HTTP 5xx
11711 - L6RSP : layer 6 invalid response - protocol error
11712 - L4CON : layer 1-4 connection problem
11713 By default "L7RSP" is used.
Christopher Fauletcf80f2f2020-04-01 11:04:52 +020011714
Christopher Fauletec07e382020-04-07 14:56:26 +020011715 tout-status <st> is optional and can be used to set the check status if
Christopher Fauletcf80f2f2020-04-01 11:04:52 +020011716 a timeout occurred during the expect rule evaluation.
Christopher Fauletd888f0f2020-05-07 07:40:17 +020011717 "L7TOUT", "L6TOUT", and "L4TOUT" are supported :
11718 - L7TOUT : layer 7 (HTTP/SMTP) timeout
11719 - L6TOUT : layer 6 (SSL) timeout
11720 - L4TOUT : layer 1-4 timeout
Christopher Fauletcf80f2f2020-04-01 11:04:52 +020011721 By default "L7TOUT" is used.
11722
Christopher Fauletbe52b4d2020-04-01 16:30:22 +020011723 on-success <fmt> is optional and can be used to customize the
11724 informational message reported in logs if the expect
11725 rule is successfully evaluated and if it is the last rule
11726 in the tcp-check ruleset. <fmt> is a log-format string.
11727
11728 on-error <fmt> is optional and can be used to customize the
11729 informational message reported in logs if an error
11730 occurred during the expect rule evaluation. <fmt> is a
11731 log-format string.
11732
Christopher Faulet98cc57c2020-04-01 20:52:31 +020011733 status-code <expr> is optional and can be used to set the check status code
11734 reported in logs, on success or on error. <expr> is a
11735 standard HAProxy expression formed by a sample-fetch
11736 followed by some converters.
11737
Willy Tarreau938c7fe2014-04-25 14:21:39 +020011738 <pattern> is the pattern to look for. It may be a string or a regular
11739 expression. If the pattern contains spaces, they must be escaped
11740 with the usual backslash ('\').
11741 If the match is set to binary, then the pattern must be passed as
Davor Ocelice9ed2812017-12-25 17:49:28 +010011742 a series of hexadecimal digits in an even number. Each sequence of
Willy Tarreau938c7fe2014-04-25 14:21:39 +020011743 two digits will represent a byte. The hexadecimal digits may be
11744 used upper or lower case.
11745
Willy Tarreau938c7fe2014-04-25 14:21:39 +020011746 The available matches are intentionally similar to their http-check cousins :
11747
11748 string <string> : test the exact string matches in the response buffer.
11749 A health check response will be considered valid if the
11750 response's buffer contains this exact string. If the
11751 "string" keyword is prefixed with "!", then the response
11752 will be considered invalid if the body contains this
11753 string. This can be used to look for a mandatory pattern
11754 in a protocol response, or to detect a failure when a
11755 specific error appears in a protocol banner.
11756
11757 rstring <regex> : test a regular expression on the response buffer.
11758 A health check response will be considered valid if the
11759 response's buffer matches this expression. If the
11760 "rstring" keyword is prefixed with "!", then the response
11761 will be considered invalid if the body matches the
11762 expression.
11763
Christopher Fauletaaab0832020-05-05 15:54:22 +020011764 string-lf <fmt> : test a log-format string match in the response's buffer.
11765 A health check response will be considered valid if the
11766 response's buffer contains the string resulting of the
11767 evaluation of <fmt>, which follows the log-format rules.
11768 If prefixed with "!", then the response will be
11769 considered invalid if the buffer contains the string.
11770
Willy Tarreau938c7fe2014-04-25 14:21:39 +020011771 binary <hexstring> : test the exact string in its hexadecimal form matches
11772 in the response buffer. A health check response will
11773 be considered valid if the response's buffer contains
11774 this exact hexadecimal string.
11775 Purpose is to match data on binary protocols.
11776
Gaetan Rivetefab6c62020-02-07 15:37:17 +010011777 rbinary <regex> : test a regular expression on the response buffer, like
11778 "rstring". However, the response buffer is transformed
11779 into its hexadecimal form, including NUL-bytes. This
11780 allows using all regex engines to match any binary
11781 content. The hexadecimal transformation takes twice the
11782 size of the original response. As such, the expected
11783 pattern should work on at-most half the response buffer
11784 size.
11785
Christopher Fauletaaab0832020-05-05 15:54:22 +020011786 binary-lf <hexfmt> : test a log-format string in its hexadecimal form
11787 match in the response's buffer. A health check response
11788 will be considered valid if the response's buffer
11789 contains the hexadecimal string resulting of the
11790 evaluation of <fmt>, which follows the log-format
11791 rules. If prefixed with "!", then the response will be
11792 considered invalid if the buffer contains the
11793 hexadecimal string. The hexadecimal string is converted
11794 in a binary string before matching the response's
11795 buffer.
11796
Willy Tarreau938c7fe2014-04-25 14:21:39 +020011797 It is important to note that the responses will be limited to a certain size
Christopher Fauletbb9fb8b2020-11-25 17:20:57 +010011798 defined by the global "tune.bufsize" option, which defaults to 16384 bytes.
Willy Tarreau938c7fe2014-04-25 14:21:39 +020011799 Thus, too large responses may not contain the mandatory pattern when using
11800 "string", "rstring" or binary. If a large response is absolutely required, it
11801 is possible to change the default max size by setting the global variable.
11802 However, it is worth keeping in mind that parsing very large responses can
11803 waste some CPU cycles, especially when regular expressions are used, and that
11804 it is always better to focus the checks on smaller resources. Also, in its
11805 current state, the check will not find any string nor regex past a null
11806 character in the response. Similarly it is not possible to request matching
11807 the null character.
11808
11809 Examples :
11810 # perform a POP check
11811 option tcp-check
11812 tcp-check expect string +OK\ POP3\ ready
11813
11814 # perform an IMAP check
11815 option tcp-check
11816 tcp-check expect string *\ OK\ IMAP4\ ready
11817
11818 # look for the redis master server
11819 option tcp-check
11820 tcp-check send PING\r\n
Baptiste Assmanna3322992015-08-04 10:12:18 +020011821 tcp-check expect string +PONG
Willy Tarreau938c7fe2014-04-25 14:21:39 +020011822 tcp-check send info\ replication\r\n
11823 tcp-check expect string role:master
11824 tcp-check send QUIT\r\n
11825 tcp-check expect string +OK
11826
11827
11828 See also : "option tcp-check", "tcp-check connect", "tcp-check send",
Christopher Fauletbb9fb8b2020-11-25 17:20:57 +010011829 "tcp-check send-binary", "http-check expect", tune.bufsize
Willy Tarreau938c7fe2014-04-25 14:21:39 +020011830
11831
Christopher Fauletb50b3e62020-05-05 18:43:43 +020011832tcp-check send <data> [comment <msg>]
11833tcp-check send-lf <fmt> [comment <msg>]
11834 Specify a string or a log-format string to be sent as a question during a
11835 generic health check
Willy Tarreau938c7fe2014-04-25 14:21:39 +020011836 May be used in sections: defaults | frontend | listen | backend
Christopher Faulet404f9192020-04-09 23:13:54 +020011837 yes | no | yes | yes
Willy Tarreau938c7fe2014-04-25 14:21:39 +020011838
Christopher Faulet4f5c2e22020-04-23 15:22:33 +020011839 Arguments :
11840 comment <msg> defines a message to report if the rule evaluation fails.
11841
Christopher Fauletb50b3e62020-05-05 18:43:43 +020011842 <data> is the string that will be sent during a generic health
11843 check session.
Christopher Faulet16fff672020-04-30 07:50:54 +020011844
Christopher Fauletb50b3e62020-05-05 18:43:43 +020011845 <fmt> is the log-format string that will be sent, once evaluated,
11846 during a generic health check session.
Willy Tarreau938c7fe2014-04-25 14:21:39 +020011847
11848 Examples :
11849 # look for the redis master server
11850 option tcp-check
11851 tcp-check send info\ replication\r\n
11852 tcp-check expect string role:master
11853
11854 See also : "option tcp-check", "tcp-check connect", "tcp-check expect",
Christopher Fauletbb9fb8b2020-11-25 17:20:57 +010011855 "tcp-check send-binary", tune.bufsize
Willy Tarreau938c7fe2014-04-25 14:21:39 +020011856
11857
Christopher Fauletb50b3e62020-05-05 18:43:43 +020011858tcp-check send-binary <hexstring> [comment <msg>]
11859tcp-check send-binary-lf <hexfmt> [comment <msg>]
11860 Specify an hex digits string or an hex digits log-format string to be sent as
11861 a binary question during a raw tcp health check
Willy Tarreau938c7fe2014-04-25 14:21:39 +020011862 May be used in sections: defaults | frontend | listen | backend
Christopher Faulet404f9192020-04-09 23:13:54 +020011863 yes | no | yes | yes
Willy Tarreau938c7fe2014-04-25 14:21:39 +020011864
Christopher Faulet4f5c2e22020-04-23 15:22:33 +020011865 Arguments :
11866 comment <msg> defines a message to report if the rule evaluation fails.
Christopher Fauletc52ea4d2020-04-23 15:43:35 +020011867
Christopher Fauletb50b3e62020-05-05 18:43:43 +020011868 <hexstring> is the hexadecimal string that will be send, once converted
11869 to binary, during a generic health check session.
Christopher Faulet16fff672020-04-30 07:50:54 +020011870
Christopher Fauletb50b3e62020-05-05 18:43:43 +020011871 <hexfmt> is the hexadecimal log-format string that will be send, once
11872 evaluated and converted to binary, during a generic health
11873 check session.
Willy Tarreau938c7fe2014-04-25 14:21:39 +020011874
11875 Examples :
11876 # redis check in binary
11877 option tcp-check
11878 tcp-check send-binary 50494e470d0a # PING\r\n
11879 tcp-check expect binary 2b504F4e47 # +PONG
11880
11881
11882 See also : "option tcp-check", "tcp-check connect", "tcp-check expect",
Christopher Fauletbb9fb8b2020-11-25 17:20:57 +010011883 "tcp-check send", tune.bufsize
Willy Tarreau938c7fe2014-04-25 14:21:39 +020011884
11885
Gaetan Rivet0c39ecc2020-02-24 17:34:11 +010011886tcp-check set-var(<var-name>) <expr>
Gaetan Rivet0c39ecc2020-02-24 17:34:11 +010011887 This operation sets the content of a variable. The variable is declared inline.
Gaetan Rivet0c39ecc2020-02-24 17:34:11 +010011888 May be used in sections: defaults | frontend | listen | backend
Christopher Faulet404f9192020-04-09 23:13:54 +020011889 yes | no | yes | yes
Gaetan Rivet0c39ecc2020-02-24 17:34:11 +010011890
Christopher Fauletc52ea4d2020-04-23 15:43:35 +020011891 Arguments :
Gaetan Rivet0c39ecc2020-02-24 17:34:11 +010011892 <var-name> The name of the variable starts with an indication about its
11893 scope. The scopes allowed for tcp-check are:
11894 "proc" : the variable is shared with the whole process.
11895 "sess" : the variable is shared with the tcp-check session.
11896 "check": the variable is declared for the lifetime of the tcp-check.
11897 This prefix is followed by a name. The separator is a '.'.
11898 The name may only contain characters 'a-z', 'A-Z', '0-9', '.',
11899 and '-'.
11900
11901 <expr> Is a sample-fetch expression potentially followed by converters.
11902
Christopher Fauletc52ea4d2020-04-23 15:43:35 +020011903 Examples :
Gaetan Rivet0c39ecc2020-02-24 17:34:11 +010011904 tcp-check set-var(check.port) int(1234)
11905
11906
11907tcp-check unset-var(<var-name>)
Gaetan Rivet0c39ecc2020-02-24 17:34:11 +010011908 Free a reference to a variable within its scope.
Gaetan Rivet0c39ecc2020-02-24 17:34:11 +010011909 May be used in sections: defaults | frontend | listen | backend
Christopher Faulet404f9192020-04-09 23:13:54 +020011910 yes | no | yes | yes
Gaetan Rivet0c39ecc2020-02-24 17:34:11 +010011911
Christopher Fauletc52ea4d2020-04-23 15:43:35 +020011912 Arguments :
Gaetan Rivet0c39ecc2020-02-24 17:34:11 +010011913 <var-name> The name of the variable starts with an indication about its
11914 scope. The scopes allowed for tcp-check are:
11915 "proc" : the variable is shared with the whole process.
11916 "sess" : the variable is shared with the tcp-check session.
11917 "check": the variable is declared for the lifetime of the tcp-check.
11918 This prefix is followed by a name. The separator is a '.'.
11919 The name may only contain characters 'a-z', 'A-Z', '0-9', '.',
11920 and '-'.
11921
Christopher Fauletc52ea4d2020-04-23 15:43:35 +020011922 Examples :
Gaetan Rivet0c39ecc2020-02-24 17:34:11 +010011923 tcp-check unset-var(check.port)
11924
11925
Willy Tarreaue9656522010-08-17 15:40:09 +020011926tcp-request connection <action> [{if | unless} <condition>]
11927 Perform an action on an incoming connection depending on a layer 4 condition
Willy Tarreau1a687942010-05-23 22:40:30 +020011928 May be used in sections : defaults | frontend | listen | backend
11929 no | yes | yes | no
Willy Tarreaue9656522010-08-17 15:40:09 +020011930 Arguments :
Willy Tarreauc870bfd2015-09-28 18:47:38 +020011931 <action> defines the action to perform if the condition applies. See
11932 below.
Willy Tarreau1a687942010-05-23 22:40:30 +020011933
Willy Tarreaue9656522010-08-17 15:40:09 +020011934 <condition> is a standard layer4-only ACL-based condition (see section 7).
Willy Tarreau68c03ab2010-08-06 15:08:45 +020011935
11936 Immediately after acceptance of a new incoming connection, it is possible to
11937 evaluate some conditions to decide whether this connection must be accepted
Willy Tarreaue9656522010-08-17 15:40:09 +020011938 or dropped or have its counters tracked. Those conditions cannot make use of
11939 any data contents because the connection has not been read from yet, and the
11940 buffers are not yet allocated. This is used to selectively and very quickly
11941 accept or drop connections from various sources with a very low overhead. If
11942 some contents need to be inspected in order to take the decision, the
11943 "tcp-request content" statements must be used instead.
Willy Tarreau68c03ab2010-08-06 15:08:45 +020011944
Willy Tarreaue9656522010-08-17 15:40:09 +020011945 The "tcp-request connection" rules are evaluated in their exact declaration
11946 order. If no rule matches or if there is no rule, the default action is to
11947 accept the incoming connection. There is no specific limit to the number of
11948 rules which may be inserted.
Willy Tarreau68c03ab2010-08-06 15:08:45 +020011949
Willy Tarreaua9083d02015-05-08 15:27:59 +020011950 Four types of actions are supported :
Willy Tarreaue9656522010-08-17 15:40:09 +020011951 - accept :
11952 accepts the connection if the condition is true (when used with "if")
11953 or false (when used with "unless"). The first such rule executed ends
11954 the rules evaluation.
Willy Tarreau68c03ab2010-08-06 15:08:45 +020011955
Willy Tarreaue9656522010-08-17 15:40:09 +020011956 - reject :
11957 rejects the connection if the condition is true (when used with "if")
11958 or false (when used with "unless"). The first such rule executed ends
11959 the rules evaluation. Rejected connections do not even become a
11960 session, which is why they are accounted separately for in the stats,
11961 as "denied connections". They are not considered for the session
11962 rate-limit and are not logged either. The reason is that these rules
11963 should only be used to filter extremely high connection rates such as
11964 the ones encountered during a massive DDoS attack. Under these extreme
11965 conditions, the simple action of logging each event would make the
11966 system collapse and would considerably lower the filtering capacity. If
11967 logging is absolutely desired, then "tcp-request content" rules should
Willy Tarreau4f614292016-10-21 17:49:36 +020011968 be used instead, as "tcp-request session" rules will not log either.
Willy Tarreau68c03ab2010-08-06 15:08:45 +020011969
Willy Tarreau4f0d9192013-06-11 20:40:55 +020011970 - expect-proxy layer4 :
11971 configures the client-facing connection to receive a PROXY protocol
11972 header before any byte is read from the socket. This is equivalent to
11973 having the "accept-proxy" keyword on the "bind" line, except that using
11974 the TCP rule allows the PROXY protocol to be accepted only for certain
11975 IP address ranges using an ACL. This is convenient when multiple layers
11976 of load balancers are passed through by traffic coming from public
11977 hosts.
11978
Bertrand Jacquin90759682016-06-06 15:35:39 +010011979 - expect-netscaler-cip layer4 :
11980 configures the client-facing connection to receive a NetScaler Client
11981 IP insertion protocol header before any byte is read from the socket.
11982 This is equivalent to having the "accept-netscaler-cip" keyword on the
11983 "bind" line, except that using the TCP rule allows the PROXY protocol
11984 to be accepted only for certain IP address ranges using an ACL. This
11985 is convenient when multiple layers of load balancers are passed
11986 through by traffic coming from public hosts.
11987
Willy Tarreau18bf01e2014-06-13 16:18:52 +020011988 - capture <sample> len <length> :
11989 This only applies to "tcp-request content" rules. It captures sample
11990 expression <sample> from the request buffer, and converts it to a
11991 string of at most <len> characters. The resulting string is stored into
11992 the next request "capture" slot, so it will possibly appear next to
11993 some captured HTTP headers. It will then automatically appear in the
11994 logs, and it will be possible to extract it using sample fetch rules to
11995 feed it into headers or anything. The length should be limited given
11996 that this size will be allocated for each capture during the whole
Willy Tarreaua9083d02015-05-08 15:27:59 +020011997 session life. Please check section 7.3 (Fetching samples) and "capture
11998 request header" for more information.
Willy Tarreau18bf01e2014-06-13 16:18:52 +020011999
Willy Tarreaube4a3ef2013-06-17 15:04:07 +020012000 - { track-sc0 | track-sc1 | track-sc2 } <key> [table <table>] :
Willy Tarreaue9656522010-08-17 15:40:09 +020012001 enables tracking of sticky counters from current connection. These
Moemen MHEDHBI9cf46342018-09-25 17:50:53 +020012002 rules do not stop evaluation and do not change default action. The
12003 number of counters that may be simultaneously tracked by the same
12004 connection is set in MAX_SESS_STKCTR at build time (reported in
John Roeslerfb2fce12019-07-10 15:45:51 -050012005 haproxy -vv) which defaults to 3, so the track-sc number is between 0
Matteo Contrini1857b8c2020-10-16 17:35:54 +020012006 and (MAX_SESS_STKCTR-1). The first "track-sc0" rule executed enables
Moemen MHEDHBI9cf46342018-09-25 17:50:53 +020012007 tracking of the counters of the specified table as the first set. The
12008 first "track-sc1" rule executed enables tracking of the counters of the
12009 specified table as the second set. The first "track-sc2" rule executed
12010 enables tracking of the counters of the specified table as the third
12011 set. It is a recommended practice to use the first set of counters for
12012 the per-frontend counters and the second set for the per-backend ones.
12013 But this is just a guideline, all may be used everywhere.
Willy Tarreau68c03ab2010-08-06 15:08:45 +020012014
Willy Tarreaue9656522010-08-17 15:40:09 +020012015 These actions take one or two arguments :
Willy Tarreaube722a22014-06-13 16:31:59 +020012016 <key> is mandatory, and is a sample expression rule as described
Willy Tarreau74ca5042013-06-11 23:12:07 +020012017 in section 7.3. It describes what elements of the incoming
Davor Ocelice9ed2812017-12-25 17:49:28 +010012018 request or connection will be analyzed, extracted, combined,
Willy Tarreau5d5b5d82012-12-09 12:00:04 +010012019 and used to select which table entry to update the counters.
12020 Note that "tcp-request connection" cannot use content-based
12021 fetches.
Willy Tarreau68c03ab2010-08-06 15:08:45 +020012022
Willy Tarreaue9656522010-08-17 15:40:09 +020012023 <table> is an optional table to be used instead of the default one,
12024 which is the stick-table declared in the current proxy. All
12025 the counters for the matches and updates for the key will
12026 then be performed in that table until the session ends.
Willy Tarreau68c03ab2010-08-06 15:08:45 +020012027
Willy Tarreaue9656522010-08-17 15:40:09 +020012028 Once a "track-sc*" rule is executed, the key is looked up in the table
12029 and if it is not found, an entry is allocated for it. Then a pointer to
12030 that entry is kept during all the session's life, and this entry's
12031 counters are updated as often as possible, every time the session's
12032 counters are updated, and also systematically when the session ends.
Willy Tarreau5d5b5d82012-12-09 12:00:04 +010012033 Counters are only updated for events that happen after the tracking has
12034 been started. For example, connection counters will not be updated when
12035 tracking layer 7 information, since the connection event happens before
12036 layer7 information is extracted.
12037
Willy Tarreaue9656522010-08-17 15:40:09 +020012038 If the entry tracks concurrent connection counters, one connection is
12039 counted for as long as the entry is tracked, and the entry will not
12040 expire during that time. Tracking counters also provides a performance
12041 advantage over just checking the keys, because only one table lookup is
12042 performed for all ACL checks that make use of it.
Willy Tarreau68c03ab2010-08-06 15:08:45 +020012043
Thierry FOURNIERe0627bd2015-08-04 08:20:33 +020012044 - sc-inc-gpc0(<sc-id>):
12045 The "sc-inc-gpc0" increments the GPC0 counter according to the sticky
12046 counter designated by <sc-id>. If an error occurs, this action silently
12047 fails and the actions evaluation continues.
12048
Frédéric Lécaille6778b272018-01-29 15:22:53 +010012049 - sc-inc-gpc1(<sc-id>):
12050 The "sc-inc-gpc1" increments the GPC1 counter according to the sticky
12051 counter designated by <sc-id>. If an error occurs, this action silently
12052 fails and the actions evaluation continues.
12053
Cédric Dufour0d7712d2019-11-06 18:38:53 +010012054 - sc-set-gpt0(<sc-id>) { <int> | <expr> }:
12055 This action sets the 32-bit unsigned GPT0 tag according to the sticky
12056 counter designated by <sc-id> and the value of <int>/<expr>. The
12057 expected result is a boolean. If an error occurs, this action silently
12058 fails and the actions evaluation continues.
Thierry FOURNIER236657b2015-08-19 08:25:14 +020012059
William Lallemand2e785f22016-05-25 01:48:42 +020012060 - set-src <expr> :
12061 Is used to set the source IP address to the value of specified
12062 expression. Useful if you want to mask source IP for privacy.
12063 If you want to provide an IP from a HTTP header use "http-request
Cyril Bonté6c81d5f2018-10-17 00:14:51 +020012064 set-src".
William Lallemand2e785f22016-05-25 01:48:42 +020012065
Cyril Bonté6c81d5f2018-10-17 00:14:51 +020012066 Arguments:
12067 <expr> Is a standard HAProxy expression formed by a sample-fetch
12068 followed by some converters.
William Lallemand2e785f22016-05-25 01:48:42 +020012069
12070 Example:
William Lallemand2e785f22016-05-25 01:48:42 +020012071 tcp-request connection set-src src,ipmask(24)
12072
Willy Tarreau0c630532016-10-21 17:52:58 +020012073 When possible, set-src preserves the original source port as long as the
12074 address family allows it, otherwise the source port is set to 0.
William Lallemand2e785f22016-05-25 01:48:42 +020012075
William Lallemand44be6402016-05-25 01:51:35 +020012076 - set-src-port <expr> :
12077 Is used to set the source port address to the value of specified
12078 expression.
12079
Cyril Bonté6c81d5f2018-10-17 00:14:51 +020012080 Arguments:
12081 <expr> Is a standard HAProxy expression formed by a sample-fetch
12082 followed by some converters.
William Lallemand44be6402016-05-25 01:51:35 +020012083
12084 Example:
William Lallemand44be6402016-05-25 01:51:35 +020012085 tcp-request connection set-src-port int(4000)
12086
Willy Tarreau0c630532016-10-21 17:52:58 +020012087 When possible, set-src-port preserves the original source address as long
12088 as the address family supports a port, otherwise it forces the source
12089 address to IPv4 "0.0.0.0" before rewriting the port.
William Lallemand44be6402016-05-25 01:51:35 +020012090
William Lallemand13e9b0c2016-05-25 02:34:07 +020012091 - set-dst <expr> :
12092 Is used to set the destination IP address to the value of specified
12093 expression. Useful if you want to mask IP for privacy in log.
12094 If you want to provide an IP from a HTTP header use "http-request
12095 set-dst". If you want to connect to the new address/port, use
12096 '0.0.0.0:0' as a server address in the backend.
12097
12098 <expr> Is a standard HAProxy expression formed by a sample-fetch
12099 followed by some converters.
12100
12101 Example:
12102
12103 tcp-request connection set-dst dst,ipmask(24)
12104 tcp-request connection set-dst ipv4(10.0.0.1)
12105
Willy Tarreau0c630532016-10-21 17:52:58 +020012106 When possible, set-dst preserves the original destination port as long as
12107 the address family allows it, otherwise the destination port is set to 0.
12108
William Lallemand13e9b0c2016-05-25 02:34:07 +020012109 - set-dst-port <expr> :
12110 Is used to set the destination port address to the value of specified
12111 expression. If you want to connect to the new address/port, use
12112 '0.0.0.0:0' as a server address in the backend.
12113
12114
12115 <expr> Is a standard HAProxy expression formed by a sample-fetch
12116 followed by some converters.
12117
12118 Example:
12119
12120 tcp-request connection set-dst-port int(4000)
12121
Willy Tarreau0c630532016-10-21 17:52:58 +020012122 When possible, set-dst-port preserves the original destination address as
12123 long as the address family supports a port, otherwise it forces the
12124 destination address to IPv4 "0.0.0.0" before rewriting the port.
12125
Willy Tarreau2d392c22015-08-24 01:43:45 +020012126 - "silent-drop" :
12127 This stops the evaluation of the rules and makes the client-facing
Davor Ocelice9ed2812017-12-25 17:49:28 +010012128 connection suddenly disappear using a system-dependent way that tries
Willy Tarreau2d392c22015-08-24 01:43:45 +020012129 to prevent the client from being notified. The effect it then that the
12130 client still sees an established connection while there's none on
12131 HAProxy. The purpose is to achieve a comparable effect to "tarpit"
12132 except that it doesn't use any local resource at all on the machine
12133 running HAProxy. It can resist much higher loads than "tarpit", and
Davor Ocelice9ed2812017-12-25 17:49:28 +010012134 slow down stronger attackers. It is important to understand the impact
12135 of using this mechanism. All stateful equipment placed between the
Willy Tarreau2d392c22015-08-24 01:43:45 +020012136 client and HAProxy (firewalls, proxies, load balancers) will also keep
12137 the established connection for a long time and may suffer from this
Davor Ocelice9ed2812017-12-25 17:49:28 +010012138 action. On modern Linux systems running with enough privileges, the
Willy Tarreau2d392c22015-08-24 01:43:45 +020012139 TCP_REPAIR socket option is used to block the emission of a TCP
12140 reset. On other systems, the socket's TTL is reduced to 1 so that the
12141 TCP reset doesn't pass the first router, though it's still delivered to
12142 local networks. Do not use it unless you fully understand how it works.
12143
Willy Tarreaue9656522010-08-17 15:40:09 +020012144 Note that the "if/unless" condition is optional. If no condition is set on
12145 the action, it is simply performed unconditionally. That can be useful for
12146 "track-sc*" actions as well as for changing the default action to a reject.
Willy Tarreau68c03ab2010-08-06 15:08:45 +020012147
Willy Tarreaue9656522010-08-17 15:40:09 +020012148 Example: accept all connections from white-listed hosts, reject too fast
12149 connection without counting them, and track accepted connections.
12150 This results in connection rate being capped from abusive sources.
Willy Tarreau68c03ab2010-08-06 15:08:45 +020012151
Willy Tarreaue9656522010-08-17 15:40:09 +020012152 tcp-request connection accept if { src -f /etc/haproxy/whitelist.lst }
Willy Tarreau68c03ab2010-08-06 15:08:45 +020012153 tcp-request connection reject if { src_conn_rate gt 10 }
Willy Tarreaube4a3ef2013-06-17 15:04:07 +020012154 tcp-request connection track-sc0 src
Willy Tarreau68c03ab2010-08-06 15:08:45 +020012155
Willy Tarreaue9656522010-08-17 15:40:09 +020012156 Example: accept all connections from white-listed hosts, count all other
12157 connections and reject too fast ones. This results in abusive ones
12158 being blocked as long as they don't slow down.
Willy Tarreau68c03ab2010-08-06 15:08:45 +020012159
Willy Tarreaue9656522010-08-17 15:40:09 +020012160 tcp-request connection accept if { src -f /etc/haproxy/whitelist.lst }
Willy Tarreaube4a3ef2013-06-17 15:04:07 +020012161 tcp-request connection track-sc0 src
12162 tcp-request connection reject if { sc0_conn_rate gt 10 }
Willy Tarreau68c03ab2010-08-06 15:08:45 +020012163
Willy Tarreau4f0d9192013-06-11 20:40:55 +020012164 Example: enable the PROXY protocol for traffic coming from all known proxies.
12165
12166 tcp-request connection expect-proxy layer4 if { src -f proxies.lst }
12167
Willy Tarreau68c03ab2010-08-06 15:08:45 +020012168 See section 7 about ACL usage.
12169
Willy Tarreau4f614292016-10-21 17:49:36 +020012170 See also : "tcp-request session", "tcp-request content", "stick-table"
Willy Tarreau68c03ab2010-08-06 15:08:45 +020012171
12172
Willy Tarreaue9656522010-08-17 15:40:09 +020012173tcp-request content <action> [{if | unless} <condition>]
12174 Perform an action on a new session depending on a layer 4-7 condition
Willy Tarreau62644772008-07-16 18:36:06 +020012175 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaufb356202010-08-03 14:02:05 +020012176 no | yes | yes | yes
Willy Tarreaue9656522010-08-17 15:40:09 +020012177 Arguments :
Willy Tarreauc870bfd2015-09-28 18:47:38 +020012178 <action> defines the action to perform if the condition applies. See
12179 below.
Willy Tarreau62644772008-07-16 18:36:06 +020012180
Willy Tarreaue9656522010-08-17 15:40:09 +020012181 <condition> is a standard layer 4-7 ACL-based condition (see section 7).
Willy Tarreau62644772008-07-16 18:36:06 +020012182
Davor Ocelice9ed2812017-12-25 17:49:28 +010012183 A request's contents can be analyzed at an early stage of request processing
Willy Tarreaue9656522010-08-17 15:40:09 +020012184 called "TCP content inspection". During this stage, ACL-based rules are
12185 evaluated every time the request contents are updated, until either an
Christopher Fauletae863c62021-03-15 12:03:44 +010012186 "accept", a "reject" or a "switch-mode" rule matches, or the TCP request
12187 inspection delay expires with no matching rule.
Willy Tarreau62644772008-07-16 18:36:06 +020012188
Willy Tarreaue9656522010-08-17 15:40:09 +020012189 The first difference between these rules and "tcp-request connection" rules
12190 is that "tcp-request content" rules can make use of contents to take a
12191 decision. Most often, these decisions will consider a protocol recognition or
12192 validity. The second difference is that content-based rules can be used in
Willy Tarreauf3338342014-01-28 21:40:28 +010012193 both frontends and backends. In case of HTTP keep-alive with the client, all
Daniel Corbett9f0843f2021-05-08 10:50:37 -040012194 tcp-request content rules are evaluated again, so HAProxy keeps a record of
Willy Tarreauf3338342014-01-28 21:40:28 +010012195 what sticky counters were assigned by a "tcp-request connection" versus a
12196 "tcp-request content" rule, and flushes all the content-related ones after
12197 processing an HTTP request, so that they may be evaluated again by the rules
12198 being evaluated again for the next request. This is of particular importance
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030012199 when the rule tracks some L7 information or when it is conditioned by an
Willy Tarreauf3338342014-01-28 21:40:28 +010012200 L7-based ACL, since tracking may change between requests.
Willy Tarreau62644772008-07-16 18:36:06 +020012201
Willy Tarreaue9656522010-08-17 15:40:09 +020012202 Content-based rules are evaluated in their exact declaration order. If no
12203 rule matches or if there is no rule, the default action is to accept the
12204 contents. There is no specific limit to the number of rules which may be
12205 inserted.
Willy Tarreau62644772008-07-16 18:36:06 +020012206
Thierry FOURNIER236657b2015-08-19 08:25:14 +020012207 Several types of actions are supported :
Willy Tarreau18bf01e2014-06-13 16:18:52 +020012208 - accept : the request is accepted
Baptiste Assmann333939c2019-01-21 08:34:50 +010012209 - do-resolve: perform a DNS resolution
Willy Tarreau18bf01e2014-06-13 16:18:52 +020012210 - reject : the request is rejected and the connection is closed
12211 - capture : the specified sample expression is captured
Patrick Hemmer268a7072018-05-11 12:52:31 -040012212 - set-priority-class <expr> | set-priority-offset <expr>
Willy Tarreaube4a3ef2013-06-17 15:04:07 +020012213 - { track-sc0 | track-sc1 | track-sc2 } <key> [table <table>]
Thierry FOURNIERe0627bd2015-08-04 08:20:33 +020012214 - sc-inc-gpc0(<sc-id>)
Frédéric Lécaille6778b272018-01-29 15:22:53 +010012215 - sc-inc-gpc1(<sc-id>)
Cédric Dufour0d7712d2019-11-06 18:38:53 +010012216 - sc-set-gpt0(<sc-id>) { <int> | <expr> }
Baptiste Assmanne1afd4f2019-04-18 16:21:13 +020012217 - set-dst <expr>
12218 - set-dst-port <expr>
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020012219 - set-var(<var-name>) <expr>
Christopher Fauletae863c62021-03-15 12:03:44 +010012220 - switch-mode http [ proto <name> ]
Christopher Faulet85d79c92016-11-09 16:54:56 +010012221 - unset-var(<var-name>)
Willy Tarreau2d392c22015-08-24 01:43:45 +020012222 - silent-drop
Davor Ocelice9ed2812017-12-25 17:49:28 +010012223 - send-spoe-group <engine-name> <group-name>
Christopher Faulet579d83b2019-11-22 15:34:17 +010012224 - use-service <service-name>
Willy Tarreau62644772008-07-16 18:36:06 +020012225
Willy Tarreaue9656522010-08-17 15:40:09 +020012226 They have the same meaning as their counter-parts in "tcp-request connection"
12227 so please refer to that section for a complete description.
Baptiste Assmann333939c2019-01-21 08:34:50 +010012228 For "do-resolve" action, please check the "http-request do-resolve"
12229 configuration section.
Willy Tarreau62644772008-07-16 18:36:06 +020012230
Willy Tarreauf3338342014-01-28 21:40:28 +010012231 While there is nothing mandatory about it, it is recommended to use the
12232 track-sc0 in "tcp-request connection" rules, track-sc1 for "tcp-request
12233 content" rules in the frontend, and track-sc2 for "tcp-request content"
12234 rules in the backend, because that makes the configuration more readable
12235 and easier to troubleshoot, but this is just a guideline and all counters
12236 may be used everywhere.
Willy Tarreau62644772008-07-16 18:36:06 +020012237
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +010012238 Note that the "if/unless" condition is optional. If no condition is set on
Willy Tarreaue9656522010-08-17 15:40:09 +020012239 the action, it is simply performed unconditionally. That can be useful for
12240 "track-sc*" actions as well as for changing the default action to a reject.
Willy Tarreau62644772008-07-16 18:36:06 +020012241
Christopher Faulet2079a4a2020-10-02 11:48:57 +020012242 Note also that it is recommended to use a "tcp-request session" rule to track
12243 information that does *not* depend on Layer 7 contents, especially for HTTP
12244 frontends. Some HTTP processing are performed at the session level and may
12245 lead to an early rejection of the requests. Thus, the tracking at the content
12246 level may be disturbed in such case. A warning is emitted during startup to
12247 prevent, as far as possible, such unreliable usage.
12248
Willy Tarreaue9656522010-08-17 15:40:09 +020012249 It is perfectly possible to match layer 7 contents with "tcp-request content"
Christopher Faulet7ea509e2020-10-02 11:38:46 +020012250 rules from a TCP proxy, since HTTP-specific ACL matches are able to
12251 preliminarily parse the contents of a buffer before extracting the required
12252 data. If the buffered contents do not parse as a valid HTTP message, then the
12253 ACL does not match. The parser which is involved there is exactly the same
12254 as for all other HTTP processing, so there is no risk of parsing something
12255 differently. In an HTTP frontend or an HTTP backend, it is guaranteed that
12256 HTTP contents will always be immediately present when the rule is evaluated
12257 first because the HTTP parsing is performed in the early stages of the
12258 connection processing, at the session level. But for such proxies, using
12259 "http-request" rules is much more natural and recommended.
Willy Tarreau62644772008-07-16 18:36:06 +020012260
Willy Tarreau5d5b5d82012-12-09 12:00:04 +010012261 Tracking layer7 information is also possible provided that the information
Willy Tarreau4d54c7c2014-09-16 15:48:15 +020012262 are present when the rule is processed. The rule processing engine is able to
12263 wait until the inspect delay expires when the data to be tracked is not yet
12264 available.
Willy Tarreau5d5b5d82012-12-09 12:00:04 +010012265
Baptiste Assmanne1afd4f2019-04-18 16:21:13 +020012266 The "set-dst" and "set-dst-port" are used to set respectively the destination
12267 IP and port. More information on how to use it at "http-request set-dst".
12268
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020012269 The "set-var" is used to set the content of a variable. The variable is
Willy Tarreau4f614292016-10-21 17:49:36 +020012270 declared inline. For "tcp-request session" rules, only session-level
12271 variables can be used, without any layer7 contents.
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020012272
Daniel Schneller0b547052016-03-21 20:46:57 +010012273 <var-name> The name of the variable starts with an indication about
12274 its scope. The scopes allowed are:
Christopher Fauletff2613e2016-11-09 11:36:17 +010012275 "proc" : the variable is shared with the whole process
Daniel Schneller0b547052016-03-21 20:46:57 +010012276 "sess" : the variable is shared with the whole session
12277 "txn" : the variable is shared with the transaction
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020012278 (request and response)
Daniel Schneller0b547052016-03-21 20:46:57 +010012279 "req" : the variable is shared only during request
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020012280 processing
Daniel Schneller0b547052016-03-21 20:46:57 +010012281 "res" : the variable is shared only during response
12282 processing
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020012283 This prefix is followed by a name. The separator is a '.'.
Christopher Fauletb71557a2016-10-31 10:49:03 +010012284 The name may only contain characters 'a-z', 'A-Z', '0-9',
12285 '.' and '_'.
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020012286
12287 <expr> Is a standard HAProxy expression formed by a sample-fetch
12288 followed by some converters.
12289
Daniel Corbett9f0843f2021-05-08 10:50:37 -040012290 The "switch-mode" is used to perform a connection upgrade. Only HTTP
Christopher Fauletae863c62021-03-15 12:03:44 +010012291 upgrades are supported for now. The protocol may optionally be
12292 specified. This action is only available for a proxy with the frontend
12293 capability. The connection upgrade is immediately performed, following
12294 "tcp-request content" rules are not evaluated. This upgrade method should be
12295 preferred to the implicit one consisting to rely on the backend mode. When
12296 used, it is possible to set HTTP directives in a frontend without any
Ilya Shipitsin3df59892021-05-10 12:50:00 +050012297 warning. These directives will be conditionally evaluated if the HTTP upgrade
Christopher Fauletae863c62021-03-15 12:03:44 +010012298 is performed. However, an HTTP backend must still be selected. It remains
12299 unsupported to route an HTTP connection (upgraded or not) to a TCP server.
12300
Christopher Faulet4d37e532021-03-26 14:44:00 +010012301 See section 4 about Proxies for more details on HTTP upgrades.
12302
Christopher Faulet85d79c92016-11-09 16:54:56 +010012303 The "unset-var" is used to unset a variable. See above for details about
12304 <var-name>.
12305
Patrick Hemmer268a7072018-05-11 12:52:31 -040012306 The "set-priority-class" is used to set the queue priority class of the
12307 current request. The value must be a sample expression which converts to an
12308 integer in the range -2047..2047. Results outside this range will be
12309 truncated. The priority class determines the order in which queued requests
12310 are processed. Lower values have higher priority.
12311
12312 The "set-priority-offset" is used to set the queue priority timestamp offset
12313 of the current request. The value must be a sample expression which converts
12314 to an integer in the range -524287..524287. Results outside this range will be
12315 truncated. When a request is queued, it is ordered first by the priority
12316 class, then by the current timestamp adjusted by the given offset in
12317 milliseconds. Lower values have higher priority.
12318 Note that the resulting timestamp is is only tracked with enough precision for
12319 524,287ms (8m44s287ms). If the request is queued long enough to where the
12320 adjusted timestamp exceeds this value, it will be misidentified as highest
12321 priority. Thus it is important to set "timeout queue" to a value, where when
12322 combined with the offset, does not exceed this limit.
12323
Christopher Faulet76c09ef2017-09-21 11:03:52 +020012324 The "send-spoe-group" is used to trigger sending of a group of SPOE
12325 messages. To do so, the SPOE engine used to send messages must be defined, as
12326 well as the SPOE group to send. Of course, the SPOE engine must refer to an
12327 existing SPOE filter. If not engine name is provided on the SPOE filter line,
12328 the SPOE agent name must be used.
12329
12330 <engine-name> The SPOE engine name.
12331
12332 <group-name> The SPOE group name as specified in the engine configuration.
12333
Christopher Faulet579d83b2019-11-22 15:34:17 +010012334 The "use-service" is used to executes a TCP service which will reply to the
12335 request and stop the evaluation of the rules. This service may choose to
12336 reply by sending any valid response or it may immediately close the
12337 connection without sending anything. Outside natives services, it is possible
12338 to write your own services in Lua. No further "tcp-request" rules are
12339 evaluated.
12340
12341 Example:
12342 tcp-request content use-service lua.deny { src -f /etc/haproxy/blacklist.lst }
12343
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020012344 Example:
12345
12346 tcp-request content set-var(sess.my_var) src
Christopher Faulet85d79c92016-11-09 16:54:56 +010012347 tcp-request content unset-var(sess.my_var2)
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020012348
Willy Tarreau62644772008-07-16 18:36:06 +020012349 Example:
Willy Tarreaue9656522010-08-17 15:40:09 +020012350 # Accept HTTP requests containing a Host header saying "example.com"
Christopher Fauletae863c62021-03-15 12:03:44 +010012351 # and reject everything else. (Only works for HTTP/1 connections)
Willy Tarreaue9656522010-08-17 15:40:09 +020012352 acl is_host_com hdr(Host) -i example.com
12353 tcp-request inspect-delay 30s
Willy Tarreauc0239e02012-04-16 14:42:55 +020012354 tcp-request content accept if is_host_com
Willy Tarreaue9656522010-08-17 15:40:09 +020012355 tcp-request content reject
12356
Christopher Fauletae863c62021-03-15 12:03:44 +010012357 # Accept HTTP requests containing a Host header saying "example.com"
12358 # and reject everything else. (works for HTTP/1 and HTTP/2 connections)
12359 acl is_host_com hdr(Host) -i example.com
12360 tcp-request inspect-delay 5s
12361 tcp-request switch-mode http if HTTP
12362 tcp-request reject # non-HTTP traffic is implicit here
12363 ...
12364 http-request reject unless is_host_com
12365
Willy Tarreaue9656522010-08-17 15:40:09 +020012366 Example:
Willy Tarreau62644772008-07-16 18:36:06 +020012367 # reject SMTP connection if client speaks first
12368 tcp-request inspect-delay 30s
12369 acl content_present req_len gt 0
Willy Tarreau68c03ab2010-08-06 15:08:45 +020012370 tcp-request content reject if content_present
Willy Tarreau62644772008-07-16 18:36:06 +020012371
12372 # Forward HTTPS connection only if client speaks
12373 tcp-request inspect-delay 30s
12374 acl content_present req_len gt 0
Willy Tarreau68c03ab2010-08-06 15:08:45 +020012375 tcp-request content accept if content_present
Willy Tarreaue9656522010-08-17 15:40:09 +020012376 tcp-request content reject
12377
Willy Tarreau5d5b5d82012-12-09 12:00:04 +010012378 Example:
Jarno Huuskonene5ae7022017-04-03 14:36:21 +030012379 # Track the last IP(stick-table type string) from X-Forwarded-For
Willy Tarreau5d5b5d82012-12-09 12:00:04 +010012380 tcp-request inspect-delay 10s
Willy Tarreau4d54c7c2014-09-16 15:48:15 +020012381 tcp-request content track-sc0 hdr(x-forwarded-for,-1)
Jarno Huuskonene5ae7022017-04-03 14:36:21 +030012382 # Or track the last IP(stick-table type ip|ipv6) from X-Forwarded-For
12383 tcp-request content track-sc0 req.hdr_ip(x-forwarded-for,-1)
Willy Tarreau5d5b5d82012-12-09 12:00:04 +010012384
12385 Example:
12386 # track request counts per "base" (concatenation of Host+URL)
12387 tcp-request inspect-delay 10s
Willy Tarreau4d54c7c2014-09-16 15:48:15 +020012388 tcp-request content track-sc0 base table req-rate
Willy Tarreau5d5b5d82012-12-09 12:00:04 +010012389
Willy Tarreaue9656522010-08-17 15:40:09 +020012390 Example: track per-frontend and per-backend counters, block abusers at the
Jarno Huuskonene5ae7022017-04-03 14:36:21 +030012391 frontend when the backend detects abuse(and marks gpc0).
Willy Tarreaue9656522010-08-17 15:40:09 +020012392
12393 frontend http
Davor Ocelice9ed2812017-12-25 17:49:28 +010012394 # Use General Purpose Counter 0 in SC0 as a global abuse counter
Willy Tarreaue9656522010-08-17 15:40:09 +020012395 # protecting all our sites
12396 stick-table type ip size 1m expire 5m store gpc0
Willy Tarreaube4a3ef2013-06-17 15:04:07 +020012397 tcp-request connection track-sc0 src
12398 tcp-request connection reject if { sc0_get_gpc0 gt 0 }
Willy Tarreaue9656522010-08-17 15:40:09 +020012399 ...
12400 use_backend http_dynamic if { path_end .php }
12401
12402 backend http_dynamic
12403 # if a source makes too fast requests to this dynamic site (tracked
Willy Tarreaube4a3ef2013-06-17 15:04:07 +020012404 # by SC1), block it globally in the frontend.
Willy Tarreaue9656522010-08-17 15:40:09 +020012405 stick-table type ip size 1m expire 5m store http_req_rate(10s)
Willy Tarreaube4a3ef2013-06-17 15:04:07 +020012406 acl click_too_fast sc1_http_req_rate gt 10
Jarno Huuskonene5ae7022017-04-03 14:36:21 +030012407 acl mark_as_abuser sc0_inc_gpc0(http) gt 0
Willy Tarreaube4a3ef2013-06-17 15:04:07 +020012408 tcp-request content track-sc1 src
Willy Tarreaue9656522010-08-17 15:40:09 +020012409 tcp-request content reject if click_too_fast mark_as_abuser
Willy Tarreau62644772008-07-16 18:36:06 +020012410
Willy Tarreauc57f0e22009-05-10 13:12:33 +020012411 See section 7 about ACL usage.
Willy Tarreau62644772008-07-16 18:36:06 +020012412
Jarno Huuskonen95b012b2017-04-06 13:59:14 +030012413 See also : "tcp-request connection", "tcp-request session",
12414 "tcp-request inspect-delay", and "http-request".
Willy Tarreau62644772008-07-16 18:36:06 +020012415
12416
12417tcp-request inspect-delay <timeout>
12418 Set the maximum allowed time to wait for data during content inspection
12419 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaufb356202010-08-03 14:02:05 +020012420 no | yes | yes | yes
Willy Tarreau62644772008-07-16 18:36:06 +020012421 Arguments :
12422 <timeout> is the timeout value specified in milliseconds by default, but
12423 can be in any other unit if the number is suffixed by the unit,
12424 as explained at the top of this document.
12425
Daniel Corbett9f0843f2021-05-08 10:50:37 -040012426 People using HAProxy primarily as a TCP relay are often worried about the
Willy Tarreau62644772008-07-16 18:36:06 +020012427 risk of passing any type of protocol to a server without any analysis. In
12428 order to be able to analyze the request contents, we must first withhold
12429 the data then analyze them. This statement simply enables withholding of
12430 data for at most the specified amount of time.
12431
Willy Tarreaufb356202010-08-03 14:02:05 +020012432 TCP content inspection applies very early when a connection reaches a
12433 frontend, then very early when the connection is forwarded to a backend. This
12434 means that a connection may experience a first delay in the frontend and a
12435 second delay in the backend if both have tcp-request rules.
12436
Daniel Corbett9f0843f2021-05-08 10:50:37 -040012437 Note that when performing content inspection, HAProxy will evaluate the whole
Willy Tarreau62644772008-07-16 18:36:06 +020012438 rules for every new chunk which gets in, taking into account the fact that
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +010012439 those data are partial. If no rule matches before the aforementioned delay,
Willy Tarreau62644772008-07-16 18:36:06 +020012440 a last check is performed upon expiration, this time considering that the
Daniel Corbett9f0843f2021-05-08 10:50:37 -040012441 contents are definitive. If no delay is set, HAProxy will not wait at all
Willy Tarreaud869b242009-03-15 14:43:58 +010012442 and will immediately apply a verdict based on the available information.
12443 Obviously this is unlikely to be very useful and might even be racy, so such
12444 setups are not recommended.
Willy Tarreau62644772008-07-16 18:36:06 +020012445
12446 As soon as a rule matches, the request is released and continues as usual. If
12447 the timeout is reached and no rule matches, the default policy will be to let
12448 it pass through unaffected.
12449
12450 For most protocols, it is enough to set it to a few seconds, as most clients
12451 send the full request immediately upon connection. Add 3 or more seconds to
12452 cover TCP retransmits but that's all. For some protocols, it may make sense
Willy Tarreaud72758d2010-01-12 10:42:19 +010012453 to use large values, for instance to ensure that the client never talks
Davor Ocelice9ed2812017-12-25 17:49:28 +010012454 before the server (e.g. SMTP), or to wait for a client to talk before passing
12455 data to the server (e.g. SSL). Note that the client timeout must cover at
Willy Tarreaub824b002010-09-29 16:36:16 +020012456 least the inspection delay, otherwise it will expire first. If the client
12457 closes the connection or if the buffer is full, the delay immediately expires
12458 since the contents will not be able to change anymore.
Willy Tarreau62644772008-07-16 18:36:06 +020012459
Willy Tarreau55165fe2009-05-10 12:02:55 +020012460 See also : "tcp-request content accept", "tcp-request content reject",
Willy Tarreau62644772008-07-16 18:36:06 +020012461 "timeout client".
12462
12463
Emeric Brun0a3b67f2010-09-24 15:34:53 +020012464tcp-response content <action> [{if | unless} <condition>]
12465 Perform an action on a session response depending on a layer 4-7 condition
12466 May be used in sections : defaults | frontend | listen | backend
12467 no | no | yes | yes
12468 Arguments :
Willy Tarreauc870bfd2015-09-28 18:47:38 +020012469 <action> defines the action to perform if the condition applies. See
12470 below.
Emeric Brun0a3b67f2010-09-24 15:34:53 +020012471
12472 <condition> is a standard layer 4-7 ACL-based condition (see section 7).
12473
Davor Ocelice9ed2812017-12-25 17:49:28 +010012474 Response contents can be analyzed at an early stage of response processing
Emeric Brun0a3b67f2010-09-24 15:34:53 +020012475 called "TCP content inspection". During this stage, ACL-based rules are
12476 evaluated every time the response contents are updated, until either an
Willy Tarreaucc1e04b2013-09-11 23:20:29 +020012477 "accept", "close" or a "reject" rule matches, or a TCP response inspection
12478 delay is set and expires with no matching rule.
Emeric Brun0a3b67f2010-09-24 15:34:53 +020012479
12480 Most often, these decisions will consider a protocol recognition or validity.
12481
12482 Content-based rules are evaluated in their exact declaration order. If no
12483 rule matches or if there is no rule, the default action is to accept the
12484 contents. There is no specific limit to the number of rules which may be
12485 inserted.
12486
Thierry FOURNIER236657b2015-08-19 08:25:14 +020012487 Several types of actions are supported :
Emeric Brun0a3b67f2010-09-24 15:34:53 +020012488 - accept :
12489 accepts the response if the condition is true (when used with "if")
12490 or false (when used with "unless"). The first such rule executed ends
12491 the rules evaluation.
12492
Willy Tarreaucc1e04b2013-09-11 23:20:29 +020012493 - close :
12494 immediately closes the connection with the server if the condition is
12495 true (when used with "if"), or false (when used with "unless"). The
12496 first such rule executed ends the rules evaluation. The main purpose of
12497 this action is to force a connection to be finished between a client
12498 and a server after an exchange when the application protocol expects
12499 some long time outs to elapse first. The goal is to eliminate idle
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030012500 connections which take significant resources on servers with certain
Willy Tarreaucc1e04b2013-09-11 23:20:29 +020012501 protocols.
12502
Emeric Brun0a3b67f2010-09-24 15:34:53 +020012503 - reject :
12504 rejects the response if the condition is true (when used with "if")
12505 or false (when used with "unless"). The first such rule executed ends
Jamie Gloudonaaa21002012-08-25 00:18:33 -040012506 the rules evaluation. Rejected session are immediately closed.
Emeric Brun0a3b67f2010-09-24 15:34:53 +020012507
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020012508 - set-var(<var-name>) <expr>
Willy Tarreau9de54ba2021-09-02 20:51:21 +020012509 Sets a variable from an expression.
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020012510
Christopher Faulet85d79c92016-11-09 16:54:56 +010012511 - unset-var(<var-name>)
12512 Unsets a variable.
12513
Thierry FOURNIERe0627bd2015-08-04 08:20:33 +020012514 - sc-inc-gpc0(<sc-id>):
12515 This action increments the GPC0 counter according to the sticky
12516 counter designated by <sc-id>. If an error occurs, this action fails
12517 silently and the actions evaluation continues.
12518
Frédéric Lécaille6778b272018-01-29 15:22:53 +010012519 - sc-inc-gpc1(<sc-id>):
12520 This action increments the GPC1 counter according to the sticky
12521 counter designated by <sc-id>. If an error occurs, this action fails
12522 silently and the actions evaluation continues.
12523
Cédric Dufour0d7712d2019-11-06 18:38:53 +010012524 - sc-set-gpt0(<sc-id>) { <int> | <expr> }
12525 This action sets the 32-bit unsigned GPT0 tag according to the sticky
12526 counter designated by <sc-id> and the value of <int>/<expr>. The
12527 expected result is a boolean. If an error occurs, this action silently
12528 fails and the actions evaluation continues.
Thierry FOURNIER236657b2015-08-19 08:25:14 +020012529
Willy Tarreau2d392c22015-08-24 01:43:45 +020012530 - "silent-drop" :
12531 This stops the evaluation of the rules and makes the client-facing
Davor Ocelice9ed2812017-12-25 17:49:28 +010012532 connection suddenly disappear using a system-dependent way that tries
Willy Tarreau2d392c22015-08-24 01:43:45 +020012533 to prevent the client from being notified. The effect it then that the
12534 client still sees an established connection while there's none on
12535 HAProxy. The purpose is to achieve a comparable effect to "tarpit"
12536 except that it doesn't use any local resource at all on the machine
12537 running HAProxy. It can resist much higher loads than "tarpit", and
Davor Ocelice9ed2812017-12-25 17:49:28 +010012538 slow down stronger attackers. It is important to understand the impact
12539 of using this mechanism. All stateful equipment placed between the
Willy Tarreau2d392c22015-08-24 01:43:45 +020012540 client and HAProxy (firewalls, proxies, load balancers) will also keep
12541 the established connection for a long time and may suffer from this
Davor Ocelice9ed2812017-12-25 17:49:28 +010012542 action. On modern Linux systems running with enough privileges, the
Willy Tarreau2d392c22015-08-24 01:43:45 +020012543 TCP_REPAIR socket option is used to block the emission of a TCP
12544 reset. On other systems, the socket's TTL is reduced to 1 so that the
12545 TCP reset doesn't pass the first router, though it's still delivered to
12546 local networks. Do not use it unless you fully understand how it works.
12547
Christopher Faulet76c09ef2017-09-21 11:03:52 +020012548 - send-spoe-group <engine-name> <group-name>
12549 Send a group of SPOE messages.
12550
Emeric Brun0a3b67f2010-09-24 15:34:53 +020012551 Note that the "if/unless" condition is optional. If no condition is set on
12552 the action, it is simply performed unconditionally. That can be useful for
12553 for changing the default action to a reject.
12554
Jamie Gloudonaaa21002012-08-25 00:18:33 -040012555 It is perfectly possible to match layer 7 contents with "tcp-response
12556 content" rules, but then it is important to ensure that a full response has
12557 been buffered, otherwise no contents will match. In order to achieve this,
12558 the best solution involves detecting the HTTP protocol during the inspection
Emeric Brun0a3b67f2010-09-24 15:34:53 +020012559 period.
12560
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020012561 The "set-var" is used to set the content of a variable. The variable is
12562 declared inline.
12563
Daniel Schneller0b547052016-03-21 20:46:57 +010012564 <var-name> The name of the variable starts with an indication about
12565 its scope. The scopes allowed are:
Christopher Fauletff2613e2016-11-09 11:36:17 +010012566 "proc" : the variable is shared with the whole process
Daniel Schneller0b547052016-03-21 20:46:57 +010012567 "sess" : the variable is shared with the whole session
12568 "txn" : the variable is shared with the transaction
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020012569 (request and response)
Daniel Schneller0b547052016-03-21 20:46:57 +010012570 "req" : the variable is shared only during request
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020012571 processing
Daniel Schneller0b547052016-03-21 20:46:57 +010012572 "res" : the variable is shared only during response
12573 processing
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020012574 This prefix is followed by a name. The separator is a '.'.
Christopher Fauletb71557a2016-10-31 10:49:03 +010012575 The name may only contain characters 'a-z', 'A-Z', '0-9',
12576 '.' and '_'.
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020012577
12578 <expr> Is a standard HAProxy expression formed by a sample-fetch
12579 followed by some converters.
12580
Christopher Faulet85d79c92016-11-09 16:54:56 +010012581 The "unset-var" is used to unset a variable. See above for details about
12582 <var-name>.
12583
Christopher Faulet76c09ef2017-09-21 11:03:52 +020012584 The "send-spoe-group" is used to trigger sending of a group of SPOE
12585 messages. To do so, the SPOE engine used to send messages must be defined, as
12586 well as the SPOE group to send. Of course, the SPOE engine must refer to an
12587 existing SPOE filter. If not engine name is provided on the SPOE filter line,
12588 the SPOE agent name must be used.
12589
12590 <engine-name> The SPOE engine name.
12591
12592 <group-name> The SPOE group name as specified in the engine configuration.
12593
Emeric Brun0a3b67f2010-09-24 15:34:53 +020012594 See section 7 about ACL usage.
12595
12596 See also : "tcp-request content", "tcp-response inspect-delay"
12597
12598
Willy Tarreau4f614292016-10-21 17:49:36 +020012599tcp-request session <action> [{if | unless} <condition>]
12600 Perform an action on a validated session depending on a layer 5 condition
12601 May be used in sections : defaults | frontend | listen | backend
12602 no | yes | yes | no
12603 Arguments :
12604 <action> defines the action to perform if the condition applies. See
12605 below.
12606
12607 <condition> is a standard layer5-only ACL-based condition (see section 7).
12608
Davor Ocelice9ed2812017-12-25 17:49:28 +010012609 Once a session is validated, (i.e. after all handshakes have been completed),
Willy Tarreau4f614292016-10-21 17:49:36 +020012610 it is possible to evaluate some conditions to decide whether this session
12611 must be accepted or dropped or have its counters tracked. Those conditions
12612 cannot make use of any data contents because no buffers are allocated yet and
12613 the processing cannot wait at this stage. The main use case it to copy some
12614 early information into variables (since variables are accessible in the
12615 session), or to keep track of some information collected after the handshake,
12616 such as SSL-level elements (SNI, ciphers, client cert's CN) or information
Davor Ocelice9ed2812017-12-25 17:49:28 +010012617 from the PROXY protocol header (e.g. track a source forwarded this way). The
Willy Tarreau4f614292016-10-21 17:49:36 +020012618 extracted information can thus be copied to a variable or tracked using
12619 "track-sc" rules. Of course it is also possible to decide to accept/reject as
12620 with other rulesets. Most operations performed here could also be performed
12621 in "tcp-request content" rules, except that in HTTP these rules are evaluated
12622 for each new request, and that might not always be acceptable. For example a
12623 rule might increment a counter on each evaluation. It would also be possible
12624 that a country is resolved by geolocation from the source IP address,
12625 assigned to a session-wide variable, then the source address rewritten from
12626 an HTTP header for all requests. If some contents need to be inspected in
12627 order to take the decision, the "tcp-request content" statements must be used
12628 instead.
12629
12630 The "tcp-request session" rules are evaluated in their exact declaration
12631 order. If no rule matches or if there is no rule, the default action is to
12632 accept the incoming session. There is no specific limit to the number of
12633 rules which may be inserted.
12634
12635 Several types of actions are supported :
12636 - accept : the request is accepted
12637 - reject : the request is rejected and the connection is closed
12638 - { track-sc0 | track-sc1 | track-sc2 } <key> [table <table>]
12639 - sc-inc-gpc0(<sc-id>)
Frédéric Lécaille6778b272018-01-29 15:22:53 +010012640 - sc-inc-gpc1(<sc-id>)
Cédric Dufour0d7712d2019-11-06 18:38:53 +010012641 - sc-set-gpt0(<sc-id>) { <int> | <expr> }
Christopher Faulet57759f32021-06-23 12:19:25 +020012642 - set-dst <expr>
12643 - set-dst-port <expr>
12644 - set-src <expr>
12645 - set-src-port <expr>
Willy Tarreau4f614292016-10-21 17:49:36 +020012646 - set-var(<var-name>) <expr>
Christopher Faulet85d79c92016-11-09 16:54:56 +010012647 - unset-var(<var-name>)
Willy Tarreau4f614292016-10-21 17:49:36 +020012648 - silent-drop
12649
12650 These actions have the same meaning as their respective counter-parts in
12651 "tcp-request connection" and "tcp-request content", so please refer to these
12652 sections for a complete description.
12653
12654 Note that the "if/unless" condition is optional. If no condition is set on
12655 the action, it is simply performed unconditionally. That can be useful for
12656 "track-sc*" actions as well as for changing the default action to a reject.
12657
12658 Example: track the original source address by default, or the one advertised
12659 in the PROXY protocol header for connection coming from the local
12660 proxies. The first connection-level rule enables receipt of the
12661 PROXY protocol for these ones, the second rule tracks whatever
12662 address we decide to keep after optional decoding.
12663
12664 tcp-request connection expect-proxy layer4 if { src -f proxies.lst }
12665 tcp-request session track-sc0 src
12666
12667 Example: accept all sessions from white-listed hosts, reject too fast
12668 sessions without counting them, and track accepted sessions.
12669 This results in session rate being capped from abusive sources.
12670
12671 tcp-request session accept if { src -f /etc/haproxy/whitelist.lst }
12672 tcp-request session reject if { src_sess_rate gt 10 }
12673 tcp-request session track-sc0 src
12674
12675 Example: accept all sessions from white-listed hosts, count all other
12676 sessions and reject too fast ones. This results in abusive ones
12677 being blocked as long as they don't slow down.
12678
12679 tcp-request session accept if { src -f /etc/haproxy/whitelist.lst }
12680 tcp-request session track-sc0 src
12681 tcp-request session reject if { sc0_sess_rate gt 10 }
12682
12683 See section 7 about ACL usage.
12684
12685 See also : "tcp-request connection", "tcp-request content", "stick-table"
12686
12687
Emeric Brun0a3b67f2010-09-24 15:34:53 +020012688tcp-response inspect-delay <timeout>
12689 Set the maximum allowed time to wait for a response during content inspection
12690 May be used in sections : defaults | frontend | listen | backend
12691 no | no | yes | yes
12692 Arguments :
12693 <timeout> is the timeout value specified in milliseconds by default, but
12694 can be in any other unit if the number is suffixed by the unit,
12695 as explained at the top of this document.
12696
12697 See also : "tcp-response content", "tcp-request inspect-delay".
12698
12699
Krzysztof Piotr Oledzki5259dfe2008-01-21 01:54:06 +010012700timeout check <timeout>
12701 Set additional check timeout, but only after a connection has been already
12702 established.
12703
12704 May be used in sections: defaults | frontend | listen | backend
12705 yes | no | yes | yes
12706 Arguments:
12707 <timeout> is the timeout value specified in milliseconds by default, but
12708 can be in any other unit if the number is suffixed by the unit,
12709 as explained at the top of this document.
12710
Daniel Corbett9f0843f2021-05-08 10:50:37 -040012711 If set, HAProxy uses min("timeout connect", "inter") as a connect timeout
Krzysztof Piotr Oledzki5259dfe2008-01-21 01:54:06 +010012712 for check and "timeout check" as an additional read timeout. The "min" is
Davor Ocelice9ed2812017-12-25 17:49:28 +010012713 used so that people running with *very* long "timeout connect" (e.g. those
Krzysztof Piotr Oledzki5259dfe2008-01-21 01:54:06 +010012714 who needed this due to the queue or tarpit) do not slow down their checks.
Willy Tarreaud7550a22010-02-10 05:10:19 +010012715 (Please also note that there is no valid reason to have such long connect
12716 timeouts, because "timeout queue" and "timeout tarpit" can always be used to
12717 avoid that).
Krzysztof Piotr Oledzki5259dfe2008-01-21 01:54:06 +010012718
Daniel Corbett9f0843f2021-05-08 10:50:37 -040012719 If "timeout check" is not set HAProxy uses "inter" for complete check
Krzysztof Piotr Oledzki5259dfe2008-01-21 01:54:06 +010012720 timeout (connect + read) exactly like all <1.3.15 version.
12721
12722 In most cases check request is much simpler and faster to handle than normal
12723 requests and people may want to kick out laggy servers so this timeout should
Willy Tarreau41a340d2008-01-22 12:25:31 +010012724 be smaller than "timeout server".
Krzysztof Piotr Oledzki5259dfe2008-01-21 01:54:06 +010012725
12726 This parameter is specific to backends, but can be specified once for all in
12727 "defaults" sections. This is in fact one of the easiest solutions not to
12728 forget about it.
12729
Willy Tarreau41a340d2008-01-22 12:25:31 +010012730 See also: "timeout connect", "timeout queue", "timeout server",
12731 "timeout tarpit".
Krzysztof Piotr Oledzki5259dfe2008-01-21 01:54:06 +010012732
12733
Willy Tarreau0ba27502007-12-24 16:55:16 +010012734timeout client <timeout>
Willy Tarreau0ba27502007-12-24 16:55:16 +010012735 Set the maximum inactivity time on the client side.
12736 May be used in sections : defaults | frontend | listen | backend
12737 yes | yes | yes | no
12738 Arguments :
Willy Tarreau844e3c52008-01-11 16:28:18 +010012739 <timeout> is the timeout value specified in milliseconds by default, but
Willy Tarreau0ba27502007-12-24 16:55:16 +010012740 can be in any other unit if the number is suffixed by the unit,
12741 as explained at the top of this document.
12742
12743 The inactivity timeout applies when the client is expected to acknowledge or
12744 send data. In HTTP mode, this timeout is particularly important to consider
12745 during the first phase, when the client sends the request, and during the
Baptiste Assmann2e1941e2016-03-06 23:24:12 +010012746 response while it is reading data sent by the server. That said, for the
12747 first phase, it is preferable to set the "timeout http-request" to better
12748 protect HAProxy from Slowloris like attacks. The value is specified in
12749 milliseconds by default, but can be in any other unit if the number is
Willy Tarreau0ba27502007-12-24 16:55:16 +010012750 suffixed by the unit, as specified at the top of this document. In TCP mode
12751 (and to a lesser extent, in HTTP mode), it is highly recommended that the
12752 client timeout remains equal to the server timeout in order to avoid complex
Willy Tarreaud2a4aa22008-01-31 15:28:22 +010012753 situations to debug. It is a good practice to cover one or several TCP packet
Willy Tarreau0ba27502007-12-24 16:55:16 +010012754 losses by specifying timeouts that are slightly above multiples of 3 seconds
Davor Ocelice9ed2812017-12-25 17:49:28 +010012755 (e.g. 4 or 5 seconds). If some long-lived sessions are mixed with short-lived
12756 sessions (e.g. WebSocket and HTTP), it's worth considering "timeout tunnel",
Willy Tarreau05cdd962014-05-10 14:30:07 +020012757 which overrides "timeout client" and "timeout server" for tunnels, as well as
12758 "timeout client-fin" for half-closed connections.
Willy Tarreau0ba27502007-12-24 16:55:16 +010012759
12760 This parameter is specific to frontends, but can be specified once for all in
12761 "defaults" sections. This is in fact one of the easiest solutions not to
12762 forget about it. An unspecified timeout results in an infinite timeout, which
12763 is not recommended. Such a usage is accepted and works but reports a warning
John Roeslerfb2fce12019-07-10 15:45:51 -050012764 during startup because it may result in accumulation of expired sessions in
Willy Tarreau0ba27502007-12-24 16:55:16 +010012765 the system if the system's timeouts are not configured either.
12766
Willy Tarreau95c4e142017-11-26 12:18:55 +010012767 This also applies to HTTP/2 connections, which will be closed with GOAWAY.
Lukas Tribus75df9d72017-11-24 19:05:12 +010012768
Tim Duesterhus86e6b6e2019-05-14 20:57:59 +020012769 See also : "timeout server", "timeout tunnel", "timeout http-request".
Willy Tarreau0ba27502007-12-24 16:55:16 +010012770
Willy Tarreau0ba27502007-12-24 16:55:16 +010012771
Willy Tarreau05cdd962014-05-10 14:30:07 +020012772timeout client-fin <timeout>
12773 Set the inactivity timeout on the client side for half-closed connections.
12774 May be used in sections : defaults | frontend | listen | backend
12775 yes | yes | yes | no
12776 Arguments :
12777 <timeout> is the timeout value specified in milliseconds by default, but
12778 can be in any other unit if the number is suffixed by the unit,
12779 as explained at the top of this document.
12780
12781 The inactivity timeout applies when the client is expected to acknowledge or
12782 send data while one direction is already shut down. This timeout is different
12783 from "timeout client" in that it only applies to connections which are closed
12784 in one direction. This is particularly useful to avoid keeping connections in
12785 FIN_WAIT state for too long when clients do not disconnect cleanly. This
12786 problem is particularly common long connections such as RDP or WebSocket.
12787 Note that this timeout can override "timeout tunnel" when a connection shuts
Willy Tarreau599391a2017-11-24 10:16:00 +010012788 down in one direction. It is applied to idle HTTP/2 connections once a GOAWAY
12789 frame was sent, often indicating an expectation that the connection quickly
12790 ends.
Willy Tarreau05cdd962014-05-10 14:30:07 +020012791
12792 This parameter is specific to frontends, but can be specified once for all in
12793 "defaults" sections. By default it is not set, so half-closed connections
12794 will use the other timeouts (timeout.client or timeout.tunnel).
12795
12796 See also : "timeout client", "timeout server-fin", and "timeout tunnel".
12797
12798
Willy Tarreau0ba27502007-12-24 16:55:16 +010012799timeout connect <timeout>
Willy Tarreau0ba27502007-12-24 16:55:16 +010012800 Set the maximum time to wait for a connection attempt to a server to succeed.
12801 May be used in sections : defaults | frontend | listen | backend
12802 yes | no | yes | yes
12803 Arguments :
Willy Tarreau844e3c52008-01-11 16:28:18 +010012804 <timeout> is the timeout value specified in milliseconds by default, but
Willy Tarreau0ba27502007-12-24 16:55:16 +010012805 can be in any other unit if the number is suffixed by the unit,
12806 as explained at the top of this document.
12807
Daniel Corbett9f0843f2021-05-08 10:50:37 -040012808 If the server is located on the same LAN as HAProxy, the connection should be
Willy Tarreaud2a4aa22008-01-31 15:28:22 +010012809 immediate (less than a few milliseconds). Anyway, it is a good practice to
Willy Tarreaud72758d2010-01-12 10:42:19 +010012810 cover one or several TCP packet losses by specifying timeouts that are
Davor Ocelice9ed2812017-12-25 17:49:28 +010012811 slightly above multiples of 3 seconds (e.g. 4 or 5 seconds). By default, the
Krzysztof Piotr Oledzki5259dfe2008-01-21 01:54:06 +010012812 connect timeout also presets both queue and tarpit timeouts to the same value
12813 if these have not been specified.
Willy Tarreau0ba27502007-12-24 16:55:16 +010012814
12815 This parameter is specific to backends, but can be specified once for all in
12816 "defaults" sections. This is in fact one of the easiest solutions not to
12817 forget about it. An unspecified timeout results in an infinite timeout, which
12818 is not recommended. Such a usage is accepted and works but reports a warning
John Roeslerfb2fce12019-07-10 15:45:51 -050012819 during startup because it may result in accumulation of failed sessions in
Willy Tarreau0ba27502007-12-24 16:55:16 +010012820 the system if the system's timeouts are not configured either.
12821
Tim Duesterhus86e6b6e2019-05-14 20:57:59 +020012822 See also: "timeout check", "timeout queue", "timeout server", "timeout tarpit".
Willy Tarreau0ba27502007-12-24 16:55:16 +010012823
Willy Tarreau0ba27502007-12-24 16:55:16 +010012824
Willy Tarreaub16a5742010-01-10 14:46:16 +010012825timeout http-keep-alive <timeout>
12826 Set the maximum allowed time to wait for a new HTTP request to appear
12827 May be used in sections : defaults | frontend | listen | backend
12828 yes | yes | yes | yes
12829 Arguments :
12830 <timeout> is the timeout value specified in milliseconds by default, but
12831 can be in any other unit if the number is suffixed by the unit,
12832 as explained at the top of this document.
12833
12834 By default, the time to wait for a new request in case of keep-alive is set
12835 by "timeout http-request". However this is not always convenient because some
12836 people want very short keep-alive timeouts in order to release connections
12837 faster, and others prefer to have larger ones but still have short timeouts
12838 once the request has started to present itself.
12839
12840 The "http-keep-alive" timeout covers these needs. It will define how long to
12841 wait for a new HTTP request to start coming after a response was sent. Once
12842 the first byte of request has been seen, the "http-request" timeout is used
12843 to wait for the complete request to come. Note that empty lines prior to a
12844 new request do not refresh the timeout and are not counted as a new request.
12845
12846 There is also another difference between the two timeouts : when a connection
12847 expires during timeout http-keep-alive, no error is returned, the connection
12848 just closes. If the connection expires in "http-request" while waiting for a
12849 connection to complete, a HTTP 408 error is returned.
12850
12851 In general it is optimal to set this value to a few tens to hundreds of
12852 milliseconds, to allow users to fetch all objects of a page at once but
Davor Ocelice9ed2812017-12-25 17:49:28 +010012853 without waiting for further clicks. Also, if set to a very small value (e.g.
Willy Tarreaub16a5742010-01-10 14:46:16 +010012854 1 millisecond) it will probably only accept pipelined requests but not the
12855 non-pipelined ones. It may be a nice trade-off for very large sites running
Patrick Mézard2382ad62010-05-09 10:43:32 +020012856 with tens to hundreds of thousands of clients.
Willy Tarreaub16a5742010-01-10 14:46:16 +010012857
12858 If this parameter is not set, the "http-request" timeout applies, and if both
12859 are not set, "timeout client" still applies at the lower level. It should be
12860 set in the frontend to take effect, unless the frontend is in TCP mode, in
12861 which case the HTTP backend's timeout will be used.
12862
Willy Tarreau95c4e142017-11-26 12:18:55 +010012863 When using HTTP/2 "timeout client" is applied instead. This is so we can keep
12864 using short keep-alive timeouts in HTTP/1.1 while using longer ones in HTTP/2
Lukas Tribus75df9d72017-11-24 19:05:12 +010012865 (where we only have one connection per client and a connection setup).
12866
Willy Tarreaub16a5742010-01-10 14:46:16 +010012867 See also : "timeout http-request", "timeout client".
12868
12869
Willy Tarreau036fae02008-01-06 13:24:40 +010012870timeout http-request <timeout>
12871 Set the maximum allowed time to wait for a complete HTTP request
12872 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaucd7afc02009-07-12 10:03:17 +020012873 yes | yes | yes | yes
Willy Tarreau036fae02008-01-06 13:24:40 +010012874 Arguments :
Willy Tarreau844e3c52008-01-11 16:28:18 +010012875 <timeout> is the timeout value specified in milliseconds by default, but
Willy Tarreau036fae02008-01-06 13:24:40 +010012876 can be in any other unit if the number is suffixed by the unit,
12877 as explained at the top of this document.
12878
12879 In order to offer DoS protection, it may be required to lower the maximum
12880 accepted time to receive a complete HTTP request without affecting the client
12881 timeout. This helps protecting against established connections on which
12882 nothing is sent. The client timeout cannot offer a good protection against
12883 this abuse because it is an inactivity timeout, which means that if the
12884 attacker sends one character every now and then, the timeout will not
12885 trigger. With the HTTP request timeout, no matter what speed the client
Willy Tarreau2705a612014-05-23 17:38:34 +020012886 types, the request will be aborted if it does not complete in time. When the
12887 timeout expires, an HTTP 408 response is sent to the client to inform it
12888 about the problem, and the connection is closed. The logs will report
12889 termination codes "cR". Some recent browsers are having problems with this
Davor Ocelice9ed2812017-12-25 17:49:28 +010012890 standard, well-documented behavior, so it might be needed to hide the 408
Willy Tarreau0f228a02015-05-01 15:37:53 +020012891 code using "option http-ignore-probes" or "errorfile 408 /dev/null". See
12892 more details in the explanations of the "cR" termination code in section 8.5.
Willy Tarreau036fae02008-01-06 13:24:40 +010012893
Baptiste Assmanneccdf432015-10-28 13:49:01 +010012894 By default, this timeout only applies to the header part of the request,
12895 and not to any data. As soon as the empty line is received, this timeout is
12896 not used anymore. When combined with "option http-buffer-request", this
12897 timeout also applies to the body of the request..
12898 It is used again on keep-alive connections to wait for a second
Willy Tarreaub16a5742010-01-10 14:46:16 +010012899 request if "timeout http-keep-alive" is not set.
Willy Tarreau036fae02008-01-06 13:24:40 +010012900
12901 Generally it is enough to set it to a few seconds, as most clients send the
12902 full request immediately upon connection. Add 3 or more seconds to cover TCP
Davor Ocelice9ed2812017-12-25 17:49:28 +010012903 retransmits but that's all. Setting it to very low values (e.g. 50 ms) will
Willy Tarreau036fae02008-01-06 13:24:40 +010012904 generally work on local networks as long as there are no packet losses. This
12905 will prevent people from sending bare HTTP requests using telnet.
12906
12907 If this parameter is not set, the client timeout still applies between each
Willy Tarreaucd7afc02009-07-12 10:03:17 +020012908 chunk of the incoming request. It should be set in the frontend to take
12909 effect, unless the frontend is in TCP mode, in which case the HTTP backend's
12910 timeout will be used.
Willy Tarreau036fae02008-01-06 13:24:40 +010012911
Willy Tarreau0f228a02015-05-01 15:37:53 +020012912 See also : "errorfile", "http-ignore-probes", "timeout http-keep-alive", and
Baptiste Assmanneccdf432015-10-28 13:49:01 +010012913 "timeout client", "option http-buffer-request".
Willy Tarreau036fae02008-01-06 13:24:40 +010012914
Willy Tarreau844e3c52008-01-11 16:28:18 +010012915
12916timeout queue <timeout>
12917 Set the maximum time to wait in the queue for a connection slot to be free
12918 May be used in sections : defaults | frontend | listen | backend
12919 yes | no | yes | yes
12920 Arguments :
12921 <timeout> is the timeout value specified in milliseconds by default, but
12922 can be in any other unit if the number is suffixed by the unit,
12923 as explained at the top of this document.
12924
12925 When a server's maxconn is reached, connections are left pending in a queue
12926 which may be server-specific or global to the backend. In order not to wait
12927 indefinitely, a timeout is applied to requests pending in the queue. If the
12928 timeout is reached, it is considered that the request will almost never be
12929 served, so it is dropped and a 503 error is returned to the client.
12930
12931 The "timeout queue" statement allows to fix the maximum time for a request to
12932 be left pending in a queue. If unspecified, the same value as the backend's
12933 connection timeout ("timeout connect") is used, for backwards compatibility
12934 with older versions with no "timeout queue" parameter.
12935
Tim Duesterhus86e6b6e2019-05-14 20:57:59 +020012936 See also : "timeout connect".
Willy Tarreau844e3c52008-01-11 16:28:18 +010012937
12938
12939timeout server <timeout>
Willy Tarreau844e3c52008-01-11 16:28:18 +010012940 Set the maximum inactivity time on the server side.
12941 May be used in sections : defaults | frontend | listen | backend
12942 yes | no | yes | yes
12943 Arguments :
12944 <timeout> is the timeout value specified in milliseconds by default, but
12945 can be in any other unit if the number is suffixed by the unit,
12946 as explained at the top of this document.
12947
12948 The inactivity timeout applies when the server is expected to acknowledge or
12949 send data. In HTTP mode, this timeout is particularly important to consider
12950 during the first phase of the server's response, when it has to send the
12951 headers, as it directly represents the server's processing time for the
12952 request. To find out what value to put there, it's often good to start with
12953 what would be considered as unacceptable response times, then check the logs
12954 to observe the response time distribution, and adjust the value accordingly.
12955
12956 The value is specified in milliseconds by default, but can be in any other
12957 unit if the number is suffixed by the unit, as specified at the top of this
12958 document. In TCP mode (and to a lesser extent, in HTTP mode), it is highly
12959 recommended that the client timeout remains equal to the server timeout in
12960 order to avoid complex situations to debug. Whatever the expected server
Willy Tarreaud2a4aa22008-01-31 15:28:22 +010012961 response times, it is a good practice to cover at least one or several TCP
Willy Tarreau844e3c52008-01-11 16:28:18 +010012962 packet losses by specifying timeouts that are slightly above multiples of 3
Davor Ocelice9ed2812017-12-25 17:49:28 +010012963 seconds (e.g. 4 or 5 seconds minimum). If some long-lived sessions are mixed
12964 with short-lived sessions (e.g. WebSocket and HTTP), it's worth considering
Willy Tarreauce887fd2012-05-12 12:50:00 +020012965 "timeout tunnel", which overrides "timeout client" and "timeout server" for
12966 tunnels.
