blob: 0f9e31aefa5503d93ea02880e382809a7fb9b36b [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
Willy Tarreau09cc6692021-03-19 17:16:18 +01007 2021/03/19
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
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200132
1331. Quick reminder about HTTP
134----------------------------
135
Davor Ocelice9ed2812017-12-25 17:49:28 +0100136When HAProxy is running in HTTP mode, both the request and the response are
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200137fully analyzed and indexed, thus it becomes possible to build matching criteria
138on almost anything found in the contents.
139
140However, it is important to understand how HTTP requests and responses are
141formed, and how HAProxy decomposes them. It will then become easier to write
142correct rules and to debug existing configurations.
143
144
1451.1. The HTTP transaction model
146-------------------------------
147
148The HTTP protocol is transaction-driven. This means that each request will lead
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +0100149to one and only one response. Traditionally, a TCP connection is established
Davor Ocelice9ed2812017-12-25 17:49:28 +0100150from the client to the server, a request is sent by the client through the
151connection, the server responds, and the connection is closed. A new request
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200152will involve a new connection :
153
154 [CON1] [REQ1] ... [RESP1] [CLO1] [CON2] [REQ2] ... [RESP2] [CLO2] ...
155
156In this mode, called the "HTTP close" mode, there are as many connection
157establishments as there are HTTP transactions. Since the connection is closed
158by the server after the response, the client does not need to know the content
159length.
160
161Due to the transactional nature of the protocol, it was possible to improve it
162to avoid closing a connection between two subsequent transactions. In this mode
163however, it is mandatory that the server indicates the content length for each
164response so that the client does not wait indefinitely. For this, a special
165header is used: "Content-length". This mode is called the "keep-alive" mode :
166
167 [CON] [REQ1] ... [RESP1] [REQ2] ... [RESP2] [CLO] ...
168
169Its advantages are a reduced latency between transactions, and less processing
170power required on the server side. It is generally better than the close mode,
171but not always because the clients often limit their concurrent connections to
Patrick Mezard9ec2ec42010-06-12 17:02:45 +0200172a smaller value.
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200173
Willy Tarreau95c4e142017-11-26 12:18:55 +0100174Another improvement in the communications is the pipelining mode. It still uses
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200175keep-alive, but the client does not wait for the first response to send the
176second request. This is useful for fetching large number of images composing a
177page :
178
179 [CON] [REQ1] [REQ2] ... [RESP1] [RESP2] [CLO] ...
180
181This can obviously have a tremendous benefit on performance because the network
182latency is eliminated between subsequent requests. Many HTTP agents do not
183correctly support pipelining since there is no way to associate a response with
184the corresponding request in HTTP. For this reason, it is mandatory for the
Cyril Bonté78caf842010-03-10 22:41:43 +0100185server to reply in the exact same order as the requests were received.
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200186
Willy Tarreau95c4e142017-11-26 12:18:55 +0100187The next improvement is the multiplexed mode, as implemented in HTTP/2. This
188time, each transaction is assigned a single stream identifier, and all streams
189are multiplexed over an existing connection. Many requests can be sent in
190parallel by the client, and responses can arrive in any order since they also
191carry the stream identifier.
192
Willy Tarreau70dffda2014-01-30 03:07:23 +0100193By default HAProxy operates in keep-alive mode with regards to persistent
194connections: for each connection it processes each request and response, and
195leaves the connection idle on both sides between the end of a response and the
Willy Tarreau95c4e142017-11-26 12:18:55 +0100196start of a new request. When it receives HTTP/2 connections from a client, it
197processes all the requests in parallel and leaves the connection idling,
198waiting for new requests, just as if it was a keep-alive HTTP connection.
Patrick Mezard9ec2ec42010-06-12 17:02:45 +0200199
Christopher Faulet315b39c2018-09-21 16:26:19 +0200200HAProxy supports 4 connection modes :
Willy Tarreau70dffda2014-01-30 03:07:23 +0100201 - keep alive : all requests and responses are processed (default)
202 - tunnel : only the first request and response are processed,
Christopher Faulet6c9bbb22019-03-26 21:37:23 +0100203 everything else is forwarded with no analysis (deprecated).
Willy Tarreau70dffda2014-01-30 03:07:23 +0100204 - server close : the server-facing connection is closed after the response.
Christopher Faulet315b39c2018-09-21 16:26:19 +0200205 - close : the connection is actively closed after end of response.
Willy Tarreau70dffda2014-01-30 03:07:23 +0100206
Willy Tarreau95c4e142017-11-26 12:18:55 +0100207
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200208
2091.2. HTTP request
210-----------------
211
212First, let's consider this HTTP request :
213
214 Line Contents
Willy Tarreaud72758d2010-01-12 10:42:19 +0100215 number
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200216 1 GET /serv/login.php?lang=en&profile=2 HTTP/1.1
217 2 Host: www.mydomain.com
218 3 User-agent: my small browser
219 4 Accept: image/jpeg, image/gif
220 5 Accept: image/png
221
222
2231.2.1. The Request line
224-----------------------
225
226Line 1 is the "request line". It is always composed of 3 fields :
227
228 - a METHOD : GET
229 - a URI : /serv/login.php?lang=en&profile=2
230 - a version tag : HTTP/1.1
231
232All of them are delimited by what the standard calls LWS (linear white spaces),
233which are commonly spaces, but can also be tabs or line feeds/carriage returns
234followed by spaces/tabs. The method itself cannot contain any colon (':') and
235is limited to alphabetic letters. All those various combinations make it
236desirable that HAProxy performs the splitting itself rather than leaving it to
237the user to write a complex or inaccurate regular expression.
238
239The URI itself can have several forms :
240
241 - A "relative URI" :
242
243 /serv/login.php?lang=en&profile=2
244
245 It is a complete URL without the host part. This is generally what is
246 received by servers, reverse proxies and transparent proxies.
247
248 - An "absolute URI", also called a "URL" :
249
250 http://192.168.0.12:8080/serv/login.php?lang=en&profile=2
251
252 It is composed of a "scheme" (the protocol name followed by '://'), a host
253 name or address, optionally a colon (':') followed by a port number, then
254 a relative URI beginning at the first slash ('/') after the address part.
255 This is generally what proxies receive, but a server supporting HTTP/1.1
256 must accept this form too.
257
258 - a star ('*') : this form is only accepted in association with the OPTIONS
259 method and is not relayable. It is used to inquiry a next hop's
260 capabilities.
Willy Tarreaud72758d2010-01-12 10:42:19 +0100261
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200262 - an address:port combination : 192.168.0.12:80
263 This is used with the CONNECT method, which is used to establish TCP
264 tunnels through HTTP proxies, generally for HTTPS, but sometimes for
265 other protocols too.
266
267In a relative URI, two sub-parts are identified. The part before the question
268mark is called the "path". It is typically the relative path to static objects
269on the server. The part after the question mark is called the "query string".
270It is mostly used with GET requests sent to dynamic scripts and is very
271specific to the language, framework or application in use.
272
Willy Tarreau95c4e142017-11-26 12:18:55 +0100273HTTP/2 doesn't convey a version information with the request, so the version is
Davor Ocelice9ed2812017-12-25 17:49:28 +0100274assumed to be the same as the one of the underlying protocol (i.e. "HTTP/2").
Willy Tarreau95c4e142017-11-26 12:18:55 +0100275
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200276
2771.2.2. The request headers
278--------------------------
279
280The headers start at the second line. They are composed of a name at the
281beginning of the line, immediately followed by a colon (':'). Traditionally,
282an LWS is added after the colon but that's not required. Then come the values.
283Multiple identical headers may be folded into one single line, delimiting the
284values with commas, provided that their order is respected. This is commonly
285encountered in the "Cookie:" field. A header may span over multiple lines if
286the subsequent lines begin with an LWS. In the example in 1.2, lines 4 and 5
287define a total of 3 values for the "Accept:" header.
288
Davor Ocelice9ed2812017-12-25 17:49:28 +0100289Contrary to a common misconception, header names are not case-sensitive, and
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200290their values are not either if they refer to other header names (such as the
Willy Tarreau95c4e142017-11-26 12:18:55 +0100291"Connection:" header). In HTTP/2, header names are always sent in lower case,
Willy Tarreau253c2512020-07-07 15:55:23 +0200292as can be seen when running in debug mode. Internally, all header names are
293normalized to lower case so that HTTP/1.x and HTTP/2 use the exact same
294representation, and they are sent as-is on the other side. This explains why an
295HTTP/1.x request typed with camel case is delivered in lower case.
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200296
297The end of the headers is indicated by the first empty line. People often say
298that it's a double line feed, which is not exact, even if a double line feed
299is one valid form of empty line.
300
301Fortunately, HAProxy takes care of all these complex combinations when indexing
302headers, checking values and counting them, so there is no reason to worry
303about the way they could be written, but it is important not to accuse an
304application of being buggy if it does unusual, valid things.
305
306Important note:
Lukas Tribus23953682017-04-28 13:24:30 +0000307 As suggested by RFC7231, HAProxy normalizes headers by replacing line breaks
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200308 in the middle of headers by LWS in order to join multi-line headers. This
309 is necessary for proper analysis and helps less capable HTTP parsers to work
310 correctly and not to be fooled by such complex constructs.
311
312
3131.3. HTTP response
314------------------
315
316An HTTP response looks very much like an HTTP request. Both are called HTTP
317messages. Let's consider this HTTP response :
318
319 Line Contents
Willy Tarreaud72758d2010-01-12 10:42:19 +0100320 number
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200321 1 HTTP/1.1 200 OK
322 2 Content-length: 350
323 3 Content-Type: text/html
324
Willy Tarreau816b9792009-09-15 21:25:21 +0200325As a special case, HTTP supports so called "Informational responses" as status
326codes 1xx. These messages are special in that they don't convey any part of the
327response, they're just used as sort of a signaling message to ask a client to
Willy Tarreau5843d1a2010-02-01 15:13:32 +0100328continue to post its request for instance. In the case of a status 100 response
329the requested information will be carried by the next non-100 response message
330following the informational one. This implies that multiple responses may be
331sent to a single request, and that this only works when keep-alive is enabled
332(1xx messages are HTTP/1.1 only). HAProxy handles these messages and is able to
333correctly forward and skip them, and only process the next non-100 response. As
334such, these messages are neither logged nor transformed, unless explicitly
335state otherwise. Status 101 messages indicate that the protocol is changing
336over the same connection and that haproxy must switch to tunnel mode, just as
337if a CONNECT had occurred. Then the Upgrade header would contain additional
338information about the type of protocol the connection is switching to.