Willy Tarreau844e3c52008-01-11 16:28:18 +010012967
12968 This parameter is specific to backends, but can be specified once for all in
12969 "defaults" sections. This is in fact one of the easiest solutions not to
12970 forget about it. An unspecified timeout results in an infinite timeout, which
12971 is not recommended. Such a usage is accepted and works but reports a warning
John Roeslerfb2fce12019-07-10 15:45:51 -050012972 during startup because it may result in accumulation of expired sessions in
Willy Tarreau844e3c52008-01-11 16:28:18 +010012973 the system if the system's timeouts are not configured either.
12974
Tim Duesterhus86e6b6e2019-05-14 20:57:59 +020012975 See also : "timeout client" and "timeout tunnel".
Willy Tarreau844e3c52008-01-11 16:28:18 +010012976
Willy Tarreau05cdd962014-05-10 14:30:07 +020012977
12978timeout server-fin <timeout>
12979 Set the inactivity timeout on the server side for half-closed connections.
12980 May be used in sections : defaults | frontend | listen | backend
12981 yes | no | yes | yes
12982 Arguments :
12983 <timeout> is the timeout value specified in milliseconds by default, but
12984 can be in any other unit if the number is suffixed by the unit,
12985 as explained at the top of this document.
12986
12987 The inactivity timeout applies when the server is expected to acknowledge or
12988 send data while one direction is already shut down. This timeout is different
12989 from "timeout server" in that it only applies to connections which are closed
12990 in one direction. This is particularly useful to avoid keeping connections in
12991 FIN_WAIT state for too long when a remote server does not disconnect cleanly.
12992 This problem is particularly common long connections such as RDP or WebSocket.
12993 Note that this timeout can override "timeout tunnel" when a connection shuts
12994 down in one direction. This setting was provided for completeness, but in most
12995 situations, it should not be needed.
12996
12997 This parameter is specific to backends, but can be specified once for all in
12998 "defaults" sections. By default it is not set, so half-closed connections
12999 will use the other timeouts (timeout.server or timeout.tunnel).
13000
13001 See also : "timeout client-fin", "timeout server", and "timeout tunnel".
13002
Willy Tarreau844e3c52008-01-11 16:28:18 +010013003
13004timeout tarpit <timeout>
Cyril Bonté78caf842010-03-10 22:41:43 +010013005 Set the duration for which tarpitted connections will be maintained
Willy Tarreau844e3c52008-01-11 16:28:18 +010013006 May be used in sections : defaults | frontend | listen | backend
13007 yes | yes | yes | yes
13008 Arguments :
13009 <timeout> is the tarpit duration specified in milliseconds by default, but
13010 can be in any other unit if the number is suffixed by the unit,
13011 as explained at the top of this document.
13012
Christopher Faulet87f1f3d2019-07-18 14:51:20 +020013013 When a connection is tarpitted using "http-request tarpit", it is maintained
13014 open with no activity for a certain amount of time, then closed. "timeout
13015 tarpit" defines how long it will be maintained open.
Willy Tarreau844e3c52008-01-11 16:28:18 +010013016
13017 The value is specified in milliseconds by default, but can be in any other
13018 unit if the number is suffixed by the unit, as specified at the top of this
13019 document. If unspecified, the same value as the backend's connection timeout
13020 ("timeout connect") is used, for backwards compatibility with older versions
Cyril Bonté78caf842010-03-10 22:41:43 +010013021 with no "timeout tarpit" parameter.
Willy Tarreau844e3c52008-01-11 16:28:18 +010013022
Tim Duesterhus86e6b6e2019-05-14 20:57:59 +020013023 See also : "timeout connect".
Willy Tarreau844e3c52008-01-11 16:28:18 +010013024
13025
Willy Tarreauce887fd2012-05-12 12:50:00 +020013026timeout tunnel <timeout>
13027 Set the maximum inactivity time on the client and server side for tunnels.
13028 May be used in sections : defaults | frontend | listen | backend
13029 yes | no | yes | yes
13030 Arguments :
13031 <timeout> is the timeout value specified in milliseconds by default, but
13032 can be in any other unit if the number is suffixed by the unit,
13033 as explained at the top of this document.
13034
Jamie Gloudonaaa21002012-08-25 00:18:33 -040013035 The tunnel timeout applies when a bidirectional connection is established
Willy Tarreauce887fd2012-05-12 12:50:00 +020013036 between a client and a server, and the connection remains inactive in both
13037 directions. This timeout supersedes both the client and server timeouts once
13038 the connection becomes a tunnel. In TCP, this timeout is used as soon as no
Davor Ocelice9ed2812017-12-25 17:49:28 +010013039 analyzer remains attached to either connection (e.g. tcp content rules are
13040 accepted). In HTTP, this timeout is used when a connection is upgraded (e.g.
Willy Tarreauce887fd2012-05-12 12:50:00 +020013041 when switching to the WebSocket protocol, or forwarding a CONNECT request
13042 to a proxy), or after the first response when no keepalive/close option is
13043 specified.
13044
Willy Tarreau05cdd962014-05-10 14:30:07 +020013045 Since this timeout is usually used in conjunction with long-lived connections,
13046 it usually is a good idea to also set "timeout client-fin" to handle the
13047 situation where a client suddenly disappears from the net and does not
13048 acknowledge a close, or sends a shutdown and does not acknowledge pending
13049 data anymore. This can happen in lossy networks where firewalls are present,
13050 and is detected by the presence of large amounts of sessions in a FIN_WAIT
13051 state.
13052
Willy Tarreauce887fd2012-05-12 12:50:00 +020013053 The value is specified in milliseconds by default, but can be in any other
13054 unit if the number is suffixed by the unit, as specified at the top of this
13055 document. Whatever the expected normal idle time, it is a good practice to
13056 cover at least one or several TCP packet losses by specifying timeouts that
Davor Ocelice9ed2812017-12-25 17:49:28 +010013057 are slightly above multiples of 3 seconds (e.g. 4 or 5 seconds minimum).
Willy Tarreauce887fd2012-05-12 12:50:00 +020013058
13059 This parameter is specific to backends, but can be specified once for all in
13060 "defaults" sections. This is in fact one of the easiest solutions not to
13061 forget about it.
13062
13063 Example :
13064 defaults http
13065 option http-server-close
13066 timeout connect 5s
13067 timeout client 30s
Willy Tarreau05cdd962014-05-10 14:30:07 +020013068 timeout client-fin 30s
Willy Tarreauce887fd2012-05-12 12:50:00 +020013069 timeout server 30s
13070 timeout tunnel 1h # timeout to use with WebSocket and CONNECT
13071
Willy Tarreau05cdd962014-05-10 14:30:07 +020013072 See also : "timeout client", "timeout client-fin", "timeout server".
Willy Tarreauce887fd2012-05-12 12:50:00 +020013073
13074
Willy Tarreau844e3c52008-01-11 16:28:18 +010013075transparent (deprecated)
13076 Enable client-side transparent proxying
13077 May be used in sections : defaults | frontend | listen | backend
Willy Tarreau4b1f8592008-12-23 23:13:55 +010013078 yes | no | yes | yes
Willy Tarreau844e3c52008-01-11 16:28:18 +010013079 Arguments : none
13080
13081 This keyword was introduced in order to provide layer 7 persistence to layer
13082 3 load balancers. The idea is to use the OS's ability to redirect an incoming
13083 connection for a remote address to a local process (here HAProxy), and let
13084 this process know what address was initially requested. When this option is
13085 used, sessions without cookies will be forwarded to the original destination
13086 IP address of the incoming request (which should match that of another
13087 equipment), while requests with cookies will still be forwarded to the
13088 appropriate server.
13089
13090 The "transparent" keyword is deprecated, use "option transparent" instead.
13091
13092 Note that contrary to a common belief, this option does NOT make HAProxy
13093 present the client's IP to the server when establishing the connection.
13094
Willy Tarreau844e3c52008-01-11 16:28:18 +010013095 See also: "option transparent"
13096
William Lallemanda73203e2012-03-12 12:48:57 +010013097unique-id-format <string>
13098 Generate a unique ID for each request.
13099 May be used in sections : defaults | frontend | listen | backend
13100 yes | yes | yes | no
13101 Arguments :
13102 <string> is a log-format string.
13103
Cyril Bonté108cf6e2012-04-21 23:30:29 +020013104 This keyword creates a ID for each request using the custom log format. A
13105 unique ID is useful to trace a request passing through many components of
13106 a complex infrastructure. The newly created ID may also be logged using the
13107 %ID tag the log-format string.
William Lallemanda73203e2012-03-12 12:48:57 +010013108
Cyril Bonté108cf6e2012-04-21 23:30:29 +020013109 The format should be composed from elements that are guaranteed to be
Daniel Corbett9f0843f2021-05-08 10:50:37 -040013110 unique when combined together. For instance, if multiple HAProxy instances
Cyril Bonté108cf6e2012-04-21 23:30:29 +020013111 are involved, it might be important to include the node name. It is often
13112 needed to log the incoming connection's source and destination addresses
13113 and ports. Note that since multiple requests may be performed over the same
13114 connection, including a request counter may help differentiate them.
13115 Similarly, a timestamp may protect against a rollover of the counter.
13116 Logging the process ID will avoid collisions after a service restart.
William Lallemanda73203e2012-03-12 12:48:57 +010013117
Cyril Bonté108cf6e2012-04-21 23:30:29 +020013118 It is recommended to use hexadecimal notation for many fields since it
13119 makes them more compact and saves space in logs.
William Lallemanda73203e2012-03-12 12:48:57 +010013120
Cyril Bonté108cf6e2012-04-21 23:30:29 +020013121 Example:
William Lallemanda73203e2012-03-12 12:48:57 +010013122
Julien Vehentf21be322014-03-07 08:27:34 -050013123 unique-id-format %{+X}o\ %ci:%cp_%fi:%fp_%Ts_%rt:%pid
William Lallemanda73203e2012-03-12 12:48:57 +010013124
13125 will generate:
13126
13127 7F000001:8296_7F00001E:1F90_4F7B0A69_0003:790A
13128
13129 See also: "unique-id-header"
13130
13131unique-id-header <name>
13132 Add a unique ID header in the HTTP request.
13133 May be used in sections : defaults | frontend | listen | backend
13134 yes | yes | yes | no
13135 Arguments :
13136 <name> is the name of the header.
13137
Cyril Bonté108cf6e2012-04-21 23:30:29 +020013138 Add a unique-id header in the HTTP request sent to the server, using the
13139 unique-id-format. It can't work if the unique-id-format doesn't exist.
William Lallemanda73203e2012-03-12 12:48:57 +010013140
Cyril Bonté108cf6e2012-04-21 23:30:29 +020013141 Example:
William Lallemanda73203e2012-03-12 12:48:57 +010013142
Julien Vehentf21be322014-03-07 08:27:34 -050013143 unique-id-format %{+X}o\ %ci:%cp_%fi:%fp_%Ts_%rt:%pid
William Lallemanda73203e2012-03-12 12:48:57 +010013144 unique-id-header X-Unique-ID
13145
13146 will generate:
13147
13148 X-Unique-ID: 7F000001:8296_7F00001E:1F90_4F7B0A69_0003:790A
13149
13150 See also: "unique-id-format"
Willy Tarreau844e3c52008-01-11 16:28:18 +010013151
Willy Tarreauf51658d2014-04-23 01:21:56 +020013152use_backend <backend> [{if | unless} <condition>]
Willy Tarreau1d0dfb12009-07-07 15:10:31 +020013153 Switch to a specific backend if/unless an ACL-based condition is matched.
Willy Tarreau844e3c52008-01-11 16:28:18 +010013154 May be used in sections : defaults | frontend | listen | backend
13155 no | yes | yes | no
13156 Arguments :
Bertrand Jacquin702d44f2013-11-19 11:43:06 +010013157 <backend> is the name of a valid backend or "listen" section, or a
13158 "log-format" string resolving to a backend name.
Willy Tarreau844e3c52008-01-11 16:28:18 +010013159
Willy Tarreauf51658d2014-04-23 01:21:56 +020013160 <condition> is a condition composed of ACLs, as described in section 7. If
13161 it is omitted, the rule is unconditionally applied.
Willy Tarreau844e3c52008-01-11 16:28:18 +010013162
13163 When doing content-switching, connections arrive on a frontend and are then
13164 dispatched to various backends depending on a number of conditions. The
13165 relation between the conditions and the backends is described with the
Willy Tarreau1d0dfb12009-07-07 15:10:31 +020013166 "use_backend" keyword. While it is normally used with HTTP processing, it can
Davor Ocelice9ed2812017-12-25 17:49:28 +010013167 also be used in pure TCP, either without content using stateless ACLs (e.g.
Willy Tarreau1d0dfb12009-07-07 15:10:31 +020013168 source address validation) or combined with a "tcp-request" rule to wait for
13169 some payload.
Willy Tarreau844e3c52008-01-11 16:28:18 +010013170
13171 There may be as many "use_backend" rules as desired. All of these rules are
13172 evaluated in their declaration order, and the first one which matches will
13173 assign the backend.
13174
13175 In the first form, the backend will be used if the condition is met. In the
13176 second form, the backend will be used if the condition is not met. If no
13177 condition is valid, the backend defined with "default_backend" will be used.
13178 If no default backend is defined, either the servers in the same section are
13179 used (in case of a "listen" section) or, in case of a frontend, no server is
13180 used and a 503 service unavailable response is returned.
13181
Willy Tarreau51aecc72009-07-12 09:47:04 +020013182 Note that it is possible to switch from a TCP frontend to an HTTP backend. In
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +010013183 this case, either the frontend has already checked that the protocol is HTTP,
Willy Tarreau51aecc72009-07-12 09:47:04 +020013184 and backend processing will immediately follow, or the backend will wait for
13185 a complete HTTP request to get in. This feature is useful when a frontend
13186 must decode several protocols on a unique port, one of them being HTTP.
13187
Bertrand Jacquin702d44f2013-11-19 11:43:06 +010013188 When <backend> is a simple name, it is resolved at configuration time, and an
13189 error is reported if the specified backend does not exist. If <backend> is
13190 a log-format string instead, no check may be done at configuration time, so
13191 the backend name is resolved dynamically at run time. If the resulting
13192 backend name does not correspond to any valid backend, no other rule is
13193 evaluated, and the default_backend directive is applied instead. Note that
13194 when using dynamic backend names, it is highly recommended to use a prefix
13195 that no other backend uses in order to ensure that an unauthorized backend
13196 cannot be forced from the request.
13197
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030013198 It is worth mentioning that "use_backend" rules with an explicit name are
Bertrand Jacquin702d44f2013-11-19 11:43:06 +010013199 used to detect the association between frontends and backends to compute the
13200 backend's "fullconn" setting. This cannot be done for dynamic names.
13201
13202 See also: "default_backend", "tcp-request", "fullconn", "log-format", and
13203 section 7 about ACLs.
Willy Tarreaud72758d2010-01-12 10:42:19 +010013204
Christopher Fauletb30b3102019-09-12 23:03:09 +020013205use-fcgi-app <name>
13206 Defines the FastCGI application to use for the backend.
13207 May be used in sections : defaults | frontend | listen | backend
13208 no | no | yes | yes
13209 Arguments :
13210 <name> is the name of the FastCGI application to use.
13211
13212 See section 10.1 about FastCGI application setup for details.
Willy Tarreau036fae02008-01-06 13:24:40 +010013213
Willy Tarreau4a5cade2012-04-05 21:09:48 +020013214use-server <server> if <condition>
13215use-server <server> unless <condition>
13216 Only use a specific server if/unless an ACL-based condition is matched.
13217 May be used in sections : defaults | frontend | listen | backend
13218 no | no | yes | yes
13219 Arguments :
Jerome Magnin824186b2020-03-29 09:37:12 +020013220 <server> is the name of a valid server in the same backend section
13221 or a "log-format" string resolving to a server name.
Willy Tarreau4a5cade2012-04-05 21:09:48 +020013222
13223 <condition> is a condition composed of ACLs, as described in section 7.
13224
13225 By default, connections which arrive to a backend are load-balanced across
13226 the available servers according to the configured algorithm, unless a
13227 persistence mechanism such as a cookie is used and found in the request.
13228
13229 Sometimes it is desirable to forward a particular request to a specific
13230 server without having to declare a dedicated backend for this server. This
13231 can be achieved using the "use-server" rules. These rules are evaluated after
13232 the "redirect" rules and before evaluating cookies, and they have precedence
13233 on them. There may be as many "use-server" rules as desired. All of these
13234 rules are evaluated in their declaration order, and the first one which
13235 matches will assign the server.
13236
13237 If a rule designates a server which is down, and "option persist" is not used
13238 and no force-persist rule was validated, it is ignored and evaluation goes on
13239 with the next rules until one matches.
13240
13241 In the first form, the server will be used if the condition is met. In the
13242 second form, the server will be used if the condition is not met. If no
13243 condition is valid, the processing continues and the server will be assigned
13244 according to other persistence mechanisms.
13245
13246 Note that even if a rule is matched, cookie processing is still performed but
13247 does not assign the server. This allows prefixed cookies to have their prefix
13248 stripped.
13249
13250 The "use-server" statement works both in HTTP and TCP mode. This makes it
13251 suitable for use with content-based inspection. For instance, a server could
Lukas Tribusa267b5d2020-07-19 00:25:06 +020013252 be selected in a farm according to the TLS SNI field when using protocols with
Alex18c88772021-06-05 13:23:08 +020013253 implicit TLS (also see "req.ssl_sni"). And if these servers have their weight
Lukas Tribusa267b5d2020-07-19 00:25:06 +020013254 set to zero, they will not be used for other traffic.
Willy Tarreau4a5cade2012-04-05 21:09:48 +020013255
13256 Example :
13257 # intercept incoming TLS requests based on the SNI field
Alex18c88772021-06-05 13:23:08 +020013258 use-server www if { req.ssl_sni -i www.example.com }
Willy Tarreau4a5cade2012-04-05 21:09:48 +020013259 server www 192.168.0.1:443 weight 0
Alex18c88772021-06-05 13:23:08 +020013260 use-server mail if { req.ssl_sni -i mail.example.com }
Lukas Tribusa267b5d2020-07-19 00:25:06 +020013261 server mail 192.168.0.1:465 weight 0
Alex18c88772021-06-05 13:23:08 +020013262 use-server imap if { req.ssl_sni -i imap.example.com }
Lukas Tribus98a3e3f2017-03-26 12:55:35 +000013263 server imap 192.168.0.1:993 weight 0
Willy Tarreau4a5cade2012-04-05 21:09:48 +020013264 # all the rest is forwarded to this server
13265 server default 192.168.0.2:443 check
13266
Jerome Magnin824186b2020-03-29 09:37:12 +020013267 When <server> is a simple name, it is checked against existing servers in the
13268 configuration and an error is reported if the specified server does not exist.
13269 If it is a log-format, no check is performed when parsing the configuration,
13270 and if we can't resolve a valid server name at runtime but the use-server rule
Ilya Shipitsin11057a32020-06-21 21:18:27 +050013271 was conditioned by an ACL returning true, no other use-server rule is applied
Jerome Magnin824186b2020-03-29 09:37:12 +020013272 and we fall back to load balancing.
13273
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030013274 See also: "use_backend", section 5 about server and section 7 about ACLs.
Willy Tarreau4a5cade2012-04-05 21:09:48 +020013275
Willy Tarreaub6205fd2012-09-24 12:27:33 +020013276
Davor Ocelice9ed2812017-12-25 17:49:28 +0100132775. Bind and server options
Willy Tarreaub6205fd2012-09-24 12:27:33 +020013278--------------------------
13279
13280The "bind", "server" and "default-server" keywords support a number of settings
13281depending on some build options and on the system HAProxy was built on. These
13282settings generally each consist in one word sometimes followed by a value,
13283written on the same line as the "bind" or "server" line. All these options are
13284described in this section.
13285
13286
132875.1. Bind options
13288-----------------
13289
13290The "bind" keyword supports a certain number of settings which are all passed
13291as arguments on the same line. The order in which those arguments appear makes
13292no importance, provided that they appear after the bind address. All of these
13293parameters are optional. Some of them consist in a single words (booleans),
13294while other ones expect a value after them. In this case, the value must be
13295provided immediately after the setting name.
13296
13297The currently supported settings are the following ones.
13298
Bertrand Jacquin93b227d2016-06-04 15:11:10 +010013299accept-netscaler-cip <magic number>
13300 Enforces the use of the NetScaler Client IP insertion protocol over any
13301 connection accepted by any of the TCP sockets declared on the same line. The
13302 NetScaler Client IP insertion protocol dictates the layer 3/4 addresses of
13303 the incoming connection to be used everywhere an address is used, with the
13304 only exception of "tcp-request connection" rules which will only see the
13305 real connection address. Logs will reflect the addresses indicated in the
13306 protocol, unless it is violated, in which case the real address will still
13307 be used. This keyword combined with support from external components can be
13308 used as an efficient and reliable alternative to the X-Forwarded-For
Bertrand Jacquin90759682016-06-06 15:35:39 +010013309 mechanism which is not always reliable and not even always usable. See also
13310 "tcp-request connection expect-netscaler-cip" for a finer-grained setting of
13311 which client is allowed to use the protocol.
Bertrand Jacquin93b227d2016-06-04 15:11:10 +010013312
Willy Tarreaub6205fd2012-09-24 12:27:33 +020013313accept-proxy
13314 Enforces the use of the PROXY protocol over any connection accepted by any of
Willy Tarreau77992672014-06-14 11:06:17 +020013315 the sockets declared on the same line. Versions 1 and 2 of the PROXY protocol
13316 are supported and correctly detected. The PROXY protocol dictates the layer
Willy Tarreaub6205fd2012-09-24 12:27:33 +020013317 3/4 addresses of the incoming connection to be used everywhere an address is
13318 used, with the only exception of "tcp-request connection" rules which will
13319 only see the real connection address. Logs will reflect the addresses
13320 indicated in the protocol, unless it is violated, in which case the real
Davor Ocelice9ed2812017-12-25 17:49:28 +010013321 address will still be used. This keyword combined with support from external
Willy Tarreaub6205fd2012-09-24 12:27:33 +020013322 components can be used as an efficient and reliable alternative to the
13323 X-Forwarded-For mechanism which is not always reliable and not even always
Willy Tarreau4f0d9192013-06-11 20:40:55 +020013324 usable. See also "tcp-request connection expect-proxy" for a finer-grained
13325 setting of which client is allowed to use the protocol.
Willy Tarreaub6205fd2012-09-24 12:27:33 +020013326
Olivier Houchardc2aae742017-09-22 18:26:28 +020013327allow-0rtt
Bertrand Jacquina25282b2018-08-14 00:56:13 +010013328 Allow receiving early data when using TLSv1.3. This is disabled by default,
Olivier Houchard69752962019-01-08 15:35:32 +010013329 due to security considerations. Because it is vulnerable to replay attacks,
John Roeslerfb2fce12019-07-10 15:45:51 -050013330 you should only allow if for requests that are safe to replay, i.e. requests
Olivier Houchard69752962019-01-08 15:35:32 +010013331 that are idempotent. You can use the "wait-for-handshake" action for any
13332 request that wouldn't be safe with early data.
Olivier Houchardc2aae742017-09-22 18:26:28 +020013333
Willy Tarreauab861d32013-04-02 02:30:41 +020013334alpn <protocols>
13335 This enables the TLS ALPN extension and advertises the specified protocol
13336 list as supported on top of ALPN. The protocol list consists in a comma-
13337 delimited list of protocol names, for instance: "http/1.1,http/1.0" (without
John Roeslerfb2fce12019-07-10 15:45:51 -050013338 quotes). This requires that the SSL library is built with support for TLS
Willy Tarreauab861d32013-04-02 02:30:41 +020013339 extensions enabled (check with haproxy -vv). The ALPN extension replaces the
Willy Tarreau95c4e142017-11-26 12:18:55 +010013340 initial NPN extension. ALPN is required to enable HTTP/2 on an HTTP frontend.
13341 Versions of OpenSSL prior to 1.0.2 didn't support ALPN and only supposed the
13342 now obsolete NPN extension. At the time of writing this, most browsers still
13343 support both ALPN and NPN for HTTP/2 so a fallback to NPN may still work for
13344 a while. But ALPN must be used whenever possible. If both HTTP/2 and HTTP/1.1
13345 are expected to be supported, both versions can be advertised, in order of
13346 preference, like below :
13347
13348 bind :443 ssl crt pub.pem alpn h2,http/1.1
Willy Tarreauab861d32013-04-02 02:30:41 +020013349
Willy Tarreaub6205fd2012-09-24 12:27:33 +020013350backlog <backlog>
Willy Tarreaue2711c72019-02-27 15:39:41 +010013351 Sets the socket's backlog to this value. If unspecified or 0, the frontend's
Willy Tarreaub6205fd2012-09-24 12:27:33 +020013352 backlog is used instead, which generally defaults to the maxconn value.
13353
Emmanuel Hocdete7f2b732017-01-09 16:15:54 +010013354curves <curves>
13355 This setting is only available when support for OpenSSL was built in. It sets
13356 the string describing the list of elliptic curves algorithms ("curve suite")
13357 that are negotiated during the SSL/TLS handshake with ECDHE. The format of the
13358 string is a colon-delimited list of curve name.
13359 Example: "X25519:P-256" (without quote)
13360 When "curves" is set, "ecdhe" parameter is ignored.
13361
Emeric Brun7fb34422012-09-28 15:26:15 +020013362ecdhe <named curve>
13363 This setting is only available when support for OpenSSL was built in. It sets
Emeric Brun6924ef82013-03-06 14:08:53 +010013364 the named curve (RFC 4492) used to generate ECDH ephemeral keys. By default,
13365 used named curve is prime256v1.
Emeric Brun7fb34422012-09-28 15:26:15 +020013366
Emeric Brunfd33a262012-10-11 16:28:27 +020013367ca-file <cafile>
Emeric Brun1a073b42012-09-28 17:07:34 +020013368 This setting is only available when support for OpenSSL was built in. It
13369 designates a PEM file from which to load CA certificates used to verify
13370 client's certificate.
13371
Emeric Brunb6dc9342012-09-28 17:55:37 +020013372ca-ignore-err [all|<errorID>,...]
13373 This setting is only available when support for OpenSSL was built in.
13374 Sets a comma separated list of errorIDs to ignore during verify at depth > 0.
13375 If set to 'all', all errors are ignored. SSL handshake is not aborted if an
13376 error is ignored.
13377
Christopher Faulet31af49d2015-06-09 17:29:50 +020013378ca-sign-file <cafile>
13379 This setting is only available when support for OpenSSL was built in. It
13380 designates a PEM file containing both the CA certificate and the CA private
13381 key used to create and sign server's certificates. This is a mandatory
13382 setting when the dynamic generation of certificates is enabled. See
13383 'generate-certificates' for details.
13384
Bertrand Jacquind4d0a232016-11-13 16:37:12 +000013385ca-sign-pass <passphrase>
Christopher Faulet31af49d2015-06-09 17:29:50 +020013386 This setting is only available when support for OpenSSL was built in. It is
13387 the CA private key passphrase. This setting is optional and used only when
13388 the dynamic generation of certificates is enabled. See
13389 'generate-certificates' for details.
13390
Emmanuel Hocdet842e94e2019-12-16 16:39:17 +010013391ca-verify-file <cafile>
13392 This setting designates a PEM file from which to load CA certificates used to
13393 verify client's certificate. It designates CA certificates which must not be
13394 included in CA names sent in server hello message. Typically, "ca-file" must
13395 be defined with intermediate certificates, and "ca-verify-file" with
13396 certificates to ending the chain, like root CA.
13397
Willy Tarreaub6205fd2012-09-24 12:27:33 +020013398ciphers <ciphers>
13399 This setting is only available when support for OpenSSL was built in. It sets
13400 the string describing the list of cipher algorithms ("cipher suite") that are
Bertrand Jacquin8cf7c1e2019-02-03 18:35:25 +000013401 negotiated during the SSL/TLS handshake up to TLSv1.2. The format of the
Bertrand Jacquin4f03ab02019-02-03 18:48:49 +000013402 string is defined in "man 1 ciphers" from OpenSSL man pages. For background
Dirkjan Bussink415150f2018-09-14 11:14:21 +020013403 information and recommendations see e.g.
13404 (https://wiki.mozilla.org/Security/Server_Side_TLS) and
13405 (https://mozilla.github.io/server-side-tls/ssl-config-generator/). For TLSv1.3
13406 cipher configuration, please check the "ciphersuites" keyword.
13407
13408ciphersuites <ciphersuites>
13409 This setting is only available when support for OpenSSL was built in and
13410 OpenSSL 1.1.1 or later was used to build HAProxy. It sets the string describing
13411 the list of cipher algorithms ("cipher suite") that are negotiated during the
13412 TLSv1.3 handshake. The format of the string is defined in "man 1 ciphers" from
Bertrand Jacquin4f03ab02019-02-03 18:48:49 +000013413 OpenSSL man pages under the "ciphersuites" section. For cipher configuration
13414 for TLSv1.2 and earlier, please check the "ciphers" keyword.
Willy Tarreaub6205fd2012-09-24 12:27:33 +020013415
Emeric Brunfd33a262012-10-11 16:28:27 +020013416crl-file <crlfile>
Emeric Brun1a073b42012-09-28 17:07:34 +020013417 This setting is only available when support for OpenSSL was built in. It
13418 designates a PEM file from which to load certificate revocation list used
Remi Tricot-Le Breton02bd6842021-05-04 12:22:34 +020013419 to verify client's certificate. You need to provide a certificate revocation
13420 list for every certificate of your certificate authority chain.
Emeric Brun1a073b42012-09-28 17:07:34 +020013421
Willy Tarreaub6205fd2012-09-24 12:27:33 +020013422crt <cert>
Alex Davies0fbf0162013-03-02 16:04:50 +000013423 This setting is only available when support for OpenSSL was built in. It
13424 designates a PEM file containing both the required certificates and any
13425 associated private keys. This file can be built by concatenating multiple
13426 PEM files into one (e.g. cat cert.pem key.pem > combined.pem). If your CA
13427 requires an intermediate certificate, this can also be concatenated into this
Emmanuel Hocdet70df7bf2019-01-04 11:08:20 +010013428 file. Intermediate certificate can also be shared in a directory via
13429 "issuers-chain-path" directive.
Alex Davies0fbf0162013-03-02 16:04:50 +000013430
William Lallemand4c5adbf2020-02-24 14:23:22 +010013431 If the file does not contain a private key, HAProxy will try to load
13432 the key at the same path suffixed by a ".key".
13433
Alex Davies0fbf0162013-03-02 16:04:50 +000013434 If the OpenSSL used supports Diffie-Hellman, parameters present in this file
13435 are loaded.
13436
13437 If a directory name is used instead of a PEM file, then all files found in
William Lallemand3f25ae32020-02-24 16:30:12 +010013438 that directory will be loaded in alphabetic order unless their name ends
13439 with '.key', '.issuer', '.ocsp' or '.sctl' (reserved extensions). This
13440 directive may be specified multiple times in order to load certificates from
13441 multiple files or directories. The certificates will be presented to clients
13442 who provide a valid TLS Server Name Indication field matching one of their
13443 CN or alt subjects. Wildcards are supported, where a wildcard character '*'
13444 is used instead of the first hostname component (e.g. *.example.org matches
Janusz Dziemidowicz2c701b52015-03-07 23:03:59 +010013445 www.example.org but not www.sub.example.org).
Alex Davies0fbf0162013-03-02 16:04:50 +000013446
13447 If no SNI is provided by the client or if the SSL library does not support
13448 TLS extensions, or if the client provides an SNI hostname which does not
13449 match any certificate, then the first loaded certificate will be presented.
13450 This means that when loading certificates from a directory, it is highly
Cyril Bonté3180f7b2015-01-25 00:16:08 +010013451 recommended to load the default one first as a file or to ensure that it will
13452 always be the first one in the directory.
Alex Davies0fbf0162013-03-02 16:04:50 +000013453
Emeric Brune032bfa2012-09-28 13:01:45 +020013454 Note that the same cert may be loaded multiple times without side effects.
Willy Tarreaub6205fd2012-09-24 12:27:33 +020013455
Davor Ocelice9ed2812017-12-25 17:49:28 +010013456 Some CAs (such as GoDaddy) offer a drop down list of server types that do not
Alex Davies0fbf0162013-03-02 16:04:50 +000013457 include HAProxy when obtaining a certificate. If this happens be sure to
Davor Ocelice9ed2812017-12-25 17:49:28 +010013458 choose a web server that the CA believes requires an intermediate CA (for
13459 GoDaddy, selection Apache Tomcat will get the correct bundle, but many
Alex Davies0fbf0162013-03-02 16:04:50 +000013460 others, e.g. nginx, result in a wrong bundle that will not work for some
13461 clients).
13462
Daniel Corbett9f0843f2021-05-08 10:50:37 -040013463 For each PEM file, HAProxy checks for the presence of file at the same path
Emeric Brun4147b2e2014-06-16 18:36:30 +020013464 suffixed by ".ocsp". If such file is found, support for the TLS Certificate
13465 Status Request extension (also known as "OCSP stapling") is automatically
13466 enabled. The content of this file is optional. If not empty, it must contain
13467 a valid OCSP Response in DER format. In order to be valid an OCSP Response
13468 must comply with the following rules: it has to indicate a good status,
13469 it has to be a single response for the certificate of the PEM file, and it
13470 has to be valid at the moment of addition. If these rules are not respected
13471 the OCSP Response is ignored and a warning is emitted. In order to identify
13472 which certificate an OCSP Response applies to, the issuer's certificate is
13473 necessary. If the issuer's certificate is not found in the PEM file, it will
13474 be loaded from a file at the same path as the PEM file suffixed by ".issuer"
13475 if it exists otherwise it will fail with an error.
13476
Daniel Corbett9f0843f2021-05-08 10:50:37 -040013477 For each PEM file, HAProxy also checks for the presence of file at the same
Janusz Dziemidowicz2c701b52015-03-07 23:03:59 +010013478 path suffixed by ".sctl". If such file is found, support for Certificate
13479 Transparency (RFC6962) TLS extension is enabled. The file must contain a
13480 valid Signed Certificate Timestamp List, as described in RFC. File is parsed
13481 to check basic syntax, but no signatures are verified.
13482
yanbzhu6c25e9e2016-01-05 12:52:02 -050013483 There are cases where it is desirable to support multiple key types, e.g. RSA
13484 and ECDSA in the cipher suites offered to the clients. This allows clients
13485 that support EC certificates to be able to use EC ciphers, while
13486 simultaneously supporting older, RSA only clients.
yanbzhud19630c2015-12-14 15:10:25 -050013487
William Lallemandf9ff3ec2020-10-02 17:57:44 +020013488 To achieve this, OpenSSL 1.1.1 is required, you can configure this behavior
13489 by providing one crt entry per certificate type, or by configuring a "cert
13490 bundle" like it was required before HAProxy 1.8. See "ssl-load-extra-files".
yanbzhud19630c2015-12-14 15:10:25 -050013491
Emeric Brunb6dc9342012-09-28 17:55:37 +020013492crt-ignore-err <errors>
Alex Davies0fbf0162013-03-02 16:04:50 +000013493 This setting is only available when support for OpenSSL was built in. Sets a
Davor Ocelice9ed2812017-12-25 17:49:28 +010013494 comma separated list of errorIDs to ignore during verify at depth == 0. If
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030013495 set to 'all', all errors are ignored. SSL handshake is not aborted if an error
Alex Davies0fbf0162013-03-02 16:04:50 +000013496 is ignored.
Emeric Brunb6dc9342012-09-28 17:55:37 +020013497
Emmanuel Hocdetfe616562013-01-22 15:31:15 +010013498crt-list <file>
13499 This setting is only available when support for OpenSSL was built in. It
Emmanuel Hocdet98263292016-12-29 18:26:15 +010013500 designates a list of PEM file with an optional ssl configuration and a SNI
13501 filter per certificate, with the following format for each line :
Emmanuel Hocdetfe616562013-01-22 15:31:15 +010013502
Emmanuel Hocdet98263292016-12-29 18:26:15 +010013503 <crtfile> [\[<sslbindconf> ...\]] [[!]<snifilter> ...]
13504
William Lallemand5d036392020-06-30 16:11:36 +020013505 sslbindconf supports "allow-0rtt", "alpn", "ca-file", "ca-verify-file",
13506 "ciphers", "ciphersuites", "crl-file", "curves", "ecdhe", "no-ca-names",
13507 "npn", "verify" configuration. With BoringSSL and Openssl >= 1.1.1
13508 "ssl-min-ver" and "ssl-max-ver" are also supported. It overrides the
13509 configuration set in bind line for the certificate.
Emmanuel Hocdetfe616562013-01-22 15:31:15 +010013510
Emmanuel Hocdet7c41a1b2013-05-07 20:20:06 +020013511 Wildcards are supported in the SNI filter. Negative filter are also supported,
Joao Moraise51fab02020-11-21 07:42:20 -030013512 useful in combination with a wildcard filter to exclude a particular SNI, or
13513 after the first certificate to exclude a pattern from its CN or Subject Alt
13514 Name (SAN). The certificates will be presented to clients who provide a valid
13515 TLS Server Name Indication field matching one of the SNI filters. If no SNI
13516 filter is specified, the CN and SAN are used. This directive may be specified
Emmanuel Hocdet7c41a1b2013-05-07 20:20:06 +020013517 multiple times. See the "crt" option for more information. The default
13518 certificate is still needed to meet OpenSSL expectations. If it is not used,
13519 the 'strict-sni' option may be used.
Emmanuel Hocdetfe616562013-01-22 15:31:15 +010013520
William Lallemandf9ff3ec2020-10-02 17:57:44 +020013521 Multi-cert bundling (see "ssl-load-extra-files") is supported with crt-list,
13522 as long as only the base name is given in the crt-list. SNI filter will do
13523 the same work on all bundled certificates.
yanbzhud19630c2015-12-14 15:10:25 -050013524
William Lallemand7c26ed72020-06-03 17:34:48 +020013525 Empty lines as well as lines beginning with a hash ('#') will be ignored.
13526
Joao Moraisaa8fcc42020-11-24 08:24:30 -030013527 The first declared certificate of a bind line is used as the default
Daniel Corbett9f0843f2021-05-08 10:50:37 -040013528 certificate, either from crt or crt-list option, which HAProxy should use in
Joao Moraisaa8fcc42020-11-24 08:24:30 -030013529 the TLS handshake if no other certificate matches. This certificate will also
13530 be used if the provided SNI matches its CN or SAN, even if a matching SNI
13531 filter is found on any crt-list. The SNI filter !* can be used after the first
13532 declared certificate to not include its CN and SAN in the SNI tree, so it will
13533 never match except if no other certificate matches. This way the first
13534 declared certificate act as a fallback.
Joao Moraise51fab02020-11-21 07:42:20 -030013535
Emmanuel Hocdet98263292016-12-29 18:26:15 +010013536 crt-list file example:
Joao Moraise51fab02020-11-21 07:42:20 -030013537 cert1.pem !*
William Lallemand7c26ed72020-06-03 17:34:48 +020013538 # comment
Emmanuel Hocdet05942112017-02-20 16:11:50 +010013539 cert2.pem [alpn h2,http/1.1]
Emmanuel Hocdet98263292016-12-29 18:26:15 +010013540 certW.pem *.domain.tld !secure.domain.tld
Emmanuel Hocdet05942112017-02-20 16:11:50 +010013541 certS.pem [curves X25519:P-256 ciphers ECDHE-ECDSA-AES256-GCM-SHA384] secure.domain.tld
Emmanuel Hocdet98263292016-12-29 18:26:15 +010013542
Willy Tarreaub6205fd2012-09-24 12:27:33 +020013543defer-accept
13544 Is an optional keyword which is supported only on certain Linux kernels. It
13545 states that a connection will only be accepted once some data arrive on it,
13546 or at worst after the first retransmit. This should be used only on protocols
Davor Ocelice9ed2812017-12-25 17:49:28 +010013547 for which the client talks first (e.g. HTTP). It can slightly improve
Willy Tarreaub6205fd2012-09-24 12:27:33 +020013548 performance by ensuring that most of the request is already available when
13549 the connection is accepted. On the other hand, it will not be able to detect
13550 connections which don't talk. It is important to note that this option is
13551 broken in all kernels up to 2.6.31, as the connection is never accepted until
13552 the client talks. This can cause issues with front firewalls which would see
13553 an established connection while the proxy will only see it in SYN_RECV. This
13554 option is only supported on TCPv4/TCPv6 sockets and ignored by other ones.
13555
William Lallemandf6975e92017-05-26 17:42:10 +020013556expose-fd listeners
13557 This option is only usable with the stats socket. It gives your stats socket
13558 the capability to pass listeners FD to another HAProxy process.
William Lallemande202b1e2017-06-01 17:38:56 +020013559 During a reload with the master-worker mode, the process is automatically
13560 reexecuted adding -x and one of the stats socket with this option.
Davor Ocelice9ed2812017-12-25 17:49:28 +010013561 See also "-x" in the management guide.
William Lallemandf6975e92017-05-26 17:42:10 +020013562
Emeric Brun2cb7ae52012-10-05 14:14:21 +020013563force-sslv3
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030013564 This option enforces use of SSLv3 only on SSL connections instantiated from
Emeric Brun2cb7ae52012-10-05 14:14:21 +020013565 this listener. SSLv3 is generally less expensive than the TLS counterparts
Emeric Brun2c86cbf2014-10-30 15:56:50 +010013566 for high connection rates. This option is also available on global statement
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020013567 "ssl-default-bind-options". See also "ssl-min-ver" and "ssl-max-ver".
Emeric Brun2cb7ae52012-10-05 14:14:21 +020013568
13569force-tlsv10
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030013570 This option enforces use of TLSv1.0 only on SSL connections instantiated from
Emeric Brun2c86cbf2014-10-30 15:56:50 +010013571 this listener. This option is also available on global statement
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020013572 "ssl-default-bind-options". See also "ssl-min-ver" and "ssl-max-ver".
Emeric Brun2cb7ae52012-10-05 14:14:21 +020013573
13574force-tlsv11
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030013575 This option enforces use of TLSv1.1 only on SSL connections instantiated from
Emeric Brun2c86cbf2014-10-30 15:56:50 +010013576 this listener. This option is also available on global statement
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020013577 "ssl-default-bind-options". See also "ssl-min-ver" and "ssl-max-ver".
Emeric Brun2cb7ae52012-10-05 14:14:21 +020013578
13579force-tlsv12
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030013580 This option enforces use of TLSv1.2 only on SSL connections instantiated from
Emeric Brun2c86cbf2014-10-30 15:56:50 +010013581 this listener. This option is also available on global statement
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020013582 "ssl-default-bind-options". See also "ssl-min-ver" and "ssl-max-ver".
Emeric Brun2cb7ae52012-10-05 14:14:21 +020013583
Emmanuel Hocdet42fb9802017-03-30 19:29:39 +020013584force-tlsv13
13585 This option enforces use of TLSv1.3 only on SSL connections instantiated from
13586 this listener. This option is also available on global statement
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020013587 "ssl-default-bind-options". See also "ssl-min-ver" and "ssl-max-ver".
Emmanuel Hocdet42fb9802017-03-30 19:29:39 +020013588
Christopher Faulet31af49d2015-06-09 17:29:50 +020013589generate-certificates
13590 This setting is only available when support for OpenSSL was built in. It
13591 enables the dynamic SSL certificates generation. A CA certificate and its
13592 private key are necessary (see 'ca-sign-file'). When HAProxy is configured as
13593 a transparent forward proxy, SSL requests generate errors because of a common
13594 name mismatch on the certificate presented to the client. With this option
13595 enabled, HAProxy will try to forge a certificate using the SNI hostname
13596 indicated by the client. This is done only if no certificate matches the SNI
13597 hostname (see 'crt-list'). If an error occurs, the default certificate is
13598 used, else the 'strict-sni' option is set.
13599 It can also be used when HAProxy is configured as a reverse proxy to ease the
13600 deployment of an architecture with many backends.
13601
13602 Creating a SSL certificate is an expensive operation, so a LRU cache is used
13603 to store forged certificates (see 'tune.ssl.ssl-ctx-cache-size'). It
Davor Ocelice9ed2812017-12-25 17:49:28 +010013604 increases the HAProxy's memory footprint to reduce latency when the same
Christopher Faulet31af49d2015-06-09 17:29:50 +020013605 certificate is used many times.
13606
Willy Tarreaub6205fd2012-09-24 12:27:33 +020013607gid <gid>
13608 Sets the group of the UNIX sockets to the designated system gid. It can also
13609 be set by default in the global section's "unix-bind" statement. Note that
13610 some platforms simply ignore this. This setting is equivalent to the "group"
13611 setting except that the group ID is used instead of its name. This setting is
13612 ignored by non UNIX sockets.
13613
13614group <group>
13615 Sets the group of the UNIX sockets to the designated system group. It can
13616 also be set by default in the global section's "unix-bind" statement. Note
13617 that some platforms simply ignore this. This setting is equivalent to the
13618 "gid" setting except that the group name is used instead of its gid. This
13619 setting is ignored by non UNIX sockets.
13620
13621id <id>
13622 Fixes the socket ID. By default, socket IDs are automatically assigned, but
13623 sometimes it is more convenient to fix them to ease monitoring. This value
13624 must be strictly positive and unique within the listener/frontend. This
13625 option can only be used when defining only a single socket.
13626
13627interface <interface>
Lukas Tribusfce2e962013-02-12 22:13:19 +010013628 Restricts the socket to a specific interface. When specified, only packets
13629 received from that particular interface are processed by the socket. This is
13630 currently only supported on Linux. The interface must be a primary system
13631 interface, not an aliased interface. It is also possible to bind multiple
13632 frontends to the same address if they are bound to different interfaces. Note
13633 that binding to a network interface requires root privileges. This parameter
Jérôme Magnin61275192018-02-07 11:39:58 +010013634 is only compatible with TCPv4/TCPv6 sockets. When specified, return traffic
13635 uses the same interface as inbound traffic, and its associated routing table,
13636 even if there are explicit routes through different interfaces configured.
13637 This can prove useful to address asymmetric routing issues when the same
13638 client IP addresses need to be able to reach frontends hosted on different
13639 interfaces.
Willy Tarreaub6205fd2012-09-24 12:27:33 +020013640
Willy Tarreauabb175f2012-09-24 12:43:26 +020013641level <level>
13642 This setting is used with the stats sockets only to restrict the nature of
13643 the commands that can be issued on the socket. It is ignored by other
13644 sockets. <level> can be one of :
Davor Ocelice9ed2812017-12-25 17:49:28 +010013645 - "user" is the least privileged level; only non-sensitive stats can be
Willy Tarreauabb175f2012-09-24 12:43:26 +020013646 read, and no change is allowed. It would make sense on systems where it
13647 is not easy to restrict access to the socket.
13648 - "operator" is the default level and fits most common uses. All data can
Davor Ocelice9ed2812017-12-25 17:49:28 +010013649 be read, and only non-sensitive changes are permitted (e.g. clear max
Willy Tarreauabb175f2012-09-24 12:43:26 +020013650 counters).
Davor Ocelice9ed2812017-12-25 17:49:28 +010013651 - "admin" should be used with care, as everything is permitted (e.g. clear
Willy Tarreauabb175f2012-09-24 12:43:26 +020013652 all counters).
13653
Andjelko Iharosc4df59e2017-07-20 11:59:48 +020013654severity-output <format>
13655 This setting is used with the stats sockets only to configure severity
13656 level output prepended to informational feedback messages. Severity
13657 level of messages can range between 0 and 7, conforming to syslog
13658 rfc5424. Valid and successful socket commands requesting data
13659 (i.e. "show map", "get acl foo" etc.) will never have a severity level
13660 prepended. It is ignored by other sockets. <format> can be one of :
13661 - "none" (default) no severity level is prepended to feedback messages.
13662 - "number" severity level is prepended as a number.
13663 - "string" severity level is prepended as a string following the
13664 rfc5424 convention.
13665
Willy Tarreaub6205fd2012-09-24 12:27:33 +020013666maxconn <maxconn>
13667 Limits the sockets to this number of concurrent connections. Extraneous
13668 connections will remain in the system's backlog until a connection is
13669 released. If unspecified, the limit will be the same as the frontend's
13670 maxconn. Note that in case of port ranges or multiple addresses, the same
13671 value will be applied to each socket. This setting enables different
13672 limitations on expensive sockets, for instance SSL entries which may easily
13673 eat all memory.
13674
13675mode <mode>
13676 Sets the octal mode used to define access permissions on the UNIX socket. It
13677 can also be set by default in the global section's "unix-bind" statement.
13678 Note that some platforms simply ignore this. This setting is ignored by non
13679 UNIX sockets.
13680
13681mss <maxseg>
13682 Sets the TCP Maximum Segment Size (MSS) value to be advertised on incoming
13683 connections. This can be used to force a lower MSS for certain specific
13684 ports, for instance for connections passing through a VPN. Note that this
13685 relies on a kernel feature which is theoretically supported under Linux but
13686 was buggy in all versions prior to 2.6.28. It may or may not work on other
13687 operating systems. It may also not change the advertised value but change the
13688 effective size of outgoing segments. The commonly advertised value for TCPv4
13689 over Ethernet networks is 1460 = 1500(MTU) - 40(IP+TCP). If this value is
13690 positive, it will be used as the advertised MSS. If it is negative, it will
13691 indicate by how much to reduce the incoming connection's advertised MSS for
13692 outgoing segments. This parameter is only compatible with TCP v4/v6 sockets.
13693
13694name <name>
13695 Sets an optional name for these sockets, which will be reported on the stats
13696 page.
13697
Willy Tarreaud72f0f32015-10-13 14:50:22 +020013698namespace <name>
13699 On Linux, it is possible to specify which network namespace a socket will
13700 belong to. This directive makes it possible to explicitly bind a listener to
13701 a namespace different from the default one. Please refer to your operating
13702 system's documentation to find more details about network namespaces.
13703
Willy Tarreaub6205fd2012-09-24 12:27:33 +020013704nice <nice>
13705 Sets the 'niceness' of connections initiated from the socket. Value must be
13706 in the range -1024..1024 inclusive, and defaults to zero. Positive values
13707 means that such connections are more friendly to others and easily offer
13708 their place in the scheduler. On the opposite, negative values mean that
13709 connections want to run with a higher priority than others. The difference
13710 only happens under high loads when the system is close to saturation.
13711 Negative values are appropriate for low-latency or administration services,
13712 and high values are generally recommended for CPU intensive tasks such as SSL
13713 processing or bulk transfers which are less sensible to latency. For example,
13714 it may make sense to use a positive value for an SMTP socket and a negative
13715 one for an RDP socket.
13716
Emmanuel Hocdet174dfe52017-07-28 15:01:05 +020013717no-ca-names
13718 This setting is only available when support for OpenSSL was built in. It
13719 prevents from send CA names in server hello message when ca-file is used.
Emmanuel Hocdet842e94e2019-12-16 16:39:17 +010013720 Use "ca-verify-file" instead of "ca-file" with "no-ca-names".
Emmanuel Hocdet174dfe52017-07-28 15:01:05 +020013721
Emeric Brun9b3009b2012-10-05 11:55:06 +020013722no-sslv3
Willy Tarreaub6205fd2012-09-24 12:27:33 +020013723 This setting is only available when support for OpenSSL was built in. It
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030013724 disables support for SSLv3 on any sockets instantiated from the listener when
Willy Tarreaub6205fd2012-09-24 12:27:33 +020013725 SSL is supported. Note that SSLv2 is forced disabled in the code and cannot
Emeric Brun2c86cbf2014-10-30 15:56:50 +010013726 be enabled using any configuration option. This option is also available on
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020013727 global statement "ssl-default-bind-options". Use "ssl-min-ver" and
13728 "ssl-max-ver" instead.
Willy Tarreaub6205fd2012-09-24 12:27:33 +020013729
Emeric Brun90ad8722012-10-02 14:00:59 +020013730no-tls-tickets
13731 This setting is only available when support for OpenSSL was built in. It
13732 disables the stateless session resumption (RFC 5077 TLS Ticket
13733 extension) and force to use stateful session resumption. Stateless
Emeric Brun2c86cbf2014-10-30 15:56:50 +010013734 session resumption is more expensive in CPU usage. This option is also
13735 available on global statement "ssl-default-bind-options".
Lukas Tribusbdb386d2020-03-10 00:56:09 +010013736 The TLS ticket mechanism is only used up to TLS 1.2.
13737 Forward Secrecy is compromised with TLS tickets, unless ticket keys
13738 are periodically rotated (via reload or by using "tls-ticket-keys").
Emeric Brun90ad8722012-10-02 14:00:59 +020013739
Emeric Brun9b3009b2012-10-05 11:55:06 +020013740no-tlsv10
Willy Tarreaub6205fd2012-09-24 12:27:33 +020013741 This setting is only available when support for OpenSSL was built in. It
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030013742 disables support for TLSv1.0 on any sockets instantiated from the listener
Emeric Brun2cb7ae52012-10-05 14:14:21 +020013743 when SSL is supported. Note that SSLv2 is forced disabled in the code and
Emeric Brun2c86cbf2014-10-30 15:56:50 +010013744 cannot be enabled using any configuration option. This option is also
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020013745 available on global statement "ssl-default-bind-options". Use "ssl-min-ver"
13746 and "ssl-max-ver" instead.
Willy Tarreaub6205fd2012-09-24 12:27:33 +020013747
Emeric Brun9b3009b2012-10-05 11:55:06 +020013748no-tlsv11
Emeric Brunf5da4932012-09-28 19:42:54 +020013749 This setting is only available when support for OpenSSL was built in. It
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030013750 disables support for TLSv1.1 on any sockets instantiated from the listener
Emeric Brun2cb7ae52012-10-05 14:14:21 +020013751 when SSL is supported. Note that SSLv2 is forced disabled in the code and
Emeric Brun2c86cbf2014-10-30 15:56:50 +010013752 cannot be enabled using any configuration option. This option is also
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020013753 available on global statement "ssl-default-bind-options". Use "ssl-min-ver"
13754 and "ssl-max-ver" instead.
Emeric Brunf5da4932012-09-28 19:42:54 +020013755
Emeric Brun9b3009b2012-10-05 11:55:06 +020013756no-tlsv12
Emeric Brunf5da4932012-09-28 19:42:54 +020013757 This setting is only available when support for OpenSSL was built in. It
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030013758 disables support for TLSv1.2 on any sockets instantiated from the listener
Emeric Brun2cb7ae52012-10-05 14:14:21 +020013759 when SSL is supported. Note that SSLv2 is forced disabled in the code and
Emeric Brun2c86cbf2014-10-30 15:56:50 +010013760 cannot be enabled using any configuration option. This option is also
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020013761 available on global statement "ssl-default-bind-options". Use "ssl-min-ver"
13762 and "ssl-max-ver" instead.
Emeric Brunf5da4932012-09-28 19:42:54 +020013763
Emmanuel Hocdet42fb9802017-03-30 19:29:39 +020013764no-tlsv13
13765 This setting is only available when support for OpenSSL was built in. It
13766 disables support for TLSv1.3 on any sockets instantiated from the listener
13767 when SSL is supported. Note that SSLv2 is forced disabled in the code and
13768 cannot be enabled using any configuration option. This option is also
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020013769 available on global statement "ssl-default-bind-options". Use "ssl-min-ver"
13770 and "ssl-max-ver" instead.
Emmanuel Hocdet42fb9802017-03-30 19:29:39 +020013771
Willy Tarreau6c9a3d52012-10-18 18:57:14 +020013772npn <protocols>
13773 This enables the NPN TLS extension and advertises the specified protocol list
13774 as supported on top of NPN. The protocol list consists in a comma-delimited
13775 list of protocol names, for instance: "http/1.1,http/1.0" (without quotes).
John Roeslerfb2fce12019-07-10 15:45:51 -050013776 This requires that the SSL library is built with support for TLS extensions
Willy Tarreauab861d32013-04-02 02:30:41 +020013777 enabled (check with haproxy -vv). Note that the NPN extension has been
Willy Tarreau95c4e142017-11-26 12:18:55 +010013778 replaced with the ALPN extension (see the "alpn" keyword), though this one is
13779 only available starting with OpenSSL 1.0.2. If HTTP/2 is desired on an older
13780 version of OpenSSL, NPN might still be used as most clients still support it
13781 at the time of writing this. It is possible to enable both NPN and ALPN
13782 though it probably doesn't make any sense out of testing.
Willy Tarreau6c9a3d52012-10-18 18:57:14 +020013783
Lukas Tribus53ae85c2017-05-04 15:45:40 +000013784prefer-client-ciphers
13785 Use the client's preference when selecting the cipher suite, by default
13786 the server's preference is enforced. This option is also available on
13787 global statement "ssl-default-bind-options".
Lukas Tribus926594f2018-05-18 17:55:57 +020013788 Note that with OpenSSL >= 1.1.1 ChaCha20-Poly1305 is reprioritized anyway
13789 (without setting this option), if a ChaCha20-Poly1305 cipher is at the top of
13790 the client cipher list.
Lukas Tribus53ae85c2017-05-04 15:45:40 +000013791
Christopher Fauletc644fa92017-11-23 22:44:11 +010013792process <process-set>[/<thread-set>]
Willy Tarreaua36b3242019-02-02 13:14:34 +010013793 This restricts the list of processes or threads on which this listener is
Christopher Fauletc644fa92017-11-23 22:44:11 +010013794 allowed to run. It does not enforce any process but eliminates those which do
Davor Ocelice9ed2812017-12-25 17:49:28 +010013795 not match. If the frontend uses a "bind-process" setting, the intersection
Christopher Fauletc644fa92017-11-23 22:44:11 +010013796 between the two is applied. If in the end the listener is not allowed to run
13797 on any remaining process, a warning is emitted, and the listener will either
13798 run on the first process of the listener if a single process was specified,
13799 or on all of its processes if multiple processes were specified. If a thread
Davor Ocelice9ed2812017-12-25 17:49:28 +010013800 set is specified, it limits the threads allowed to process incoming
Willy Tarreaua36b3242019-02-02 13:14:34 +010013801 connections for this listener, for the the process set. If multiple processes
13802 and threads are configured, a warning is emitted, as it either results from a
13803 configuration error or a misunderstanding of these models. For the unlikely
13804 case where several ranges are needed, this directive may be repeated.
13805 <process-set> and <thread-set> must use the format
Christopher Fauletc644fa92017-11-23 22:44:11 +010013806
13807 all | odd | even | number[-[number]]
13808
13809 Ranges can be partially defined. The higher bound can be omitted. In such
13810 case, it is replaced by the corresponding maximum value. The main purpose of
13811 this directive is to be used with the stats sockets and have one different
13812 socket per process. The second purpose is to have multiple bind lines sharing
13813 the same IP:port but not the same process in a listener, so that the system
13814 can distribute the incoming connections into multiple queues and allow a
13815 smoother inter-process load balancing. Currently Linux 3.9 and above is known
13816 for supporting this. See also "bind-process" and "nbproc".
Willy Tarreau6ae1ba62014-05-07 19:01:58 +020013817
Christopher Fauleta717b992018-04-10 14:43:00 +020013818proto <name>
13819 Forces the multiplexer's protocol to use for the incoming connections. It
13820 must be compatible with the mode of the frontend (TCP or HTTP). It must also
13821 be usable on the frontend side. The list of available protocols is reported
Christopher Faulet982e17d2021-03-26 14:44:18 +010013822 in haproxy -vv. The protocols properties are reported : the mode (TCP/HTTP),
13823 the side (FE/BE), the mux name and its flags.
13824
13825 Some protocols report errors on aborts (flag=CLEAN_ABRT). Some others are
13826 subject to the head-of-line blocking on server side (flag=HOL_RISK). Finally
13827 some protocols don't support upgrades (flag=NO_UPG). The HTX compatibility is
13828 also reported (flag=HTX).
13829
13830 Here are the protocols that may be used as argument to a "proto" directive on
13831 a bind line :
13832
13833 h2 : mode=HTTP side=FE|BE mux=H2 flags=HTX|CLEAN_ABRT|HOL_RISK|NO_UPG
13834 h1 : mode=HTTP side=FE|BE mux=H1 flags=HTX|NO_UPG
13835 none : mode=TCP side=FE|BE mux=PASS flags=NO_UPG
13836
Daniel Corbett67a82712020-07-06 23:01:19 -040013837 Idea behind this option is to bypass the selection of the best multiplexer's
Christopher Fauleta717b992018-04-10 14:43:00 +020013838 protocol for all connections instantiated from this listening socket. For
Joseph Herlant71b4b152018-11-13 16:55:16 -080013839 instance, it is possible to force the http/2 on clear TCP by specifying "proto
Christopher Fauleta717b992018-04-10 14:43:00 +020013840 h2" on the bind line.
13841
Willy Tarreaub6205fd2012-09-24 12:27:33 +020013842ssl
13843 This setting is only available when support for OpenSSL was built in. It
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030013844 enables SSL deciphering on connections instantiated from this listener. A
Willy Tarreaub6205fd2012-09-24 12:27:33 +020013845 certificate is necessary (see "crt" above). All contents in the buffers will
13846 appear in clear text, so that ACLs and HTTP processing will only have access
Emmanuel Hocdetbd695fe2017-05-15 15:53:41 +020013847 to deciphered contents. SSLv3 is disabled per default, use "ssl-min-ver SSLv3"
13848 to enable it.
Willy Tarreaub6205fd2012-09-24 12:27:33 +020013849
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020013850ssl-max-ver [ SSLv3 | TLSv1.0 | TLSv1.1 | TLSv1.2 | TLSv1.3 ]
13851 This option enforces use of <version> or lower on SSL connections instantiated
William Lallemand50df1cb2020-06-02 10:52:24 +020013852 from this listener. Using this setting without "ssl-min-ver" can be
13853 ambiguous because the default ssl-min-ver value could change in future HAProxy
13854 versions. This option is also available on global statement
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020013855 "ssl-default-bind-options". See also "ssl-min-ver".
13856
13857ssl-min-ver [ SSLv3 | TLSv1.0 | TLSv1.1 | TLSv1.2 | TLSv1.3 ]
William Lallemand50df1cb2020-06-02 10:52:24 +020013858 This option enforces use of <version> or upper on SSL connections
13859 instantiated from this listener. The default value is "TLSv1.2". This option
13860 is also available on global statement "ssl-default-bind-options".
13861 See also "ssl-max-ver".
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020013862
Emmanuel Hocdet65623372013-01-24 17:17:15 +010013863strict-sni
13864 This setting is only available when support for OpenSSL was built in. The
13865 SSL/TLS negotiation is allow only if the client provided an SNI which match
13866 a certificate. The default certificate is not used.
13867 See the "crt" option for more information.
13868
Willy Tarreau2af207a2015-02-04 00:45:58 +010013869tcp-ut <delay>
Tim Düsterhus4896c442016-11-29 02:15:19 +010013870 Sets the TCP User Timeout for all incoming connections instantiated from this
Willy Tarreau2af207a2015-02-04 00:45:58 +010013871 listening socket. This option is available on Linux since version 2.6.37. It
Daniel Corbett9f0843f2021-05-08 10:50:37 -040013872 allows HAProxy to configure a timeout for sockets which contain data not
Davor Ocelice9ed2812017-12-25 17:49:28 +010013873 receiving an acknowledgment for the configured delay. This is especially
Willy Tarreau2af207a2015-02-04 00:45:58 +010013874 useful on long-lived connections experiencing long idle periods such as
13875 remote terminals or database connection pools, where the client and server
13876 timeouts must remain high to allow a long period of idle, but where it is
13877 important to detect that the client has disappeared in order to release all
13878 resources associated with its connection (and the server's session). The
13879 argument is a delay expressed in milliseconds by default. This only works
13880 for regular TCP connections, and is ignored for other protocols.
13881
Willy Tarreau1c862c52012-10-05 16:21:00 +020013882tfo
Lukas Tribus0defb902013-02-13 23:35:39 +010013883 Is an optional keyword which is supported only on Linux kernels >= 3.7. It
Willy Tarreau1c862c52012-10-05 16:21:00 +020013884 enables TCP Fast Open on the listening socket, which means that clients which
13885 support this feature will be able to send a request and receive a response
13886 during the 3-way handshake starting from second connection, thus saving one
13887 round-trip after the first connection. This only makes sense with protocols
13888 that use high connection rates and where each round trip matters. This can
13889 possibly cause issues with many firewalls which do not accept data on SYN
13890 packets, so this option should only be enabled once well tested. This option
Lukas Tribus0999f762013-04-02 16:43:24 +020013891 is only supported on TCPv4/TCPv6 sockets and ignored by other ones. You may
13892 need to build HAProxy with USE_TFO=1 if your libc doesn't define
13893 TCP_FASTOPEN.
Willy Tarreau1c862c52012-10-05 16:21:00 +020013894
Nenad Merdanovic188ad3e2015-02-27 19:56:50 +010013895tls-ticket-keys <keyfile>
13896 Sets the TLS ticket keys file to load the keys from. The keys need to be 48
Emeric Brun9e754772019-01-10 17:51:55 +010013897 or 80 bytes long, depending if aes128 or aes256 is used, encoded with base64
13898 with one line per key (ex. openssl rand 80 | openssl base64 -A | xargs echo).
13899 The first key determines the key length used for next keys: you can't mix
13900 aes128 and aes256 keys. Number of keys is specified by the TLS_TICKETS_NO
13901 build option (default 3) and at least as many keys need to be present in
13902 the file. Last TLS_TICKETS_NO keys will be used for decryption and the
13903 penultimate one for encryption. This enables easy key rotation by just
13904 appending new key to the file and reloading the process. Keys must be
13905 periodically rotated (ex. every 12h) or Perfect Forward Secrecy is
13906 compromised. It is also a good idea to keep the keys off any permanent
Nenad Merdanovic188ad3e2015-02-27 19:56:50 +010013907 storage such as hard drives (hint: use tmpfs and don't swap those files).
13908 Lifetime hint can be changed using tune.ssl.timeout.
13909
Willy Tarreaub6205fd2012-09-24 12:27:33 +020013910transparent
13911 Is an optional keyword which is supported only on certain Linux kernels. It
13912 indicates that the addresses will be bound even if they do not belong to the
13913 local machine, and that packets targeting any of these addresses will be
13914 intercepted just as if the addresses were locally configured. This normally
13915 requires that IP forwarding is enabled. Caution! do not use this with the
13916 default address '*', as it would redirect any traffic for the specified port.
13917 This keyword is available only when HAProxy is built with USE_LINUX_TPROXY=1.
13918 This parameter is only compatible with TCPv4 and TCPv6 sockets, depending on
13919 kernel version. Some distribution kernels include backports of the feature,
13920 so check for support with your vendor.
13921
Willy Tarreau77e3af92012-11-24 15:07:23 +010013922v4v6
13923 Is an optional keyword which is supported only on most recent systems
13924 including Linux kernels >= 2.4.21. It is used to bind a socket to both IPv4
13925 and IPv6 when it uses the default address. Doing so is sometimes necessary
13926 on systems which bind to IPv6 only by default. It has no effect on non-IPv6
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030013927 sockets, and is overridden by the "v6only" option.
Willy Tarreau77e3af92012-11-24 15:07:23 +010013928
Willy Tarreau9b6700f2012-11-24 11:55:28 +010013929v6only
13930 Is an optional keyword which is supported only on most recent systems
13931 including Linux kernels >= 2.4.21. It is used to bind a socket to IPv6 only
13932 when it uses the default address. Doing so is sometimes preferred to doing it
Willy Tarreau77e3af92012-11-24 15:07:23 +010013933 system-wide as it is per-listener. It has no effect on non-IPv6 sockets and
13934 has precedence over the "v4v6" option.
Willy Tarreau9b6700f2012-11-24 11:55:28 +010013935
Willy Tarreaub6205fd2012-09-24 12:27:33 +020013936uid <uid>
13937 Sets the owner of the UNIX sockets to the designated system uid. It can also
13938 be set by default in the global section's "unix-bind" statement. Note that
13939 some platforms simply ignore this. This setting is equivalent to the "user"
13940 setting except that the user numeric ID is used instead of its name. This
13941 setting is ignored by non UNIX sockets.
13942
13943user <user>
13944 Sets the owner of the UNIX sockets to the designated system user. It can also
13945 be set by default in the global section's "unix-bind" statement. Note that
13946 some platforms simply ignore this. This setting is equivalent to the "uid"
13947 setting except that the user name is used instead of its uid. This setting is
13948 ignored by non UNIX sockets.
13949
Emeric Brun1a073b42012-09-28 17:07:34 +020013950verify [none|optional|required]
13951 This setting is only available when support for OpenSSL was built in. If set
13952 to 'none', client certificate is not requested. This is the default. In other
13953 cases, a client certificate is requested. If the client does not provide a
13954 certificate after the request and if 'verify' is set to 'required', then the
13955 handshake is aborted, while it would have succeeded if set to 'optional'. The
Emeric Brunfd33a262012-10-11 16:28:27 +020013956 certificate provided by the client is always verified using CAs from
13957 'ca-file' and optional CRLs from 'crl-file'. On verify failure the handshake
13958 is aborted, regardless of the 'verify' option, unless the error code exactly
13959 matches one of those listed with 'ca-ignore-err' or 'crt-ignore-err'.
Willy Tarreau4a5cade2012-04-05 21:09:48 +020013960
Willy Tarreaub6205fd2012-09-24 12:27:33 +0200139615.2. Server and default-server options
Cyril Bontéf0c60612010-02-06 14:44:47 +010013962------------------------------------
Willy Tarreau6a06a402007-07-15 20:15:28 +020013963
Krzysztof Piotr Oledzkic6df0662010-01-05 16:38:49 +010013964The "server" and "default-server" keywords support a certain number of settings
13965which are all passed as arguments on the server line. The order in which those
13966arguments appear does not count, and they are all optional. Some of those
13967settings are single words (booleans) while others expect one or several values
13968after them. In this case, the values must immediately follow the setting name.
13969Except default-server, all those settings must be specified after the server's
13970address if they are used:
Willy Tarreau6a06a402007-07-15 20:15:28 +020013971
Willy Tarreauc57f0e22009-05-10 13:12:33 +020013972 server <name> <address>[:port] [settings ...]
Krzysztof Piotr Oledzkic6df0662010-01-05 16:38:49 +010013973 default-server [settings ...]
Willy Tarreau6a06a402007-07-15 20:15:28 +020013974
Frédéric Lécailled2376272017-03-21 18:52:12 +010013975Note that all these settings are supported both by "server" and "default-server"
13976keywords, except "id" which is only supported by "server".
13977
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010013978The currently supported settings are the following ones.
Willy Tarreau0ba27502007-12-24 16:55:16 +010013979
Willy Tarreauceb4ac92012-04-28 00:41:46 +020013980addr <ipv4|ipv6>
Willy Tarreauc57f0e22009-05-10 13:12:33 +020013981 Using the "addr" parameter, it becomes possible to use a different IP address
Baptiste Assmann13f83532016-03-06 23:14:36 +010013982 to send health-checks or to probe the agent-check. On some servers, it may be
13983 desirable to dedicate an IP address to specific component able to perform
13984 complex tests which are more suitable to health-checks than the application.
13985 This parameter is ignored if the "check" parameter is not set. See also the
13986 "port" parameter.
Willy Tarreau6a06a402007-07-15 20:15:28 +020013987
Simon Hormand60d6912013-11-25 10:46:36 +090013988agent-check
13989 Enable an auxiliary agent check which is run independently of a regular
Willy Tarreau81f5d942013-12-09 20:51:51 +010013990 health check. An agent health check is performed by making a TCP connection
Willy Tarreau7a0139e2018-12-16 08:42:56 +010013991 to the port set by the "agent-port" parameter and reading an ASCII string
13992 terminated by the first '\r' or '\n' met. The string is made of a series of
13993 words delimited by spaces, tabs or commas in any order, each consisting of :
Simon Hormand60d6912013-11-25 10:46:36 +090013994
Willy Tarreau81f5d942013-12-09 20:51:51 +010013995 - An ASCII representation of a positive integer percentage, e.g. "75%".
Simon Hormand60d6912013-11-25 10:46:36 +090013996 Values in this format will set the weight proportional to the initial
Daniel Corbett9f0843f2021-05-08 10:50:37 -040013997 weight of a server as configured when HAProxy starts. Note that a zero
Willy Tarreauc5af3a62014-10-07 15:27:33 +020013998 weight is reported on the stats page as "DRAIN" since it has the same
13999 effect on the server (it's removed from the LB farm).
Simon Hormand60d6912013-11-25 10:46:36 +090014000
Davor Ocelice9ed2812017-12-25 17:49:28 +010014001 - The string "maxconn:" followed by an integer (no space between). Values
14002 in this format will set the maxconn of a server. The maximum number of
14003 connections advertised needs to be multiplied by the number of load
14004 balancers and different backends that use this health check to get the
14005 total number of connections the server might receive. Example: maxconn:30
Nenad Merdanovic174dd372016-04-24 23:10:06 +020014006
Willy Tarreau81f5d942013-12-09 20:51:51 +010014007 - The word "ready". This will turn the server's administrative state to the
Davor Ocelice9ed2812017-12-25 17:49:28 +010014008 READY mode, thus canceling any DRAIN or MAINT state
Simon Hormand60d6912013-11-25 10:46:36 +090014009
Willy Tarreau81f5d942013-12-09 20:51:51 +010014010 - The word "drain". This will turn the server's administrative state to the
14011 DRAIN mode, thus it will not accept any new connections other than those
14012 that are accepted via persistence.
Simon Hormand60d6912013-11-25 10:46:36 +090014013
Willy Tarreau81f5d942013-12-09 20:51:51 +010014014 - The word "maint". This will turn the server's administrative state to the
14015 MAINT mode, thus it will not accept any new connections at all, and health
14016 checks will be stopped.
Simon Hormand60d6912013-11-25 10:46:36 +090014017
William Dauchyf8e795c2020-09-26 13:35:51 +020014018 - The words "down", "fail", or "stopped", optionally followed by a
Willy Tarreau81f5d942013-12-09 20:51:51 +010014019 description string after a sharp ('#'). All of these mark the server's
14020 operating state as DOWN, but since the word itself is reported on the stats
14021 page, the difference allows an administrator to know if the situation was
14022 expected or not : the service may intentionally be stopped, may appear up
Davor Ocelice9ed2812017-12-25 17:49:28 +010014023 but fail some validity tests, or may be seen as down (e.g. missing process,
Willy Tarreau81f5d942013-12-09 20:51:51 +010014024 or port not responding).
Simon Hormand60d6912013-11-25 10:46:36 +090014025
Willy Tarreau81f5d942013-12-09 20:51:51 +010014026 - The word "up" sets back the server's operating state as UP if health checks
14027 also report that the service is accessible.
Simon Hormand60d6912013-11-25 10:46:36 +090014028
Willy Tarreau81f5d942013-12-09 20:51:51 +010014029 Parameters which are not advertised by the agent are not changed. For
14030 example, an agent might be designed to monitor CPU usage and only report a
14031 relative weight and never interact with the operating status. Similarly, an
14032 agent could be designed as an end-user interface with 3 radio buttons
14033 allowing an administrator to change only the administrative state. However,
14034 it is important to consider that only the agent may revert its own actions,
14035 so if a server is set to DRAIN mode or to DOWN state using the agent, the
14036 agent must implement the other equivalent actions to bring the service into
14037 operations again.
Simon Hormand60d6912013-11-25 10:46:36 +090014038
Simon Horman2f1f9552013-11-25 10:46:37 +090014039 Failure to connect to the agent is not considered an error as connectivity
14040 is tested by the regular health check which is enabled by the "check"
Willy Tarreau81f5d942013-12-09 20:51:51 +010014041 parameter. Warning though, it is not a good idea to stop an agent after it
14042 reports "down", since only an agent reporting "up" will be able to turn the
14043 server up again. Note that the CLI on the Unix stats socket is also able to
Willy Tarreau989222a2016-01-15 10:26:26 +010014044 force an agent's result in order to work around a bogus agent if needed.
Simon Horman2f1f9552013-11-25 10:46:37 +090014045
Willy Tarreau81f5d942013-12-09 20:51:51 +010014046 Requires the "agent-port" parameter to be set. See also the "agent-inter"
Frédéric Lécailled2376272017-03-21 18:52:12 +010014047 and "no-agent-check" parameters.
Simon Hormand60d6912013-11-25 10:46:36 +090014048
James Brown55f9ff12015-10-21 18:19:05 -070014049agent-send <string>
Daniel Corbett9f0843f2021-05-08 10:50:37 -040014050 If this option is specified, HAProxy will send the given string (verbatim)
James Brown55f9ff12015-10-21 18:19:05 -070014051 to the agent server upon connection. You could, for example, encode
14052 the backend name into this string, which would enable your agent to send
14053 different responses based on the backend. Make sure to include a '\n' if
14054 you want to terminate your request with a newline.
14055
Simon Hormand60d6912013-11-25 10:46:36 +090014056agent-inter <delay>
14057 The "agent-inter" parameter sets the interval between two agent checks
14058 to <delay> milliseconds. If left unspecified, the delay defaults to 2000 ms.
14059
14060 Just as with every other time-based parameter, it may be entered in any
14061 other explicit unit among { us, ms, s, m, h, d }. The "agent-inter"
14062 parameter also serves as a timeout for agent checks "timeout check" is
14063 not set. In order to reduce "resonance" effects when multiple servers are
14064 hosted on the same hardware, the agent and health checks of all servers
14065 are started with a small time offset between them. It is also possible to
14066 add some random noise in the agent and health checks interval using the
14067 global "spread-checks" keyword. This makes sense for instance when a lot
14068 of backends use the same servers.
14069
14070 See also the "agent-check" and "agent-port" parameters.
14071
Misiek768d8602017-01-09 09:52:43 +010014072agent-addr <addr>
14073 The "agent-addr" parameter sets address for agent check.
14074
14075 You can offload agent-check to another target, so you can make single place
Daniel Corbett9f0843f2021-05-08 10:50:37 -040014076 managing status and weights of servers defined in HAProxy in case you can't
Misiek768d8602017-01-09 09:52:43 +010014077 make self-aware and self-managing services. You can specify both IP or
14078 hostname, it will be resolved.
14079
Simon Hormand60d6912013-11-25 10:46:36 +090014080agent-port <port>
14081 The "agent-port" parameter sets the TCP port used for agent checks.
14082
14083 See also the "agent-check" and "agent-inter" parameters.
14084
Olivier Houchard8cb2d2e2019-05-06 18:58:48 +020014085allow-0rtt
14086 Allow sending early data to the server when using TLS 1.3.
Olivier Houchard22c9b442019-05-06 19:01:04 +020014087 Note that early data will be sent only if the client used early data, or
14088 if the backend uses "retry-on" with the "0rtt-rejected" keyword.
Olivier Houchard8cb2d2e2019-05-06 18:58:48 +020014089
Olivier Houchardc7566002018-11-20 23:33:50 +010014090alpn <protocols>
14091 This enables the TLS ALPN extension and advertises the specified protocol
14092 list as supported on top of ALPN. The protocol list consists in a comma-
14093 delimited list of protocol names, for instance: "http/1.1,http/1.0" (without
John Roeslerfb2fce12019-07-10 15:45:51 -050014094 quotes). This requires that the SSL library is built with support for TLS
Olivier Houchardc7566002018-11-20 23:33:50 +010014095 extensions enabled (check with haproxy -vv). The ALPN extension replaces the
14096 initial NPN extension. ALPN is required to connect to HTTP/2 servers.
14097 Versions of OpenSSL prior to 1.0.2 didn't support ALPN and only supposed the
14098 now obsolete NPN extension.
14099 If both HTTP/2 and HTTP/1.1 are expected to be supported, both versions can
14100 be advertised, in order of preference, like below :
14101
14102 server 127.0.0.1:443 ssl crt pub.pem alpn h2,http/1.1
14103
Willy Tarreauc57f0e22009-05-10 13:12:33 +020014104backup
14105 When "backup" is present on a server line, the server is only used in load
14106 balancing when all other non-backup servers are unavailable. Requests coming
14107 with a persistence cookie referencing the server will always be served
14108 though. By default, only the first operational backup server is used, unless
Frédéric Lécailled2376272017-03-21 18:52:12 +010014109 the "allbackups" option is set in the backend. See also the "no-backup" and
14110 "allbackups" options.
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010014111
Emeric Brunef42d922012-10-11 16:11:36 +020014112ca-file <cafile>
14113 This setting is only available when support for OpenSSL was built in. It
14114 designates a PEM file from which to load CA certificates used to verify
14115 server's certificate.
14116
Willy Tarreauc57f0e22009-05-10 13:12:33 +020014117check
Jerome Magnin90702bc2020-04-26 14:23:04 +020014118 This option enables health checks on a server:
14119 - when not set, no health checking is performed, and the server is always
14120 considered available.
14121 - when set and no other check method is configured, the server is considered
14122 available when a connection can be established at the highest configured
14123 transport layer. This means TCP by default, or SSL/TLS when "ssl" or
14124 "check-ssl" are set, both possibly combined with connection prefixes such
14125 as a PROXY protocol header when "send-proxy" or "check-send-proxy" are
14126 set.
14127 - when set and an application-level health check is defined, the
14128 application-level exchanges are performed on top of the configured
14129 transport layer and the server is considered available if all of the
14130 exchanges succeed.
14131
14132 By default, health checks are performed on the same address and port as
14133 configured on the server, using the same encapsulation parameters (SSL/TLS,
14134 proxy-protocol header, etc... ). It is possible to change the destination
14135 address using "addr" and the port using "port". When done, it is assumed the
14136 server isn't checked on the service port, and configured encapsulation
Ilya Shipitsin4329a9a2020-05-05 21:17:10 +050014137 parameters are not reused. One must explicitly set "check-send-proxy" to send
Jerome Magnin90702bc2020-04-26 14:23:04 +020014138 connection headers, "check-ssl" to use SSL/TLS.
14139
14140 When "sni" or "alpn" are set on the server line, their value is not used for
14141 health checks and one must use "check-sni" or "check-alpn".
14142
14143 The default source address for health check traffic is the same as the one
14144 defined in the backend. It can be changed with the "source" keyword.
14145
14146 The interval between checks can be set using the "inter" keyword, and the
14147 "rise" and "fall" keywords can be used to define how many successful or
14148 failed health checks are required to flag a server available or not
14149 available.
14150
14151 Optional application-level health checks can be configured with "option
14152 httpchk", "option mysql-check" "option smtpchk", "option pgsql-check",
14153 "option ldap-check", or "option redis-check".
14154
14155 Example:
14156 # simple tcp check
14157 backend foo
14158 server s1 192.168.0.1:80 check
14159 # this does a tcp connect + tls handshake
14160 backend foo
14161 server s1 192.168.0.1:443 ssl check
14162 # simple tcp check is enough for check success
14163 backend foo
14164 option tcp-check
14165 tcp-check connect
14166 server s1 192.168.0.1:443 ssl check
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010014167
Willy Tarreau6c16adc2012-10-05 00:04:16 +020014168check-send-proxy
14169 This option forces emission of a PROXY protocol line with outgoing health
14170 checks, regardless of whether the server uses send-proxy or not for the
14171 normal traffic. By default, the PROXY protocol is enabled for health checks
14172 if it is already enabled for normal traffic and if no "port" nor "addr"
14173 directive is present. However, if such a directive is present, the
14174 "check-send-proxy" option needs to be used to force the use of the
14175 protocol. See also the "send-proxy" option for more information.
14176
Olivier Houchard92150142018-12-21 19:47:01 +010014177check-alpn <protocols>
14178 Defines which protocols to advertise with ALPN. The protocol list consists in
14179 a comma-delimited list of protocol names, for instance: "http/1.1,http/1.0"
14180 (without quotes). If it is not set, the server ALPN is used.
14181
Christopher Fauletedc6ed92020-04-23 16:27:59 +020014182check-proto <name>
14183 Forces the multiplexer's protocol to use for the server's health-check
14184 connections. It must be compatible with the health-check type (TCP or
14185 HTTP). It must also be usable on the backend side. The list of available
Christopher Faulet982e17d2021-03-26 14:44:18 +010014186 protocols is reported in haproxy -vv. The protocols properties are
14187 reported : the mode (TCP/HTTP), the side (FE/BE), the mux name and its flags.
14188
14189 Some protocols report errors on aborts (flag=CLEAN_ABRT). Some others are
14190 subject to the head-of-line blocking on server side (flag=HOL_RISK). Finally
14191 some protocols don't support upgrades (flag=NO_UPG). The HTX compatibility is
14192 also reported (flag=HTX).
14193
14194 Here are the protocols that may be used as argument to a "check-proto"
14195 directive on a server line:
14196
14197 h2 : mode=HTTP side=FE|BE mux=H2 flags=HTX|CLEAN_ABRT|HOL_RISK|NO_UPG
14198 fcgi : mode=HTTP side=BE mux=FCGI flags=HTX|HOL_RISK|NO_UPG
14199 h1 : mode=HTTP side=FE|BE mux=H1 flags=HTX|NO_UPG
14200 none : mode=TCP side=FE|BE mux=PASS flags=NO_UPG
14201
Daniel Corbett67a82712020-07-06 23:01:19 -040014202 Idea behind this option is to bypass the selection of the best multiplexer's
Christopher Fauletedc6ed92020-04-23 16:27:59 +020014203 protocol for health-check connections established to this server.
14204 If not defined, the server one will be used, if set.
14205
Jérôme Magninae9bb762018-12-09 16:08:26 +010014206check-sni <sni>
Olivier Houchard9130a962017-10-17 17:33:43 +020014207 This option allows you to specify the SNI to be used when doing health checks
Jérôme Magninae9bb762018-12-09 16:08:26 +010014208 over SSL. It is only possible to use a string to set <sni>. If you want to
14209 set a SNI for proxied traffic, see "sni".
Olivier Houchard9130a962017-10-17 17:33:43 +020014210
Willy Tarreau763a95b2012-10-04 23:15:39 +020014211check-ssl
14212 This option forces encryption of all health checks over SSL, regardless of
14213 whether the server uses SSL or not for the normal traffic. This is generally
14214 used when an explicit "port" or "addr" directive is specified and SSL health
14215 checks are not inherited. It is important to understand that this option
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030014216 inserts an SSL transport layer below the checks, so that a simple TCP connect
Willy Tarreau763a95b2012-10-04 23:15:39 +020014217 check becomes an SSL connect, which replaces the old ssl-hello-chk. The most
14218 common use is to send HTTPS checks by combining "httpchk" with SSL checks.
Davor Ocelice9ed2812017-12-25 17:49:28 +010014219 All SSL settings are common to health checks and traffic (e.g. ciphers).
Frédéric Lécailled2376272017-03-21 18:52:12 +010014220 See the "ssl" option for more information and "no-check-ssl" to disable
14221 this option.
Willy Tarreau763a95b2012-10-04 23:15:39 +020014222
Alexander Liu2a54bb72019-05-22 19:44:48 +080014223check-via-socks4
John Roeslerfb2fce12019-07-10 15:45:51 -050014224 This option enables outgoing health checks using upstream socks4 proxy. By
Alexander Liu2a54bb72019-05-22 19:44:48 +080014225 default, the health checks won't go through socks tunnel even it was enabled
14226 for normal traffic.
14227
Willy Tarreaua0ee1d02012-09-10 09:01:23 +020014228ciphers <ciphers>
Dirkjan Bussink415150f2018-09-14 11:14:21 +020014229 This setting is only available when support for OpenSSL was built in. This
14230 option sets the string describing the list of cipher algorithms that is
14231 negotiated during the SSL/TLS handshake with the server. The format of the
Bertrand Jacquin4f03ab02019-02-03 18:48:49 +000014232 string is defined in "man 1 ciphers" from OpenSSL man pages. For background
14233 information and recommendations see e.g.
14234 (https://wiki.mozilla.org/Security/Server_Side_TLS) and
14235 (https://mozilla.github.io/server-side-tls/ssl-config-generator/). For TLSv1.3
14236 cipher configuration, please check the "ciphersuites" keyword.
Willy Tarreaua0ee1d02012-09-10 09:01:23 +020014237
Dirkjan Bussink415150f2018-09-14 11:14:21 +020014238ciphersuites <ciphersuites>
14239 This setting is only available when support for OpenSSL was built in and
14240 OpenSSL 1.1.1 or later was used to build HAProxy. This option sets the string
14241 describing the list of cipher algorithms that is negotiated during the TLS
14242 1.3 handshake with the server. The format of the string is defined in
Bertrand Jacquin4f03ab02019-02-03 18:48:49 +000014243 "man 1 ciphers" from OpenSSL man pages under the "ciphersuites" section.
14244 For cipher configuration for TLSv1.2 and earlier, please check the "ciphers"
14245 keyword.
Dirkjan Bussink415150f2018-09-14 11:14:21 +020014246
Willy Tarreauc57f0e22009-05-10 13:12:33 +020014247cookie <value>
14248 The "cookie" parameter sets the cookie value assigned to the server to
14249 <value>. This value will be checked in incoming requests, and the first
14250 operational server possessing the same value will be selected. In return, in
14251 cookie insertion or rewrite modes, this value will be assigned to the cookie
14252 sent to the client. There is nothing wrong in having several servers sharing
14253 the same cookie value, and it is in fact somewhat common between normal and
14254 backup servers. See also the "cookie" keyword in backend section.
14255
Emeric Brunef42d922012-10-11 16:11:36 +020014256crl-file <crlfile>
14257 This setting is only available when support for OpenSSL was built in. It
14258 designates a PEM file from which to load certificate revocation list used
14259 to verify server's certificate.
14260
Emeric Bruna7aa3092012-10-26 12:58:00 +020014261crt <cert>
14262 This setting is only available when support for OpenSSL was built in.
14263 It designates a PEM file from which to load both a certificate and the
14264 associated private key. This file can be built by concatenating both PEM
14265 files into one. This certificate will be sent if the server send a client
14266 certificate request.
14267
Remi Tricot-Le Breton7c980df2021-05-07 15:28:08 +020014268 If the file does not contain a private key, HAProxy will try to load the key
14269 at the same path suffixed by a ".key" (provided the "ssl-load-extra-files"
14270 option is set accordingly).
14271
Willy Tarreau96839092010-03-29 10:02:24 +020014272disabled
14273 The "disabled" keyword starts the server in the "disabled" state. That means
14274 that it is marked down in maintenance mode, and no connection other than the
14275 ones allowed by persist mode will reach it. It is very well suited to setup
14276 new servers, because normal traffic will never reach them, while it is still
14277 possible to test the service by making use of the force-persist mechanism.
Frédéric Lécailled2376272017-03-21 18:52:12 +010014278 See also "enabled" setting.
Willy Tarreau96839092010-03-29 10:02:24 +020014279
Frédéric Lécailled2376272017-03-21 18:52:12 +010014280enabled
14281 This option may be used as 'server' setting to reset any 'disabled'
14282 setting which would have been inherited from 'default-server' directive as
14283 default value.
14284 It may also be used as 'default-server' setting to reset any previous
14285 'default-server' 'disabled' setting.
Willy Tarreau96839092010-03-29 10:02:24 +020014286
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010014287error-limit <count>
Willy Tarreau983e01e2010-01-11 18:42:06 +010014288 If health observing is enabled, the "error-limit" parameter specifies the
14289 number of consecutive errors that triggers event selected by the "on-error"
14290 option. By default it is set to 10 consecutive errors.
Krzysztof Piotr Oledzki97f07b82009-12-15 22:31:24 +010014291
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010014292 See also the "check", "error-limit" and "on-error".
Krzysztof Piotr Oledzki97f07b82009-12-15 22:31:24 +010014293
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010014294fall <count>
Willy Tarreauc57f0e22009-05-10 13:12:33 +020014295 The "fall" parameter states that a server will be considered as dead after
14296 <count> consecutive unsuccessful health checks. This value defaults to 3 if
14297 unspecified. See also the "check", "inter" and "rise" parameters.
14298
Emeric Brun8694b9a2012-10-05 14:39:07 +020014299force-sslv3
14300 This option enforces use of SSLv3 only when SSL is used to communicate with
14301 the server. SSLv3 is generally less expensive than the TLS counterparts for
Emeric Brun2c86cbf2014-10-30 15:56:50 +010014302 high connection rates. This option is also available on global statement
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020014303 "ssl-default-server-options". See also "ssl-min-ver" and ssl-max-ver".
Emeric Brun8694b9a2012-10-05 14:39:07 +020014304
14305force-tlsv10
14306 This option enforces use of TLSv1.0 only when SSL is used to communicate with
Emeric Brun2c86cbf2014-10-30 15:56:50 +010014307 the server. This option is also available on global statement
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020014308 "ssl-default-server-options". See also "ssl-min-ver" and ssl-max-ver".
Emeric Brun8694b9a2012-10-05 14:39:07 +020014309
14310force-tlsv11
14311 This option enforces use of TLSv1.1 only when SSL is used to communicate with
Emeric Brun2c86cbf2014-10-30 15:56:50 +010014312 the server. This option is also available on global statement
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020014313 "ssl-default-server-options". See also "ssl-min-ver" and ssl-max-ver".
Emeric Brun8694b9a2012-10-05 14:39:07 +020014314
14315force-tlsv12
14316 This option enforces use of TLSv1.2 only when SSL is used to communicate with
Emeric Brun2c86cbf2014-10-30 15:56:50 +010014317 the server. This option is also available on global statement
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020014318 "ssl-default-server-options". See also "ssl-min-ver" and ssl-max-ver".
Emeric Brun8694b9a2012-10-05 14:39:07 +020014319
Emmanuel Hocdet42fb9802017-03-30 19:29:39 +020014320force-tlsv13
14321 This option enforces use of TLSv1.3 only when SSL is used to communicate with
14322 the server. This option is also available on global statement
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020014323 "ssl-default-server-options". See also "ssl-min-ver" and ssl-max-ver".
Emmanuel Hocdet42fb9802017-03-30 19:29:39 +020014324
Willy Tarreauc57f0e22009-05-10 13:12:33 +020014325id <value>
Willy Tarreau53fb4ae2009-10-04 23:04:08 +020014326 Set a persistent ID for the server. This ID must be positive and unique for
14327 the proxy. An unused ID will automatically be assigned if unset. The first
14328 assigned value will be 1. This ID is currently only returned in statistics.
Willy Tarreauc57f0e22009-05-10 13:12:33 +020014329
Willy Tarreau6a031d12016-11-07 19:42:35 +010014330init-addr {last | libc | none | <ip>},[...]*
14331 Indicate in what order the server's address should be resolved upon startup
14332 if it uses an FQDN. Attempts are made to resolve the address by applying in
Davor Ocelice9ed2812017-12-25 17:49:28 +010014333 turn each of the methods mentioned in the comma-delimited list. The first
Willy Tarreau6a031d12016-11-07 19:42:35 +010014334 method which succeeds is used. If the end of the list is reached without
14335 finding a working method, an error is thrown. Method "last" suggests to pick
14336 the address which appears in the state file (see "server-state-file"). Method
14337 "libc" uses the libc's internal resolver (gethostbyname() or getaddrinfo()
14338 depending on the operating system and build options). Method "none"
14339 specifically indicates that the server should start without any valid IP
14340 address in a down state. It can be useful to ignore some DNS issues upon
14341 startup, waiting for the situation to get fixed later. Finally, an IP address
14342 (IPv4 or IPv6) may be provided. It can be the currently known address of the
Davor Ocelice9ed2812017-12-25 17:49:28 +010014343 server (e.g. filled by a configuration generator), or the address of a dummy
Willy Tarreau6a031d12016-11-07 19:42:35 +010014344 server used to catch old sessions and present them with a decent error
14345 message for example. When the "first" load balancing algorithm is used, this
14346 IP address could point to a fake server used to trigger the creation of new
14347 instances on the fly. This option defaults to "last,libc" indicating that the
14348 previous address found in the state file (if any) is used first, otherwise
14349 the libc's resolver is used. This ensures continued compatibility with the
Davor Ocelice9ed2812017-12-25 17:49:28 +010014350 historic behavior.
Willy Tarreau6a031d12016-11-07 19:42:35 +010014351
14352 Example:
14353 defaults
14354 # never fail on address resolution
14355 default-server init-addr last,libc,none
14356
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010014357inter <delay>
14358fastinter <delay>
14359downinter <delay>
Willy Tarreauc57f0e22009-05-10 13:12:33 +020014360 The "inter" parameter sets the interval between two consecutive health checks
14361 to <delay> milliseconds. If left unspecified, the delay defaults to 2000 ms.
14362 It is also possible to use "fastinter" and "downinter" to optimize delays
14363 between checks depending on the server state :
14364
Pieter Baauw44fc9df2015-09-17 21:30:46 +020014365 Server state | Interval used
14366 ----------------------------------------+----------------------------------
14367 UP 100% (non-transitional) | "inter"
14368 ----------------------------------------+----------------------------------
14369 Transitionally UP (going down "fall"), | "fastinter" if set,
14370 Transitionally DOWN (going up "rise"), | "inter" otherwise.
14371 or yet unchecked. |
14372 ----------------------------------------+----------------------------------
14373 DOWN 100% (non-transitional) | "downinter" if set,
14374 | "inter" otherwise.
14375 ----------------------------------------+----------------------------------
Willy Tarreaud72758d2010-01-12 10:42:19 +010014376
Willy Tarreauc57f0e22009-05-10 13:12:33 +020014377 Just as with every other time-based parameter, they can be entered in any
14378 other explicit unit among { us, ms, s, m, h, d }. The "inter" parameter also
14379 serves as a timeout for health checks sent to servers if "timeout check" is
14380 not set. In order to reduce "resonance" effects when multiple servers are
Simon Hormand60d6912013-11-25 10:46:36 +090014381 hosted on the same hardware, the agent and health checks of all servers
14382 are started with a small time offset between them. It is also possible to
14383 add some random noise in the agent and health checks interval using the
14384 global "spread-checks" keyword. This makes sense for instance when a lot
14385 of backends use the same servers.
Willy Tarreauc57f0e22009-05-10 13:12:33 +020014386
Emeric Brun97556472020-05-30 01:42:45 +020014387log-proto <logproto>
14388 The "log-proto" specifies the protocol used to forward event messages to
14389 a server configured in a ring section. Possible values are "legacy"
14390 and "octet-count" corresponding respectively to "Non-transparent-framing"
14391 and "Octet counting" in rfc6587. "legacy" is the default.
14392
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010014393maxconn <maxconn>
Willy Tarreauc57f0e22009-05-10 13:12:33 +020014394 The "maxconn" parameter specifies the maximal number of concurrent
14395 connections that will be sent to this server. If the number of incoming
Tim Duesterhuscefbbd92019-11-27 22:35:27 +010014396 concurrent connections goes higher than this value, they will be queued,
14397 waiting for a slot to be released. This parameter is very important as it can
Willy Tarreauc57f0e22009-05-10 13:12:33 +020014398 save fragile servers from going down under extreme loads. If a "minconn"
14399 parameter is specified, the limit becomes dynamic. The default value is "0"
14400 which means unlimited. See also the "minconn" and "maxqueue" parameters, and
14401 the backend's "fullconn" keyword.
14402
Tim Duesterhuscefbbd92019-11-27 22:35:27 +010014403 In HTTP mode this parameter limits the number of concurrent requests instead
14404 of the number of connections. Multiple requests might be multiplexed over a
14405 single TCP connection to the server. As an example if you specify a maxconn
14406 of 50 you might see between 1 and 50 actual server connections, but no more
14407 than 50 concurrent requests.
14408
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010014409maxqueue <maxqueue>
Willy Tarreauc57f0e22009-05-10 13:12:33 +020014410 The "maxqueue" parameter specifies the maximal number of connections which
14411 will wait in the queue for this server. If this limit is reached, next
14412 requests will be redispatched to other servers instead of indefinitely
14413 waiting to be served. This will break persistence but may allow people to
Willy Tarreau8ae8c482020-10-22 17:19:07 +020014414 quickly re-log in when the server they try to connect to is dying. Some load
14415 balancing algorithms such as leastconn take this into account and accept to
14416 add requests into a server's queue up to this value if it is explicitly set
14417 to a value greater than zero, which often allows to better smooth the load
14418 when dealing with single-digit maxconn values. The default value is "0" which
14419 means the queue is unlimited. See also the "maxconn" and "minconn" parameters
14420 and "balance leastconn".
Willy Tarreauc57f0e22009-05-10 13:12:33 +020014421
Willy Tarreau9c538e02019-01-23 10:21:49 +010014422max-reuse <count>
14423 The "max-reuse" argument indicates the HTTP connection processors that they
14424 should not reuse a server connection more than this number of times to send
14425 new requests. Permitted values are -1 (the default), which disables this
14426 limit, or any positive value. Value zero will effectively disable keep-alive.
14427 This is only used to work around certain server bugs which cause them to leak
14428 resources over time. The argument is not necessarily respected by the lower
14429 layers as there might be technical limitations making it impossible to
14430 enforce. At least HTTP/2 connections to servers will respect it.
14431
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010014432minconn <minconn>
Willy Tarreauc57f0e22009-05-10 13:12:33 +020014433 When the "minconn" parameter is set, the maxconn limit becomes a dynamic
14434 limit following the backend's load. The server will always accept at least
14435 <minconn> connections, never more than <maxconn>, and the limit will be on
14436 the ramp between both values when the backend has less than <fullconn>
14437 concurrent connections. This makes it possible to limit the load on the
14438 server during normal loads, but push it further for important loads without
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +010014439 overloading the server during exceptional loads. See also the "maxconn"
Willy Tarreauc57f0e22009-05-10 13:12:33 +020014440 and "maxqueue" parameters, as well as the "fullconn" backend keyword.
Krzysztof Piotr Oledzki97f07b82009-12-15 22:31:24 +010014441
Willy Tarreaud72f0f32015-10-13 14:50:22 +020014442namespace <name>
14443 On Linux, it is possible to specify which network namespace a socket will
14444 belong to. This directive makes it possible to explicitly bind a server to
14445 a namespace different from the default one. Please refer to your operating
14446 system's documentation to find more details about network namespaces.
14447
Frédéric Lécailled2376272017-03-21 18:52:12 +010014448no-agent-check
14449 This option may be used as "server" setting to reset any "agent-check"
14450 setting which would have been inherited from "default-server" directive as
14451 default value.
14452 It may also be used as "default-server" setting to reset any previous
14453 "default-server" "agent-check" setting.
14454
14455no-backup
14456 This option may be used as "server" setting to reset any "backup"
14457 setting which would have been inherited from "default-server" directive as
14458 default value.
14459 It may also be used as "default-server" setting to reset any previous
14460 "default-server" "backup" setting.
14461
14462no-check
14463 This option may be used as "server" setting to reset any "check"
14464 setting which would have been inherited from "default-server" directive as
14465 default value.
14466 It may also be used as "default-server" setting to reset any previous
14467 "default-server" "check" setting.
14468
14469no-check-ssl
14470 This option may be used as "server" setting to reset any "check-ssl"
14471 setting which would have been inherited from "default-server" directive as
14472 default value.
14473 It may also be used as "default-server" setting to reset any previous
14474 "default-server" "check-ssl" setting.
14475
Frédéric Lécailled2376272017-03-21 18:52:12 +010014476no-send-proxy
14477 This option may be used as "server" setting to reset any "send-proxy"
14478 setting which would have been inherited from "default-server" directive as
14479 default value.
14480 It may also be used as "default-server" setting to reset any previous
14481 "default-server" "send-proxy" setting.
14482
14483no-send-proxy-v2
14484 This option may be used as "server" setting to reset any "send-proxy-v2"
14485 setting which would have been inherited from "default-server" directive as
14486 default value.
14487 It may also be used as "default-server" setting to reset any previous
14488 "default-server" "send-proxy-v2" setting.
14489
14490no-send-proxy-v2-ssl
14491 This option may be used as "server" setting to reset any "send-proxy-v2-ssl"
14492 setting which would have been inherited from "default-server" directive as
14493 default value.
14494 It may also be used as "default-server" setting to reset any previous
14495 "default-server" "send-proxy-v2-ssl" setting.
14496
14497no-send-proxy-v2-ssl-cn
14498 This option may be used as "server" setting to reset any "send-proxy-v2-ssl-cn"
14499 setting which would have been inherited from "default-server" directive as
14500 default value.
14501 It may also be used as "default-server" setting to reset any previous
14502 "default-server" "send-proxy-v2-ssl-cn" setting.
14503
14504no-ssl
14505 This option may be used as "server" setting to reset any "ssl"
14506 setting which would have been inherited from "default-server" directive as
14507 default value.
14508 It may also be used as "default-server" setting to reset any previous
14509 "default-server" "ssl" setting.
14510
William Dauchyf6370442020-11-14 19:25:33 +010014511 Note that using `default-server ssl` setting and `no-ssl` on server will
14512 however init SSL connection, so it can be later be enabled through the
14513 runtime API: see `set server` commands in management doc.
14514
Willy Tarreau2a3fb1c2015-02-05 16:47:07 +010014515no-ssl-reuse
14516 This option disables SSL session reuse when SSL is used to communicate with
14517 the server. It will force the server to perform a full handshake for every
14518 new connection. It's probably only useful for benchmarking, troubleshooting,
14519 and for paranoid users.
14520
Emeric Brun9b3009b2012-10-05 11:55:06 +020014521no-sslv3
Willy Tarreaua0ee1d02012-09-10 09:01:23 +020014522 This option disables support for SSLv3 when SSL is used to communicate with
14523 the server. Note that SSLv2 is disabled in the code and cannot be enabled
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020014524 using any configuration option. Use "ssl-min-ver" and "ssl-max-ver" instead.
Willy Tarreaua0ee1d02012-09-10 09:01:23 +020014525
Emmanuel Hocdet6cb2d1e2017-03-30 14:43:31 +020014526 Supported in default-server: No
14527
Emeric Brunf9c5c472012-10-11 15:28:34 +020014528no-tls-tickets
14529 This setting is only available when support for OpenSSL was built in. It
14530 disables the stateless session resumption (RFC 5077 TLS Ticket
14531 extension) and force to use stateful session resumption. Stateless
Emeric Brun2c86cbf2014-10-30 15:56:50 +010014532 session resumption is more expensive in CPU usage for servers. This option
14533 is also available on global statement "ssl-default-server-options".
Lukas Tribusbdb386d2020-03-10 00:56:09 +010014534 The TLS ticket mechanism is only used up to TLS 1.2.
14535 Forward Secrecy is compromised with TLS tickets, unless ticket keys
14536 are periodically rotated (via reload or by using "tls-ticket-keys").
Frédéric Lécailled2376272017-03-21 18:52:12 +010014537 See also "tls-tickets".
Emeric Brunf9c5c472012-10-11 15:28:34 +020014538
Emeric Brun9b3009b2012-10-05 11:55:06 +020014539no-tlsv10
Emeric Brun8694b9a2012-10-05 14:39:07 +020014540 This option disables support for TLSv1.0 when SSL is used to communicate with
Emeric Brunf5da4932012-09-28 19:42:54 +020014541 the server. Note that SSLv2 is disabled in the code and cannot be enabled
14542 using any configuration option. TLSv1 is more expensive than SSLv3 so it
Emeric Brun2c86cbf2014-10-30 15:56:50 +010014543 often makes sense to disable it when communicating with local servers. This
14544 option is also available on global statement "ssl-default-server-options".
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020014545 Use "ssl-min-ver" and "ssl-max-ver" instead.
Willy Tarreau763a95b2012-10-04 23:15:39 +020014546
Emmanuel Hocdet6cb2d1e2017-03-30 14:43:31 +020014547 Supported in default-server: No
14548
Emeric Brun9b3009b2012-10-05 11:55:06 +020014549no-tlsv11
Emeric Brun8694b9a2012-10-05 14:39:07 +020014550 This option disables support for TLSv1.1 when SSL is used to communicate with
Emeric Brunf5da4932012-09-28 19:42:54 +020014551 the server. Note that SSLv2 is disabled in the code and cannot be enabled
14552 using any configuration option. TLSv1 is more expensive than SSLv3 so it
Emeric Brun2c86cbf2014-10-30 15:56:50 +010014553 often makes sense to disable it when communicating with local servers. This
14554 option is also available on global statement "ssl-default-server-options".
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020014555 Use "ssl-min-ver" and "ssl-max-ver" instead.
Willy Tarreau763a95b2012-10-04 23:15:39 +020014556
Emmanuel Hocdet6cb2d1e2017-03-30 14:43:31 +020014557 Supported in default-server: No
14558
Emeric Brun9b3009b2012-10-05 11:55:06 +020014559no-tlsv12
Emeric Brun8694b9a2012-10-05 14:39:07 +020014560 This option disables support for TLSv1.2 when SSL is used to communicate with
Willy Tarreaua0ee1d02012-09-10 09:01:23 +020014561 the server. Note that SSLv2 is disabled in the code and cannot be enabled
14562 using any configuration option. TLSv1 is more expensive than SSLv3 so it
Emeric Brun2c86cbf2014-10-30 15:56:50 +010014563 often makes sense to disable it when communicating with local servers. This
14564 option is also available on global statement "ssl-default-server-options".
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020014565 Use "ssl-min-ver" and "ssl-max-ver" instead.
Emmanuel Hocdet42fb9802017-03-30 19:29:39 +020014566
14567 Supported in default-server: No
14568
14569no-tlsv13
14570 This option disables support for TLSv1.3 when SSL is used to communicate with
14571 the server. Note that SSLv2 is disabled in the code and cannot be enabled
14572 using any configuration option. TLSv1 is more expensive than SSLv3 so it
14573 often makes sense to disable it when communicating with local servers. This
14574 option is also available on global statement "ssl-default-server-options".
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020014575 Use "ssl-min-ver" and "ssl-max-ver" instead.
Willy Tarreaua0ee1d02012-09-10 09:01:23 +020014576
Emmanuel Hocdet6cb2d1e2017-03-30 14:43:31 +020014577 Supported in default-server: No
14578
Frédéric Lécailled2376272017-03-21 18:52:12 +010014579no-verifyhost
14580 This option may be used as "server" setting to reset any "verifyhost"
14581 setting which would have been inherited from "default-server" directive as
14582 default value.
14583 It may also be used as "default-server" setting to reset any previous
14584 "default-server" "verifyhost" setting.
Willy Tarreau763a95b2012-10-04 23:15:39 +020014585
Frédéric Lécaille1b9423d2019-07-04 14:19:06 +020014586no-tfo
14587 This option may be used as "server" setting to reset any "tfo"
14588 setting which would have been inherited from "default-server" directive as
14589 default value.
14590 It may also be used as "default-server" setting to reset any previous
14591 "default-server" "tfo" setting.
14592
Simon Hormanfa461682011-06-25 09:39:49 +090014593non-stick
14594 Never add connections allocated to this sever to a stick-table.
14595 This may be used in conjunction with backup to ensure that
14596 stick-table persistence is disabled for backup servers.
14597
Olivier Houchardc7566002018-11-20 23:33:50 +010014598npn <protocols>
14599 This enables the NPN TLS extension and advertises the specified protocol list
14600 as supported on top of NPN. The protocol list consists in a comma-delimited
14601 list of protocol names, for instance: "http/1.1,http/1.0" (without quotes).
John Roeslerfb2fce12019-07-10 15:45:51 -050014602 This requires that the SSL library is built with support for TLS extensions
Olivier Houchardc7566002018-11-20 23:33:50 +010014603 enabled (check with haproxy -vv). Note that the NPN extension has been
14604 replaced with the ALPN extension (see the "alpn" keyword), though this one is
14605 only available starting with OpenSSL 1.0.2.
14606
Krzysztof Piotr Oledzki97f07b82009-12-15 22:31:24 +010014607observe <mode>
14608 This option enables health adjusting based on observing communication with
14609 the server. By default this functionality is disabled and enabling it also
14610 requires to enable health checks. There are two supported modes: "layer4" and
14611 "layer7". In layer4 mode, only successful/unsuccessful tcp connections are
14612 significant. In layer7, which is only allowed for http proxies, responses
14613 received from server are verified, like valid/wrong http code, unparsable
Willy Tarreau150d1462012-03-10 08:19:02 +010014614 headers, a timeout, etc. Valid status codes include 100 to 499, 501 and 505.
Krzysztof Piotr Oledzki97f07b82009-12-15 22:31:24 +010014615
14616 See also the "check", "on-error" and "error-limit".
14617
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010014618on-error <mode>
Krzysztof Piotr Oledzki97f07b82009-12-15 22:31:24 +010014619 Select what should happen when enough consecutive errors are detected.
14620 Currently, four modes are available:
14621 - fastinter: force fastinter
14622 - fail-check: simulate a failed check, also forces fastinter (default)
14623 - sudden-death: simulate a pre-fatal failed health check, one more failed
14624 check will mark a server down, forces fastinter
14625 - mark-down: mark the server immediately down and force fastinter
14626
14627 See also the "check", "observe" and "error-limit".
14628
Simon Hormane0d1bfb2011-06-21 14:34:58 +090014629on-marked-down <action>
14630 Modify what occurs when a server is marked down.
14631 Currently one action is available:
Justin Karnegeseb2c24a2012-05-24 15:28:52 -070014632 - shutdown-sessions: Shutdown peer sessions. When this setting is enabled,
14633 all connections to the server are immediately terminated when the server
14634 goes down. It might be used if the health check detects more complex cases
14635 than a simple connection status, and long timeouts would cause the service
14636 to remain unresponsive for too long a time. For instance, a health check
14637 might detect that a database is stuck and that there's no chance to reuse
14638 existing connections anymore. Connections killed this way are logged with
14639 a 'D' termination code (for "Down").
Simon Hormane0d1bfb2011-06-21 14:34:58 +090014640
14641 Actions are disabled by default
14642
Justin Karnegeseb2c24a2012-05-24 15:28:52 -070014643on-marked-up <action>
14644 Modify what occurs when a server is marked up.
14645 Currently one action is available:
14646 - shutdown-backup-sessions: Shutdown sessions on all backup servers. This is
14647 done only if the server is not in backup state and if it is not disabled
14648 (it must have an effective weight > 0). This can be used sometimes to force
14649 an active server to take all the traffic back after recovery when dealing
Davor Ocelice9ed2812017-12-25 17:49:28 +010014650 with long sessions (e.g. LDAP, SQL, ...). Doing this can cause more trouble
14651 than it tries to solve (e.g. incomplete transactions), so use this feature
Justin Karnegeseb2c24a2012-05-24 15:28:52 -070014652 with extreme care. Sessions killed because a server comes up are logged
14653 with an 'U' termination code (for "Up").
14654
14655 Actions are disabled by default
14656
Willy Tarreau2f3f4d32020-07-01 07:43:51 +020014657pool-low-conn <max>
14658 Set a low threshold on the number of idling connections for a server, below
14659 which a thread will not try to steal a connection from another thread. This
14660 can be useful to improve CPU usage patterns in scenarios involving many very
14661 fast servers, in order to ensure all threads will keep a few idle connections
14662 all the time instead of letting them accumulate over one thread and migrating
14663 them from thread to thread. Typical values of twice the number of threads
14664 seem to show very good performance already with sub-millisecond response
14665 times. The default is zero, indicating that any idle connection can be used
14666 at any time. It is the recommended setting for normal use. This only applies
14667 to connections that can be shared according to the same principles as those
Willy Tarreau0784db82021-02-19 11:45:22 +010014668 applying to "http-reuse". In case connection sharing between threads would
14669 be disabled via "tune.idle-pool.shared", it can become very important to use
14670 this setting to make sure each thread always has a few connections, or the
14671 connection reuse rate will decrease as thread count increases.
Willy Tarreau2f3f4d32020-07-01 07:43:51 +020014672
Olivier Houchard006e3102018-12-10 18:30:32 +010014673pool-max-conn <max>
14674 Set the maximum number of idling connections for a server. -1 means unlimited
14675 connections, 0 means no idle connections. The default is -1. When idle
14676 connections are enabled, orphaned idle connections which do not belong to any
14677 client session anymore are moved to a dedicated pool so that they remain
14678 usable by future clients. This only applies to connections that can be shared
14679 according to the same principles as those applying to "http-reuse".
14680
Olivier Houchardb7b3faa2018-12-14 18:15:36 +010014681pool-purge-delay <delay>
14682 Sets the delay to start purging idle connections. Each <delay> interval, half
Olivier Houcharda56eebf2019-03-19 16:44:02 +010014683 of the idle connections are closed. 0 means we don't keep any idle connection.
Willy Tarreaufb553652019-06-04 14:06:31 +020014684 The default is 5s.
Olivier Houchardb7b3faa2018-12-14 18:15:36 +010014685
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010014686port <port>
Willy Tarreauc57f0e22009-05-10 13:12:33 +020014687 Using the "port" parameter, it becomes possible to use a different port to
William Dauchy4858fb22021-02-03 22:30:09 +010014688 send health-checks or to probe the agent-check. On some servers, it may be
14689 desirable to dedicate a port to a specific component able to perform complex
14690 tests which are more suitable to health-checks than the application. It is
14691 common to run a simple script in inetd for instance. This parameter is
14692 ignored if the "check" parameter is not set. See also the "addr" parameter.
Willy Tarreauc57f0e22009-05-10 13:12:33 +020014693
Christopher Faulet8ed0a3e2018-04-10 14:45:45 +020014694proto <name>
Christopher Faulet8ed0a3e2018-04-10 14:45:45 +020014695 Forces the multiplexer's protocol to use for the outgoing connections to this
14696 server. It must be compatible with the mode of the backend (TCP or HTTP). It
14697 must also be usable on the backend side. The list of available protocols is
Christopher Faulet982e17d2021-03-26 14:44:18 +010014698 reported in haproxy -vv.The protocols properties are reported : the mode
14699 (TCP/HTTP), the side (FE/BE), the mux name and its flags.
14700
14701 Some protocols report errors on aborts (flag=CLEAN_ABRT). Some others are
14702 subject to the head-of-line blocking on server side (flag=HOL_RISK). Finally
14703 some protocols don't support upgrades (flag=NO_UPG). The HTX compatibility is
14704 also reported (flag=HTX).
14705
14706 Here are the protocols that may be used as argument to a "proto" directive on
14707 a server line :
14708
14709 h2 : mode=HTTP side=FE|BE mux=H2 flags=HTX|CLEAN_ABRT|HOL_RISK|NO_UPG
14710 fcgi : mode=HTTP side=BE mux=FCGI flags=HTX|HOL_RISK|NO_UPG
14711 h1 : mode=HTTP side=FE|BE mux=H1 flags=HTX|NO_UPG
14712 none : mode=TCP side=FE|BE mux=PASS flags=NO_UPG
14713
Daniel Corbett67a82712020-07-06 23:01:19 -040014714 Idea behind this option is to bypass the selection of the best multiplexer's
Christopher Faulet8ed0a3e2018-04-10 14:45:45 +020014715 protocol for all connections established to this server.
14716
Willy Tarreauc57f0e22009-05-10 13:12:33 +020014717redir <prefix>
14718 The "redir" parameter enables the redirection mode for all GET and HEAD
14719 requests addressing this server. This means that instead of having HAProxy
14720 forward the request to the server, it will send an "HTTP 302" response with
14721 the "Location" header composed of this prefix immediately followed by the
14722 requested URI beginning at the leading '/' of the path component. That means
14723 that no trailing slash should be used after <prefix>. All invalid requests
14724 will be rejected, and all non-GET or HEAD requests will be normally served by
14725 the server. Note that since the response is completely forged, no header
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +010014726 mangling nor cookie insertion is possible in the response. However, cookies in
Davor Ocelice9ed2812017-12-25 17:49:28 +010014727 requests are still analyzed, making this solution completely usable to direct
Willy Tarreauc57f0e22009-05-10 13:12:33 +020014728 users to a remote location in case of local disaster. Main use consists in
14729 increasing bandwidth for static servers by having the clients directly
14730 connect to them. Note: never use a relative location here, it would cause a
14731 loop between the client and HAProxy!
14732
14733 Example : server srv1 192.168.1.1:80 redir http://image1.mydomain.com check
14734
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010014735rise <count>
Willy Tarreauc57f0e22009-05-10 13:12:33 +020014736 The "rise" parameter states that a server will be considered as operational
14737 after <count> consecutive successful health checks. This value defaults to 2
14738 if unspecified. See also the "check", "inter" and "fall" parameters.
14739
Baptiste Assmann8e2d9432018-06-22 15:04:43 +020014740resolve-opts <option>,<option>,...
14741 Comma separated list of options to apply to DNS resolution linked to this
14742 server.
14743
14744 Available options:
14745
14746 * allow-dup-ip
14747 By default, HAProxy prevents IP address duplication in a backend when DNS
14748 resolution at runtime is in operation.
14749 That said, for some cases, it makes sense that two servers (in the same
14750 backend, being resolved by the same FQDN) have the same IP address.
14751 For such case, simply enable this option.
14752 This is the opposite of prevent-dup-ip.
14753
Daniel Corbettf8716912019-11-17 09:48:56 -050014754 * ignore-weight
14755 Ignore any weight that is set within an SRV record. This is useful when
14756 you would like to control the weights using an alternate method, such as
14757 using an "agent-check" or through the runtime api.
14758
Baptiste Assmann8e2d9432018-06-22 15:04:43 +020014759 * prevent-dup-ip
14760 Ensure HAProxy's default behavior is enforced on a server: prevent re-using
14761 an IP address already set to a server in the same backend and sharing the
14762 same fqdn.
14763 This is the opposite of allow-dup-ip.
14764
14765 Example:
14766 backend b_myapp
14767 default-server init-addr none resolvers dns
14768 server s1 myapp.example.com:80 check resolve-opts allow-dup-ip
14769 server s2 myapp.example.com:81 check resolve-opts allow-dup-ip
14770
14771 With the option allow-dup-ip set:
14772 * if the nameserver returns a single IP address, then both servers will use
14773 it
14774 * If the nameserver returns 2 IP addresses, then each server will pick up a
14775 different address
14776
14777 Default value: not set
14778
Baptiste Assmann1fa66662015-04-14 00:28:47 +020014779resolve-prefer <family>
14780 When DNS resolution is enabled for a server and multiple IP addresses from
14781 different families are returned, HAProxy will prefer using an IP address
14782 from the family mentioned in the "resolve-prefer" parameter.
14783 Available families: "ipv4" and "ipv6"
14784
Baptiste Assmannc4aabae2015-08-04 22:43:06 +020014785 Default value: ipv6
14786
Olivier Doucetaa1ea8a2016-08-05 17:15:20 +020014787 Example:
14788
14789 server s1 app1.domain.com:80 resolvers mydns resolve-prefer ipv6
Baptiste Assmann1fa66662015-04-14 00:28:47 +020014790
Thierry Fournierac88cfe2016-02-17 22:05:30 +010014791resolve-net <network>[,<network[,...]]
John Roeslerfb2fce12019-07-10 15:45:51 -050014792 This option prioritizes the choice of an ip address matching a network. This is
Thierry Fournierac88cfe2016-02-17 22:05:30 +010014793 useful with clouds to prefer a local ip. In some cases, a cloud high
Tim Düsterhus4896c442016-11-29 02:15:19 +010014794 availability service can be announced with many ip addresses on many
Davor Ocelice9ed2812017-12-25 17:49:28 +010014795 different datacenters. The latency between datacenter is not negligible, so
14796 this patch permits to prefer a local datacenter. If no address matches the
Thierry Fournierac88cfe2016-02-17 22:05:30 +010014797 configured network, another address is selected.
14798
Olivier Doucetaa1ea8a2016-08-05 17:15:20 +020014799 Example:
14800
14801 server s1 app1.domain.com:80 resolvers mydns resolve-net 10.0.0.0/8
Thierry Fournierac88cfe2016-02-17 22:05:30 +010014802
Baptiste Assmann1fa66662015-04-14 00:28:47 +020014803resolvers <id>
14804 Points to an existing "resolvers" section to resolve current server's
14805 hostname.
14806
Olivier Doucetaa1ea8a2016-08-05 17:15:20 +020014807 Example:
14808
14809 server s1 app1.domain.com:80 check resolvers mydns
Baptiste Assmann1fa66662015-04-14 00:28:47 +020014810
Olivier Doucetaa1ea8a2016-08-05 17:15:20 +020014811 See also section 5.3
Baptiste Assmann1fa66662015-04-14 00:28:47 +020014812
Willy Tarreau5ab04ec2011-03-20 10:32:26 +010014813send-proxy
14814 The "send-proxy" parameter enforces use of the PROXY protocol over any
14815 connection established to this server. The PROXY protocol informs the other
14816 end about the layer 3/4 addresses of the incoming connection, so that it can
14817 know the client's address or the public address it accessed to, whatever the
Bertrand Jacquin93b227d2016-06-04 15:11:10 +010014818 upper layer protocol. For connections accepted by an "accept-proxy" or
14819 "accept-netscaler-cip" listener, the advertised address will be used. Only
14820 TCPv4 and TCPv6 address families are supported. Other families such as
14821 Unix sockets, will report an UNKNOWN family. Servers using this option can
Daniel Corbett9f0843f2021-05-08 10:50:37 -040014822 fully be chained to another instance of HAProxy listening with an
Bertrand Jacquin93b227d2016-06-04 15:11:10 +010014823 "accept-proxy" setting. This setting must not be used if the server isn't
14824 aware of the protocol. When health checks are sent to the server, the PROXY
14825 protocol is automatically used when this option is set, unless there is an
14826 explicit "port" or "addr" directive, in which case an explicit
14827 "check-send-proxy" directive would also be needed to use the PROXY protocol.
Frédéric Lécailled2376272017-03-21 18:52:12 +010014828 See also the "no-send-proxy" option of this section and "accept-proxy" and
14829 "accept-netscaler-cip" option of the "bind" keyword.
Willy Tarreau5ab04ec2011-03-20 10:32:26 +010014830
David Safb76832014-05-08 23:42:08 -040014831send-proxy-v2
14832 The "send-proxy-v2" parameter enforces use of the PROXY protocol version 2
14833 over any connection established to this server. The PROXY protocol informs
14834 the other end about the layer 3/4 addresses of the incoming connection, so
14835 that it can know the client's address or the public address it accessed to,
Emmanuel Hocdet404d9782017-10-24 10:55:14 +020014836 whatever the upper layer protocol. It also send ALPN information if an alpn
14837 have been negotiated. This setting must not be used if the server isn't aware
14838 of this version of the protocol. See also the "no-send-proxy-v2" option of
14839 this section and send-proxy" option of the "bind" keyword.
David Safb76832014-05-08 23:42:08 -040014840
Emmanuel Hocdetf643b802018-02-01 15:20:32 +010014841proxy-v2-options <option>[,<option>]*
Tim Duesterhuscf6e0c82020-03-13 12:34:24 +010014842 The "proxy-v2-options" parameter add options to send in PROXY protocol
14843 version 2 when "send-proxy-v2" is used. Options available are:
14844
14845 - ssl : See also "send-proxy-v2-ssl".
14846 - cert-cn : See also "send-proxy-v2-ssl-cn".
14847 - ssl-cipher: Name of the used cipher.
14848 - cert-sig : Signature algorithm of the used certificate.
14849 - cert-key : Key algorithm of the used certificate
14850 - authority : Host name value passed by the client (only SNI from a TLS
14851 connection is supported).
14852 - crc32c : Checksum of the PROXYv2 header.
14853 - unique-id : Send a unique ID generated using the frontend's
14854 "unique-id-format" within the PROXYv2 header.
14855 This unique-id is primarily meant for "mode tcp". It can
14856 lead to unexpected results in "mode http", because the
14857 generated unique ID is also used for the first HTTP request
14858 within a Keep-Alive connection.
Emmanuel Hocdetf643b802018-02-01 15:20:32 +010014859
David Safb76832014-05-08 23:42:08 -040014860send-proxy-v2-ssl
14861 The "send-proxy-v2-ssl" parameter enforces use of the PROXY protocol version
14862 2 over any connection established to this server. The PROXY protocol informs
14863 the other end about the layer 3/4 addresses of the incoming connection, so
14864 that it can know the client's address or the public address it accessed to,
14865 whatever the upper layer protocol. In addition, the SSL information extension
14866 of the PROXY protocol is added to the PROXY protocol header. This setting
14867 must not be used if the server isn't aware of this version of the protocol.
Frédéric Lécailled2376272017-03-21 18:52:12 +010014868 See also the "no-send-proxy-v2-ssl" option of this section and the
14869 "send-proxy-v2" option of the "bind" keyword.
David Safb76832014-05-08 23:42:08 -040014870
14871send-proxy-v2-ssl-cn
14872 The "send-proxy-v2-ssl" parameter enforces use of the PROXY protocol version
14873 2 over any connection established to this server. The PROXY protocol informs
14874 the other end about the layer 3/4 addresses of the incoming connection, so
14875 that it can know the client's address or the public address it accessed to,
14876 whatever the upper layer protocol. In addition, the SSL information extension
14877 of the PROXY protocol, along along with the Common Name from the subject of
14878 the client certificate (if any), is added to the PROXY protocol header. This
14879 setting must not be used if the server isn't aware of this version of the
Davor Ocelice9ed2812017-12-25 17:49:28 +010014880 protocol. See also the "no-send-proxy-v2-ssl-cn" option of this section and
14881 the "send-proxy-v2" option of the "bind" keyword.
David Safb76832014-05-08 23:42:08 -040014882
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010014883slowstart <start_time_in_ms>
Willy Tarreauc57f0e22009-05-10 13:12:33 +020014884 The "slowstart" parameter for a server accepts a value in milliseconds which
14885 indicates after how long a server which has just come back up will run at
14886 full speed. Just as with every other time-based parameter, it can be entered
14887 in any other explicit unit among { us, ms, s, m, h, d }. The speed grows
14888 linearly from 0 to 100% during this time. The limitation applies to two
14889 parameters :
14890
14891 - maxconn: the number of connections accepted by the server will grow from 1
14892 to 100% of the usual dynamic limit defined by (minconn,maxconn,fullconn).
14893
14894 - weight: when the backend uses a dynamic weighted algorithm, the weight
14895 grows linearly from 1 to 100%. In this case, the weight is updated at every
14896 health-check. For this reason, it is important that the "inter" parameter
14897 is smaller than the "slowstart", in order to maximize the number of steps.
14898
Daniel Corbett9f0843f2021-05-08 10:50:37 -040014899 The slowstart never applies when HAProxy starts, otherwise it would cause
Willy Tarreauc57f0e22009-05-10 13:12:33 +020014900 trouble to running servers. It only applies when a server has been previously
14901 seen as failed.
14902
Willy Tarreau732eac42015-07-09 11:40:25 +020014903sni <expression>
14904 The "sni" parameter evaluates the sample fetch expression, converts it to a
14905 string and uses the result as the host name sent in the SNI TLS extension to
14906 the server. A typical use case is to send the SNI received from the client in
14907 a bridged HTTPS scenario, using the "ssl_fc_sni" sample fetch for the
Willy Tarreau2ab88672017-07-05 18:23:03 +020014908 expression, though alternatives such as req.hdr(host) can also make sense. If
14909 "verify required" is set (which is the recommended setting), the resulting
Willy Tarreauad92a9a2017-07-28 11:38:41 +020014910 name will also be matched against the server certificate's names. See the
Jérôme Magninb36a6d22018-12-09 16:03:40 +010014911 "verify" directive for more details. If you want to set a SNI for health
14912 checks, see the "check-sni" directive for more details.
Willy Tarreau732eac42015-07-09 11:40:25 +020014913
Willy Tarreauc6f4ce82009-06-10 11:09:37 +020014914source <addr>[:<pl>[-<ph>]] [usesrc { <addr2>[:<port2>] | client | clientip } ]
Willy Tarreaubce70882009-09-07 11:51:47 +020014915source <addr>[:<port>] [usesrc { <addr2>[:<port2>] | hdr_ip(<hdr>[,<occ>]) } ]
Willy Tarreauc6f4ce82009-06-10 11:09:37 +020014916source <addr>[:<pl>[-<ph>]] [interface <name>] ...
Willy Tarreauc57f0e22009-05-10 13:12:33 +020014917 The "source" parameter sets the source address which will be used when
14918 connecting to the server. It follows the exact same parameters and principle
14919 as the backend "source" keyword, except that it only applies to the server
14920 referencing it. Please consult the "source" keyword for details.
14921
Willy Tarreauc6f4ce82009-06-10 11:09:37 +020014922 Additionally, the "source" statement on a server line allows one to specify a
14923 source port range by indicating the lower and higher bounds delimited by a
14924 dash ('-'). Some operating systems might require a valid IP address when a
14925 source port range is specified. It is permitted to have the same IP/range for
14926 several servers. Doing so makes it possible to bypass the maximum of 64k
14927 total concurrent connections. The limit will then reach 64k connections per
14928 server.
14929
Lukas Tribus7d56c6d2016-09-13 09:51:15 +000014930 Since Linux 4.2/libc 2.23 IP_BIND_ADDRESS_NO_PORT is set for connections
14931 specifying the source address without port(s).
14932
Willy Tarreaua0ee1d02012-09-10 09:01:23 +020014933ssl
Willy Tarreau44f65392013-06-25 07:56:20 +020014934 This option enables SSL ciphering on outgoing connections to the server. It
14935 is critical to verify server certificates using "verify" when using SSL to
14936 connect to servers, otherwise the communication is prone to trivial man in
14937 the-middle attacks rendering SSL useless. When this option is used, health
14938 checks are automatically sent in SSL too unless there is a "port" or an
14939 "addr" directive indicating the check should be sent to a different location.
Frédéric Lécailled2376272017-03-21 18:52:12 +010014940 See the "no-ssl" to disable "ssl" option and "check-ssl" option to force
14941 SSL health checks.
Willy Tarreau763a95b2012-10-04 23:15:39 +020014942
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020014943ssl-max-ver [ SSLv3 | TLSv1.0 | TLSv1.1 | TLSv1.2 | TLSv1.3 ]
14944 This option enforces use of <version> or lower when SSL is used to communicate
14945 with the server. This option is also available on global statement
14946 "ssl-default-server-options". See also "ssl-min-ver".
14947
14948ssl-min-ver [ SSLv3 | TLSv1.0 | TLSv1.1 | TLSv1.2 | TLSv1.3 ]
14949 This option enforces use of <version> or upper when SSL is used to communicate
14950 with the server. This option is also available on global statement
14951 "ssl-default-server-options". See also "ssl-max-ver".
14952
Frédéric Lécailled2376272017-03-21 18:52:12 +010014953ssl-reuse
14954 This option may be used as "server" setting to reset any "no-ssl-reuse"
14955 setting which would have been inherited from "default-server" directive as
14956 default value.
14957 It may also be used as "default-server" setting to reset any previous
14958 "default-server" "no-ssl-reuse" setting.
14959
14960stick
14961 This option may be used as "server" setting to reset any "non-stick"
14962 setting which would have been inherited from "default-server" directive as
14963 default value.
14964 It may also be used as "default-server" setting to reset any previous
14965 "default-server" "non-stick" setting.
Willy Tarreaua0ee1d02012-09-10 09:01:23 +020014966
Alexander Liu2a54bb72019-05-22 19:44:48 +080014967socks4 <addr>:<port>
John Roeslerfb2fce12019-07-10 15:45:51 -050014968 This option enables upstream socks4 tunnel for outgoing connections to the
Alexander Liu2a54bb72019-05-22 19:44:48 +080014969 server. Using this option won't force the health check to go via socks4 by
14970 default. You will have to use the keyword "check-via-socks4" to enable it.
14971
Willy Tarreau163d4622015-10-13 16:16:41 +020014972tcp-ut <delay>
14973 Sets the TCP User Timeout for all outgoing connections to this server. This
Daniel Corbett9f0843f2021-05-08 10:50:37 -040014974 option is available on Linux since version 2.6.37. It allows HAProxy to
Willy Tarreau163d4622015-10-13 16:16:41 +020014975 configure a timeout for sockets which contain data not receiving an
Davor Ocelice9ed2812017-12-25 17:49:28 +010014976 acknowledgment for the configured delay. This is especially useful on
Willy Tarreau163d4622015-10-13 16:16:41 +020014977 long-lived connections experiencing long idle periods such as remote
14978 terminals or database connection pools, where the client and server timeouts
14979 must remain high to allow a long period of idle, but where it is important to
14980 detect that the server has disappeared in order to release all resources
14981 associated with its connection (and the client's session). One typical use
14982 case is also to force dead server connections to die when health checks are
14983 too slow or during a soft reload since health checks are then disabled. The
14984 argument is a delay expressed in milliseconds by default. This only works for
14985 regular TCP connections, and is ignored for other protocols.
14986
Willy Tarreau034c88c2017-01-23 23:36:45 +010014987tfo
14988 This option enables using TCP fast open when connecting to servers, on
14989 systems that support it (currently only the Linux kernel >= 4.11).
14990 See the "tfo" bind option for more information about TCP fast open.
14991 Please note that when using tfo, you should also use the "conn-failure",
Daniel Corbett9f0843f2021-05-08 10:50:37 -040014992 "empty-response" and "response-timeout" keywords for "retry-on", or HAProxy
Frédéric Lécaille1b9423d2019-07-04 14:19:06 +020014993 won't be able to retry the connection on failure. See also "no-tfo".
Willy Tarreau034c88c2017-01-23 23:36:45 +010014994
Willy Tarreauc57f0e22009-05-10 13:12:33 +020014995track [<proxy>/]<server>
Willy Tarreau32091232014-05-16 13:52:00 +020014996 This option enables ability to set the current state of the server by tracking
14997 another one. It is possible to track a server which itself tracks another
14998 server, provided that at the end of the chain, a server has health checks
14999 enabled. If <proxy> is omitted the current one is used. If disable-on-404 is
Willy Tarreauc57f0e22009-05-10 13:12:33 +020015000 used, it has to be enabled on both proxies.
15001
Frédéric Lécailled2376272017-03-21 18:52:12 +010015002tls-tickets
15003 This option may be used as "server" setting to reset any "no-tls-tickets"
15004 setting which would have been inherited from "default-server" directive as
15005 default value.
Lukas Tribusbdb386d2020-03-10 00:56:09 +010015006 The TLS ticket mechanism is only used up to TLS 1.2.
15007 Forward Secrecy is compromised with TLS tickets, unless ticket keys
15008 are periodically rotated (via reload or by using "tls-ticket-keys").
Frédéric Lécailled2376272017-03-21 18:52:12 +010015009 It may also be used as "default-server" setting to reset any previous
Bjoern Jacke5ab7eb62020-02-13 14:16:16 +010015010 "default-server" "no-tls-tickets" setting.
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010015011
Emeric Brunef42d922012-10-11 16:11:36 +020015012verify [none|required]
15013 This setting is only available when support for OpenSSL was built in. If set
Emeric Brun850efd52014-01-29 12:24:34 +010015014 to 'none', server certificate is not verified. In the other case, The
Willy Tarreauad92a9a2017-07-28 11:38:41 +020015015 certificate provided by the server is verified using CAs from 'ca-file' and
15016 optional CRLs from 'crl-file' after having checked that the names provided in
Davor Ocelice9ed2812017-12-25 17:49:28 +010015017 the certificate's subject and subjectAlternateNames attributes match either
Willy Tarreauad92a9a2017-07-28 11:38:41 +020015018 the name passed using the "sni" directive, or if not provided, the static
15019 host name passed using the "verifyhost" directive. When no name is found, the
15020 certificate's names are ignored. For this reason, without SNI it's important
15021 to use "verifyhost". On verification failure the handshake is aborted. It is
15022 critically important to verify server certificates when using SSL to connect
15023 to servers, otherwise the communication is prone to trivial man-in-the-middle
15024 attacks rendering SSL totally useless. Unless "ssl_server_verify" appears in
15025 the global section, "verify" is set to "required" by default.
Emeric Brunef42d922012-10-11 16:11:36 +020015026
Evan Broderbe554312013-06-27 00:05:25 -070015027verifyhost <hostname>
15028 This setting is only available when support for OpenSSL was built in, and
Willy Tarreauad92a9a2017-07-28 11:38:41 +020015029 only takes effect if 'verify required' is also specified. This directive sets
15030 a default static hostname to check the server's certificate against when no
15031 SNI was used to connect to the server. If SNI is not used, this is the only
15032 way to enable hostname verification. This static hostname, when set, will
15033 also be used for health checks (which cannot provide an SNI value). If none
15034 of the hostnames in the certificate match the specified hostname, the
15035 handshake is aborted. The hostnames in the server-provided certificate may
15036 include wildcards. See also "verify", "sni" and "no-verifyhost" options.
Evan Broderbe554312013-06-27 00:05:25 -070015037
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010015038weight <weight>
Willy Tarreauc57f0e22009-05-10 13:12:33 +020015039 The "weight" parameter is used to adjust the server's weight relative to
15040 other servers. All servers will receive a load proportional to their weight
15041 relative to the sum of all weights, so the higher the weight, the higher the
Willy Tarreau6704d672009-06-15 10:56:05 +020015042 load. The default weight is 1, and the maximal value is 256. A value of 0
15043 means the server will not participate in load-balancing but will still accept
15044 persistent connections. If this parameter is used to distribute the load
15045 according to server's capacity, it is recommended to start with values which
15046 can both grow and shrink, for instance between 10 and 100 to leave enough
15047 room above and below for later adjustments.
Willy Tarreauc57f0e22009-05-10 13:12:33 +020015048
15049
Cyril Bonté46175dd2015-07-02 22:45:32 +0200150505.3. Server IP address resolution using DNS
15051-------------------------------------------
Baptiste Assmann1fa66662015-04-14 00:28:47 +020015052
Baptiste Assmann62b75b42015-09-09 01:11:36 +020015053HAProxy allows using a host name on the server line to retrieve its IP address
15054using name servers. By default, HAProxy resolves the name when parsing the
Thayne McCombscdbcca92021-01-07 21:24:41 -070015055configuration file, at startup and cache the result for the process's life.
Baptiste Assmann1fa66662015-04-14 00:28:47 +020015056This is not sufficient in some cases, such as in Amazon where a server's IP
15057can change after a reboot or an ELB Virtual IP can change based on current
15058workload.
15059This chapter describes how HAProxy can be configured to process server's name
15060resolution at run time.
15061Whether run time server name resolution has been enable or not, HAProxy will
15062carry on doing the first resolution when parsing the configuration.
15063
15064
Cyril Bonté46175dd2015-07-02 22:45:32 +0200150655.3.1. Global overview
15066----------------------
Baptiste Assmann1fa66662015-04-14 00:28:47 +020015067
15068As we've seen in introduction, name resolution in HAProxy occurs at two
15069different steps of the process life:
15070
15071 1. when starting up, HAProxy parses the server line definition and matches a
15072 host name. It uses libc functions to get the host name resolved. This
15073 resolution relies on /etc/resolv.conf file.
15074
Christopher Faulet67957bd2017-09-27 11:00:59 +020015075 2. at run time, HAProxy performs periodically name resolutions for servers
15076 requiring DNS resolutions.
Baptiste Assmann1fa66662015-04-14 00:28:47 +020015077
15078A few other events can trigger a name resolution at run time:
15079 - when a server's health check ends up in a connection timeout: this may be
15080 because the server has a new IP address. So we need to trigger a name
15081 resolution to know this new IP.
15082
Christopher Faulet67957bd2017-09-27 11:00:59 +020015083When using resolvers, the server name can either be a hostname, or a SRV label.
Davor Ocelice9ed2812017-12-25 17:49:28 +010015084HAProxy considers anything that starts with an underscore as a SRV label. If a
Christopher Faulet67957bd2017-09-27 11:00:59 +020015085SRV label is specified, then the corresponding SRV records will be retrieved
15086from the DNS server, and the provided hostnames will be used. The SRV label
Daniel Corbett9f0843f2021-05-08 10:50:37 -040015087will be checked periodically, and if any server are added or removed, HAProxy
Christopher Faulet67957bd2017-09-27 11:00:59 +020015088will automatically do the same.
Olivier Houchardecfa18d2017-08-07 17:30:03 +020015089
Baptiste Assmann1fa66662015-04-14 00:28:47 +020015090A few things important to notice:
John Roeslerfb2fce12019-07-10 15:45:51 -050015091 - all the name servers are queried in the meantime. HAProxy will process the
Baptiste Assmann1fa66662015-04-14 00:28:47 +020015092 first valid response.
15093
15094 - a resolution is considered as invalid (NX, timeout, refused), when all the
15095 servers return an error.
15096
15097
Cyril Bonté46175dd2015-07-02 22:45:32 +0200150985.3.2. The resolvers section
15099----------------------------
Baptiste Assmann1fa66662015-04-14 00:28:47 +020015100
15101This section is dedicated to host information related to name resolution in
Christopher Faulet67957bd2017-09-27 11:00:59 +020015102HAProxy. There can be as many as resolvers section as needed. Each section can
15103contain many name servers.
Baptiste Assmann1fa66662015-04-14 00:28:47 +020015104
Baptiste Assmann62b75b42015-09-09 01:11:36 +020015105When multiple name servers are configured in a resolvers section, then HAProxy
15106uses the first valid response. In case of invalid responses, only the last one
15107is treated. Purpose is to give the chance to a slow server to deliver a valid
15108answer after a fast faulty or outdated server.
15109
15110When each server returns a different error type, then only the last error is
Christopher Faulet67957bd2017-09-27 11:00:59 +020015111used by HAProxy. The following processing is applied on this error:
Baptiste Assmann62b75b42015-09-09 01:11:36 +020015112
Christopher Faulet67957bd2017-09-27 11:00:59 +020015113 1. HAProxy retries the same DNS query with a new query type. The A queries are
15114 switch to AAAA or the opposite. SRV queries are not concerned here. Timeout
15115 errors are also excluded.
Baptiste Assmann62b75b42015-09-09 01:11:36 +020015116
Christopher Faulet67957bd2017-09-27 11:00:59 +020015117 2. When the fallback on the query type was done (or not applicable), HAProxy
15118 retries the original DNS query, with the preferred query type.
Baptiste Assmann62b75b42015-09-09 01:11:36 +020015119
Christopher Faulet67957bd2017-09-27 11:00:59 +020015120 3. HAProxy retries previous steps <resolve_retires> times. If no valid
15121 response is received after that, it stops the DNS resolution and reports
15122 the error.
Baptiste Assmann62b75b42015-09-09 01:11:36 +020015123
Christopher Faulet67957bd2017-09-27 11:00:59 +020015124For example, with 2 name servers configured in a resolvers section, the
15125following scenarios are possible:
15126
15127 - First response is valid and is applied directly, second response is
15128 ignored
15129
15130 - First response is invalid and second one is valid, then second response is
15131 applied
15132
15133 - First response is a NX domain and second one a truncated response, then
15134 HAProxy retries the query with a new type
15135
15136 - First response is a NX domain and second one is a timeout, then HAProxy
15137 retries the query with a new type
15138
15139 - Query timed out for both name servers, then HAProxy retries it with the
15140 same query type
Baptiste Assmann62b75b42015-09-09 01:11:36 +020015141
Daniel Corbett9f0843f2021-05-08 10:50:37 -040015142As a DNS server may not answer all the IPs in one DNS request, HAProxy keeps
Olivier Houcharda8c6db82017-07-06 18:46:47 +020015143a cache of previous answers, an answer will be considered obsolete after
Christopher Faulet67957bd2017-09-27 11:00:59 +020015144<hold obsolete> seconds without the IP returned.
Olivier Houcharda8c6db82017-07-06 18:46:47 +020015145
Baptiste Assmann62b75b42015-09-09 01:11:36 +020015146
Baptiste Assmann1fa66662015-04-14 00:28:47 +020015147resolvers <resolvers id>
Davor Ocelice9ed2812017-12-25 17:49:28 +010015148 Creates a new name server list labeled <resolvers id>
Baptiste Assmann1fa66662015-04-14 00:28:47 +020015149
15150A resolvers section accept the following parameters:
15151
Baptiste Assmann2af08fe2017-08-14 00:13:01 +020015152accepted_payload_size <nb>
Davor Ocelice9ed2812017-12-25 17:49:28 +010015153 Defines the maximum payload size accepted by HAProxy and announced to all the
Christopher Faulet67957bd2017-09-27 11:00:59 +020015154 name servers configured in this resolvers section.
Baptiste Assmann2af08fe2017-08-14 00:13:01 +020015155 <nb> is in bytes. If not set, HAProxy announces 512. (minimal value defined
15156 by RFC 6891)
15157
Emeric Brun4c751952021-03-08 16:41:29 +010015158 Note: the maximum allowed value is 65535. Recommended value for UDP is
15159 4096 and it is not recommended to exceed 8192 except if you are sure
15160 that your system and network can handle this (over 65507 makes no sense
15161 since is the maximum UDP payload size). If you are using only TCP
15162 nameservers to handle huge DNS responses, you should put this value
15163 to the max: 65535.
Baptiste Assmann9d8dbbc2017-08-18 23:35:08 +020015164
Emeric Brunc8f3e452021-04-07 16:04:54 +020015165nameserver <name> <address>[:port] [param*]
15166 Used to configure a nameserver. <name> of the nameserver should ne unique.
15167 By default the <address> is considered of type datagram. This means if an
15168 IPv4 or IPv6 is configured without special address prefixes (paragraph 11.)
15169 the UDP protocol will be used. If an stream protocol address prefix is used,
15170 the nameserver will be considered as a stream server (TCP for instance) and
15171 "server" parameters found in 5.2 paragraph which are relevant for DNS
15172 resolving will be considered. Note: currently, in TCP mode, 4 queries are
15173 pipelined on the same connections. A batch of idle connections are removed
15174 every 5 seconds. "maxconn" can be configured to limit the amount of those
Emeric Brun56fc5d92021-02-12 20:05:45 +010015175 concurrent connections and TLS should also usable if the server supports.
15176
Ben Draut44e609b2018-05-29 15:40:08 -060015177parse-resolv-conf
15178 Adds all nameservers found in /etc/resolv.conf to this resolvers nameservers
15179 list. Ordered as if each nameserver in /etc/resolv.conf was individually
15180 placed in the resolvers section in place of this directive.
15181
Baptiste Assmann1fa66662015-04-14 00:28:47 +020015182hold <status> <period>
15183 Defines <period> during which the last name resolution should be kept based
15184 on last resolution <status>
Baptiste Assmann987e16d2016-11-02 22:23:31 +010015185 <status> : last name resolution status. Acceptable values are "nx",
Olivier Houcharda8c6db82017-07-06 18:46:47 +020015186 "other", "refused", "timeout", "valid", "obsolete".
Baptiste Assmann1fa66662015-04-14 00:28:47 +020015187 <period> : interval between two successive name resolution when the last
15188 answer was in <status>. It follows the HAProxy time format.
15189 <period> is in milliseconds by default.
15190
Baptiste Assmann686408b2017-08-18 10:15:42 +020015191 Default value is 10s for "valid", 0s for "obsolete" and 30s for others.
Baptiste Assmann1fa66662015-04-14 00:28:47 +020015192
Baptiste Assmann1fa66662015-04-14 00:28:47 +020015193resolve_retries <nb>
15194 Defines the number <nb> of queries to send to resolve a server name before
15195 giving up.
15196 Default value: 3
15197
Baptiste Assmann62b75b42015-09-09 01:11:36 +020015198 A retry occurs on name server timeout or when the full sequence of DNS query
15199 type failover is over and we need to start up from the default ANY query
15200 type.
15201
Baptiste Assmann1fa66662015-04-14 00:28:47 +020015202timeout <event> <time>
15203 Defines timeouts related to name resolution
15204 <event> : the event on which the <time> timeout period applies to.
15205 events available are:
Frédéric Lécaille93d33162019-03-06 09:35:59 +010015206 - resolve : default time to trigger name resolutions when no
15207 other time applied.
Christopher Faulet67957bd2017-09-27 11:00:59 +020015208 Default value: 1s
15209 - retry : time between two DNS queries, when no valid response
Frédéric Lécaille93d33162019-03-06 09:35:59 +010015210 have been received.
Christopher Faulet67957bd2017-09-27 11:00:59 +020015211 Default value: 1s
Baptiste Assmann1fa66662015-04-14 00:28:47 +020015212 <time> : time related to the event. It follows the HAProxy time format.
15213 <time> is expressed in milliseconds.
15214
Olivier Doucetaa1ea8a2016-08-05 17:15:20 +020015215 Example:
Baptiste Assmann1fa66662015-04-14 00:28:47 +020015216
15217 resolvers mydns
15218 nameserver dns1 10.0.0.1:53
15219 nameserver dns2 10.0.0.2:53
Emeric Brunc8f3e452021-04-07 16:04:54 +020015220 nameserver dns3 tcp@10.0.0.3:53
Ben Draut44e609b2018-05-29 15:40:08 -060015221 parse-resolv-conf
Baptiste Assmann1fa66662015-04-14 00:28:47 +020015222 resolve_retries 3
Christopher Faulet67957bd2017-09-27 11:00:59 +020015223 timeout resolve 1s
Baptiste Assmann1fa66662015-04-14 00:28:47 +020015224 timeout retry 1s
Baptiste Assmann987e16d2016-11-02 22:23:31 +010015225 hold other 30s
15226 hold refused 30s
15227 hold nx 30s
15228 hold timeout 30s
Baptiste Assmann1fa66662015-04-14 00:28:47 +020015229 hold valid 10s
Olivier Houcharda8c6db82017-07-06 18:46:47 +020015230 hold obsolete 30s
Baptiste Assmann1fa66662015-04-14 00:28:47 +020015231
15232
Christopher Faulet87f1f3d2019-07-18 14:51:20 +0200152336. Cache
15234---------
15235
15236HAProxy provides a cache, which was designed to perform cache on small objects
15237(favicon, css...). This is a minimalist low-maintenance cache which runs in
15238RAM.
15239
15240The cache is based on a memory which is shared between processes and threads,
15241this memory is split in blocks of 1k.
15242
15243If an object is not used anymore, it can be deleted to store a new object
15244independently of its expiration date. The oldest objects are deleted first
15245when we try to allocate a new one.
15246
15247The cache uses a hash of the host header and the URI as the key.
15248
15249It's possible to view the status of a cache using the Unix socket command
15250"show cache" consult section 9.3 "Unix Socket commands" of Management Guide
15251for more details.
15252
15253When an object is delivered from the cache, the server name in the log is
15254replaced by "<CACHE>".
15255
15256
152576.1. Limitation
15258----------------
15259
15260The cache won't store and won't deliver objects in these cases:
15261
15262- If the response is not a 200
Remi Tricot-Le Breton4f730832020-11-26 15:51:50 +010015263- If the response contains a Vary header and either the process-vary option is
15264 disabled, or a currently unmanaged header is specified in the Vary value (only
15265 accept-encoding and referer are managed for now)
Christopher Faulet87f1f3d2019-07-18 14:51:20 +020015266- If the Content-Length + the headers size is greater than "max-object-size"
15267- If the response is not cacheable
Remi Tricot-Le Bretond493bc82020-11-26 15:51:29 +010015268- If the response does not have an explicit expiration time (s-maxage or max-age
15269 Cache-Control directives or Expires header) or a validator (ETag or Last-Modified
15270 headers)
Remi Tricot-Le Breton5853c0c2020-12-10 17:58:43 +010015271- If the process-vary option is enabled and there are already max-secondary-entries
15272 entries with the same primary key as the current response
Remi Tricot-Le Breton6ca89162021-01-07 14:50:51 +010015273- If the process-vary option is enabled and the response has an unknown encoding (not
15274 mentioned in https://www.iana.org/assignments/http-parameters/http-parameters.xhtml)
15275 while varying on the accept-encoding client header
Christopher Faulet87f1f3d2019-07-18 14:51:20 +020015276
15277- If the request is not a GET
15278- If the HTTP version of the request is smaller than 1.1
15279- If the request contains an Authorization header
15280
15281
152826.2. Setup
15283-----------
15284
15285To setup a cache, you must define a cache section and use it in a proxy with
15286the corresponding http-request and response actions.
15287
15288
152896.2.1. Cache section
15290---------------------
15291
15292cache <name>
15293 Declare a cache section, allocate a shared cache memory named <name>, the
15294 size of cache is mandatory.
15295
15296total-max-size <megabytes>
15297 Define the size in RAM of the cache in megabytes. This size is split in
15298 blocks of 1kB which are used by the cache entries. Its maximum value is 4095.
15299
15300max-object-size <bytes>
15301 Define the maximum size of the objects to be cached. Must not be greater than
15302 an half of "total-max-size". If not set, it equals to a 256th of the cache size.
15303 All objects with sizes larger than "max-object-size" will not be cached.
15304
15305max-age <seconds>
Remi Tricot-Le Breton5853c0c2020-12-10 17:58:43 +010015306 Define the maximum expiration duration. The expiration is set as the lowest
Christopher Faulet87f1f3d2019-07-18 14:51:20 +020015307 value between the s-maxage or max-age (in this order) directive in the
15308 Cache-Control response header and this value. The default value is 60
15309 seconds, which means that you can't cache an object more than 60 seconds by
15310 default.
15311
Remi Tricot-Le Bretone6cc5b52020-12-23 18:13:53 +010015312process-vary <on/off>
15313 Enable or disable the processing of the Vary header. When disabled, a response
Remi Tricot-Le Breton754b2422020-11-16 15:56:10 +010015314 containing such a header will never be cached. When enabled, we need to calculate
15315 a preliminary hash for a subset of request headers on all the incoming requests
15316 (which might come with a cpu cost) which will be used to build a secondary key
Remi Tricot-Le Bretone6cc5b52020-12-23 18:13:53 +010015317 for a given request (see RFC 7234#4.1). The default value is off (disabled).
Remi Tricot-Le Breton754b2422020-11-16 15:56:10 +010015318
Remi Tricot-Le Breton5853c0c2020-12-10 17:58:43 +010015319max-secondary-entries <number>
15320 Define the maximum number of simultaneous secondary entries with the same primary
15321 key in the cache. This needs the vary support to be enabled. Its default value is 10
15322 and should be passed a strictly positive integer.
15323
Christopher Faulet87f1f3d2019-07-18 14:51:20 +020015324
153256.2.2. Proxy section
15326---------------------
15327
15328http-request cache-use <name> [ { if | unless } <condition> ]
15329 Try to deliver a cached object from the cache <name>. This directive is also
15330 mandatory to store the cache as it calculates the cache hash. If you want to
15331 use a condition for both storage and delivering that's a good idea to put it
15332 after this one.
15333
15334http-response cache-store <name> [ { if | unless } <condition> ]
15335 Store an http-response within the cache. The storage of the response headers
15336 is done at this step, which means you can use others http-response actions
15337 to modify headers before or after the storage of the response. This action
15338 is responsible for the setup of the cache storage filter.
15339
15340
15341Example:
15342
15343 backend bck1
15344 mode http
15345
15346 http-request cache-use foobar
15347 http-response cache-store foobar
15348 server srv1 127.0.0.1:80
15349
15350 cache foobar
15351 total-max-size 4
15352 max-age 240
15353
15354
Willy Tarreau74ca5042013-06-11 23:12:07 +0200153557. Using ACLs and fetching samples
15356----------------------------------
15357
Davor Ocelice9ed2812017-12-25 17:49:28 +010015358HAProxy is capable of extracting data from request or response streams, from
Willy Tarreau74ca5042013-06-11 23:12:07 +020015359client or server information, from tables, environmental information etc...
15360The action of extracting such data is called fetching a sample. Once retrieved,
15361these samples may be used for various purposes such as a key to a stick-table,
15362but most common usages consist in matching them against predefined constant
15363data called patterns.
15364
15365
153667.1. ACL basics
15367---------------
Willy Tarreauc57f0e22009-05-10 13:12:33 +020015368
15369The use of Access Control Lists (ACL) provides a flexible solution to perform
15370content switching and generally to take decisions based on content extracted
15371from the request, the response or any environmental status. The principle is
15372simple :
15373
Willy Tarreau74ca5042013-06-11 23:12:07 +020015374 - extract a data sample from a stream, table or the environment
Willy Tarreaue6b11e42013-11-26 19:02:32 +010015375 - optionally apply some format conversion to the extracted sample
Willy Tarreau74ca5042013-06-11 23:12:07 +020015376 - apply one or multiple pattern matching methods on this sample
15377 - perform actions only when a pattern matches the sample
Willy Tarreauc57f0e22009-05-10 13:12:33 +020015378
Willy Tarreau74ca5042013-06-11 23:12:07 +020015379The actions generally consist in blocking a request, selecting a backend, or
15380adding a header.
Willy Tarreauc57f0e22009-05-10 13:12:33 +020015381
15382In order to define a test, the "acl" keyword is used. The syntax is :
15383
Willy Tarreau74ca5042013-06-11 23:12:07 +020015384 acl <aclname> <criterion> [flags] [operator] [<value>] ...
Willy Tarreauc57f0e22009-05-10 13:12:33 +020015385
15386This creates a new ACL <aclname> or completes an existing one with new tests.
15387Those tests apply to the portion of request/response specified in <criterion>
15388and may be adjusted with optional flags [flags]. Some criteria also support
Willy Tarreaue6b11e42013-11-26 19:02:32 +010015389an operator which may be specified before the set of values. Optionally some
15390conversion operators may be applied to the sample, and they will be specified
15391as a comma-delimited list of keywords just after the first keyword. The values
15392are of the type supported by the criterion, and are separated by spaces.
Willy Tarreauc57f0e22009-05-10 13:12:33 +020015393
15394ACL names must be formed from upper and lower case letters, digits, '-' (dash),
15395'_' (underscore) , '.' (dot) and ':' (colon). ACL names are case-sensitive,
15396which means that "my_acl" and "My_Acl" are two different ACLs.
15397
15398There is no enforced limit to the number of ACLs. The unused ones do not affect
15399performance, they just consume a small amount of memory.
15400
Willy Tarreau74ca5042013-06-11 23:12:07 +020015401The criterion generally is the name of a sample fetch method, or one of its ACL
15402specific declinations. The default test method is implied by the output type of
15403this sample fetch method. The ACL declinations can describe alternate matching
Willy Tarreaue6b11e42013-11-26 19:02:32 +010015404methods of a same sample fetch method. The sample fetch methods are the only
15405ones supporting a conversion.
Willy Tarreau74ca5042013-06-11 23:12:07 +020015406
15407Sample fetch methods return data which can be of the following types :
15408 - boolean
15409 - integer (signed or unsigned)
15410 - IPv4 or IPv6 address
15411 - string
15412 - data block
15413
Willy Tarreaue6b11e42013-11-26 19:02:32 +010015414Converters transform any of these data into any of these. For example, some
15415converters might convert a string to a lower-case string while other ones
15416would turn a string to an IPv4 address, or apply a netmask to an IP address.
15417The resulting sample is of the type of the last converter applied to the list,
15418which defaults to the type of the sample fetch method.
15419
Thierry FOURNIER2a06e392014-05-11 15:49:55 +020015420Each sample or converter returns data of a specific type, specified with its
15421keyword in this documentation. When an ACL is declared using a standard sample
15422fetch method, certain types automatically involved a default matching method
15423which are summarized in the table below :
15424
15425 +---------------------+-----------------+
15426 | Sample or converter | Default |
15427 | output type | matching method |
15428 +---------------------+-----------------+
15429 | boolean | bool |
15430 +---------------------+-----------------+
15431 | integer | int |
15432 +---------------------+-----------------+
15433 | ip | ip |
15434 +---------------------+-----------------+
15435 | string | str |
15436 +---------------------+-----------------+
15437 | binary | none, use "-m" |
15438 +---------------------+-----------------+
15439
15440Note that in order to match a binary samples, it is mandatory to specify a
15441matching method, see below.
15442
Willy Tarreau74ca5042013-06-11 23:12:07 +020015443The ACL engine can match these types against patterns of the following types :
15444 - boolean
15445 - integer or integer range
15446 - IP address / network
15447 - string (exact, substring, suffix, prefix, subdir, domain)
15448 - regular expression
15449 - hex block
15450
Willy Tarreauc57f0e22009-05-10 13:12:33 +020015451The following ACL flags are currently supported :
15452
Willy Tarreau2b5285d2010-05-09 23:45:24 +020015453 -i : ignore case during matching of all subsequent patterns.
15454 -f : load patterns from a file.
Willy Tarreau74ca5042013-06-11 23:12:07 +020015455 -m : use a specific pattern matching method
Thierry FOURNIERb7729c92014-02-11 16:24:41 +010015456 -n : forbid the DNS resolutions
Thierry FOURNIER9860c412014-01-29 14:23:29 +010015457 -M : load the file pointed by -f like a map file.
Thierry FOURNIER3534d882014-01-20 17:01:44 +010015458 -u : force the unique id of the ACL
Willy Tarreau6a06a402007-07-15 20:15:28 +020015459 -- : force end of flags. Useful when a string looks like one of the flags.
15460
Willy Tarreau74ca5042013-06-11 23:12:07 +020015461The "-f" flag is followed by the name of a file from which all lines will be
15462read as individual values. It is even possible to pass multiple "-f" arguments
15463if the patterns are to be loaded from multiple files. Empty lines as well as
15464lines beginning with a sharp ('#') will be ignored. All leading spaces and tabs
15465will be stripped. If it is absolutely necessary to insert a valid pattern
15466beginning with a sharp, just prefix it with a space so that it is not taken for
Daniel Corbett9f0843f2021-05-08 10:50:37 -040015467a comment. Depending on the data type and match method, HAProxy may load the
Willy Tarreau74ca5042013-06-11 23:12:07 +020015468lines into a binary tree, allowing very fast lookups. This is true for IPv4 and
15469exact string matching. In this case, duplicates will automatically be removed.