Willy Tarreau816b9792009-09-15 21:25:21 +0200339
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200340
Davor Ocelice9ed2812017-12-25 17:49:28 +01003411.3.1. The response line
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200342------------------------
343
344Line 1 is the "response line". It is always composed of 3 fields :
345
346 - a version tag : HTTP/1.1
347 - a status code : 200
348 - a reason : OK
349
350The status code is always 3-digit. The first digit indicates a general status :
Davor Ocelice9ed2812017-12-25 17:49:28 +0100351 - 1xx = informational message to be skipped (e.g. 100, 101)
352 - 2xx = OK, content is following (e.g. 200, 206)
353 - 3xx = OK, no content following (e.g. 302, 304)
354 - 4xx = error caused by the client (e.g. 401, 403, 404)
355 - 5xx = error caused by the server (e.g. 500, 502, 503)
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200356
Lukas Tribus23953682017-04-28 13:24:30 +0000357Please refer to RFC7231 for the detailed meaning of all such codes. The
Willy Tarreaud72758d2010-01-12 10:42:19 +0100358"reason" field is just a hint, but is not parsed by clients. Anything can be
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200359found there, but it's a common practice to respect the well-established
360messages. It can be composed of one or multiple words, such as "OK", "Found",
361or "Authentication Required".
362
Davor Ocelice9ed2812017-12-25 17:49:28 +0100363HAProxy may emit the following status codes by itself :
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200364
365 Code When / reason
366 200 access to stats page, and when replying to monitoring requests
367 301 when performing a redirection, depending on the configured code
368 302 when performing a redirection, depending on the configured code
369 303 when performing a redirection, depending on the configured code
Willy Tarreaub67fdc42013-03-29 19:28:11 +0100370 307 when performing a redirection, depending on the configured code
371 308 when performing a redirection, depending on the configured code
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200372 400 for an invalid or too large request
373 401 when an authentication is required to perform the action (when
374 accessing the stats page)
Christopher Faulet87f1f3d2019-07-18 14:51:20 +0200375 403 when a request is forbidden by a "http-request deny" rule
Florian Tham9205fea2020-01-08 13:35:30 +0100376 404 when the requested resource could not be found
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200377 408 when the request timeout strikes before the request is complete
Florian Tham272e29b2020-01-08 10:19:05 +0100378 410 when the requested resource is no longer available and will not
379 be available again
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200380 500 when haproxy encounters an unrecoverable internal error, such as a
381 memory allocation failure, which should never happen
Christopher Faulete095f312020-12-07 11:22:24 +0100382 501 when haproxy is unable to satisfy a client request because of an
383 unsupported feature
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200384 502 when the server returns an empty, invalid or incomplete response, or
Christopher Faulet87f1f3d2019-07-18 14:51:20 +0200385 when an "http-response deny" rule blocks the response.
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200386 503 when no server was available to handle the request, or in response to
387 monitoring requests which match the "monitor fail" condition
388 504 when the response timeout strikes before the server responds
389
390The error 4xx and 5xx codes above may be customized (see "errorloc" in section
3914.2).
392
393
3941.3.2. The response headers
395---------------------------
396
397Response headers work exactly like request headers, and as such, HAProxy uses
398the same parsing function for both. Please refer to paragraph 1.2.2 for more
399details.
400
401
4022. Configuring HAProxy
403----------------------
404
4052.1. Configuration file format
406------------------------------
Willy Tarreau6a06a402007-07-15 20:15:28 +0200407
408HAProxy's configuration process involves 3 major sources of parameters :
409
410 - the arguments from the command-line, which always take precedence
Willy Tarreau6f1129d2020-11-25 19:58:20 +0100411 - the configuration file(s), whose format is described here
Thayne McCombscdbcca92021-01-07 21:24:41 -0700412 - the running process's environment, in case some environment variables are
Willy Tarreau6f1129d2020-11-25 19:58:20 +0100413 explicitly referenced
Willy Tarreau6a06a402007-07-15 20:15:28 +0200414
Willy Tarreau6f1129d2020-11-25 19:58:20 +0100415The configuration file follows a fairly simple hierarchical format which obey
416a few basic rules:
Willy Tarreau0ba27502007-12-24 16:55:16 +0100417
Willy Tarreau6f1129d2020-11-25 19:58:20 +0100418 1. a configuration file is an ordered sequence of statements
419
420 2. a statement is a single non-empty line before any unprotected "#" (hash)
421
422 3. a line is a series of tokens or "words" delimited by unprotected spaces or
423 tab characters
424
425 4. the first word or sequence of words of a line is one of the keywords or
426 keyword sequences listed in this document
427
428 5. all other words are all arguments of the first one, some being well-known
429 keywords listed in this document, others being values, references to other
430 parts of the configuration, or expressions
431
432 6. certain keywords delimit a section inside which only a subset of keywords
433 are supported
434
435 7. a section ends at the end of a file or on a special keyword starting a new
436 section
437
438This is all that is needed to know to write a simple but reliable configuration
439generator, but this is not enough to reliably parse any configuration nor to
440figure how to deal with certain corner cases.
441
442First, there are a few consequences of the rules above. Rule 6 and 7 imply that
443the keywords used to define a new section are valid everywhere and cannot have
444a different meaning in a specific section. These keywords are always a single
445word (as opposed to a sequence of words), and traditionally the section that
446follows them is designated using the same name. For example when speaking about
447the "global section", it designates the section of configuration that follows
448the "global" keyword. This usage is used a lot in error messages to help locate
449the parts that need to be addressed.
450
451A number of sections create an internal object or configuration space, which
452requires to be distinguished from other ones. In this case they will take an
453extra word which will set the name of this particular section. For some of them
454the section name is mandatory. For example "frontend foo" will create a new
455section of type "frontend" named "foo". Usually a name is specific to its
456section and two sections of different types may use the same name, but this is
457not recommended as it tends to complexify configuration management.
458
459A direct consequence of rule 7 is that when multiple files are read at once,
460each of them must start with a new section, and the end of each file will end
461a section. A file cannot contain sub-sections nor end an existing section and
462start a new one.
463
464Rule 1 mentioned that ordering matters. Indeed, some keywords create directives
465that can be repeated multiple times to create ordered sequences of rules to be
466applied in a certain order. For example "tcp-request" can be used to alternate
467"accept" and "reject" rules on varying criteria. As such, a configuration file
468processor must always preserve a section's ordering when editing a file. The
469ordering of sections usually does not matter except for the global section
470which must be placed before other sections, but it may be repeated if needed.
471In addition, some automatic identifiers may automatically be assigned to some
472of the created objects (e.g. proxies), and by reordering sections, their
473identifiers will change. These ones appear in the statistics for example. As
474such, the configuration below will assign "foo" ID number 1 and "bar" ID number
4752, which will be swapped if the two sections are reversed:
476
477 listen foo
478 bind :80
479
480 listen bar
481 bind :81
482
483Another important point is that according to rules 2 and 3 above, empty lines,
484spaces, tabs, and comments following and unprotected "#" character are not part
485of the configuration as they are just used as delimiters. This implies that the
486following configurations are strictly equivalent:
487
488 global#this is the global section
489 daemon#daemonize
490 frontend foo
491 mode http # or tcp
492
493and:
494
495 global
496 daemon
497
498 # this is the public web frontend
499 frontend foo
500 mode http
501
502The common practice is to align to the left only the keyword that initiates a
503new section, and indent (i.e. prepend a tab character or a few spaces) all
504other keywords so that it's instantly visible that they belong to the same
505section (as done in the second example above). Placing comments before a new
506section helps the reader decide if it's the desired one. Leaving a blank line
507at the end of a section also visually helps spotting the end when editing it.
508
509Tabs are very convenient for indent but they do not copy-paste well. If spaces
510are used instead, it is recommended to avoid placing too many (2 to 4) so that
511editing in field doesn't become a burden with limited editors that do not
512support automatic indent.
513
514In the early days it used to be common to see arguments split at fixed tab
515positions because most keywords would not take more than two arguments. With
516modern versions featuring complex expressions this practice does not stand
517anymore, and is not recommended.
518
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200519
William Lallemandf9873ba2015-05-05 17:37:14 +02005202.2. Quoting and escaping
521-------------------------
522
Willy Tarreau6f1129d2020-11-25 19:58:20 +0100523In modern configurations, some arguments require the use of some characters
524that were previously considered as pure delimiters. In order to make this
525possible, HAProxy supports character escaping by prepending a backslash ('\')
526in front of the character to be escaped, weak quoting within double quotes
527('"') and strong quoting within single quotes ("'").
William Lallemandf9873ba2015-05-05 17:37:14 +0200528
Willy Tarreau6f1129d2020-11-25 19:58:20 +0100529This is pretty similar to what is done in a number of programming languages and
530very close to what is commonly encountered in Bourne shell. The principle is
531the following: while the configuration parser cuts the lines into words, it
532also takes care of quotes and backslashes to decide whether a character is a
533delimiter or is the raw representation of this character within the current
534word. The escape character is then removed, the quotes are removed, and the
535remaining word is used as-is as a keyword or argument for example.
William Lallemandf9873ba2015-05-05 17:37:14 +0200536
Willy Tarreau6f1129d2020-11-25 19:58:20 +0100537If a backslash is needed in a word, it must either be escaped using itself
538(i.e. double backslash) or be strongly quoted.
539
540Escaping outside quotes is achieved by preceding a special character by a
541backslash ('\'):
William Lallemandf9873ba2015-05-05 17:37:14 +0200542
543 \ to mark a space and differentiate it from a delimiter
544 \# to mark a hash and differentiate it from a comment
545 \\ to use a backslash
546 \' to use a single quote and differentiate it from strong quoting
547 \" to use a double quote and differentiate it from weak quoting
548
Willy Tarreau6f1129d2020-11-25 19:58:20 +0100549In addition, a few non-printable characters may be emitted using their usual
550C-language representation:
551
552 \n to insert a line feed (LF, character \x0a or ASCII 10 decimal)
553 \r to insert a carriage return (CR, character \x0d or ASCII 13 decimal)
554 \t to insert a tab (character \x09 or ASCII 9 decimal)
555 \xNN to insert character having ASCII code hex NN (e.g \x0a for LF).
556
557Weak quoting is achieved by surrounding double quotes ("") around the character
558or sequence of characters to protect. Weak quoting prevents the interpretation
559of:
William Lallemandf9873ba2015-05-05 17:37:14 +0200560
Willy Tarreau6f1129d2020-11-25 19:58:20 +0100561 space or tab as a word separator
William Lallemandf9873ba2015-05-05 17:37:14 +0200562 ' single quote as a strong quoting delimiter
563 # hash as a comment start
564
Willy Tarreau6f1129d2020-11-25 19:58:20 +0100565Weak quoting permits the interpretation of environment variables (which are not
566evaluated outside of quotes) by preceding them with a dollar sign ('$'). If a
567dollar character is needed inside double quotes, it must be escaped using a
568backslash.