15470
Thierry FOURNIER9860c412014-01-29 14:23:29 +010015471The "-M" flag allows an ACL to use a map file. If this flag is set, the file is
15472parsed as two column file. The first column contains the patterns used by the
15473ACL, and the second column contain the samples. The sample can be used later by
15474a map. This can be useful in some rare cases where an ACL would just be used to
15475check for the existence of a pattern in a map before a mapping is applied.
15476
Thierry FOURNIER3534d882014-01-20 17:01:44 +010015477The "-u" flag forces the unique id of the ACL. This unique id is used with the
15478socket interface to identify ACL and dynamically change its values. Note that a
15479file is always identified by its name even if an id is set.
15480
Willy Tarreau74ca5042013-06-11 23:12:07 +020015481Also, note that the "-i" flag applies to subsequent entries and not to entries
15482loaded from files preceding it. For instance :
15483
15484 acl valid-ua hdr(user-agent) -f exact-ua.lst -i -f generic-ua.lst test
15485
15486In this example, each line of "exact-ua.lst" will be exactly matched against
15487the "user-agent" header of the request. Then each line of "generic-ua" will be
15488case-insensitively matched. Then the word "test" will be insensitively matched
15489as well.
15490
15491The "-m" flag is used to select a specific pattern matching method on the input
15492sample. All ACL-specific criteria imply a pattern matching method and generally
15493do not need this flag. However, this flag is useful with generic sample fetch
15494methods to describe how they're going to be matched against the patterns. This
15495is required for sample fetches which return data type for which there is no
Davor Ocelice9ed2812017-12-25 17:49:28 +010015496obvious matching method (e.g. string or binary). When "-m" is specified and
Willy Tarreau74ca5042013-06-11 23:12:07 +020015497followed by a pattern matching method name, this method is used instead of the
15498default one for the criterion. This makes it possible to match contents in ways
15499that were not initially planned, or with sample fetch methods which return a
15500string. The matching method also affects the way the patterns are parsed.
15501
Thierry FOURNIERb7729c92014-02-11 16:24:41 +010015502The "-n" flag forbids the dns resolutions. It is used with the load of ip files.
15503By default, if the parser cannot parse ip address it considers that the parsed
15504string is maybe a domain name and try dns resolution. The flag "-n" disable this
15505resolution. It is useful for detecting malformed ip lists. Note that if the DNS
Daniel Corbett9f0843f2021-05-08 10:50:37 -040015506server is not reachable, the HAProxy configuration parsing may last many minutes
John Roeslerfb2fce12019-07-10 15:45:51 -050015507waiting for the timeout. During this time no error messages are displayed. The
Thierry FOURNIERb7729c92014-02-11 16:24:41 +010015508flag "-n" disable this behavior. Note also that during the runtime, this
15509function is disabled for the dynamic acl modifications.
15510
Willy Tarreau74ca5042013-06-11 23:12:07 +020015511There are some restrictions however. Not all methods can be used with all
15512sample fetch methods. Also, if "-m" is used in conjunction with "-f", it must
15513be placed first. The pattern matching method must be one of the following :
Willy Tarreau5adeda12013-03-31 22:13:34 +020015514
15515 - "found" : only check if the requested sample could be found in the stream,
15516 but do not compare it against any pattern. It is recommended not
Willy Tarreau74ca5042013-06-11 23:12:07 +020015517 to pass any pattern to avoid confusion. This matching method is
15518 particularly useful to detect presence of certain contents such
15519 as headers, cookies, etc... even if they are empty and without
15520 comparing them to anything nor counting them.
Willy Tarreau5adeda12013-03-31 22:13:34 +020015521
15522 - "bool" : check the value as a boolean. It can only be applied to fetches
15523 which return a boolean or integer value, and takes no pattern.
Willy Tarreau74ca5042013-06-11 23:12:07 +020015524 Value zero or false does not match, all other values do match.
Willy Tarreau5adeda12013-03-31 22:13:34 +020015525
15526 - "int" : match the value as an integer. It can be used with integer and
Willy Tarreau74ca5042013-06-11 23:12:07 +020015527 boolean samples. Boolean false is integer 0, true is integer 1.
Willy Tarreau5adeda12013-03-31 22:13:34 +020015528
15529 - "ip" : match the value as an IPv4 or IPv6 address. It is compatible
Willy Tarreau74ca5042013-06-11 23:12:07 +020015530 with IP address samples only, so it is implied and never needed.
Willy Tarreau5adeda12013-03-31 22:13:34 +020015531
Davor Ocelice9ed2812017-12-25 17:49:28 +010015532 - "bin" : match the contents against a hexadecimal string representing a
Willy Tarreau5adeda12013-03-31 22:13:34 +020015533 binary sequence. This may be used with binary or string samples.
15534
15535 - "len" : match the sample's length as an integer. This may be used with
15536 binary or string samples.
15537
Willy Tarreau74ca5042013-06-11 23:12:07 +020015538 - "str" : exact match : match the contents against a string. This may be
15539 used with binary or string samples.
Willy Tarreau5adeda12013-03-31 22:13:34 +020015540
Willy Tarreau74ca5042013-06-11 23:12:07 +020015541 - "sub" : substring match : check that the contents contain at least one of
15542 the provided string patterns. This may be used with binary or
15543 string samples.
Willy Tarreau5adeda12013-03-31 22:13:34 +020015544
Willy Tarreau74ca5042013-06-11 23:12:07 +020015545 - "reg" : regex match : match the contents against a list of regular
15546 expressions. This may be used with binary or string samples.
Willy Tarreau5adeda12013-03-31 22:13:34 +020015547
Willy Tarreau74ca5042013-06-11 23:12:07 +020015548 - "beg" : prefix match : check that the contents begin like the provided
15549 string patterns. This may be used with binary or string samples.
Willy Tarreau5adeda12013-03-31 22:13:34 +020015550
Willy Tarreau74ca5042013-06-11 23:12:07 +020015551 - "end" : suffix match : check that the contents end like the provided
15552 string patterns. This may be used with binary or string samples.
Willy Tarreau5adeda12013-03-31 22:13:34 +020015553
Willy Tarreau74ca5042013-06-11 23:12:07 +020015554 - "dir" : subdir match : check that a slash-delimited portion of the
15555 contents exactly matches one of the provided string patterns.
Willy Tarreau5adeda12013-03-31 22:13:34 +020015556 This may be used with binary or string samples.
15557
Willy Tarreau74ca5042013-06-11 23:12:07 +020015558 - "dom" : domain match : check that a dot-delimited portion of the contents
15559 exactly match one of the provided string patterns. This may be
15560 used with binary or string samples.
Willy Tarreau5adeda12013-03-31 22:13:34 +020015561
15562For example, to quickly detect the presence of cookie "JSESSIONID" in an HTTP
15563request, it is possible to do :
15564
15565 acl jsess_present cook(JSESSIONID) -m found
15566
15567In order to apply a regular expression on the 500 first bytes of data in the
15568buffer, one would use the following acl :
15569
15570 acl script_tag payload(0,500) -m reg -i <script>
15571
Willy Tarreaue6b11e42013-11-26 19:02:32 +010015572On systems where the regex library is much slower when using "-i", it is
15573possible to convert the sample to lowercase before matching, like this :
15574
15575 acl script_tag payload(0,500),lower -m reg <script>
15576
Willy Tarreau74ca5042013-06-11 23:12:07 +020015577All ACL-specific criteria imply a default matching method. Most often, these
15578criteria are composed by concatenating the name of the original sample fetch
15579method and the matching method. For example, "hdr_beg" applies the "beg" match
15580to samples retrieved using the "hdr" fetch method. Since all ACL-specific
15581criteria rely on a sample fetch method, it is always possible instead to use
15582the original sample fetch method and the explicit matching method using "-m".
Willy Tarreau2b5285d2010-05-09 23:45:24 +020015583
Willy Tarreau74ca5042013-06-11 23:12:07 +020015584If an alternate match is specified using "-m" on an ACL-specific criterion,
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030015585the matching method is simply applied to the underlying sample fetch method.
15586For example, all ACLs below are exact equivalent :
Willy Tarreau2b5285d2010-05-09 23:45:24 +020015587
Willy Tarreau74ca5042013-06-11 23:12:07 +020015588 acl short_form hdr_beg(host) www.
15589 acl alternate1 hdr_beg(host) -m beg www.
15590 acl alternate2 hdr_dom(host) -m beg www.
15591 acl alternate3 hdr(host) -m beg www.
Willy Tarreau2b5285d2010-05-09 23:45:24 +020015592
Willy Tarreau2b5285d2010-05-09 23:45:24 +020015593
Thierry FOURNIER2a06e392014-05-11 15:49:55 +020015594The table below summarizes the compatibility matrix between sample or converter
15595types and the pattern types to fetch against. It indicates for each compatible
15596combination the name of the matching method to be used, surrounded with angle
15597brackets ">" and "<" when the method is the default one and will work by
15598default without "-m".
Willy Tarreau0ba27502007-12-24 16:55:16 +010015599
Willy Tarreau74ca5042013-06-11 23:12:07 +020015600 +-------------------------------------------------+
15601 | Input sample type |
15602 +----------------------+---------+---------+---------+---------+---------+
Thierry FOURNIER2a06e392014-05-11 15:49:55 +020015603 | pattern type | boolean | integer | ip | string | binary |
Willy Tarreau74ca5042013-06-11 23:12:07 +020015604 +----------------------+---------+---------+---------+---------+---------+
15605 | none (presence only) | found | found | found | found | found |
15606 +----------------------+---------+---------+---------+---------+---------+
Thierry FOURNIER2a06e392014-05-11 15:49:55 +020015607 | none (boolean value) |> bool <| bool | | bool | |
Willy Tarreau74ca5042013-06-11 23:12:07 +020015608 +----------------------+---------+---------+---------+---------+---------+
Thierry FOURNIER2a06e392014-05-11 15:49:55 +020015609 | integer (value) | int |> int <| int | int | |
Willy Tarreau74ca5042013-06-11 23:12:07 +020015610 +----------------------+---------+---------+---------+---------+---------+
Thierry FOURNIERe3ded592013-12-06 15:36:54 +010015611 | integer (length) | len | len | len | len | len |
Willy Tarreau74ca5042013-06-11 23:12:07 +020015612 +----------------------+---------+---------+---------+---------+---------+
Thierry FOURNIER2a06e392014-05-11 15:49:55 +020015613 | IP address | | |> ip <| ip | ip |
Willy Tarreau74ca5042013-06-11 23:12:07 +020015614 +----------------------+---------+---------+---------+---------+---------+
Thierry FOURNIER2a06e392014-05-11 15:49:55 +020015615 | exact string | str | str | str |> str <| str |
Willy Tarreau74ca5042013-06-11 23:12:07 +020015616 +----------------------+---------+---------+---------+---------+---------+
Thierry FOURNIERe3ded592013-12-06 15:36:54 +010015617 | prefix | beg | beg | beg | beg | beg |
Willy Tarreau74ca5042013-06-11 23:12:07 +020015618 +----------------------+---------+---------+---------+---------+---------+
Thierry FOURNIERe3ded592013-12-06 15:36:54 +010015619 | suffix | end | end | end | end | end |
Willy Tarreau74ca5042013-06-11 23:12:07 +020015620 +----------------------+---------+---------+---------+---------+---------+
Thierry FOURNIERe3ded592013-12-06 15:36:54 +010015621 | substring | sub | sub | sub | sub | sub |
Willy Tarreau74ca5042013-06-11 23:12:07 +020015622 +----------------------+---------+---------+---------+---------+---------+
Thierry FOURNIERe3ded592013-12-06 15:36:54 +010015623 | subdir | dir | dir | dir | dir | dir |
Willy Tarreau74ca5042013-06-11 23:12:07 +020015624 +----------------------+---------+---------+---------+---------+---------+
Thierry FOURNIERe3ded592013-12-06 15:36:54 +010015625 | domain | dom | dom | dom | dom | dom |
Willy Tarreau74ca5042013-06-11 23:12:07 +020015626 +----------------------+---------+---------+---------+---------+---------+
Thierry FOURNIERe3ded592013-12-06 15:36:54 +010015627 | regex | reg | reg | reg | reg | reg |
Willy Tarreau74ca5042013-06-11 23:12:07 +020015628 +----------------------+---------+---------+---------+---------+---------+
15629 | hex block | | | | bin | bin |
15630 +----------------------+---------+---------+---------+---------+---------+
Willy Tarreau6a06a402007-07-15 20:15:28 +020015631
15632
Willy Tarreau74ca5042013-06-11 23:12:07 +0200156337.1.1. Matching booleans
15634------------------------
15635
15636In order to match a boolean, no value is needed and all values are ignored.
15637Boolean matching is used by default for all fetch methods of type "boolean".
15638When boolean matching is used, the fetched value is returned as-is, which means
15639that a boolean "true" will always match and a boolean "false" will never match.
15640
15641Boolean matching may also be enforced using "-m bool" on fetch methods which
15642return an integer value. Then, integer value 0 is converted to the boolean
15643"false" and all other values are converted to "true".
15644
Willy Tarreau6a06a402007-07-15 20:15:28 +020015645
Willy Tarreau74ca5042013-06-11 23:12:07 +0200156467.1.2. Matching integers
15647------------------------
15648
15649Integer matching applies by default to integer fetch methods. It can also be
15650enforced on boolean fetches using "-m int". In this case, "false" is converted
15651to the integer 0, and "true" is converted to the integer 1.
15652
15653Integer matching also supports integer ranges and operators. Note that integer
15654matching only applies to positive values. A range is a value expressed with a
15655lower and an upper bound separated with a colon, both of which may be omitted.
Willy Tarreau6a06a402007-07-15 20:15:28 +020015656
15657For instance, "1024:65535" is a valid range to represent a range of
15658unprivileged ports, and "1024:" would also work. "0:1023" is a valid
15659representation of privileged ports, and ":1023" would also work.
15660
Willy Tarreau62644772008-07-16 18:36:06 +020015661As a special case, some ACL functions support decimal numbers which are in fact
15662two integers separated by a dot. This is used with some version checks for
15663instance. All integer properties apply to those decimal numbers, including
15664ranges and operators.
15665
Willy Tarreau6a06a402007-07-15 20:15:28 +020015666For an easier usage, comparison operators are also supported. Note that using
Willy Tarreau0ba27502007-12-24 16:55:16 +010015667operators with ranges does not make much sense and is strongly discouraged.
15668Similarly, it does not make much sense to perform order comparisons with a set
15669of values.
Willy Tarreau6a06a402007-07-15 20:15:28 +020015670
Willy Tarreau0ba27502007-12-24 16:55:16 +010015671Available operators for integer matching are :
Willy Tarreau6a06a402007-07-15 20:15:28 +020015672
15673 eq : true if the tested value equals at least one value
15674 ge : true if the tested value is greater than or equal to at least one value
15675 gt : true if the tested value is greater than at least one value
15676 le : true if the tested value is less than or equal to at least one value
15677 lt : true if the tested value is less than at least one value
15678
Willy Tarreau0ba27502007-12-24 16:55:16 +010015679For instance, the following ACL matches any negative Content-Length header :
Willy Tarreau6a06a402007-07-15 20:15:28 +020015680
15681 acl negative-length hdr_val(content-length) lt 0
15682
Willy Tarreau62644772008-07-16 18:36:06 +020015683This one matches SSL versions between 3.0 and 3.1 (inclusive) :
15684
15685 acl sslv3 req_ssl_ver 3:3.1
15686
Willy Tarreau6a06a402007-07-15 20:15:28 +020015687
Willy Tarreau74ca5042013-06-11 23:12:07 +0200156887.1.3. Matching strings
15689-----------------------
15690
15691String matching applies to string or binary fetch methods, and exists in 6
15692different forms :
15693
15694 - exact match (-m str) : the extracted string must exactly match the
Davor Ocelice9ed2812017-12-25 17:49:28 +010015695 patterns;
Willy Tarreau74ca5042013-06-11 23:12:07 +020015696
15697 - substring match (-m sub) : the patterns are looked up inside the
Davor Ocelice9ed2812017-12-25 17:49:28 +010015698 extracted string, and the ACL matches if any of them is found inside;
Willy Tarreau74ca5042013-06-11 23:12:07 +020015699
15700 - prefix match (-m beg) : the patterns are compared with the beginning of
15701 the extracted string, and the ACL matches if any of them matches.
15702
15703 - suffix match (-m end) : the patterns are compared with the end of the
15704 extracted string, and the ACL matches if any of them matches.
15705
Baptiste Assmann33db6002016-03-06 23:32:10 +010015706 - subdir match (-m dir) : the patterns are looked up inside the extracted
Willy Tarreau74ca5042013-06-11 23:12:07 +020015707 string, delimited with slashes ("/"), and the ACL matches if any of them
15708 matches.
15709
15710 - domain match (-m dom) : the patterns are looked up inside the extracted
15711 string, delimited with dots ("."), and the ACL matches if any of them
15712 matches.
Willy Tarreau6a06a402007-07-15 20:15:28 +020015713
15714String matching applies to verbatim strings as they are passed, with the
15715exception of the backslash ("\") which makes it possible to escape some
15716characters such as the space. If the "-i" flag is passed before the first
15717string, then the matching will be performed ignoring the case. In order
15718to match the string "-i", either set it second, or pass the "--" flag
Willy Tarreau0ba27502007-12-24 16:55:16 +010015719before the first string. Same applies of course to match the string "--".
Willy Tarreau6a06a402007-07-15 20:15:28 +020015720
Mathias Weiersmuellercb250fc2019-12-02 09:43:40 +010015721Do not use string matches for binary fetches which might contain null bytes
15722(0x00), as the comparison stops at the occurrence of the first null byte.
15723Instead, convert the binary fetch to a hex string with the hex converter first.
15724
15725Example:
15726 # matches if the string <tag> is present in the binary sample
15727 acl tag_found req.payload(0,0),hex -m sub 3C7461673E
15728
Willy Tarreau6a06a402007-07-15 20:15:28 +020015729
Willy Tarreau74ca5042013-06-11 23:12:07 +0200157307.1.4. Matching regular expressions (regexes)
15731---------------------------------------------
Willy Tarreau6a06a402007-07-15 20:15:28 +020015732
15733Just like with string matching, regex matching applies to verbatim strings as
15734they are passed, with the exception of the backslash ("\") which makes it
15735possible to escape some characters such as the space. If the "-i" flag is
15736passed before the first regex, then the matching will be performed ignoring
15737the case. In order to match the string "-i", either set it second, or pass
Willy Tarreau0ba27502007-12-24 16:55:16 +010015738the "--" flag before the first string. Same principle applies of course to
15739match the string "--".
Willy Tarreau6a06a402007-07-15 20:15:28 +020015740
15741
Willy Tarreau74ca5042013-06-11 23:12:07 +0200157427.1.5. Matching arbitrary data blocks
15743-------------------------------------
15744
15745It is possible to match some extracted samples against a binary block which may
15746not safely be represented as a string. For this, the patterns must be passed as
15747a series of hexadecimal digits in an even number, when the match method is set
15748to binary. Each sequence of two digits will represent a byte. The hexadecimal
15749digits may be used upper or lower case.
15750
15751Example :
15752 # match "Hello\n" in the input stream (\x48 \x65 \x6c \x6c \x6f \x0a)
15753 acl hello payload(0,6) -m bin 48656c6c6f0a
15754
15755
157567.1.6. Matching IPv4 and IPv6 addresses
15757---------------------------------------
Willy Tarreau6a06a402007-07-15 20:15:28 +020015758
15759IPv4 addresses values can be specified either as plain addresses or with a
15760netmask appended, in which case the IPv4 address matches whenever it is
15761within the network. Plain addresses may also be replaced with a resolvable
Willy Tarreaud2a4aa22008-01-31 15:28:22 +010015762host name, but this practice is generally discouraged as it makes it more
Willy Tarreau0ba27502007-12-24 16:55:16 +010015763difficult to read and debug configurations. If hostnames are used, you should
15764at least ensure that they are present in /etc/hosts so that the configuration
15765does not depend on any random DNS match at the moment the configuration is
15766parsed.
Willy Tarreau6a06a402007-07-15 20:15:28 +020015767
Daniel Schnellereba56342016-04-13 00:26:52 +020015768The dotted IPv4 address notation is supported in both regular as well as the
15769abbreviated form with all-0-octets omitted:
15770
15771 +------------------+------------------+------------------+
15772 | Example 1 | Example 2 | Example 3 |
15773 +------------------+------------------+------------------+
15774 | 192.168.0.1 | 10.0.0.12 | 127.0.0.1 |
15775 | 192.168.1 | 10.12 | 127.1 |
15776 | 192.168.0.1/22 | 10.0.0.12/8 | 127.0.0.1/8 |
15777 | 192.168.1/22 | 10.12/8 | 127.1/8 |
15778 +------------------+------------------+------------------+
15779
15780Notice that this is different from RFC 4632 CIDR address notation in which
15781192.168.42/24 would be equivalent to 192.168.42.0/24.
15782
Willy Tarreauceb4ac92012-04-28 00:41:46 +020015783IPv6 may be entered in their usual form, with or without a netmask appended.
15784Only bit counts are accepted for IPv6 netmasks. In order to avoid any risk of
15785trouble with randomly resolved IP addresses, host names are never allowed in
15786IPv6 patterns.
15787
15788HAProxy is also able to match IPv4 addresses with IPv6 addresses in the
15789following situations :
15790 - tested address is IPv4, pattern address is IPv4, the match applies
15791 in IPv4 using the supplied mask if any.
15792 - tested address is IPv6, pattern address is IPv6, the match applies
15793 in IPv6 using the supplied mask if any.
15794 - tested address is IPv6, pattern address is IPv4, the match applies in IPv4
15795 using the pattern's mask if the IPv6 address matches with 2002:IPV4::,
15796 ::IPV4 or ::ffff:IPV4, otherwise it fails.
15797 - tested address is IPv4, pattern address is IPv6, the IPv4 address is first
15798 converted to IPv6 by prefixing ::ffff: in front of it, then the match is
15799 applied in IPv6 using the supplied IPv6 mask.
15800
Willy Tarreau74ca5042013-06-11 23:12:07 +020015801
158027.2. Using ACLs to form conditions
15803----------------------------------
15804
15805Some actions are only performed upon a valid condition. A condition is a
15806combination of ACLs with operators. 3 operators are supported :
15807
15808 - AND (implicit)
15809 - OR (explicit with the "or" keyword or the "||" operator)
15810 - Negation with the exclamation mark ("!")
Willy Tarreau6a06a402007-07-15 20:15:28 +020015811
Willy Tarreau74ca5042013-06-11 23:12:07 +020015812A condition is formed as a disjunctive form:
Willy Tarreau6a06a402007-07-15 20:15:28 +020015813
Willy Tarreau74ca5042013-06-11 23:12:07 +020015814 [!]acl1 [!]acl2 ... [!]acln { or [!]acl1 [!]acl2 ... [!]acln } ...
Willy Tarreaubef91e72013-03-31 23:14:46 +020015815
Willy Tarreau74ca5042013-06-11 23:12:07 +020015816Such conditions are generally used after an "if" or "unless" statement,
15817indicating when the condition will trigger the action.
Willy Tarreaubef91e72013-03-31 23:14:46 +020015818
Willy Tarreau74ca5042013-06-11 23:12:07 +020015819For instance, to block HTTP requests to the "*" URL with methods other than
15820"OPTIONS", as well as POST requests without content-length, and GET or HEAD
15821requests with a content-length greater than 0, and finally every request which
15822is not either GET/HEAD/POST/OPTIONS !
15823
15824 acl missing_cl hdr_cnt(Content-length) eq 0
Jarno Huuskonen84c51ec2017-04-03 14:20:34 +030015825 http-request deny if HTTP_URL_STAR !METH_OPTIONS || METH_POST missing_cl
15826 http-request deny if METH_GET HTTP_CONTENT
15827 http-request deny unless METH_GET or METH_POST or METH_OPTIONS
Willy Tarreau74ca5042013-06-11 23:12:07 +020015828
15829To select a different backend for requests to static contents on the "www" site
15830and to every request on the "img", "video", "download" and "ftp" hosts :
15831
15832 acl url_static path_beg /static /images /img /css
15833 acl url_static path_end .gif .png .jpg .css .js
15834 acl host_www hdr_beg(host) -i www
15835 acl host_static hdr_beg(host) -i img. video. download. ftp.
15836
Davor Ocelice9ed2812017-12-25 17:49:28 +010015837 # now use backend "static" for all static-only hosts, and for static URLs
Willy Tarreau74ca5042013-06-11 23:12:07 +020015838 # of host "www". Use backend "www" for the rest.
15839 use_backend static if host_static or host_www url_static
15840 use_backend www if host_www
15841
15842It is also possible to form rules using "anonymous ACLs". Those are unnamed ACL
15843expressions that are built on the fly without needing to be declared. They must
15844be enclosed between braces, with a space before and after each brace (because
15845the braces must be seen as independent words). Example :
15846
15847 The following rule :
15848
15849 acl missing_cl hdr_cnt(Content-length) eq 0
Jarno Huuskonen84c51ec2017-04-03 14:20:34 +030015850 http-request deny if METH_POST missing_cl
Willy Tarreau74ca5042013-06-11 23:12:07 +020015851
15852 Can also be written that way :
15853
Jarno Huuskonen84c51ec2017-04-03 14:20:34 +030015854 http-request deny if METH_POST { hdr_cnt(Content-length) eq 0 }
Willy Tarreau74ca5042013-06-11 23:12:07 +020015855
15856It is generally not recommended to use this construct because it's a lot easier
15857to leave errors in the configuration when written that way. However, for very
15858simple rules matching only one source IP address for instance, it can make more
15859sense to use them than to declare ACLs with random names. Another example of
15860good use is the following :
15861
15862 With named ACLs :
15863
15864 acl site_dead nbsrv(dynamic) lt 2
15865 acl site_dead nbsrv(static) lt 2
15866 monitor fail if site_dead
15867
15868 With anonymous ACLs :
15869
15870 monitor fail if { nbsrv(dynamic) lt 2 } || { nbsrv(static) lt 2 }
15871
Jarno Huuskonen84c51ec2017-04-03 14:20:34 +030015872See section 4.2 for detailed help on the "http-request deny" and "use_backend"
15873keywords.
Willy Tarreau74ca5042013-06-11 23:12:07 +020015874
15875
158767.3. Fetching samples
15877---------------------
15878
15879Historically, sample fetch methods were only used to retrieve data to match
15880against patterns using ACLs. With the arrival of stick-tables, a new class of
15881sample fetch methods was created, most often sharing the same syntax as their
15882ACL counterpart. These sample fetch methods are also known as "fetches". As
15883of now, ACLs and fetches have converged. All ACL fetch methods have been made
15884available as fetch methods, and ACLs may use any sample fetch method as well.
15885
15886This section details all available sample fetch methods and their output type.
15887Some sample fetch methods have deprecated aliases that are used to maintain
15888compatibility with existing configurations. They are then explicitly marked as
15889deprecated and should not be used in new setups.
15890
15891The ACL derivatives are also indicated when available, with their respective
15892matching methods. These ones all have a well defined default pattern matching
15893method, so it is never necessary (though allowed) to pass the "-m" option to
15894indicate how the sample will be matched using ACLs.
15895
15896As indicated in the sample type versus matching compatibility matrix above,
15897when using a generic sample fetch method in an ACL, the "-m" option is
15898mandatory unless the sample type is one of boolean, integer, IPv4 or IPv6. When
15899the same keyword exists as an ACL keyword and as a standard fetch method, the
15900ACL engine will automatically pick the ACL-only one by default.
15901
15902Some of these keywords support one or multiple mandatory arguments, and one or
15903multiple optional arguments. These arguments are strongly typed and are checked
15904when the configuration is parsed so that there is no risk of running with an
Davor Ocelice9ed2812017-12-25 17:49:28 +010015905incorrect argument (e.g. an unresolved backend name). Fetch function arguments
15906are passed between parenthesis and are delimited by commas. When an argument
Willy Tarreau74ca5042013-06-11 23:12:07 +020015907is optional, it will be indicated below between square brackets ('[ ]'). When
15908all arguments are optional, the parenthesis may be omitted.
15909
15910Thus, the syntax of a standard sample fetch method is one of the following :
15911 - name
15912 - name(arg1)
15913 - name(arg1,arg2)
15914
Thierry FOURNIER060762e2014-04-23 13:29:15 +020015915
159167.3.1. Converters
15917-----------------
15918
Willy Tarreaue6b11e42013-11-26 19:02:32 +010015919Sample fetch methods may be combined with transformations to be applied on top
15920of the fetched sample (also called "converters"). These combinations form what
15921is called "sample expressions" and the result is a "sample". Initially this
15922was only supported by "stick on" and "stick store-request" directives but this
Davor Ocelice9ed2812017-12-25 17:49:28 +010015923has now be extended to all places where samples may be used (ACLs, log-format,
Willy Tarreaue6b11e42013-11-26 19:02:32 +010015924unique-id-format, add-header, ...).
15925
15926These transformations are enumerated as a series of specific keywords after the
15927sample fetch method. These keywords may equally be appended immediately after
15928the fetch keyword's argument, delimited by a comma. These keywords can also
Davor Ocelice9ed2812017-12-25 17:49:28 +010015929support some arguments (e.g. a netmask) which must be passed in parenthesis.
Willy Tarreau0ba27502007-12-24 16:55:16 +010015930
Willy Tarreau97707872015-01-27 15:12:13 +010015931A certain category of converters are bitwise and arithmetic operators which
15932support performing basic operations on integers. Some bitwise operations are
15933supported (and, or, xor, cpl) and some arithmetic operations are supported
15934(add, sub, mul, div, mod, neg). Some comparators are provided (odd, even, not,
15935bool) which make it possible to report a match without having to write an ACL.
15936
Willy Tarreau74ca5042013-06-11 23:12:07 +020015937The currently available list of transformation keywords include :
Willy Tarreau0ba27502007-12-24 16:55:16 +010015938
Ben Shillitof25e8e52016-12-02 14:25:37 +00001593951d.single(<prop>[,<prop>*])
15940 Returns values for the properties requested as a string, where values are
15941 separated by the delimiter specified with "51degrees-property-separator".
15942 The device is identified using the User-Agent header passed to the
15943 converter. The function can be passed up to five property names, and if a
15944 property name can't be found, the value "NoData" is returned.
15945
15946 Example :
Davor Ocelice9ed2812017-12-25 17:49:28 +010015947 # Here the header "X-51D-DeviceTypeMobileTablet" is added to the request,
15948 # containing values for the three properties requested by using the
Ben Shillitof25e8e52016-12-02 14:25:37 +000015949 # User-Agent passed to the converter.
15950 frontend http-in
15951 bind *:8081
15952 default_backend servers
15953 http-request set-header X-51D-DeviceTypeMobileTablet \
15954 %[req.fhdr(User-Agent),51d.single(DeviceType,IsMobile,IsTablet)]
15955
Willy Tarreau97707872015-01-27 15:12:13 +010015956add(<value>)
Thierry FOURNIER07ee64e2015-07-06 23:43:03 +020015957 Adds <value> to the input value of type signed integer, and returns the
Thierry FOURNIER5d86fae2015-07-07 21:10:16 +020015958 result as a signed integer. <value> can be a numeric value or a variable
Daniel Schneller0b547052016-03-21 20:46:57 +010015959 name. The name of the variable starts with an indication about its scope. The
15960 scopes allowed are:
Christopher Fauletff2613e2016-11-09 11:36:17 +010015961 "proc" : the variable is shared with the whole process
Daniel Schneller0b547052016-03-21 20:46:57 +010015962 "sess" : the variable is shared with the whole session
15963 "txn" : the variable is shared with the transaction (request and response)
15964 "req" : the variable is shared only during request processing
15965 "res" : the variable is shared only during response processing
Davor Ocelice9ed2812017-12-25 17:49:28 +010015966 This prefix is followed by a name. The separator is a '.'. The name may only
Christopher Fauletb71557a2016-10-31 10:49:03 +010015967 contain characters 'a-z', 'A-Z', '0-9', '.' and '_'.
Willy Tarreau97707872015-01-27 15:12:13 +010015968
Nenad Merdanovicc31499d2019-03-23 11:00:32 +010015969aes_gcm_dec(<bits>,<nonce>,<key>,<aead_tag>)
15970 Decrypts the raw byte input using the AES128-GCM, AES192-GCM or
15971 AES256-GCM algorithm, depending on the <bits> parameter. All other parameters
15972 need to be base64 encoded and the returned result is in raw byte format.
15973 If the <aead_tag> validation fails, the converter doesn't return any data.
15974 The <nonce>, <key> and <aead_tag> can either be strings or variables. This
15975 converter requires at least OpenSSL 1.0.1.
15976
15977 Example:
15978 http-response set-header X-Decrypted-Text %[var(txn.enc),\
15979 aes_gcm_dec(128,txn.nonce,Zm9vb2Zvb29mb29wZm9vbw==,txn.aead_tag)]
15980
Willy Tarreau97707872015-01-27 15:12:13 +010015981and(<value>)
Thierry FOURNIER07ee64e2015-07-06 23:43:03 +020015982 Performs a bitwise "AND" between <value> and the input value of type signed
Thierry FOURNIER5d86fae2015-07-07 21:10:16 +020015983 integer, and returns the result as an signed integer. <value> can be a
Daniel Schneller0b547052016-03-21 20:46:57 +010015984 numeric value or a variable name. The name of the variable starts with an
15985 indication about its scope. The scopes allowed are:
Christopher Fauletff2613e2016-11-09 11:36:17 +010015986 "proc" : the variable is shared with the whole process
Daniel Schneller0b547052016-03-21 20:46:57 +010015987 "sess" : the variable is shared with the whole session
15988 "txn" : the variable is shared with the transaction (request and response)
15989 "req" : the variable is shared only during request processing
15990 "res" : the variable is shared only during response processing
Davor Ocelice9ed2812017-12-25 17:49:28 +010015991 This prefix is followed by a name. The separator is a '.'. The name may only
Christopher Fauletb71557a2016-10-31 10:49:03 +010015992 contain characters 'a-z', 'A-Z', '0-9', '.' and '_'.
Willy Tarreau97707872015-01-27 15:12:13 +010015993
Holger Just1bfc24b2017-05-06 00:56:53 +020015994b64dec
15995 Converts (decodes) a base64 encoded input string to its binary
15996 representation. It performs the inverse operation of base64().
Moemen MHEDHBI92f7d432021-04-01 20:53:59 +020015997 For base64url("URL and Filename Safe Alphabet" (RFC 4648)) variant
15998 see "ub64dec".
Holger Just1bfc24b2017-05-06 00:56:53 +020015999
Emeric Brun53d1a982014-04-30 18:21:37 +020016000base64
16001 Converts a binary input sample to a base64 string. It is used to log or
Davor Ocelice9ed2812017-12-25 17:49:28 +010016002 transfer binary content in a way that can be reliably transferred (e.g.
Moemen MHEDHBI92f7d432021-04-01 20:53:59 +020016003 an SSL ID can be copied in a header). For base64url("URL and Filename
16004 Safe Alphabet" (RFC 4648)) variant see "ub64enc".
Emeric Brun53d1a982014-04-30 18:21:37 +020016005
Willy Tarreau97707872015-01-27 15:12:13 +010016006bool
Thierry FOURNIER07ee64e2015-07-06 23:43:03 +020016007 Returns a boolean TRUE if the input value of type signed integer is
Willy Tarreau97707872015-01-27 15:12:13 +010016008 non-null, otherwise returns FALSE. Used in conjunction with and(), it can be
Davor Ocelice9ed2812017-12-25 17:49:28 +010016009 used to report true/false for bit testing on input values (e.g. verify the
Willy Tarreau97707872015-01-27 15:12:13 +010016010 presence of a flag).
16011
Emeric Brun54c4ac82014-11-03 15:32:43 +010016012bytes(<offset>[,<length>])
16013 Extracts some bytes from an input binary sample. The result is a binary
16014 sample starting at an offset (in bytes) of the original sample and
Tim Düsterhus4896c442016-11-29 02:15:19 +010016015 optionally truncated at the given length.
Emeric Brun54c4ac82014-11-03 15:32:43 +010016016
Willy Tarreau280f42b2018-02-19 15:34:12 +010016017concat([<start>],[<var>],[<end>])
16018 Concatenates up to 3 fields after the current sample which is then turned to
16019 a string. The first one, <start>, is a constant string, that will be appended
16020 immediately after the existing sample. It may be omitted if not used. The
16021 second one, <var>, is a variable name. The variable will be looked up, its
16022 contents converted to a string, and it will be appended immediately after the
16023 <first> part. If the variable is not found, nothing is appended. It may be
16024 omitted as well. The third field, <end> is a constant string that will be
16025 appended after the variable. It may also be omitted. Together, these elements
16026 allow to concatenate variables with delimiters to an existing set of
16027 variables. This can be used to build new variables made of a succession of
Willy Tarreauef21fac2020-02-14 13:37:20 +010016028 other variables, such as colon-delimited values. If commas or closing
Daniel Corbett67a82712020-07-06 23:01:19 -040016029 parenthesis are needed as delimiters, they must be protected by quotes or
Willy Tarreauef21fac2020-02-14 13:37:20 +010016030 backslashes, themselves protected so that they are not stripped by the first
16031 level parser. See examples below.
Willy Tarreau280f42b2018-02-19 15:34:12 +010016032
16033 Example:
16034 tcp-request session set-var(sess.src) src
16035 tcp-request session set-var(sess.dn) ssl_c_s_dn
16036 tcp-request session set-var(txn.sig) str(),concat(<ip=,sess.ip,>),concat(<dn=,sess.dn,>)
Willy Tarreauef21fac2020-02-14 13:37:20 +010016037 tcp-request session set-var(txn.ipport) "str(),concat('addr=(',sess.ip),concat(',',sess.port,')')"
Willy Tarreau280f42b2018-02-19 15:34:12 +010016038 http-request set-header x-hap-sig %[var(txn.sig)]
16039
Willy Tarreau97707872015-01-27 15:12:13 +010016040cpl
Thierry FOURNIER07ee64e2015-07-06 23:43:03 +020016041 Takes the input value of type signed integer, applies a ones-complement
16042 (flips all bits) and returns the result as an signed integer.
Willy Tarreau97707872015-01-27 15:12:13 +010016043
Willy Tarreau80599772015-01-20 19:35:24 +010016044crc32([<avalanche>])
16045 Hashes a binary input sample into an unsigned 32-bit quantity using the CRC32
16046 hash function. Optionally, it is possible to apply a full avalanche hash
16047 function to the output if the optional <avalanche> argument equals 1. This
16048 converter uses the same functions as used by the various hash-based load
16049 balancing algorithms, so it will provide exactly the same results. It is
16050 provided for compatibility with other software which want a CRC32 to be
16051 computed on some input keys, so it follows the most common implementation as
16052 found in Ethernet, Gzip, PNG, etc... It is slower than the other algorithms
16053 but may provide a better or at least less predictable distribution. It must
16054 not be used for security purposes as a 32-bit hash is trivial to break. See
Emmanuel Hocdet50791a72018-03-21 11:19:01 +010016055 also "djb2", "sdbm", "wt6", "crc32c" and the "hash-type" directive.
16056
16057crc32c([<avalanche>])
16058 Hashes a binary input sample into an unsigned 32-bit quantity using the CRC32C
16059 hash function. Optionally, it is possible to apply a full avalanche hash
16060 function to the output if the optional <avalanche> argument equals 1. This
16061 converter uses the same functions as described in RFC4960, Appendix B [8].
16062 It is provided for compatibility with other software which want a CRC32C to be
16063 computed on some input keys. It is slower than the other algorithms and it must
16064 not be used for security purposes as a 32-bit hash is trivial to break. See
16065 also "djb2", "sdbm", "wt6", "crc32" and the "hash-type" directive.
Willy Tarreau80599772015-01-20 19:35:24 +010016066
Christopher Fauletea159d62020-04-01 16:21:44 +020016067cut_crlf
16068 Cuts the string representation of the input sample on the first carriage
16069 return ('\r') or newline ('\n') character found. Only the string length is
16070 updated.
16071
David Carlier29b3ca32015-09-25 14:09:21 +010016072da-csv-conv(<prop>[,<prop>*])
David Carlier4542b102015-06-01 13:54:29 +020016073 Asks the DeviceAtlas converter to identify the User Agent string passed on
16074 input, and to emit a string made of the concatenation of the properties
16075 enumerated in argument, delimited by the separator defined by the global
16076 keyword "deviceatlas-property-separator", or by default the pipe character
Daniel Corbett9f0843f2021-05-08 10:50:37 -040016077 ('|'). There's a limit of 12 different properties imposed by the HAProxy
David Carlier4542b102015-06-01 13:54:29 +020016078 configuration language.
16079
16080 Example:
16081 frontend www
Cyril Bonté307ee1e2015-09-28 23:16:06 +020016082 bind *:8881
16083 default_backend servers
David Carlier840b0242016-03-16 10:09:55 +000016084 http-request set-header X-DeviceAtlas-Data %[req.fhdr(User-Agent),da-csv(primaryHardwareType,osName,osVersion,browserName,browserVersion,browserRenderingEngine)]
David Carlier4542b102015-06-01 13:54:29 +020016085
Willy Tarreau0851fd52019-12-17 10:07:25 +010016086debug([<prefix][,<destination>])
16087 This converter is used as debug tool. It takes a capture of the input sample
16088 and sends it to event sink <destination>, which may designate a ring buffer
16089 such as "buf0", as well as "stdout", or "stderr". Available sinks may be
16090 checked at run time by issuing "show events" on the CLI. When not specified,
16091 the output will be "buf0", which may be consulted via the CLI's "show events"
16092 command. An optional prefix <prefix> may be passed to help distinguish
16093 outputs from multiple expressions. It will then appear before the colon in
16094 the output message. The input sample is passed as-is on the output, so that
16095 it is safe to insert the debug converter anywhere in a chain, even with non-
16096 printable sample types.
16097
16098 Example:
16099 tcp-request connection track-sc0 src,debug(track-sc)
Thierry FOURNIER9687c772015-05-07 15:46:29 +020016100
Patrick Gansterer8e366512020-04-22 16:47:57 +020016101digest(<algorithm>)
16102 Converts a binary input sample to a message digest. The result is a binary
16103 sample. The <algorithm> must be an OpenSSL message digest name (e.g. sha256).
16104
Daniel Corbett9f0843f2021-05-08 10:50:37 -040016105 Please note that this converter is only available when HAProxy has been
Patrick Gansterer8e366512020-04-22 16:47:57 +020016106 compiled with USE_OPENSSL.
16107
Willy Tarreau97707872015-01-27 15:12:13 +010016108div(<value>)
Thierry FOURNIER07ee64e2015-07-06 23:43:03 +020016109 Divides the input value of type signed integer by <value>, and returns the
16110 result as an signed integer. If <value> is null, the largest unsigned
Thierry FOURNIER5d86fae2015-07-07 21:10:16 +020016111 integer is returned (typically 2^63-1). <value> can be a numeric value or a
Daniel Schneller0b547052016-03-21 20:46:57 +010016112 variable name. The name of the variable starts with an indication about its
16113 scope. The scopes allowed are:
Christopher Fauletff2613e2016-11-09 11:36:17 +010016114 "proc" : the variable is shared with the whole process
Daniel Schneller0b547052016-03-21 20:46:57 +010016115 "sess" : the variable is shared with the whole session
16116 "txn" : the variable is shared with the transaction (request and response)
16117 "req" : the variable is shared only during request processing
16118 "res" : the variable is shared only during response processing
Davor Ocelice9ed2812017-12-25 17:49:28 +010016119 This prefix is followed by a name. The separator is a '.'. The name may only
Christopher Fauletb71557a2016-10-31 10:49:03 +010016120 contain characters 'a-z', 'A-Z', '0-9', '.' and '_'.
Willy Tarreau97707872015-01-27 15:12:13 +010016121
Willy Tarreau23ec4ca2014-07-15 20:15:37 +020016122djb2([<avalanche>])
16123 Hashes a binary input sample into an unsigned 32-bit quantity using the DJB2
16124 hash function. Optionally, it is possible to apply a full avalanche hash
16125 function to the output if the optional <avalanche> argument equals 1. This
16126 converter uses the same functions as used by the various hash-based load
16127 balancing algorithms, so it will provide exactly the same results. It is
16128 mostly intended for debugging, but can be used as a stick-table entry to
16129 collect rough statistics. It must not be used for security purposes as a
Emmanuel Hocdet50791a72018-03-21 11:19:01 +010016130 32-bit hash is trivial to break. See also "crc32", "sdbm", "wt6", "crc32c",
16131 and the "hash-type" directive.
Willy Tarreau23ec4ca2014-07-15 20:15:37 +020016132
Willy Tarreau97707872015-01-27 15:12:13 +010016133even
Thierry FOURNIER07ee64e2015-07-06 23:43:03 +020016134 Returns a boolean TRUE if the input value of type signed integer is even
Willy Tarreau97707872015-01-27 15:12:13 +010016135 otherwise returns FALSE. It is functionally equivalent to "not,and(1),bool".
16136
Marcin Deranek9631a282018-04-16 14:30:46 +020016137field(<index>,<delimiters>[,<count>])
16138 Extracts the substring at the given index counting from the beginning
16139 (positive index) or from the end (negative index) considering given delimiters
16140 from an input string. Indexes start at 1 or -1 and delimiters are a string
16141 formatted list of chars. Optionally you can specify <count> of fields to
16142 extract (default: 1). Value of 0 indicates extraction of all remaining
16143 fields.
16144
16145 Example :
16146 str(f1_f2_f3__f5),field(5,_) # f5
16147 str(f1_f2_f3__f5),field(2,_,0) # f2_f3__f5
16148 str(f1_f2_f3__f5),field(2,_,2) # f2_f3
16149 str(f1_f2_f3__f5),field(-2,_,3) # f2_f3_
16150 str(f1_f2_f3__f5),field(-3,_,0) # f1_f2_f3
Emeric Brunf399b0d2014-11-03 17:07:03 +010016151
Baptiste Assmanne138dda2020-10-22 15:39:03 +020016152fix_is_valid
16153 Parses a binary payload and performs sanity checks regarding FIX (Financial
16154 Information eXchange):
16155
16156 - checks that all tag IDs and values are not empty and the tags IDs are well
16157 numeric
Ilya Shipitsin2272d8a2020-12-21 01:22:40 +050016158 - checks the BeginString tag is the first tag with a valid FIX version
Baptiste Assmanne138dda2020-10-22 15:39:03 +020016159 - checks the BodyLength tag is the second one with the right body length
Christopher Fauleted4bef72021-03-18 17:40:56 +010016160 - checks the MsgType tag is the third tag.
Baptiste Assmanne138dda2020-10-22 15:39:03 +020016161 - checks that last tag in the message is the CheckSum tag with a valid
16162 checksum
16163
16164 Due to current HAProxy design, only the first message sent by the client and
16165 the server can be parsed.
16166
16167 This converter returns a boolean, true if the payload contains a valid FIX
16168 message, false if not.
16169
16170 See also the fix_tag_value converter.
16171
16172 Example:
16173 tcp-request inspect-delay 10s
16174 tcp-request content reject unless { req.payload(0,0),fix_is_valid }
16175
16176fix_tag_value(<tag>)
16177 Parses a FIX (Financial Information eXchange) message and extracts the value
16178 from the tag <tag>. <tag> can be a string or an integer pointing to the
16179 desired tag. Any integer value is accepted, but only the following strings
16180 are translated into their integer equivalent: BeginString, BodyLength,
Daniel Corbettbefef702021-03-09 23:00:34 -050016181 MsgType, SenderCompID, TargetCompID, CheckSum. More tag names can be easily
Baptiste Assmanne138dda2020-10-22 15:39:03 +020016182 added.
16183
16184 Due to current HAProxy design, only the first message sent by the client and
16185 the server can be parsed. No message validation is performed by this
16186 converter. It is highly recommended to validate the message first using
16187 fix_is_valid converter.
16188
16189 See also the fix_is_valid converter.
16190
16191 Example:
16192 tcp-request inspect-delay 10s
16193 tcp-request content reject unless { req.payload(0,0),fix_is_valid }
16194 # MsgType tag ID is 35, so both lines below will return the same content
16195 tcp-request content set-var(txn.foo) req.payload(0,0),fix_tag_value(35)
16196 tcp-request content set-var(txn.bar) req.payload(0,0),fix_tag_value(MsgType)
16197
Thierry FOURNIER060762e2014-04-23 13:29:15 +020016198hex
Davor Ocelice9ed2812017-12-25 17:49:28 +010016199 Converts a binary input sample to a hex string containing two hex digits per
Thierry FOURNIER060762e2014-04-23 13:29:15 +020016200 input byte. It is used to log or transfer hex dumps of some binary input data
Davor Ocelice9ed2812017-12-25 17:49:28 +010016201 in a way that can be reliably transferred (e.g. an SSL ID can be copied in a
Thierry FOURNIER060762e2014-04-23 13:29:15 +020016202 header).
Thierry FOURNIER2f49d6d2014-03-12 15:01:52 +010016203
Dragan Dosen3f957b22017-10-24 09:27:34 +020016204hex2i
16205 Converts a hex string containing two hex digits per input byte to an
John Roeslerfb2fce12019-07-10 15:45:51 -050016206 integer. If the input value cannot be converted, then zero is returned.
Dragan Dosen3f957b22017-10-24 09:27:34 +020016207
Christopher Faulet4ccc12f2020-04-01 09:08:32 +020016208htonl
16209 Converts the input integer value to its 32-bit binary representation in the
16210 network byte order. Because sample fetches own signed 64-bit integer, when
16211 this converter is used, the input integer value is first casted to an
16212 unsigned 32-bit integer.
16213
Tim Duesterhusa3082092021-01-21 17:40:49 +010016214hmac(<algorithm>,<key>)
Patrick Gansterer8e366512020-04-22 16:47:57 +020016215 Converts a binary input sample to a message authentication code with the given
16216 key. The result is a binary sample. The <algorithm> must be one of the
16217 registered OpenSSL message digest names (e.g. sha256). The <key> parameter must
16218 be base64 encoded and can either be a string or a variable.
16219
Daniel Corbett9f0843f2021-05-08 10:50:37 -040016220 Please note that this converter is only available when HAProxy has been
Patrick Gansterer8e366512020-04-22 16:47:57 +020016221 compiled with USE_OPENSSL.
16222
Cyril Bonté6bcd1822019-11-05 23:13:59 +010016223http_date([<offset],[<unit>])
Thierry FOURNIER060762e2014-04-23 13:29:15 +020016224 Converts an integer supposed to contain a date since epoch to a string
16225 representing this date in a format suitable for use in HTTP header fields. If
Damien Claisseae6f1252019-10-30 15:57:28 +000016226 an offset value is specified, then it is added to the date before the
16227 conversion is operated. This is particularly useful to emit Date header fields,
16228 Expires values in responses when combined with a positive offset, or
16229 Last-Modified values when the offset is negative.
16230 If a unit value is specified, then consider the timestamp as either
16231 "s" for seconds (default behavior), "ms" for milliseconds, or "us" for
16232 microseconds since epoch. Offset is assumed to have the same unit as
16233 input timestamp.
Willy Tarreau74ca5042013-06-11 23:12:07 +020016234
Tim Duesterhus3943e4f2020-09-11 14:25:23 +020016235iif(<true>,<false>)
16236 Returns the <true> string if the input value is true. Returns the <false>
16237 string otherwise.
16238
16239 Example:
Tim Duesterhus870713b2020-09-11 17:13:12 +020016240 http-request set-header x-forwarded-proto %[ssl_fc,iif(https,http)]
Tim Duesterhus3943e4f2020-09-11 14:25:23 +020016241
Willy Tarreaud9f316a2014-07-10 14:03:38 +020016242in_table(<table>)
16243 Uses the string representation of the input sample to perform a look up in
16244 the specified table. If the key is not found in the table, a boolean false
16245 is returned. Otherwise a boolean true is returned. This can be used to verify
Davor Ocelice9ed2812017-12-25 17:49:28 +010016246 the presence of a certain key in a table tracking some elements (e.g. whether
Willy Tarreaud9f316a2014-07-10 14:03:38 +020016247 or not a source IP address or an Authorization header was already seen).
16248
Tim Duesterhusa3082092021-01-21 17:40:49 +010016249ipmask(<mask4>,[<mask6>])
Tim Duesterhus1478aa72018-01-25 16:24:51 +010016250 Apply a mask to an IP address, and use the result for lookups and storage.
Willy Tarreauffcb2e42014-07-10 16:29:08 +020016251 This can be used to make all hosts within a certain mask to share the same
Tim Duesterhus1478aa72018-01-25 16:24:51 +010016252 table entries and as such use the same server. The mask4 can be passed in
16253 dotted form (e.g. 255.255.255.0) or in CIDR form (e.g. 24). The mask6 can
16254 be passed in quadruplet form (e.g. ffff:ffff::) or in CIDR form (e.g. 64).
16255 If no mask6 is given IPv6 addresses will fail to convert for backwards
16256 compatibility reasons.
Willy Tarreauffcb2e42014-07-10 16:29:08 +020016257
Thierry FOURNIER317e1c42014-08-12 10:20:47 +020016258json([<input-code>])
Davor Ocelice9ed2812017-12-25 17:49:28 +010016259 Escapes the input string and produces an ASCII output string ready to use as a
Thierry FOURNIER317e1c42014-08-12 10:20:47 +020016260 JSON string. The converter tries to decode the input string according to the
Herve COMMOWICK8dfe8632016-08-05 12:01:20 +020016261 <input-code> parameter. It can be "ascii", "utf8", "utf8s", "utf8p" or
Thierry FOURNIER317e1c42014-08-12 10:20:47 +020016262 "utf8ps". The "ascii" decoder never fails. The "utf8" decoder detects 3 types
16263 of errors:
16264 - bad UTF-8 sequence (lone continuation byte, bad number of continuation
16265 bytes, ...)
16266 - invalid range (the decoded value is within a UTF-8 prohibited range),
16267 - code overlong (the value is encoded with more bytes than necessary).
16268
16269 The UTF-8 JSON encoding can produce a "too long value" error when the UTF-8
16270 character is greater than 0xffff because the JSON string escape specification
16271 only authorizes 4 hex digits for the value encoding. The UTF-8 decoder exists
16272 in 4 variants designated by a combination of two suffix letters : "p" for
16273 "permissive" and "s" for "silently ignore". The behaviors of the decoders
16274 are :
Davor Ocelice9ed2812017-12-25 17:49:28 +010016275 - "ascii" : never fails;
16276 - "utf8" : fails on any detected errors;
16277 - "utf8s" : never fails, but removes characters corresponding to errors;
Thierry FOURNIER317e1c42014-08-12 10:20:47 +020016278 - "utf8p" : accepts and fixes the overlong errors, but fails on any other
Davor Ocelice9ed2812017-12-25 17:49:28 +010016279 error;
Thierry FOURNIER317e1c42014-08-12 10:20:47 +020016280 - "utf8ps" : never fails, accepts and fixes the overlong errors, but removes
16281 characters corresponding to the other errors.
16282
16283 This converter is particularly useful for building properly escaped JSON for
Davor Ocelice9ed2812017-12-25 17:49:28 +010016284 logging to servers which consume JSON-formatted traffic logs.
Thierry FOURNIER317e1c42014-08-12 10:20:47 +020016285
16286 Example:
Thierry FOURNIER317e1c42014-08-12 10:20:47 +020016287 capture request header Host len 15
Herve COMMOWICK8dfe8632016-08-05 12:01:20 +020016288 capture request header user-agent len 150
16289 log-format '{"ip":"%[src]","user-agent":"%[capture.req.hdr(1),json(utf8s)]"}'
Thierry FOURNIER317e1c42014-08-12 10:20:47 +020016290
16291 Input request from client 127.0.0.1:
16292 GET / HTTP/1.0
16293 User-Agent: Very "Ugly" UA 1/2
16294
16295 Output log:
16296 {"ip":"127.0.0.1","user-agent":"Very \"Ugly\" UA 1\/2"}
16297
Alex51c8ad42021-04-15 16:45:15 +020016298json_query(<json_path>,[<output_type>])
16299 The json_query converter supports the JSON types string, boolean and
16300 number. Floating point numbers will be returned as a string. By
16301 specifying the output_type 'int' the value will be converted to an
16302 Integer. If conversion is not possible the json_query converter fails.
16303
16304 <json_path> must be a valid JSON Path string as defined in
16305 https://datatracker.ietf.org/doc/draft-ietf-jsonpath-base/
16306
16307 Example:
16308 # get a integer value from the request body
16309 # "{"integer":4}" => 5
16310 http-request set-var(txn.pay_int) req.body,json_query('$.integer','int'),add(1)
16311
16312 # get a key with '.' in the name
16313 # {"my.key":"myvalue"} => myvalue
16314 http-request set-var(txn.pay_mykey) req.body,json_query('$.my\\.key')
16315
16316 # {"boolean-false":false} => 0
16317 http-request set-var(txn.pay_boolean_false) req.body,json_query('$.boolean-false')
16318
16319 # get the value of the key 'iss' from a JWT Bearer token
16320 http-request set-var(txn.token_payload) req.hdr(Authorization),word(2,.),ub64dec,json_query('$.iss')
16321
Thierry FOURNIER060762e2014-04-23 13:29:15 +020016322language(<value>[,<default>])
16323 Returns the value with the highest q-factor from a list as extracted from the
16324 "accept-language" header using "req.fhdr". Values with no q-factor have a
16325 q-factor of 1. Values with a q-factor of 0 are dropped. Only values which
16326 belong to the list of semi-colon delimited <values> will be considered. The
16327 argument <value> syntax is "lang[;lang[;lang[;...]]]". If no value matches the
16328 given list and a default value is provided, it is returned. Note that language
16329 names may have a variant after a dash ('-'). If this variant is present in the
16330 list, it will be matched, but if it is not, only the base language is checked.
16331 The match is case-sensitive, and the output string is always one of those
Davor Ocelice9ed2812017-12-25 17:49:28 +010016332 provided in arguments. The ordering of arguments is meaningless, only the
Thierry FOURNIER060762e2014-04-23 13:29:15 +020016333 ordering of the values in the request counts, as the first value among
16334 multiple sharing the same q-factor is used.
Thierry FOURNIERad903512014-04-11 17:51:01 +020016335
Thierry FOURNIER060762e2014-04-23 13:29:15 +020016336 Example :
Thierry FOURNIERad903512014-04-11 17:51:01 +020016337
Thierry FOURNIER060762e2014-04-23 13:29:15 +020016338 # this configuration switches to the backend matching a
16339 # given language based on the request :
Thierry FOURNIERad903512014-04-11 17:51:01 +020016340
Thierry FOURNIER060762e2014-04-23 13:29:15 +020016341 acl es req.fhdr(accept-language),language(es;fr;en) -m str es
16342 acl fr req.fhdr(accept-language),language(es;fr;en) -m str fr
16343 acl en req.fhdr(accept-language),language(es;fr;en) -m str en
16344 use_backend spanish if es
16345 use_backend french if fr
16346 use_backend english if en
16347 default_backend choose_your_language
Thierry FOURNIERad903512014-04-11 17:51:01 +020016348
Willy Tarreau60a2ee72017-12-15 07:13:48 +010016349length
Etienne Carriereed0d24e2017-12-13 13:41:34 +010016350 Get the length of the string. This can only be placed after a string
16351 sample fetch function or after a transformation keyword returning a string
16352 type. The result is of type integer.
16353
Willy Tarreauffcb2e42014-07-10 16:29:08 +020016354lower
16355 Convert a string sample to lower case. This can only be placed after a string
16356 sample fetch function or after a transformation keyword returning a string
16357 type. The result is of type string.
16358
Willy Tarreau0dbfdba2014-07-10 16:37:47 +020016359ltime(<format>[,<offset>])
16360 Converts an integer supposed to contain a date since epoch to a string
16361 representing this date in local time using a format defined by the <format>
16362 string using strftime(3). The purpose is to allow any date format to be used
16363 in logs. An optional <offset> in seconds may be applied to the input date
16364 (positive or negative). See the strftime() man page for the format supported
16365 by your operating system. See also the utime converter.
16366
16367 Example :
16368
16369 # Emit two colons, one with the local time and another with ip:port
Davor Ocelice9ed2812017-12-25 17:49:28 +010016370 # e.g. 20140710162350 127.0.0.1:57325
Willy Tarreau0dbfdba2014-07-10 16:37:47 +020016371 log-format %[date,ltime(%Y%m%d%H%M%S)]\ %ci:%cp
16372
Christopher Faulet51fc9d12020-04-01 17:24:41 +020016373ltrim(<chars>)
16374 Skips any characters from <chars> from the beginning of the string
16375 representation of the input sample.
16376
Thierry FOURNIER060762e2014-04-23 13:29:15 +020016377map(<map_file>[,<default_value>])
16378map_<match_type>(<map_file>[,<default_value>])
16379map_<match_type>_<output_type>(<map_file>[,<default_value>])
16380 Search the input value from <map_file> using the <match_type> matching method,
16381 and return the associated value converted to the type <output_type>. If the
16382 input value cannot be found in the <map_file>, the converter returns the
16383 <default_value>. If the <default_value> is not set, the converter fails and
16384 acts as if no input value could be fetched. If the <match_type> is not set, it
16385 defaults to "str". Likewise, if the <output_type> is not set, it defaults to
16386 "str". For convenience, the "map" keyword is an alias for "map_str" and maps a
16387 string to another string.
Thierry FOURNIERd5f624d2013-11-26 11:52:33 +010016388
Thierry FOURNIER060762e2014-04-23 13:29:15 +020016389 It is important to avoid overlapping between the keys : IP addresses and
16390 strings are stored in trees, so the first of the finest match will be used.
16391 Other keys are stored in lists, so the first matching occurrence will be used.
Thierry FOURNIERd5f624d2013-11-26 11:52:33 +010016392
Tim Düsterhus4896c442016-11-29 02:15:19 +010016393 The following array contains the list of all map functions available sorted by
Thierry FOURNIER060762e2014-04-23 13:29:15 +020016394 input type, match type and output type.
Thierry FOURNIERd5f624d2013-11-26 11:52:33 +010016395
Thierry FOURNIER060762e2014-04-23 13:29:15 +020016396 input type | match method | output type str | output type int | output type ip
16397 -----------+--------------+-----------------+-----------------+---------------
16398 str | str | map_str | map_str_int | map_str_ip
16399 -----------+--------------+-----------------+-----------------+---------------
Willy Tarreau787a4c02014-05-10 07:55:30 +020016400 str | beg | map_beg | map_beg_int | map_end_ip
16401 -----------+--------------+-----------------+-----------------+---------------
Thierry FOURNIER060762e2014-04-23 13:29:15 +020016402 str | sub | map_sub | map_sub_int | map_sub_ip
16403 -----------+--------------+-----------------+-----------------+---------------
16404 str | dir | map_dir | map_dir_int | map_dir_ip
16405 -----------+--------------+-----------------+-----------------+---------------
16406 str | dom | map_dom | map_dom_int | map_dom_ip
16407 -----------+--------------+-----------------+-----------------+---------------
16408 str | end | map_end | map_end_int | map_end_ip
16409 -----------+--------------+-----------------+-----------------+---------------
Ruoshan Huang3c5e3742016-12-02 16:25:31 +080016410 str | reg | map_reg | map_reg_int | map_reg_ip
16411 -----------+--------------+-----------------+-----------------+---------------
16412 str | reg | map_regm | map_reg_int | map_reg_ip
Thierry FOURNIER060762e2014-04-23 13:29:15 +020016413 -----------+--------------+-----------------+-----------------+---------------
16414 int | int | map_int | map_int_int | map_int_ip
16415 -----------+--------------+-----------------+-----------------+---------------
16416 ip | ip | map_ip | map_ip_int | map_ip_ip
16417 -----------+--------------+-----------------+-----------------+---------------
Thierry FOURNIERd5f624d2013-11-26 11:52:33 +010016418
Thierry Fournier8feaa662016-02-10 22:55:20 +010016419 The special map called "map_regm" expect matching zone in the regular
16420 expression and modify the output replacing back reference (like "\1") by
16421 the corresponding match text.
16422
Thierry FOURNIER060762e2014-04-23 13:29:15 +020016423 The file contains one key + value per line. Lines which start with '#' are
16424 ignored, just like empty lines. Leading tabs and spaces are stripped. The key
16425 is then the first "word" (series of non-space/tabs characters), and the value
16426 is what follows this series of space/tab till the end of the line excluding
16427 trailing spaces/tabs.
Thierry FOURNIERd5f624d2013-11-26 11:52:33 +010016428
Thierry FOURNIER060762e2014-04-23 13:29:15 +020016429 Example :
16430
16431 # this is a comment and is ignored
16432 2.22.246.0/23 United Kingdom \n
16433 <-><-----------><--><------------><---->
16434 | | | | `- trailing spaces ignored
16435 | | | `---------- value
16436 | | `-------------------- middle spaces ignored
16437 | `---------------------------- key
16438 `------------------------------------ leading spaces ignored
16439
Willy Tarreau97707872015-01-27 15:12:13 +010016440mod(<value>)
Thierry FOURNIER07ee64e2015-07-06 23:43:03 +020016441 Divides the input value of type signed integer by <value>, and returns the
16442 remainder as an signed integer. If <value> is null, then zero is returned.
Thierry FOURNIER5d86fae2015-07-07 21:10:16 +020016443 <value> can be a numeric value or a variable name. The name of the variable
Daniel Schneller0b547052016-03-21 20:46:57 +010016444 starts with an indication about its scope. The scopes allowed are:
Christopher Fauletff2613e2016-11-09 11:36:17 +010016445 "proc" : the variable is shared with the whole process
Daniel Schneller0b547052016-03-21 20:46:57 +010016446 "sess" : the variable is shared with the whole session
16447 "txn" : the variable is shared with the transaction (request and response)
16448 "req" : the variable is shared only during request processing
16449 "res" : the variable is shared only during response processing
Davor Ocelice9ed2812017-12-25 17:49:28 +010016450 This prefix is followed by a name. The separator is a '.'. The name may only
Christopher Fauletb71557a2016-10-31 10:49:03 +010016451 contain characters 'a-z', 'A-Z', '0-9', '.' and '_'.
Willy Tarreau97707872015-01-27 15:12:13 +010016452
Alexbf1bd5a2021-04-24 13:02:21 +020016453mqtt_field_value(<packettype>,<fieldname_or_property_ID>)
Baptiste Assmanne279ca62020-10-27 18:10:06 +010016454 Returns value of <fieldname> found in input MQTT payload of type
16455 <packettype>.
16456 <packettype> can be either a string (case insensitive matching) or a numeric
16457 value corresponding to the type of packet we're supposed to extract data
16458 from.
16459 Supported string and integers can be found here:
16460 https://docs.oasis-open.org/mqtt/mqtt/v3.1.1/os/mqtt-v3.1.1-os.html#_Toc398718021
16461 https://docs.oasis-open.org/mqtt/mqtt/v5.0/os/mqtt-v5.0-os.html#_Toc3901022
16462
16463 <fieldname> depends on <packettype> and can be any of the following below.
16464 (note that <fieldname> matching is case insensitive).
16465 <property id> can only be found in MQTT v5.0 streams. check this table:
16466 https://docs.oasis-open.org/mqtt/mqtt/v5.0/os/mqtt-v5.0-os.html#_Toc3901029
16467
16468 - CONNECT (or 1): flags, protocol_name, protocol_version, client_identifier,
16469 will_topic, will_payload, username, password, keepalive
16470 OR any property ID as a numeric value (for MQTT v5.0
16471 packets only):
16472 17: Session Expiry Interval
16473 33: Receive Maximum
16474 39: Maximum Packet Size
16475 34: Topic Alias Maximum
16476 25: Request Response Information
16477 23: Request Problem Information
16478 21: Authentication Method
16479 22: Authentication Data
16480 18: Will Delay Interval
16481 1: Payload Format Indicator
16482 2: Message Expiry Interval
16483 3: Content Type
16484 8: Response Topic
16485 9: Correlation Data
16486 Not supported yet:
16487 38: User Property
16488
16489 - CONNACK (or 2): flags, protocol_version, reason_code
16490 OR any property ID as a numeric value (for MQTT v5.0
16491 packets only):
16492 17: Session Expiry Interval
16493 33: Receive Maximum
16494 36: Maximum QoS
16495 37: Retain Available
16496 39: Maximum Packet Size
16497 18: Assigned Client Identifier
16498 34: Topic Alias Maximum
16499 31: Reason String
16500 40; Wildcard Subscription Available
16501 41: Subscription Identifiers Available
16502 42: Shared Subscription Available
16503 19: Server Keep Alive
16504 26: Response Information
16505 28: Server Reference
16506 21: Authentication Method
16507 22: Authentication Data
16508 Not supported yet:
16509 38: User Property
16510
16511 Due to current HAProxy design, only the first message sent by the client and
16512 the server can be parsed. Thus this converter can extract data only from
16513 CONNECT and CONNACK packet types. CONNECT is the first message sent by the
16514 client and CONNACK is the first response sent by the server.
16515
16516 Example:
16517
16518 acl data_in_buffer req.len ge 4
16519 tcp-request content set-var(txn.username) \
16520 req.payload(0,0),mqtt_field_value(connect,protocol_name) \
16521 if data_in_buffer
16522 # do the same as above
16523 tcp-request content set-var(txn.username) \
16524 req.payload(0,0),mqtt_field_value(1,protocol_name) \
16525 if data_in_buffer
16526
16527mqtt_is_valid
16528 Checks that the binary input is a valid MQTT packet. It returns a boolean.
16529
16530 Due to current HAProxy design, only the first message sent by the client and
16531 the server can be parsed. Thus this converter can extract data only from
16532 CONNECT and CONNACK packet types. CONNECT is the first message sent by the
16533 client and CONNACK is the first response sent by the server.
16534
16535 Example:
16536
16537 acl data_in_buffer req.len ge 4
Daniel Corbettcc9d9b02021-05-13 10:46:07 -040016538 tcp-request content reject unless { req.payload(0,0),mqtt_is_valid }
Baptiste Assmanne279ca62020-10-27 18:10:06 +010016539
Willy Tarreau97707872015-01-27 15:12:13 +010016540mul(<value>)
Thierry FOURNIER07ee64e2015-07-06 23:43:03 +020016541 Multiplies the input value of type signed integer by <value>, and returns
Thierry FOURNIER00c005c2015-07-08 01:10:21 +020016542 the product as an signed integer. In case of overflow, the largest possible
16543 value for the sign is returned so that the operation doesn't wrap around.
Thierry FOURNIER5d86fae2015-07-07 21:10:16 +020016544 <value> can be a numeric value or a variable name. The name of the variable
Daniel Schneller0b547052016-03-21 20:46:57 +010016545 starts with an indication about its scope. The scopes allowed are:
Christopher Fauletff2613e2016-11-09 11:36:17 +010016546 "proc" : the variable is shared with the whole process
Daniel Schneller0b547052016-03-21 20:46:57 +010016547 "sess" : the variable is shared with the whole session
16548 "txn" : the variable is shared with the transaction (request and response)
16549 "req" : the variable is shared only during request processing
16550 "res" : the variable is shared only during response processing
Davor Ocelice9ed2812017-12-25 17:49:28 +010016551 This prefix is followed by a name. The separator is a '.'. The name may only
Christopher Fauletb71557a2016-10-31 10:49:03 +010016552 contain characters 'a-z', 'A-Z', '0-9', '.' and '_'.
Willy Tarreau97707872015-01-27 15:12:13 +010016553
Nenad Merdanovicb7e7c472017-03-12 21:56:55 +010016554nbsrv
16555 Takes an input value of type string, interprets it as a backend name and
16556 returns the number of usable servers in that backend. Can be used in places
16557 where we want to look up a backend from a dynamic name, like a result of a
16558 map lookup.
16559
Willy Tarreau97707872015-01-27 15:12:13 +010016560neg
Thierry FOURNIER07ee64e2015-07-06 23:43:03 +020016561 Takes the input value of type signed integer, computes the opposite value,
16562 and returns the remainder as an signed integer. 0 is identity. This operator
16563 is provided for reversed subtracts : in order to subtract the input from a
16564 constant, simply perform a "neg,add(value)".
Willy Tarreau97707872015-01-27 15:12:13 +010016565
16566not
Thierry FOURNIER07ee64e2015-07-06 23:43:03 +020016567 Returns a boolean FALSE if the input value of type signed integer is
Willy Tarreau97707872015-01-27 15:12:13 +010016568 non-null, otherwise returns TRUE. Used in conjunction with and(), it can be
Davor Ocelice9ed2812017-12-25 17:49:28 +010016569 used to report true/false for bit testing on input values (e.g. verify the
Willy Tarreau97707872015-01-27 15:12:13 +010016570 absence of a flag).
16571
16572odd
Thierry FOURNIER07ee64e2015-07-06 23:43:03 +020016573 Returns a boolean TRUE if the input value of type signed integer is odd
Willy Tarreau97707872015-01-27 15:12:13 +010016574 otherwise returns FALSE. It is functionally equivalent to "and(1),bool".
16575
16576or(<value>)
Thierry FOURNIER07ee64e2015-07-06 23:43:03 +020016577 Performs a bitwise "OR" between <value> and the input value of type signed
Thierry FOURNIER5d86fae2015-07-07 21:10:16 +020016578 integer, and returns the result as an signed integer. <value> can be a
Daniel Schneller0b547052016-03-21 20:46:57 +010016579 numeric value or a variable name. The name of the variable starts with an
16580 indication about its scope. The scopes allowed are:
Christopher Fauletff2613e2016-11-09 11:36:17 +010016581 "proc" : the variable is shared with the whole process
Daniel Schneller0b547052016-03-21 20:46:57 +010016582 "sess" : the variable is shared with the whole session
16583 "txn" : the variable is shared with the transaction (request and response)
16584 "req" : the variable is shared only during request processing
16585 "res" : the variable is shared only during response processing
Davor Ocelice9ed2812017-12-25 17:49:28 +010016586 This prefix is followed by a name. The separator is a '.'. The name may only
Christopher Fauletb71557a2016-10-31 10:49:03 +010016587 contain characters 'a-z', 'A-Z', '0-9', '.' and '_'.
Willy Tarreau97707872015-01-27 15:12:13 +010016588
Frédéric Lécaillebfe61382019-03-06 14:34:36 +010016589protobuf(<field_number>,[<field_type>])
16590 This extracts the protocol buffers message field in raw mode of an input binary
16591 sample representation of a protocol buffer message with <field_number> as field
16592 number (dotted notation) if <field_type> is not present, or as an integer sample
16593 if this field is present (see also "ungrpc" below).
16594 The list of the authorized types is the following one: "int32", "int64", "uint32",
16595 "uint64", "sint32", "sint64", "bool", "enum" for the "varint" wire type 0
16596 "fixed64", "sfixed64", "double" for the 64bit wire type 1, "fixed32", "sfixed32",
16597 "float" for the wire type 5. Note that "string" is considered as a length-delimited
16598 type, so it does not require any <field_type> argument to be extracted.
16599 More information may be found here about the protocol buffers message field types:
16600 https://developers.google.com/protocol-buffers/docs/encoding
16601
Willy Tarreauc4dc3502015-01-23 20:39:28 +010016602regsub(<regex>,<subst>[,<flags>])
Willy Tarreau7eda8492015-01-20 19:47:06 +010016603 Applies a regex-based substitution to the input string. It does the same
16604 operation as the well-known "sed" utility with "s/<regex>/<subst>/". By
16605 default it will replace in the input string the first occurrence of the
16606 largest part matching the regular expression <regex> with the substitution
16607 string <subst>. It is possible to replace all occurrences instead by adding
16608 the flag "g" in the third argument <flags>. It is also possible to make the
16609 regex case insensitive by adding the flag "i" in <flags>. Since <flags> is a
16610 string, it is made up from the concatenation of all desired flags. Thus if
16611 both "i" and "g" are desired, using "gi" or "ig" will have the same effect.
Willy Tarreauef21fac2020-02-14 13:37:20 +010016612 The first use of this converter is to replace certain characters or sequence
16613 of characters with other ones.
16614
16615 It is highly recommended to enclose the regex part using protected quotes to
16616 improve clarity and never have a closing parenthesis from the regex mixed up
16617 with the parenthesis from the function. Just like in Bourne shell, the first
16618 level of quotes is processed when delimiting word groups on the line, a
16619 second level is usable for argument. It is recommended to use single quotes
16620 outside since these ones do not try to resolve backslashes nor dollar signs.
Willy Tarreau7eda8492015-01-20 19:47:06 +010016621
Willy Tarreaucd0d2ed2020-02-14 17:33:06 +010016622 Examples:
Willy Tarreau7eda8492015-01-20 19:47:06 +010016623
16624 # de-duplicate "/" in header "x-path".
16625 # input: x-path: /////a///b/c/xzxyz/
16626 # output: x-path: /a/b/c/xzxyz/
Willy Tarreauef21fac2020-02-14 13:37:20 +010016627 http-request set-header x-path "%[hdr(x-path),regsub('/+','/','g')]"
Willy Tarreau7eda8492015-01-20 19:47:06 +010016628
Willy Tarreaucd0d2ed2020-02-14 17:33:06 +010016629 # copy query string to x-query and drop all leading '?', ';' and '&'
16630 http-request set-header x-query "%[query,regsub([?;&]*,'')]"
16631
Jerome Magnin07e1e3c2020-02-16 19:20:19 +010016632 # capture groups and backreferences
16633 # both lines do the same.
Willy Tarreau465dc7d2020-10-08 18:05:56 +020016634 http-request redirect location %[url,'regsub("(foo|bar)([0-9]+)?","\2\1",i)']
Jerome Magnin07e1e3c2020-02-16 19:20:19 +010016635 http-request redirect location %[url,regsub(\"(foo|bar)([0-9]+)?\",\"\2\1\",i)]
16636
Thierry FOURNIER35ab2752015-05-28 13:22:03 +020016637capture-req(<id>)
16638 Capture the string entry in the request slot <id> and returns the entry as
16639 is. If the slot doesn't exist, the capture fails silently.
16640
16641 See also: "declare capture", "http-request capture",
Baptiste Assmann5ac425c2015-10-21 23:13:46 +020016642 "http-response capture", "capture.req.hdr" and
16643 "capture.res.hdr" (sample fetches).
Thierry FOURNIER35ab2752015-05-28 13:22:03 +020016644
16645capture-res(<id>)
16646 Capture the string entry in the response slot <id> and returns the entry as
16647 is. If the slot doesn't exist, the capture fails silently.
16648
16649 See also: "declare capture", "http-request capture",
Baptiste Assmann5ac425c2015-10-21 23:13:46 +020016650 "http-response capture", "capture.req.hdr" and
16651 "capture.res.hdr" (sample fetches).
Thierry FOURNIER35ab2752015-05-28 13:22:03 +020016652
Christopher Faulet568415a2020-04-01 17:24:47 +020016653rtrim(<chars>)
16654 Skips any characters from <chars> from the end of the string representation
16655 of the input sample.
16656
Willy Tarreau23ec4ca2014-07-15 20:15:37 +020016657sdbm([<avalanche>])
16658 Hashes a binary input sample into an unsigned 32-bit quantity using the SDBM
16659 hash function. Optionally, it is possible to apply a full avalanche hash
16660 function to the output if the optional <avalanche> argument equals 1. This
16661 converter uses the same functions as used by the various hash-based load
16662 balancing algorithms, so it will provide exactly the same results. It is
16663 mostly intended for debugging, but can be used as a stick-table entry to
16664 collect rough statistics. It must not be used for security purposes as a
Emmanuel Hocdet50791a72018-03-21 11:19:01 +010016665 32-bit hash is trivial to break. See also "crc32", "djb2", "wt6", "crc32c",
16666 and the "hash-type" directive.
Willy Tarreau23ec4ca2014-07-15 20:15:37 +020016667
Tim Duesterhusf38175c2020-06-09 11:48:42 +020016668secure_memcmp(<var>)
16669 Compares the contents of <var> with the input value. Both values are treated
16670 as a binary string. Returns a boolean indicating whether both binary strings
16671 match.
16672
16673 If both binary strings have the same length then the comparison will be
16674 performed in constant time.
16675
Daniel Corbett9f0843f2021-05-08 10:50:37 -040016676 Please note that this converter is only available when HAProxy has been
Tim Duesterhusf38175c2020-06-09 11:48:42 +020016677 compiled with USE_OPENSSL.
16678
16679 Example :
16680
16681 http-request set-var(txn.token) hdr(token)
16682 # Check whether the token sent by the client matches the secret token
16683 # value, without leaking the contents using a timing attack.
16684 acl token_given str(my_secret_token),secure_memcmp(txn.token)
16685
Tim Duesterhusef4e45c2021-01-21 17:40:50 +010016686set-var(<var>)
Davor Ocelice9ed2812017-12-25 17:49:28 +010016687 Sets a variable with the input content and returns the content on the output
16688 as-is. The variable keeps the value and the associated input type. The name of
16689 the variable starts with an indication about its scope. The scopes allowed are:
Christopher Fauletff2613e2016-11-09 11:36:17 +010016690 "proc" : the variable is shared with the whole process
Daniel Schneller0b547052016-03-21 20:46:57 +010016691 "sess" : the variable is shared with the whole session
16692 "txn" : the variable is shared with the transaction (request and
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020016693 response),
Daniel Schneller0b547052016-03-21 20:46:57 +010016694 "req" : the variable is shared only during request processing,
16695 "res" : the variable is shared only during response processing.
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020016696 This prefix is followed by a name. The separator is a '.'. The name may only
Christopher Fauletb71557a2016-10-31 10:49:03 +010016697 contain characters 'a-z', 'A-Z', '0-9', '.' and '_'.
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020016698
Dragan Dosen6e5a9ca2017-10-24 09:18:23 +020016699sha1
Tim Duesterhusd4376302019-06-17 12:41:44 +020016700 Converts a binary input sample to a SHA-1 digest. The result is a binary
Dragan Dosen6e5a9ca2017-10-24 09:18:23 +020016701 sample with length of 20 bytes.
16702
Tim Duesterhusd4376302019-06-17 12:41:44 +020016703sha2([<bits>])
16704 Converts a binary input sample to a digest in the SHA-2 family. The result
16705 is a binary sample with length of <bits>/8 bytes.
16706
16707 Valid values for <bits> are 224, 256, 384, 512, each corresponding to
16708 SHA-<bits>. The default value is 256.
16709
Daniel Corbett9f0843f2021-05-08 10:50:37 -040016710 Please note that this converter is only available when HAProxy has been
Tim Duesterhusd4376302019-06-17 12:41:44 +020016711 compiled with USE_OPENSSL.
16712
Nenad Merdanovic177adc92019-08-27 01:58:13 +020016713srv_queue
16714 Takes an input value of type string, either a server name or <backend>/<server>
16715 format and returns the number of queued sessions on that server. Can be used
16716 in places where we want to look up queued sessions from a dynamic name, like a
16717 cookie value (e.g. req.cook(SRVID),srv_queue) and then make a decision to break
16718 persistence or direct a request elsewhere.