William Lallemandb2f07452015-05-12 14:27:13 +0200569
Willy Tarreau6f1129d2020-11-25 19:58:20 +0100570Strong quoting is achieved by surrounding single quotes ('') around the
571character or sequence of characters to protect. Inside single quotes, nothing
572is interpreted, it's the efficient way to quote regular expressions.
William Lallemandf9873ba2015-05-05 17:37:14 +0200573
Willy Tarreau6f1129d2020-11-25 19:58:20 +0100574As a result, here is the matrix indicating how special characters can be
575entered in different contexts (unprintable characters are replaced with their
576name within angle brackets). Note that some characters that may only be
577represented escaped have no possible representation inside single quotes,
578hence the '-' there:
William Lallemandf9873ba2015-05-05 17:37:14 +0200579
Willy Tarreau6f1129d2020-11-25 19:58:20 +0100580 Character | Unquoted | Weakly quoted | Strongly quoted
581 -----------+---------------+-----------------------------+-----------------
582 <TAB> | \<TAB>, \x09 | "<TAB>", "\<TAB>", "\x09" | '<TAB>'
583 <LF> | \n, \x0a | "\n", "\x0a" | -
584 <CR> | \r, \x0d | "\r", "\x0d" | -
585 <SPC> | \<SPC>, \x20 | "<SPC>", "\<SPC>", "\x20" | '<SPC>'
586 " | \", \x22 | "\"", "\x22" | '"'
587 # | \#, \x23 | "#", "\#", "\x23" | '#'
588 $ | $, \$, \x24 | "\$", "\x24" | '$'
589 ' | \', \x27 | "'", "\'", "\x27" | -
590 \ | \\, \x5c | "\\", "\x5c" | '\'
William Lallemandf9873ba2015-05-05 17:37:14 +0200591
592 Example:
Willy Tarreau6f1129d2020-11-25 19:58:20 +0100593 # those are all strictly equivalent:
William Lallemandf9873ba2015-05-05 17:37:14 +0200594 log-format %{+Q}o\ %t\ %s\ %{-Q}r
595 log-format "%{+Q}o %t %s %{-Q}r"
596 log-format '%{+Q}o %t %s %{-Q}r'
597 log-format "%{+Q}o %t"' %s %{-Q}r'
598 log-format "%{+Q}o %t"' %s'\ %{-Q}r
599
Willy Tarreau6f1129d2020-11-25 19:58:20 +0100600There is one particular case where a second level of quoting or escaping may be
601necessary. Some keywords take arguments within parenthesis, sometimes delimited
602by commas. These arguments are commonly integers or predefined words, but when
603they are arbitrary strings, it may be required to perform a separate level of
604escaping to disambiguate the characters that belong to the argument from the
605characters that are used to delimit the arguments themselves. A pretty common
606case is the "regsub" converter. It takes a regular expression in argument, and
607if a closing parenthesis is needed inside, this one will require to have its
608own quotes.
609
610The keyword argument parser is exactly the same as the top-level one regarding
611quotes, except that is will not make special cases of backslashes. But what is
Ilya Shipitsin2272d8a2020-12-21 01:22:40 +0500612not always obvious is that the delimiters used inside must first be escaped or
Willy Tarreau6f1129d2020-11-25 19:58:20 +0100613quoted so that they are not resolved at the top level.
614
615Let's take this example making use of the "regsub" converter which takes 3
616arguments, one regular expression, one replacement string and one set of flags:
617
618 # replace all occurrences of "foo" with "blah" in the path:
619 http-request set-path %[path,regsub(foo,blah,g)]
620
621Here no special quoting was necessary. But if now we want to replace either
622"foo" or "bar" with "blah", we'll need the regular expression "(foo|bar)". We
623cannot write:
624
625 http-request set-path %[path,regsub((foo|bar),blah,g)]
626
627because we would like the string to cut like this:
628
629 http-request set-path %[path,regsub((foo|bar),blah,g)]
630 |---------|----|-|
631 arg1 _/ / /
632 arg2 __________/ /
633 arg3 ______________/
634
635but actually what is passed is a string between the opening and closing
636parenthesis then garbage:
637
638 http-request set-path %[path,regsub((foo|bar),blah,g)]
639 |--------|--------|
640 arg1=(foo|bar _/ /
641 trailing garbage _________/
642
643The obvious solution here seems to be that the closing parenthesis needs to be
644quoted, but alone this will not work, because as mentioned above, quotes are
645processed by the top-level parser which will resolve them before processing
646this word:
647
648 http-request set-path %[path,regsub("(foo|bar)",blah,g)]
649 ------------ -------- ----------------------------------
650 word1 word2 word3=%[path,regsub((foo|bar),blah,g)]
651
652So we didn't change anything for the argument parser at the second level which
653still sees a truncated regular expression as the only argument, and garbage at
654the end of the string. By escaping the quotes they will be passed unmodified to
655the second level:
656
657 http-request set-path %[path,regsub(\"(foo|bar)\",blah,g)]
658 ------------ -------- ------------------------------------
659 word1 word2 word3=%[path,regsub("(foo|bar)",blah,g)]
660 |---------||----|-|
661 arg1=(foo|bar) _/ / /
662 arg2=blah ___________/ /
663 arg3=g _______________/
664
Ilya Shipitsin2272d8a2020-12-21 01:22:40 +0500665Another approach consists in using single quotes outside the whole string and
Willy Tarreau6f1129d2020-11-25 19:58:20 +0100666double quotes inside (so that the double quotes are not stripped again):
667
668 http-request set-path '%[path,regsub("(foo|bar)",blah,g)]'
669 ------------ -------- ----------------------------------
670 word1 word2 word3=%[path,regsub("(foo|bar)",blah,g)]
671 |---------||----|-|
672 arg1=(foo|bar) _/ / /
673 arg2 ___________/ /
674 arg3 _______________/
675
676When using regular expressions, it can happen that the dollar ('$') character
677appears in the expression or that a backslash ('\') is used in the replacement
678string. In this case these ones will also be processed inside the double quotes
679thus single quotes are preferred (or double escaping). Example:
680
681 http-request set-path '%[path,regsub("^/(here)(/|$)","my/\1",g)]'
682 ------------ -------- -----------------------------------------
683 word1 word2 word3=%[path,regsub("^/(here)(/|$)","my/\1",g)]
684 |-------------| |-----||-|
685 arg1=(here)(/|$) _/ / /
686 arg2=my/\1 ________________/ /
687 arg3 ______________________/
688
Ilya Shipitsin2272d8a2020-12-21 01:22:40 +0500689Remember that backslahes are not escape characters within single quotes and
Willy Tarreau6f1129d2020-11-25 19:58:20 +0100690that the whole word3 above is already protected against them using the single
691quotes. Conversely, if double quotes had been used around the whole expression,
692single the dollar character and the backslashes would have been resolved at top
693level, breaking the argument contents at the second level.
694
695When in doubt, simply do not use quotes anywhere, and start to place single or
696double quotes around arguments that require a comma or a closing parenthesis,
697and think about escaping these quotes using a backslash of the string contains
698a dollar or a backslash. Again, this is pretty similar to what is used under
699a Bourne shell when double-escaping a command passed to "eval". For API writers
700the best is probably to place escaped quotes around each and every argument,
701regardless of their contents. Users will probably find that using single quotes
702around the whole expression and double quotes around each argument provides
703more readable configurations.
William Lallemandf9873ba2015-05-05 17:37:14 +0200704
705
William Lallemandb2f07452015-05-12 14:27:13 +02007062.3. Environment variables
707--------------------------
708
709HAProxy's configuration supports environment variables. Those variables are
710interpreted only within double quotes. Variables are expanded during the
711configuration parsing. Variable names must be preceded by a dollar ("$") and
712optionally enclosed with braces ("{}") similarly to what is done in Bourne
713shell. Variable names can contain alphanumerical characters or the character
Amaury Denoyellefa41cb62020-10-01 14:32:35 +0200714underscore ("_") but should not start with a digit. If the variable contains a
715list of several values separated by spaces, it can be expanded as individual
716arguments by enclosing the variable with braces and appending the suffix '[*]'
717before the closing brace.
William Lallemandb2f07452015-05-12 14:27:13 +0200718
719 Example:
720
721 bind "fd@${FD_APP1}"
722
723 log "${LOCAL_SYSLOG}:514" local0 notice # send to local server
724
725 user "$HAPROXY_USER"
726
William Lallemand4d03e432019-06-14 15:35:37 +0200727Some variables are defined by HAProxy, they can be used in the configuration
728file, or could be inherited by a program (See 3.7. Programs):
William Lallemanddaf4cd22018-04-17 16:46:13 +0200729
William Lallemand4d03e432019-06-14 15:35:37 +0200730* HAPROXY_LOCALPEER: defined at the startup of the process which contains the
731 name of the local peer. (See "-L" in the management guide.)
732
733* HAPROXY_CFGFILES: list of the configuration files loaded by HAProxy,
734 separated by semicolons. Can be useful in the case you specified a
735 directory.
736
737* HAPROXY_MWORKER: In master-worker mode, this variable is set to 1.
738
John Roeslerfb2fce12019-07-10 15:45:51 -0500739* HAPROXY_CLI: configured listeners addresses of the stats socket for every
William Lallemand4d03e432019-06-14 15:35:37 +0200740 processes, separated by semicolons.
741
John Roeslerfb2fce12019-07-10 15:45:51 -0500742* HAPROXY_MASTER_CLI: In master-worker mode, listeners addresses of the master
William Lallemand4d03e432019-06-14 15:35:37 +0200743 CLI, separated by semicolons.
744
745See also "external-check command" for other variables.
William Lallemandb2f07452015-05-12 14:27:13 +0200746
Willy Tarreau4b103022021-02-12 17:59:10 +0100747
7482.4. Conditional blocks
749-----------------------
750
751It may sometimes be convenient to be able to conditionally enable or disable
752some arbitrary parts of the configuration, for example to enable/disable SSL or
753ciphers, enable or disable some pre-production listeners without modifying the
754configuration, or adjust the configuration's syntax to support two distinct
755versions of HAProxy during a migration.. HAProxy brings a set of nestable
756preprocessor-like directives which allow to integrate or ignore some blocks of
757text. These directives must be placed on their own line and they act on the
758lines that follow them. Two of them support an expression, the other ones only
759switch to an alternate block or end a current level. The 4 following directives
760are defined to form conditional blocks:
761
762 - .if <condition>
763 - .elif <condition>
764 - .else
765 - .endif
766
767The ".if" directive nests a new level, ".elif" stays at the same level, ".else"
768as well, and ".endif" closes a level. Each ".if" must be terminated by a
769matching ".endif". The ".elif" may only be placed after ".if" or ".elif", and
770there is no limit to the number of ".elif" that may be chained. There may be
771only one ".else" per ".if" and it must always be after the ".if" or the last
772".elif" of a block.
773
774Comments may be placed on the same line if needed after a '#', they will be
775ignored. The directives are tokenized like other configuration directives, and
776as such it is possible to use environment variables in conditions.
777
778The conditions are currently limited to:
779
780 - an empty string, always returns "false"
781 - the integer zero ('0'), always returns "false"
782 - a non-nul integer (e.g. '1'), always returns "true".
783
784Other patterns are not supported yet but the purpose is to bring a few
785functions to test for certain build options and supported features.
786
787Three other directives are provided to report some status:
788
789 - .notice "message" : emit this message at level NOTICE
790 - .warning "message" : emit this message at level WARNING
791 - .alert "message" : emit this message at level ALERT
792
793Messages emitted at level WARNING may cause the process to fail to start if the
794"strict-mode" is enabled. Messages emitted at level ALERT will always cause a
795fatal error. These can be used to detect some inappropriate conditions and
796provide advice to the user.
797
798Example:
799
800 .if "${A}"
801 .if "${B}"
802 .notice "A=1, B=1"
803 .elif "${C}"
804 .notice "A=1, B=0, C=1"
805 .elif "${D}"
806 .warning "A=1, B=0, C=0, D=1"
807 .else
808 .alert "A=1, B=0, C=0, D=0"
809 .endif
810 .else
811 .notice "A=0"
812 .endif
813
814
8152.5. Time format
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200816----------------
817
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +0100818Some parameters involve values representing time, such as timeouts. These
Willy Tarreau0ba27502007-12-24 16:55:16 +0100819values are generally expressed in milliseconds (unless explicitly stated
820otherwise) but may be expressed in any other unit by suffixing the unit to the
821numeric value. It is important to consider this because it will not be repeated
822for every keyword. Supported units are :
823
824 - us : microseconds. 1 microsecond = 1/1000000 second
825 - ms : milliseconds. 1 millisecond = 1/1000 second. This is the default.
826 - s : seconds. 1s = 1000ms
827 - m : minutes. 1m = 60s = 60000ms
828 - h : hours. 1h = 60m = 3600s = 3600000ms
829 - d : days. 1d = 24h = 1440m = 86400s = 86400000ms
830
831
Willy Tarreau4b103022021-02-12 17:59:10 +01008322.6. Examples
Patrick Mezard35da19c2010-06-12 17:02:47 +0200833-------------
834
835 # Simple configuration for an HTTP proxy listening on port 80 on all
836 # interfaces and forwarding requests to a single backend "servers" with a
837 # single server "server1" listening on 127.0.0.1:8000
838 global
839 daemon
840 maxconn 256
841
842 defaults
843 mode http
844 timeout connect 5000ms
845 timeout client 50000ms
846 timeout server 50000ms
847
848 frontend http-in
849 bind *:80
850 default_backend servers
851
852 backend servers
853 server server1 127.0.0.1:8000 maxconn 32
854
855
856 # The same configuration defined with a single listen block. Shorter but
857 # less expressive, especially in HTTP mode.
858 global
859 daemon
860 maxconn 256
861
862 defaults
863 mode http
864 timeout connect 5000ms
865 timeout client 50000ms
866 timeout server 50000ms
867
868 listen http-in
869 bind *:80
870 server server1 127.0.0.1:8000 maxconn 32
871
872
873Assuming haproxy is in $PATH, test these configurations in a shell with:
874
Willy Tarreauccb289d2010-12-11 20:19:38 +0100875 $ sudo haproxy -f configuration.conf -c
Patrick Mezard35da19c2010-06-12 17:02:47 +0200876
877
Willy Tarreauc57f0e22009-05-10 13:12:33 +02008783. Global parameters
Willy Tarreau6a06a402007-07-15 20:15:28 +0200879--------------------
880
881Parameters in the "global" section are process-wide and often OS-specific. They
882are generally set once for all and do not need being changed once correct. Some
883of them have command-line equivalents.
884
885The following keywords are supported in the "global" section :
886
887 * Process management and security
Emeric Brunc8e8d122012-10-02 18:42:10 +0200888 - ca-base
Willy Tarreau6a06a402007-07-15 20:15:28 +0200889 - chroot
Emeric Brunc8e8d122012-10-02 18:42:10 +0200890 - crt-base
Baptiste Assmann3493d0f2015-10-12 20:21:23 +0200891 - cpu-map
Willy Tarreau6a06a402007-07-15 20:15:28 +0200892 - daemon
Baptiste Assmann3493d0f2015-10-12 20:21:23 +0200893 - description
894 - deviceatlas-json-file
895 - deviceatlas-log-level
896 - deviceatlas-separator
897 - deviceatlas-properties-cookie
Simon Horman98637e52014-06-20 12:30:16 +0900898 - external-check
Willy Tarreau6a06a402007-07-15 20:15:28 +0200899 - gid
900 - group
Cyril Bonté203ec5a2017-03-23 22:44:13 +0100901 - hard-stop-after
Christopher Faulet98fbe952019-07-22 16:18:24 +0200902 - h1-case-adjust
903 - h1-case-adjust-file
Willy Tarreaud96f1122019-12-03 07:07:36 +0100904 - insecure-fork-wanted
Willy Tarreaua45a8b52019-12-06 16:31:45 +0100905 - insecure-setuid-wanted
Emmanuel Hocdet70df7bf2019-01-04 11:08:20 +0100906 - issuers-chain-path
Dragan Dosen13cd54c2020-06-18 18:24:05 +0200907 - localpeer
Willy Tarreau6a06a402007-07-15 20:15:28 +0200908 - log
Baptiste Assmann3493d0f2015-10-12 20:21:23 +0200909 - log-tag
Joe Williamsdf5b38f2010-12-29 17:05:48 +0100910 - log-send-hostname
Baptiste Assmann3493d0f2015-10-12 20:21:23 +0200911 - lua-load
Thierry Fournier59f11be2020-11-29 00:37:41 +0100912 - lua-load-per-thread
Tim Duesterhusdd74b5f2020-01-12 13:55:40 +0100913 - lua-prepend-path
William Lallemand27edc4b2019-05-07 17:49:33 +0200914 - mworker-max-reloads
Willy Tarreau6a06a402007-07-15 20:15:28 +0200915 - nbproc
Christopher Fauletbe0faa22017-08-29 15:37:10 +0200916 - nbthread
Baptiste Assmann3493d0f2015-10-12 20:21:23 +0200917 - node
Willy Tarreau6a06a402007-07-15 20:15:28 +0200918 - pidfile
Willy Tarreau119e50e2020-05-22 13:53:29 +0200919 - pp2-never-send-local
Willy Tarreau1d549722016-02-16 12:41:57 +0100920 - presetenv
921 - resetenv
Willy Tarreau6a06a402007-07-15 20:15:28 +0200922 - uid
923 - ulimit-n
924 - user
Willy Tarreau636848a2019-04-15 19:38:50 +0200925 - set-dumpable
Willy Tarreau13d2ba22021-03-26 11:38:08 +0100926 - set-var
Willy Tarreau1d549722016-02-16 12:41:57 +0100927 - setenv
Willy Tarreaufbee7132007-10-18 13:53:22 +0200928 - stats
Baptiste Assmann3493d0f2015-10-12 20:21:23 +0200929 - ssl-default-bind-ciphers
Dirkjan Bussink415150f2018-09-14 11:14:21 +0200930 - ssl-default-bind-ciphersuites
Jerome Magninb203ff62020-04-03 15:28:22 +0200931 - ssl-default-bind-curves
Baptiste Assmann3493d0f2015-10-12 20:21:23 +0200932 - ssl-default-bind-options
933 - ssl-default-server-ciphers
Dirkjan Bussink415150f2018-09-14 11:14:21 +0200934 - ssl-default-server-ciphersuites
Baptiste Assmann3493d0f2015-10-12 20:21:23 +0200935 - ssl-default-server-options
936 - ssl-dh-param-file
Emeric Brun850efd52014-01-29 12:24:34 +0100937 - ssl-server-verify
Emmanuel Hocdetc3b7e742020-04-22 11:06:19 +0200938 - ssl-skip-self-issued-ca
Willy Tarreauceb24bc2010-11-09 12:46:41 +0100939 - unix-bind
Willy Tarreau1d549722016-02-16 12:41:57 +0100940 - unsetenv
Thomas Holmesdb04f192015-05-18 13:21:39 +0100941 - 51degrees-data-file
942 - 51degrees-property-name-list
Dragan Dosen93b38d92015-06-29 16:43:25 +0200943 - 51degrees-property-separator
Dragan Dosenae6d39a2015-06-29 16:43:27 +0200944 - 51degrees-cache-size
Willy Tarreaub3cc9f22019-04-19 16:03:32 +0200945 - wurfl-data-file
946 - wurfl-information-list
947 - wurfl-information-list-separator
Willy Tarreaub3cc9f22019-04-19 16:03:32 +0200948 - wurfl-cache-size
William Dauchy0fec3ab2019-10-27 20:08:11 +0100949 - strict-limits
Willy Tarreaud72758d2010-01-12 10:42:19 +0100950
Willy Tarreau6a06a402007-07-15 20:15:28 +0200951 * Performance tuning