16719
Tim Duesterhusca097c12018-04-27 21:18:45 +020016720strcmp(<var>)
16721 Compares the contents of <var> with the input value of type string. Returns
16722 the result as a signed integer compatible with strcmp(3): 0 if both strings
16723 are identical. A value less than 0 if the left string is lexicographically
16724 smaller than the right string or if the left string is shorter. A value greater
16725 than 0 otherwise (right string greater than left string or the right string is
16726 shorter).
16727
Tim Duesterhusf38175c2020-06-09 11:48:42 +020016728 See also the secure_memcmp converter if you need to compare two binary
16729 strings in constant time.
16730
Tim Duesterhusca097c12018-04-27 21:18:45 +020016731 Example :
16732
16733 http-request set-var(txn.host) hdr(host)
16734 # Check whether the client is attempting domain fronting.
16735 acl ssl_sni_http_host_match ssl_fc_sni,strcmp(txn.host) eq 0
16736
16737
Willy Tarreau97707872015-01-27 15:12:13 +010016738sub(<value>)
Thierry FOURNIER07ee64e2015-07-06 23:43:03 +020016739 Subtracts <value> from the input value of type signed integer, and returns
16740 the result as an signed integer. Note: in order to subtract the input from
Thierry FOURNIER5d86fae2015-07-07 21:10:16 +020016741 a constant, simply perform a "neg,add(value)". <value> can be a numeric value
Daniel Schneller0b547052016-03-21 20:46:57 +010016742 or a variable name. The name of the variable starts with an indication about
16743 its scope. The scopes allowed are:
Christopher Fauletff2613e2016-11-09 11:36:17 +010016744 "proc" : the variable is shared with the whole process
Daniel Schneller0b547052016-03-21 20:46:57 +010016745 "sess" : the variable is shared with the whole session
16746 "txn" : the variable is shared with the transaction (request and
Thierry FOURNIER5d86fae2015-07-07 21:10:16 +020016747 response),
Daniel Schneller0b547052016-03-21 20:46:57 +010016748 "req" : the variable is shared only during request processing,
16749 "res" : the variable is shared only during response processing.
Thierry FOURNIER5d86fae2015-07-07 21:10:16 +020016750 This prefix is followed by a name. The separator is a '.'. The name may only
Christopher Fauletb71557a2016-10-31 10:49:03 +010016751 contain characters 'a-z', 'A-Z', '0-9', '.' and '_'.
Willy Tarreau97707872015-01-27 15:12:13 +010016752
Willy Tarreaud9f316a2014-07-10 14:03:38 +020016753table_bytes_in_rate(<table>)
16754 Uses the string representation of the input sample to perform a look up in
16755 the specified table. If the key is not found in the table, integer value zero
16756 is returned. Otherwise the converter returns the average client-to-server
16757 bytes rate associated with the input sample in the designated table, measured
16758 in amount of bytes over the period configured in the table. See also the
16759 sc_bytes_in_rate sample fetch keyword.
16760
16761
16762table_bytes_out_rate(<table>)
16763 Uses the string representation of the input sample to perform a look up in
16764 the specified table. If the key is not found in the table, integer value zero
16765 is returned. Otherwise the converter returns the average server-to-client
16766 bytes rate associated with the input sample in the designated table, measured
16767 in amount of bytes over the period configured in the table. See also the
16768 sc_bytes_out_rate sample fetch keyword.
16769
16770table_conn_cnt(<table>)
16771 Uses the string representation of the input sample to perform a look up in
16772 the specified table. If the key is not found in the table, integer value zero
Davor Ocelice9ed2812017-12-25 17:49:28 +010016773 is returned. Otherwise the converter returns the cumulative number of incoming
Willy Tarreaud9f316a2014-07-10 14:03:38 +020016774 connections associated with the input sample in the designated table. See
16775 also the sc_conn_cnt sample fetch keyword.
16776
16777table_conn_cur(<table>)
16778 Uses the string representation of the input sample to perform a look up in
16779 the specified table. If the key is not found in the table, integer value zero
16780 is returned. Otherwise the converter returns the current amount of concurrent
16781 tracked connections associated with the input sample in the designated table.
16782 See also the sc_conn_cur sample fetch keyword.
16783
16784table_conn_rate(<table>)
16785 Uses the string representation of the input sample to perform a look up in
16786 the specified table. If the key is not found in the table, integer value zero
16787 is returned. Otherwise the converter returns the average incoming connection
16788 rate associated with the input sample in the designated table. See also the
16789 sc_conn_rate sample fetch keyword.
16790
Thierry FOURNIER236657b2015-08-19 08:25:14 +020016791table_gpt0(<table>)
16792 Uses the string representation of the input sample to perform a look up in
16793 the specified table. If the key is not found in the table, boolean value zero
16794 is returned. Otherwise the converter returns the current value of the first
16795 general purpose tag associated with the input sample in the designated table.
16796 See also the sc_get_gpt0 sample fetch keyword.
16797
Willy Tarreaud9f316a2014-07-10 14:03:38 +020016798table_gpc0(<table>)
16799 Uses the string representation of the input sample to perform a look up in
16800 the specified table. If the key is not found in the table, integer value zero
16801 is returned. Otherwise the converter returns the current value of the first
16802 general purpose counter associated with the input sample in the designated
16803 table. See also the sc_get_gpc0 sample fetch keyword.
16804
16805table_gpc0_rate(<table>)
16806 Uses the string representation of the input sample to perform a look up in
16807 the specified table. If the key is not found in the table, integer value zero
16808 is returned. Otherwise the converter returns the frequency which the gpc0
16809 counter was incremented over the configured period in the table, associated
16810 with the input sample in the designated table. See also the sc_get_gpc0_rate
16811 sample fetch keyword.
16812
Frédéric Lécaille6778b272018-01-29 15:22:53 +010016813table_gpc1(<table>)
16814 Uses the string representation of the input sample to perform a look up in
16815 the specified table. If the key is not found in the table, integer value zero
16816 is returned. Otherwise the converter returns the current value of the second
16817 general purpose counter associated with the input sample in the designated
16818 table. See also the sc_get_gpc1 sample fetch keyword.
16819
16820table_gpc1_rate(<table>)
16821 Uses the string representation of the input sample to perform a look up in
16822 the specified table. If the key is not found in the table, integer value zero
16823 is returned. Otherwise the converter returns the frequency which the gpc1
16824 counter was incremented over the configured period in the table, associated
16825 with the input sample in the designated table. See also the sc_get_gpc1_rate
16826 sample fetch keyword.
16827
Willy Tarreaud9f316a2014-07-10 14:03:38 +020016828table_http_err_cnt(<table>)
16829 Uses the string representation of the input sample to perform a look up in
16830 the specified table. If the key is not found in the table, integer value zero
Davor Ocelice9ed2812017-12-25 17:49:28 +010016831 is returned. Otherwise the converter returns the cumulative number of HTTP
Willy Tarreaud9f316a2014-07-10 14:03:38 +020016832 errors associated with the input sample in the designated table. See also the
16833 sc_http_err_cnt sample fetch keyword.
16834
16835table_http_err_rate(<table>)
16836 Uses the string representation of the input sample to perform a look up in
16837 the specified table. If the key is not found in the table, integer value zero
16838 is returned. Otherwise the average rate of HTTP errors associated with the
16839 input sample in the designated table, measured in amount of errors over the
16840 period configured in the table. See also the sc_http_err_rate sample fetch
16841 keyword.
16842
Willy Tarreau826f3ab2021-02-10 12:07:15 +010016843table_http_fail_cnt(<table>)
16844 Uses the string representation of the input sample to perform a look up in
16845 the specified table. If the key is not found in the table, integer value zero
16846 is returned. Otherwise the converter returns the cumulative number of HTTP
16847 failures associated with the input sample in the designated table. See also
16848 the sc_http_fail_cnt sample fetch keyword.
16849
16850table_http_fail_rate(<table>)
16851 Uses the string representation of the input sample to perform a look up in
16852 the specified table. If the key is not found in the table, integer value zero
16853 is returned. Otherwise the average rate of HTTP failures associated with the
16854 input sample in the designated table, measured in amount of failures over the
16855 period configured in the table. See also the sc_http_fail_rate sample fetch
16856 keyword.
16857
Willy Tarreaud9f316a2014-07-10 14:03:38 +020016858table_http_req_cnt(<table>)
16859 Uses the string representation of the input sample to perform a look up in
16860 the specified table. If the key is not found in the table, integer value zero
Davor Ocelice9ed2812017-12-25 17:49:28 +010016861 is returned. Otherwise the converter returns the cumulative number of HTTP
Willy Tarreaud9f316a2014-07-10 14:03:38 +020016862 requests associated with the input sample in the designated table. See also
16863 the sc_http_req_cnt sample fetch keyword.
16864
16865table_http_req_rate(<table>)
16866 Uses the string representation of the input sample to perform a look up in
16867 the specified table. If the key is not found in the table, integer value zero
16868 is returned. Otherwise the average rate of HTTP requests associated with the
16869 input sample in the designated table, measured in amount of requests over the
16870 period configured in the table. See also the sc_http_req_rate sample fetch
16871 keyword.
16872
16873table_kbytes_in(<table>)
16874 Uses the string representation of the input sample to perform a look up in
16875 the specified table. If the key is not found in the table, integer value zero
Davor Ocelice9ed2812017-12-25 17:49:28 +010016876 is returned. Otherwise the converter returns the cumulative number of client-
Willy Tarreaud9f316a2014-07-10 14:03:38 +020016877 to-server data associated with the input sample in the designated table,
16878 measured in kilobytes. The test is currently performed on 32-bit integers,
16879 which limits values to 4 terabytes. See also the sc_kbytes_in sample fetch
16880 keyword.
16881
16882table_kbytes_out(<table>)
16883 Uses the string representation of the input sample to perform a look up in
16884 the specified table. If the key is not found in the table, integer value zero
Davor Ocelice9ed2812017-12-25 17:49:28 +010016885 is returned. Otherwise the converter returns the cumulative number of server-
Willy Tarreaud9f316a2014-07-10 14:03:38 +020016886 to-client data associated with the input sample in the designated table,
16887 measured in kilobytes. The test is currently performed on 32-bit integers,
16888 which limits values to 4 terabytes. See also the sc_kbytes_out sample fetch
16889 keyword.
16890
16891table_server_id(<table>)
16892 Uses the string representation of the input sample to perform a look up in
16893 the specified table. If the key is not found in the table, integer value zero
16894 is returned. Otherwise the converter returns the server ID associated with
16895 the input sample in the designated table. A server ID is associated to a
16896 sample by a "stick" rule when a connection to a server succeeds. A server ID
16897 zero means that no server is associated with this key.
16898
16899table_sess_cnt(<table>)
16900 Uses the string representation of the input sample to perform a look up in
16901 the specified table. If the key is not found in the table, integer value zero
Davor Ocelice9ed2812017-12-25 17:49:28 +010016902 is returned. Otherwise the converter returns the cumulative number of incoming
Willy Tarreaud9f316a2014-07-10 14:03:38 +020016903 sessions associated with the input sample in the designated table. Note that
16904 a session here refers to an incoming connection being accepted by the
16905 "tcp-request connection" rulesets. See also the sc_sess_cnt sample fetch
16906 keyword.
16907
16908table_sess_rate(<table>)
16909 Uses the string representation of the input sample to perform a look up in
16910 the specified table. If the key is not found in the table, integer value zero
16911 is returned. Otherwise the converter returns the average incoming session
16912 rate associated with the input sample in the designated table. Note that a
16913 session here refers to an incoming connection being accepted by the
16914 "tcp-request connection" rulesets. See also the sc_sess_rate sample fetch
16915 keyword.
16916
16917table_trackers(<table>)
16918 Uses the string representation of the input sample to perform a look up in
16919 the specified table. If the key is not found in the table, integer value zero
16920 is returned. Otherwise the converter returns the current amount of concurrent
16921 connections tracking the same key as the input sample in the designated
16922 table. It differs from table_conn_cur in that it does not rely on any stored
16923 information but on the table's reference count (the "use" value which is
16924 returned by "show table" on the CLI). This may sometimes be more suited for
16925 layer7 tracking. It can be used to tell a server how many concurrent
16926 connections there are from a given address for example. See also the
16927 sc_trackers sample fetch keyword.
16928
Moemen MHEDHBI92f7d432021-04-01 20:53:59 +020016929ub64dec
16930 This converter is the base64url variant of b64dec converter. base64url
16931 encoding is the "URL and Filename Safe Alphabet" variant of base64 encoding.
16932 It is also the encoding used in JWT (JSON Web Token) standard.
16933
16934 Example:
16935 # Decoding a JWT payload:
16936 http-request set-var(txn.token_payload) req.hdr(Authorization),word(2,.),ub64dec
16937
16938ub64enc
16939 This converter is the base64url variant of base64 converter.
16940
Willy Tarreauffcb2e42014-07-10 16:29:08 +020016941upper
16942 Convert a string sample to upper case. This can only be placed after a string
16943 sample fetch function or after a transformation keyword returning a string
16944 type. The result is of type string.
16945
Willy Tarreau62ba9ba2020-04-23 17:54:47 +020016946url_dec([<in_form>])
16947 Takes an url-encoded string provided as input and returns the decoded version
16948 as output. The input and the output are of type string. If the <in_form>
16949 argument is set to a non-zero integer value, the input string is assumed to
16950 be part of a form or query string and the '+' character will be turned into a
16951 space (' '). Otherwise this will only happen after a question mark indicating
16952 a query string ('?').
Thierry FOURNIER82ff3c92015-05-07 15:46:20 +020016953
William Dauchy888b0ae2021-01-06 23:39:50 +010016954url_enc([<enc_type>])
16955 Takes a string provided as input and returns the encoded version as output.
16956 The input and the output are of type string. By default the type of encoding
16957 is meant for `query` type. There is no other type supported for now but the
16958 optional argument is here for future changes.
16959
Frédéric Lécaille756d97f2019-03-04 19:03:48 +010016960ungrpc(<field_number>,[<field_type>])
Frédéric Lécaille50290fb2019-02-27 14:34:51 +010016961 This extracts the protocol buffers message field in raw mode of an input binary
Frédéric Lécaillebfe61382019-03-06 14:34:36 +010016962 sample representation of a gRPC message with <field_number> as field number
16963 (dotted notation) if <field_type> is not present, or as an integer sample if this
16964 field is present.
Frédéric Lécaille756d97f2019-03-04 19:03:48 +010016965 The list of the authorized types is the following one: "int32", "int64", "uint32",
16966 "uint64", "sint32", "sint64", "bool", "enum" for the "varint" wire type 0
16967 "fixed64", "sfixed64", "double" for the 64bit wire type 1, "fixed32", "sfixed32",
16968 "float" for the wire type 5. Note that "string" is considered as a length-delimited
Frédéric Lécaille93d33162019-03-06 09:35:59 +010016969 type, so it does not require any <field_type> argument to be extracted.
Frédéric Lécaille756d97f2019-03-04 19:03:48 +010016970 More information may be found here about the protocol buffers message field types:
16971 https://developers.google.com/protocol-buffers/docs/encoding
Frédéric Lécaille50290fb2019-02-27 14:34:51 +010016972
16973 Example:
16974 // with such a protocol buffer .proto file content adapted from
16975 // https://github.com/grpc/grpc/blob/master/examples/protos/route_guide.proto
16976
16977 message Point {
16978 int32 latitude = 1;
16979 int32 longitude = 2;
16980 }
16981
16982 message PPoint {
16983 Point point = 59;
16984 }
16985
16986 message Rectangle {
16987 // One corner of the rectangle.
16988 PPoint lo = 48;
16989 // The other corner of the rectangle.
16990 PPoint hi = 49;
16991 }
16992
Peter Gervaidf4c9d22020-06-11 18:05:11 +020016993 let's say a body request is made of a "Rectangle" object value (two PPoint
16994 protocol buffers messages), the four protocol buffers fields could be
16995 extracted with these "ungrpc" directives:
Frédéric Lécaille50290fb2019-02-27 14:34:51 +010016996
Frédéric Lécaille756d97f2019-03-04 19:03:48 +010016997 req.body,ungrpc(48.59.1,int32) # "latitude" of "lo" first PPoint
16998 req.body,ungrpc(48.59.2,int32) # "longitude" of "lo" first PPoint
John Roeslerfb2fce12019-07-10 15:45:51 -050016999 req.body,ungrpc(49.59.1,int32) # "latitude" of "hi" second PPoint
Frédéric Lécaille756d97f2019-03-04 19:03:48 +010017000 req.body,ungrpc(49.59.2,int32) # "longitude" of "hi" second PPoint
17001
Peter Gervaidf4c9d22020-06-11 18:05:11 +020017002 We could also extract the intermediary 48.59 field as a binary sample as follows:
Frédéric Lécaille756d97f2019-03-04 19:03:48 +010017003
Frédéric Lécaille93d33162019-03-06 09:35:59 +010017004 req.body,ungrpc(48.59)
Frédéric Lécaille756d97f2019-03-04 19:03:48 +010017005
Peter Gervaidf4c9d22020-06-11 18:05:11 +020017006 As a gRPC message is always made of a gRPC header followed by protocol buffers
17007 messages, in the previous example the "latitude" of "lo" first PPoint
17008 could be extracted with these equivalent directives:
Frédéric Lécaillebfe61382019-03-06 14:34:36 +010017009
17010 req.body,ungrpc(48.59),protobuf(1,int32)
17011 req.body,ungrpc(48),protobuf(59.1,int32)
17012 req.body,ungrpc(48),protobuf(59),protobuf(1,int32)
17013
Peter Gervaidf4c9d22020-06-11 18:05:11 +020017014 Note that the first convert must be "ungrpc", the remaining ones must be
17015 "protobuf" and only the last one may have or not a second argument to
17016 interpret the previous binary sample.
Frédéric Lécaillebfe61382019-03-06 14:34:36 +010017017
Frédéric Lécaille50290fb2019-02-27 14:34:51 +010017018
Tim Duesterhusef4e45c2021-01-21 17:40:50 +010017019unset-var(<var>)
Christopher Faulet85d79c92016-11-09 16:54:56 +010017020 Unsets a variable if the input content is defined. The name of the variable
17021 starts with an indication about its scope. The scopes allowed are:
17022 "proc" : the variable is shared with the whole process
17023 "sess" : the variable is shared with the whole session
17024 "txn" : the variable is shared with the transaction (request and
17025 response),
17026 "req" : the variable is shared only during request processing,
17027 "res" : the variable is shared only during response processing.
17028 This prefix is followed by a name. The separator is a '.'. The name may only
17029 contain characters 'a-z', 'A-Z', '0-9', '.' and '_'.
17030
Willy Tarreau0dbfdba2014-07-10 16:37:47 +020017031utime(<format>[,<offset>])
17032 Converts an integer supposed to contain a date since epoch to a string
17033 representing this date in UTC time using a format defined by the <format>
17034 string using strftime(3). The purpose is to allow any date format to be used
17035 in logs. An optional <offset> in seconds may be applied to the input date
17036 (positive or negative). See the strftime() man page for the format supported
17037 by your operating system. See also the ltime converter.
17038
17039 Example :
17040
17041 # Emit two colons, one with the UTC time and another with ip:port
Davor Ocelice9ed2812017-12-25 17:49:28 +010017042 # e.g. 20140710162350 127.0.0.1:57325
Willy Tarreau0dbfdba2014-07-10 16:37:47 +020017043 log-format %[date,utime(%Y%m%d%H%M%S)]\ %ci:%cp
17044
Marcin Deranek9631a282018-04-16 14:30:46 +020017045word(<index>,<delimiters>[,<count>])
17046 Extracts the nth word counting from the beginning (positive index) or from
17047 the end (negative index) considering given delimiters from an input string.
17048 Indexes start at 1 or -1 and delimiters are a string formatted list of chars.
Jerome Magnin88209322020-01-28 13:33:44 +010017049 Delimiters at the beginning or end of the input string are ignored.
Marcin Deranek9631a282018-04-16 14:30:46 +020017050 Optionally you can specify <count> of words to extract (default: 1).
17051 Value of 0 indicates extraction of all remaining words.
17052
17053 Example :
17054 str(f1_f2_f3__f5),word(4,_) # f5
17055 str(f1_f2_f3__f5),word(2,_,0) # f2_f3__f5
17056 str(f1_f2_f3__f5),word(3,_,2) # f3__f5
17057 str(f1_f2_f3__f5),word(-2,_,3) # f1_f2_f3
17058 str(f1_f2_f3__f5),word(-3,_,0) # f1_f2
Jerome Magnin88209322020-01-28 13:33:44 +010017059 str(/f1/f2/f3/f4),word(1,/) # f1
Emeric Brunc9a0f6d2014-11-25 14:09:01 +010017060
Willy Tarreau23ec4ca2014-07-15 20:15:37 +020017061wt6([<avalanche>])
17062 Hashes a binary input sample into an unsigned 32-bit quantity using the WT6
17063 hash function. Optionally, it is possible to apply a full avalanche hash
17064 function to the output if the optional <avalanche> argument equals 1. This
17065 converter uses the same functions as used by the various hash-based load
17066 balancing algorithms, so it will provide exactly the same results. It is
17067 mostly intended for debugging, but can be used as a stick-table entry to
17068 collect rough statistics. It must not be used for security purposes as a
Emmanuel Hocdet50791a72018-03-21 11:19:01 +010017069 32-bit hash is trivial to break. See also "crc32", "djb2", "sdbm", "crc32c",
17070 and the "hash-type" directive.
Willy Tarreau23ec4ca2014-07-15 20:15:37 +020017071
Willy Tarreau97707872015-01-27 15:12:13 +010017072xor(<value>)
17073 Performs a bitwise "XOR" (exclusive OR) between <value> and the input value
Thierry FOURNIER07ee64e2015-07-06 23:43:03 +020017074 of type signed integer, and returns the result as an signed integer.
Thierry FOURNIER5d86fae2015-07-07 21:10:16 +020017075 <value> can be a numeric value or a variable name. The name of the variable
Daniel Schneller0b547052016-03-21 20:46:57 +010017076 starts with an indication about its scope. The scopes allowed are:
Christopher Fauletff2613e2016-11-09 11:36:17 +010017077 "proc" : the variable is shared with the whole process
Daniel Schneller0b547052016-03-21 20:46:57 +010017078 "sess" : the variable is shared with the whole session
17079 "txn" : the variable is shared with the transaction (request and
Thierry FOURNIER5d86fae2015-07-07 21:10:16 +020017080 response),
Daniel Schneller0b547052016-03-21 20:46:57 +010017081 "req" : the variable is shared only during request processing,
17082 "res" : the variable is shared only during response processing.
Thierry FOURNIER5d86fae2015-07-07 21:10:16 +020017083 This prefix is followed by a name. The separator is a '.'. The name may only
Christopher Fauletb71557a2016-10-31 10:49:03 +010017084 contain characters 'a-z', 'A-Z', '0-9', '.' and '_'.
Willy Tarreau97707872015-01-27 15:12:13 +010017085
Dragan Dosen04bf0cc2020-12-22 21:44:33 +010017086xxh3([<seed>])
17087 Hashes a binary input sample into a signed 64-bit quantity using the XXH3
17088 64-bit variant of the XXhash hash function. This hash supports a seed which
17089 defaults to zero but a different value maybe passed as the <seed> argument.
17090 This hash is known to be very good and very fast so it can be used to hash
17091 URLs and/or URL parameters for use as stick-table keys to collect statistics
17092 with a low collision rate, though care must be taken as the algorithm is not
17093 considered as cryptographically secure.
17094
Thierry FOURNIER01e09742016-12-26 11:46:11 +010017095xxh32([<seed>])
17096 Hashes a binary input sample into an unsigned 32-bit quantity using the 32-bit
17097 variant of the XXHash hash function. This hash supports a seed which defaults
17098 to zero but a different value maybe passed as the <seed> argument. This hash
17099 is known to be very good and very fast so it can be used to hash URLs and/or
17100 URL parameters for use as stick-table keys to collect statistics with a low
17101 collision rate, though care must be taken as the algorithm is not considered
17102 as cryptographically secure.
17103
17104xxh64([<seed>])
17105 Hashes a binary input sample into a signed 64-bit quantity using the 64-bit
17106 variant of the XXHash hash function. This hash supports a seed which defaults
17107 to zero but a different value maybe passed as the <seed> argument. This hash
17108 is known to be very good and very fast so it can be used to hash URLs and/or
17109 URL parameters for use as stick-table keys to collect statistics with a low
17110 collision rate, though care must be taken as the algorithm is not considered
17111 as cryptographically secure.
17112
Thierry FOURNIERd5f624d2013-11-26 11:52:33 +010017113
Thierry FOURNIER060762e2014-04-23 13:29:15 +0200171147.3.2. Fetching samples from internal states
Willy Tarreau74ca5042013-06-11 23:12:07 +020017115--------------------------------------------
17116
17117A first set of sample fetch methods applies to internal information which does
17118not even relate to any client information. These ones are sometimes used with
17119"monitor-fail" directives to report an internal status to external watchers.
17120The sample fetch methods described in this section are usable anywhere.
17121
17122always_false : boolean
17123 Always returns the boolean "false" value. It may be used with ACLs as a
17124 temporary replacement for another one when adjusting configurations.
17125
17126always_true : boolean
17127 Always returns the boolean "true" value. It may be used with ACLs as a
17128 temporary replacement for another one when adjusting configurations.
17129
17130avg_queue([<backend>]) : integer
Willy Tarreaud63335a2010-02-26 12:56:52 +010017131 Returns the total number of queued connections of the designated backend
Willy Tarreau74ca5042013-06-11 23:12:07 +020017132 divided by the number of active servers. The current backend is used if no
17133 backend is specified. This is very similar to "queue" except that the size of
17134 the farm is considered, in order to give a more accurate measurement of the
17135 time it may take for a new connection to be processed. The main usage is with
17136 ACL to return a sorry page to new users when it becomes certain they will get
17137 a degraded service, or to pass to the backend servers in a header so that
17138 they decide to work in degraded mode or to disable some functions to speed up
17139 the processing a bit. Note that in the event there would not be any active
17140 server anymore, twice the number of queued connections would be considered as
17141 the measured value. This is a fair estimate, as we expect one server to get
17142 back soon anyway, but we still prefer to send new traffic to another backend
17143 if in better shape. See also the "queue", "be_conn", and "be_sess_rate"
17144 sample fetches.
Krzysztof Piotr Oledzki346f76d2010-01-12 21:59:30 +010017145
Willy Tarreau74ca5042013-06-11 23:12:07 +020017146be_conn([<backend>]) : integer
Willy Tarreaua36af912009-10-10 12:02:45 +020017147 Applies to the number of currently established connections on the backend,
17148 possibly including the connection being evaluated. If no backend name is
17149 specified, the current one is used. But it is also possible to check another
17150 backend. It can be used to use a specific farm when the nominal one is full.
Patrick Hemmer4cdf3ab2018-06-14 17:10:27 -040017151 See also the "fe_conn", "queue", "be_conn_free", and "be_sess_rate" criteria.
17152
17153be_conn_free([<backend>]) : integer
17154 Returns an integer value corresponding to the number of available connections
17155 across available servers in the backend. Queue slots are not included. Backup
17156 servers are also not included, unless all other servers are down. If no
17157 backend name is specified, the current one is used. But it is also possible
17158 to check another backend. It can be used to use a specific farm when the
Patrick Hemmer155e93e2018-06-14 18:01:35 -040017159 nominal one is full. See also the "be_conn", "connslots", and "srv_conn_free"
17160 criteria.
Patrick Hemmer4cdf3ab2018-06-14 17:10:27 -040017161
17162 OTHER CAVEATS AND NOTES: if any of the server maxconn, or maxqueue is 0
17163 (meaning unlimited), then this fetch clearly does not make sense, in which
17164 case the value returned will be -1.
Willy Tarreau6a06a402007-07-15 20:15:28 +020017165
Willy Tarreau74ca5042013-06-11 23:12:07 +020017166be_sess_rate([<backend>]) : integer
17167 Returns an integer value corresponding to the sessions creation rate on the
17168 backend, in number of new sessions per second. This is used with ACLs to
17169 switch to an alternate backend when an expensive or fragile one reaches too
Davor Ocelice9ed2812017-12-25 17:49:28 +010017170 high a session rate, or to limit abuse of service (e.g. prevent sucking of an
Willy Tarreau74ca5042013-06-11 23:12:07 +020017171 online dictionary). It can also be useful to add this element to logs using a
17172 log-format directive.
Willy Tarreaud63335a2010-02-26 12:56:52 +010017173
17174 Example :
17175 # Redirect to an error page if the dictionary is requested too often
17176 backend dynamic
17177 mode http
17178 acl being_scanned be_sess_rate gt 100
17179 redirect location /denied.html if being_scanned
Willy Tarreau0ba27502007-12-24 16:55:16 +010017180
Davor Ocelice9ed2812017-12-25 17:49:28 +010017181bin(<hex>) : bin
Thierry FOURNIERcc103292015-06-06 19:30:17 +020017182 Returns a binary chain. The input is the hexadecimal representation
17183 of the string.
17184
17185bool(<bool>) : bool
17186 Returns a boolean value. <bool> can be 'true', 'false', '1' or '0'.
17187 'false' and '0' are the same. 'true' and '1' are the same.
17188
Willy Tarreau74ca5042013-06-11 23:12:07 +020017189connslots([<backend>]) : integer
17190 Returns an integer value corresponding to the number of connection slots
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030017191 still available in the backend, by totaling the maximum amount of
Willy Tarreau74ca5042013-06-11 23:12:07 +020017192 connections on all servers and the maximum queue size. This is probably only
17193 used with ACLs.
Tait Clarridge7896d522012-12-05 21:39:31 -050017194
Jeffrey 'jf' Lim5051d7b2008-09-04 01:03:03 +080017195 The basic idea here is to be able to measure the number of connection "slots"
Willy Tarreau55165fe2009-05-10 12:02:55 +020017196 still available (connection + queue), so that anything beyond that (intended
Jeffrey 'jf' Lim5051d7b2008-09-04 01:03:03 +080017197 usage; see "use_backend" keyword) can be redirected to a different backend.
17198
Willy Tarreau55165fe2009-05-10 12:02:55 +020017199 'connslots' = number of available server connection slots, + number of
17200 available server queue slots.
Jeffrey 'jf' Lim5051d7b2008-09-04 01:03:03 +080017201
Willy Tarreaua36af912009-10-10 12:02:45 +020017202 Note that while "fe_conn" may be used, "connslots" comes in especially
Willy Tarreau55165fe2009-05-10 12:02:55 +020017203 useful when you have a case of traffic going to one single ip, splitting into
Willy Tarreau74ca5042013-06-11 23:12:07 +020017204 multiple backends (perhaps using ACLs to do name-based load balancing) and
Willy Tarreau55165fe2009-05-10 12:02:55 +020017205 you want to be able to differentiate between different backends, and their
Davor Ocelice9ed2812017-12-25 17:49:28 +010017206 available "connslots". Also, whereas "nbsrv" only measures servers that are
Willy Tarreau74ca5042013-06-11 23:12:07 +020017207 actually *down*, this fetch is more fine-grained and looks into the number of
Willy Tarreaua36af912009-10-10 12:02:45 +020017208 available connection slots as well. See also "queue" and "avg_queue".
Jeffrey 'jf' Lim5051d7b2008-09-04 01:03:03 +080017209
Willy Tarreau55165fe2009-05-10 12:02:55 +020017210 OTHER CAVEATS AND NOTES: at this point in time, the code does not take care
17211 of dynamic connections. Also, if any of the server maxconn, or maxqueue is 0,
Willy Tarreau74ca5042013-06-11 23:12:07 +020017212 then this fetch clearly does not make sense, in which case the value returned
Willy Tarreau55165fe2009-05-10 12:02:55 +020017213 will be -1.
Jeffrey 'jf' Lim5051d7b2008-09-04 01:03:03 +080017214
Willy Tarreau70fe9442018-11-22 16:07:39 +010017215cpu_calls : integer
17216 Returns the number of calls to the task processing the stream or current
17217 request since it was allocated. This number is reset for each new request on
17218 the same connections in case of HTTP keep-alive. This value should usually be
17219 low and stable (around 2 calls for a typically simple request) but may become
17220 high if some processing (compression, caching or analysis) is performed. This
17221 is purely for performance monitoring purposes.
17222
17223cpu_ns_avg : integer
17224 Returns the average number of nanoseconds spent in each call to the task
17225 processing the stream or current request. This number is reset for each new
17226 request on the same connections in case of HTTP keep-alive. This value
17227 indicates the overall cost of processing the request or the connection for
17228 each call. There is no good nor bad value but the time spent in a call
17229 automatically causes latency for other processing (see lat_ns_avg below),
17230 and may affect other connection's apparent response time. Certain operations
17231 like compression, complex regex matching or heavy Lua operations may directly
17232 affect this value, and having it in the logs will make it easier to spot the
17233 faulty processing that needs to be fixed to recover decent performance.
17234 Note: this value is exactly cpu_ns_tot divided by cpu_calls.
17235
17236cpu_ns_tot : integer
17237 Returns the total number of nanoseconds spent in each call to the task
17238 processing the stream or current request. This number is reset for each new
17239 request on the same connections in case of HTTP keep-alive. This value
17240 indicates the overall cost of processing the request or the connection for
17241 each call. There is no good nor bad value but the time spent in a call
17242 automatically causes latency for other processing (see lat_ns_avg below),
17243 induces CPU costs on the machine, and may affect other connection's apparent
17244 response time. Certain operations like compression, complex regex matching or
17245 heavy Lua operations may directly affect this value, and having it in the
17246 logs will make it easier to spot the faulty processing that needs to be fixed
17247 to recover decent performance. The value may be artificially high due to a
17248 high cpu_calls count, for example when processing many HTTP chunks, and for
17249 this reason it is often preferred to log cpu_ns_avg instead.
17250
Cyril Bonté6bcd1822019-11-05 23:13:59 +010017251date([<offset>],[<unit>]) : integer
Willy Tarreau6236d3a2013-07-25 14:28:25 +020017252 Returns the current date as the epoch (number of seconds since 01/01/1970).
Damien Claisseae6f1252019-10-30 15:57:28 +000017253
17254 If an offset value is specified, then it is added to the current date before
17255 returning the value. This is particularly useful to compute relative dates,
17256 as both positive and negative offsets are allowed.
Willy Tarreau276fae92013-07-25 14:36:01 +020017257 It is useful combined with the http_date converter.
17258
Damien Claisseae6f1252019-10-30 15:57:28 +000017259 <unit> is facultative, and can be set to "s" for seconds (default behavior),
17260 "ms" for milliseconds or "us" for microseconds.
17261 If unit is set, return value is an integer reflecting either seconds,
17262 milliseconds or microseconds since epoch, plus offset.
17263 It is useful when a time resolution of less than a second is needed.
17264
Willy Tarreau276fae92013-07-25 14:36:01 +020017265 Example :
17266
17267 # set an expires header to now+1 hour in every response
17268 http-response set-header Expires %[date(3600),http_date]
Willy Tarreau6236d3a2013-07-25 14:28:25 +020017269
Damien Claisseae6f1252019-10-30 15:57:28 +000017270 # set an expires header to now+1 hour in every response, with
17271 # millisecond granularity
17272 http-response set-header Expires %[date(3600000,ms),http_date(0,ms)]
17273
Etienne Carrierea792a0a2018-01-17 13:43:24 +010017274date_us : integer
17275 Return the microseconds part of the date (the "second" part is returned by
17276 date sample). This sample is coherent with the date sample as it is comes
17277 from the same timeval structure.
17278
Willy Tarreaud716f9b2017-10-13 11:03:15 +020017279distcc_body(<token>[,<occ>]) : binary
17280 Parses a distcc message and returns the body associated to occurrence #<occ>
17281 of the token <token>. Occurrences start at 1, and when unspecified, any may
17282 match though in practice only the first one is checked for now. This can be
17283 used to extract file names or arguments in files built using distcc through
Daniel Corbett9f0843f2021-05-08 10:50:37 -040017284 HAProxy. Please refer to distcc's protocol documentation for the complete
Willy Tarreaud716f9b2017-10-13 11:03:15 +020017285 list of supported tokens.
17286
17287distcc_param(<token>[,<occ>]) : integer
17288 Parses a distcc message and returns the parameter associated to occurrence
17289 #<occ> of the token <token>. Occurrences start at 1, and when unspecified,
17290 any may match though in practice only the first one is checked for now. This
17291 can be used to extract certain information such as the protocol version, the
Daniel Corbett9f0843f2021-05-08 10:50:37 -040017292 file size or the argument in files built using distcc through HAProxy.
Willy Tarreaud716f9b2017-10-13 11:03:15 +020017293 Another use case consists in waiting for the start of the preprocessed file
17294 contents before connecting to the server to avoid keeping idle connections.
17295 Please refer to distcc's protocol documentation for the complete list of
17296 supported tokens.
17297
17298 Example :
17299 # wait up to 20s for the pre-processed file to be uploaded
17300 tcp-request inspect-delay 20s
17301 tcp-request content accept if { distcc_param(DOTI) -m found }
17302 # send large files to the big farm
17303 use_backend big_farm if { distcc_param(DOTI) gt 1000000 }
17304
Willy Tarreau595ec542013-06-12 21:34:28 +020017305env(<name>) : string
17306 Returns a string containing the value of environment variable <name>. As a
17307 reminder, environment variables are per-process and are sampled when the
17308 process starts. This can be useful to pass some information to a next hop
17309 server, or with ACLs to take specific action when the process is started a
17310 certain way.
17311
17312 Examples :
17313 # Pass the Via header to next hop with the local hostname in it
17314 http-request add-header Via 1.1\ %[env(HOSTNAME)]
17315
17316 # reject cookie-less requests when the STOP environment variable is set
17317 http-request deny if !{ cook(SESSIONID) -m found } { env(STOP) -m found }
17318
Willy Tarreau74ca5042013-06-11 23:12:07 +020017319fe_conn([<frontend>]) : integer
17320 Returns the number of currently established connections on the frontend,
Willy Tarreaud63335a2010-02-26 12:56:52 +010017321 possibly including the connection being evaluated. If no frontend name is
17322 specified, the current one is used. But it is also possible to check another
Willy Tarreau74ca5042013-06-11 23:12:07 +020017323 frontend. It can be used to return a sorry page before hard-blocking, or to
17324 use a specific backend to drain new requests when the farm is considered
Davor Ocelice9ed2812017-12-25 17:49:28 +010017325 full. This is mostly used with ACLs but can also be used to pass some
Willy Tarreau74ca5042013-06-11 23:12:07 +020017326 statistics to servers in HTTP headers. See also the "dst_conn", "be_conn",
17327 "fe_sess_rate" fetches.
Willy Tarreaua36af912009-10-10 12:02:45 +020017328
Nenad Merdanovicad9a7e92016-10-03 04:57:37 +020017329fe_req_rate([<frontend>]) : integer
17330 Returns an integer value corresponding to the number of HTTP requests per
17331 second sent to a frontend. This number can differ from "fe_sess_rate" in
17332 situations where client-side keep-alive is enabled.
17333
Willy Tarreau74ca5042013-06-11 23:12:07 +020017334fe_sess_rate([<frontend>]) : integer
17335 Returns an integer value corresponding to the sessions creation rate on the
17336 frontend, in number of new sessions per second. This is used with ACLs to
17337 limit the incoming session rate to an acceptable range in order to prevent
17338 abuse of service at the earliest moment, for example when combined with other
17339 layer 4 ACLs in order to force the clients to wait a bit for the rate to go
17340 down below the limit. It can also be useful to add this element to logs using
17341 a log-format directive. See also the "rate-limit sessions" directive for use
17342 in frontends.
Willy Tarreau079ff0a2009-03-05 21:34:28 +010017343
17344 Example :
17345 # This frontend limits incoming mails to 10/s with a max of 100
17346 # concurrent connections. We accept any connection below 10/s, and
17347 # force excess clients to wait for 100 ms. Since clients are limited to
17348 # 100 max, there cannot be more than 10 incoming mails per second.
17349 frontend mail
17350 bind :25
17351 mode tcp
17352 maxconn 100
17353 acl too_fast fe_sess_rate ge 10
17354 tcp-request inspect-delay 100ms
17355 tcp-request content accept if ! too_fast
17356 tcp-request content accept if WAIT_END
Willy Tarreaud72758d2010-01-12 10:42:19 +010017357
Nenad Merdanovic807a6e72017-03-12 22:00:00 +010017358hostname : string
17359 Returns the system hostname.
17360
Thierry FOURNIER07ee64e2015-07-06 23:43:03 +020017361int(<integer>) : signed integer
17362 Returns a signed integer.
17363
Thierry FOURNIERcc103292015-06-06 19:30:17 +020017364ipv4(<ipv4>) : ipv4
17365 Returns an ipv4.
17366
17367ipv6(<ipv6>) : ipv6
17368 Returns an ipv6.
17369
Willy Tarreau70fe9442018-11-22 16:07:39 +010017370lat_ns_avg : integer
17371 Returns the average number of nanoseconds spent between the moment the task
17372 handling the stream is woken up and the moment it is effectively called. This
17373 number is reset for each new request on the same connections in case of HTTP
17374 keep-alive. This value indicates the overall latency inflicted to the current
17375 request by all other requests being processed in parallel, and is a direct
17376 indicator of perceived performance due to noisy neighbours. In order to keep
17377 the value low, it is possible to reduce the scheduler's run queue depth using
17378 "tune.runqueue-depth", to reduce the number of concurrent events processed at
17379 once using "tune.maxpollevents", to decrease the stream's nice value using
Willy Tarreaue7723bd2020-06-24 11:11:02 +020017380 the "nice" option on the "bind" lines or in the frontend, to enable low
17381 latency scheduling using "tune.sched.low-latency", or to look for other heavy
17382 requests in logs (those exhibiting large values of "cpu_ns_avg"), whose
17383 processing needs to be adjusted or fixed. Compression of large buffers could
17384 be a culprit, like heavy regex or long lists of regex. Note: this value is
17385 exactly lat_ns_tot divided by cpu_calls.
Willy Tarreau70fe9442018-11-22 16:07:39 +010017386
17387lat_ns_tot : integer
17388 Returns the total number of nanoseconds spent between the moment the task
17389 handling the stream is woken up and the moment it is effectively called. This
17390 number is reset for each new request on the same connections in case of HTTP
17391 keep-alive. This value indicates the overall latency inflicted to the current
17392 request by all other requests being processed in parallel, and is a direct
17393 indicator of perceived performance due to noisy neighbours. In order to keep
17394 the value low, it is possible to reduce the scheduler's run queue depth using
17395 "tune.runqueue-depth", to reduce the number of concurrent events processed at
17396 once using "tune.maxpollevents", to decrease the stream's nice value using
Willy Tarreaue7723bd2020-06-24 11:11:02 +020017397 the "nice" option on the "bind" lines or in the frontend, to enable low
17398 latency scheduling using "tune.sched.low-latency", or to look for other heavy
17399 requests in logs (those exhibiting large values of "cpu_ns_avg"), whose
17400 processing needs to be adjusted or fixed. Compression of large buffers could
17401 be a culprit, like heavy regex or long lists of regex. Note: while it
Willy Tarreau70fe9442018-11-22 16:07:39 +010017402 may intuitively seem that the total latency adds to a transfer time, it is
17403 almost never true because while a task waits for the CPU, network buffers
17404 continue to fill up and the next call will process more at once. The value
17405 may be artificially high due to a high cpu_calls count, for example when
17406 processing many HTTP chunks, and for this reason it is often preferred to log
17407 lat_ns_avg instead, which is a more relevant performance indicator.
17408
Thierry FOURNIERcc103292015-06-06 19:30:17 +020017409meth(<method>) : method
17410 Returns a method.
17411
Willy Tarreau0f30d262014-11-24 16:02:05 +010017412nbproc : integer
17413 Returns an integer value corresponding to the number of processes that were
17414 started (it equals the global "nbproc" setting). This is useful for logging
17415 and debugging purposes.
17416
Willy Tarreau74ca5042013-06-11 23:12:07 +020017417nbsrv([<backend>]) : integer
17418 Returns an integer value corresponding to the number of usable servers of
17419 either the current backend or the named backend. This is mostly used with
17420 ACLs but can also be useful when added to logs. This is normally used to
Willy Tarreaud63335a2010-02-26 12:56:52 +010017421 switch to an alternate backend when the number of servers is too low to
17422 to handle some load. It is useful to report a failure when combined with
17423 "monitor fail".
Willy Tarreau079ff0a2009-03-05 21:34:28 +010017424
Patrick Hemmerfabb24f2018-08-13 14:07:57 -040017425prio_class : integer
17426 Returns the priority class of the current session for http mode or connection
17427 for tcp mode. The value will be that set by the last call to "http-request
17428 set-priority-class" or "tcp-request content set-priority-class".
17429
17430prio_offset : integer
17431 Returns the priority offset of the current session for http mode or
17432 connection for tcp mode. The value will be that set by the last call to
17433 "http-request set-priority-offset" or "tcp-request content
17434 set-priority-offset".
17435
Willy Tarreau0f30d262014-11-24 16:02:05 +010017436proc : integer
17437 Returns an integer value corresponding to the position of the process calling
17438 the function, between 1 and global.nbproc. This is useful for logging and
17439 debugging purposes.
17440
Willy Tarreau74ca5042013-06-11 23:12:07 +020017441queue([<backend>]) : integer
Willy Tarreaud63335a2010-02-26 12:56:52 +010017442 Returns the total number of queued connections of the designated backend,
17443 including all the connections in server queues. If no backend name is
17444 specified, the current one is used, but it is also possible to check another
Willy Tarreau74ca5042013-06-11 23:12:07 +020017445 one. This is useful with ACLs or to pass statistics to backend servers. This
17446 can be used to take actions when queuing goes above a known level, generally
17447 indicating a surge of traffic or a massive slowdown on the servers. One
17448 possible action could be to reject new users but still accept old ones. See
17449 also the "avg_queue", "be_conn", and "be_sess_rate" fetches.
17450
Willy Tarreau84310e22014-02-14 11:59:04 +010017451rand([<range>]) : integer
17452 Returns a random integer value within a range of <range> possible values,
17453 starting at zero. If the range is not specified, it defaults to 2^32, which
17454 gives numbers between 0 and 4294967295. It can be useful to pass some values
17455 needed to take some routing decisions for example, or just for debugging
17456 purposes. This random must not be used for security purposes.
17457
Willy Tarreau74ca5042013-06-11 23:12:07 +020017458srv_conn([<backend>/]<server>) : integer
17459 Returns an integer value corresponding to the number of currently established
17460 connections on the designated server, possibly including the connection being
17461 evaluated. If <backend> is omitted, then the server is looked up in the
17462 current backend. It can be used to use a specific farm when one server is
17463 full, or to inform the server about our view of the number of active
Patrick Hemmer155e93e2018-06-14 18:01:35 -040017464 connections with it. See also the "fe_conn", "be_conn", "queue", and
17465 "srv_conn_free" fetch methods.
17466
17467srv_conn_free([<backend>/]<server>) : integer
17468 Returns an integer value corresponding to the number of available connections
17469 on the designated server, possibly including the connection being evaluated.
17470 The value does not include queue slots. If <backend> is omitted, then the
17471 server is looked up in the current backend. It can be used to use a specific
17472 farm when one server is full, or to inform the server about our view of the
17473 number of active connections with it. See also the "be_conn_free" and
17474 "srv_conn" fetch methods.
17475
17476 OTHER CAVEATS AND NOTES: If the server maxconn is 0, then this fetch clearly
17477 does not make sense, in which case the value returned will be -1.
Willy Tarreau74ca5042013-06-11 23:12:07 +020017478
17479srv_is_up([<backend>/]<server>) : boolean
17480 Returns true when the designated server is UP, and false when it is either
17481 DOWN or in maintenance mode. If <backend> is omitted, then the server is
17482 looked up in the current backend. It is mainly used to take action based on
Davor Ocelice9ed2812017-12-25 17:49:28 +010017483 an external status reported via a health check (e.g. a geographical site's
Willy Tarreau74ca5042013-06-11 23:12:07 +020017484 availability). Another possible use which is more of a hack consists in
17485 using dummy servers as boolean variables that can be enabled or disabled from
17486 the CLI, so that rules depending on those ACLs can be tweaked in realtime.
17487
Willy Tarreauff2b7af2017-10-13 11:46:26 +020017488srv_queue([<backend>/]<server>) : integer
17489 Returns an integer value corresponding to the number of connections currently
17490 pending in the designated server's queue. If <backend> is omitted, then the
17491 server is looked up in the current backend. It can sometimes be used together
17492 with the "use-server" directive to force to use a known faster server when it
17493 is not much loaded. See also the "srv_conn", "avg_queue" and "queue" sample
17494 fetch methods.
17495
Willy Tarreau74ca5042013-06-11 23:12:07 +020017496srv_sess_rate([<backend>/]<server>) : integer
17497 Returns an integer corresponding to the sessions creation rate on the
17498 designated server, in number of new sessions per second. If <backend> is
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030017499 omitted, then the server is looked up in the current backend. This is mostly
Willy Tarreau74ca5042013-06-11 23:12:07 +020017500 used with ACLs but can make sense with logs too. This is used to switch to an
17501 alternate backend when an expensive or fragile one reaches too high a session
Davor Ocelice9ed2812017-12-25 17:49:28 +010017502 rate, or to limit abuse of service (e.g. prevent latent requests from
Willy Tarreau74ca5042013-06-11 23:12:07 +020017503 overloading servers).
17504
17505 Example :
17506 # Redirect to a separate back
17507 acl srv1_full srv_sess_rate(be1/srv1) gt 50
17508 acl srv2_full srv_sess_rate(be1/srv2) gt 50
17509 use_backend be2 if srv1_full or srv2_full
17510
Alexbf1bd5a2021-04-24 13:02:21 +020017511srv_iweight([<backend>/]<server>) : integer
Christopher Faulet1bea8652020-07-10 16:03:45 +020017512 Returns an integer corresponding to the server's initial weight. If <backend>
17513 is omitted, then the server is looked up in the current backend. See also
17514 "srv_weight" and "srv_uweight".
17515
Alexbf1bd5a2021-04-24 13:02:21 +020017516srv_uweight([<backend>/]<server>) : integer
Christopher Faulet1bea8652020-07-10 16:03:45 +020017517 Returns an integer corresponding to the user visible server's weight. If
17518 <backend> is omitted, then the server is looked up in the current
17519 backend. See also "srv_weight" and "srv_iweight".
17520
Alexbf1bd5a2021-04-24 13:02:21 +020017521srv_weight([<backend>/]<server>) : integer
Christopher Faulet1bea8652020-07-10 16:03:45 +020017522 Returns an integer corresponding to the current (or effective) server's
17523 weight. If <backend> is omitted, then the server is looked up in the current
17524 backend. See also "srv_iweight" and "srv_uweight".
17525
Willy Tarreau0f30d262014-11-24 16:02:05 +010017526stopping : boolean
17527 Returns TRUE if the process calling the function is currently stopping. This
17528 can be useful for logging, or for relaxing certain checks or helping close
17529 certain connections upon graceful shutdown.
17530
Thierry FOURNIERcc103292015-06-06 19:30:17 +020017531str(<string>) : string
17532 Returns a string.
17533
Willy Tarreau74ca5042013-06-11 23:12:07 +020017534table_avl([<table>]) : integer
17535 Returns the total number of available entries in the current proxy's
17536 stick-table or in the designated stick-table. See also table_cnt.
17537
17538table_cnt([<table>]) : integer
17539 Returns the total number of entries currently in use in the current proxy's
17540 stick-table or in the designated stick-table. See also src_conn_cnt and
17541 table_avl for other entry counting methods.
17542
Christopher Faulet34adb2a2017-11-21 21:45:38 +010017543thread : integer
17544 Returns an integer value corresponding to the position of the thread calling
17545 the function, between 0 and (global.nbthread-1). This is useful for logging
17546 and debugging purposes.
17547
Alexandar Lazica429ad32021-06-01 00:27:01 +020017548uuid([<version>]) : string
17549 Returns a UUID following the RFC4122 standard. If the version is not
17550 specified, a UUID version 4 (fully random) is returned.
17551 Currently, only version 4 is supported.
17552
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020017553var(<var-name>) : undefined
17554 Returns a variable with the stored type. If the variable is not set, the
Daniel Schneller0b547052016-03-21 20:46:57 +010017555 sample fetch fails. The name of the variable starts with an indication
17556 about its scope. The scopes allowed are:
Christopher Fauletff2613e2016-11-09 11:36:17 +010017557 "proc" : the variable is shared with the whole process
Daniel Schneller0b547052016-03-21 20:46:57 +010017558 "sess" : the variable is shared with the whole session
17559 "txn" : the variable is shared with the transaction (request and
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020017560 response),
Daniel Schneller0b547052016-03-21 20:46:57 +010017561 "req" : the variable is shared only during request processing,
17562 "res" : the variable is shared only during response processing.
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020017563 This prefix is followed by a name. The separator is a '.'. The name may only
Christopher Fauletb71557a2016-10-31 10:49:03 +010017564 contain characters 'a-z', 'A-Z', '0-9', '.' and '_'.
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020017565
Thierry FOURNIER060762e2014-04-23 13:29:15 +0200175667.3.3. Fetching samples at Layer 4
Willy Tarreau74ca5042013-06-11 23:12:07 +020017567----------------------------------
17568
Daniel Corbett9f0843f2021-05-08 10:50:37 -040017569The layer 4 usually describes just the transport layer which in HAProxy is
Willy Tarreau74ca5042013-06-11 23:12:07 +020017570closest to the connection, where no content is yet made available. The fetch
17571methods described here are usable as low as the "tcp-request connection" rule
17572sets unless they require some future information. Those generally include
17573TCP/IP addresses and ports, as well as elements from stick-tables related to
Willy Tarreau4d4149c2013-07-23 19:33:46 +020017574the incoming connection. For retrieving a value from a sticky counters, the
17575counter number can be explicitly set as 0, 1, or 2 using the pre-defined
Moemen MHEDHBI9cf46342018-09-25 17:50:53 +020017576"sc0_", "sc1_", or "sc2_" prefix. These three pre-defined prefixes can only be
17577used if MAX_SESS_STKCTR value does not exceed 3, otherwise the counter number
17578can be specified as the first integer argument when using the "sc_" prefix.
17579Starting from "sc_0" to "sc_N" where N is (MAX_SESS_STKCTR-1). An optional
17580table may be specified with the "sc*" form, in which case the currently
17581tracked key will be looked up into this alternate table instead of the table
17582currently being tracked.
Willy Tarreau74ca5042013-06-11 23:12:07 +020017583
Christopher Faulet7d081f02021-04-15 09:38:37 +020017584bc_dst : ip
17585 This is the destination ip address of the connection on the server side,
17586 which is the server address HAProxy connected to. It is of type IP and works
17587 on both IPv4 and IPv6 tables. On IPv6 tables, IPv4 address is mapped to its
17588 IPv6 equivalent, according to RFC 4291.
17589
17590bc_dst_port : integer
17591 Returns an integer value corresponding to the destination TCP port of the
Daniel Corbett9f0843f2021-05-08 10:50:37 -040017592 connection on the server side, which is the port HAProxy connected to.
Christopher Faulet7d081f02021-04-15 09:38:37 +020017593
Jérôme Magnin35e53a62019-01-16 14:38:37 +010017594bc_http_major : integer
Jérôme Magnin86577422018-12-07 09:03:11 +010017595 Returns the backend connection's HTTP major version encoding, which may be 1
17596 for HTTP/0.9 to HTTP/1.1 or 2 for HTTP/2. Note, this is based on the on-wire
17597 encoding and not the version present in the request header.
17598
Christopher Faulet7d081f02021-04-15 09:38:37 +020017599bc_src : ip
17600 This is the source ip address of the connection on the server side, which is
Daniel Corbett9f0843f2021-05-08 10:50:37 -040017601 the server address HAProxy connected from. It is of type IP and works on both
Christopher Faulet7d081f02021-04-15 09:38:37 +020017602 IPv4 and IPv6 tables. On IPv6 tables, IPv4 addresses are mapped to their IPv6
17603 equivalent, according to RFC 4291.
17604
17605bc_src_port : integer
17606 Returns an integer value corresponding to the TCP source port of the
Daniel Corbett9f0843f2021-05-08 10:50:37 -040017607 connection on the server side, which is the port HAProxy connected from.
Christopher Faulet7d081f02021-04-15 09:38:37 +020017608
Willy Tarreau74ca5042013-06-11 23:12:07 +020017609be_id : integer
17610 Returns an integer containing the current backend's id. It can be used in
Christopher Fauletd1b44642020-04-30 09:51:15 +020017611 frontends with responses to check which backend processed the request. It can
17612 also be used in a tcp-check or an http-check ruleset.
Willy Tarreau74ca5042013-06-11 23:12:07 +020017613
Marcin Deranekd2471c22016-12-12 14:08:05 +010017614be_name : string
17615 Returns a string containing the current backend's name. It can be used in
Christopher Fauletd1b44642020-04-30 09:51:15 +020017616 frontends with responses to check which backend processed the request. It can
17617 also be used in a tcp-check or an http-check ruleset.
Marcin Deranekd2471c22016-12-12 14:08:05 +010017618
Amaury Denoyelled91d7792020-12-10 13:43:56 +010017619be_server_timeout : integer
17620 Returns the configuration value in millisecond for the server timeout of the
17621 current backend. This timeout can be overwritten by a "set-timeout" rule. See
17622 also the "cur_server_timeout".
17623
17624be_tunnel_timeout : integer
17625 Returns the configuration value in millisecond for the tunnel timeout of the
17626 current backend. This timeout can be overwritten by a "set-timeout" rule. See
17627 also the "cur_tunnel_timeout".
17628
Amaury Denoyellef7719a22020-12-10 13:43:58 +010017629cur_server_timeout : integer
17630 Returns the currently applied server timeout in millisecond for the stream.
17631 In the default case, this will be equal to be_server_timeout unless a
17632 "set-timeout" rule has been applied. See also "be_server_timeout".
17633
17634cur_tunnel_timeout : integer
17635 Returns the currently applied tunnel timeout in millisecond for the stream.
17636 In the default case, this will be equal to be_tunnel_timeout unless a
17637 "set-timeout" rule has been applied. See also "be_tunnel_timeout".
17638
Willy Tarreau74ca5042013-06-11 23:12:07 +020017639dst : ip
17640 This is the destination IPv4 address of the connection on the client side,
17641 which is the address the client connected to. It can be useful when running
17642 in transparent mode. It is of type IP and works on both IPv4 and IPv6 tables.
17643 On IPv6 tables, IPv4 address is mapped to its IPv6 equivalent, according to
Willy Tarreau64ded3d2019-01-23 10:02:15 +010017644 RFC 4291. When the incoming connection passed through address translation or
17645 redirection involving connection tracking, the original destination address
17646 before the redirection will be reported. On Linux systems, the source and
17647 destination may seldom appear reversed if the nf_conntrack_tcp_loose sysctl
17648 is set, because a late response may reopen a timed out connection and switch
17649 what is believed to be the source and the destination.
Willy Tarreau74ca5042013-06-11 23:12:07 +020017650
17651dst_conn : integer
17652 Returns an integer value corresponding to the number of currently established
17653 connections on the same socket including the one being evaluated. It is
17654 normally used with ACLs but can as well be used to pass the information to
17655 servers in an HTTP header or in logs. It can be used to either return a sorry
17656 page before hard-blocking, or to use a specific backend to drain new requests
17657 when the socket is considered saturated. This offers the ability to assign
17658 different limits to different listening ports or addresses. See also the
17659 "fe_conn" and "be_conn" fetches.
Willy Tarreaud63335a2010-02-26 12:56:52 +010017660
Willy Tarreau16e01562016-08-09 16:46:18 +020017661dst_is_local : boolean
17662 Returns true if the destination address of the incoming connection is local
17663 to the system, or false if the address doesn't exist on the system, meaning
17664 that it was intercepted in transparent mode. It can be useful to apply
17665 certain rules by default to forwarded traffic and other rules to the traffic
Davor Ocelice9ed2812017-12-25 17:49:28 +010017666 targeting the real address of the machine. For example the stats page could
Willy Tarreau16e01562016-08-09 16:46:18 +020017667 be delivered only on this address, or SSH access could be locally redirected.
17668 Please note that the check involves a few system calls, so it's better to do
17669 it only once per connection.
17670
Willy Tarreau74ca5042013-06-11 23:12:07 +020017671dst_port : integer
17672 Returns an integer value corresponding to the destination TCP port of the
17673 connection on the client side, which is the port the client connected to.
17674 This might be used when running in transparent mode, when assigning dynamic
17675 ports to some clients for a whole application session, to stick all users to
17676 a same server, or to pass the destination port information to a server using
17677 an HTTP header.
17678
Willy Tarreau60ca10a2017-08-18 15:26:54 +020017679fc_http_major : integer
17680 Reports the front connection's HTTP major version encoding, which may be 1
17681 for HTTP/0.9 to HTTP/1.1 or 2 for HTTP/2. Note, this is based on the on-wire
17682 encoding and not on the version present in the request header.
17683
Geoff Simmons7185b782019-08-27 18:31:16 +020017684fc_pp_authority : string
17685 Returns the authority TLV sent by the client in the PROXY protocol header,
17686 if any.
17687
Tim Duesterhusd1b15b62020-03-13 12:34:23 +010017688fc_pp_unique_id : string
17689 Returns the unique ID TLV sent by the client in the PROXY protocol header,
17690 if any.
17691
Emeric Brun4f603012017-01-05 15:11:44 +010017692fc_rcvd_proxy : boolean
17693 Returns true if the client initiated the connection with a PROXY protocol
17694 header.
17695
Thierry Fournier / OZON.IO6310bef2016-07-24 20:16:50 +020017696fc_rtt(<unit>) : integer
17697 Returns the Round Trip Time (RTT) measured by the kernel for the client
17698 connection. <unit> is facultative, by default the unit is milliseconds. <unit>
17699 can be set to "ms" for milliseconds or "us" for microseconds. If the server
17700 connection is not established, if the connection is not TCP or if the
17701 operating system does not support TCP_INFO, for example Linux kernels before
17702 2.4, the sample fetch fails.
17703
17704fc_rttvar(<unit>) : integer
17705 Returns the Round Trip Time (RTT) variance measured by the kernel for the
17706 client connection. <unit> is facultative, by default the unit is milliseconds.
17707 <unit> can be set to "ms" for milliseconds or "us" for microseconds. If the
17708 server connection is not established, if the connection is not TCP or if the
17709 operating system does not support TCP_INFO, for example Linux kernels before
17710 2.4, the sample fetch fails.
17711
Christopher Fauletba0c53e2019-10-17 14:40:48 +020017712fc_unacked : integer
Joe Williams30fcd392016-08-10 07:06:44 -070017713 Returns the unacked counter measured by the kernel for the client connection.
17714 If the server connection is not established, if the connection is not TCP or
17715 if the operating system does not support TCP_INFO, for example Linux kernels
17716 before 2.4, the sample fetch fails.
17717
Christopher Fauletba0c53e2019-10-17 14:40:48 +020017718fc_sacked : integer
Joe Williams30fcd392016-08-10 07:06:44 -070017719 Returns the sacked counter measured by the kernel for the client connection.
17720 If the server connection is not established, if the connection is not TCP or
17721 if the operating system does not support TCP_INFO, for example Linux kernels
17722 before 2.4, the sample fetch fails.
17723
Christopher Fauletba0c53e2019-10-17 14:40:48 +020017724fc_retrans : integer
Joe Williams30fcd392016-08-10 07:06:44 -070017725 Returns the retransmits counter measured by the kernel for the client
17726 connection. If the server connection is not established, if the connection is
17727 not TCP or if the operating system does not support TCP_INFO, for example
17728 Linux kernels before 2.4, the sample fetch fails.
17729
Christopher Fauletba0c53e2019-10-17 14:40:48 +020017730fc_fackets : integer
Joe Williams30fcd392016-08-10 07:06:44 -070017731 Returns the fack counter measured by the kernel for the client
17732 connection. If the server connection is not established, if the connection is
17733 not TCP or if the operating system does not support TCP_INFO, for example
17734 Linux kernels before 2.4, the sample fetch fails.
17735
Christopher Fauletba0c53e2019-10-17 14:40:48 +020017736fc_lost : integer
Joe Williams30fcd392016-08-10 07:06:44 -070017737 Returns the lost counter measured by the kernel for the client
17738 connection. If the server connection is not established, if the connection is
17739 not TCP or if the operating system does not support TCP_INFO, for example
17740 Linux kernels before 2.4, the sample fetch fails.
17741
Christopher Fauletba0c53e2019-10-17 14:40:48 +020017742fc_reordering : integer
Joe Williams30fcd392016-08-10 07:06:44 -070017743 Returns the reordering counter measured by the kernel for the client
17744 connection. If the server connection is not established, if the connection is
17745 not TCP or if the operating system does not support TCP_INFO, for example
17746 Linux kernels before 2.4, the sample fetch fails.
17747
Marcin Deranek9a66dfb2018-04-13 14:37:50 +020017748fe_defbe : string
17749 Returns a string containing the frontend's default backend name. It can be
17750 used in frontends to check which backend will handle requests by default.
17751
Willy Tarreau74ca5042013-06-11 23:12:07 +020017752fe_id : integer
17753 Returns an integer containing the current frontend's id. It can be used in
Marcin Deranek6e413ed2016-12-13 12:40:01 +010017754 backends to check from which frontend it was called, or to stick all users
Willy Tarreau74ca5042013-06-11 23:12:07 +020017755 coming via a same frontend to the same server.
17756
Marcin Deranekd2471c22016-12-12 14:08:05 +010017757fe_name : string
17758 Returns a string containing the current frontend's name. It can be used in
17759 backends to check from which frontend it was called, or to stick all users
17760 coming via a same frontend to the same server.
17761
Amaury Denoyelleda184d52020-12-10 13:43:55 +010017762fe_client_timeout : integer
17763 Returns the configuration value in millisecond for the client timeout of the
17764 current frontend.
17765
Cyril Bonté62ba8702014-04-22 23:52:25 +020017766sc_bytes_in_rate(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020017767sc0_bytes_in_rate([<table>]) : integer
17768sc1_bytes_in_rate([<table>]) : integer
17769sc2_bytes_in_rate([<table>]) : integer
Willy Tarreaue9656522010-08-17 15:40:09 +020017770 Returns the average client-to-server bytes rate from the currently tracked
17771 counters, measured in amount of bytes over the period configured in the
17772 table. See also src_bytes_in_rate.
17773
Cyril Bonté62ba8702014-04-22 23:52:25 +020017774sc_bytes_out_rate(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020017775sc0_bytes_out_rate([<table>]) : integer
17776sc1_bytes_out_rate([<table>]) : integer
17777sc2_bytes_out_rate([<table>]) : integer
Willy Tarreaue9656522010-08-17 15:40:09 +020017778 Returns the average server-to-client bytes rate from the currently tracked
17779 counters, measured in amount of bytes over the period configured in the
17780 table. See also src_bytes_out_rate.
17781
Cyril Bonté62ba8702014-04-22 23:52:25 +020017782sc_clr_gpc0(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020017783sc0_clr_gpc0([<table>]) : integer
17784sc1_clr_gpc0([<table>]) : integer
17785sc2_clr_gpc0([<table>]) : integer
Willy Tarreauf73cd112011-08-13 01:45:16 +020017786 Clears the first General Purpose Counter associated to the currently tracked
17787 counters, and returns its previous value. Before the first invocation, the
Willy Tarreau869948b2013-01-04 14:14:57 +010017788 stored value is zero, so first invocation will always return zero. This is
17789 typically used as a second ACL in an expression in order to mark a connection
17790 when a first ACL was verified :
Willy Tarreauf73cd112011-08-13 01:45:16 +020017791
Jarno Huuskonen676f6222017-03-30 09:19:45 +030017792 Example:
Willy Tarreauf73cd112011-08-13 01:45:16 +020017793 # block if 5 consecutive requests continue to come faster than 10 sess
17794 # per second, and reset the counter as soon as the traffic slows down.
Willy Tarreaube4a3ef2013-06-17 15:04:07 +020017795 acl abuse sc0_http_req_rate gt 10
17796 acl kill sc0_inc_gpc0 gt 5
17797 acl save sc0_clr_gpc0 ge 0
Willy Tarreauf73cd112011-08-13 01:45:16 +020017798 tcp-request connection accept if !abuse save
17799 tcp-request connection reject if abuse kill
17800
Frédéric Lécaille6778b272018-01-29 15:22:53 +010017801sc_clr_gpc1(<ctr>[,<table>]) : integer
17802sc0_clr_gpc1([<table>]) : integer
17803sc1_clr_gpc1([<table>]) : integer
17804sc2_clr_gpc1([<table>]) : integer
17805 Clears the second General Purpose Counter associated to the currently tracked
17806 counters, and returns its previous value. Before the first invocation, the
17807 stored value is zero, so first invocation will always return zero. This is
17808 typically used as a second ACL in an expression in order to mark a connection
17809 when a first ACL was verified.
17810
Cyril Bonté62ba8702014-04-22 23:52:25 +020017811sc_conn_cnt(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020017812sc0_conn_cnt([<table>]) : integer
17813sc1_conn_cnt([<table>]) : integer
17814sc2_conn_cnt([<table>]) : integer
Davor Ocelice9ed2812017-12-25 17:49:28 +010017815 Returns the cumulative number of incoming connections from currently tracked
Willy Tarreaue9656522010-08-17 15:40:09 +020017816 counters. See also src_conn_cnt.
17817
Cyril Bonté62ba8702014-04-22 23:52:25 +020017818sc_conn_cur(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020017819sc0_conn_cur([<table>]) : integer
17820sc1_conn_cur([<table>]) : integer
17821sc2_conn_cur([<table>]) : integer
Willy Tarreaue9656522010-08-17 15:40:09 +020017822 Returns the current amount of concurrent connections tracking the same
17823 tracked counters. This number is automatically incremented when tracking
17824 begins and decremented when tracking stops. See also src_conn_cur.
17825
Cyril Bonté62ba8702014-04-22 23:52:25 +020017826sc_conn_rate(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020017827sc0_conn_rate([<table>]) : integer
17828sc1_conn_rate([<table>]) : integer
17829sc2_conn_rate([<table>]) : integer
Willy Tarreaue9656522010-08-17 15:40:09 +020017830 Returns the average connection rate from the currently tracked counters,
17831 measured in amount of connections over the period configured in the table.
17832 See also src_conn_rate.
17833
Cyril Bonté62ba8702014-04-22 23:52:25 +020017834sc_get_gpc0(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020017835sc0_get_gpc0([<table>]) : integer
17836sc1_get_gpc0([<table>]) : integer
17837sc2_get_gpc0([<table>]) : integer
Willy Tarreaue9656522010-08-17 15:40:09 +020017838 Returns the value of the first General Purpose Counter associated to the
Willy Tarreau4d4149c2013-07-23 19:33:46 +020017839 currently tracked counters. See also src_get_gpc0 and sc/sc0/sc1/sc2_inc_gpc0.
Willy Tarreauba2ffd12013-05-29 15:54:14 +020017840
Frédéric Lécaille6778b272018-01-29 15:22:53 +010017841sc_get_gpc1(<ctr>[,<table>]) : integer
17842sc0_get_gpc1([<table>]) : integer
17843sc1_get_gpc1([<table>]) : integer
17844sc2_get_gpc1([<table>]) : integer
17845 Returns the value of the second General Purpose Counter associated to the
17846 currently tracked counters. See also src_get_gpc1 and sc/sc0/sc1/sc2_inc_gpc1.
17847
Thierry FOURNIER236657b2015-08-19 08:25:14 +020017848sc_get_gpt0(<ctr>[,<table>]) : integer
17849sc0_get_gpt0([<table>]) : integer
17850sc1_get_gpt0([<table>]) : integer
17851sc2_get_gpt0([<table>]) : integer
17852 Returns the value of the first General Purpose Tag associated to the
17853 currently tracked counters. See also src_get_gpt0.
17854
Cyril Bonté62ba8702014-04-22 23:52:25 +020017855sc_gpc0_rate(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020017856sc0_gpc0_rate([<table>]) : integer
17857sc1_gpc0_rate([<table>]) : integer
17858sc2_gpc0_rate([<table>]) : integer
Willy Tarreauba2ffd12013-05-29 15:54:14 +020017859 Returns the average increment rate of the first General Purpose Counter
17860 associated to the currently tracked counters. It reports the frequency
17861 which the gpc0 counter was incremented over the configured period. See also
Willy Tarreau4d4149c2013-07-23 19:33:46 +020017862 src_gpc0_rate, sc/sc0/sc1/sc2_get_gpc0, and sc/sc0/sc1/sc2_inc_gpc0. Note
17863 that the "gpc0_rate" counter must be stored in the stick-table for a value to
17864 be returned, as "gpc0" only holds the event count.
Willy Tarreaue9656522010-08-17 15:40:09 +020017865
Frédéric Lécaille6778b272018-01-29 15:22:53 +010017866sc_gpc1_rate(<ctr>[,<table>]) : integer
17867sc0_gpc1_rate([<table>]) : integer
17868sc1_gpc1_rate([<table>]) : integer
17869sc2_gpc1_rate([<table>]) : integer
17870 Returns the average increment rate of the second General Purpose Counter
17871 associated to the currently tracked counters. It reports the frequency
17872 which the gpc1 counter was incremented over the configured period. See also
17873 src_gpcA_rate, sc/sc0/sc1/sc2_get_gpc1, and sc/sc0/sc1/sc2_inc_gpc1. Note
17874 that the "gpc1_rate" counter must be stored in the stick-table for a value to
17875 be returned, as "gpc1" only holds the event count.
17876
Cyril Bonté62ba8702014-04-22 23:52:25 +020017877sc_http_err_cnt(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020017878sc0_http_err_cnt([<table>]) : integer
17879sc1_http_err_cnt([<table>]) : integer
17880sc2_http_err_cnt([<table>]) : integer
Davor Ocelice9ed2812017-12-25 17:49:28 +010017881 Returns the cumulative number of HTTP errors from the currently tracked
Willy Tarreaue9656522010-08-17 15:40:09 +020017882 counters. This includes the both request errors and 4xx error responses.
17883 See also src_http_err_cnt.
17884
Cyril Bonté62ba8702014-04-22 23:52:25 +020017885sc_http_err_rate(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020017886sc0_http_err_rate([<table>]) : integer
17887sc1_http_err_rate([<table>]) : integer
17888sc2_http_err_rate([<table>]) : integer
Willy Tarreaue9656522010-08-17 15:40:09 +020017889 Returns the average rate of HTTP errors from the currently tracked counters,
17890 measured in amount of errors over the period configured in the table. This
17891 includes the both request errors and 4xx error responses. See also
17892 src_http_err_rate.
17893
Willy Tarreau826f3ab2021-02-10 12:07:15 +010017894sc_http_fail_cnt(<ctr>[,<table>]) : integer
17895sc0_http_fail_cnt([<table>]) : integer
17896sc1_http_fail_cnt([<table>]) : integer
17897sc2_http_fail_cnt([<table>]) : integer
17898 Returns the cumulative number of HTTP response failures from the currently
17899 tracked counters. This includes the both response errors and 5xx status codes
17900 other than 501 and 505. See also src_http_fail_cnt.
17901
17902sc_http_fail_rate(<ctr>[,<table>]) : integer
17903sc0_http_fail_rate([<table>]) : integer
17904sc1_http_fail_rate([<table>]) : integer
17905sc2_http_fail_rate([<table>]) : integer
17906 Returns the average rate of HTTP response failures from the currently tracked
17907 counters, measured in amount of failures over the period configured in the
17908 table. This includes the both response errors and 5xx status codes other than
17909 501 and 505. See also src_http_fail_rate.
17910
Cyril Bonté62ba8702014-04-22 23:52:25 +020017911sc_http_req_cnt(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020017912sc0_http_req_cnt([<table>]) : integer
17913sc1_http_req_cnt([<table>]) : integer
17914sc2_http_req_cnt([<table>]) : integer
Davor Ocelice9ed2812017-12-25 17:49:28 +010017915 Returns the cumulative number of HTTP requests from the currently tracked
Willy Tarreaue9656522010-08-17 15:40:09 +020017916 counters. This includes every started request, valid or not. See also
17917 src_http_req_cnt.
17918
Cyril Bonté62ba8702014-04-22 23:52:25 +020017919sc_http_req_rate(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020017920sc0_http_req_rate([<table>]) : integer
17921sc1_http_req_rate([<table>]) : integer
17922sc2_http_req_rate([<table>]) : integer
Willy Tarreaue9656522010-08-17 15:40:09 +020017923 Returns the average rate of HTTP requests from the currently tracked
17924 counters, measured in amount of requests over the period configured in
17925 the table. This includes every started request, valid or not. See also
17926 src_http_req_rate.
17927
Cyril Bonté62ba8702014-04-22 23:52:25 +020017928sc_inc_gpc0(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020017929sc0_inc_gpc0([<table>]) : integer
17930sc1_inc_gpc0([<table>]) : integer
17931sc2_inc_gpc0([<table>]) : integer
Willy Tarreaue9656522010-08-17 15:40:09 +020017932 Increments the first General Purpose Counter associated to the currently
Willy Tarreau869948b2013-01-04 14:14:57 +010017933 tracked counters, and returns its new value. Before the first invocation,
17934 the stored value is zero, so first invocation will increase it to 1 and will
17935 return 1. This is typically used as a second ACL in an expression in order
17936 to mark a connection when a first ACL was verified :
Willy Tarreaue9656522010-08-17 15:40:09 +020017937
Jarno Huuskonen676f6222017-03-30 09:19:45 +030017938 Example:
Willy Tarreaube4a3ef2013-06-17 15:04:07 +020017939 acl abuse sc0_http_req_rate gt 10
17940 acl kill sc0_inc_gpc0 gt 0
Willy Tarreaue9656522010-08-17 15:40:09 +020017941 tcp-request connection reject if abuse kill
17942
Frédéric Lécaille6778b272018-01-29 15:22:53 +010017943sc_inc_gpc1(<ctr>[,<table>]) : integer
17944sc0_inc_gpc1([<table>]) : integer
17945sc1_inc_gpc1([<table>]) : integer
17946sc2_inc_gpc1([<table>]) : integer
17947 Increments the second General Purpose Counter associated to the currently
17948 tracked counters, and returns its new value. Before the first invocation,
17949 the stored value is zero, so first invocation will increase it to 1 and will
17950 return 1. This is typically used as a second ACL in an expression in order
17951 to mark a connection when a first ACL was verified.
17952
Cyril Bonté62ba8702014-04-22 23:52:25 +020017953sc_kbytes_in(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020017954sc0_kbytes_in([<table>]) : integer
17955sc1_kbytes_in([<table>]) : integer
17956sc2_kbytes_in([<table>]) : integer
Willy Tarreaua01b9742014-07-10 15:29:24 +020017957 Returns the total amount of client-to-server data from the currently tracked
17958 counters, measured in kilobytes. The test is currently performed on 32-bit
17959 integers, which limits values to 4 terabytes. See also src_kbytes_in.
Willy Tarreaue9656522010-08-17 15:40:09 +020017960
Cyril Bonté62ba8702014-04-22 23:52:25 +020017961sc_kbytes_out(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020017962sc0_kbytes_out([<table>]) : integer
17963sc1_kbytes_out([<table>]) : integer
17964sc2_kbytes_out([<table>]) : integer
Willy Tarreaua01b9742014-07-10 15:29:24 +020017965 Returns the total amount of server-to-client data from the currently tracked
17966 counters, measured in kilobytes. The test is currently performed on 32-bit
17967 integers, which limits values to 4 terabytes. See also src_kbytes_out.
Willy Tarreaue9656522010-08-17 15:40:09 +020017968
Cyril Bonté62ba8702014-04-22 23:52:25 +020017969sc_sess_cnt(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020017970sc0_sess_cnt([<table>]) : integer
17971sc1_sess_cnt([<table>]) : integer
17972sc2_sess_cnt([<table>]) : integer
Davor Ocelice9ed2812017-12-25 17:49:28 +010017973 Returns the cumulative number of incoming connections that were transformed
Willy Tarreaue9656522010-08-17 15:40:09 +020017974 into sessions, which means that they were accepted by a "tcp-request
17975 connection" rule, from the currently tracked counters. A backend may count
17976 more sessions than connections because each connection could result in many
Jamie Gloudonaaa21002012-08-25 00:18:33 -040017977 backend sessions if some HTTP keep-alive is performed over the connection
Willy Tarreaue9656522010-08-17 15:40:09 +020017978 with the client. See also src_sess_cnt.
17979
Cyril Bonté62ba8702014-04-22 23:52:25 +020017980sc_sess_rate(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020017981sc0_sess_rate([<table>]) : integer
17982sc1_sess_rate([<table>]) : integer
17983sc2_sess_rate([<table>]) : integer
Willy Tarreaue9656522010-08-17 15:40:09 +020017984 Returns the average session rate from the currently tracked counters,
17985 measured in amount of sessions over the period configured in the table. A
17986 session is a connection that got past the early "tcp-request connection"
17987 rules. A backend may count more sessions than connections because each
17988 connection could result in many backend sessions if some HTTP keep-alive is
Jamie Gloudonaaa21002012-08-25 00:18:33 -040017989 performed over the connection with the client. See also src_sess_rate.
Willy Tarreaue9656522010-08-17 15:40:09 +020017990
Cyril Bonté62ba8702014-04-22 23:52:25 +020017991sc_tracked(<ctr>[,<table>]) : boolean
Willy Tarreau0f791d42013-07-23 19:56:43 +020017992sc0_tracked([<table>]) : boolean
17993sc1_tracked([<table>]) : boolean
17994sc2_tracked([<table>]) : boolean
Willy Tarreau6f1615f2013-06-03 15:15:22 +020017995 Returns true if the designated session counter is currently being tracked by
17996 the current session. This can be useful when deciding whether or not we want
17997 to set some values in a header passed to the server.
17998
Cyril Bonté62ba8702014-04-22 23:52:25 +020017999sc_trackers(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020018000sc0_trackers([<table>]) : integer
18001sc1_trackers([<table>]) : integer
18002sc2_trackers([<table>]) : integer
Willy Tarreau2406db42012-12-09 12:16:43 +010018003 Returns the current amount of concurrent connections tracking the same
18004 tracked counters. This number is automatically incremented when tracking
Willy Tarreaube4a3ef2013-06-17 15:04:07 +020018005 begins and decremented when tracking stops. It differs from sc0_conn_cur in
Willy Tarreau2406db42012-12-09 12:16:43 +010018006 that it does not rely on any stored information but on the table's reference
18007 count (the "use" value which is returned by "show table" on the CLI). This
Willy Tarreau74ca5042013-06-11 23:12:07 +020018008 may sometimes be more suited for layer7 tracking. It can be used to tell a
18009 server how many concurrent connections there are from a given address for
18010 example.