William Dauchy0a8824f2019-10-27 20:08:09 +0100952 - busy-polling
Willy Tarreau1746eec2014-04-25 10:46:47 +0200953 - max-spread-checks
Willy Tarreau6a06a402007-07-15 20:15:28 +0200954 - maxconn
Willy Tarreau81c25d02011-09-07 15:17:21 +0200955 - maxconnrate
William Lallemandd85f9172012-11-09 17:05:39 +0100956 - maxcomprate
William Lallemand072a2bf2012-11-20 17:01:01 +0100957 - maxcompcpuusage
Willy Tarreauff4f82d2009-02-06 11:28:13 +0100958 - maxpipes
Willy Tarreau93e7c002013-10-07 18:51:07 +0200959 - maxsessrate
Willy Tarreau403edff2012-09-06 11:58:37 +0200960 - maxsslconn
Willy Tarreaue43d5322013-10-07 20:01:52 +0200961 - maxsslrate
Baptiste Assmann3493d0f2015-10-12 20:21:23 +0200962 - maxzlibmem
Willy Tarreau6a06a402007-07-15 20:15:28 +0200963 - noepoll
964 - nokqueue
Emmanuel Hocdet0ba4f482019-04-08 16:53:32 +0000965 - noevports
Willy Tarreau6a06a402007-07-15 20:15:28 +0200966 - nopoll
Willy Tarreauff4f82d2009-02-06 11:28:13 +0100967 - nosplice
Jarno Huuskonen0e82b922014-04-12 18:22:19 +0300968 - nogetaddrinfo
Lukas Tribusa0bcbdc2016-09-12 21:42:20 +0000969 - noreuseport
Willy Tarreau75c62c22018-11-22 11:02:09 +0100970 - profiling.tasks
Willy Tarreaufe255b72007-10-14 23:09:26 +0200971 - spread-checks
Baptiste Assmann5626f482015-08-23 10:00:10 +0200972 - server-state-base
Baptiste Assmannef1f0fc2015-08-23 10:06:39 +0200973 - server-state-file
Grant Zhang872f9c22017-01-21 01:10:18 +0000974 - ssl-engine
Grant Zhangfa6c7ee2017-01-14 01:42:15 +0000975 - ssl-mode-async
Baptiste Assmann3493d0f2015-10-12 20:21:23 +0200976 - tune.buffers.limit
977 - tune.buffers.reserve
Willy Tarreau27a674e2009-08-17 07:23:33 +0200978 - tune.bufsize
Willy Tarreau43961d52010-10-04 20:39:20 +0200979 - tune.chksize
William Lallemandf3747832012-11-09 12:33:10 +0100980 - tune.comp.maxlevel
Willy Tarreaubc52bec2020-06-18 08:58:47 +0200981 - tune.fd.edge-triggered
Willy Tarreaufe20e5b2017-07-27 11:42:14 +0200982 - tune.h2.header-table-size
Willy Tarreaue6baec02017-07-27 11:45:11 +0200983 - tune.h2.initial-window-size
Willy Tarreau5242ef82017-07-27 11:47:28 +0200984 - tune.h2.max-concurrent-streams
Willy Tarreau193b8c62012-11-22 00:17:38 +0100985 - tune.http.cookielen
Stéphane Cottin23e9e932017-05-18 08:58:41 +0200986 - tune.http.logurilen
Willy Tarreauac1932d2011-10-24 19:14:41 +0200987 - tune.http.maxhdr
Willy Tarreau76cc6992020-07-01 18:49:24 +0200988 - tune.idle-pool.shared
Willy Tarreau7e312732014-02-12 16:35:14 +0100989 - tune.idletimer
Thierry FOURNIER90da1912015-03-05 11:17:06 +0100990 - tune.lua.forced-yield
Willy Tarreau32f61e22015-03-18 17:54:59 +0100991 - tune.lua.maxmem
Thierry FOURNIER90da1912015-03-05 11:17:06 +0100992 - tune.lua.session-timeout
993 - tune.lua.task-timeout
Thierry FOURNIER7dd784b2015-10-01 14:49:33 +0200994 - tune.lua.service-timeout
Willy Tarreaua0250ba2008-01-06 11:22:57 +0100995 - tune.maxaccept
996 - tune.maxpollevents
Willy Tarreau27a674e2009-08-17 07:23:33 +0200997 - tune.maxrewrite
Willy Tarreauf3045d22015-04-29 16:24:50 +0200998 - tune.pattern.cache-size
Willy Tarreaubd9a0a72011-10-23 21:14:29 +0200999 - tune.pipesize
Willy Tarreaua8e2d972020-07-01 18:27:16 +02001000 - tune.pool-high-fd-ratio
1001 - tune.pool-low-fd-ratio
Willy Tarreaue803de22010-01-21 17:43:04 +01001002 - tune.rcvbuf.client
1003 - tune.rcvbuf.server
Willy Tarreaub22fc302015-12-14 12:04:35 +01001004 - tune.recv_enough
Olivier Houchard1599b802018-05-24 18:59:04 +02001005 - tune.runqueue-depth
Willy Tarreaue7723bd2020-06-24 11:11:02 +02001006 - tune.sched.low-latency
Willy Tarreaue803de22010-01-21 17:43:04 +01001007 - tune.sndbuf.client
1008 - tune.sndbuf.server
Willy Tarreau6ec58db2012-11-16 16:32:15 +01001009 - tune.ssl.cachesize
William Lallemand7d42ef52020-07-06 11:41:30 +02001010 - tune.ssl.keylog
Willy Tarreaubfd59462013-02-21 07:46:09 +01001011 - tune.ssl.lifetime
Emeric Brun8dc60392014-05-09 13:52:00 +02001012 - tune.ssl.force-private-cache
Willy Tarreaubfd59462013-02-21 07:46:09 +01001013 - tune.ssl.maxrecord
Remi Gacognef46cd6e2014-06-12 14:58:40 +02001014 - tune.ssl.default-dh-param
Christopher Faulet31af49d2015-06-09 17:29:50 +02001015 - tune.ssl.ssl-ctx-cache-size
Thierry FOURNIER5bf77322017-02-25 12:45:22 +01001016 - tune.ssl.capture-cipherlist-size
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02001017 - tune.vars.global-max-size
Christopher Fauletff2613e2016-11-09 11:36:17 +01001018 - tune.vars.proc-max-size
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02001019 - tune.vars.reqres-max-size
1020 - tune.vars.sess-max-size
1021 - tune.vars.txn-max-size
William Lallemanda509e4c2012-11-07 16:54:34 +01001022 - tune.zlib.memlevel
1023 - tune.zlib.windowsize
Willy Tarreaud72758d2010-01-12 10:42:19 +01001024
Willy Tarreau6a06a402007-07-15 20:15:28 +02001025 * Debugging
Willy Tarreau6a06a402007-07-15 20:15:28 +02001026 - quiet
Willy Tarreau3eb10b82020-04-15 16:42:39 +02001027 - zero-warning
Willy Tarreau6a06a402007-07-15 20:15:28 +02001028
1029
Willy Tarreauc57f0e22009-05-10 13:12:33 +020010303.1. Process management and security
Willy Tarreau6a06a402007-07-15 20:15:28 +02001031------------------------------------
1032
Emeric Brunc8e8d122012-10-02 18:42:10 +02001033ca-base <dir>
1034 Assigns a default directory to fetch SSL CA certificates and CRLs from when a
Emmanuel Hocdet842e94e2019-12-16 16:39:17 +01001035 relative path is used with "ca-file", "ca-verify-file" or "crl-file"
1036 directives. Absolute locations specified in "ca-file", "ca-verify-file" and
1037 "crl-file" prevail and ignore "ca-base".
Emeric Brunc8e8d122012-10-02 18:42:10 +02001038
Willy Tarreau6a06a402007-07-15 20:15:28 +02001039chroot <jail dir>
1040 Changes current directory to <jail dir> and performs a chroot() there before
1041 dropping privileges. This increases the security level in case an unknown
1042 vulnerability would be exploited, since it would make it very hard for the
1043 attacker to exploit the system. This only works when the process is started
1044 with superuser privileges. It is important to ensure that <jail_dir> is both
Davor Ocelice9ed2812017-12-25 17:49:28 +01001045 empty and non-writable to anyone.
Willy Tarreaud72758d2010-01-12 10:42:19 +01001046
Christopher Fauletcb6a9452017-11-22 16:50:41 +01001047cpu-map [auto:]<process-set>[/<thread-set>] <cpu-set>...
1048 On Linux 2.6 and above, it is possible to bind a process or a thread to a
1049 specific CPU set. This means that the process or the thread will never run on
1050 other CPUs. The "cpu-map" directive specifies CPU sets for process or thread
1051 sets. The first argument is a process set, eventually followed by a thread
1052 set. These sets have the format
1053
1054 all | odd | even | number[-[number]]
1055
1056 <number>> must be a number between 1 and 32 or 64, depending on the machine's
Davor Ocelice9ed2812017-12-25 17:49:28 +01001057 word size. Any process IDs above nbproc and any thread IDs above nbthread are
Christopher Fauletcb6a9452017-11-22 16:50:41 +01001058 ignored. It is possible to specify a range with two such number delimited by
1059 a dash ('-'). It also is possible to specify all processes at once using
Christopher Faulet1dcb9cb2017-11-22 10:24:40 +01001060 "all", only odd numbers using "odd" or even numbers using "even", just like
1061 with the "bind-process" directive. The second and forthcoming arguments are
Davor Ocelice9ed2812017-12-25 17:49:28 +01001062 CPU sets. Each CPU set is either a unique number between 0 and 31 or 63 or a
Christopher Faulet1dcb9cb2017-11-22 10:24:40 +01001063 range with two such numbers delimited by a dash ('-'). Multiple CPU numbers
Christopher Fauletcb6a9452017-11-22 16:50:41 +01001064 or ranges may be specified, and the processes or threads will be allowed to
Davor Ocelice9ed2812017-12-25 17:49:28 +01001065 bind to all of them. Obviously, multiple "cpu-map" directives may be
Christopher Fauletcb6a9452017-11-22 16:50:41 +01001066 specified. Each "cpu-map" directive will replace the previous ones when they
1067 overlap. A thread will be bound on the intersection of its mapping and the
1068 one of the process on which it is attached. If the intersection is null, no
1069 specific binding will be set for the thread.
Willy Tarreaufc6c0322012-11-16 16:12:27 +01001070
Christopher Fauletff4121f2017-11-22 16:38:49 +01001071 Ranges can be partially defined. The higher bound can be omitted. In such
1072 case, it is replaced by the corresponding maximum value, 32 or 64 depending
1073 on the machine's word size.
1074
Christopher Faulet26028f62017-11-22 15:01:51 +01001075 The prefix "auto:" can be added before the process set to let HAProxy
Christopher Fauletcb6a9452017-11-22 16:50:41 +01001076 automatically bind a process or a thread to a CPU by incrementing
1077 process/thread and CPU sets. To be valid, both sets must have the same
1078 size. No matter the declaration order of the CPU sets, it will be bound from
1079 the lowest to the highest bound. Having a process and a thread range with the
1080 "auto:" prefix is not supported. Only one range is supported, the other one
1081 must be a fixed number.
Christopher Faulet26028f62017-11-22 15:01:51 +01001082
1083 Examples:
Christopher Fauletcb6a9452017-11-22 16:50:41 +01001084 cpu-map 1-4 0-3 # bind processes 1 to 4 on the first 4 CPUs
1085
1086 cpu-map 1/all 0-3 # bind all threads of the first process on the
1087 # first 4 CPUs
1088
1089 cpu-map 1- 0- # will be replaced by "cpu-map 1-64 0-63"
1090 # or "cpu-map 1-32 0-31" depending on the machine's
1091 # word size.
1092
Christopher Faulet26028f62017-11-22 15:01:51 +01001093 # all these lines bind the process 1 to the cpu 0, the process 2 to cpu 1
Christopher Fauletcb6a9452017-11-22 16:50:41 +01001094 # and so on.
Christopher Faulet26028f62017-11-22 15:01:51 +01001095 cpu-map auto:1-4 0-3
1096 cpu-map auto:1-4 0-1 2-3
1097 cpu-map auto:1-4 3 2 1 0
1098
Christopher Fauletcb6a9452017-11-22 16:50:41 +01001099 # all these lines bind the thread 1 to the cpu 0, the thread 2 to cpu 1
1100 # and so on.
1101 cpu-map auto:1/1-4 0-3
1102 cpu-map auto:1/1-4 0-1 2-3
1103 cpu-map auto:1/1-4 3 2 1 0
1104
Davor Ocelice9ed2812017-12-25 17:49:28 +01001105 # bind each process to exactly one CPU using all/odd/even keyword
Christopher Faulet26028f62017-11-22 15:01:51 +01001106 cpu-map auto:all 0-63
1107 cpu-map auto:even 0-31
1108 cpu-map auto:odd 32-63
1109
1110 # invalid cpu-map because process and CPU sets have different sizes.
1111 cpu-map auto:1-4 0 # invalid
1112 cpu-map auto:1 0-3 # invalid
1113
Christopher Fauletcb6a9452017-11-22 16:50:41 +01001114 # invalid cpu-map because automatic binding is used with a process range
1115 # and a thread range.
1116 cpu-map auto:all/all 0 # invalid
1117 cpu-map auto:all/1-4 0 # invalid
1118 cpu-map auto:1-4/all 0 # invalid
1119
Emeric Brunc8e8d122012-10-02 18:42:10 +02001120crt-base <dir>
1121 Assigns a default directory to fetch SSL certificates from when a relative
William Dauchy238ea3b2020-01-11 13:09:12 +01001122 path is used with "crtfile" or "crt" directives. Absolute locations specified
1123 prevail and ignore "crt-base".
Emeric Brunc8e8d122012-10-02 18:42:10 +02001124
Willy Tarreau6a06a402007-07-15 20:15:28 +02001125daemon
1126 Makes the process fork into background. This is the recommended mode of
1127 operation. It is equivalent to the command line "-D" argument. It can be
Lukas Tribusf46bf952017-11-21 12:39:34 +01001128 disabled by the command line "-db" argument. This option is ignored in
1129 systemd mode.
Willy Tarreau6a06a402007-07-15 20:15:28 +02001130
David Carlier8167f302015-06-01 13:50:06 +02001131deviceatlas-json-file <path>
1132 Sets the path of the DeviceAtlas JSON data file to be loaded by the API.
Davor Ocelice9ed2812017-12-25 17:49:28 +01001133 The path must be a valid JSON data file and accessible by HAProxy process.
David Carlier8167f302015-06-01 13:50:06 +02001134
1135deviceatlas-log-level <value>
Davor Ocelice9ed2812017-12-25 17:49:28 +01001136 Sets the level of information returned by the API. This directive is
David Carlier8167f302015-06-01 13:50:06 +02001137 optional and set to 0 by default if not set.
1138
1139deviceatlas-separator <char>
1140 Sets the character separator for the API properties results. This directive
1141 is optional and set to | by default if not set.
1142
Cyril Bonté0306c4a2015-10-26 22:37:38 +01001143deviceatlas-properties-cookie <name>
Cyril Bonté307ee1e2015-09-28 23:16:06 +02001144 Sets the client cookie's name used for the detection if the DeviceAtlas
1145 Client-side component was used during the request. This directive is optional
1146 and set to DAPROPS by default if not set.
David Carlier29b3ca32015-09-25 14:09:21 +01001147
Simon Horman98637e52014-06-20 12:30:16 +09001148external-check
Willy Tarreaud96f1122019-12-03 07:07:36 +01001149 Allows the use of an external agent to perform health checks. This is
1150 disabled by default as a security precaution, and even when enabled, checks
Willy Tarreaua45a8b52019-12-06 16:31:45 +01001151 may still fail unless "insecure-fork-wanted" is enabled as well. If the
1152 program launched makes use of a setuid executable (it should really not),
1153 you may also need to set "insecure-setuid-wanted" in the global section.
1154 See "option external-check", and "insecure-fork-wanted", and
1155 "insecure-setuid-wanted".
Simon Horman98637e52014-06-20 12:30:16 +09001156
Willy Tarreau6a06a402007-07-15 20:15:28 +02001157gid <number>
Thayne McCombscdbcca92021-01-07 21:24:41 -07001158 Changes the process's group ID to <number>. It is recommended that the group
Willy Tarreau6a06a402007-07-15 20:15:28 +02001159 ID is dedicated to HAProxy or to a small set of similar daemons. HAProxy must
1160 be started with a user belonging to this group, or with superuser privileges.
Michael Schererab012dd2013-01-12 18:35:19 +01001161 Note that if haproxy is started from a user having supplementary groups, it
1162 will only be able to drop these groups if started with superuser privileges.
Willy Tarreau6a06a402007-07-15 20:15:28 +02001163 See also "group" and "uid".
Willy Tarreaud72758d2010-01-12 10:42:19 +01001164
Willy Tarreau11770ce2019-12-03 08:29:22 +01001165group <group name>
1166 Similar to "gid" but uses the GID of group name <group name> from /etc/group.
1167 See also "gid" and "user".
1168
Cyril Bonté203ec5a2017-03-23 22:44:13 +01001169hard-stop-after <time>
1170 Defines the maximum time allowed to perform a clean soft-stop.
1171
1172 Arguments :
1173 <time> is the maximum time (by default in milliseconds) for which the
1174 instance will remain alive when a soft-stop is received via the
1175 SIGUSR1 signal.
1176
1177 This may be used to ensure that the instance will quit even if connections
1178 remain opened during a soft-stop (for example with long timeouts for a proxy
1179 in tcp mode). It applies both in TCP and HTTP mode.
1180
1181 Example:
1182 global
1183 hard-stop-after 30s
1184
Christopher Faulet98fbe952019-07-22 16:18:24 +02001185h1-case-adjust <from> <to>
1186 Defines the case adjustment to apply, when enabled, to the header name
1187 <from>, to change it to <to> before sending it to HTTP/1 clients or
1188 servers. <from> must be in lower case, and <from> and <to> must not differ
1189 except for their case. It may be repeated if several header names need to be
Ilya Shipitsin8525fd92020-02-29 12:34:59 +05001190 adjusted. Duplicate entries are not allowed. If a lot of header names have to
Christopher Faulet98fbe952019-07-22 16:18:24 +02001191 be adjusted, it might be more convenient to use "h1-case-adjust-file".
1192 Please note that no transformation will be applied unless "option
1193 h1-case-adjust-bogus-client" or "option h1-case-adjust-bogus-server" is
1194 specified in a proxy.
1195
1196 There is no standard case for header names because, as stated in RFC7230,
1197 they are case-insensitive. So applications must handle them in a case-
1198 insensitive manner. But some bogus applications violate the standards and
1199 erroneously rely on the cases most commonly used by browsers. This problem
1200 becomes critical with HTTP/2 because all header names must be exchanged in
1201 lower case, and HAProxy follows the same convention. All header names are
1202 sent in lower case to clients and servers, regardless of the HTTP version.
1203
1204 Applications which fail to properly process requests or responses may require
1205 to temporarily use such workarounds to adjust header names sent to them for
1206 the time it takes the application to be fixed. Please note that an
1207 application which requires such workarounds might be vulnerable to content
1208 smuggling attacks and must absolutely be fixed.
1209
1210 Example:
1211 global
1212 h1-case-adjust content-length Content-Length
1213
1214 See "h1-case-adjust-file", "option h1-case-adjust-bogus-client" and
1215 "option h1-case-adjust-bogus-server".
1216
1217h1-case-adjust-file <hdrs-file>
1218 Defines a file containing a list of key/value pairs used to adjust the case
1219 of some header names before sending them to HTTP/1 clients or servers. The
1220 file <hdrs-file> must contain 2 header names per line. The first one must be
1221 in lower case and both must not differ except for their case. Lines which
1222 start with '#' are ignored, just like empty lines. Leading and trailing tabs
1223 and spaces are stripped. Duplicate entries are not allowed. Please note that
1224 no transformation will be applied unless "option h1-case-adjust-bogus-client"
1225 or "option h1-case-adjust-bogus-server" is specified in a proxy.
1226
1227 If this directive is repeated, only the last one will be processed. It is an
1228 alternative to the directive "h1-case-adjust" if a lot of header names need
1229 to be adjusted. Please read the risks associated with using this.
1230
1231 See "h1-case-adjust", "option h1-case-adjust-bogus-client" and
1232 "option h1-case-adjust-bogus-server".
1233
Willy Tarreaud96f1122019-12-03 07:07:36 +01001234insecure-fork-wanted
1235 By default haproxy tries hard to prevent any thread and process creation
1236 after it starts. Doing so is particularly important when using Lua files of
1237 uncertain origin, and when experimenting with development versions which may
1238 still contain bugs whose exploitability is uncertain. And generally speaking
1239 it's good hygiene to make sure that no unexpected background activity can be
1240 triggered by traffic. But this prevents external checks from working, and may
1241 break some very specific Lua scripts which actively rely on the ability to
1242 fork. This option is there to disable this protection. Note that it is a bad
1243 idea to disable it, as a vulnerability in a library or within haproxy itself
1244 will be easier to exploit once disabled. In addition, forking from Lua or
1245 anywhere else is not reliable as the forked process may randomly embed a lock
1246 set by another thread and never manage to finish an operation. As such it is
1247 highly recommended that this option is never used and that any workload
1248 requiring such a fork be reconsidered and moved to a safer solution (such as
1249 agents instead of external checks). This option supports the "no" prefix to
1250 disable it.