Willy Tarreau2406db42012-12-09 12:16:43 +010018011
Willy Tarreau74ca5042013-06-11 23:12:07 +020018012so_id : integer
18013 Returns an integer containing the current listening socket's id. It is useful
18014 in frontends involving many "bind" lines, or to stick all users coming via a
18015 same socket to the same server.
Willy Tarreaud63335a2010-02-26 12:56:52 +010018016
Jerome Magnineb421b22020-03-27 22:08:40 +010018017so_name : string
18018 Returns a string containing the current listening socket's name, as defined
18019 with name on a "bind" line. It can serve the same purposes as so_id but with
18020 strings instead of integers.
18021
Willy Tarreau74ca5042013-06-11 23:12:07 +020018022src : ip
Davor Ocelice9ed2812017-12-25 17:49:28 +010018023 This is the source IPv4 address of the client of the session. It is of type
Willy Tarreau74ca5042013-06-11 23:12:07 +020018024 IP and works on both IPv4 and IPv6 tables. On IPv6 tables, IPv4 addresses are
18025 mapped to their IPv6 equivalent, according to RFC 4291. Note that it is the
18026 TCP-level source address which is used, and not the address of a client
Bertrand Jacquin93b227d2016-06-04 15:11:10 +010018027 behind a proxy. However if the "accept-proxy" or "accept-netscaler-cip" bind
18028 directive is used, it can be the address of a client behind another
18029 PROXY-protocol compatible component for all rule sets except
Willy Tarreau64ded3d2019-01-23 10:02:15 +010018030 "tcp-request connection" which sees the real address. When the incoming
18031 connection passed through address translation or redirection involving
18032 connection tracking, the original destination address before the redirection
18033 will be reported. On Linux systems, the source and destination may seldom
18034 appear reversed if the nf_conntrack_tcp_loose sysctl is set, because a late
18035 response may reopen a timed out connection and switch what is believed to be
18036 the source and the destination.
Willy Tarreaud63335a2010-02-26 12:56:52 +010018037
Thierry FOURNIERd5f624d2013-11-26 11:52:33 +010018038 Example:
18039 # add an HTTP header in requests with the originating address' country
18040 http-request set-header X-Country %[src,map_ip(geoip.lst)]
18041
Willy Tarreau74ca5042013-06-11 23:12:07 +020018042src_bytes_in_rate([<table>]) : integer
18043 Returns the average bytes rate from the incoming connection's source address
18044 in the current proxy's stick-table or in the designated stick-table, measured
18045 in amount of bytes over the period configured in the table. If the address is
Willy Tarreau4d4149c2013-07-23 19:33:46 +020018046 not found, zero is returned. See also sc/sc0/sc1/sc2_bytes_in_rate.
Willy Tarreauc9705a12010-07-27 20:05:50 +020018047
Willy Tarreau74ca5042013-06-11 23:12:07 +020018048src_bytes_out_rate([<table>]) : integer
18049 Returns the average bytes rate to the incoming connection's source address in
18050 the current proxy's stick-table or in the designated stick-table, measured in
Willy Tarreauc9705a12010-07-27 20:05:50 +020018051 amount of bytes over the period configured in the table. If the address is
Willy Tarreau4d4149c2013-07-23 19:33:46 +020018052 not found, zero is returned. See also sc/sc0/sc1/sc2_bytes_out_rate.
Willy Tarreauc9705a12010-07-27 20:05:50 +020018053
Willy Tarreau74ca5042013-06-11 23:12:07 +020018054src_clr_gpc0([<table>]) : integer
18055 Clears the first General Purpose Counter associated to the incoming
18056 connection's source address in the current proxy's stick-table or in the
18057 designated stick-table, and returns its previous value. If the address is not
18058 found, an entry is created and 0 is returned. This is typically used as a
18059 second ACL in an expression in order to mark a connection when a first ACL
18060 was verified :
Willy Tarreauf73cd112011-08-13 01:45:16 +020018061
Jarno Huuskonen676f6222017-03-30 09:19:45 +030018062 Example:
Willy Tarreauf73cd112011-08-13 01:45:16 +020018063 # block if 5 consecutive requests continue to come faster than 10 sess
18064 # per second, and reset the counter as soon as the traffic slows down.
18065 acl abuse src_http_req_rate gt 10
18066 acl kill src_inc_gpc0 gt 5
Willy Tarreau869948b2013-01-04 14:14:57 +010018067 acl save src_clr_gpc0 ge 0
Willy Tarreauf73cd112011-08-13 01:45:16 +020018068 tcp-request connection accept if !abuse save
18069 tcp-request connection reject if abuse kill
18070
Frédéric Lécaille6778b272018-01-29 15:22:53 +010018071src_clr_gpc1([<table>]) : integer
18072 Clears the second General Purpose Counter associated to the incoming
18073 connection's source address in the current proxy's stick-table or in the
18074 designated stick-table, and returns its previous value. If the address is not
18075 found, an entry is created and 0 is returned. This is typically used as a
18076 second ACL in an expression in order to mark a connection when a first ACL
18077 was verified.
18078
Willy Tarreau74ca5042013-06-11 23:12:07 +020018079src_conn_cnt([<table>]) : integer
Davor Ocelice9ed2812017-12-25 17:49:28 +010018080 Returns the cumulative number of connections initiated from the current
Willy Tarreau74ca5042013-06-11 23:12:07 +020018081 incoming connection's source address in the current proxy's stick-table or in
Willy Tarreauc9705a12010-07-27 20:05:50 +020018082 the designated stick-table. If the address is not found, zero is returned.
Willy Tarreau4d4149c2013-07-23 19:33:46 +020018083 See also sc/sc0/sc1/sc2_conn_cnt.
Willy Tarreauc9705a12010-07-27 20:05:50 +020018084
Willy Tarreau74ca5042013-06-11 23:12:07 +020018085src_conn_cur([<table>]) : integer
Willy Tarreauc9705a12010-07-27 20:05:50 +020018086 Returns the current amount of concurrent connections initiated from the
Willy Tarreau74ca5042013-06-11 23:12:07 +020018087 current incoming connection's source address in the current proxy's
18088 stick-table or in the designated stick-table. If the address is not found,
Willy Tarreau4d4149c2013-07-23 19:33:46 +020018089 zero is returned. See also sc/sc0/sc1/sc2_conn_cur.
Willy Tarreauc9705a12010-07-27 20:05:50 +020018090
Willy Tarreau74ca5042013-06-11 23:12:07 +020018091src_conn_rate([<table>]) : integer
18092 Returns the average connection rate from the incoming connection's source
18093 address in the current proxy's stick-table or in the designated stick-table,
18094 measured in amount of connections over the period configured in the table. If
Willy Tarreau4d4149c2013-07-23 19:33:46 +020018095 the address is not found, zero is returned. See also sc/sc0/sc1/sc2_conn_rate.
Willy Tarreauc9705a12010-07-27 20:05:50 +020018096
Willy Tarreau74ca5042013-06-11 23:12:07 +020018097src_get_gpc0([<table>]) : integer
Willy Tarreauc9705a12010-07-27 20:05:50 +020018098 Returns the value of the first General Purpose Counter associated to the
Willy Tarreau74ca5042013-06-11 23:12:07 +020018099 incoming connection's source address in the current proxy's stick-table or in
Willy Tarreauc9705a12010-07-27 20:05:50 +020018100 the designated stick-table. If the address is not found, zero is returned.
Willy Tarreau4d4149c2013-07-23 19:33:46 +020018101 See also sc/sc0/sc1/sc2_get_gpc0 and src_inc_gpc0.
Willy Tarreauc9705a12010-07-27 20:05:50 +020018102
Frédéric Lécaille6778b272018-01-29 15:22:53 +010018103src_get_gpc1([<table>]) : integer
18104 Returns the value of the second General Purpose Counter associated to the
18105 incoming connection's source address in the current proxy's stick-table or in
18106 the designated stick-table. If the address is not found, zero is returned.
18107 See also sc/sc0/sc1/sc2_get_gpc1 and src_inc_gpc1.
18108
Thierry FOURNIER236657b2015-08-19 08:25:14 +020018109src_get_gpt0([<table>]) : integer
18110 Returns the value of the first General Purpose Tag associated to the
18111 incoming connection's source address in the current proxy's stick-table or in
18112 the designated stick-table. If the address is not found, zero is returned.
18113 See also sc/sc0/sc1/sc2_get_gpt0.
18114
Willy Tarreau74ca5042013-06-11 23:12:07 +020018115src_gpc0_rate([<table>]) : integer
Willy Tarreauba2ffd12013-05-29 15:54:14 +020018116 Returns the average increment rate of the first General Purpose Counter
Willy Tarreau74ca5042013-06-11 23:12:07 +020018117 associated to the incoming connection's source address in the current proxy's
Willy Tarreauba2ffd12013-05-29 15:54:14 +020018118 stick-table or in the designated stick-table. It reports the frequency
18119 which the gpc0 counter was incremented over the configured period. See also
Willy Tarreau4d4149c2013-07-23 19:33:46 +020018120 sc/sc0/sc1/sc2_gpc0_rate, src_get_gpc0, and sc/sc0/sc1/sc2_inc_gpc0. Note
18121 that the "gpc0_rate" counter must be stored in the stick-table for a value to
18122 be returned, as "gpc0" only holds the event count.
Willy Tarreauba2ffd12013-05-29 15:54:14 +020018123
Frédéric Lécaille6778b272018-01-29 15:22:53 +010018124src_gpc1_rate([<table>]) : integer
18125 Returns the average increment rate of the second General Purpose Counter
18126 associated to the incoming connection's source address in the current proxy's
18127 stick-table or in the designated stick-table. It reports the frequency
18128 which the gpc1 counter was incremented over the configured period. See also
18129 sc/sc0/sc1/sc2_gpc1_rate, src_get_gpc1, and sc/sc0/sc1/sc2_inc_gpc1. Note
18130 that the "gpc1_rate" counter must be stored in the stick-table for a value to
18131 be returned, as "gpc1" only holds the event count.
18132
Willy Tarreau74ca5042013-06-11 23:12:07 +020018133src_http_err_cnt([<table>]) : integer
Davor Ocelice9ed2812017-12-25 17:49:28 +010018134 Returns the cumulative number of HTTP errors from the incoming connection's
Willy Tarreau74ca5042013-06-11 23:12:07 +020018135 source address in the current proxy's stick-table or in the designated
Willy Tarreauc9705a12010-07-27 20:05:50 +020018136 stick-table. This includes the both request errors and 4xx error responses.
Willy Tarreau4d4149c2013-07-23 19:33:46 +020018137 See also sc/sc0/sc1/sc2_http_err_cnt. If the address is not found, zero is
Willy Tarreau74ca5042013-06-11 23:12:07 +020018138 returned.
Willy Tarreauc9705a12010-07-27 20:05:50 +020018139
Willy Tarreau74ca5042013-06-11 23:12:07 +020018140src_http_err_rate([<table>]) : integer
18141 Returns the average rate of HTTP errors from the incoming connection's source
18142 address in the current proxy's stick-table or in the designated stick-table,
18143 measured in amount of errors over the period configured in the table. This
18144 includes the both request errors and 4xx error responses. If the address is
Willy Tarreau4d4149c2013-07-23 19:33:46 +020018145 not found, zero is returned. See also sc/sc0/sc1/sc2_http_err_rate.
Willy Tarreauc9705a12010-07-27 20:05:50 +020018146
Willy Tarreau826f3ab2021-02-10 12:07:15 +010018147src_http_fail_cnt([<table>]) : integer
18148 Returns the cumulative number of HTTP response failures triggered by the
18149 incoming connection's source address in the current proxy's stick-table or in
Ilya Shipitsin0de36ad2021-02-20 00:23:36 +050018150 the designated stick-table. This includes the both response errors and 5xx
Willy Tarreau826f3ab2021-02-10 12:07:15 +010018151 status codes other than 501 and 505. See also sc/sc0/sc1/sc2_http_fail_cnt.
18152 If the address is not found, zero is returned.
18153
18154src_http_fail_rate([<table>]) : integer
18155 Returns the average rate of HTTP response failures triggered by the incoming
18156 connection's source address in the current proxy's stick-table or in the
18157 designated stick-table, measured in amount of failures over the period
18158 configured in the table. This includes the both response errors and 5xx
18159 status codes other than 501 and 505. If the address is not found, zero is
18160 returned. See also sc/sc0/sc1/sc2_http_fail_rate.
18161
Willy Tarreau74ca5042013-06-11 23:12:07 +020018162src_http_req_cnt([<table>]) : integer
Davor Ocelice9ed2812017-12-25 17:49:28 +010018163 Returns the cumulative number of HTTP requests from the incoming connection's
Willy Tarreau74ca5042013-06-11 23:12:07 +020018164 source address in the current proxy's stick-table or in the designated stick-
18165 table. This includes every started request, valid or not. If the address is
Willy Tarreau4d4149c2013-07-23 19:33:46 +020018166 not found, zero is returned. See also sc/sc0/sc1/sc2_http_req_cnt.
Willy Tarreauc9705a12010-07-27 20:05:50 +020018167
Willy Tarreau74ca5042013-06-11 23:12:07 +020018168src_http_req_rate([<table>]) : integer
18169 Returns the average rate of HTTP requests from the incoming connection's
18170 source address in the current proxy's stick-table or in the designated stick-
18171 table, measured in amount of requests over the period configured in the
Willy Tarreauc9705a12010-07-27 20:05:50 +020018172 table. This includes every started request, valid or not. If the address is
Willy Tarreau4d4149c2013-07-23 19:33:46 +020018173 not found, zero is returned. See also sc/sc0/sc1/sc2_http_req_rate.
Willy Tarreauc9705a12010-07-27 20:05:50 +020018174
Willy Tarreau74ca5042013-06-11 23:12:07 +020018175src_inc_gpc0([<table>]) : integer
18176 Increments the first General Purpose Counter associated to the incoming
18177 connection's source address in the current proxy's stick-table or in the
18178 designated stick-table, and returns its new value. If the address is not
Willy Tarreaube4a3ef2013-06-17 15:04:07 +020018179 found, an entry is created and 1 is returned. See also sc0/sc2/sc2_inc_gpc0.
Willy Tarreau74ca5042013-06-11 23:12:07 +020018180 This is typically used as a second ACL in an expression in order to mark a
18181 connection when a first ACL was verified :
Willy Tarreauc9705a12010-07-27 20:05:50 +020018182
Jarno Huuskonen676f6222017-03-30 09:19:45 +030018183 Example:
Willy Tarreauc9705a12010-07-27 20:05:50 +020018184 acl abuse src_http_req_rate gt 10
Willy Tarreau869948b2013-01-04 14:14:57 +010018185 acl kill src_inc_gpc0 gt 0
Willy Tarreaue9656522010-08-17 15:40:09 +020018186 tcp-request connection reject if abuse kill
Willy Tarreauc9705a12010-07-27 20:05:50 +020018187
Frédéric Lécaille6778b272018-01-29 15:22:53 +010018188src_inc_gpc1([<table>]) : integer
18189 Increments the second General Purpose Counter associated to the incoming
18190 connection's source address in the current proxy's stick-table or in the
18191 designated stick-table, and returns its new value. If the address is not
18192 found, an entry is created and 1 is returned. See also sc0/sc2/sc2_inc_gpc1.
18193 This is typically used as a second ACL in an expression in order to mark a
18194 connection when a first ACL was verified.
18195
Willy Tarreau16e01562016-08-09 16:46:18 +020018196src_is_local : boolean
18197 Returns true if the source address of the incoming connection is local to the
18198 system, or false if the address doesn't exist on the system, meaning that it
18199 comes from a remote machine. Note that UNIX addresses are considered local.
18200 It can be useful to apply certain access restrictions based on where the
Davor Ocelice9ed2812017-12-25 17:49:28 +010018201 client comes from (e.g. require auth or https for remote machines). Please
Willy Tarreau16e01562016-08-09 16:46:18 +020018202 note that the check involves a few system calls, so it's better to do it only
18203 once per connection.
18204
Willy Tarreau74ca5042013-06-11 23:12:07 +020018205src_kbytes_in([<table>]) : integer
Willy Tarreaua01b9742014-07-10 15:29:24 +020018206 Returns the total amount of data received from the incoming connection's
18207 source address in the current proxy's stick-table or in the designated
18208 stick-table, measured in kilobytes. If the address is not found, zero is
18209 returned. The test is currently performed on 32-bit integers, which limits
18210 values to 4 terabytes. See also sc/sc0/sc1/sc2_kbytes_in.
Willy Tarreauc9705a12010-07-27 20:05:50 +020018211
Willy Tarreau74ca5042013-06-11 23:12:07 +020018212src_kbytes_out([<table>]) : integer
Willy Tarreaua01b9742014-07-10 15:29:24 +020018213 Returns the total amount of data sent to the incoming connection's source
18214 address in the current proxy's stick-table or in the designated stick-table,
18215 measured in kilobytes. If the address is not found, zero is returned. The
18216 test is currently performed on 32-bit integers, which limits values to 4
18217 terabytes. See also sc/sc0/sc1/sc2_kbytes_out.
Willy Tarreaua975b8f2010-06-05 19:13:27 +020018218
Willy Tarreau74ca5042013-06-11 23:12:07 +020018219src_port : integer
18220 Returns an integer value corresponding to the TCP source port of the
18221 connection on the client side, which is the port the client connected from.
18222 Usage of this function is very limited as modern protocols do not care much
18223 about source ports nowadays.
Willy Tarreau079ff0a2009-03-05 21:34:28 +010018224
Willy Tarreau74ca5042013-06-11 23:12:07 +020018225src_sess_cnt([<table>]) : integer
Davor Ocelice9ed2812017-12-25 17:49:28 +010018226 Returns the cumulative number of connections initiated from the incoming
Willy Tarreauc9705a12010-07-27 20:05:50 +020018227 connection's source IPv4 address in the current proxy's stick-table or in the
18228 designated stick-table, that were transformed into sessions, which means that
18229 they were accepted by "tcp-request" rules. If the address is not found, zero
Willy Tarreau4d4149c2013-07-23 19:33:46 +020018230 is returned. See also sc/sc0/sc1/sc2_sess_cnt.
Willy Tarreauc9705a12010-07-27 20:05:50 +020018231
Willy Tarreau74ca5042013-06-11 23:12:07 +020018232src_sess_rate([<table>]) : integer
18233 Returns the average session rate from the incoming connection's source
18234 address in the current proxy's stick-table or in the designated stick-table,
18235 measured in amount of sessions over the period configured in the table. A
18236 session is a connection that went past the early "tcp-request" rules. If the
Willy Tarreau4d4149c2013-07-23 19:33:46 +020018237 address is not found, zero is returned. See also sc/sc0/sc1/sc2_sess_rate.
Willy Tarreauc9705a12010-07-27 20:05:50 +020018238
Willy Tarreau74ca5042013-06-11 23:12:07 +020018239src_updt_conn_cnt([<table>]) : integer
18240 Creates or updates the entry associated to the incoming connection's source
18241 address in the current proxy's stick-table or in the designated stick-table.
18242 This table must be configured to store the "conn_cnt" data type, otherwise
18243 the match will be ignored. The current count is incremented by one, and the
18244 expiration timer refreshed. The updated count is returned, so this match
18245 can't return zero. This was used to reject service abusers based on their
18246 source address. Note: it is recommended to use the more complete "track-sc*"
18247 actions in "tcp-request" rules instead.
Willy Tarreaua975b8f2010-06-05 19:13:27 +020018248
18249 Example :
18250 # This frontend limits incoming SSH connections to 3 per 10 second for
18251 # each source address, and rejects excess connections until a 10 second
18252 # silence is observed. At most 20 addresses are tracked.
18253 listen ssh
18254 bind :22
18255 mode tcp
18256 maxconn 100
Willy Tarreauc9705a12010-07-27 20:05:50 +020018257 stick-table type ip size 20 expire 10s store conn_cnt
Willy Tarreau74ca5042013-06-11 23:12:07 +020018258 tcp-request content reject if { src_updt_conn_cnt gt 3 }
Willy Tarreaua975b8f2010-06-05 19:13:27 +020018259 server local 127.0.0.1:22
18260
Willy Tarreau74ca5042013-06-11 23:12:07 +020018261srv_id : integer
18262 Returns an integer containing the server's id when processing the response.
18263 While it's almost only used with ACLs, it may be used for logging or
Christopher Fauletd1b44642020-04-30 09:51:15 +020018264 debugging. It can also be used in a tcp-check or an http-check ruleset.
Hervé COMMOWICKdaa824e2011-08-05 12:09:44 +020018265
vkill1dfd1652019-10-30 16:58:14 +080018266srv_name : string
18267 Returns a string containing the server's name when processing the response.
18268 While it's almost only used with ACLs, it may be used for logging or
Christopher Fauletd1b44642020-04-30 09:51:15 +020018269 debugging. It can also be used in a tcp-check or an http-check ruleset.
vkill1dfd1652019-10-30 16:58:14 +080018270
Thierry FOURNIER060762e2014-04-23 13:29:15 +0200182717.3.4. Fetching samples at Layer 5
Willy Tarreau74ca5042013-06-11 23:12:07 +020018272----------------------------------
Willy Tarreau0b1cd942010-05-16 22:18:27 +020018273
Daniel Corbett9f0843f2021-05-08 10:50:37 -040018274The layer 5 usually describes just the session layer which in HAProxy is
Willy Tarreau74ca5042013-06-11 23:12:07 +020018275closest to the session once all the connection handshakes are finished, but
18276when no content is yet made available. The fetch methods described here are
18277usable as low as the "tcp-request content" rule sets unless they require some
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030018278future information. Those generally include the results of SSL negotiations.
Willy Tarreauc735a072011-03-29 00:57:02 +020018279
Ben Shillitof25e8e52016-12-02 14:25:37 +00001828051d.all(<prop>[,<prop>*]) : string
18281 Returns values for the properties requested as a string, where values are
18282 separated by the delimiter specified with "51degrees-property-separator".
18283 The device is identified using all the important HTTP headers from the
18284 request. The function can be passed up to five property names, and if a
18285 property name can't be found, the value "NoData" is returned.
18286
18287 Example :
18288 # Here the header "X-51D-DeviceTypeMobileTablet" is added to the request
18289 # containing the three properties requested using all relevant headers from
18290 # the request.
18291 frontend http-in
18292 bind *:8081
18293 default_backend servers
18294 http-request set-header X-51D-DeviceTypeMobileTablet \
18295 %[51d.all(DeviceType,IsMobile,IsTablet)]
18296
Emeric Brun645ae792014-04-30 14:21:06 +020018297ssl_bc : boolean
18298 Returns true when the back connection was made via an SSL/TLS transport
18299 layer and is locally deciphered. This means the outgoing connection was made
Christopher Fauletd92ea7f2020-04-30 10:03:55 +020018300 other a server with the "ssl" option. It can be used in a tcp-check or an
18301 http-check ruleset.
Emeric Brun645ae792014-04-30 14:21:06 +020018302
18303ssl_bc_alg_keysize : integer
18304 Returns the symmetric cipher key size supported in bits when the outgoing
Christopher Fauletd92ea7f2020-04-30 10:03:55 +020018305 connection was made over an SSL/TLS transport layer. It can be used in a
18306 tcp-check or an http-check ruleset.
Emeric Brun645ae792014-04-30 14:21:06 +020018307
Olivier Houchard6b77f492018-11-22 18:18:29 +010018308ssl_bc_alpn : string
18309 This extracts the Application Layer Protocol Negotiation field from an
18310 outgoing connection made via a TLS transport layer.
Michael Prokop4438c602019-05-24 10:25:45 +020018311 The result is a string containing the protocol name negotiated with the
Olivier Houchard6b77f492018-11-22 18:18:29 +010018312 server. The SSL library must have been built with support for TLS
18313 extensions enabled (check haproxy -vv). Note that the TLS ALPN extension is
18314 not advertised unless the "alpn" keyword on the "server" line specifies a
18315 protocol list. Also, nothing forces the server to pick a protocol from this
18316 list, any other one may be requested. The TLS ALPN extension is meant to
Christopher Fauletd92ea7f2020-04-30 10:03:55 +020018317 replace the TLS NPN extension. See also "ssl_bc_npn". It can be used in a
18318 tcp-check or an http-check ruleset.
Olivier Houchard6b77f492018-11-22 18:18:29 +010018319
Emeric Brun645ae792014-04-30 14:21:06 +020018320ssl_bc_cipher : string
18321 Returns the name of the used cipher when the outgoing connection was made
Christopher Fauletd92ea7f2020-04-30 10:03:55 +020018322 over an SSL/TLS transport layer. It can be used in a tcp-check or an
18323 http-check ruleset.
Emeric Brun645ae792014-04-30 14:21:06 +020018324
Patrick Hemmer65674662019-06-04 08:13:03 -040018325ssl_bc_client_random : binary
18326 Returns the client random of the back connection when the incoming connection
18327 was made over an SSL/TLS transport layer. It is useful to to decrypt traffic
18328 sent using ephemeral ciphers. This requires OpenSSL >= 1.1.0, or BoringSSL.
Christopher Fauletd92ea7f2020-04-30 10:03:55 +020018329 It can be used in a tcp-check or an http-check ruleset.
Patrick Hemmer65674662019-06-04 08:13:03 -040018330
Emeric Brun74f7ffa2018-02-19 16:14:12 +010018331ssl_bc_is_resumed : boolean
18332 Returns true when the back connection was made over an SSL/TLS transport
18333 layer and the newly created SSL session was resumed using a cached
Christopher Fauletd92ea7f2020-04-30 10:03:55 +020018334 session or a TLS ticket. It can be used in a tcp-check or an http-check
18335 ruleset.
Emeric Brun74f7ffa2018-02-19 16:14:12 +010018336
Olivier Houchard6b77f492018-11-22 18:18:29 +010018337ssl_bc_npn : string
18338 This extracts the Next Protocol Negotiation field from an outgoing connection
18339 made via a TLS transport layer. The result is a string containing the
Michael Prokop4438c602019-05-24 10:25:45 +020018340 protocol name negotiated with the server . The SSL library must have been
Olivier Houchard6b77f492018-11-22 18:18:29 +010018341 built with support for TLS extensions enabled (check haproxy -vv). Note that
18342 the TLS NPN extension is not advertised unless the "npn" keyword on the
18343 "server" line specifies a protocol list. Also, nothing forces the server to
18344 pick a protocol from this list, any other one may be used. Please note that
Christopher Fauletd92ea7f2020-04-30 10:03:55 +020018345 the TLS NPN extension was replaced with ALPN. It can be used in a tcp-check
18346 or an http-check ruleset.
Olivier Houchard6b77f492018-11-22 18:18:29 +010018347
Emeric Brun645ae792014-04-30 14:21:06 +020018348ssl_bc_protocol : string
18349 Returns the name of the used protocol when the outgoing connection was made
Christopher Fauletd92ea7f2020-04-30 10:03:55 +020018350 over an SSL/TLS transport layer. It can be used in a tcp-check or an
18351 http-check ruleset.
Emeric Brun645ae792014-04-30 14:21:06 +020018352
Emeric Brunb73a9b02014-04-30 18:49:19 +020018353ssl_bc_unique_id : binary
Emeric Brun645ae792014-04-30 14:21:06 +020018354 When the outgoing connection was made over an SSL/TLS transport layer,
Emeric Brunb73a9b02014-04-30 18:49:19 +020018355 returns the TLS unique ID as defined in RFC5929 section 3. The unique id
Christopher Fauletd92ea7f2020-04-30 10:03:55 +020018356 can be encoded to base64 using the converter: "ssl_bc_unique_id,base64". It
18357 can be used in a tcp-check or an http-check ruleset.
Emeric Brun645ae792014-04-30 14:21:06 +020018358
Patrick Hemmer65674662019-06-04 08:13:03 -040018359ssl_bc_server_random : binary
18360 Returns the server random of the back connection when the incoming connection
18361 was made over an SSL/TLS transport layer. It is useful to to decrypt traffic
18362 sent using ephemeral ciphers. This requires OpenSSL >= 1.1.0, or BoringSSL.
Christopher Fauletd92ea7f2020-04-30 10:03:55 +020018363 It can be used in a tcp-check or an http-check ruleset.
Patrick Hemmer65674662019-06-04 08:13:03 -040018364
Emeric Brun645ae792014-04-30 14:21:06 +020018365ssl_bc_session_id : binary
18366 Returns the SSL ID of the back connection when the outgoing connection was
18367 made over an SSL/TLS transport layer. It is useful to log if we want to know
Christopher Fauletd92ea7f2020-04-30 10:03:55 +020018368 if session was reused or not. It can be used in a tcp-check or an http-check
18369 ruleset.
Emeric Brun645ae792014-04-30 14:21:06 +020018370
Patrick Hemmere0275472018-04-28 19:15:51 -040018371ssl_bc_session_key : binary
18372 Returns the SSL session master key of the back connection when the outgoing
18373 connection was made over an SSL/TLS transport layer. It is useful to decrypt
18374 traffic sent using ephemeral ciphers. This requires OpenSSL >= 1.1.0, or
Christopher Fauletd92ea7f2020-04-30 10:03:55 +020018375 BoringSSL. It can be used in a tcp-check or an http-check ruleset.
Patrick Hemmere0275472018-04-28 19:15:51 -040018376
Emeric Brun645ae792014-04-30 14:21:06 +020018377ssl_bc_use_keysize : integer
18378 Returns the symmetric cipher key size used in bits when the outgoing
Christopher Fauletd92ea7f2020-04-30 10:03:55 +020018379 connection was made over an SSL/TLS transport layer. It can be used in a
18380 tcp-check or an http-check ruleset.
Emeric Brun645ae792014-04-30 14:21:06 +020018381
Willy Tarreau74ca5042013-06-11 23:12:07 +020018382ssl_c_ca_err : integer
18383 When the incoming connection was made over an SSL/TLS transport layer,
18384 returns the ID of the first error detected during verification of the client
18385 certificate at depth > 0, or 0 if no error was encountered during this
18386 verification process. Please refer to your SSL library's documentation to
18387 find the exhaustive list of error codes.
Willy Tarreauc735a072011-03-29 00:57:02 +020018388
Willy Tarreau74ca5042013-06-11 23:12:07 +020018389ssl_c_ca_err_depth : integer
18390 When the incoming connection was made over an SSL/TLS transport layer,
18391 returns the depth in the CA chain of the first error detected during the
18392 verification of the client certificate. If no error is encountered, 0 is
18393 returned.
Willy Tarreau0ba27502007-12-24 16:55:16 +010018394
Christopher Faulet70d10d12020-11-06 12:10:33 +010018395ssl_c_chain_der : binary
William Dauchya598b502020-08-06 18:11:38 +020018396 Returns the DER formatted chain certificate presented by the client when the
18397 incoming connection was made over an SSL/TLS transport layer. When used for
18398 an ACL, the value(s) to match against can be passed in hexadecimal form. One
Ilya Shipitsin2272d8a2020-12-21 01:22:40 +050018399 can parse the result with any lib accepting ASN.1 DER data. It currently
William Dauchya598b502020-08-06 18:11:38 +020018400 does not support resumed sessions.
18401
Christopher Faulet70d10d12020-11-06 12:10:33 +010018402ssl_c_der : binary
18403 Returns the DER formatted certificate presented by the client when the
18404 incoming connection was made over an SSL/TLS transport layer. When used for
18405 an ACL, the value(s) to match against can be passed in hexadecimal form.
18406
Willy Tarreau74ca5042013-06-11 23:12:07 +020018407ssl_c_err : integer
18408 When the incoming connection was made over an SSL/TLS transport layer,
18409 returns the ID of the first error detected during verification at depth 0, or
18410 0 if no error was encountered during this verification process. Please refer
18411 to your SSL library's documentation to find the exhaustive list of error
18412 codes.
Willy Tarreau62644772008-07-16 18:36:06 +020018413
Elliot Otchet71f82972020-01-15 08:12:14 -050018414ssl_c_i_dn([<entry>[,<occ>[,<format>]]]) : string
Willy Tarreau74ca5042013-06-11 23:12:07 +020018415 When the incoming connection was made over an SSL/TLS transport layer,
18416 returns the full distinguished name of the issuer of the certificate
18417 presented by the client when no <entry> is specified, or the value of the
18418 first given entry found from the beginning of the DN. If a positive/negative
18419 occurrence number is specified as the optional second argument, it returns
18420 the value of the nth given entry value from the beginning/end of the DN.
18421 For instance, "ssl_c_i_dn(OU,2)" the second organization unit, and
18422 "ssl_c_i_dn(CN)" retrieves the common name.
Elliot Otchet71f82972020-01-15 08:12:14 -050018423 The <format> parameter allows you to receive the DN suitable for
18424 consumption by different protocols. Currently supported is rfc2253 for
18425 LDAP v3.
18426 If you'd like to modify the format only you can specify an empty string
18427 and zero for the first two parameters. Example: ssl_c_i_dn(,0,rfc2253)
Willy Tarreau62644772008-07-16 18:36:06 +020018428
Willy Tarreau74ca5042013-06-11 23:12:07 +020018429ssl_c_key_alg : string
18430 Returns the name of the algorithm used to generate the key of the certificate
18431 presented by the client when the incoming connection was made over an SSL/TLS
18432 transport layer.
Willy Tarreau62644772008-07-16 18:36:06 +020018433
Willy Tarreau74ca5042013-06-11 23:12:07 +020018434ssl_c_notafter : string
18435 Returns the end date presented by the client as a formatted string
18436 YYMMDDhhmmss[Z] when the incoming connection was made over an SSL/TLS
18437 transport layer.
Emeric Brunbede3d02009-06-30 17:54:00 +020018438
Willy Tarreau74ca5042013-06-11 23:12:07 +020018439ssl_c_notbefore : string
18440 Returns the start date presented by the client as a formatted string
18441 YYMMDDhhmmss[Z] when the incoming connection was made over an SSL/TLS
18442 transport layer.
Willy Tarreaub6672b52011-12-12 17:23:41 +010018443
Elliot Otchet71f82972020-01-15 08:12:14 -050018444ssl_c_s_dn([<entry>[,<occ>[,<format>]]]) : string
Willy Tarreau74ca5042013-06-11 23:12:07 +020018445 When the incoming connection was made over an SSL/TLS transport layer,
18446 returns the full distinguished name of the subject of the certificate
18447 presented by the client when no <entry> is specified, or the value of the
18448 first given entry found from the beginning of the DN. If a positive/negative
18449 occurrence number is specified as the optional second argument, it returns
18450 the value of the nth given entry value from the beginning/end of the DN.
18451 For instance, "ssl_c_s_dn(OU,2)" the second organization unit, and
18452 "ssl_c_s_dn(CN)" retrieves the common name.
Elliot Otchet71f82972020-01-15 08:12:14 -050018453 The <format> parameter allows you to receive the DN suitable for
18454 consumption by different protocols. Currently supported is rfc2253 for
18455 LDAP v3.
18456 If you'd like to modify the format only you can specify an empty string
18457 and zero for the first two parameters. Example: ssl_c_s_dn(,0,rfc2253)
Willy Tarreaub6672b52011-12-12 17:23:41 +010018458
Willy Tarreau74ca5042013-06-11 23:12:07 +020018459ssl_c_serial : binary
18460 Returns the serial of the certificate presented by the client when the
18461 incoming connection was made over an SSL/TLS transport layer. When used for
18462 an ACL, the value(s) to match against can be passed in hexadecimal form.
Emeric Brun2525b6b2012-10-18 15:59:43 +020018463
Willy Tarreau74ca5042013-06-11 23:12:07 +020018464ssl_c_sha1 : binary
18465 Returns the SHA-1 fingerprint of the certificate presented by the client when
18466 the incoming connection was made over an SSL/TLS transport layer. This can be
18467 used to stick a client to a server, or to pass this information to a server.
Willy Tarreau2d0caa32014-07-02 19:01:22 +020018468 Note that the output is binary, so if you want to pass that signature to the
18469 server, you need to encode it in hex or base64, such as in the example below:
18470
Jarno Huuskonen676f6222017-03-30 09:19:45 +030018471 Example:
Willy Tarreau2d0caa32014-07-02 19:01:22 +020018472 http-request set-header X-SSL-Client-SHA1 %[ssl_c_sha1,hex]
Emeric Brun2525b6b2012-10-18 15:59:43 +020018473
Willy Tarreau74ca5042013-06-11 23:12:07 +020018474ssl_c_sig_alg : string
18475 Returns the name of the algorithm used to sign the certificate presented by
18476 the client when the incoming connection was made over an SSL/TLS transport
18477 layer.
Emeric Brun87855892012-10-17 17:39:35 +020018478
Willy Tarreau74ca5042013-06-11 23:12:07 +020018479ssl_c_used : boolean
18480 Returns true if current SSL session uses a client certificate even if current
18481 connection uses SSL session resumption. See also "ssl_fc_has_crt".
Emeric Brun7f56e742012-10-19 18:15:40 +020018482
Willy Tarreau74ca5042013-06-11 23:12:07 +020018483ssl_c_verify : integer
18484 Returns the verify result error ID when the incoming connection was made over
18485 an SSL/TLS transport layer, otherwise zero if no error is encountered. Please
18486 refer to your SSL library's documentation for an exhaustive list of error
18487 codes.
Emeric Brunce5ad802012-10-22 14:11:22 +020018488
Willy Tarreau74ca5042013-06-11 23:12:07 +020018489ssl_c_version : integer
18490 Returns the version of the certificate presented by the client when the
18491 incoming connection was made over an SSL/TLS transport layer.
Emeric Brunce5ad802012-10-22 14:11:22 +020018492
Emeric Brun43e79582014-10-29 19:03:26 +010018493ssl_f_der : binary
18494 Returns the DER formatted certificate presented by the frontend when the
18495 incoming connection was made over an SSL/TLS transport layer. When used for
18496 an ACL, the value(s) to match against can be passed in hexadecimal form.
18497
Elliot Otchet71f82972020-01-15 08:12:14 -050018498ssl_f_i_dn([<entry>[,<occ>[,<format>]]]) : string
Willy Tarreau74ca5042013-06-11 23:12:07 +020018499 When the incoming connection was made over an SSL/TLS transport layer,
18500 returns the full distinguished name of the issuer of the certificate
18501 presented by the frontend when no <entry> is specified, or the value of the
18502 first given entry found from the beginning of the DN. If a positive/negative
Emeric Brun87855892012-10-17 17:39:35 +020018503 occurrence number is specified as the optional second argument, it returns
Willy Tarreau74ca5042013-06-11 23:12:07 +020018504 the value of the nth given entry value from the beginning/end of the DN.
18505 For instance, "ssl_f_i_dn(OU,2)" the second organization unit, and
18506 "ssl_f_i_dn(CN)" retrieves the common name.
Elliot Otchet71f82972020-01-15 08:12:14 -050018507 The <format> parameter allows you to receive the DN suitable for
18508 consumption by different protocols. Currently supported is rfc2253 for
18509 LDAP v3.
18510 If you'd like to modify the format only you can specify an empty string
18511 and zero for the first two parameters. Example: ssl_f_i_dn(,0,rfc2253)
Emeric Brun87855892012-10-17 17:39:35 +020018512
Willy Tarreau74ca5042013-06-11 23:12:07 +020018513ssl_f_key_alg : string
18514 Returns the name of the algorithm used to generate the key of the certificate
18515 presented by the frontend when the incoming connection was made over an
18516 SSL/TLS transport layer.
Emeric Brun7f56e742012-10-19 18:15:40 +020018517
Willy Tarreau74ca5042013-06-11 23:12:07 +020018518ssl_f_notafter : string
18519 Returns the end date presented by the frontend as a formatted string
18520 YYMMDDhhmmss[Z] when the incoming connection was made over an SSL/TLS
18521 transport layer.
Emeric Brun2525b6b2012-10-18 15:59:43 +020018522
Willy Tarreau74ca5042013-06-11 23:12:07 +020018523ssl_f_notbefore : string
18524 Returns the start date presented by the frontend as a formatted string
18525 YYMMDDhhmmss[Z] when the incoming connection was made over an SSL/TLS
18526 transport layer.
Emeric Brun87855892012-10-17 17:39:35 +020018527
Elliot Otchet71f82972020-01-15 08:12:14 -050018528ssl_f_s_dn([<entry>[,<occ>[,<format>]]]) : string
Willy Tarreau74ca5042013-06-11 23:12:07 +020018529 When the incoming connection was made over an SSL/TLS transport layer,
18530 returns the full distinguished name of the subject of the certificate
18531 presented by the frontend when no <entry> is specified, or the value of the
18532 first given entry found from the beginning of the DN. If a positive/negative
18533 occurrence number is specified as the optional second argument, it returns
18534 the value of the nth given entry value from the beginning/end of the DN.
18535 For instance, "ssl_f_s_dn(OU,2)" the second organization unit, and
18536 "ssl_f_s_dn(CN)" retrieves the common name.
Elliot Otchet71f82972020-01-15 08:12:14 -050018537 The <format> parameter allows you to receive the DN suitable for
18538 consumption by different protocols. Currently supported is rfc2253 for
18539 LDAP v3.
18540 If you'd like to modify the format only you can specify an empty string
18541 and zero for the first two parameters. Example: ssl_f_s_dn(,0,rfc2253)
Emeric Brunce5ad802012-10-22 14:11:22 +020018542
Willy Tarreau74ca5042013-06-11 23:12:07 +020018543ssl_f_serial : binary
18544 Returns the serial of the certificate presented by the frontend when the
18545 incoming connection was made over an SSL/TLS transport layer. When used for
18546 an ACL, the value(s) to match against can be passed in hexadecimal form.
Emeric Brun87855892012-10-17 17:39:35 +020018547
Emeric Brun55f4fa82014-04-30 17:11:25 +020018548ssl_f_sha1 : binary
18549 Returns the SHA-1 fingerprint of the certificate presented by the frontend
18550 when the incoming connection was made over an SSL/TLS transport layer. This
18551 can be used to know which certificate was chosen using SNI.
18552
Willy Tarreau74ca5042013-06-11 23:12:07 +020018553ssl_f_sig_alg : string
18554 Returns the name of the algorithm used to sign the certificate presented by
18555 the frontend when the incoming connection was made over an SSL/TLS transport
18556 layer.
Emeric Brun7f56e742012-10-19 18:15:40 +020018557
Willy Tarreau74ca5042013-06-11 23:12:07 +020018558ssl_f_version : integer
18559 Returns the version of the certificate presented by the frontend when the
18560 incoming connection was made over an SSL/TLS transport layer.
18561
18562ssl_fc : boolean
Emeric Brun2525b6b2012-10-18 15:59:43 +020018563 Returns true when the front connection was made via an SSL/TLS transport
18564 layer and is locally deciphered. This means it has matched a socket declared
18565 with a "bind" line having the "ssl" option.
18566
Willy Tarreau74ca5042013-06-11 23:12:07 +020018567 Example :
18568 # This passes "X-Proto: https" to servers when client connects over SSL
18569 listen http-https
18570 bind :80
18571 bind :443 ssl crt /etc/haproxy.pem
18572 http-request add-header X-Proto https if { ssl_fc }
18573
18574ssl_fc_alg_keysize : integer
18575 Returns the symmetric cipher key size supported in bits when the incoming
18576 connection was made over an SSL/TLS transport layer.
18577
18578ssl_fc_alpn : string
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030018579 This extracts the Application Layer Protocol Negotiation field from an
Willy Tarreau74ca5042013-06-11 23:12:07 +020018580 incoming connection made via a TLS transport layer and locally deciphered by
Daniel Corbett9f0843f2021-05-08 10:50:37 -040018581 HAProxy. The result is a string containing the protocol name advertised by
Willy Tarreau74ca5042013-06-11 23:12:07 +020018582 the client. The SSL library must have been built with support for TLS
18583 extensions enabled (check haproxy -vv). Note that the TLS ALPN extension is
18584 not advertised unless the "alpn" keyword on the "bind" line specifies a
18585 protocol list. Also, nothing forces the client to pick a protocol from this
18586 list, any other one may be requested. The TLS ALPN extension is meant to
18587 replace the TLS NPN extension. See also "ssl_fc_npn".
18588
Willy Tarreau74ca5042013-06-11 23:12:07 +020018589ssl_fc_cipher : string
18590 Returns the name of the used cipher when the incoming connection was made
18591 over an SSL/TLS transport layer.
Willy Tarreauab861d32013-04-02 02:30:41 +020018592
Thierry FOURNIER5bf77322017-02-25 12:45:22 +010018593ssl_fc_cipherlist_bin : binary
18594 Returns the binary form of the client hello cipher list. The maximum returned
18595 value length is according with the value of
Emmanuel Hocdetaaee7502017-03-07 18:34:58 +010018596 "tune.ssl.capture-cipherlist-size".
Thierry FOURNIER5bf77322017-02-25 12:45:22 +010018597
18598ssl_fc_cipherlist_hex : string
18599 Returns the binary form of the client hello cipher list encoded as
18600 hexadecimal. The maximum returned value length is according with the value of
Emmanuel Hocdetaaee7502017-03-07 18:34:58 +010018601 "tune.ssl.capture-cipherlist-size".
Thierry FOURNIER5bf77322017-02-25 12:45:22 +010018602
18603ssl_fc_cipherlist_str : string
18604 Returns the decoded text form of the client hello cipher list. The maximum
18605 number of ciphers returned is according with the value of
18606 "tune.ssl.capture-cipherlist-size". Note that this sample-fetch is only
Davor Ocelice9ed2812017-12-25 17:49:28 +010018607 available with OpenSSL >= 1.0.2. If the function is not enabled, this
Emmanuel Hocdetddcde192017-09-01 17:32:08 +020018608 sample-fetch returns the hash like "ssl_fc_cipherlist_xxh".
Thierry FOURNIER5bf77322017-02-25 12:45:22 +010018609
18610ssl_fc_cipherlist_xxh : integer
18611 Returns a xxh64 of the cipher list. This hash can be return only is the value
18612 "tune.ssl.capture-cipherlist-size" is set greater than 0, however the hash
Emmanuel Hocdetaaee7502017-03-07 18:34:58 +010018613 take in account all the data of the cipher list.
Thierry FOURNIER5bf77322017-02-25 12:45:22 +010018614
Patrick Hemmer65674662019-06-04 08:13:03 -040018615ssl_fc_client_random : binary
18616 Returns the client random of the front connection when the incoming connection
18617 was made over an SSL/TLS transport layer. It is useful to to decrypt traffic
18618 sent using ephemeral ciphers. This requires OpenSSL >= 1.1.0, or BoringSSL.
18619
William Lallemand7d42ef52020-07-06 11:41:30 +020018620ssl_fc_client_early_traffic_secret : string
18621 Return the CLIENT_EARLY_TRAFFIC_SECRET as an hexadecimal string for the
18622 front connection when the incoming connection was made over a TLS 1.3
18623 transport layer.
18624 Require OpenSSL >= 1.1.1. This is one of the keys dumped by the OpenSSL
18625 keylog callback to generate the SSLKEYLOGFILE. The SSL Key logging must be
18626 activated with "tune.ssl.keylog on" in the global section. See also
18627 "tune.ssl.keylog"
18628
18629ssl_fc_client_handshake_traffic_secret : string
18630 Return the CLIENT_HANDSHAKE_TRAFFIC_SECRET as an hexadecimal string for the
18631 front connection when the incoming connection was made over a TLS 1.3
18632 transport layer.
18633 Require OpenSSL >= 1.1.1. This is one of the keys dumped by the OpenSSL
18634 keylog callback to generate the SSLKEYLOGFILE. The SSL Key logging must be
18635 activated with "tune.ssl.keylog on" in the global section. See also
18636 "tune.ssl.keylog"
18637
18638ssl_fc_client_traffic_secret_0 : string
18639 Return the CLIENT_TRAFFIC_SECRET_0 as an hexadecimal string for the
18640 front connection when the incoming connection was made over a TLS 1.3
18641 transport layer.
18642 Require OpenSSL >= 1.1.1. This is one of the keys dumped by the OpenSSL
18643 keylog callback to generate the SSLKEYLOGFILE. The SSL Key logging must be
18644 activated with "tune.ssl.keylog on" in the global section. See also
18645 "tune.ssl.keylog"
18646
18647ssl_fc_exporter_secret : string
18648 Return the EXPORTER_SECRET as an hexadecimal string for the
18649 front connection when the incoming connection was made over a TLS 1.3
18650 transport layer.
18651 Require OpenSSL >= 1.1.1. This is one of the keys dumped by the OpenSSL
18652 keylog callback to generate the SSLKEYLOGFILE. The SSL Key logging must be
18653 activated with "tune.ssl.keylog on" in the global section. See also
18654 "tune.ssl.keylog"
18655
18656ssl_fc_early_exporter_secret : string
18657 Return the EARLY_EXPORTER_SECRET as an hexadecimal string for the
18658 front connection when the incoming connection was made over an TLS 1.3
18659 transport layer.
18660 Require OpenSSL >= 1.1.1. This is one of the keys dumped by the OpenSSL
18661 keylog callback to generate the SSLKEYLOGFILE. The SSL Key logging must be
18662 activated with "tune.ssl.keylog on" in the global section. See also
18663 "tune.ssl.keylog"
18664
Willy Tarreau74ca5042013-06-11 23:12:07 +020018665ssl_fc_has_crt : boolean
Emeric Brun2525b6b2012-10-18 15:59:43 +020018666 Returns true if a client certificate is present in an incoming connection over
18667 SSL/TLS transport layer. Useful if 'verify' statement is set to 'optional'.
Emeric Brun9143d372012-12-20 15:44:16 +010018668 Note: on SSL session resumption with Session ID or TLS ticket, client
18669 certificate is not present in the current connection but may be retrieved
18670 from the cache or the ticket. So prefer "ssl_c_used" if you want to check if
18671 current SSL session uses a client certificate.
Emeric Brun2525b6b2012-10-18 15:59:43 +020018672
Olivier Houchardccaa7de2017-10-02 11:51:03 +020018673ssl_fc_has_early : boolean
18674 Returns true if early data were sent, and the handshake didn't happen yet. As
18675 it has security implications, it is useful to be able to refuse those, or
18676 wait until the handshake happened.
18677
Willy Tarreau74ca5042013-06-11 23:12:07 +020018678ssl_fc_has_sni : boolean
18679 This checks for the presence of a Server Name Indication TLS extension (SNI)
Willy Tarreauf7bc57c2012-10-03 00:19:48 +020018680 in an incoming connection was made over an SSL/TLS transport layer. Returns
18681 true when the incoming connection presents a TLS SNI field. This requires
John Roeslerfb2fce12019-07-10 15:45:51 -050018682 that the SSL library is built with support for TLS extensions enabled (check
Willy Tarreauf7bc57c2012-10-03 00:19:48 +020018683 haproxy -vv).
Willy Tarreau7875d092012-09-10 08:20:03 +020018684
Nenad Merdanovic1516fe32016-05-17 03:31:21 +020018685ssl_fc_is_resumed : boolean
Nenad Merdanovic26ea8222015-05-18 02:28:57 +020018686 Returns true if the SSL/TLS session has been resumed through the use of
Jérôme Magnin4a326cb2018-01-15 14:01:17 +010018687 SSL session cache or TLS tickets on an incoming connection over an SSL/TLS
18688 transport layer.
Nenad Merdanovic26ea8222015-05-18 02:28:57 +020018689
Willy Tarreau74ca5042013-06-11 23:12:07 +020018690ssl_fc_npn : string
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030018691 This extracts the Next Protocol Negotiation field from an incoming connection
Daniel Corbett9f0843f2021-05-08 10:50:37 -040018692 made via a TLS transport layer and locally deciphered by HAProxy. The result
Willy Tarreau74ca5042013-06-11 23:12:07 +020018693 is a string containing the protocol name advertised by the client. The SSL
18694 library must have been built with support for TLS extensions enabled (check
18695 haproxy -vv). Note that the TLS NPN extension is not advertised unless the
18696 "npn" keyword on the "bind" line specifies a protocol list. Also, nothing
18697 forces the client to pick a protocol from this list, any other one may be
18698 requested. Please note that the TLS NPN extension was replaced with ALPN.
Willy Tarreaua33c6542012-10-15 13:19:06 +020018699
Willy Tarreau74ca5042013-06-11 23:12:07 +020018700ssl_fc_protocol : string
18701 Returns the name of the used protocol when the incoming connection was made
18702 over an SSL/TLS transport layer.
Willy Tarreau7875d092012-09-10 08:20:03 +020018703
Emeric Brunb73a9b02014-04-30 18:49:19 +020018704ssl_fc_unique_id : binary
David Sc1ad52e2014-04-08 18:48:47 -040018705 When the incoming connection was made over an SSL/TLS transport layer,
Emeric Brunb73a9b02014-04-30 18:49:19 +020018706 returns the TLS unique ID as defined in RFC5929 section 3. The unique id
18707 can be encoded to base64 using the converter: "ssl_bc_unique_id,base64".
David Sc1ad52e2014-04-08 18:48:47 -040018708
William Lallemand7d42ef52020-07-06 11:41:30 +020018709ssl_fc_server_handshake_traffic_secret : string
18710 Return the SERVER_HANDSHAKE_TRAFFIC_SECRET as an hexadecimal string for the
18711 front connection when the incoming connection was made over a TLS 1.3
18712 transport layer.
18713 Require OpenSSL >= 1.1.1. This is one of the keys dumped by the OpenSSL
18714 keylog callback to generate the SSLKEYLOGFILE. The SSL Key logging must be
18715 activated with "tune.ssl.keylog on" in the global section. See also
18716 "tune.ssl.keylog"
18717
18718ssl_fc_server_traffic_secret_0 : string
18719 Return the SERVER_TRAFFIC_SECRET_0 as an hexadecimal string for the
18720 front connection when the incoming connection was made over an TLS 1.3
18721 transport layer.
18722 Require OpenSSL >= 1.1.1. This is one of the keys dumped by the OpenSSL
18723 keylog callback to generate the SSLKEYLOGFILE. The SSL Key logging must be
18724 activated with "tune.ssl.keylog on" in the global section. See also
18725 "tune.ssl.keylog"
18726
Patrick Hemmer65674662019-06-04 08:13:03 -040018727ssl_fc_server_random : binary
18728 Returns the server random of the front connection when the incoming connection
18729 was made over an SSL/TLS transport layer. It is useful to to decrypt traffic
18730 sent using ephemeral ciphers. This requires OpenSSL >= 1.1.0, or BoringSSL.
18731
Willy Tarreau74ca5042013-06-11 23:12:07 +020018732ssl_fc_session_id : binary
18733 Returns the SSL ID of the front connection when the incoming connection was
18734 made over an SSL/TLS transport layer. It is useful to stick a given client to
18735 a server. It is important to note that some browsers refresh their session ID
18736 every few minutes.
Willy Tarreau7875d092012-09-10 08:20:03 +020018737
Patrick Hemmere0275472018-04-28 19:15:51 -040018738ssl_fc_session_key : binary
18739 Returns the SSL session master key of the front connection when the incoming
18740 connection was made over an SSL/TLS transport layer. It is useful to decrypt
18741 traffic sent using ephemeral ciphers. This requires OpenSSL >= 1.1.0, or
18742 BoringSSL.
18743
18744
Willy Tarreau74ca5042013-06-11 23:12:07 +020018745ssl_fc_sni : string
18746 This extracts the Server Name Indication TLS extension (SNI) field from an
18747 incoming connection made via an SSL/TLS transport layer and locally
Daniel Corbett9f0843f2021-05-08 10:50:37 -040018748 deciphered by HAProxy. The result (when present) typically is a string
Willy Tarreau74ca5042013-06-11 23:12:07 +020018749 matching the HTTPS host name (253 chars or less). The SSL library must have
18750 been built with support for TLS extensions enabled (check haproxy -vv).
18751
Alex18c88772021-06-05 13:23:08 +020018752 This fetch is different from "req.ssl_sni" above in that it applies to the
Daniel Corbett9f0843f2021-05-08 10:50:37 -040018753 connection being deciphered by HAProxy and not to SSL contents being blindly
Willy Tarreau74ca5042013-06-11 23:12:07 +020018754 forwarded. See also "ssl_fc_sni_end" and "ssl_fc_sni_reg" below. This
John Roeslerfb2fce12019-07-10 15:45:51 -050018755 requires that the SSL library is built with support for TLS extensions
Cyril Bonté9c1eb1e2012-10-09 22:45:34 +020018756 enabled (check haproxy -vv).
Willy Tarreau62644772008-07-16 18:36:06 +020018757
Willy Tarreau74ca5042013-06-11 23:12:07 +020018758 ACL derivatives :
Willy Tarreau74ca5042013-06-11 23:12:07 +020018759 ssl_fc_sni_end : suffix match
18760 ssl_fc_sni_reg : regex match
Emeric Brun589fcad2012-10-16 14:13:26 +020018761
Willy Tarreau74ca5042013-06-11 23:12:07 +020018762ssl_fc_use_keysize : integer
18763 Returns the symmetric cipher key size used in bits when the incoming
18764 connection was made over an SSL/TLS transport layer.
Willy Tarreaub6fb4202008-07-20 11:18:28 +020018765
William Lallemandbfa3e812020-06-25 20:07:18 +020018766ssl_s_der : binary
18767 Returns the DER formatted certificate presented by the server when the
18768 outgoing connection was made over an SSL/TLS transport layer. When used for
18769 an ACL, the value(s) to match against can be passed in hexadecimal form.
18770
William Dauchya598b502020-08-06 18:11:38 +020018771ssl_s_chain_der : binary
18772 Returns the DER formatted chain certificate presented by the server when the
18773 outgoing connection was made over an SSL/TLS transport layer. When used for
18774 an ACL, the value(s) to match against can be passed in hexadecimal form. One
Ilya Shipitsin2272d8a2020-12-21 01:22:40 +050018775 can parse the result with any lib accepting ASN.1 DER data. It currently
William Dauchya598b502020-08-06 18:11:38 +020018776 does not support resumed sessions.
18777
William Lallemandbfa3e812020-06-25 20:07:18 +020018778ssl_s_key_alg : string
18779 Returns the name of the algorithm used to generate the key of the certificate
18780 presented by the server when the outgoing connection was made over an
18781 SSL/TLS transport layer.
18782
18783ssl_s_notafter : string
18784 Returns the end date presented by the server as a formatted string
18785 YYMMDDhhmmss[Z] when the outgoing connection was made over an SSL/TLS
18786 transport layer.
18787
18788ssl_s_notbefore : string
18789 Returns the start date presented by the server as a formatted string
18790 YYMMDDhhmmss[Z] when the outgoing connection was made over an SSL/TLS
18791 transport layer.
18792
18793ssl_s_i_dn([<entry>[,<occ>[,<format>]]]) : string
18794 When the outgoing connection was made over an SSL/TLS transport layer,
18795 returns the full distinguished name of the issuer of the certificate
18796 presented by the server when no <entry> is specified, or the value of the
18797 first given entry found from the beginning of the DN. If a positive/negative
18798 occurrence number is specified as the optional second argument, it returns
18799 the value of the nth given entry value from the beginning/end of the DN.
William Lallemand8f600c82020-06-26 09:55:06 +020018800 For instance, "ssl_s_i_dn(OU,2)" the second organization unit, and
18801 "ssl_s_i_dn(CN)" retrieves the common name.
William Lallemandbfa3e812020-06-25 20:07:18 +020018802 The <format> parameter allows you to receive the DN suitable for
18803 consumption by different protocols. Currently supported is rfc2253 for
18804 LDAP v3.
18805 If you'd like to modify the format only you can specify an empty string
18806 and zero for the first two parameters. Example: ssl_s_i_dn(,0,rfc2253)
18807
18808ssl_s_s_dn([<entry>[,<occ>[,<format>]]]) : string
18809 When the outgoing connection was made over an SSL/TLS transport layer,
18810 returns the full distinguished name of the subject of the certificate
18811 presented by the server when no <entry> is specified, or the value of the
18812 first given entry found from the beginning of the DN. If a positive/negative
18813 occurrence number is specified as the optional second argument, it returns
18814 the value of the nth given entry value from the beginning/end of the DN.
William Lallemand8f600c82020-06-26 09:55:06 +020018815 For instance, "ssl_s_s_dn(OU,2)" the second organization unit, and
18816 "ssl_s_s_dn(CN)" retrieves the common name.
William Lallemandbfa3e812020-06-25 20:07:18 +020018817 The <format> parameter allows you to receive the DN suitable for
18818 consumption by different protocols. Currently supported is rfc2253 for
18819 LDAP v3.
18820 If you'd like to modify the format only you can specify an empty string
18821 and zero for the first two parameters. Example: ssl_s_s_dn(,0,rfc2253)
18822
18823ssl_s_serial : binary
18824 Returns the serial of the certificate presented by the server when the
18825 outgoing connection was made over an SSL/TLS transport layer. When used for
18826 an ACL, the value(s) to match against can be passed in hexadecimal form.
18827
18828ssl_s_sha1 : binary
18829 Returns the SHA-1 fingerprint of the certificate presented by the server
18830 when the outgoing connection was made over an SSL/TLS transport layer. This
18831 can be used to know which certificate was chosen using SNI.
18832
18833ssl_s_sig_alg : string
18834 Returns the name of the algorithm used to sign the certificate presented by
18835 the server when the outgoing connection was made over an SSL/TLS transport
18836 layer.
18837
18838ssl_s_version : integer
18839 Returns the version of the certificate presented by the server when the
18840 outgoing connection was made over an SSL/TLS transport layer.
Willy Tarreaub6fb4202008-07-20 11:18:28 +020018841
Thierry FOURNIER060762e2014-04-23 13:29:15 +0200188427.3.5. Fetching samples from buffer contents (Layer 6)
Willy Tarreau74ca5042013-06-11 23:12:07 +020018843------------------------------------------------------
Willy Tarreaub6fb4202008-07-20 11:18:28 +020018844
Willy Tarreau74ca5042013-06-11 23:12:07 +020018845Fetching samples from buffer contents is a bit different from the previous
18846sample fetches above because the sampled data are ephemeral. These data can
18847only be used when they're available and will be lost when they're forwarded.
18848For this reason, samples fetched from buffer contents during a request cannot
18849be used in a response for example. Even while the data are being fetched, they
18850can change. Sometimes it is necessary to set some delays or combine multiple
18851sample fetch methods to ensure that the expected data are complete and usable,
18852for example through TCP request content inspection. Please see the "tcp-request
18853content" keyword for more detailed information on the subject.
Willy Tarreau62644772008-07-16 18:36:06 +020018854
Christopher Fauleta434a002021-03-25 11:58:51 +010018855Warning : Following sample fetches are ignored if used from HTTP proxies. They
18856 only deal with raw contents found in the buffers. On their side,
Daniel Corbett9f0843f2021-05-08 10:50:37 -040018857 HTTP proxies use structured content. Thus raw representation of
Christopher Fauleta434a002021-03-25 11:58:51 +010018858 these data are meaningless. A warning is emitted if an ACL relies on
18859 one of the following sample fetches. But it is not possible to detect
18860 all invalid usage (for instance inside a log-format string or a
18861 sample expression). So be careful.
18862
Willy Tarreau74ca5042013-06-11 23:12:07 +020018863payload(<offset>,<length>) : binary (deprecated)
Davor Ocelice9ed2812017-12-25 17:49:28 +010018864 This is an alias for "req.payload" when used in the context of a request (e.g.
Willy Tarreau74ca5042013-06-11 23:12:07 +020018865 "stick on", "stick match"), and for "res.payload" when used in the context of
18866 a response such as in "stick store response".
Willy Tarreau0ba27502007-12-24 16:55:16 +010018867
Willy Tarreau74ca5042013-06-11 23:12:07 +020018868payload_lv(<offset1>,<length>[,<offset2>]) : binary (deprecated)
18869 This is an alias for "req.payload_lv" when used in the context of a request
Davor Ocelice9ed2812017-12-25 17:49:28 +010018870 (e.g. "stick on", "stick match"), and for "res.payload_lv" when used in the
Willy Tarreau74ca5042013-06-11 23:12:07 +020018871 context of a response such as in "stick store response".
Willy Tarreau0ba27502007-12-24 16:55:16 +010018872
Willy Tarreau74ca5042013-06-11 23:12:07 +020018873req.len : integer
18874req_len : integer (deprecated)
18875 Returns an integer value corresponding to the number of bytes present in the
18876 request buffer. This is mostly used in ACL. It is important to understand
18877 that this test does not return false as long as the buffer is changing. This
18878 means that a check with equality to zero will almost always immediately match
18879 at the beginning of the session, while a test for more data will wait for
Daniel Corbett9f0843f2021-05-08 10:50:37 -040018880 that data to come in and return false only when HAProxy is certain that no
Willy Tarreau74ca5042013-06-11 23:12:07 +020018881 more data will come in. This test was designed to be used with TCP request
18882 content inspection.
Willy Tarreaua7ad50c2012-04-29 15:39:40 +020018883
Willy Tarreau74ca5042013-06-11 23:12:07 +020018884req.payload(<offset>,<length>) : binary
18885 This extracts a binary block of <length> bytes and starting at byte <offset>
Willy Tarreau00f00842013-08-02 11:07:32 +020018886 in the request buffer. As a special case, if the <length> argument is zero,
18887 the the whole buffer from <offset> to the end is extracted. This can be used
18888 with ACLs in order to check for the presence of some content in a buffer at
18889 any location.
Willy Tarreaua7ad50c2012-04-29 15:39:40 +020018890
Willy Tarreau74ca5042013-06-11 23:12:07 +020018891 ACL alternatives :
18892 payload(<offset>,<length>) : hex binary match
Willy Tarreaua7ad50c2012-04-29 15:39:40 +020018893
Willy Tarreau74ca5042013-06-11 23:12:07 +020018894req.payload_lv(<offset1>,<length>[,<offset2>]) : binary
18895 This extracts a binary block whose size is specified at <offset1> for <length>
18896 bytes, and which starts at <offset2> if specified or just after the length in
18897 the request buffer. The <offset2> parameter also supports relative offsets if
18898 prepended with a '+' or '-' sign.
Willy Tarreaua7ad50c2012-04-29 15:39:40 +020018899
Willy Tarreau74ca5042013-06-11 23:12:07 +020018900 ACL alternatives :
18901 payload_lv(<offset1>,<length>[,<offset2>]) : hex binary match
Willy Tarreaua7ad50c2012-04-29 15:39:40 +020018902
Willy Tarreau74ca5042013-06-11 23:12:07 +020018903 Example : please consult the example from the "stick store-response" keyword.
Willy Tarreaua7ad50c2012-04-29 15:39:40 +020018904
Willy Tarreau74ca5042013-06-11 23:12:07 +020018905req.proto_http : boolean
18906req_proto_http : boolean (deprecated)
18907 Returns true when data in the request buffer look like HTTP and correctly
18908 parses as such. It is the same parser as the common HTTP request parser which
18909 is used so there should be no surprises. The test does not match until the
18910 request is complete, failed or timed out. This test may be used to report the
18911 protocol in TCP logs, but the biggest use is to block TCP request analysis
18912 until a complete HTTP request is present in the buffer, for example to track
18913 a header.
Willy Tarreaua7ad50c2012-04-29 15:39:40 +020018914
Willy Tarreau74ca5042013-06-11 23:12:07 +020018915 Example:
18916 # track request counts per "base" (concatenation of Host+URL)
18917 tcp-request inspect-delay 10s
18918 tcp-request content reject if !HTTP
Willy Tarreaube4a3ef2013-06-17 15:04:07 +020018919 tcp-request content track-sc0 base table req-rate
Willy Tarreaua7ad50c2012-04-29 15:39:40 +020018920
Willy Tarreau74ca5042013-06-11 23:12:07 +020018921req.rdp_cookie([<name>]) : string
18922rdp_cookie([<name>]) : string (deprecated)
18923 When the request buffer looks like the RDP protocol, extracts the RDP cookie
18924 <name>, or any cookie if unspecified. The parser only checks for the first
18925 cookie, as illustrated in the RDP protocol specification. The cookie name is
18926 case insensitive. Generally the "MSTS" cookie name will be used, as it can
18927 contain the user name of the client connecting to the server if properly
18928 configured on the client. The "MSTSHASH" cookie is often used as well for
18929 session stickiness to servers.
Willy Tarreau04aa6a92012-04-06 18:57:55 +020018930
Willy Tarreau74ca5042013-06-11 23:12:07 +020018931 This differs from "balance rdp-cookie" in that any balancing algorithm may be
18932 used and thus the distribution of clients to backend servers is not linked to
18933 a hash of the RDP cookie. It is envisaged that using a balancing algorithm
18934 such as "balance roundrobin" or "balance leastconn" will lead to a more even
18935 distribution of clients to backend servers than the hash used by "balance
18936 rdp-cookie".
Willy Tarreau04aa6a92012-04-06 18:57:55 +020018937
Willy Tarreau74ca5042013-06-11 23:12:07 +020018938 ACL derivatives :
18939 req_rdp_cookie([<name>]) : exact string match
Willy Tarreau04aa6a92012-04-06 18:57:55 +020018940
Willy Tarreau74ca5042013-06-11 23:12:07 +020018941 Example :
18942 listen tse-farm
18943 bind 0.0.0.0:3389
18944 # wait up to 5s for an RDP cookie in the request
18945 tcp-request inspect-delay 5s
18946 tcp-request content accept if RDP_COOKIE
18947 # apply RDP cookie persistence
18948 persist rdp-cookie
18949 # Persist based on the mstshash cookie
18950 # This is only useful makes sense if
18951 # balance rdp-cookie is not used
18952 stick-table type string size 204800
18953 stick on req.rdp_cookie(mstshash)
18954 server srv1 1.1.1.1:3389
18955 server srv1 1.1.1.2:3389
Willy Tarreau04aa6a92012-04-06 18:57:55 +020018956
Willy Tarreau74ca5042013-06-11 23:12:07 +020018957 See also : "balance rdp-cookie", "persist rdp-cookie", "tcp-request" and the
18958 "req_rdp_cookie" ACL.
Willy Tarreau04aa6a92012-04-06 18:57:55 +020018959
Willy Tarreau74ca5042013-06-11 23:12:07 +020018960req.rdp_cookie_cnt([name]) : integer
18961rdp_cookie_cnt([name]) : integer (deprecated)
18962 Tries to parse the request buffer as RDP protocol, then returns an integer
18963 corresponding to the number of RDP cookies found. If an optional cookie name
18964 is passed, only cookies matching this name are considered. This is mostly
18965 used in ACL.
Willy Tarreau04aa6a92012-04-06 18:57:55 +020018966
Willy Tarreau74ca5042013-06-11 23:12:07 +020018967 ACL derivatives :
18968 req_rdp_cookie_cnt([<name>]) : integer match
Willy Tarreau04aa6a92012-04-06 18:57:55 +020018969
Alex Zorin4afdd132018-12-30 13:56:28 +110018970req.ssl_alpn : string
18971 Returns a string containing the values of the Application-Layer Protocol
18972 Negotiation (ALPN) TLS extension (RFC7301), sent by the client within the SSL
18973 ClientHello message. Note that this only applies to raw contents found in the
18974 request buffer and not to the contents deciphered via an SSL data layer, so
18975 this will not work with "bind" lines having the "ssl" option. This is useful
18976 in ACL to make a routing decision based upon the ALPN preferences of a TLS
Jarno Huuskonene504f812019-01-03 07:56:49 +020018977 client, like in the example below. See also "ssl_fc_alpn".
Alex Zorin4afdd132018-12-30 13:56:28 +110018978
18979 Examples :
18980 # Wait for a client hello for at most 5 seconds
18981 tcp-request inspect-delay 5s
18982 tcp-request content accept if { req_ssl_hello_type 1 }
Jarno Huuskonene504f812019-01-03 07:56:49 +020018983 use_backend bk_acme if { req.ssl_alpn acme-tls/1 }
Alex Zorin4afdd132018-12-30 13:56:28 +110018984 default_backend bk_default
18985
Nenad Merdanovic5fc7d7e2015-07-07 22:00:17 +020018986req.ssl_ec_ext : boolean
18987 Returns a boolean identifying if client sent the Supported Elliptic Curves
18988 Extension as defined in RFC4492, section 5.1. within the SSL ClientHello
Cyril Bonté307ee1e2015-09-28 23:16:06 +020018989 message. This can be used to present ECC compatible clients with EC
18990 certificate and to use RSA for all others, on the same IP address. Note that
18991 this only applies to raw contents found in the request buffer and not to
18992 contents deciphered via an SSL data layer, so this will not work with "bind"
18993 lines having the "ssl" option.
Nenad Merdanovic5fc7d7e2015-07-07 22:00:17 +020018994
Willy Tarreau74ca5042013-06-11 23:12:07 +020018995req.ssl_hello_type : integer
18996req_ssl_hello_type : integer (deprecated)
18997 Returns an integer value containing the type of the SSL hello message found
18998 in the request buffer if the buffer contains data that parse as a complete
18999 SSL (v3 or superior) client hello message. Note that this only applies to raw
19000 contents found in the request buffer and not to contents deciphered via an
19001 SSL data layer, so this will not work with "bind" lines having the "ssl"
19002 option. This is mostly used in ACL to detect presence of an SSL hello message
19003 that is supposed to contain an SSL session ID usable for stickiness.
Willy Tarreau04aa6a92012-04-06 18:57:55 +020019004
Willy Tarreau74ca5042013-06-11 23:12:07 +020019005req.ssl_sni : string
19006req_ssl_sni : string (deprecated)
19007 Returns a string containing the value of the Server Name TLS extension sent
19008 by a client in a TLS stream passing through the request buffer if the buffer
19009 contains data that parse as a complete SSL (v3 or superior) client hello
19010 message. Note that this only applies to raw contents found in the request
19011 buffer and not to contents deciphered via an SSL data layer, so this will not
Lukas Tribusa267b5d2020-07-19 00:25:06 +020019012 work with "bind" lines having the "ssl" option. This will only work for actual
19013 implicit TLS based protocols like HTTPS (443), IMAPS (993), SMTPS (465),
19014 however it will not work for explicit TLS based protocols, like SMTP (25/587)
19015 or IMAP (143). SNI normally contains the name of the host the client tries to
19016 connect to (for recent browsers). SNI is useful for allowing or denying access
19017 to certain hosts when SSL/TLS is used by the client. This test was designed to
19018 be used with TCP request content inspection. If content switching is needed,
19019 it is recommended to first wait for a complete client hello (type 1), like in
19020 the example below. See also "ssl_fc_sni".
Willy Tarreau04aa6a92012-04-06 18:57:55 +020019021
Willy Tarreau74ca5042013-06-11 23:12:07 +020019022 ACL derivatives :
Alex18c88772021-06-05 13:23:08 +020019023 req.ssl_sni : exact string match
Willy Tarreau04aa6a92012-04-06 18:57:55 +020019024
Willy Tarreau74ca5042013-06-11 23:12:07 +020019025 Examples :
19026 # Wait for a client hello for at most 5 seconds
19027 tcp-request inspect-delay 5s
19028 tcp-request content accept if { req_ssl_hello_type 1 }
Alex18c88772021-06-05 13:23:08 +020019029 use_backend bk_allow if { req.ssl_sni -f allowed_sites }
Willy Tarreau74ca5042013-06-11 23:12:07 +020019030 default_backend bk_sorry_page
Willy Tarreau04aa6a92012-04-06 18:57:55 +020019031
Pradeep Jindalbb2acf52015-09-29 10:12:57 +053019032req.ssl_st_ext : integer
19033 Returns 0 if the client didn't send a SessionTicket TLS Extension (RFC5077)
19034 Returns 1 if the client sent SessionTicket TLS Extension
19035 Returns 2 if the client also sent non-zero length TLS SessionTicket
19036 Note that this only applies to raw contents found in the request buffer and
19037 not to contents deciphered via an SSL data layer, so this will not work with
19038 "bind" lines having the "ssl" option. This can for example be used to detect
19039 whether the client sent a SessionTicket or not and stick it accordingly, if
19040 no SessionTicket then stick on SessionID or don't stick as there's no server
19041 side state is there when SessionTickets are in use.
19042
Willy Tarreau74ca5042013-06-11 23:12:07 +020019043req.ssl_ver : integer
19044req_ssl_ver : integer (deprecated)
19045 Returns an integer value containing the version of the SSL/TLS protocol of a
19046 stream present in the request buffer. Both SSLv2 hello messages and SSLv3
19047 messages are supported. TLSv1 is announced as SSL version 3.1. The value is
19048 composed of the major version multiplied by 65536, added to the minor
19049 version. Note that this only applies to raw contents found in the request
19050 buffer and not to contents deciphered via an SSL data layer, so this will not
19051 work with "bind" lines having the "ssl" option. The ACL version of the test
Davor Ocelice9ed2812017-12-25 17:49:28 +010019052 matches against a decimal notation in the form MAJOR.MINOR (e.g. 3.1). This
Willy Tarreau74ca5042013-06-11 23:12:07 +020019053 fetch is mostly used in ACL.
Willy Tarreaud63335a2010-02-26 12:56:52 +010019054
Willy Tarreau74ca5042013-06-11 23:12:07 +020019055 ACL derivatives :
19056 req_ssl_ver : decimal match
Willy Tarreaud63335a2010-02-26 12:56:52 +010019057
Willy Tarreau47e8eba2013-09-11 23:28:46 +020019058res.len : integer
19059 Returns an integer value corresponding to the number of bytes present in the
19060 response buffer. This is mostly used in ACL. It is important to understand
19061 that this test does not return false as long as the buffer is changing. This
19062 means that a check with equality to zero will almost always immediately match
19063 at the beginning of the session, while a test for more data will wait for
Daniel Corbett9f0843f2021-05-08 10:50:37 -040019064 that data to come in and return false only when HAProxy is certain that no
Willy Tarreau47e8eba2013-09-11 23:28:46 +020019065 more data will come in. This test was designed to be used with TCP response
Christopher Faulete596d182020-05-05 17:46:34 +020019066 content inspection. But it may also be used in tcp-check based expect rules.
Willy Tarreau47e8eba2013-09-11 23:28:46 +020019067
Willy Tarreau74ca5042013-06-11 23:12:07 +020019068res.payload(<offset>,<length>) : binary
19069 This extracts a binary block of <length> bytes and starting at byte <offset>
Willy Tarreau00f00842013-08-02 11:07:32 +020019070 in the response buffer. As a special case, if the <length> argument is zero,
Christopher Faulete596d182020-05-05 17:46:34 +020019071 the whole buffer from <offset> to the end is extracted. This can be used
Willy Tarreau00f00842013-08-02 11:07:32 +020019072 with ACLs in order to check for the presence of some content in a buffer at
Christopher Faulete596d182020-05-05 17:46:34 +020019073 any location. It may also be used in tcp-check based expect rules.
Willy Tarreaud63335a2010-02-26 12:56:52 +010019074
Willy Tarreau74ca5042013-06-11 23:12:07 +020019075res.payload_lv(<offset1>,<length>[,<offset2>]) : binary
19076 This extracts a binary block whose size is specified at <offset1> for <length>
19077 bytes, and which starts at <offset2> if specified or just after the length in
19078 the response buffer. The <offset2> parameter also supports relative offsets
Christopher Faulete596d182020-05-05 17:46:34 +020019079 if prepended with a '+' or '-' sign. It may also be used in tcp-check based
19080 expect rules.
Willy Tarreaud63335a2010-02-26 12:56:52 +010019081
Willy Tarreau74ca5042013-06-11 23:12:07 +020019082 Example : please consult the example from the "stick store-response" keyword.
Willy Tarreaud63335a2010-02-26 12:56:52 +010019083
Willy Tarreau971f7b62015-09-29 14:06:59 +020019084res.ssl_hello_type : integer
19085rep_ssl_hello_type : integer (deprecated)
19086 Returns an integer value containing the type of the SSL hello message found
19087 in the response buffer if the buffer contains data that parses as a complete
19088 SSL (v3 or superior) hello message. Note that this only applies to raw
19089 contents found in the response buffer and not to contents deciphered via an
19090 SSL data layer, so this will not work with "server" lines having the "ssl"
19091 option. This is mostly used in ACL to detect presence of an SSL hello message
19092 that is supposed to contain an SSL session ID usable for stickiness.
19093
Willy Tarreau74ca5042013-06-11 23:12:07 +020019094wait_end : boolean
19095 This fetch either returns true when the inspection period is over, or does
19096 not fetch. It is only used in ACLs, in conjunction with content analysis to
Davor Ocelice9ed2812017-12-25 17:49:28 +010019097 avoid returning a wrong verdict early. It may also be used to delay some
Willy Tarreau74ca5042013-06-11 23:12:07 +020019098 actions, such as a delayed reject for some special addresses. Since it either
19099 stops the rules evaluation or immediately returns true, it is recommended to
Davor Ocelice9ed2812017-12-25 17:49:28 +010019100 use this acl as the last one in a rule. Please note that the default ACL
Willy Tarreau74ca5042013-06-11 23:12:07 +020019101 "WAIT_END" is always usable without prior declaration. This test was designed
19102 to be used with TCP request content inspection.
Willy Tarreaud63335a2010-02-26 12:56:52 +010019103
Willy Tarreau74ca5042013-06-11 23:12:07 +020019104 Examples :
19105 # delay every incoming request by 2 seconds
19106 tcp-request inspect-delay 2s
19107 tcp-request content accept if WAIT_END
Willy Tarreaud63335a2010-02-26 12:56:52 +010019108
Willy Tarreau74ca5042013-06-11 23:12:07 +020019109 # don't immediately tell bad guys they are rejected
19110 tcp-request inspect-delay 10s
19111 acl goodguys src 10.0.0.0/24
19112 acl badguys src 10.0.1.0/24
19113 tcp-request content accept if goodguys
19114 tcp-request content reject if badguys WAIT_END
19115 tcp-request content reject
19116
19117
Thierry FOURNIER060762e2014-04-23 13:29:15 +0200191187.3.6. Fetching HTTP samples (Layer 7)
Willy Tarreau74ca5042013-06-11 23:12:07 +020019119--------------------------------------
19120
19121It is possible to fetch samples from HTTP contents, requests and responses.
19122This application layer is also called layer 7. It is only possible to fetch the
19123data in this section when a full HTTP request or response has been parsed from
19124its respective request or response buffer. This is always the case with all
19125HTTP specific rules and for sections running with "mode http". When using TCP
19126content inspection, it may be necessary to support an inspection delay in order
19127to let the request or response come in first. These fetches may require a bit
19128more CPU resources than the layer 4 ones, but not much since the request and
19129response are indexed.
19130
Christopher Faulet4d37e532021-03-26 14:44:00 +010019131Note : Regarding HTTP processing from the tcp-request content rules, everything
19132 will work as expected from an HTTP proxy. However, from a TCP proxy,
19133 without an HTTP upgrade, it will only work for HTTP/1 content. For
19134 HTTP/2 content, only the preface is visible. Thus, it is only possible
19135 to rely to "req.proto_http", "req.ver" and eventually "method" sample
19136 fetches. All other L7 sample fetches will fail. After an HTTP upgrade,
19137 they will work in the same manner than from an HTTP proxy.