1251
Willy Tarreaua45a8b52019-12-06 16:31:45 +01001252insecure-setuid-wanted
1253 HAProxy doesn't need to call executables at run time (except when using
1254 external checks which are strongly recommended against), and is even expected
1255 to isolate itself into an empty chroot. As such, there basically is no valid
1256 reason to allow a setuid executable to be called without the user being fully
1257 aware of the risks. In a situation where haproxy would need to call external
1258 checks and/or disable chroot, exploiting a vulnerability in a library or in
1259 haproxy itself could lead to the execution of an external program. On Linux
1260 it is possible to lock the process so that any setuid bit present on such an
1261 executable is ignored. This significantly reduces the risk of privilege
1262 escalation in such a situation. This is what haproxy does by default. In case
1263 this causes a problem to an external check (for example one which would need
1264 the "ping" command), then it is possible to disable this protection by
1265 explicitly adding this directive in the global section. If enabled, it is
1266 possible to turn it back off by prefixing it with the "no" keyword.
1267
Emmanuel Hocdet70df7bf2019-01-04 11:08:20 +01001268issuers-chain-path <dir>
1269 Assigns a directory to load certificate chain for issuer completion. All
1270 files must be in PEM format. For certificates loaded with "crt" or "crt-list",
1271 if certificate chain is not included in PEM (also commonly known as
1272 intermediate certificate), haproxy will complete chain if the issuer of the
1273 certificate corresponds to the first certificate of the chain loaded with
1274 "issuers-chain-path".
1275 A "crt" file with PrivateKey+Certificate+IntermediateCA2+IntermediateCA1
1276 could be replaced with PrivateKey+Certificate. HAProxy will complete the
1277 chain if a file with IntermediateCA2+IntermediateCA1 is present in
1278 "issuers-chain-path" directory. All other certificates with the same issuer
1279 will share the chain in memory.
1280
Dragan Dosen13cd54c2020-06-18 18:24:05 +02001281localpeer <name>
1282 Sets the local instance's peer name. It will be ignored if the "-L"
1283 command line argument is specified or if used after "peers" section
1284 definitions. In such cases, a warning message will be emitted during
1285 the configuration parsing.
1286
1287 This option will also set the HAPROXY_LOCALPEER environment variable.
1288 See also "-L" in the management guide and "peers" section below.
1289
Jan Wagner3e678602020-12-17 22:22:32 +01001290log <address> [len <length>] [format <format>] [sample <ranges>:<sample_size>]
Frédéric Lécailled690dfa2019-04-25 10:52:17 +02001291 <facility> [max level [min level]]
Cyril Bonté3e954872018-03-20 23:30:27 +01001292 Adds a global syslog server. Several global servers can be defined. They
Davor Ocelice9ed2812017-12-25 17:49:28 +01001293 will receive logs for starts and exits, as well as all logs from proxies
Robert Tsai81ae1952007-12-05 10:47:29 +01001294 configured with "log global".
1295
1296 <address> can be one of:
1297
Willy Tarreau2769aa02007-12-27 18:26:09 +01001298 - An IPv4 address optionally followed by a colon and a UDP port. If
Robert Tsai81ae1952007-12-05 10:47:29 +01001299 no port is specified, 514 is used by default (the standard syslog
1300 port).
1301
David du Colombier24bb5f52011-03-17 10:40:23 +01001302 - An IPv6 address followed by a colon and optionally a UDP port. If
1303 no port is specified, 514 is used by default (the standard syslog
1304 port).
1305
Willy Tarreau5a32ecc2018-11-12 07:34:59 +01001306 - A filesystem path to a datagram UNIX domain socket, keeping in mind
Robert Tsai81ae1952007-12-05 10:47:29 +01001307 considerations for chroot (be sure the path is accessible inside
1308 the chroot) and uid/gid (be sure the path is appropriately
Davor Ocelice9ed2812017-12-25 17:49:28 +01001309 writable).
Robert Tsai81ae1952007-12-05 10:47:29 +01001310
Willy Tarreau5a32ecc2018-11-12 07:34:59 +01001311 - A file descriptor number in the form "fd@<number>", which may point
1312 to a pipe, terminal, or socket. In this case unbuffered logs are used
1313 and one writev() call per log is performed. This is a bit expensive
1314 but acceptable for most workloads. Messages sent this way will not be
1315 truncated but may be dropped, in which case the DroppedLogs counter
1316 will be incremented. The writev() call is atomic even on pipes for
1317 messages up to PIPE_BUF size, which POSIX recommends to be at least
1318 512 and which is 4096 bytes on most modern operating systems. Any
1319 larger message may be interleaved with messages from other processes.
1320 Exceptionally for debugging purposes the file descriptor may also be
1321 directed to a file, but doing so will significantly slow haproxy down
1322 as non-blocking calls will be ignored. Also there will be no way to
1323 purge nor rotate this file without restarting the process. Note that
1324 the configured syslog format is preserved, so the output is suitable
Willy Tarreauc1b06452018-11-12 11:57:56 +01001325 for use with a TCP syslog server. See also the "short" and "raw"
1326 format below.
Willy Tarreau5a32ecc2018-11-12 07:34:59 +01001327
1328 - "stdout" / "stderr", which are respectively aliases for "fd@1" and
1329 "fd@2", see above.
1330
Willy Tarreauc046d162019-08-30 15:24:59 +02001331 - A ring buffer in the form "ring@<name>", which will correspond to an
1332 in-memory ring buffer accessible over the CLI using the "show events"
1333 command, which will also list existing rings and their sizes. Such
1334 buffers are lost on reload or restart but when used as a complement
1335 this can help troubleshooting by having the logs instantly available.
1336
William Lallemandb2f07452015-05-12 14:27:13 +02001337 You may want to reference some environment variables in the address
1338 parameter, see section 2.3 about environment variables.
Willy Tarreaudad36a32013-03-11 01:20:04 +01001339
Willy Tarreau18324f52014-06-27 18:10:07 +02001340 <length> is an optional maximum line length. Log lines larger than this value
1341 will be truncated before being sent. The reason is that syslog
1342 servers act differently on log line length. All servers support the
1343 default value of 1024, but some servers simply drop larger lines
1344 while others do log them. If a server supports long lines, it may
1345 make sense to set this value here in order to avoid truncating long
1346 lines. Similarly, if a server drops long lines, it is preferable to
1347 truncate them before sending them. Accepted values are 80 to 65535
1348 inclusive. The default value of 1024 is generally fine for all
1349 standard usages. Some specific cases of long captures or
Davor Ocelice9ed2812017-12-25 17:49:28 +01001350 JSON-formatted logs may require larger values. You may also need to
1351 increase "tune.http.logurilen" if your request URIs are truncated.
Willy Tarreau18324f52014-06-27 18:10:07 +02001352
Dragan Dosen7ad31542015-09-28 17:16:47 +02001353 <format> is the log format used when generating syslog messages. It may be
1354 one of the following :
1355
Emeric Brun0237c4e2020-11-27 16:24:34 +01001356 local Analog to rfc3164 syslog message format except that hostname
1357 field is stripped. This is the default.
1358 Note: option "log-send-hostname" switches the default to
1359 rfc3164.
1360
1361 rfc3164 The RFC3164 syslog message format.
Dragan Dosen7ad31542015-09-28 17:16:47 +02001362 (https://tools.ietf.org/html/rfc3164)
1363
1364 rfc5424 The RFC5424 syslog message format.
1365 (https://tools.ietf.org/html/rfc5424)
1366
Emeric Brun54648852020-07-06 15:54:06 +02001367 priority A message containing only a level plus syslog facility between
1368 angle brackets such as '<63>', followed by the text. The PID,
1369 date, time, process name and system name are omitted. This is
1370 designed to be used with a local log server.
1371
Willy Tarreaue8746a02018-11-12 08:45:00 +01001372 short A message containing only a level between angle brackets such as
1373 '<3>', followed by the text. The PID, date, time, process name
1374 and system name are omitted. This is designed to be used with a
1375 local log server. This format is compatible with what the systemd
1376 logger consumes.
1377
Emeric Brun54648852020-07-06 15:54:06 +02001378 timed A message containing only a level between angle brackets such as
1379 '<3>', followed by ISO date and by the text. The PID, process
1380 name and system name are omitted. This is designed to be
1381 used with a local log server.
1382
1383 iso A message containing only the ISO date, followed by the text.
1384 The PID, process name and system name are omitted. This is
1385 designed to be used with a local log server.
1386
Willy Tarreauc1b06452018-11-12 11:57:56 +01001387 raw A message containing only the text. The level, PID, date, time,
1388 process name and system name are omitted. This is designed to be
1389 used in containers or during development, where the severity only
1390 depends on the file descriptor used (stdout/stderr).
1391
Frédéric Lécailled690dfa2019-04-25 10:52:17 +02001392 <ranges> A list of comma-separated ranges to identify the logs to sample.
1393 This is used to balance the load of the logs to send to the log
1394 server. The limits of the ranges cannot be null. They are numbered
1395 from 1. The size or period (in number of logs) of the sample must be
1396 set with <sample_size> parameter.
1397
1398 <sample_size>
1399 The size of the sample in number of logs to consider when balancing
1400 their logging loads. It is used to balance the load of the logs to
1401 send to the syslog server. This size must be greater or equal to the
1402 maximum of the high limits of the ranges.
1403 (see also <ranges> parameter).
1404
Robert Tsai81ae1952007-12-05 10:47:29 +01001405 <facility> must be one of the 24 standard syslog facilities :
Willy Tarreau6a06a402007-07-15 20:15:28 +02001406
Willy Tarreaue8746a02018-11-12 08:45:00 +01001407 kern user mail daemon auth syslog lpr news
1408 uucp cron auth2 ftp ntp audit alert cron2
1409 local0 local1 local2 local3 local4 local5 local6 local7
1410
Willy Tarreauc1b06452018-11-12 11:57:56 +01001411 Note that the facility is ignored for the "short" and "raw"
1412 formats, but still required as a positional field. It is
1413 recommended to use "daemon" in this case to make it clear that
1414 it's only supposed to be used locally.
Willy Tarreau6a06a402007-07-15 20:15:28 +02001415
1416 An optional level can be specified to filter outgoing messages. By default,
Willy Tarreauf7edefa2009-05-10 17:20:05 +02001417 all messages are sent. If a maximum level is specified, only messages with a
1418 severity at least as important as this level will be sent. An optional minimum
1419 level can be specified. If it is set, logs emitted with a more severe level
1420 than this one will be capped to this level. This is used to avoid sending
1421 "emerg" messages on all terminals on some default syslog configurations.