19138
Willy Tarreau74ca5042013-06-11 23:12:07 +020019139base : string
19140 This returns the concatenation of the first Host header and the path part of
19141 the request, which starts at the first slash and ends before the question
19142 mark. It can be useful in virtual hosted environments to detect URL abuses as
19143 well as to improve shared caches efficiency. Using this with a limited size
19144 stick table also allows one to collect statistics about most commonly
19145 requested objects by host/path. With ACLs it can allow simple content
19146 switching rules involving the host and the path at the same time, such as
19147 "www.example.com/favicon.ico". See also "path" and "uri".
19148
19149 ACL derivatives :
19150 base : exact string match
19151 base_beg : prefix match
19152 base_dir : subdir match
19153 base_dom : domain match
19154 base_end : suffix match
19155 base_len : length match
19156 base_reg : regex match
19157 base_sub : substring match
19158
19159base32 : integer
19160 This returns a 32-bit hash of the value returned by the "base" fetch method
19161 above. This is useful to track per-URL activity on high traffic sites without
19162 having to store all URLs. Instead a shorter hash is stored, saving a lot of
Willy Tarreau23ec4ca2014-07-15 20:15:37 +020019163 memory. The output type is an unsigned integer. The hash function used is
19164 SDBM with full avalanche on the output. Technically, base32 is exactly equal
19165 to "base,sdbm(1)".
Willy Tarreau74ca5042013-06-11 23:12:07 +020019166
19167base32+src : binary
19168 This returns the concatenation of the base32 fetch above and the src fetch
19169 below. The resulting type is of type binary, with a size of 8 or 20 bytes
19170 depending on the source address family. This can be used to track per-IP,
19171 per-URL counters.
19172
Yves Lafonb4d37082021-02-11 11:01:28 +010019173baseq : string
19174 This returns the concatenation of the first Host header and the path part of
19175 the request with the query-string, which starts at the first slash. Using this
19176 instead of "base" allows one to properly identify the target resource, for
19177 statistics or caching use cases. See also "path", "pathq" and "base".
19178
William Lallemand65ad6e12014-01-31 15:08:02 +010019179capture.req.hdr(<idx>) : string
19180 This extracts the content of the header captured by the "capture request
19181 header", idx is the position of the capture keyword in the configuration.
19182 The first entry is an index of 0. See also: "capture request header".
19183
19184capture.req.method : string
19185 This extracts the METHOD of an HTTP request. It can be used in both request
19186 and response. Unlike "method", it can be used in both request and response
19187 because it's allocated.
19188
19189capture.req.uri : string
19190 This extracts the request's URI, which starts at the first slash and ends
19191 before the first space in the request (without the host part). Unlike "path"
19192 and "url", it can be used in both request and response because it's
19193 allocated.
19194
Willy Tarreau3c1b5ec2014-04-24 23:41:57 +020019195capture.req.ver : string
19196 This extracts the request's HTTP version and returns either "HTTP/1.0" or
19197 "HTTP/1.1". Unlike "req.ver", it can be used in both request, response, and
19198 logs because it relies on a persistent flag.
19199
William Lallemand65ad6e12014-01-31 15:08:02 +010019200capture.res.hdr(<idx>) : string
19201 This extracts the content of the header captured by the "capture response
19202 header", idx is the position of the capture keyword in the configuration.
19203 The first entry is an index of 0.
19204 See also: "capture response header"
19205
Willy Tarreau3c1b5ec2014-04-24 23:41:57 +020019206capture.res.ver : string
19207 This extracts the response's HTTP version and returns either "HTTP/1.0" or
19208 "HTTP/1.1". Unlike "res.ver", it can be used in logs because it relies on a
19209 persistent flag.
19210
Willy Tarreaua5910cc2015-05-02 00:46:08 +020019211req.body : binary
Christopher Fauletaf4dc4c2020-05-05 17:33:25 +020019212 This returns the HTTP request's available body as a block of data. It is
19213 recommended to use "option http-buffer-request" to be sure to wait, as much
19214 as possible, for the request's body.
Willy Tarreaua5910cc2015-05-02 00:46:08 +020019215
Thierry FOURNIER9826c772015-05-20 15:50:54 +020019216req.body_param([<name>) : string
19217 This fetch assumes that the body of the POST request is url-encoded. The user
19218 can check if the "content-type" contains the value
19219 "application/x-www-form-urlencoded". This extracts the first occurrence of the
19220 parameter <name> in the body, which ends before '&'. The parameter name is
19221 case-sensitive. If no name is given, any parameter will match, and the first
19222 one will be returned. The result is a string corresponding to the value of the
19223 parameter <name> as presented in the request body (no URL decoding is
19224 performed). Note that the ACL version of this fetch iterates over multiple
19225 parameters and will iteratively report all parameters values if no name is
19226 given.
19227
Willy Tarreaua5910cc2015-05-02 00:46:08 +020019228req.body_len : integer
19229 This returns the length of the HTTP request's available body in bytes. It may
19230 be lower than the advertised length if the body is larger than the buffer. It
Christopher Fauletaf4dc4c2020-05-05 17:33:25 +020019231 is recommended to use "option http-buffer-request" to be sure to wait, as
19232 much as possible, for the request's body.
Willy Tarreaua5910cc2015-05-02 00:46:08 +020019233
19234req.body_size : integer
19235 This returns the advertised length of the HTTP request's body in bytes. It
Christopher Fauletaf4dc4c2020-05-05 17:33:25 +020019236 will represent the advertised Content-Length header, or the size of the
19237 available data in case of chunked encoding.
Willy Tarreaua5910cc2015-05-02 00:46:08 +020019238
Willy Tarreau74ca5042013-06-11 23:12:07 +020019239req.cook([<name>]) : string
19240cook([<name>]) : string (deprecated)
19241 This extracts the last occurrence of the cookie name <name> on a "Cookie"
19242 header line from the request, and returns its value as string. If no name is
19243 specified, the first cookie value is returned. When used with ACLs, all
19244 matching cookies are evaluated. Spaces around the name and the value are
19245 ignored as requested by the Cookie header specification (RFC6265). The cookie
19246 name is case-sensitive. Empty cookies are valid, so an empty cookie may very
19247 well return an empty value if it is present. Use the "found" match to detect
19248 presence. Use the res.cook() variant for response cookies sent by the server.
19249
19250 ACL derivatives :
19251 cook([<name>]) : exact string match
19252 cook_beg([<name>]) : prefix match
19253 cook_dir([<name>]) : subdir match
19254 cook_dom([<name>]) : domain match
19255 cook_end([<name>]) : suffix match
19256 cook_len([<name>]) : length match
19257 cook_reg([<name>]) : regex match
19258 cook_sub([<name>]) : substring match
Willy Tarreaud63335a2010-02-26 12:56:52 +010019259
Willy Tarreau74ca5042013-06-11 23:12:07 +020019260req.cook_cnt([<name>]) : integer
19261cook_cnt([<name>]) : integer (deprecated)
19262 Returns an integer value representing the number of occurrences of the cookie
19263 <name> in the request, or all cookies if <name> is not specified.
Willy Tarreaud63335a2010-02-26 12:56:52 +010019264
Willy Tarreau74ca5042013-06-11 23:12:07 +020019265req.cook_val([<name>]) : integer
19266cook_val([<name>]) : integer (deprecated)
19267 This extracts the last occurrence of the cookie name <name> on a "Cookie"
19268 header line from the request, and converts its value to an integer which is
19269 returned. If no name is specified, the first cookie value is returned. When
19270 used in ACLs, all matching names are iterated over until a value matches.
Willy Tarreau0e698542011-09-16 08:32:32 +020019271
Willy Tarreau74ca5042013-06-11 23:12:07 +020019272cookie([<name>]) : string (deprecated)
19273 This extracts the last occurrence of the cookie name <name> on a "Cookie"
19274 header line from the request, or a "Set-Cookie" header from the response, and
19275 returns its value as a string. A typical use is to get multiple clients
19276 sharing a same profile use the same server. This can be similar to what
Willy Tarreau294d0f02015-08-10 19:40:12 +020019277 "appsession" did with the "request-learn" statement, but with support for
Willy Tarreau74ca5042013-06-11 23:12:07 +020019278 multi-peer synchronization and state keeping across restarts. If no name is
19279 specified, the first cookie value is returned. This fetch should not be used
19280 anymore and should be replaced by req.cook() or res.cook() instead as it
19281 ambiguously uses the direction based on the context where it is used.
Willy Tarreaud63335a2010-02-26 12:56:52 +010019282
Willy Tarreau74ca5042013-06-11 23:12:07 +020019283hdr([<name>[,<occ>]]) : string
19284 This is equivalent to req.hdr() when used on requests, and to res.hdr() when
19285 used on responses. Please refer to these respective fetches for more details.
19286 In case of doubt about the fetch direction, please use the explicit ones.
19287 Note that contrary to the hdr() sample fetch method, the hdr_* ACL keywords
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030019288 unambiguously apply to the request headers.
Willy Tarreaud63335a2010-02-26 12:56:52 +010019289
Willy Tarreau74ca5042013-06-11 23:12:07 +020019290req.fhdr(<name>[,<occ>]) : string
Tim Duesterhus27c70ae2021-01-23 17:50:21 +010019291 This returns the full value of the last occurrence of header <name> in an
19292 HTTP request. It differs from req.hdr() in that any commas present in the
19293 value are returned and are not used as delimiters. This is sometimes useful
19294 with headers such as User-Agent.
19295
19296 When used from an ACL, all occurrences are iterated over until a match is
19297 found.
19298
Willy Tarreau74ca5042013-06-11 23:12:07 +020019299 Optionally, a specific occurrence might be specified as a position number.
19300 Positive values indicate a position from the first occurrence, with 1 being
19301 the first one. Negative values indicate positions relative to the last one,
Tim Duesterhus27c70ae2021-01-23 17:50:21 +010019302 with -1 being the last one.
Willy Tarreaud63335a2010-02-26 12:56:52 +010019303
Willy Tarreau74ca5042013-06-11 23:12:07 +020019304req.fhdr_cnt([<name>]) : integer
19305 Returns an integer value representing the number of occurrences of request
19306 header field name <name>, or the total number of header fields if <name> is
Tim Duesterhus27c70ae2021-01-23 17:50:21 +010019307 not specified. Like req.fhdr() it differs from res.hdr_cnt() by not splitting
19308 headers at commas.
Willy Tarreaud63335a2010-02-26 12:56:52 +010019309
Willy Tarreau74ca5042013-06-11 23:12:07 +020019310req.hdr([<name>[,<occ>]]) : string
Tim Duesterhus27c70ae2021-01-23 17:50:21 +010019311 This returns the last comma-separated value of the header <name> in an HTTP
19312 request. The fetch considers any comma as a delimiter for distinct values.
19313 This is useful if you need to process headers that are defined to be a list
19314 of values, such as Accept, or X-Forwarded-For. If full-line headers are
19315 desired instead, use req.fhdr(). Please carefully check RFC 7231 to know how
19316 certain headers are supposed to be parsed. Also, some of them are case
19317 insensitive (e.g. Connection).
19318
19319 When used from an ACL, all occurrences are iterated over until a match is
19320 found.
19321
Willy Tarreau74ca5042013-06-11 23:12:07 +020019322 Optionally, a specific occurrence might be specified as a position number.
19323 Positive values indicate a position from the first occurrence, with 1 being
19324 the first one. Negative values indicate positions relative to the last one,
Tim Duesterhus27c70ae2021-01-23 17:50:21 +010019325 with -1 being the last one.
19326
19327 A typical use is with the X-Forwarded-For header once converted to IP,
19328 associated with an IP stick-table.
Willy Tarreaud63335a2010-02-26 12:56:52 +010019329
Willy Tarreau74ca5042013-06-11 23:12:07 +020019330 ACL derivatives :
19331 hdr([<name>[,<occ>]]) : exact string match
19332 hdr_beg([<name>[,<occ>]]) : prefix match
19333 hdr_dir([<name>[,<occ>]]) : subdir match
19334 hdr_dom([<name>[,<occ>]]) : domain match
19335 hdr_end([<name>[,<occ>]]) : suffix match
19336 hdr_len([<name>[,<occ>]]) : length match
19337 hdr_reg([<name>[,<occ>]]) : regex match
19338 hdr_sub([<name>[,<occ>]]) : substring match
19339
19340req.hdr_cnt([<name>]) : integer
19341hdr_cnt([<header>]) : integer (deprecated)
19342 Returns an integer value representing the number of occurrences of request
19343 header field name <name>, or the total number of header field values if
Tim Duesterhus27c70ae2021-01-23 17:50:21 +010019344 <name> is not specified. Like req.hdr() it counts each comma separated
19345 part of the header's value. If counting of full-line headers is desired,
19346 then req.fhdr_cnt() should be used instead.
19347
19348 With ACLs, it can be used to detect presence, absence or abuse of a specific
19349 header, as well as to block request smuggling attacks by rejecting requests
19350 which contain more than one of certain headers.
19351
19352 Refer to req.hdr() for more information on header matching.
Willy Tarreau74ca5042013-06-11 23:12:07 +020019353
19354req.hdr_ip([<name>[,<occ>]]) : ip
19355hdr_ip([<name>[,<occ>]]) : ip (deprecated)
19356 This extracts the last occurrence of header <name> in an HTTP request,
19357 converts it to an IPv4 or IPv6 address and returns this address. When used
19358 with ACLs, all occurrences are checked, and if <name> is omitted, every value
Willy Tarreau7b0e00d2021-03-25 14:12:29 +010019359 of every header is checked. The parser strictly adheres to the format
19360 described in RFC7239, with the extension that IPv4 addresses may optionally
19361 be followed by a colon (':') and a valid decimal port number (0 to 65535),
19362 which will be silently dropped. All other forms will not match and will
19363 cause the address to be ignored.
Tim Duesterhus27c70ae2021-01-23 17:50:21 +010019364
19365 The <occ> parameter is processed as with req.hdr().
19366
19367 A typical use is with the X-Forwarded-For and X-Client-IP headers.
Willy Tarreau74ca5042013-06-11 23:12:07 +020019368
19369req.hdr_val([<name>[,<occ>]]) : integer
19370hdr_val([<name>[,<occ>]]) : integer (deprecated)
19371 This extracts the last occurrence of header <name> in an HTTP request, and
19372 converts it to an integer value. When used with ACLs, all occurrences are
19373 checked, and if <name> is omitted, every value of every header is checked.
Tim Duesterhus27c70ae2021-01-23 17:50:21 +010019374
19375 The <occ> parameter is processed as with req.hdr().
19376
19377 A typical use is with the X-Forwarded-For header.
Willy Tarreau74ca5042013-06-11 23:12:07 +020019378
Christopher Faulet687a68e2020-11-24 17:13:24 +010019379req.hdrs : string
19380 Returns the current request headers as string including the last empty line
19381 separating headers from the request body. The last empty line can be used to
19382 detect a truncated header block. This sample fetch is useful for some SPOE
19383 headers analyzers and for advanced logging.
19384
19385req.hdrs_bin : binary
19386 Returns the current request headers contained in preparsed binary form. This
19387 is useful for offloading some processing with SPOE. Each string is described
19388 by a length followed by the number of bytes indicated in the length. The
19389 length is represented using the variable integer encoding detailed in the
19390 SPOE documentation. The end of the list is marked by a couple of empty header
19391 names and values (length of 0 for both).
19392
19393 *(<str:header-name><str:header-value>)<empty string><empty string>
Frédéric Lécailleec891192019-02-26 15:02:35 +010019394
Christopher Faulet687a68e2020-11-24 17:13:24 +010019395 int: refer to the SPOE documentation for the encoding
19396 str: <int:length><bytes>
Frédéric Lécailleec891192019-02-26 15:02:35 +010019397
Willy Tarreau74ca5042013-06-11 23:12:07 +020019398http_auth(<userlist>) : boolean
19399 Returns a boolean indicating whether the authentication data received from
19400 the client match a username & password stored in the specified userlist. This
19401 fetch function is not really useful outside of ACLs. Currently only http
19402 basic auth is supported.
19403
Thierry FOURNIER9eec0a62014-01-22 18:38:02 +010019404http_auth_group(<userlist>) : string
19405 Returns a string corresponding to the user name found in the authentication
19406 data received from the client if both the user name and password are valid
19407 according to the specified userlist. The main purpose is to use it in ACLs
19408 where it is then checked whether the user belongs to any group within a list.
Willy Tarreau74ca5042013-06-11 23:12:07 +020019409 This fetch function is not really useful outside of ACLs. Currently only http
19410 basic auth is supported.
19411
19412 ACL derivatives :
Thierry FOURNIER9eec0a62014-01-22 18:38:02 +010019413 http_auth_group(<userlist>) : group ...
19414 Returns true when the user extracted from the request and whose password is
19415 valid according to the specified userlist belongs to at least one of the
19416 groups.
Willy Tarreau74ca5042013-06-11 23:12:07 +020019417
Christopher Fauleta4063562019-08-02 11:51:37 +020019418http_auth_pass : string
Willy Tarreauc9c6cdb2020-03-05 16:03:58 +010019419 Returns the user's password found in the authentication data received from
19420 the client, as supplied in the Authorization header. Not checks are
19421 performed by this sample fetch. Only Basic authentication is supported.
Christopher Fauleta4063562019-08-02 11:51:37 +020019422
19423http_auth_type : string
Willy Tarreauc9c6cdb2020-03-05 16:03:58 +010019424 Returns the authentication method found in the authentication data received from
19425 the client, as supplied in the Authorization header. Not checks are
19426 performed by this sample fetch. Only Basic authentication is supported.
Christopher Fauleta4063562019-08-02 11:51:37 +020019427
19428http_auth_user : string
Willy Tarreauc9c6cdb2020-03-05 16:03:58 +010019429 Returns the user name found in the authentication data received from the
19430 client, as supplied in the Authorization header. Not checks are performed by
19431 this sample fetch. Only Basic authentication is supported.
Christopher Fauleta4063562019-08-02 11:51:37 +020019432
Willy Tarreau74ca5042013-06-11 23:12:07 +020019433http_first_req : boolean
Willy Tarreau7f18e522010-10-22 20:04:13 +020019434 Returns true when the request being processed is the first one of the
19435 connection. This can be used to add or remove headers that may be missing
Willy Tarreau74ca5042013-06-11 23:12:07 +020019436 from some requests when a request is not the first one, or to help grouping
19437 requests in the logs.
Willy Tarreau7f18e522010-10-22 20:04:13 +020019438
Willy Tarreau74ca5042013-06-11 23:12:07 +020019439method : integer + string
19440 Returns an integer value corresponding to the method in the HTTP request. For
19441 example, "GET" equals 1 (check sources to establish the matching). Value 9
19442 means "other method" and may be converted to a string extracted from the
19443 stream. This should not be used directly as a sample, this is only meant to
19444 be used from ACLs, which transparently convert methods from patterns to these
19445 integer + string values. Some predefined ACL already check for most common
19446 methods.
Willy Tarreau6a06a402007-07-15 20:15:28 +020019447
Willy Tarreau74ca5042013-06-11 23:12:07 +020019448 ACL derivatives :
19449 method : case insensitive method match
Willy Tarreau6a06a402007-07-15 20:15:28 +020019450
Willy Tarreau74ca5042013-06-11 23:12:07 +020019451 Example :
19452 # only accept GET and HEAD requests
19453 acl valid_method method GET HEAD
19454 http-request deny if ! valid_method
Willy Tarreau6a06a402007-07-15 20:15:28 +020019455
Willy Tarreau74ca5042013-06-11 23:12:07 +020019456path : string
19457 This extracts the request's URL path, which starts at the first slash and
19458 ends before the question mark (without the host part). A typical use is with
19459 prefetch-capable caches, and with portals which need to aggregate multiple
19460 information from databases and keep them in caches. Note that with outgoing
19461 caches, it would be wiser to use "url" instead. With ACLs, it's typically
Davor Ocelice9ed2812017-12-25 17:49:28 +010019462 used to match exact file names (e.g. "/login.php"), or directory parts using
Willy Tarreau74ca5042013-06-11 23:12:07 +020019463 the derivative forms. See also the "url" and "base" fetch methods.
Willy Tarreau6a06a402007-07-15 20:15:28 +020019464
Willy Tarreau74ca5042013-06-11 23:12:07 +020019465 ACL derivatives :
19466 path : exact string match
19467 path_beg : prefix match
19468 path_dir : subdir match
19469 path_dom : domain match
19470 path_end : suffix match
19471 path_len : length match
19472 path_reg : regex match
19473 path_sub : substring match
Willy Tarreau6a06a402007-07-15 20:15:28 +020019474
Christopher Faulete720c322020-09-02 17:25:18 +020019475pathq : string
19476 This extracts the request's URL path with the query-string, which starts at
19477 the first slash. This sample fetch is pretty handy to always retrieve a
19478 relative URI, excluding the scheme and the authority part, if any. Indeed,
19479 while it is the common representation for an HTTP/1.1 request target, in
19480 HTTP/2, an absolute URI is often used. This sample fetch will return the same
19481 result in both cases.
19482
Willy Tarreau49ad95c2015-01-19 15:06:26 +010019483query : string
19484 This extracts the request's query string, which starts after the first
19485 question mark. If no question mark is present, this fetch returns nothing. If
19486 a question mark is present but nothing follows, it returns an empty string.
19487 This means it's possible to easily know whether a query string is present
Tim Düsterhus4896c442016-11-29 02:15:19 +010019488 using the "found" matching method. This fetch is the complement of "path"
Willy Tarreau49ad95c2015-01-19 15:06:26 +010019489 which stops before the question mark.
19490
Willy Tarreaueb27ec72015-02-20 13:55:29 +010019491req.hdr_names([<delim>]) : string
19492 This builds a string made from the concatenation of all header names as they
19493 appear in the request when the rule is evaluated. The default delimiter is
19494 the comma (',') but it may be overridden as an optional argument <delim>. In
19495 this case, only the first character of <delim> is considered.
19496
Willy Tarreau74ca5042013-06-11 23:12:07 +020019497req.ver : string
19498req_ver : string (deprecated)
19499 Returns the version string from the HTTP request, for example "1.1". This can
19500 be useful for logs, but is mostly there for ACL. Some predefined ACL already
19501 check for versions 1.0 and 1.1.
Willy Tarreaud63335a2010-02-26 12:56:52 +010019502
Willy Tarreau74ca5042013-06-11 23:12:07 +020019503 ACL derivatives :
19504 req_ver : exact string match
Willy Tarreau0e698542011-09-16 08:32:32 +020019505
Christopher Faulete596d182020-05-05 17:46:34 +020019506res.body : binary
19507 This returns the HTTP response's available body as a block of data. Unlike
19508 the request side, there is no directive to wait for the response's body. This
Tim Duesterhus27c70ae2021-01-23 17:50:21 +010019509 sample fetch is really useful (and usable) in the health-check context.
19510
19511 It may be used in tcp-check based expect rules.
Christopher Faulete596d182020-05-05 17:46:34 +020019512
19513res.body_len : integer
19514 This returns the length of the HTTP response available body in bytes. Unlike
19515 the request side, there is no directive to wait for the response's body. This
Tim Duesterhus27c70ae2021-01-23 17:50:21 +010019516 sample fetch is really useful (and usable) in the health-check context.
19517
19518 It may be used in tcp-check based expect rules.
Christopher Faulete596d182020-05-05 17:46:34 +020019519
19520res.body_size : integer
19521 This returns the advertised length of the HTTP response body in bytes. It
19522 will represent the advertised Content-Length header, or the size of the
19523 available data in case of chunked encoding. Unlike the request side, there is
19524 no directive to wait for the response body. This sample fetch is really
Tim Duesterhus27c70ae2021-01-23 17:50:21 +010019525 useful (and usable) in the health-check context.
19526
19527 It may be used in tcp-check based expect rules.
Christopher Faulete596d182020-05-05 17:46:34 +020019528
Remi Tricot-Le Bretonbf971212020-10-27 11:55:57 +010019529res.cache_hit : boolean
19530 Returns the boolean "true" value if the response has been built out of an
19531 HTTP cache entry, otherwise returns boolean "false".
19532
19533res.cache_name : string
19534 Returns a string containing the name of the HTTP cache that was used to
19535 build the HTTP response if res.cache_hit is true, otherwise returns an
19536 empty string.
19537
Willy Tarreau74ca5042013-06-11 23:12:07 +020019538res.comp : boolean
19539 Returns the boolean "true" value if the response has been compressed by
19540 HAProxy, otherwise returns boolean "false". This may be used to add
19541 information in the logs.
Willy Tarreau6a06a402007-07-15 20:15:28 +020019542
Willy Tarreau74ca5042013-06-11 23:12:07 +020019543res.comp_algo : string
19544 Returns a string containing the name of the algorithm used if the response
19545 was compressed by HAProxy, for example : "deflate". This may be used to add
19546 some information in the logs.
Willy Tarreaud63335a2010-02-26 12:56:52 +010019547
Willy Tarreau74ca5042013-06-11 23:12:07 +020019548res.cook([<name>]) : string
19549scook([<name>]) : string (deprecated)
19550 This extracts the last occurrence of the cookie name <name> on a "Set-Cookie"
19551 header line from the response, and returns its value as string. If no name is
Tim Duesterhus27c70ae2021-01-23 17:50:21 +010019552 specified, the first cookie value is returned.
19553
19554 It may be used in tcp-check based expect rules.
Willy Tarreau0ce3aa02012-04-25 18:46:33 +020019555
Willy Tarreau74ca5042013-06-11 23:12:07 +020019556 ACL derivatives :
19557 scook([<name>] : exact string match
Willy Tarreau0ce3aa02012-04-25 18:46:33 +020019558
Willy Tarreau74ca5042013-06-11 23:12:07 +020019559res.cook_cnt([<name>]) : integer
19560scook_cnt([<name>]) : integer (deprecated)
19561 Returns an integer value representing the number of occurrences of the cookie
19562 <name> in the response, or all cookies if <name> is not specified. This is
Tim Duesterhus27c70ae2021-01-23 17:50:21 +010019563 mostly useful when combined with ACLs to detect suspicious responses.
19564
19565 It may be used in tcp-check based expect rules.
Willy Tarreaud63335a2010-02-26 12:56:52 +010019566
Willy Tarreau74ca5042013-06-11 23:12:07 +020019567res.cook_val([<name>]) : integer
19568scook_val([<name>]) : integer (deprecated)
19569 This extracts the last occurrence of the cookie name <name> on a "Set-Cookie"
19570 header line from the response, and converts its value to an integer which is
Tim Duesterhus27c70ae2021-01-23 17:50:21 +010019571 returned. If no name is specified, the first cookie value is returned.
19572
19573 It may be used in tcp-check based expect rules.
Willy Tarreaud63335a2010-02-26 12:56:52 +010019574
Willy Tarreau74ca5042013-06-11 23:12:07 +020019575res.fhdr([<name>[,<occ>]]) : string
Tim Duesterhus27c70ae2021-01-23 17:50:21 +010019576 This fetch works like the req.fhdr() fetch with the difference that it acts
19577 on the headers within an HTTP response.
19578
19579 Like req.fhdr() the res.fhdr() fetch returns full values. If the header is
19580 defined to be a list you should use res.hdr().
19581
19582 This fetch is sometimes useful with headers such as Date or Expires.
19583
19584 It may be used in tcp-check based expect rules.
Willy Tarreau6a06a402007-07-15 20:15:28 +020019585
Willy Tarreau74ca5042013-06-11 23:12:07 +020019586res.fhdr_cnt([<name>]) : integer
Tim Duesterhus27c70ae2021-01-23 17:50:21 +010019587 This fetch works like the req.fhdr_cnt() fetch with the difference that it
19588 acts on the headers within an HTTP response.
19589
19590 Like req.fhdr_cnt() the res.fhdr_cnt() fetch acts on full values. If the
19591 header is defined to be a list you should use res.hdr_cnt().
19592
19593 It may be used in tcp-check based expect rules.
Willy Tarreau6a06a402007-07-15 20:15:28 +020019594
Willy Tarreau74ca5042013-06-11 23:12:07 +020019595res.hdr([<name>[,<occ>]]) : string
19596shdr([<name>[,<occ>]]) : string (deprecated)
Tim Duesterhus27c70ae2021-01-23 17:50:21 +010019597 This fetch works like the req.hdr() fetch with the difference that it acts
19598 on the headers within an HTTP response.
19599
Ilya Shipitsinacf84592021-02-06 22:29:08 +050019600 Like req.hdr() the res.hdr() fetch considers the comma to be a delimiter. If
Tim Duesterhus27c70ae2021-01-23 17:50:21 +010019601 this is not desired res.fhdr() should be used.
19602
19603 It may be used in tcp-check based expect rules.
Willy Tarreau6a06a402007-07-15 20:15:28 +020019604
Willy Tarreau74ca5042013-06-11 23:12:07 +020019605 ACL derivatives :
19606 shdr([<name>[,<occ>]]) : exact string match
19607 shdr_beg([<name>[,<occ>]]) : prefix match
19608 shdr_dir([<name>[,<occ>]]) : subdir match
19609 shdr_dom([<name>[,<occ>]]) : domain match
19610 shdr_end([<name>[,<occ>]]) : suffix match
19611 shdr_len([<name>[,<occ>]]) : length match
19612 shdr_reg([<name>[,<occ>]]) : regex match
19613 shdr_sub([<name>[,<occ>]]) : substring match
19614
19615res.hdr_cnt([<name>]) : integer
19616shdr_cnt([<name>]) : integer (deprecated)
Tim Duesterhus27c70ae2021-01-23 17:50:21 +010019617 This fetch works like the req.hdr_cnt() fetch with the difference that it
19618 acts on the headers within an HTTP response.
19619
19620 Like req.hdr_cnt() the res.hdr_cnt() fetch considers the comma to be a
Ilya Shipitsinacf84592021-02-06 22:29:08 +050019621 delimiter. If this is not desired res.fhdr_cnt() should be used.
Tim Duesterhus27c70ae2021-01-23 17:50:21 +010019622
19623 It may be used in tcp-check based expect rules.
Willy Tarreau6a06a402007-07-15 20:15:28 +020019624
Willy Tarreau74ca5042013-06-11 23:12:07 +020019625res.hdr_ip([<name>[,<occ>]]) : ip
19626shdr_ip([<name>[,<occ>]]) : ip (deprecated)
Tim Duesterhus27c70ae2021-01-23 17:50:21 +010019627 This fetch works like the req.hdr_ip() fetch with the difference that it
19628 acts on the headers within an HTTP response.
19629
19630 This can be useful to learn some data into a stick table.
19631
19632 It may be used in tcp-check based expect rules.
Willy Tarreau6a06a402007-07-15 20:15:28 +020019633
Willy Tarreaueb27ec72015-02-20 13:55:29 +010019634res.hdr_names([<delim>]) : string
19635 This builds a string made from the concatenation of all header names as they
19636 appear in the response when the rule is evaluated. The default delimiter is
19637 the comma (',') but it may be overridden as an optional argument <delim>. In
Tim Duesterhus27c70ae2021-01-23 17:50:21 +010019638 this case, only the first character of <delim> is considered.
19639
19640 It may be used in tcp-check based expect rules.
Willy Tarreaueb27ec72015-02-20 13:55:29 +010019641
Willy Tarreau74ca5042013-06-11 23:12:07 +020019642res.hdr_val([<name>[,<occ>]]) : integer
19643shdr_val([<name>[,<occ>]]) : integer (deprecated)
Tim Duesterhus27c70ae2021-01-23 17:50:21 +010019644 This fetch works like the req.hdr_val() fetch with the difference that it
19645 acts on the headers within an HTTP response.
19646
19647 This can be useful to learn some data into a stick table.
19648
19649 It may be used in tcp-check based expect rules.
Christopher Faulete596d182020-05-05 17:46:34 +020019650
19651res.hdrs : string
19652 Returns the current response headers as string including the last empty line
19653 separating headers from the request body. The last empty line can be used to
19654 detect a truncated header block. This sample fetch is useful for some SPOE
Tim Duesterhus27c70ae2021-01-23 17:50:21 +010019655 headers analyzers and for advanced logging.
19656
19657 It may also be used in tcp-check based expect rules.
Christopher Faulete596d182020-05-05 17:46:34 +020019658
19659res.hdrs_bin : binary
19660 Returns the current response headers contained in preparsed binary form. This
19661 is useful for offloading some processing with SPOE. It may be used in
19662 tcp-check based expect rules. Each string is described by a length followed
19663 by the number of bytes indicated in the length. The length is represented
19664 using the variable integer encoding detailed in the SPOE documentation. The
19665 end of the list is marked by a couple of empty header names and values
19666 (length of 0 for both).
19667
19668 *(<str:header-name><str:header-value>)<empty string><empty string>
19669
19670 int: refer to the SPOE documentation for the encoding
19671 str: <int:length><bytes>
Alexandre Cassen5eb1a902007-11-29 15:43:32 +010019672
Willy Tarreau74ca5042013-06-11 23:12:07 +020019673res.ver : string
19674resp_ver : string (deprecated)
19675 Returns the version string from the HTTP response, for example "1.1". This
Tim Duesterhus27c70ae2021-01-23 17:50:21 +010019676 can be useful for logs, but is mostly there for ACL.
19677
19678 It may be used in tcp-check based expect rules.
Willy Tarreau0e698542011-09-16 08:32:32 +020019679
Willy Tarreau74ca5042013-06-11 23:12:07 +020019680 ACL derivatives :
19681 resp_ver : exact string match
Alexandre Cassen5eb1a902007-11-29 15:43:32 +010019682
Willy Tarreau74ca5042013-06-11 23:12:07 +020019683set-cookie([<name>]) : string (deprecated)
19684 This extracts the last occurrence of the cookie name <name> on a "Set-Cookie"
19685 header line from the response and uses the corresponding value to match. This
Willy Tarreau294d0f02015-08-10 19:40:12 +020019686 can be comparable to what "appsession" did with default options, but with
Willy Tarreau74ca5042013-06-11 23:12:07 +020019687 support for multi-peer synchronization and state keeping across restarts.
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +010019688
Willy Tarreau74ca5042013-06-11 23:12:07 +020019689 This fetch function is deprecated and has been superseded by the "res.cook"
19690 fetch. This keyword will disappear soon.
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +010019691
Willy Tarreau74ca5042013-06-11 23:12:07 +020019692status : integer
19693 Returns an integer containing the HTTP status code in the HTTP response, for
19694 example, 302. It is mostly used within ACLs and integer ranges, for example,
Tim Duesterhus27c70ae2021-01-23 17:50:21 +010019695 to remove any Location header if the response is not a 3xx.
19696
19697 It may be used in tcp-check based expect rules.
Willy Tarreau25c1ebc2012-04-25 16:21:44 +020019698
Thierry Fournier0e00dca2016-04-07 15:47:40 +020019699unique-id : string
19700 Returns the unique-id attached to the request. The directive
19701 "unique-id-format" must be set. If it is not set, the unique-id sample fetch
19702 fails. Note that the unique-id is usually used with HTTP requests, however this
19703 sample fetch can be used with other protocols. Obviously, if it is used with
19704 other protocols than HTTP, the unique-id-format directive must not contain
19705 HTTP parts. See: unique-id-format and unique-id-header
19706
Willy Tarreau74ca5042013-06-11 23:12:07 +020019707url : string
19708 This extracts the request's URL as presented in the request. A typical use is
19709 with prefetch-capable caches, and with portals which need to aggregate
19710 multiple information from databases and keep them in caches. With ACLs, using
19711 "path" is preferred over using "url", because clients may send a full URL as
19712 is normally done with proxies. The only real use is to match "*" which does
19713 not match in "path", and for which there is already a predefined ACL. See
19714 also "path" and "base".
Willy Tarreau25c1ebc2012-04-25 16:21:44 +020019715
Willy Tarreau74ca5042013-06-11 23:12:07 +020019716 ACL derivatives :
19717 url : exact string match
19718 url_beg : prefix match
19719 url_dir : subdir match
19720 url_dom : domain match
19721 url_end : suffix match
19722 url_len : length match
19723 url_reg : regex match
19724 url_sub : substring match
Willy Tarreau25c1ebc2012-04-25 16:21:44 +020019725
Willy Tarreau74ca5042013-06-11 23:12:07 +020019726url_ip : ip
19727 This extracts the IP address from the request's URL when the host part is
19728 presented as an IP address. Its use is very limited. For instance, a
19729 monitoring system might use this field as an alternative for the source IP in
19730 order to test what path a given source address would follow, or to force an
19731 entry in a table for a given source address. With ACLs it can be used to
19732 restrict access to certain systems through a proxy, for example when combined
19733 with option "http_proxy".
Willy Tarreau25c1ebc2012-04-25 16:21:44 +020019734
Willy Tarreau74ca5042013-06-11 23:12:07 +020019735url_port : integer
19736 This extracts the port part from the request's URL. Note that if the port is
19737 not specified in the request, port 80 is assumed. With ACLs it can be used to
19738 restrict access to certain systems through a proxy, for example when combined
19739 with option "http_proxy".
Willy Tarreau25c1ebc2012-04-25 16:21:44 +020019740
Willy Tarreau1ede1da2015-05-07 16:06:18 +020019741urlp([<name>[,<delim>]]) : string
19742url_param([<name>[,<delim>]]) : string
Willy Tarreau74ca5042013-06-11 23:12:07 +020019743 This extracts the first occurrence of the parameter <name> in the query
19744 string, which begins after either '?' or <delim>, and which ends before '&',
Willy Tarreau1ede1da2015-05-07 16:06:18 +020019745 ';' or <delim>. The parameter name is case-sensitive. If no name is given,
19746 any parameter will match, and the first one will be returned. The result is
19747 a string corresponding to the value of the parameter <name> as presented in
19748 the request (no URL decoding is performed). This can be used for session
Willy Tarreau74ca5042013-06-11 23:12:07 +020019749 stickiness based on a client ID, to extract an application cookie passed as a
19750 URL parameter, or in ACLs to apply some checks. Note that the ACL version of
Willy Tarreau1ede1da2015-05-07 16:06:18 +020019751 this fetch iterates over multiple parameters and will iteratively report all
19752 parameters values if no name is given
Willy Tarreau25c1ebc2012-04-25 16:21:44 +020019753
Willy Tarreau74ca5042013-06-11 23:12:07 +020019754 ACL derivatives :
19755 urlp(<name>[,<delim>]) : exact string match
19756 urlp_beg(<name>[,<delim>]) : prefix match
19757 urlp_dir(<name>[,<delim>]) : subdir match
19758 urlp_dom(<name>[,<delim>]) : domain match
19759 urlp_end(<name>[,<delim>]) : suffix match
19760 urlp_len(<name>[,<delim>]) : length match
19761 urlp_reg(<name>[,<delim>]) : regex match
19762 urlp_sub(<name>[,<delim>]) : substring match
Willy Tarreau25c1ebc2012-04-25 16:21:44 +020019763
Willy Tarreau25c1ebc2012-04-25 16:21:44 +020019764
Willy Tarreau74ca5042013-06-11 23:12:07 +020019765 Example :
19766 # match http://example.com/foo?PHPSESSIONID=some_id
19767 stick on urlp(PHPSESSIONID)
19768 # match http://example.com/foo;JSESSIONID=some_id
19769 stick on urlp(JSESSIONID,;)
Willy Tarreau25c1ebc2012-04-25 16:21:44 +020019770
Jarno Huuskonen676f6222017-03-30 09:19:45 +030019771urlp_val([<name>[,<delim>]]) : integer
Willy Tarreau74ca5042013-06-11 23:12:07 +020019772 See "urlp" above. This one extracts the URL parameter <name> in the request
19773 and converts it to an integer value. This can be used for session stickiness
19774 based on a user ID for example, or with ACLs to match a page number or price.
Willy Tarreaua9fddca2012-07-31 07:51:48 +020019775
Dragan Dosen0070cd52016-06-16 12:19:49 +020019776url32 : integer
19777 This returns a 32-bit hash of the value obtained by concatenating the first
19778 Host header and the whole URL including parameters (not only the path part of
19779 the request, as in the "base32" fetch above). This is useful to track per-URL
19780 activity. A shorter hash is stored, saving a lot of memory. The output type
19781 is an unsigned integer.
19782
19783url32+src : binary
19784 This returns the concatenation of the "url32" fetch and the "src" fetch. The
19785 resulting type is of type binary, with a size of 8 or 20 bytes depending on
19786 the source address family. This can be used to track per-IP, per-URL counters.
19787
Christopher Faulet16032ab2020-04-30 11:30:00 +020019788
Christopher Faulete596d182020-05-05 17:46:34 +0200197897.3.7. Fetching samples for developers
Christopher Fauletd47941d2020-01-08 14:40:19 +010019790---------------------------------------
19791
19792This set of sample fetch methods is reserved to developers and must never be
19793used on a production environment, except on developer demand, for debugging
19794purposes. Moreover, no special care will be taken on backwards compatibility.
19795There is no warranty the following sample fetches will never change, be renamed
19796or simply removed. So be really careful if you should use one of them. To avoid
19797any ambiguity, these sample fetches are placed in the dedicated scope "internal",
19798for instance "internal.strm.is_htx".
19799
19800internal.htx.data : integer
19801 Returns the size in bytes used by data in the HTX message associated to a
19802 channel. The channel is chosen depending on the sample direction.
19803
19804internal.htx.free : integer
19805 Returns the free space (size - used) in bytes in the HTX message associated
19806 to a channel. The channel is chosen depending on the sample direction.
19807
19808internal.htx.free_data : integer
19809 Returns the free space for the data in bytes in the HTX message associated to
19810 a channel. The channel is chosen depending on the sample direction.
19811
19812internal.htx.has_eom : boolean
Christopher Fauletd1ac2b92020-12-02 19:12:22 +010019813 Returns true if the HTX message associated to a channel contains the
19814 end-of-message flag (EOM). Otherwise, it returns false. The channel is chosen
19815 depending on the sample direction.
Christopher Fauletd47941d2020-01-08 14:40:19 +010019816
19817internal.htx.nbblks : integer
19818 Returns the number of blocks present in the HTX message associated to a
19819 channel. The channel is chosen depending on the sample direction.
19820
19821internal.htx.size : integer
19822 Returns the total size in bytes of the HTX message associated to a
19823 channel. The channel is chosen depending on the sample direction.
19824
19825internal.htx.used : integer
19826 Returns the total size used in bytes (data + metadata) in the HTX message
19827 associated to a channel. The channel is chosen depending on the sample
19828 direction.
19829
19830internal.htx_blk.size(<idx>) : integer
19831 Returns the size of the block at the position <idx> in the HTX message
19832 associated to a channel or 0 if it does not exist. The channel is chosen
19833 depending on the sample direction. <idx> may be any positive integer or one
19834 of the special value :
19835 * head : The oldest inserted block
19836 * tail : The newest inserted block
Ilya Shipitsin8525fd92020-02-29 12:34:59 +050019837 * first : The first block where to (re)start the analysis
Christopher Fauletd47941d2020-01-08 14:40:19 +010019838
19839internal.htx_blk.type(<idx>) : string
19840 Returns the type of the block at the position <idx> in the HTX message
19841 associated to a channel or "HTX_BLK_UNUSED" if it does not exist. The channel
19842 is chosen depending on the sample direction. <idx> may be any positive
19843 integer or one of the special value :
19844 * head : The oldest inserted block
19845 * tail : The newest inserted block
Ilya Shipitsin8525fd92020-02-29 12:34:59 +050019846 * first : The first block where to (re)start the analysis
Christopher Fauletd47941d2020-01-08 14:40:19 +010019847
19848internal.htx_blk.data(<idx>) : binary
19849 Returns the value of the DATA block at the position <idx> in the HTX message
19850 associated to a channel or an empty string if it does not exist or if it is
19851 not a DATA block. The channel is chosen depending on the sample direction.
19852 <idx> may be any positive integer or one of the special value :
19853
19854 * head : The oldest inserted block
19855 * tail : The newest inserted block
Ilya Shipitsin8525fd92020-02-29 12:34:59 +050019856 * first : The first block where to (re)start the analysis
Christopher Fauletd47941d2020-01-08 14:40:19 +010019857
19858internal.htx_blk.hdrname(<idx>) : string
19859 Returns the header name of the HEADER block at the position <idx> in the HTX
19860 message associated to a channel or an empty string if it does not exist or if
19861 it is not an HEADER block. The channel is chosen depending on the sample
19862 direction. <idx> may be any positive integer or one of the special value :
19863
19864 * head : The oldest inserted block
19865 * tail : The newest inserted block
Ilya Shipitsin8525fd92020-02-29 12:34:59 +050019866 * first : The first block where to (re)start the analysis
Christopher Fauletd47941d2020-01-08 14:40:19 +010019867
19868internal.htx_blk.hdrval(<idx>) : string
19869 Returns the header value of the HEADER block at the position <idx> in the HTX
19870 message associated to a channel or an empty string if it does not exist or if
19871 it is not an HEADER block. The channel is chosen depending on the sample
19872 direction. <idx> may be any positive integer or one of the special value :
19873
19874 * head : The oldest inserted block
19875 * tail : The newest inserted block
Ilya Shipitsin8525fd92020-02-29 12:34:59 +050019876 * first : The first block where to (re)start the analysis
Christopher Fauletd47941d2020-01-08 14:40:19 +010019877
19878internal.htx_blk.start_line(<idx>) : string
19879 Returns the value of the REQ_SL or RES_SL block at the position <idx> in the
19880 HTX message associated to a channel or an empty string if it does not exist
19881 or if it is not a SL block. The channel is chosen depending on the sample
19882 direction. <idx> may be any positive integer or one of the special value :
19883
19884 * head : The oldest inserted block
19885 * tail : The newest inserted block
Ilya Shipitsin8525fd92020-02-29 12:34:59 +050019886 * first : The first block where to (re)start the analysis
Christopher Fauletd47941d2020-01-08 14:40:19 +010019887
19888internal.strm.is_htx : boolean
19889 Returns true if the current stream is an HTX stream. It means the data in the
19890 channels buffers are stored using the internal HTX representation. Otherwise,
19891 it returns false.
19892
19893
Willy Tarreau74ca5042013-06-11 23:12:07 +0200198947.4. Pre-defined ACLs
Willy Tarreauc57f0e22009-05-10 13:12:33 +020019895---------------------
Willy Tarreauced27012008-01-17 20:35:34 +010019896
Willy Tarreauc57f0e22009-05-10 13:12:33 +020019897Some predefined ACLs are hard-coded so that they do not have to be declared in
19898every frontend which needs them. They all have their names in upper case in
Patrick Mézard2382ad62010-05-09 10:43:32 +020019899order to avoid confusion. Their equivalence is provided below.
Willy Tarreauced27012008-01-17 20:35:34 +010019900
Willy Tarreauc57f0e22009-05-10 13:12:33 +020019901ACL name Equivalent to Usage
Christopher Faulet779184e2021-04-01 17:24:04 +020019902---------------+----------------------------------+------------------------------------------------------
19903FALSE always_false never match
19904HTTP req.proto_http match if request protocol is valid HTTP
19905HTTP_1.0 req.ver 1.0 match if HTTP request version is 1.0
19906HTTP_1.1 req.ver 1.1 match if HTTP request version is 1.1
Christopher Faulet8043e832021-03-26 16:00:54 +010019907HTTP_2.0 req.ver 2.0 match if HTTP request version is 2.0
Christopher Faulet779184e2021-04-01 17:24:04 +020019908HTTP_CONTENT req.hdr_val(content-length) gt 0 match an existing content-length in the HTTP request
19909HTTP_URL_ABS url_reg ^[^/:]*:// match absolute URL with scheme
19910HTTP_URL_SLASH url_beg / match URL beginning with "/"
19911HTTP_URL_STAR url * match URL equal to "*"
19912LOCALHOST src 127.0.0.1/8 match connection from local host
19913METH_CONNECT method CONNECT match HTTP CONNECT method
19914METH_DELETE method DELETE match HTTP DELETE method
19915METH_GET method GET HEAD match HTTP GET or HEAD method
19916METH_HEAD method HEAD match HTTP HEAD method
19917METH_OPTIONS method OPTIONS match HTTP OPTIONS method
19918METH_POST method POST match HTTP POST method
19919METH_PUT method PUT match HTTP PUT method
19920METH_TRACE method TRACE match HTTP TRACE method
19921RDP_COOKIE req.rdp_cookie_cnt gt 0 match presence of an RDP cookie in the request buffer
19922REQ_CONTENT req.len gt 0 match data in the request buffer
19923TRUE always_true always match
19924WAIT_END wait_end wait for end of content analysis
19925---------------+----------------------------------+------------------------------------------------------
Willy Tarreauced27012008-01-17 20:35:34 +010019926
Willy Tarreaub937b7e2010-01-12 15:27:54 +010019927
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200199288. Logging
19929----------
Willy Tarreau844e3c52008-01-11 16:28:18 +010019930
Willy Tarreaucc6c8912009-02-22 10:53:55 +010019931One of HAProxy's strong points certainly lies is its precise logs. It probably
19932provides the finest level of information available for such a product, which is
19933very important for troubleshooting complex environments. Standard information
19934provided in logs include client ports, TCP/HTTP state timers, precise session
19935state at termination and precise termination cause, information about decisions
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +010019936to direct traffic to a server, and of course the ability to capture arbitrary
Willy Tarreaucc6c8912009-02-22 10:53:55 +010019937headers.
19938
19939In order to improve administrators reactivity, it offers a great transparency
19940about encountered problems, both internal and external, and it is possible to
19941send logs to different sources at the same time with different level filters :
19942
19943 - global process-level logs (system errors, start/stop, etc..)
19944 - per-instance system and internal errors (lack of resource, bugs, ...)
19945 - per-instance external troubles (servers up/down, max connections)
19946 - per-instance activity (client connections), either at the establishment or
19947 at the termination.
Davor Ocelice9ed2812017-12-25 17:49:28 +010019948 - per-request control of log-level, e.g.
Jim Freeman9e8714b2015-05-26 09:16:34 -060019949 http-request set-log-level silent if sensitive_request
Willy Tarreaucc6c8912009-02-22 10:53:55 +010019950
19951The ability to distribute different levels of logs to different log servers
19952allow several production teams to interact and to fix their problems as soon
19953as possible. For example, the system team might monitor system-wide errors,
19954while the application team might be monitoring the up/down for their servers in
19955real time, and the security team might analyze the activity logs with one hour
19956delay.
19957
19958
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200199598.1. Log levels
19960---------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010019961
Simon Hormandf791f52011-05-29 15:01:10 +090019962TCP and HTTP connections can be logged with information such as the date, time,
Willy Tarreaucc6c8912009-02-22 10:53:55 +010019963source IP address, destination address, connection duration, response times,
Simon Hormandf791f52011-05-29 15:01:10 +090019964HTTP request, HTTP return code, number of bytes transmitted, conditions
19965in which the session ended, and even exchanged cookies values. For example
19966track a particular user's problems. All messages may be sent to up to two
19967syslog servers. Check the "log" keyword in section 4.2 for more information
19968about log facilities.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010019969
19970
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200199718.2. Log formats
19972----------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010019973
William Lallemand48940402012-01-30 16:47:22 +010019974HAProxy supports 5 log formats. Several fields are common between these formats
Simon Hormandf791f52011-05-29 15:01:10 +090019975and will be detailed in the following sections. A few of them may vary
19976slightly with the configuration, due to indicators specific to certain
19977options. The supported formats are as follows :
Willy Tarreaucc6c8912009-02-22 10:53:55 +010019978
19979 - the default format, which is very basic and very rarely used. It only
19980 provides very basic information about the incoming connection at the moment
19981 it is accepted : source IP:port, destination IP:port, and frontend-name.
19982 This mode will eventually disappear so it will not be described to great
19983 extents.
19984
19985 - the TCP format, which is more advanced. This format is enabled when "option
19986 tcplog" is set on the frontend. HAProxy will then usually wait for the
19987 connection to terminate before logging. This format provides much richer
19988 information, such as timers, connection counts, queue size, etc... This
19989 format is recommended for pure TCP proxies.
19990
19991 - the HTTP format, which is the most advanced for HTTP proxying. This format
19992 is enabled when "option httplog" is set on the frontend. It provides the
19993 same information as the TCP format with some HTTP-specific fields such as
19994 the request, the status code, and captures of headers and cookies. This
19995 format is recommended for HTTP proxies.
19996
Emeric Brun3a058f32009-06-30 18:26:00 +020019997 - the CLF HTTP format, which is equivalent to the HTTP format, but with the
19998 fields arranged in the same order as the CLF format. In this mode, all
19999 timers, captures, flags, etc... appear one per field after the end of the
20000 common fields, in the same order they appear in the standard HTTP format.
20001
William Lallemand48940402012-01-30 16:47:22 +010020002 - the custom log format, allows you to make your own log line.
20003
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020004Next sections will go deeper into details for each of these formats. Format
20005specification will be performed on a "field" basis. Unless stated otherwise, a
20006field is a portion of text delimited by any number of spaces. Since syslog
20007servers are susceptible of inserting fields at the beginning of a line, it is
20008always assumed that the first field is the one containing the process name and
20009identifier.
20010
20011Note : Since log lines may be quite long, the log examples in sections below
20012 might be broken into multiple lines. The example log lines will be
20013 prefixed with 3 closing angle brackets ('>>>') and each time a log is
20014 broken into multiple lines, each non-final line will end with a
20015 backslash ('\') and the next line will start indented by two characters.
20016
20017
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200200188.2.1. Default log format
20019-------------------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020020
20021This format is used when no specific option is set. The log is emitted as soon
20022as the connection is accepted. One should note that this currently is the only
20023format which logs the request's destination IP and ports.
20024
20025 Example :
20026 listen www
20027 mode http
20028 log global
20029 server srv1 127.0.0.1:8000
20030
20031 >>> Feb 6 12:12:09 localhost \
20032 haproxy[14385]: Connect from 10.0.1.2:33312 to 10.0.3.31:8012 \
20033 (www/HTTP)
20034
20035 Field Format Extract from the example above
20036 1 process_name '[' pid ']:' haproxy[14385]:
20037 2 'Connect from' Connect from
20038 3 source_ip ':' source_port 10.0.1.2:33312
20039 4 'to' to
20040 5 destination_ip ':' destination_port 10.0.3.31:8012
20041 6 '(' frontend_name '/' mode ')' (www/HTTP)
20042
20043Detailed fields description :
20044 - "source_ip" is the IP address of the client which initiated the connection.
20045 - "source_port" is the TCP port of the client which initiated the connection.
20046 - "destination_ip" is the IP address the client connected to.
20047 - "destination_port" is the TCP port the client connected to.
20048 - "frontend_name" is the name of the frontend (or listener) which received
20049 and processed the connection.
20050 - "mode is the mode the frontend is operating (TCP or HTTP).
20051
Willy Tarreauceb24bc2010-11-09 12:46:41 +010020052In case of a UNIX socket, the source and destination addresses are marked as
20053"unix:" and the ports reflect the internal ID of the socket which accepted the
20054connection (the same ID as reported in the stats).
20055
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020056It is advised not to use this deprecated format for newer installations as it
20057will eventually disappear.
20058
20059
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200200608.2.2. TCP log format
20061---------------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020062
20063The TCP format is used when "option tcplog" is specified in the frontend, and
20064is the recommended format for pure TCP proxies. It provides a lot of precious
20065information for troubleshooting. Since this format includes timers and byte
20066counts, the log is normally emitted at the end of the session. It can be
20067emitted earlier if "option logasap" is specified, which makes sense in most
20068environments with long sessions such as remote terminals. Sessions which match
20069the "monitor" rules are never logged. It is also possible not to emit logs for
20070sessions for which no data were exchanged between the client and the server, by
Willy Tarreauc9bd0cc2009-05-10 11:57:02 +020020071specifying "option dontlognull" in the frontend. Successful connections will
20072not be logged if "option dontlog-normal" is specified in the frontend. A few
20073fields may slightly vary depending on some configuration options, those are
20074marked with a star ('*') after the field name below.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020075
20076 Example :
20077 frontend fnt
20078 mode tcp
20079 option tcplog
20080 log global
20081 default_backend bck
20082
20083 backend bck
20084 server srv1 127.0.0.1:8000
20085
20086 >>> Feb 6 12:12:56 localhost \
20087 haproxy[14387]: 10.0.1.2:33313 [06/Feb/2009:12:12:51.443] fnt \
20088 bck/srv1 0/0/5007 212 -- 0/0/0/0/3 0/0
20089
20090 Field Format Extract from the example above
20091 1 process_name '[' pid ']:' haproxy[14387]:
20092 2 client_ip ':' client_port 10.0.1.2:33313
20093 3 '[' accept_date ']' [06/Feb/2009:12:12:51.443]
20094 4 frontend_name fnt
20095 5 backend_name '/' server_name bck/srv1
20096 6 Tw '/' Tc '/' Tt* 0/0/5007
20097 7 bytes_read* 212
20098 8 termination_state --
20099 9 actconn '/' feconn '/' beconn '/' srv_conn '/' retries* 0/0/0/0/3
20100 10 srv_queue '/' backend_queue 0/0
20101
20102Detailed fields description :
20103 - "client_ip" is the IP address of the client which initiated the TCP
Daniel Corbett9f0843f2021-05-08 10:50:37 -040020104 connection to HAProxy. If the connection was accepted on a UNIX socket
Willy Tarreauceb24bc2010-11-09 12:46:41 +010020105 instead, the IP address would be replaced with the word "unix". Note that
20106 when the connection is accepted on a socket configured with "accept-proxy"
Bertrand Jacquin93b227d2016-06-04 15:11:10 +010020107 and the PROXY protocol is correctly used, or with a "accept-netscaler-cip"
Davor Ocelice9ed2812017-12-25 17:49:28 +010020108 and the NetScaler Client IP insertion protocol is correctly used, then the
Bertrand Jacquin93b227d2016-06-04 15:11:10 +010020109 logs will reflect the forwarded connection's information.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020110
20111 - "client_port" is the TCP port of the client which initiated the connection.
Willy Tarreauceb24bc2010-11-09 12:46:41 +010020112 If the connection was accepted on a UNIX socket instead, the port would be
20113 replaced with the ID of the accepting socket, which is also reported in the
20114 stats interface.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020115
Daniel Corbett9f0843f2021-05-08 10:50:37 -040020116 - "accept_date" is the exact date when the connection was received by HAProxy
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020117 (which might be very slightly different from the date observed on the
20118 network if there was some queuing in the system's backlog). This is usually
Willy Tarreau590a0512018-09-05 11:56:48 +020020119 the same date which may appear in any upstream firewall's log. When used in
20120 HTTP mode, the accept_date field will be reset to the first moment the
20121 connection is ready to receive a new request (end of previous response for
20122 HTTP/1, immediately after previous request for HTTP/2).
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020123
20124 - "frontend_name" is the name of the frontend (or listener) which received
20125 and processed the connection.
20126
20127 - "backend_name" is the name of the backend (or listener) which was selected
20128 to manage the connection to the server. This will be the same as the
20129 frontend if no switching rule has been applied, which is common for TCP
20130 applications.
20131
20132 - "server_name" is the name of the last server to which the connection was
20133 sent, which might differ from the first one if there were connection errors
20134 and a redispatch occurred. Note that this server belongs to the backend
20135 which processed the request. If the connection was aborted before reaching
20136 a server, "<NOSRV>" is indicated instead of a server name.
20137
20138 - "Tw" is the total time in milliseconds spent waiting in the various queues.
20139 It can be "-1" if the connection was aborted before reaching the queue.
20140 See "Timers" below for more details.
20141
20142 - "Tc" is the total time in milliseconds spent waiting for the connection to
20143 establish to the final server, including retries. It can be "-1" if the
20144 connection was aborted before a connection could be established. See
20145 "Timers" below for more details.
20146
20147 - "Tt" is the total time in milliseconds elapsed between the accept and the
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030020148 last close. It covers all possible processing. There is one exception, if
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020149 "option logasap" was specified, then the time counting stops at the moment
20150 the log is emitted. In this case, a '+' sign is prepended before the value,
20151 indicating that the final one will be larger. See "Timers" below for more
20152 details.
20153
20154 - "bytes_read" is the total number of bytes transmitted from the server to
20155 the client when the log is emitted. If "option logasap" is specified, the
20156 this value will be prefixed with a '+' sign indicating that the final one
20157 may be larger. Please note that this value is a 64-bit counter, so log
20158 analysis tools must be able to handle it without overflowing.
20159
20160 - "termination_state" is the condition the session was in when the session
20161 ended. This indicates the session state, which side caused the end of
20162 session to happen, and for what reason (timeout, error, ...). The normal
20163 flags should be "--", indicating the session was closed by either end with
20164 no data remaining in buffers. See below "Session state at disconnection"
20165 for more details.
20166
20167 - "actconn" is the total number of concurrent connections on the process when
Jamie Gloudonaaa21002012-08-25 00:18:33 -040020168 the session was logged. It is useful to detect when some per-process system
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020169 limits have been reached. For instance, if actconn is close to 512 when
20170 multiple connection errors occur, chances are high that the system limits
20171 the process to use a maximum of 1024 file descriptors and that all of them
Willy Tarreauc57f0e22009-05-10 13:12:33 +020020172 are used. See section 3 "Global parameters" to find how to tune the system.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020173
20174 - "feconn" is the total number of concurrent connections on the frontend when
20175 the session was logged. It is useful to estimate the amount of resource
20176 required to sustain high loads, and to detect when the frontend's "maxconn"
20177 has been reached. Most often when this value increases by huge jumps, it is
20178 because there is congestion on the backend servers, but sometimes it can be
20179 caused by a denial of service attack.
20180
20181 - "beconn" is the total number of concurrent connections handled by the
20182 backend when the session was logged. It includes the total number of
20183 concurrent connections active on servers as well as the number of
20184 connections pending in queues. It is useful to estimate the amount of
20185 additional servers needed to support high loads for a given application.
20186 Most often when this value increases by huge jumps, it is because there is
20187 congestion on the backend servers, but sometimes it can be caused by a
20188 denial of service attack.
20189
20190 - "srv_conn" is the total number of concurrent connections still active on
20191 the server when the session was logged. It can never exceed the server's
20192 configured "maxconn" parameter. If this value is very often close or equal
20193 to the server's "maxconn", it means that traffic regulation is involved a
20194 lot, meaning that either the server's maxconn value is too low, or that
20195 there aren't enough servers to process the load with an optimal response
20196 time. When only one of the server's "srv_conn" is high, it usually means
20197 that this server has some trouble causing the connections to take longer to
20198 be processed than on other servers.
20199
20200 - "retries" is the number of connection retries experienced by this session
20201 when trying to connect to the server. It must normally be zero, unless a
20202 server is being stopped at the same moment the connection was attempted.
20203 Frequent retries generally indicate either a network problem between
Daniel Corbett9f0843f2021-05-08 10:50:37 -040020204 HAProxy and the server, or a misconfigured system backlog on the server
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020205 preventing new connections from being queued. This field may optionally be
20206 prefixed with a '+' sign, indicating that the session has experienced a
20207 redispatch after the maximal retry count has been reached on the initial
20208 server. In this case, the server name appearing in the log is the one the
20209 connection was redispatched to, and not the first one, though both may
20210 sometimes be the same in case of hashing for instance. So as a general rule
20211 of thumb, when a '+' is present in front of the retry count, this count
20212 should not be attributed to the logged server.
20213
20214 - "srv_queue" is the total number of requests which were processed before
20215 this one in the server queue. It is zero when the request has not gone
20216 through the server queue. It makes it possible to estimate the approximate
20217 server's response time by dividing the time spent in queue by the number of
20218 requests in the queue. It is worth noting that if a session experiences a
20219 redispatch and passes through two server queues, their positions will be
Davor Ocelice9ed2812017-12-25 17:49:28 +010020220 cumulative. A request should not pass through both the server queue and the
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020221 backend queue unless a redispatch occurs.
20222
20223 - "backend_queue" is the total number of requests which were processed before
20224 this one in the backend's global queue. It is zero when the request has not
20225 gone through the global queue. It makes it possible to estimate the average
20226 queue length, which easily translates into a number of missing servers when
20227 divided by a server's "maxconn" parameter. It is worth noting that if a
20228 session experiences a redispatch, it may pass twice in the backend's queue,
Davor Ocelice9ed2812017-12-25 17:49:28 +010020229 and then both positions will be cumulative. A request should not pass
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020230 through both the server queue and the backend queue unless a redispatch
20231 occurs.
20232
20233
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200202348.2.3. HTTP log format
20235----------------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020236
20237The HTTP format is the most complete and the best suited for HTTP proxies. It
20238is enabled by when "option httplog" is specified in the frontend. It provides
20239the same level of information as the TCP format with additional features which
20240are specific to the HTTP protocol. Just like the TCP format, the log is usually
20241emitted at the end of the session, unless "option logasap" is specified, which
20242generally only makes sense for download sites. A session which matches the
20243"monitor" rules will never logged. It is also possible not to log sessions for
20244which no data were sent by the client by specifying "option dontlognull" in the
Willy Tarreauc9bd0cc2009-05-10 11:57:02 +020020245frontend. Successful connections will not be logged if "option dontlog-normal"
20246is specified in the frontend.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020247
20248Most fields are shared with the TCP log, some being different. A few fields may
20249slightly vary depending on some configuration options. Those ones are marked
20250with a star ('*') after the field name below.
20251
20252 Example :
20253 frontend http-in
20254 mode http
20255 option httplog
20256 log global
20257 default_backend bck
20258
20259 backend static
20260 server srv1 127.0.0.1:8000
20261
20262 >>> Feb 6 12:14:14 localhost \
20263 haproxy[14389]: 10.0.1.2:33317 [06/Feb/2009:12:14:14.655] http-in \
20264 static/srv1 10/0/30/69/109 200 2750 - - ---- 1/1/1/1/0 0/0 {1wt.eu} \
Willy Tarreaud72758d2010-01-12 10:42:19 +010020265 {} "GET /index.html HTTP/1.1"
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020266
20267 Field Format Extract from the example above
20268 1 process_name '[' pid ']:' haproxy[14389]:
20269 2 client_ip ':' client_port 10.0.1.2:33317
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020020270 3 '[' request_date ']' [06/Feb/2009:12:14:14.655]
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020271 4 frontend_name http-in
20272 5 backend_name '/' server_name static/srv1
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020020273 6 TR '/' Tw '/' Tc '/' Tr '/' Ta* 10/0/30/69/109
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020274 7 status_code 200
20275 8 bytes_read* 2750
20276 9 captured_request_cookie -
20277 10 captured_response_cookie -
20278 11 termination_state ----
20279 12 actconn '/' feconn '/' beconn '/' srv_conn '/' retries* 1/1/1/1/0
20280 13 srv_queue '/' backend_queue 0/0
20281 14 '{' captured_request_headers* '}' {haproxy.1wt.eu}
20282 15 '{' captured_response_headers* '}' {}
20283 16 '"' http_request '"' "GET /index.html HTTP/1.1"
Willy Tarreaud72758d2010-01-12 10:42:19 +010020284
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020285Detailed fields description :
20286 - "client_ip" is the IP address of the client which initiated the TCP
Daniel Corbett9f0843f2021-05-08 10:50:37 -040020287 connection to HAProxy. If the connection was accepted on a UNIX socket
Willy Tarreauceb24bc2010-11-09 12:46:41 +010020288 instead, the IP address would be replaced with the word "unix". Note that
20289 when the connection is accepted on a socket configured with "accept-proxy"
Bertrand Jacquin93b227d2016-06-04 15:11:10 +010020290 and the PROXY protocol is correctly used, or with a "accept-netscaler-cip"
Davor Ocelice9ed2812017-12-25 17:49:28 +010020291 and the NetScaler Client IP insertion protocol is correctly used, then the
Bertrand Jacquin93b227d2016-06-04 15:11:10 +010020292 logs will reflect the forwarded connection's information.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020293
20294 - "client_port" is the TCP port of the client which initiated the connection.
Willy Tarreauceb24bc2010-11-09 12:46:41 +010020295 If the connection was accepted on a UNIX socket instead, the port would be
20296 replaced with the ID of the accepting socket, which is also reported in the
20297 stats interface.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020298
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020020299 - "request_date" is the exact date when the first byte of the HTTP request
Daniel Corbett9f0843f2021-05-08 10:50:37 -040020300 was received by HAProxy (log field %tr).
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020301
20302 - "frontend_name" is the name of the frontend (or listener) which received
20303 and processed the connection.
20304
20305 - "backend_name" is the name of the backend (or listener) which was selected
20306 to manage the connection to the server. This will be the same as the
20307 frontend if no switching rule has been applied.
20308
20309 - "server_name" is the name of the last server to which the connection was
20310 sent, which might differ from the first one if there were connection errors
20311 and a redispatch occurred. Note that this server belongs to the backend
20312 which processed the request. If the request was aborted before reaching a
20313 server, "<NOSRV>" is indicated instead of a server name. If the request was
20314 intercepted by the stats subsystem, "<STATS>" is indicated instead.
20315
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020020316 - "TR" is the total time in milliseconds spent waiting for a full HTTP
20317 request from the client (not counting body) after the first byte was
20318 received. It can be "-1" if the connection was aborted before a complete
John Roeslerfb2fce12019-07-10 15:45:51 -050020319 request could be received or a bad request was received. It should
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020020320 always be very small because a request generally fits in one single packet.
20321 Large times here generally indicate network issues between the client and
Daniel Corbett9f0843f2021-05-08 10:50:37 -040020322 HAProxy or requests being typed by hand. See section 8.4 "Timing Events"
Willy Tarreau590a0512018-09-05 11:56:48 +020020323 for more details.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020324
20325 - "Tw" is the total time in milliseconds spent waiting in the various queues.
20326 It can be "-1" if the connection was aborted before reaching the queue.
Willy Tarreau590a0512018-09-05 11:56:48 +020020327 See section 8.4 "Timing Events" for more details.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020328
20329 - "Tc" is the total time in milliseconds spent waiting for the connection to
20330 establish to the final server, including retries. It can be "-1" if the
Willy Tarreau590a0512018-09-05 11:56:48 +020020331 request was aborted before a connection could be established. See section
20332 8.4 "Timing Events" for more details.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020333
20334 - "Tr" is the total time in milliseconds spent waiting for the server to send
20335 a full HTTP response, not counting data. It can be "-1" if the request was
20336 aborted before a complete response could be received. It generally matches
20337 the server's processing time for the request, though it may be altered by
20338 the amount of data sent by the client to the server. Large times here on
Willy Tarreau590a0512018-09-05 11:56:48 +020020339 "GET" requests generally indicate an overloaded server. See section 8.4
20340 "Timing Events" for more details.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020341
Daniel Corbett9f0843f2021-05-08 10:50:37 -040020342 - "Ta" is the time the request remained active in HAProxy, which is the total
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020020343 time in milliseconds elapsed between the first byte of the request was
20344 received and the last byte of response was sent. It covers all possible
20345 processing except the handshake (see Th) and idle time (see Ti). There is
20346 one exception, if "option logasap" was specified, then the time counting
20347 stops at the moment the log is emitted. In this case, a '+' sign is
20348 prepended before the value, indicating that the final one will be larger.
Willy Tarreau590a0512018-09-05 11:56:48 +020020349 See section 8.4 "Timing Events" for more details.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020350
20351 - "status_code" is the HTTP status code returned to the client. This status
Daniel Corbett9f0843f2021-05-08 10:50:37 -040020352 is generally set by the server, but it might also be set by HAProxy when
20353 the server cannot be reached or when its response is blocked by HAProxy.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020354
20355 - "bytes_read" is the total number of bytes transmitted to the client when
20356 the log is emitted. This does include HTTP headers. If "option logasap" is
John Roeslerfb2fce12019-07-10 15:45:51 -050020357 specified, this value will be prefixed with a '+' sign indicating that
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020358 the final one may be larger. Please note that this value is a 64-bit
20359 counter, so log analysis tools must be able to handle it without
20360 overflowing.
20361
20362 - "captured_request_cookie" is an optional "name=value" entry indicating that
20363 the client had this cookie in the request. The cookie name and its maximum
20364 length are defined by the "capture cookie" statement in the frontend
20365 configuration. The field is a single dash ('-') when the option is not
20366 set. Only one cookie may be captured, it is generally used to track session
20367 ID exchanges between a client and a server to detect session crossing
20368 between clients due to application bugs. For more details, please consult
20369 the section "Capturing HTTP headers and cookies" below.
20370
20371 - "captured_response_cookie" is an optional "name=value" entry indicating
20372 that the server has returned a cookie with its response. The cookie name
20373 and its maximum length are defined by the "capture cookie" statement in the
20374 frontend configuration. The field is a single dash ('-') when the option is
20375 not set. Only one cookie may be captured, it is generally used to track
20376 session ID exchanges between a client and a server to detect session
20377 crossing between clients due to application bugs. For more details, please
20378 consult the section "Capturing HTTP headers and cookies" below.
20379
20380 - "termination_state" is the condition the session was in when the session
20381 ended. This indicates the session state, which side caused the end of
20382 session to happen, for what reason (timeout, error, ...), just like in TCP
20383 logs, and information about persistence operations on cookies in the last
20384 two characters. The normal flags should begin with "--", indicating the
20385 session was closed by either end with no data remaining in buffers. See
20386 below "Session state at disconnection" for more details.
20387
20388 - "actconn" is the total number of concurrent connections on the process when
Jamie Gloudonaaa21002012-08-25 00:18:33 -040020389 the session was logged. It is useful to detect when some per-process system
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020390 limits have been reached. For instance, if actconn is close to 512 or 1024
20391 when multiple connection errors occur, chances are high that the system
20392 limits the process to use a maximum of 1024 file descriptors and that all
Willy Tarreauc57f0e22009-05-10 13:12:33 +020020393 of them are used. See section 3 "Global parameters" to find how to tune the
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020394 system.
20395
20396 - "feconn" is the total number of concurrent connections on the frontend when
20397 the session was logged. It is useful to estimate the amount of resource
20398 required to sustain high loads, and to detect when the frontend's "maxconn"
20399 has been reached. Most often when this value increases by huge jumps, it is
20400 because there is congestion on the backend servers, but sometimes it can be
20401 caused by a denial of service attack.
20402
20403 - "beconn" is the total number of concurrent connections handled by the
20404 backend when the session was logged. It includes the total number of
20405 concurrent connections active on servers as well as the number of
20406 connections pending in queues. It is useful to estimate the amount of
20407 additional servers needed to support high loads for a given application.
20408 Most often when this value increases by huge jumps, it is because there is
20409 congestion on the backend servers, but sometimes it can be caused by a
20410 denial of service attack.
20411
20412 - "srv_conn" is the total number of concurrent connections still active on
20413 the server when the session was logged. It can never exceed the server's
20414 configured "maxconn" parameter. If this value is very often close or equal
20415 to the server's "maxconn", it means that traffic regulation is involved a
20416 lot, meaning that either the server's maxconn value is too low, or that
20417 there aren't enough servers to process the load with an optimal response
20418 time. When only one of the server's "srv_conn" is high, it usually means
20419 that this server has some trouble causing the requests to take longer to be
20420 processed than on other servers.
20421
20422 - "retries" is the number of connection retries experienced by this session
20423 when trying to connect to the server. It must normally be zero, unless a
20424 server is being stopped at the same moment the connection was attempted.
20425 Frequent retries generally indicate either a network problem between
Daniel Corbett9f0843f2021-05-08 10:50:37 -040020426 HAProxy and the server, or a misconfigured system backlog on the server
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020427 preventing new connections from being queued. This field may optionally be
20428 prefixed with a '+' sign, indicating that the session has experienced a
20429 redispatch after the maximal retry count has been reached on the initial
20430 server. In this case, the server name appearing in the log is the one the
20431 connection was redispatched to, and not the first one, though both may
20432 sometimes be the same in case of hashing for instance. So as a general rule
20433 of thumb, when a '+' is present in front of the retry count, this count
20434 should not be attributed to the logged server.
20435
20436 - "srv_queue" is the total number of requests which were processed before
20437 this one in the server queue. It is zero when the request has not gone
20438 through the server queue. It makes it possible to estimate the approximate
20439 server's response time by dividing the time spent in queue by the number of
20440 requests in the queue. It is worth noting that if a session experiences a
20441 redispatch and passes through two server queues, their positions will be
Davor Ocelice9ed2812017-12-25 17:49:28 +010020442 cumulative. A request should not pass through both the server queue and the
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020443 backend queue unless a redispatch occurs.
20444
20445 - "backend_queue" is the total number of requests which were processed before
20446 this one in the backend's global queue. It is zero when the request has not
20447 gone through the global queue. It makes it possible to estimate the average
20448 queue length, which easily translates into a number of missing servers when
20449 divided by a server's "maxconn" parameter. It is worth noting that if a
20450 session experiences a redispatch, it may pass twice in the backend's queue,
Davor Ocelice9ed2812017-12-25 17:49:28 +010020451 and then both positions will be cumulative. A request should not pass
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020452 through both the server queue and the backend queue unless a redispatch
20453 occurs.
20454
20455 - "captured_request_headers" is a list of headers captured in the request due
20456 to the presence of the "capture request header" statement in the frontend.
20457 Multiple headers can be captured, they will be delimited by a vertical bar
20458 ('|'). When no capture is enabled, the braces do not appear, causing a
20459 shift of remaining fields. It is important to note that this field may
20460 contain spaces, and that using it requires a smarter log parser than when
20461 it's not used. Please consult the section "Capturing HTTP headers and
20462 cookies" below for more details.
20463
20464 - "captured_response_headers" is a list of headers captured in the response
20465 due to the presence of the "capture response header" statement in the
20466 frontend. Multiple headers can be captured, they will be delimited by a
20467 vertical bar ('|'). When no capture is enabled, the braces do not appear,
20468 causing a shift of remaining fields. It is important to note that this
20469 field may contain spaces, and that using it requires a smarter log parser
20470 than when it's not used. Please consult the section "Capturing HTTP headers
20471 and cookies" below for more details.
20472
20473 - "http_request" is the complete HTTP request line, including the method,
20474 request and HTTP version string. Non-printable characters are encoded (see
20475 below the section "Non-printable characters"). This is always the last
20476 field, and it is always delimited by quotes and is the only one which can
20477 contain quotes. If new fields are added to the log format, they will be
20478 added before this field. This field might be truncated if the request is
20479 huge and does not fit in the standard syslog buffer (1024 characters). This
20480 is the reason why this field must always remain the last one.
20481
20482
Cyril Bontédc4d9032012-04-08 21:57:39 +0200204838.2.4. Custom log format
20484------------------------
William Lallemand48940402012-01-30 16:47:22 +010020485
Willy Tarreau2beef582012-12-20 17:22:52 +010020486The directive log-format allows you to customize the logs in http mode and tcp
William Lallemandbddd4fd2012-02-27 11:23:10 +010020487mode. It takes a string as argument.
William Lallemand48940402012-01-30 16:47:22 +010020488
Davor Ocelice9ed2812017-12-25 17:49:28 +010020489HAProxy understands some log format variables. % precedes log format variables.
William Lallemand48940402012-01-30 16:47:22 +010020490Variables can take arguments using braces ('{}'), and multiple arguments are
20491separated by commas within the braces. Flags may be added or removed by
20492prefixing them with a '+' or '-' sign.
20493
20494Special variable "%o" may be used to propagate its flags to all other
20495variables on the same format string. This is particularly handy with quoted
Dragan Dosen835b9212016-02-12 13:23:03 +010020496("Q") and escaped ("E") string formats.
William Lallemand48940402012-01-30 16:47:22 +010020497
Willy Tarreauc8368452012-12-21 00:09:23 +010020498If a variable is named between square brackets ('[' .. ']') then it is used
Willy Tarreaube722a22014-06-13 16:31:59 +020020499as a sample expression rule (see section 7.3). This it useful to add some
Willy Tarreauc8368452012-12-21 00:09:23 +010020500less common information such as the client's SSL certificate's DN, or to log
20501the key that would be used to store an entry into a stick table.
20502
Dragan Dosen1e3b16f2020-06-23 18:16:44 +020020503Note: spaces must be escaped. In configuration directives "log-format",
20504"log-format-sd" and "unique-id-format", spaces are considered as
20505delimiters and are merged. In order to emit a verbatim '%', it must be
20506preceded by another '%' resulting in '%%'.
William Lallemand48940402012-01-30 16:47:22 +010020507
Dragan Dosen835b9212016-02-12 13:23:03 +010020508Note: when using the RFC5424 syslog message format, the characters '"',
20509'\' and ']' inside PARAM-VALUE should be escaped with '\' as prefix (see
20510https://tools.ietf.org/html/rfc5424#section-6.3.3 for more details). In
20511such cases, the use of the flag "E" should be considered.
20512
William Lallemand48940402012-01-30 16:47:22 +010020513Flags are :
20514 * Q: quote a string
Jamie Gloudonaaa21002012-08-25 00:18:33 -040020515 * X: hexadecimal representation (IPs, Ports, %Ts, %rt, %pid)
Dragan Dosen835b9212016-02-12 13:23:03 +010020516 * E: escape characters '"', '\' and ']' in a string with '\' as prefix
20517 (intended purpose is for the RFC5424 structured-data log formats)
William Lallemand48940402012-01-30 16:47:22 +010020518
20519 Example:
20520
20521 log-format %T\ %t\ Some\ Text
20522 log-format %{+Q}o\ %t\ %s\ %{-Q}r
20523
Dragan Dosen835b9212016-02-12 13:23:03 +010020524 log-format-sd %{+Q,+E}o\ [exampleSDID@1234\ header=%[capture.req.hdr(0)]]
20525
William Lallemand48940402012-01-30 16:47:22 +010020526At the moment, the default HTTP format is defined this way :
20527
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020020528 log-format "%ci:%cp [%tr] %ft %b/%s %TR/%Tw/%Tc/%Tr/%Ta %ST %B %CC \
20529 %CS %tsc %ac/%fc/%bc/%sc/%rc %sq/%bq %hr %hs %{+Q}r"
William Lallemand48940402012-01-30 16:47:22 +010020530
William Lallemandbddd4fd2012-02-27 11:23:10 +010020531the default CLF format is defined this way :
William Lallemand48940402012-01-30 16:47:22 +010020532
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020020533 log-format "%{+Q}o %{-Q}ci - - [%trg] %r %ST %B \"\" \"\" %cp \
20534 %ms %ft %b %s %TR %Tw %Tc %Tr %Ta %tsc %ac %fc \
20535 %bc %sc %rc %sq %bq %CC %CS %hrl %hsl"
William Lallemand48940402012-01-30 16:47:22 +010020536
William Lallemandbddd4fd2012-02-27 11:23:10 +010020537and the default TCP format is defined this way :
20538
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020020539 log-format "%ci:%cp [%t] %ft %b/%s %Tw/%Tc/%Tt %B %ts \
20540 %ac/%fc/%bc/%sc/%rc %sq/%bq"
William Lallemandbddd4fd2012-02-27 11:23:10 +010020541
William Lallemand48940402012-01-30 16:47:22 +010020542Please refer to the table below for currently defined variables :
20543
William Lallemandbddd4fd2012-02-27 11:23:10 +010020544 +---+------+-----------------------------------------------+-------------+
Willy Tarreauffc3fcd2012-10-12 20:17:54 +020020545 | R | var | field name (8.2.2 and 8.2.3 for description) | type |
William Lallemandbddd4fd2012-02-27 11:23:10 +010020546 +---+------+-----------------------------------------------+-------------+
20547 | | %o | special variable, apply flags on all next var | |
20548 +---+------+-----------------------------------------------+-------------+
Willy Tarreau2beef582012-12-20 17:22:52 +010020549 | | %B | bytes_read (from server to client) | numeric |
20550 | H | %CC | captured_request_cookie | string |
20551 | H | %CS | captured_response_cookie | string |
William Lallemand5f232402012-04-05 18:02:55 +020020552 | | %H | hostname | string |
Andrew Hayworth0ebc55f2015-04-27 21:37:03 +000020553 | H | %HM | HTTP method (ex: POST) | string |
Maciej Zdeb21acc332020-11-26 10:45:52 +000020554 | H | %HP | HTTP request URI without query string | string |
Maciej Zdebfcdfd852020-11-30 18:27:47 +000020555 | H | %HPO | HTTP path only (without host nor query string)| string |
Andrew Hayworthe63ac872015-07-31 16:14:16 +000020556 | H | %HQ | HTTP request URI query string (ex: ?bar=baz) | string |
Andrew Hayworth0ebc55f2015-04-27 21:37:03 +000020557 | H | %HU | HTTP request URI (ex: /foo?bar=baz) | string |
20558 | H | %HV | HTTP version (ex: HTTP/1.0) | string |
William Lallemanda73203e2012-03-12 12:48:57 +010020559 | | %ID | unique-id | string |
Willy Tarreau4bf99632014-06-13 12:21:40 +020020560 | | %ST | status_code | numeric |
William Lallemand5f232402012-04-05 18:02:55 +020020561 | | %T | gmt_date_time | date |
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020020562 | | %Ta | Active time of the request (from TR to end) | numeric |
William Lallemandbddd4fd2012-02-27 11:23:10 +010020563 | | %Tc | Tc | numeric |
Willy Tarreau27b639d2016-05-17 17:55:27 +020020564 | | %Td | Td = Tt - (Tq + Tw + Tc + Tr) | numeric |
Yuxans Yao4e25b012012-10-19 10:36:09 +080020565 | | %Tl | local_date_time | date |
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020020566 | | %Th | connection handshake time (SSL, PROXY proto) | numeric |
20567 | H | %Ti | idle time before the HTTP request | numeric |
20568 | H | %Tq | Th + Ti + TR | numeric |
20569 | H | %TR | time to receive the full request from 1st byte| numeric |
20570 | H | %Tr | Tr (response time) | numeric |
William Lallemand5f232402012-04-05 18:02:55 +020020571 | | %Ts | timestamp | numeric |
William Lallemandbddd4fd2012-02-27 11:23:10 +010020572 | | %Tt | Tt | numeric |
Damien Claisse57c8eb92020-04-28 12:09:19 +000020573 | | %Tu | Tu | numeric |
William Lallemandbddd4fd2012-02-27 11:23:10 +010020574 | | %Tw | Tw | numeric |
Willy Tarreau2beef582012-12-20 17:22:52 +010020575 | | %U | bytes_uploaded (from client to server) | numeric |
William Lallemandbddd4fd2012-02-27 11:23:10 +010020576 | | %ac | actconn | numeric |
20577 | | %b | backend_name | string |
Willy Tarreau2beef582012-12-20 17:22:52 +010020578 | | %bc | beconn (backend concurrent connections) | numeric |
20579 | | %bi | backend_source_ip (connecting address) | IP |
20580 | | %bp | backend_source_port (connecting address) | numeric |
William Lallemandbddd4fd2012-02-27 11:23:10 +010020581 | | %bq | backend_queue | numeric |
Willy Tarreau2beef582012-12-20 17:22:52 +010020582 | | %ci | client_ip (accepted address) | IP |
20583 | | %cp | client_port (accepted address) | numeric |
William Lallemandbddd4fd2012-02-27 11:23:10 +010020584 | | %f | frontend_name | string |
Willy Tarreau2beef582012-12-20 17:22:52 +010020585 | | %fc | feconn (frontend concurrent connections) | numeric |
20586 | | %fi | frontend_ip (accepting address) | IP |
20587 | | %fp | frontend_port (accepting address) | numeric |
Willy Tarreau773d65f2012-10-12 14:56:11 +020020588 | | %ft | frontend_name_transport ('~' suffix for SSL) | string |
Willy Tarreau7346acb2014-08-28 15:03:15 +020020589 | | %lc | frontend_log_counter | numeric |
Willy Tarreaud9ed3d22014-06-13 12:23:06 +020020590 | | %hr | captured_request_headers default style | string |
20591 | | %hrl | captured_request_headers CLF style | string list |
20592 | | %hs | captured_response_headers default style | string |
20593 | | %hsl | captured_response_headers CLF style | string list |
Willy Tarreau812c88e2015-08-09 10:56:35 +020020594 | | %ms | accept date milliseconds (left-padded with 0) | numeric |
William Lallemand5f232402012-04-05 18:02:55 +020020595 | | %pid | PID | numeric |
Willy Tarreauffc3fcd2012-10-12 20:17:54 +020020596 | H | %r | http_request | string |
William Lallemandbddd4fd2012-02-27 11:23:10 +010020597 | | %rc | retries | numeric |
Willy Tarreau1f0da242014-01-25 11:01:50 +010020598 | | %rt | request_counter (HTTP req or TCP session) | numeric |
William Lallemandbddd4fd2012-02-27 11:23:10 +010020599 | | %s | server_name | string |
Willy Tarreau2beef582012-12-20 17:22:52 +010020600 | | %sc | srv_conn (server concurrent connections) | numeric |
20601 | | %si | server_IP (target address) | IP |
20602 | | %sp | server_port (target address) | numeric |
William Lallemandbddd4fd2012-02-27 11:23:10 +010020603 | | %sq | srv_queue | numeric |
Willy Tarreauffc3fcd2012-10-12 20:17:54 +020020604 | S | %sslc| ssl_ciphers (ex: AES-SHA) | string |
20605 | S | %sslv| ssl_version (ex: TLSv1) | string |
Willy Tarreau2beef582012-12-20 17:22:52 +010020606 | | %t | date_time (with millisecond resolution) | date |
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020020607 | H | %tr | date_time of HTTP request | date |
20608 | H | %trg | gmt_date_time of start of HTTP request | date |
Jens Bissinger15c64ff2018-08-23 14:11:27 +020020609 | H | %trl | local_date_time of start of HTTP request | date |
William Lallemandbddd4fd2012-02-27 11:23:10 +010020610 | | %ts | termination_state | string |
Willy Tarreauffc3fcd2012-10-12 20:17:54 +020020611 | H | %tsc | termination_state with cookie status | string |
William Lallemandbddd4fd2012-02-27 11:23:10 +010020612 +---+------+-----------------------------------------------+-------------+
William Lallemand48940402012-01-30 16:47:22 +010020613
Willy Tarreauffc3fcd2012-10-12 20:17:54 +020020614 R = Restrictions : H = mode http only ; S = SSL only
William Lallemand48940402012-01-30 16:47:22 +010020615
Willy Tarreau5f51e1a2012-12-03 18:40:10 +010020616
206178.2.5. Error log format
20618-----------------------
20619
20620When an incoming connection fails due to an SSL handshake or an invalid PROXY
Daniel Corbett9f0843f2021-05-08 10:50:37 -040020621protocol header, HAProxy will log the event using a shorter, fixed line format.
Willy Tarreau5f51e1a2012-12-03 18:40:10 +010020622By default, logs are emitted at the LOG_INFO level, unless the option
20623"log-separate-errors" is set in the backend, in which case the LOG_ERR level
Davor Ocelice9ed2812017-12-25 17:49:28 +010020624will be used. Connections on which no data are exchanged (e.g. probes) are not
Willy Tarreau5f51e1a2012-12-03 18:40:10 +010020625logged if the "dontlognull" option is set.
20626
20627The format looks like this :
20628
20629 >>> Dec 3 18:27:14 localhost \
20630 haproxy[6103]: 127.0.0.1:56059 [03/Dec/2012:17:35:10.380] frt/f1: \
20631 Connection error during SSL handshake
20632
20633 Field Format Extract from the example above
20634 1 process_name '[' pid ']:' haproxy[6103]:
20635 2 client_ip ':' client_port 127.0.0.1:56059
20636 3 '[' accept_date ']' [03/Dec/2012:17:35:10.380]
20637 4 frontend_name "/" bind_name ":" frt/f1:
20638 5 message Connection error during SSL handshake
20639
20640These fields just provide minimal information to help debugging connection
20641failures.
20642
20643
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200206448.3. Advanced logging options
20645-----------------------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020646
20647Some advanced logging options are often looked for but are not easy to find out
20648just by looking at the various options. Here is an entry point for the few
20649options which can enable better logging. Please refer to the keywords reference
20650for more information about their usage.
20651
20652
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200206538.3.1. Disabling logging of external tests
20654------------------------------------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020655
20656It is quite common to have some monitoring tools perform health checks on
Daniel Corbett9f0843f2021-05-08 10:50:37 -040020657HAProxy. Sometimes it will be a layer 3 load-balancer such as LVS or any
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020658commercial load-balancer, and sometimes it will simply be a more complete
20659monitoring system such as Nagios. When the tests are very frequent, users often
20660ask how to disable logging for those checks. There are three possibilities :
20661
20662 - if connections come from everywhere and are just TCP probes, it is often
20663 desired to simply disable logging of connections without data exchange, by
20664 setting "option dontlognull" in the frontend. It also disables logging of
20665 port scans, which may or may not be desired.
20666
Willy Tarreau9e9919d2020-10-14 15:55:23 +020020667 - it is possible to use the "http-request set-log-level silent" action using
20668 a variety of conditions (source networks, paths, user-agents, etc).
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020669
20670 - if the tests are performed on a known URI, use "monitor-uri" to declare
20671 this URI as dedicated to monitoring. Any host sending this request will
20672 only get the result of a health-check, and the request will not be logged.
20673
20674
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200206758.3.2. Logging before waiting for the session to terminate
20676----------------------------------------------------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020677
20678The problem with logging at end of connection is that you have no clue about
20679what is happening during very long sessions, such as remote terminal sessions
20680or large file downloads. This problem can be worked around by specifying
Davor Ocelice9ed2812017-12-25 17:49:28 +010020681"option logasap" in the frontend. HAProxy will then log as soon as possible,
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020682just before data transfer begins. This means that in case of TCP, it will still
20683log the connection status to the server, and in case of HTTP, it will log just
20684after processing the server headers. In this case, the number of bytes reported
20685is the number of header bytes sent to the client. In order to avoid confusion
20686with normal logs, the total time field and the number of bytes are prefixed
20687with a '+' sign which means that real numbers are certainly larger.
20688
20689
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200206908.3.3. Raising log level upon errors
20691------------------------------------
Willy Tarreauc9bd0cc2009-05-10 11:57:02 +020020692
20693Sometimes it is more convenient to separate normal traffic from errors logs,
20694for instance in order to ease error monitoring from log files. When the option
20695"log-separate-errors" is used, connections which experience errors, timeouts,
20696retries, redispatches or HTTP status codes 5xx will see their syslog level
20697raised from "info" to "err". This will help a syslog daemon store the log in
20698a separate file. It is very important to keep the errors in the normal traffic
20699file too, so that log ordering is not altered. You should also be careful if
20700you already have configured your syslog daemon to store all logs higher than
20701"notice" in an "admin" file, because the "err" level is higher than "notice".
20702
20703
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200207048.3.4. Disabling logging of successful connections
20705--------------------------------------------------
Willy Tarreauc9bd0cc2009-05-10 11:57:02 +020020706
20707Although this may sound strange at first, some large sites have to deal with
20708multiple thousands of logs per second and are experiencing difficulties keeping
20709them intact for a long time or detecting errors within them. If the option
20710"dontlog-normal" is set on the frontend, all normal connections will not be
20711logged. In this regard, a normal connection is defined as one without any
20712error, timeout, retry nor redispatch. In HTTP, the status code is checked too,
20713and a response with a status 5xx is not considered normal and will be logged
20714too. Of course, doing is is really discouraged as it will remove most of the
20715useful information from the logs. Do this only if you have no other
20716alternative.
20717
20718
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200207198.4. Timing events
20720------------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020721
20722Timers provide a great help in troubleshooting network problems. All values are
20723reported in milliseconds (ms). These timers should be used in conjunction with
20724the session termination flags. In TCP mode with "option tcplog" set on the
20725frontend, 3 control points are reported under the form "Tw/Tc/Tt", and in HTTP
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020020726mode, 5 control points are reported under the form "TR/Tw/Tc/Tr/Ta". In
20727addition, three other measures are provided, "Th", "Ti", and "Tq".
20728
Guillaume de Lafondf27cddc2016-12-23 17:32:43 +010020729Timings events in HTTP mode:
20730
20731 first request 2nd request
20732 |<-------------------------------->|<-------------- ...
20733 t tr t tr ...
20734 ---|----|----|----|----|----|----|----|----|--
20735 : Th Ti TR Tw Tc Tr Td : Ti ...
20736 :<---- Tq ---->: :
20737 :<-------------- Tt -------------->:
Damien Claisse57c8eb92020-04-28 12:09:19 +000020738 :<-- -----Tu--------------->:
Guillaume de Lafondf27cddc2016-12-23 17:32:43 +010020739 :<--------- Ta --------->:
20740
20741Timings events in TCP mode:
20742
20743 TCP session
20744 |<----------------->|
20745 t t
20746 ---|----|----|----|----|---
20747 | Th Tw Tc Td |
20748 |<------ Tt ------->|
20749
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020020750 - Th: total time to accept tcp connection and execute handshakes for low level
Davor Ocelice9ed2812017-12-25 17:49:28 +010020751 protocols. Currently, these protocols are proxy-protocol and SSL. This may
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020020752 only happen once during the whole connection's lifetime. A large time here
20753 may indicate that the client only pre-established the connection without
20754 speaking, that it is experiencing network issues preventing it from
Davor Ocelice9ed2812017-12-25 17:49:28 +010020755 completing a handshake in a reasonable time (e.g. MTU issues), or that an
Willy Tarreau590a0512018-09-05 11:56:48 +020020756 SSL handshake was very expensive to compute. Please note that this time is
20757 reported only before the first request, so it is safe to average it over
20758 all request to calculate the amortized value. The second and subsequent
20759 request will always report zero here.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020760
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020020761 - Ti: is the idle time before the HTTP request (HTTP mode only). This timer
20762 counts between the end of the handshakes and the first byte of the HTTP
20763 request. When dealing with a second request in keep-alive mode, it starts
Willy Tarreau590a0512018-09-05 11:56:48 +020020764 to count after the end of the transmission the previous response. When a
20765 multiplexed protocol such as HTTP/2 is used, it starts to count immediately
20766 after the previous request. Some browsers pre-establish connections to a
20767 server in order to reduce the latency of a future request, and keep them
20768 pending until they need it. This delay will be reported as the idle time. A
20769 value of -1 indicates that nothing was received on the connection.
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020020770
20771 - TR: total time to get the client request (HTTP mode only). It's the time
20772 elapsed between the first bytes received and the moment the proxy received
20773 the empty line marking the end of the HTTP headers. The value "-1"
20774 indicates that the end of headers has never been seen. This happens when
20775 the client closes prematurely or times out. This time is usually very short
20776 since most requests fit in a single packet. A large time may indicate a
20777 request typed by hand during a test.
20778
20779 - Tq: total time to get the client request from the accept date or since the
20780 emission of the last byte of the previous response (HTTP mode only). It's
Davor Ocelice9ed2812017-12-25 17:49:28 +010020781 exactly equal to Th + Ti + TR unless any of them is -1, in which case it
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020020782 returns -1 as well. This timer used to be very useful before the arrival of
20783 HTTP keep-alive and browsers' pre-connect feature. It's recommended to drop
20784 it in favor of TR nowadays, as the idle time adds a lot of noise to the
20785 reports.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020786
20787 - Tw: total time spent in the queues waiting for a connection slot. It
20788 accounts for backend queue as well as the server queues, and depends on the
20789 queue size, and the time needed for the server to complete previous
20790 requests. The value "-1" means that the request was killed before reaching
20791 the queue, which is generally what happens with invalid or denied requests.
20792
20793 - Tc: total time to establish the TCP connection to the server. It's the time
20794 elapsed between the moment the proxy sent the connection request, and the
20795 moment it was acknowledged by the server, or between the TCP SYN packet and
20796 the matching SYN/ACK packet in return. The value "-1" means that the
20797 connection never established.
20798
20799 - Tr: server response time (HTTP mode only). It's the time elapsed between
20800 the moment the TCP connection was established to the server and the moment
20801 the server sent its complete response headers. It purely shows its request
20802 processing time, without the network overhead due to the data transmission.
20803 It is worth noting that when the client has data to send to the server, for
20804 instance during a POST request, the time already runs, and this can distort
20805 apparent response time. For this reason, it's generally wise not to trust
20806 too much this field for POST requests initiated from clients behind an
20807 untrusted network. A value of "-1" here means that the last the response
20808 header (empty line) was never seen, most likely because the server timeout
20809 stroke before the server managed to process the request.
20810
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020020811 - Ta: total active time for the HTTP request, between the moment the proxy
20812 received the first byte of the request header and the emission of the last
20813 byte of the response body. The exception is when the "logasap" option is
20814 specified. In this case, it only equals (TR+Tw+Tc+Tr), and is prefixed with
20815 a '+' sign. From this field, we can deduce "Td", the data transmission time,
20816 by subtracting other timers when valid :
20817
20818 Td = Ta - (TR + Tw + Tc + Tr)
20819
20820 Timers with "-1" values have to be excluded from this equation. Note that
20821 "Ta" can never be negative.
20822
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020823 - Tt: total session duration time, between the moment the proxy accepted it
20824 and the moment both ends were closed. The exception is when the "logasap"
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020020825 option is specified. In this case, it only equals (Th+Ti+TR+Tw+Tc+Tr), and
20826 is prefixed with a '+' sign. From this field, we can deduce "Td", the data
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030020827 transmission time, by subtracting other timers when valid :
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020828
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020020829 Td = Tt - (Th + Ti + TR + Tw + Tc + Tr)
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020830
20831 Timers with "-1" values have to be excluded from this equation. In TCP
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020020832 mode, "Ti", "Tq" and "Tr" have to be excluded too. Note that "Tt" can never
20833 be negative and that for HTTP, Tt is simply equal to (Th+Ti+Ta).
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020834
Damien Claisse57c8eb92020-04-28 12:09:19 +000020835 - Tu: total estimated time as seen from client, between the moment the proxy
20836 accepted it and the moment both ends were closed, without idle time.
20837 This is useful to roughly measure end-to-end time as a user would see it,
20838 without idle time pollution from keep-alive time between requests. This
20839 timer in only an estimation of time seen by user as it assumes network
20840 latency is the same in both directions. The exception is when the "logasap"
20841 option is specified. In this case, it only equals (Th+TR+Tw+Tc+Tr), and is
20842 prefixed with a '+' sign.
20843
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020844These timers provide precious indications on trouble causes. Since the TCP
20845protocol defines retransmit delays of 3, 6, 12... seconds, we know for sure
20846that timers close to multiples of 3s are nearly always related to lost packets
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020020847due to network problems (wires, negotiation, congestion). Moreover, if "Ta" or
20848"Tt" is close to a timeout value specified in the configuration, it often means
20849that a session has been aborted on timeout.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020850
20851Most common cases :
20852
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020020853 - If "Th" or "Ti" are close to 3000, a packet has probably been lost between
20854 the client and the proxy. This is very rare on local networks but might
20855 happen when clients are on far remote networks and send large requests. It
20856 may happen that values larger than usual appear here without any network
20857 cause. Sometimes, during an attack or just after a resource starvation has
Daniel Corbett9f0843f2021-05-08 10:50:37 -040020858 ended, HAProxy may accept thousands of connections in a few milliseconds.
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020020859 The time spent accepting these connections will inevitably slightly delay
20860 processing of other connections, and it can happen that request times in the
20861 order of a few tens of milliseconds are measured after a few thousands of
20862 new connections have been accepted at once. Using one of the keep-alive
20863 modes may display larger idle times since "Ti" measures the time spent
Patrick Mezard105faca2010-06-12 17:02:46 +020020864 waiting for additional requests.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020865
20866 - If "Tc" is close to 3000, a packet has probably been lost between the
20867 server and the proxy during the server connection phase. This value should
20868 always be very low, such as 1 ms on local networks and less than a few tens
20869 of ms on remote networks.
20870
Willy Tarreau55165fe2009-05-10 12:02:55 +020020871 - If "Tr" is nearly always lower than 3000 except some rare values which seem
20872 to be the average majored by 3000, there are probably some packets lost
20873 between the proxy and the server.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020874
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020020875 - If "Ta" is large even for small byte counts, it generally is because
20876 neither the client nor the server decides to close the connection while
Daniel Corbett9f0843f2021-05-08 10:50:37 -040020877 HAProxy is running in tunnel mode and both have agreed on a keep-alive
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020020878 connection mode. In order to solve this issue, it will be needed to specify
20879 one of the HTTP options to manipulate keep-alive or close options on either
20880 the frontend or the backend. Having the smallest possible 'Ta' or 'Tt' is
20881 important when connection regulation is used with the "maxconn" option on
20882 the servers, since no new connection will be sent to the server until
20883 another one is released.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020884
20885Other noticeable HTTP log cases ('xx' means any value to be ignored) :
20886
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020020887 TR/Tw/Tc/Tr/+Ta The "option logasap" is present on the frontend and the log
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020888 was emitted before the data phase. All the timers are valid
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020020889 except "Ta" which is shorter than reality.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020890
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020020891 -1/xx/xx/xx/Ta The client was not able to send a complete request in time
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020892 or it aborted too early. Check the session termination flags
20893 then "timeout http-request" and "timeout client" settings.
20894
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020020895 TR/-1/xx/xx/Ta It was not possible to process the request, maybe because
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020896 servers were out of order, because the request was invalid
20897 or forbidden by ACL rules. Check the session termination
20898 flags.
20899
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020020900 TR/Tw/-1/xx/Ta The connection could not establish on the server. Either it
20901 actively refused it or it timed out after Ta-(TR+Tw) ms.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020902 Check the session termination flags, then check the
20903 "timeout connect" setting. Note that the tarpit action might
20904 return similar-looking patterns, with "Tw" equal to the time
20905 the client connection was maintained open.
20906
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020020907 TR/Tw/Tc/-1/Ta The server has accepted the connection but did not return
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030020908 a complete response in time, or it closed its connection
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020020909 unexpectedly after Ta-(TR+Tw+Tc) ms. Check the session
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020910 termination flags, then check the "timeout server" setting.
20911
20912
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200209138.5. Session state at disconnection
20914-----------------------------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020915
20916TCP and HTTP logs provide a session termination indicator in the
20917"termination_state" field, just before the number of active connections. It is
209182-characters long in TCP mode, and is extended to 4 characters in HTTP mode,
20919each of which has a special meaning :
20920
20921 - On the first character, a code reporting the first event which caused the
20922 session to terminate :
20923
20924 C : the TCP session was unexpectedly aborted by the client.
20925
20926 S : the TCP session was unexpectedly aborted by the server, or the
20927 server explicitly refused it.
20928
20929 P : the session was prematurely aborted by the proxy, because of a
20930 connection limit enforcement, because a DENY filter was matched,
20931 because of a security check which detected and blocked a dangerous
20932 error in server response which might have caused information leak
Davor Ocelice9ed2812017-12-25 17:49:28 +010020933 (e.g. cacheable cookie).
Willy Tarreau570f2212013-06-10 16:42:09 +020020934
Daniel Corbett9f0843f2021-05-08 10:50:37 -040020935 L : the session was locally processed by HAProxy and was not passed to
Willy Tarreau570f2212013-06-10 16:42:09 +020020936 a server. This is what happens for stats and redirects.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020937
20938 R : a resource on the proxy has been exhausted (memory, sockets, source
20939 ports, ...). Usually, this appears during the connection phase, and
20940 system logs should contain a copy of the precise error. If this
20941 happens, it must be considered as a very serious anomaly which
20942 should be fixed as soon as possible by any means.
20943
20944 I : an internal error was identified by the proxy during a self-check.
20945 This should NEVER happen, and you are encouraged to report any log
20946 containing this, because this would almost certainly be a bug. It
20947 would be wise to preventively restart the process after such an
20948 event too, in case it would be caused by memory corruption.
20949
Daniel Corbett9f0843f2021-05-08 10:50:37 -040020950 D : the session was killed by HAProxy because the server was detected
Simon Horman752dc4a2011-06-21 14:34:59 +090020951 as down and was configured to kill all connections when going down.
20952
Daniel Corbett9f0843f2021-05-08 10:50:37 -040020953 U : the session was killed by HAProxy on this backup server because an
Justin Karnegeseb2c24a2012-05-24 15:28:52 -070020954 active server was detected as up and was configured to kill all
20955 backup connections when going up.
20956
Daniel Corbett9f0843f2021-05-08 10:50:37 -040020957 K : the session was actively killed by an admin operating on HAProxy.
Willy Tarreaua2a64e92011-09-07 23:01:56 +020020958
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020959 c : the client-side timeout expired while waiting for the client to
20960 send or receive data.
20961
20962 s : the server-side timeout expired while waiting for the server to
20963 send or receive data.
20964
20965 - : normal session completion, both the client and the server closed
20966 with nothing left in the buffers.
20967
20968 - on the second character, the TCP or HTTP session state when it was closed :
20969
Willy Tarreauf7b30a92010-12-06 22:59:17 +010020970 R : the proxy was waiting for a complete, valid REQUEST from the client
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020971 (HTTP mode only). Nothing was sent to any server.
20972
20973 Q : the proxy was waiting in the QUEUE for a connection slot. This can
20974 only happen when servers have a 'maxconn' parameter set. It can
20975 also happen in the global queue after a redispatch consecutive to
20976 a failed attempt to connect to a dying server. If no redispatch is
20977 reported, then no connection attempt was made to any server.
20978
20979 C : the proxy was waiting for the CONNECTION to establish on the
20980 server. The server might at most have noticed a connection attempt.
20981
20982 H : the proxy was waiting for complete, valid response HEADERS from the
20983 server (HTTP only).
20984
20985 D : the session was in the DATA phase.
20986
20987 L : the proxy was still transmitting LAST data to the client while the
20988 server had already finished. This one is very rare as it can only
20989 happen when the client dies while receiving the last packets.
20990
20991 T : the request was tarpitted. It has been held open with the client
20992 during the whole "timeout tarpit" duration or until the client
20993 closed, both of which will be reported in the "Tw" timer.
20994
20995 - : normal session completion after end of data transfer.
20996
20997 - the third character tells whether the persistence cookie was provided by
20998 the client (only in HTTP mode) :
20999
21000 N : the client provided NO cookie. This is usually the case for new
21001 visitors, so counting the number of occurrences of this flag in the
21002 logs generally indicate a valid trend for the site frequentation.
21003
21004 I : the client provided an INVALID cookie matching no known server.
21005 This might be caused by a recent configuration change, mixed
Cyril Bontéa8e7bbc2010-04-25 22:29:29 +020021006 cookies between HTTP/HTTPS sites, persistence conditionally
21007 ignored, or an attack.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010021008
21009 D : the client provided a cookie designating a server which was DOWN,
21010 so either "option persist" was used and the client was sent to
21011 this server, or it was not set and the client was redispatched to
21012 another server.
21013
Willy Tarreau996a92c2010-10-13 19:30:47 +020021014 V : the client provided a VALID cookie, and was sent to the associated
Willy Tarreaucc6c8912009-02-22 10:53:55 +010021015 server.
21016
Willy Tarreau996a92c2010-10-13 19:30:47 +020021017 E : the client provided a valid cookie, but with a last date which was
21018 older than what is allowed by the "maxidle" cookie parameter, so
21019 the cookie is consider EXPIRED and is ignored. The request will be
21020 redispatched just as if there was no cookie.
21021
21022 O : the client provided a valid cookie, but with a first date which was
21023 older than what is allowed by the "maxlife" cookie parameter, so
21024 the cookie is consider too OLD and is ignored. The request will be
21025 redispatched just as if there was no cookie.
21026
Willy Tarreauc89ccb62012-04-05 21:18:22 +020021027 U : a cookie was present but was not used to select the server because
21028 some other server selection mechanism was used instead (typically a
21029 "use-server" rule).
21030
Willy Tarreaucc6c8912009-02-22 10:53:55 +010021031 - : does not apply (no cookie set in configuration).
21032
21033 - the last character reports what operations were performed on the persistence
21034 cookie returned by the server (only in HTTP mode) :
21035
21036 N : NO cookie was provided by the server, and none was inserted either.
21037
21038 I : no cookie was provided by the server, and the proxy INSERTED one.
21039 Note that in "cookie insert" mode, if the server provides a cookie,
21040 it will still be overwritten and reported as "I" here.
21041
Willy Tarreau996a92c2010-10-13 19:30:47 +020021042 U : the proxy UPDATED the last date in the cookie that was presented by
21043 the client. This can only happen in insert mode with "maxidle". It
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030021044 happens every time there is activity at a different date than the
Willy Tarreau996a92c2010-10-13 19:30:47 +020021045 date indicated in the cookie. If any other change happens, such as
21046 a redispatch, then the cookie will be marked as inserted instead.
21047
Willy Tarreaucc6c8912009-02-22 10:53:55 +010021048 P : a cookie was PROVIDED by the server and transmitted as-is.
21049
21050 R : the cookie provided by the server was REWRITTEN by the proxy, which
21051 happens in "cookie rewrite" or "cookie prefix" modes.
21052
21053 D : the cookie provided by the server was DELETED by the proxy.
21054
21055 - : does not apply (no cookie set in configuration).
21056
Willy Tarreau996a92c2010-10-13 19:30:47 +020021057The combination of the two first flags gives a lot of information about what
21058was happening when the session terminated, and why it did terminate. It can be
Willy Tarreaucc6c8912009-02-22 10:53:55 +010021059helpful to detect server saturation, network troubles, local system resource
21060starvation, attacks, etc...
21061
21062The most common termination flags combinations are indicated below. They are
21063alphabetically sorted, with the lowercase set just after the upper case for
21064easier finding and understanding.
21065
21066 Flags Reason
21067
21068 -- Normal termination.
21069
21070 CC The client aborted before the connection could be established to the
Daniel Corbett9f0843f2021-05-08 10:50:37 -040021071 server. This can happen when HAProxy tries to connect to a recently
21072 dead (or unchecked) server, and the client aborts while HAProxy is
Willy Tarreaucc6c8912009-02-22 10:53:55 +010021073 waiting for the server to respond or for "timeout connect" to expire.
21074
21075 CD The client unexpectedly aborted during data transfer. This can be
21076 caused by a browser crash, by an intermediate equipment between the
Daniel Corbett9f0843f2021-05-08 10:50:37 -040021077 client and HAProxy which decided to actively break the connection,
21078 by network routing issues between the client and HAProxy, or by a
Willy Tarreaucc6c8912009-02-22 10:53:55 +010021079 keep-alive session between the server and the client terminated first
21080 by the client.
Willy Tarreaud72758d2010-01-12 10:42:19 +010021081
Willy Tarreaucc6c8912009-02-22 10:53:55 +010021082 cD The client did not send nor acknowledge any data for as long as the
21083 "timeout client" delay. This is often caused by network failures on
Cyril Bontédc4d9032012-04-08 21:57:39 +020021084 the client side, or the client simply leaving the net uncleanly.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010021085
21086 CH The client aborted while waiting for the server to start responding.
21087 It might be the server taking too long to respond or the client
21088 clicking the 'Stop' button too fast.
21089
21090 cH The "timeout client" stroke while waiting for client data during a
21091 POST request. This is sometimes caused by too large TCP MSS values
21092 for PPPoE networks which cannot transport full-sized packets. It can
21093 also happen when client timeout is smaller than server timeout and
21094 the server takes too long to respond.
21095
21096 CQ The client aborted while its session was queued, waiting for a server
21097 with enough empty slots to accept it. It might be that either all the
21098 servers were saturated or that the assigned server was taking too
21099 long a time to respond.
21100
21101 CR The client aborted before sending a full HTTP request. Most likely
21102 the request was typed by hand using a telnet client, and aborted
21103 too early. The HTTP status code is likely a 400 here. Sometimes this
Daniel Corbett9f0843f2021-05-08 10:50:37 -040021104 might also be caused by an IDS killing the connection between HAProxy
Willy Tarreau0f228a02015-05-01 15:37:53 +020021105 and the client. "option http-ignore-probes" can be used to ignore
21106 connections without any data transfer.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010021107
21108 cR The "timeout http-request" stroke before the client sent a full HTTP
21109 request. This is sometimes caused by too large TCP MSS values on the
21110 client side for PPPoE networks which cannot transport full-sized
21111 packets, or by clients sending requests by hand and not typing fast
21112 enough, or forgetting to enter the empty line at the end of the
Willy Tarreau2705a612014-05-23 17:38:34 +020021113 request. The HTTP status code is likely a 408 here. Note: recently,
Willy Tarreau0f228a02015-05-01 15:37:53 +020021114 some browsers started to implement a "pre-connect" feature consisting
21115 in speculatively connecting to some recently visited web sites just
21116 in case the user would like to visit them. This results in many
21117 connections being established to web sites, which end up in 408
21118 Request Timeout if the timeout strikes first, or 400 Bad Request when
21119 the browser decides to close them first. These ones pollute the log
21120 and feed the error counters. Some versions of some browsers have even
21121 been reported to display the error code. It is possible to work
Davor Ocelice9ed2812017-12-25 17:49:28 +010021122 around the undesirable effects of this behavior by adding "option
Willy Tarreau0f228a02015-05-01 15:37:53 +020021123 http-ignore-probes" in the frontend, resulting in connections with
21124 zero data transfer to be totally ignored. This will definitely hide
21125 the errors of people experiencing connectivity issues though.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010021126
21127 CT The client aborted while its session was tarpitted. It is important to
21128 check if this happens on valid requests, in order to be sure that no
Willy Tarreau55165fe2009-05-10 12:02:55 +020021129 wrong tarpit rules have been written. If a lot of them happen, it
21130 might make sense to lower the "timeout tarpit" value to something
21131 closer to the average reported "Tw" timer, in order not to consume
21132 resources for just a few attackers.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010021133
Daniel Corbett9f0843f2021-05-08 10:50:37 -040021134 LR The request was intercepted and locally handled by HAProxy. Generally
Willy Tarreau570f2212013-06-10 16:42:09 +020021135 it means that this was a redirect or a stats request.
21136
Daniel Corbett9f0843f2021-05-08 10:50:37 -040021137 SC The server or an equipment between it and HAProxy explicitly refused
Willy Tarreaucc6c8912009-02-22 10:53:55 +010021138 the TCP connection (the proxy received a TCP RST or an ICMP message
21139 in return). Under some circumstances, it can also be the network
Davor Ocelice9ed2812017-12-25 17:49:28 +010021140 stack telling the proxy that the server is unreachable (e.g. no route,
Willy Tarreaucc6c8912009-02-22 10:53:55 +010021141 or no ARP response on local network). When this happens in HTTP mode,
21142 the status code is likely a 502 or 503 here.
21143
21144 sC The "timeout connect" stroke before a connection to the server could
21145 complete. When this happens in HTTP mode, the status code is likely a
21146 503 or 504 here.
21147
21148 SD The connection to the server died with an error during the data
Daniel Corbett9f0843f2021-05-08 10:50:37 -040021149 transfer. This usually means that HAProxy has received an RST from
Willy Tarreaucc6c8912009-02-22 10:53:55 +010021150 the server or an ICMP message from an intermediate equipment while
21151 exchanging data with the server. This can be caused by a server crash
21152 or by a network issue on an intermediate equipment.
21153
21154 sD The server did not send nor acknowledge any data for as long as the
21155 "timeout server" setting during the data phase. This is often caused
Davor Ocelice9ed2812017-12-25 17:49:28 +010021156 by too short timeouts on L4 equipment before the server (firewalls,
Willy Tarreaucc6c8912009-02-22 10:53:55 +010021157 load-balancers, ...), as well as keep-alive sessions maintained
Daniel Corbett9f0843f2021-05-08 10:50:37 -040021158 between the client and the server expiring first on HAProxy.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010021159
21160 SH The server aborted before sending its full HTTP response headers, or
21161 it crashed while processing the request. Since a server aborting at
21162 this moment is very rare, it would be wise to inspect its logs to
21163 control whether it crashed and why. The logged request may indicate a
21164 small set of faulty requests, demonstrating bugs in the application.
21165 Sometimes this might also be caused by an IDS killing the connection
Daniel Corbett9f0843f2021-05-08 10:50:37 -040021166 between HAProxy and the server.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010021167
21168 sH The "timeout server" stroke before the server could return its
21169 response headers. This is the most common anomaly, indicating too
21170 long transactions, probably caused by server or database saturation.
21171 The immediate workaround consists in increasing the "timeout server"
21172 setting, but it is important to keep in mind that the user experience
21173 will suffer from these long response times. The only long term
21174 solution is to fix the application.
21175
21176 sQ The session spent too much time in queue and has been expired. See
21177 the "timeout queue" and "timeout connect" settings to find out how to
21178 fix this if it happens too often. If it often happens massively in
21179 short periods, it may indicate general problems on the affected
21180 servers due to I/O or database congestion, or saturation caused by
21181 external attacks.
21182
21183 PC The proxy refused to establish a connection to the server because the
Thayne McCombscdbcca92021-01-07 21:24:41 -070021184 process's socket limit has been reached while attempting to connect.
Cyril Bontédc4d9032012-04-08 21:57:39 +020021185 The global "maxconn" parameter may be increased in the configuration
Willy Tarreaucc6c8912009-02-22 10:53:55 +010021186 so that it does not happen anymore. This status is very rare and
21187 might happen when the global "ulimit-n" parameter is forced by hand.
21188
Willy Tarreaued2fd2d2010-12-29 11:23:27 +010021189 PD The proxy blocked an incorrectly formatted chunked encoded message in
21190 a request or a response, after the server has emitted its headers. In
21191 most cases, this will indicate an invalid message from the server to
Davor Ocelice9ed2812017-12-25 17:49:28 +010021192 the client. HAProxy supports chunk sizes of up to 2GB - 1 (2147483647
Willy Tarreauf3a3e132013-08-31 08:16:26 +020021193 bytes). Any larger size will be considered as an error.
Willy Tarreaued2fd2d2010-12-29 11:23:27 +010021194
Willy Tarreaucc6c8912009-02-22 10:53:55 +010021195 PH The proxy blocked the server's response, because it was invalid,
21196 incomplete, dangerous (cache control), or matched a security filter.
21197 In any case, an HTTP 502 error is sent to the client. One possible
21198 cause for this error is an invalid syntax in an HTTP header name
Willy Tarreaued2fd2d2010-12-29 11:23:27 +010021199 containing unauthorized characters. It is also possible but quite
21200 rare, that the proxy blocked a chunked-encoding request from the
21201 client due to an invalid syntax, before the server responded. In this
21202 case, an HTTP 400 error is sent to the client and reported in the
21203 logs.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010021204
21205 PR The proxy blocked the client's HTTP request, either because of an
21206 invalid HTTP syntax, in which case it returned an HTTP 400 error to
21207 the client, or because a deny filter matched, in which case it
21208 returned an HTTP 403 error.
21209
21210 PT The proxy blocked the client's request and has tarpitted its
21211 connection before returning it a 500 server error. Nothing was sent
21212 to the server. The connection was maintained open for as long as
21213 reported by the "Tw" timer field.
21214
21215 RC A local resource has been exhausted (memory, sockets, source ports)
21216 preventing the connection to the server from establishing. The error
21217 logs will tell precisely what was missing. This is very rare and can
21218 only be solved by proper system tuning.
21219
Willy Tarreau996a92c2010-10-13 19:30:47 +020021220The combination of the two last flags gives a lot of information about how
Daniel Corbett9f0843f2021-05-08 10:50:37 -040021221persistence was handled by the client, the server and by HAProxy. This is very
Willy Tarreau996a92c2010-10-13 19:30:47 +020021222important to troubleshoot disconnections, when users complain they have to
21223re-authenticate. The commonly encountered flags are :
21224
21225 -- Persistence cookie is not enabled.
21226
21227 NN No cookie was provided by the client, none was inserted in the
21228 response. For instance, this can be in insert mode with "postonly"
21229 set on a GET request.
21230
21231 II A cookie designating an invalid server was provided by the client,
21232 a valid one was inserted in the response. This typically happens when
Jamie Gloudonaaa21002012-08-25 00:18:33 -040021233 a "server" entry is removed from the configuration, since its cookie
Willy Tarreau996a92c2010-10-13 19:30:47 +020021234 value can be presented by a client when no other server knows it.
21235
21236 NI No cookie was provided by the client, one was inserted in the
21237 response. This typically happens for first requests from every user
21238 in "insert" mode, which makes it an easy way to count real users.
21239
21240 VN A cookie was provided by the client, none was inserted in the
21241 response. This happens for most responses for which the client has
21242 already got a cookie.
21243
21244 VU A cookie was provided by the client, with a last visit date which is
21245 not completely up-to-date, so an updated cookie was provided in
21246 response. This can also happen if there was no date at all, or if
21247 there was a date but the "maxidle" parameter was not set, so that the
21248 cookie can be switched to unlimited time.
21249
21250 EI A cookie was provided by the client, with a last visit date which is
21251 too old for the "maxidle" parameter, so the cookie was ignored and a
21252 new cookie was inserted in the response.
21253
21254 OI A cookie was provided by the client, with a first visit date which is
21255 too old for the "maxlife" parameter, so the cookie was ignored and a
21256 new cookie was inserted in the response.
21257
21258 DI The server designated by the cookie was down, a new server was
21259 selected and a new cookie was emitted in the response.
21260
21261 VI The server designated by the cookie was not marked dead but could not
21262 be reached. A redispatch happened and selected another one, which was
21263 then advertised in the response.
21264
Willy Tarreaucc6c8912009-02-22 10:53:55 +010021265
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200212668.6. Non-printable characters
21267-----------------------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010021268
21269In order not to cause trouble to log analysis tools or terminals during log
21270consulting, non-printable characters are not sent as-is into log files, but are
21271converted to the two-digits hexadecimal representation of their ASCII code,
21272prefixed by the character '#'. The only characters that can be logged without
21273being escaped are comprised between 32 and 126 (inclusive). Obviously, the
21274escape character '#' itself is also encoded to avoid any ambiguity ("#23"). It
21275is the same for the character '"' which becomes "#22", as well as '{', '|' and
21276'}' when logging headers.
21277
21278Note that the space character (' ') is not encoded in headers, which can cause
21279issues for tools relying on space count to locate fields. A typical header
21280containing spaces is "User-Agent".
21281
21282Last, it has been observed that some syslog daemons such as syslog-ng escape
21283the quote ('"') with a backslash ('\'). The reverse operation can safely be
21284performed since no quote may appear anywhere else in the logs.
21285
21286
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200212878.7. Capturing HTTP cookies
21288---------------------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010021289
21290Cookie capture simplifies the tracking a complete user session. This can be
21291achieved using the "capture cookie" statement in the frontend. Please refer to
Willy Tarreauc57f0e22009-05-10 13:12:33 +020021292section 4.2 for more details. Only one cookie can be captured, and the same
Willy Tarreaucc6c8912009-02-22 10:53:55 +010021293cookie will simultaneously be checked in the request ("Cookie:" header) and in
21294the response ("Set-Cookie:" header). The respective values will be reported in
21295the HTTP logs at the "captured_request_cookie" and "captured_response_cookie"
Willy Tarreauc57f0e22009-05-10 13:12:33 +020021296locations (see section 8.2.3 about HTTP log format). When either cookie is
Willy Tarreaucc6c8912009-02-22 10:53:55 +010021297not seen, a dash ('-') replaces the value. This way, it's easy to detect when a
21298user switches to a new session for example, because the server will reassign it
21299a new cookie. It is also possible to detect if a server unexpectedly sets a
21300wrong cookie to a client, leading to session crossing.
21301
21302 Examples :
21303 # capture the first cookie whose name starts with "ASPSESSION"
21304 capture cookie ASPSESSION len 32
21305
21306 # capture the first cookie whose name is exactly "vgnvisitor"
21307 capture cookie vgnvisitor= len 32
21308
21309
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200213108.8. Capturing HTTP headers
21311---------------------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010021312
21313Header captures are useful to track unique request identifiers set by an upper
21314proxy, virtual host names, user-agents, POST content-length, referrers, etc. In
21315the response, one can search for information about the response length, how the
21316server asked the cache to behave, or an object location during a redirection.
21317
21318Header captures are performed using the "capture request header" and "capture
21319response header" statements in the frontend. Please consult their definition in
Willy Tarreauc57f0e22009-05-10 13:12:33 +020021320section 4.2 for more details.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010021321
21322It is possible to include both request headers and response headers at the same
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +010021323time. Non-existent headers are logged as empty strings, and if one header
21324appears more than once, only its last occurrence will be logged. Request headers
Willy Tarreaucc6c8912009-02-22 10:53:55 +010021325are grouped within braces '{' and '}' in the same order as they were declared,
21326and delimited with a vertical bar '|' without any space. Response headers
21327follow the same representation, but are displayed after a space following the
21328request headers block. These blocks are displayed just before the HTTP request
21329in the logs.
21330
Willy Tarreaud9ed3d22014-06-13 12:23:06 +020021331As a special case, it is possible to specify an HTTP header capture in a TCP
21332frontend. The purpose is to enable logging of headers which will be parsed in
21333an HTTP backend if the request is then switched to this HTTP backend.
21334
Willy Tarreaucc6c8912009-02-22 10:53:55 +010021335 Example :
21336 # This instance chains to the outgoing proxy
21337 listen proxy-out
21338 mode http
21339 option httplog
21340 option logasap
21341 log global
21342 server cache1 192.168.1.1:3128
21343
21344 # log the name of the virtual server
21345 capture request header Host len 20
21346
21347 # log the amount of data uploaded during a POST
21348 capture request header Content-Length len 10
21349
21350 # log the beginning of the referrer
21351 capture request header Referer len 20
21352
21353 # server name (useful for outgoing proxies only)
21354 capture response header Server len 20
21355
21356 # logging the content-length is useful with "option logasap"
21357 capture response header Content-Length len 10
21358
Davor Ocelice9ed2812017-12-25 17:49:28 +010021359 # log the expected cache behavior on the response
Willy Tarreaucc6c8912009-02-22 10:53:55 +010021360 capture response header Cache-Control len 8
21361
21362 # the Via header will report the next proxy's name
21363 capture response header Via len 20
21364
21365 # log the URL location during a redirection
21366 capture response header Location len 20
21367
21368 >>> Aug 9 20:26:09 localhost \
21369 haproxy[2022]: 127.0.0.1:34014 [09/Aug/2004:20:26:09] proxy-out \
21370 proxy-out/cache1 0/0/0/162/+162 200 +350 - - ---- 0/0/0/0/0 0/0 \
21371 {fr.adserver.yahoo.co||http://fr.f416.mail.} {|864|private||} \
21372 "GET http://fr.adserver.yahoo.com/"
21373
21374 >>> Aug 9 20:30:46 localhost \
21375 haproxy[2022]: 127.0.0.1:34020 [09/Aug/2004:20:30:46] proxy-out \
21376 proxy-out/cache1 0/0/0/182/+182 200 +279 - - ---- 0/0/0/0/0 0/0 \
21377 {w.ods.org||} {Formilux/0.1.8|3495|||} \
Willy Tarreaud72758d2010-01-12 10:42:19 +010021378 "GET http://trafic.1wt.eu/ HTTP/1.1"
Willy Tarreaucc6c8912009-02-22 10:53:55 +010021379
21380 >>> Aug 9 20:30:46 localhost \
21381 haproxy[2022]: 127.0.0.1:34028 [09/Aug/2004:20:30:46] proxy-out \
21382 proxy-out/cache1 0/0/2/126/+128 301 +223 - - ---- 0/0/0/0/0 0/0 \
21383 {www.sytadin.equipement.gouv.fr||http://trafic.1wt.eu/} \
21384 {Apache|230|||http://www.sytadin.} \
Willy Tarreaud72758d2010-01-12 10:42:19 +010021385 "GET http://www.sytadin.equipement.gouv.fr/ HTTP/1.1"
Willy Tarreaucc6c8912009-02-22 10:53:55 +010021386
21387
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200213888.9. Examples of logs
21389---------------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010021390
21391These are real-world examples of logs accompanied with an explanation. Some of
21392them have been made up by hand. The syslog part has been removed for better
21393reading. Their sole purpose is to explain how to decipher them.
21394
21395 >>> haproxy[674]: 127.0.0.1:33318 [15/Oct/2003:08:31:57.130] px-http \
21396 px-http/srv1 6559/0/7/147/6723 200 243 - - ---- 5/3/3/1/0 0/0 \
21397 "HEAD / HTTP/1.0"
21398
21399 => long request (6.5s) entered by hand through 'telnet'. The server replied
21400 in 147 ms, and the session ended normally ('----')
21401
21402 >>> haproxy[674]: 127.0.0.1:33319 [15/Oct/2003:08:31:57.149] px-http \
21403 px-http/srv1 6559/1230/7/147/6870 200 243 - - ---- 324/239/239/99/0 \
21404 0/9 "HEAD / HTTP/1.0"
21405
21406 => Idem, but the request was queued in the global queue behind 9 other
21407 requests, and waited there for 1230 ms.
21408
21409 >>> haproxy[674]: 127.0.0.1:33320 [15/Oct/2003:08:32:17.654] px-http \
21410 px-http/srv1 9/0/7/14/+30 200 +243 - - ---- 3/3/3/1/0 0/0 \
21411 "GET /image.iso HTTP/1.0"
21412
21413 => request for a long data transfer. The "logasap" option was specified, so
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +010021414 the log was produced just before transferring data. The server replied in
Willy Tarreaucc6c8912009-02-22 10:53:55 +010021415 14 ms, 243 bytes of headers were sent to the client, and total time from
21416 accept to first data byte is 30 ms.
21417
21418 >>> haproxy[674]: 127.0.0.1:33320 [15/Oct/2003:08:32:17.925] px-http \
21419 px-http/srv1 9/0/7/14/30 502 243 - - PH-- 3/2/2/0/0 0/0 \
21420 "GET /cgi-bin/bug.cgi? HTTP/1.0"
21421
Christopher Faulet87f1f3d2019-07-18 14:51:20 +020021422 => the proxy blocked a server response either because of an "http-response
21423 deny" rule, or because the response was improperly formatted and not
21424 HTTP-compliant, or because it blocked sensitive information which risked
21425 being cached. In this case, the response is replaced with a "502 bad
Daniel Corbett9f0843f2021-05-08 10:50:37 -040021426 gateway". The flags ("PH--") tell us that it was HAProxy who decided to
Christopher Faulet87f1f3d2019-07-18 14:51:20 +020021427 return the 502 and not the server.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010021428
21429 >>> haproxy[18113]: 127.0.0.1:34548 [15/Oct/2003:15:18:55.798] px-http \
Willy Tarreaud72758d2010-01-12 10:42:19 +010021430 px-http/<NOSRV> -1/-1/-1/-1/8490 -1 0 - - CR-- 2/2/2/0/0 0/0 ""
Willy Tarreaucc6c8912009-02-22 10:53:55 +010021431
21432 => the client never completed its request and aborted itself ("C---") after
21433 8.5s, while the proxy was waiting for the request headers ("-R--").
21434 Nothing was sent to any server.
21435
21436 >>> haproxy[18113]: 127.0.0.1:34549 [15/Oct/2003:15:19:06.103] px-http \
21437 px-http/<NOSRV> -1/-1/-1/-1/50001 408 0 - - cR-- 2/2/2/0/0 0/0 ""
21438
21439 => The client never completed its request, which was aborted by the
21440 time-out ("c---") after 50s, while the proxy was waiting for the request
Davor Ocelice9ed2812017-12-25 17:49:28 +010021441 headers ("-R--"). Nothing was sent to any server, but the proxy could
Willy Tarreaucc6c8912009-02-22 10:53:55 +010021442 send a 408 return code to the client.
21443
21444 >>> haproxy[18989]: 127.0.0.1:34550 [15/Oct/2003:15:24:28.312] px-tcp \
21445 px-tcp/srv1 0/0/5007 0 cD 0/0/0/0/0 0/0
21446
21447 => This log was produced with "option tcplog". The client timed out after
21448 5 seconds ("c----").
21449
21450 >>> haproxy[18989]: 10.0.0.1:34552 [15/Oct/2003:15:26:31.462] px-http \
21451 px-http/srv1 3183/-1/-1/-1/11215 503 0 - - SC-- 205/202/202/115/3 \
Willy Tarreaud72758d2010-01-12 10:42:19 +010021452 0/0 "HEAD / HTTP/1.0"
Willy Tarreaucc6c8912009-02-22 10:53:55 +010021453
21454 => The request took 3s to complete (probably a network problem), and the
Willy Tarreauc57f0e22009-05-10 13:12:33 +020021455 connection to the server failed ('SC--') after 4 attempts of 2 seconds
Willy Tarreaucc6c8912009-02-22 10:53:55 +010021456 (config says 'retries 3'), and no redispatch (otherwise we would have
21457 seen "/+3"). Status code 503 was returned to the client. There were 115
21458 connections on this server, 202 connections on this proxy, and 205 on
21459 the global process. It is possible that the server refused the
21460 connection because of too many already established.
Willy Tarreau844e3c52008-01-11 16:28:18 +010021461
Willy Tarreau52b2d222011-09-07 23:48:48 +020021462
Christopher Fauletc3fe5332016-04-07 15:30:10 +0200214639. Supported filters
21464--------------------
21465
21466Here are listed officially supported filters with the list of parameters they
21467accept. Depending on compile options, some of these filters might be
21468unavailable. The list of available filters is reported in haproxy -vv.
21469
21470See also : "filter"
21471
214729.1. Trace
21473----------
21474
Christopher Fauletc41d8bd2020-11-17 10:43:26 +010021475filter trace [name <name>] [random-forwarding] [hexdump]
Christopher Fauletc3fe5332016-04-07 15:30:10 +020021476
21477 Arguments:
21478 <name> is an arbitrary name that will be reported in
21479 messages. If no name is provided, "TRACE" is used.
21480
Christopher Faulet96a577a2020-11-17 10:45:05 +010021481 <quiet> inhibits trace messages.
Christopher Fauletc3fe5332016-04-07 15:30:10 +020021482
Davor Ocelice9ed2812017-12-25 17:49:28 +010021483 <random-forwarding> enables the random forwarding of parsed data. By
Christopher Fauletc3fe5332016-04-07 15:30:10 +020021484 default, this filter forwards all previously parsed
21485 data. With this parameter, it only forwards a random
21486 amount of the parsed data.
21487
Davor Ocelice9ed2812017-12-25 17:49:28 +010021488 <hexdump> dumps all forwarded data to the server and the client.
Christopher Faulet31bfe1f2016-12-09 17:42:38 +010021489
Christopher Fauletc3fe5332016-04-07 15:30:10 +020021490This filter can be used as a base to develop new filters. It defines all
21491callbacks and print a message on the standard error stream (stderr) with useful
21492information for all of them. It may be useful to debug the activity of other
21493filters or, quite simply, HAProxy's activity.
21494
21495Using <random-parsing> and/or <random-forwarding> parameters is a good way to
21496tests the behavior of a filter that parses data exchanged between a client and
21497a server by adding some latencies in the processing.
21498
21499
215009.2. HTTP compression
21501---------------------
21502
21503filter compression
21504
21505The HTTP compression has been moved in a filter in HAProxy 1.7. "compression"
21506keyword must still be used to enable and configure the HTTP compression. And
Christopher Fauletb30b3102019-09-12 23:03:09 +020021507when no other filter is used, it is enough. When used with the cache or the
21508fcgi-app enabled, it is also enough. In this case, the compression is always
21509done after the response is stored in the cache. But it is mandatory to
21510explicitly use a filter line to enable the HTTP compression when at least one
21511filter other than the cache or the fcgi-app is used for the same
21512listener/frontend/backend. This is important to know the filters evaluation
21513order.
Christopher Fauletc3fe5332016-04-07 15:30:10 +020021514
Christopher Fauletb30b3102019-09-12 23:03:09 +020021515See also : "compression", section 9.4 about the cache filter and section 9.5
21516 about the fcgi-app filter.
Christopher Fauletc3fe5332016-04-07 15:30:10 +020021517
21518
Christopher Fauletf7e4e7e2016-10-27 22:29:49 +0200215199.3. Stream Processing Offload Engine (SPOE)
21520--------------------------------------------
21521
21522filter spoe [engine <name>] config <file>
21523
21524 Arguments :
21525
21526 <name> is the engine name that will be used to find the right scope in
21527 the configuration file. If not provided, all the file will be
21528 parsed.
21529
21530 <file> is the path of the engine configuration file. This file can
21531 contain configuration of several engines. In this case, each
21532 part must be placed in its own scope.
21533
21534The Stream Processing Offload Engine (SPOE) is a filter communicating with
21535external components. It allows the offload of some specifics processing on the
Davor Ocelice9ed2812017-12-25 17:49:28 +010021536streams in tiered applications. These external components and information
Christopher Fauletf7e4e7e2016-10-27 22:29:49 +020021537exchanged with them are configured in dedicated files, for the main part. It
21538also requires dedicated backends, defined in HAProxy configuration.
21539
21540SPOE communicates with external components using an in-house binary protocol,
21541the Stream Processing Offload Protocol (SPOP).
21542
Tim Düsterhus4896c442016-11-29 02:15:19 +010021543For all information about the SPOE configuration and the SPOP specification, see
Christopher Fauletf7e4e7e2016-10-27 22:29:49 +020021544"doc/SPOE.txt".
21545
Christopher Faulet99a17a22018-12-11 09:18:27 +0100215469.4. Cache
21547----------
21548
21549filter cache <name>
21550
21551 Arguments :
21552
21553 <name> is name of the cache section this filter will use.
21554
21555The cache uses a filter to store cacheable responses. The HTTP rules
21556"cache-store" and "cache-use" must be used to define how and when to use a
John Roeslerfb2fce12019-07-10 15:45:51 -050021557cache. By default the corresponding filter is implicitly defined. And when no
Christopher Fauletb30b3102019-09-12 23:03:09 +020021558other filters than fcgi-app or compression are used, it is enough. In such
21559case, the compression filter is always evaluated after the cache filter. But it
21560is mandatory to explicitly use a filter line to use a cache when at least one
21561filter other than the compression or the fcgi-app is used for the same
Christopher Faulet27d93c32018-12-15 22:32:02 +010021562listener/frontend/backend. This is important to know the filters evaluation
21563order.
Christopher Faulet99a17a22018-12-11 09:18:27 +010021564
Christopher Fauletb30b3102019-09-12 23:03:09 +020021565See also : section 9.2 about the compression filter, section 9.5 about the
21566 fcgi-app filter and section 6 about cache.
21567
21568
215699.5. Fcgi-app
21570-------------
21571
Daniel Corbett67a82712020-07-06 23:01:19 -040021572filter fcgi-app <name>
Christopher Fauletb30b3102019-09-12 23:03:09 +020021573
21574 Arguments :
21575
21576 <name> is name of the fcgi-app section this filter will use.
21577
21578The FastCGI application uses a filter to evaluate all custom parameters on the
21579request path, and to process the headers on the response path. the <name> must
21580reference an existing fcgi-app section. The directive "use-fcgi-app" should be
21581used to define the application to use. By default the corresponding filter is
21582implicitly defined. And when no other filters than cache or compression are
21583used, it is enough. But it is mandatory to explicitly use a filter line to a
21584fcgi-app when at least one filter other than the compression or the cache is
21585used for the same backend. This is important to know the filters evaluation
21586order.
21587
21588See also: "use-fcgi-app", section 9.2 about the compression filter, section 9.4
21589 about the cache filter and section 10 about FastCGI application.
21590
21591
Miroslav Zagoracdc32cd92020-12-13 18:32:57 +0100215929.6. OpenTracing
21593----------------
21594
21595The OpenTracing filter adds native support for using distributed tracing in
21596HAProxy. This is enabled by sending an OpenTracing compliant request to one
21597of the supported tracers such as Datadog, Jaeger, Lightstep and Zipkin tracers.
21598Please note: tracers are not listed by any preference, but alphabetically.
21599
Daniel Corbett9f0843f2021-05-08 10:50:37 -040021600This feature is only enabled when HAProxy was built with USE_OT=1.
Miroslav Zagoracdc32cd92020-12-13 18:32:57 +010021601
21602The OpenTracing filter activation is done explicitly by specifying it in the
21603HAProxy configuration. If this is not done, the OpenTracing filter in no way
21604participates in the work of HAProxy.
21605
21606filter opentracing [id <id>] config <file>
21607
21608 Arguments :
21609
21610 <id> is the OpenTracing filter id that will be used to find the
21611 right scope in the configuration file. If no filter id is
21612 specified, 'ot-filter' is used as default. If scope is not
21613 specified in the configuration file, it applies to all defined
21614 OpenTracing filters.
21615
21616 <file> is the path of the OpenTracing configuration file. The same
21617 file can contain configurations for multiple OpenTracing
21618 filters simultaneously. In that case we do not need to define
21619 scope so the same configuration applies to all filters or each
21620 filter must have its own scope defined.
21621
21622More detailed documentation related to the operation, configuration and use
Willy Tarreaua63d1a02021-04-02 17:16:46 +020021623of the filter can be found in the addons/ot directory.
Miroslav Zagoracdc32cd92020-12-13 18:32:57 +010021624
21625
Christopher Fauletb30b3102019-09-12 23:03:09 +02002162610. FastCGI applications
21627-------------------------
21628
21629HAProxy is able to send HTTP requests to Responder FastCGI applications. This
21630feature was added in HAProxy 2.1. To do so, servers must be configured to use
21631the FastCGI protocol (using the keyword "proto fcgi" on the server line) and a
21632FastCGI application must be configured and used by the backend managing these
21633servers (using the keyword "use-fcgi-app" into the proxy section). Several
21634FastCGI applications may be defined, but only one can be used at a time by a
21635backend.
21636
21637HAProxy implements all features of the FastCGI specification for Responder
21638application. Especially it is able to multiplex several requests on a simple
21639connection.
21640
2164110.1. Setup
21642-----------
21643
2164410.1.1. Fcgi-app section
21645--------------------------
21646
21647fcgi-app <name>
21648 Declare a FastCGI application named <name>. To be valid, at least the
21649 document root must be defined.
21650
21651acl <aclname> <criterion> [flags] [operator] <value> ...
21652 Declare or complete an access list.
21653
21654 See "acl" keyword in section 4.2 and section 7 about ACL usage for
21655 details. ACLs defined for a FastCGI application are private. They cannot be
21656 used by any other application or by any proxy. In the same way, ACLs defined
21657 in any other section are not usable by a FastCGI application. However,
21658 Pre-defined ACLs are available.
21659
21660docroot <path>
21661 Define the document root on the remote host. <path> will be used to build
21662 the default value of FastCGI parameters SCRIPT_FILENAME and
21663 PATH_TRANSLATED. It is a mandatory setting.
21664
21665index <script-name>
21666 Define the script name that will be appended after an URI that ends with a
21667 slash ("/") to set the default value of the FastCGI parameter SCRIPT_NAME. It
21668 is an optional setting.
21669
21670 Example :
21671 index index.php
21672
21673log-stderr global
21674log-stderr <address> [len <length>] [format <format>]
Jan Wagner3e678602020-12-17 22:22:32 +010021675 [sample <ranges>:<sample_size>] <facility> [<level> [<minlevel>]]
Christopher Fauletb30b3102019-09-12 23:03:09 +020021676 Enable logging of STDERR messages reported by the FastCGI application.
21677
21678 See "log" keyword in section 4.2 for details. It is an optional setting. By
21679 default STDERR messages are ignored.
21680
21681pass-header <name> [ { if | unless } <condition> ]
21682 Specify the name of a request header which will be passed to the FastCGI
21683 application. It may optionally be followed by an ACL-based condition, in
21684 which case it will only be evaluated if the condition is true.
21685
21686 Most request headers are already available to the FastCGI application,
21687 prefixed with "HTTP_". Thus, this directive is only required to pass headers
21688 that are purposefully omitted. Currently, the headers "Authorization",
21689 "Proxy-Authorization" and hop-by-hop headers are omitted.
21690
21691 Note that the headers "Content-type" and "Content-length" are never passed to
21692 the FastCGI application because they are already converted into parameters.
21693
21694path-info <regex>
Christopher Faulet28cb3662020-02-14 14:47:37 +010021695 Define a regular expression to extract the script-name and the path-info from
Christopher Faulet6c57f2d2020-02-14 16:55:52 +010021696 the URL-decoded path. Thus, <regex> may have two captures: the first one to
21697 capture the script name and the second one to capture the path-info. The
21698 first one is mandatory, the second one is optional. This way, it is possible
21699 to extract the script-name from the path ignoring the path-info. It is an
21700 optional setting. If it is not defined, no matching is performed on the
21701 path. and the FastCGI parameters PATH_INFO and PATH_TRANSLATED are not
21702 filled.
Christopher Faulet28cb3662020-02-14 14:47:37 +010021703
21704 For security reason, when this regular expression is defined, the newline and
Ilya Shipitsin8525fd92020-02-29 12:34:59 +050021705 the null characters are forbidden from the path, once URL-decoded. The reason
Christopher Faulet28cb3662020-02-14 14:47:37 +010021706 to such limitation is because otherwise the matching always fails (due to a
21707 limitation one the way regular expression are executed in HAProxy). So if one
21708 of these two characters is found in the URL-decoded path, an error is
21709 returned to the client. The principle of least astonishment is applied here.
Christopher Fauletb30b3102019-09-12 23:03:09 +020021710
21711 Example :
Christopher Faulet6c57f2d2020-02-14 16:55:52 +010021712 path-info ^(/.+\.php)(/.*)?$ # both script-name and path-info may be set
21713 path-info ^(/.+\.php) # the path-info is ignored
Christopher Fauletb30b3102019-09-12 23:03:09 +020021714
21715option get-values
21716no option get-values
21717 Enable or disable the retrieve of variables about connection management.
21718
Daniel Corbett67a82712020-07-06 23:01:19 -040021719 HAProxy is able to send the record FCGI_GET_VALUES on connection
Christopher Fauletb30b3102019-09-12 23:03:09 +020021720 establishment to retrieve the value for following variables:
21721
21722 * FCGI_MAX_REQS The maximum number of concurrent requests this
21723 application will accept.
21724
William Lallemand93e548e2019-09-30 13:54:02 +020021725 * FCGI_MPXS_CONNS "0" if this application does not multiplex connections,
21726 "1" otherwise.
Christopher Fauletb30b3102019-09-12 23:03:09 +020021727
21728 Some FastCGI applications does not support this feature. Some others close
Ilya Shipitsin11057a32020-06-21 21:18:27 +050021729 the connection immediately after sending their response. So, by default, this
Christopher Fauletb30b3102019-09-12 23:03:09 +020021730 option is disabled.
21731
21732 Note that the maximum number of concurrent requests accepted by a FastCGI
21733 application is a connection variable. It only limits the number of streams
21734 per connection. If the global load must be limited on the application, the
21735 server parameters "maxconn" and "pool-max-conn" must be set. In addition, if
21736 an application does not support connection multiplexing, the maximum number
21737 of concurrent requests is automatically set to 1.
21738
21739option keep-conn
21740no option keep-conn
21741 Instruct the FastCGI application to keep the connection open or not after
21742 sending a response.
21743
21744 If disabled, the FastCGI application closes the connection after responding
21745 to this request. By default, this option is enabled.
21746
21747option max-reqs <reqs>
21748 Define the maximum number of concurrent requests this application will
21749 accept.
21750
21751 This option may be overwritten if the variable FCGI_MAX_REQS is retrieved
21752 during connection establishment. Furthermore, if the application does not
21753 support connection multiplexing, this option will be ignored. By default set
21754 to 1.
21755
21756option mpxs-conns
21757no option mpxs-conns
21758 Enable or disable the support of connection multiplexing.
21759
21760 This option may be overwritten if the variable FCGI_MPXS_CONNS is retrieved
21761 during connection establishment. It is disabled by default.
21762
21763set-param <name> <fmt> [ { if | unless } <condition> ]
21764 Set a FastCGI parameter that should be passed to this application. Its
21765 value, defined by <fmt> must follows the log-format rules (see section 8.2.4
21766 "Custom Log format"). It may optionally be followed by an ACL-based
21767 condition, in which case it will only be evaluated if the condition is true.
21768
21769 With this directive, it is possible to overwrite the value of default FastCGI
21770 parameters. If the value is evaluated to an empty string, the rule is
21771 ignored. These directives are evaluated in their declaration order.
21772
21773 Example :
21774 # PHP only, required if PHP was built with --enable-force-cgi-redirect
21775 set-param REDIRECT_STATUS 200
21776
21777 set-param PHP_AUTH_DIGEST %[req.hdr(Authorization)]
21778
21779
2178010.1.2. Proxy section
21781---------------------
21782
21783use-fcgi-app <name>
21784 Define the FastCGI application to use for the backend.
21785
21786 Arguments :
21787 <name> is the name of the FastCGI application to use.
21788
21789 This keyword is only available for HTTP proxies with the backend capability
21790 and with at least one FastCGI server. However, FastCGI servers can be mixed
21791 with HTTP servers. But except there is a good reason to do so, it is not
21792 recommended (see section 10.3 about the limitations for details). Only one
21793 application may be defined at a time per backend.
21794
21795 Note that, once a FastCGI application is referenced for a backend, depending
21796 on the configuration some processing may be done even if the request is not
21797 sent to a FastCGI server. Rules to set parameters or pass headers to an
21798 application are evaluated.
21799
21800
2180110.1.3. Example
21802---------------
21803
21804 frontend front-http
21805 mode http
21806 bind *:80
21807 bind *:
21808
21809 use_backend back-dynamic if { path_reg ^/.+\.php(/.*)?$ }
21810 default_backend back-static
21811
21812 backend back-static
21813 mode http
21814 server www A.B.C.D:80
21815
21816 backend back-dynamic
21817 mode http
21818 use-fcgi-app php-fpm
21819 server php-fpm A.B.C.D:9000 proto fcgi
21820
21821 fcgi-app php-fpm
21822 log-stderr global
21823 option keep-conn
21824
21825 docroot /var/www/my-app
21826 index index.php
21827 path-info ^(/.+\.php)(/.*)?$
21828
21829
2183010.2. Default parameters
21831------------------------
21832
21833A Responder FastCGI application has the same purpose as a CGI/1.1 program. In
21834the CGI/1.1 specification (RFC3875), several variables must be passed to the
Ilya Shipitsin8525fd92020-02-29 12:34:59 +050021835script. So HAProxy set them and some others commonly used by FastCGI
Christopher Fauletb30b3102019-09-12 23:03:09 +020021836applications. All these variables may be overwritten, with caution though.
21837
21838 +-------------------+-----------------------------------------------------+
21839 | AUTH_TYPE | Identifies the mechanism, if any, used by HAProxy |
21840 | | to authenticate the user. Concretely, only the |
21841 | | BASIC authentication mechanism is supported. |
21842 | | |
21843 +-------------------+-----------------------------------------------------+
21844 | CONTENT_LENGTH | Contains the size of the message-body attached to |
21845 | | the request. It means only requests with a known |
21846 | | size are considered as valid and sent to the |
21847 | | application. |
21848 | | |
21849 +-------------------+-----------------------------------------------------+
21850 | CONTENT_TYPE | Contains the type of the message-body attached to |
21851 | | the request. It may not be set. |
21852 | | |
21853 +-------------------+-----------------------------------------------------+
21854 | DOCUMENT_ROOT | Contains the document root on the remote host under |
21855 | | which the script should be executed, as defined in |
21856 | | the application's configuration. |
21857 | | |
21858 +-------------------+-----------------------------------------------------+
21859 | GATEWAY_INTERFACE | Contains the dialect of CGI being used by HAProxy |
21860 | | to communicate with the FastCGI application. |
21861 | | Concretely, it is set to "CGI/1.1". |
21862 | | |
21863 +-------------------+-----------------------------------------------------+
21864 | PATH_INFO | Contains the portion of the URI path hierarchy |
21865 | | following the part that identifies the script |
21866 | | itself. To be set, the directive "path-info" must |
21867 | | be defined. |
21868 | | |
21869 +-------------------+-----------------------------------------------------+
21870 | PATH_TRANSLATED | If PATH_INFO is set, it is its translated version. |
21871 | | It is the concatenation of DOCUMENT_ROOT and |
21872 | | PATH_INFO. If PATH_INFO is not set, this parameters |
21873 | | is not set too. |
21874 | | |
21875 +-------------------+-----------------------------------------------------+
21876 | QUERY_STRING | Contains the request's query string. It may not be |
21877 | | set. |
21878 | | |
21879 +-------------------+-----------------------------------------------------+
21880 | REMOTE_ADDR | Contains the network address of the client sending |
21881 | | the request. |
21882 | | |
21883 +-------------------+-----------------------------------------------------+
21884 | REMOTE_USER | Contains the user identification string supplied by |
21885 | | client as part of user authentication. |
21886 | | |
21887 +-------------------+-----------------------------------------------------+
21888 | REQUEST_METHOD | Contains the method which should be used by the |
21889 | | script to process the request. |
21890 | | |
21891 +-------------------+-----------------------------------------------------+
21892 | REQUEST_URI | Contains the request's URI. |
21893 | | |
21894 +-------------------+-----------------------------------------------------+
21895 | SCRIPT_FILENAME | Contains the absolute pathname of the script. it is |
21896 | | the concatenation of DOCUMENT_ROOT and SCRIPT_NAME. |
21897 | | |
21898 +-------------------+-----------------------------------------------------+
21899 | SCRIPT_NAME | Contains the name of the script. If the directive |
21900 | | "path-info" is defined, it is the first part of the |
21901 | | URI path hierarchy, ending with the script name. |
21902 | | Otherwise, it is the entire URI path. |
21903 | | |
21904 +-------------------+-----------------------------------------------------+
21905 | SERVER_NAME | Contains the name of the server host to which the |
21906 | | client request is directed. It is the value of the |
21907 | | header "Host", if defined. Otherwise, the |
21908 | | destination address of the connection on the client |
21909 | | side. |
21910 | | |
21911 +-------------------+-----------------------------------------------------+
21912 | SERVER_PORT | Contains the destination TCP port of the connection |
21913 | | on the client side, which is the port the client |
21914 | | connected to. |
21915 | | |
21916 +-------------------+-----------------------------------------------------+
21917 | SERVER_PROTOCOL | Contains the request's protocol. |
21918 | | |
21919 +-------------------+-----------------------------------------------------+
Christopher Fauletb2a50292021-06-11 13:34:42 +020021920 | SERVER_SOFTWARE | Contains the string "HAProxy" followed by the |
21921 | | current HAProxy version. |
21922 | | |
21923 +-------------------+-----------------------------------------------------+
Christopher Fauletb30b3102019-09-12 23:03:09 +020021924 | HTTPS | Set to a non-empty value ("on") if the script was |
21925 | | queried through the HTTPS protocol. |
21926 | | |
21927 +-------------------+-----------------------------------------------------+
21928
21929
2193010.3. Limitations
21931------------------
21932
21933The current implementation have some limitations. The first one is about the
21934way some request headers are hidden to the FastCGI applications. This happens
21935during the headers analysis, on the backend side, before the connection
21936establishment. At this stage, HAProxy know the backend is using a FastCGI
21937application but it don't know if the request will be routed to a FastCGI server
21938or not. But to hide request headers, it simply removes them from the HTX
21939message. So, if the request is finally routed to an HTTP server, it never see
21940these headers. For this reason, it is not recommended to mix FastCGI servers
21941and HTTP servers under the same backend.
21942
21943Similarly, the rules "set-param" and "pass-header" are evaluated during the
21944request headers analysis. So the evaluation is always performed, even if the
21945requests is finally forwarded to an HTTP server.
21946
21947About the rules "set-param", when a rule is applied, a pseudo header is added
21948into the HTX message. So, the same way than for HTTP header rewrites, it may
21949fail if the buffer is full. The rules "set-param" will compete with
21950"http-request" ones.
21951
21952Finally, all FastCGI params and HTTP headers are sent into a unique record
21953FCGI_PARAM. Encoding of this record must be done in one pass, otherwise a
21954processing error is returned. It means the record FCGI_PARAM, once encoded,
21955must not exceeds the size of a buffer. However, there is no reserve to respect
21956here.
William Lallemand86d0df02017-11-24 21:36:45 +010021957
Emeric Brunce325c42021-04-02 17:05:09 +020021958
2195911. Address formats
21960-------------------
21961
21962Several statements as "bind, "server", "nameserver" and "log" requires an
21963address.
21964
21965This address can be a host name, an IPv4 address, an IPv6 address, or '*'.
21966The '*' is equal to the special address "0.0.0.0" and can be used, in the case
21967of "bind" or "dgram-bind" to listen on all IPv4 of the system.The IPv6
21968equivalent is '::'.
21969
21970Depending of the statement, a port or port range follows the IP address. This
21971is mandatory on 'bind' statement, optional on 'server'.
21972
21973This address can also begin with a slash '/'. It is considered as the "unix"
21974family, and '/' and following characters must be present the path.
21975
21976Default socket type or transport method "datagram" or "stream" depends on the
21977configuration statement showing the address. Indeed, 'bind' and 'server' will
21978use a "stream" socket type by default whereas 'log', 'nameserver' or
21979'dgram-bind' will use a "datagram".
21980
21981Optionally, a prefix could be used to force the address family and/or the
21982socket type and the transport method.
21983
21984
2198511.1 Address family prefixes
21986----------------------------
21987
21988'abns@<name>' following <name> is an abstract namespace (Linux only).
21989
21990'fd@<n>' following address is a file descriptor <n> inherited from the
21991 parent. The fd must be bound and may or may not already be
21992 listening.
21993
21994'ip@<address>[:port1[-port2]]' following <address> is considered as an IPv4 or
21995 IPv6 address depending on the syntax. Depending
21996 on the statement using this address, a port or
21997 a port range may or must be specified.
21998
21999'ipv4@<address>[:port1[-port2]]' following <address> is always considered as
22000 an IPv4 address. Depending on the statement
22001 using this address, a port or a port range
22002 may or must be specified.
22003
22004'ipv6@<address>[:port1[-port2]]' following <address> is always considered as
22005 an IPv6 address. Depending on the statement
22006 using this address, a port or a port range
22007 may or must be specified.
22008
22009'sockpair@<n>' following address is the file descriptor of a connected unix
22010 socket or of a socketpair. During a connection, the initiator
22011 creates a pair of connected sockets, and passes one of them
22012 over the FD to the other end. The listener waits to receive
22013 the FD from the unix socket and uses it as if it were the FD
22014 of an accept(). Should be used carefully.
22015
22016'unix@<path>' following string is considered as a UNIX socket <path>. this
22017 prefix is useful to declare an UNIX socket path which don't
22018 start by slash '/'.
22019
22020
2202111.2 Socket type prefixes
22022-------------------------
22023
22024Previous "Address family prefixes" can also be prefixed to force the socket
22025type and the transport method. The default depends of the statement using
22026this address but in some cases the user may force it to a different one.
22027This is the case for "log" statement where the default is syslog over UDP
22028but we could force to use syslog over TCP.
22029
22030Those prefixes were designed for internal purpose and users should
22031instead use aliases of the next section "11.5.3 Protocol prefixes".
22032
22033If users need one those prefixes to perform what they expect because
22034they can not configure the same using the protocol prefixes, they should
22035report this to the maintainers.
22036
22037'stream+<family>@<address>' forces socket type and transport method
22038 to "stream"
22039
22040'dgram+<family>@<address>' forces socket type and transport method
22041 to "datagram".
22042
22043
2204411.3 Protocol prefixes
22045----------------------
22046
22047'tcp@<address>[:port1[-port2]]' following <address> is considered as an IPv4
22048 or IPv6 address depending of the syntax but
22049 socket type and transport method is forced to
22050 "stream". Depending on the statement using
22051 this address, a port or a port range can or
22052 must be specified. It is considered as an alias
22053 of 'stream+ip@'.
22054
22055'tcp4@<address>[:port1[-port2]]' following <address> is always considered as
22056 an IPv4 address but socket type and transport
22057 method is forced to "stream". Depending on the
22058 statement using this address, a port or port
22059 range can or must be specified.
22060 It is considered as an alias of 'stream+ipv4@'.
22061
22062'tcp6@<address>[:port1[-port2]]' following <address> is always considered as
22063 an IPv6 address but socket type and transport
22064 method is forced to "stream". Depending on the
22065 statement using this address, a port or port
22066 range can or must be specified.
22067 It is considered as an alias of 'stream+ipv4@'.
22068
22069'udp@<address>[:port1[-port2]]' following <address> is considered as an IPv4
22070 or IPv6 address depending of the syntax but
22071 socket type and transport method is forced to
22072 "datagram". Depending on the statement using
22073 this address, a port or a port range can or
22074 must be specified. It is considered as an alias
22075 of 'dgram+ip@'.
22076
22077'udp4@<address>[:port1[-port2]]' following <address> is always considered as
22078 an IPv4 address but socket type and transport
22079 method is forced to "datagram". Depending on
22080 the statement using this address, a port or
22081 port range can or must be specified.
22082 It is considered as an alias of 'stream+ipv4@'.
22083
22084'udp6@<address>[:port1[-port2]]' following <address> is always considered as
22085 an IPv6 address but socket type and transport
22086 method is forced to "datagram". Depending on
22087 the statement using this address, a port or
22088 port range can or must be specified.
22089 It is considered as an alias of 'stream+ipv4@'.
22090
22091'uxdg@<path>' following string is considered as a unix socket <path> but
22092 transport method is forced to "datagram". It is considered as
22093 an alias of 'dgram+unix@'.
22094
22095'uxst@<path>' following string is considered as a unix socket <path> but
22096 transport method is forced to "stream". It is considered as
22097 an alias of 'stream+unix@'.
22098
22099In future versions, other prefixes could be used to specify protocols like
22100QUIC which proposes stream transport based on socket of type "datagram".
22101
Willy Tarreau0ba27502007-12-24 16:55:16 +010022102/*
22103 * Local variables:
22104 * fill-column: 79
22105 * End:
22106 */