1422 Eight levels are known :
Willy Tarreau6a06a402007-07-15 20:15:28 +02001423
Cyril Bontédc4d9032012-04-08 21:57:39 +02001424 emerg alert crit err warning notice info debug
Willy Tarreau6a06a402007-07-15 20:15:28 +02001425
Joe Williamsdf5b38f2010-12-29 17:05:48 +01001426log-send-hostname [<string>]
1427 Sets the hostname field in the syslog header. If optional "string" parameter
1428 is set the header is set to the string contents, otherwise uses the hostname
1429 of the system. Generally used if one is not relaying logs through an
1430 intermediate syslog server or for simply customizing the hostname printed in
1431 the logs.
1432
Kevinm48936af2010-12-22 16:08:21 +00001433log-tag <string>
1434 Sets the tag field in the syslog header to this string. It defaults to the
1435 program name as launched from the command line, which usually is "haproxy".
1436 Sometimes it can be useful to differentiate between multiple processes
Willy Tarreau094af4e2015-01-07 15:03:42 +01001437 running on the same host. See also the per-proxy "log-tag" directive.
Kevinm48936af2010-12-22 16:08:21 +00001438
Thierry FOURNIER90da1912015-03-05 11:17:06 +01001439lua-load <file>
Thierry Fournier59f11be2020-11-29 00:37:41 +01001440 This global directive loads and executes a Lua file in the shared context
1441 that is visible to all threads. Any variable set in such a context is visible
1442 from any thread. This is the easiest and recommended way to load Lua programs
1443 but it will not scale well if a lot of Lua calls are performed, as only one
1444 thread may be running on the global state at a time. A program loaded this
1445 way will always see 0 in the "core.thread" variable. This directive can be
Thierry FOURNIER90da1912015-03-05 11:17:06 +01001446 used multiple times.
1447
Thierry Fournier59f11be2020-11-29 00:37:41 +01001448lua-load-per-thread <file>
1449 This global directive loads and executes a Lua file into each started thread.
1450 Any global variable has a thread-local visibility so that each thread could
1451 see a different value. As such it is strongly recommended not to use global
1452 variables in programs loaded this way. An independent copy is loaded and
1453 initialized for each thread, everything is done sequentially and in the
1454 thread's numeric order from 1 to nbthread. If some operations need to be
1455 performed only once, the program should check the "core.thread" variable to
1456 figure what thread is being initialized. Programs loaded this way will run
1457 concurrently on all threads and will be highly scalable. This is the
1458 recommended way to load simple functions that register sample-fetches,
1459 converters, actions or services once it is certain the program doesn't depend
1460 on global variables. For the sake of simplicity, the directive is available
1461 even if only one thread is used and even if threads are disabled (in which
1462 case it will be equivalent to lua-load). This directive can be used multiple
1463 times.
1464
Tim Duesterhusdd74b5f2020-01-12 13:55:40 +01001465lua-prepend-path <string> [<type>]
1466 Prepends the given string followed by a semicolon to Lua's package.<type>
1467 variable.
1468 <type> must either be "path" or "cpath". If <type> is not given it defaults
1469 to "path".
1470
1471 Lua's paths are semicolon delimited lists of patterns that specify how the
1472 `require` function attempts to find the source file of a library. Question
1473 marks (?) within a pattern will be replaced by module name. The path is
1474 evaluated left to right. This implies that paths that are prepended later
1475 will be checked earlier.
1476
1477 As an example by specifying the following path:
1478
1479 lua-prepend-path /usr/share/haproxy-lua/?/init.lua
1480 lua-prepend-path /usr/share/haproxy-lua/?.lua
1481
1482 When `require "example"` is being called Lua will first attempt to load the
1483 /usr/share/haproxy-lua/example.lua script, if that does not exist the
1484 /usr/share/haproxy-lua/example/init.lua will be attempted and the default
1485 paths if that does not exist either.
1486
1487 See https://www.lua.org/pil/8.1.html for the details within the Lua
1488 documentation.
1489
William Lallemand4cfede82017-11-24 22:02:34 +01001490master-worker [no-exit-on-failure]
William Lallemande202b1e2017-06-01 17:38:56 +02001491 Master-worker mode. It is equivalent to the command line "-W" argument.
1492 This mode will launch a "master" which will monitor the "workers". Using
1493 this mode, you can reload HAProxy directly by sending a SIGUSR2 signal to
Davor Ocelice9ed2812017-12-25 17:49:28 +01001494 the master. The master-worker mode is compatible either with the foreground
William Lallemande202b1e2017-06-01 17:38:56 +02001495 or daemon mode. It is recommended to use this mode with multiprocess and
1496 systemd.
William Lallemand4cfede82017-11-24 22:02:34 +01001497 By default, if a worker exits with a bad return code, in the case of a
1498 segfault for example, all workers will be killed, and the master will leave.
1499 It is convenient to combine this behavior with Restart=on-failure in a
1500 systemd unit file in order to relaunch the whole process. If you don't want
1501 this behavior, you must use the keyword "no-exit-on-failure".
William Lallemande202b1e2017-06-01 17:38:56 +02001502
William Lallemand4cfede82017-11-24 22:02:34 +01001503 See also "-W" in the management guide.
William Lallemande202b1e2017-06-01 17:38:56 +02001504
William Lallemand27edc4b2019-05-07 17:49:33 +02001505mworker-max-reloads <number>
1506 In master-worker mode, this option limits the number of time a worker can
John Roeslerfb2fce12019-07-10 15:45:51 -05001507 survive to a reload. If the worker did not leave after a reload, once its
William Lallemand27edc4b2019-05-07 17:49:33 +02001508 number of reloads is greater than this number, the worker will receive a
1509 SIGTERM. This option helps to keep under control the number of workers.
1510 See also "show proc" in the Management Guide.
1511
Willy Tarreauf42d7942020-10-20 11:54:49 +02001512nbproc <number> (deprecated)
Willy Tarreau6a06a402007-07-15 20:15:28 +02001513 Creates <number> processes when going daemon. This requires the "daemon"
1514 mode. By default, only one process is created, which is the recommended mode
1515 of operation. For systems limited to small sets of file descriptors per
Willy Tarreau149ab772019-01-26 14:27:06 +01001516 process, it may be needed to fork multiple daemons. When set to a value
1517 larger than 1, threads are automatically disabled. USING MULTIPLE PROCESSES
Willy Tarreauf42d7942020-10-20 11:54:49 +02001518 IS HARDER TO DEBUG AND IS REALLY DISCOURAGED. This directive is deprecated
1519 and scheduled for removal in 2.5. Please use "nbthread" instead. See also
1520 "daemon" and "nbthread".
Willy Tarreau6a06a402007-07-15 20:15:28 +02001521
Christopher Fauletbe0faa22017-08-29 15:37:10 +02001522nbthread <number>
1523 This setting is only available when support for threads was built in. It
Willy Tarreau26f6ae12019-02-02 12:56:15 +01001524 makes haproxy run on <number> threads. This is exclusive with "nbproc". While
1525 "nbproc" historically used to be the only way to use multiple processors, it
1526 also involved a number of shortcomings related to the lack of synchronization
1527 between processes (health-checks, peers, stick-tables, stats, ...) which do
1528 not affect threads. As such, any modern configuration is strongly encouraged
Willy Tarreau149ab772019-01-26 14:27:06 +01001529 to migrate away from "nbproc" to "nbthread". "nbthread" also works when
1530 HAProxy is started in foreground. On some platforms supporting CPU affinity,
1531 when nbproc is not used, the default "nbthread" value is automatically set to
1532 the number of CPUs the process is bound to upon startup. This means that the
1533 thread count can easily be adjusted from the calling process using commands
1534 like "taskset" or "cpuset". Otherwise, this value defaults to 1. The default
1535 value is reported in the output of "haproxy -vv". See also "nbproc".
Christopher Fauletbe0faa22017-08-29 15:37:10 +02001536
Willy Tarreau6a06a402007-07-15 20:15:28 +02001537pidfile <pidfile>
MIZUTA Takeshic32f3942020-08-26 13:46:19 +09001538 Writes PIDs of all daemons into file <pidfile> when daemon mode or writes PID
1539 of master process into file <pidfile> when master-worker mode. This option is
1540 equivalent to the "-p" command line argument. The file must be accessible to
1541 the user starting the process. See also "daemon" and "master-worker".
Willy Tarreau6a06a402007-07-15 20:15:28 +02001542
Willy Tarreau119e50e2020-05-22 13:53:29 +02001543pp2-never-send-local
1544 A bug in the PROXY protocol v2 implementation was present in HAProxy up to
1545 version 2.1, causing it to emit a PROXY command instead of a LOCAL command
1546 for health checks. This is particularly minor but confuses some servers'
1547 logs. Sadly, the bug was discovered very late and revealed that some servers
1548 which possibly only tested their PROXY protocol implementation against
1549 HAProxy fail to properly handle the LOCAL command, and permanently remain in
1550 the "down" state when HAProxy checks them. When this happens, it is possible
1551 to enable this global option to revert to the older (bogus) behavior for the
1552 time it takes to contact the affected components' vendors and get them fixed.
1553 This option is disabled by default and acts on all servers having the
1554 "send-proxy-v2" statement.
1555
Willy Tarreau1d549722016-02-16 12:41:57 +01001556presetenv <name> <value>
1557 Sets environment variable <name> to value <value>. If the variable exists, it
1558 is NOT overwritten. The changes immediately take effect so that the next line
1559 in the configuration file sees the new value. See also "setenv", "resetenv",
1560 and "unsetenv".
1561
1562resetenv [<name> ...]
1563 Removes all environment variables except the ones specified in argument. It
1564 allows to use a clean controlled environment before setting new values with
1565 setenv or unsetenv. Please note that some internal functions may make use of
1566 some environment variables, such as time manipulation functions, but also
1567 OpenSSL or even external checks. This must be used with extreme care and only
1568 after complete validation. The changes immediately take effect so that the
1569 next line in the configuration file sees the new environment. See also
1570 "setenv", "presetenv", and "unsetenv".
1571
Christopher Fauletff4121f2017-11-22 16:38:49 +01001572stats bind-process [ all | odd | even | <process_num>[-[process_num>]] ] ...
Willy Tarreau35b7b162012-10-22 23:17:18 +02001573 Limits the stats socket to a certain set of processes numbers. By default the
1574 stats socket is bound to all processes, causing a warning to be emitted when
1575 nbproc is greater than 1 because there is no way to select the target process
1576 when connecting. However, by using this setting, it becomes possible to pin
1577 the stats socket to a specific set of processes, typically the first one. The
1578 warning will automatically be disabled when this setting is used, whatever
Willy Tarreaua9db57e2013-01-18 11:29:29 +01001579 the number of processes used. The maximum process ID depends on the machine's
Christopher Fauletff4121f2017-11-22 16:38:49 +01001580 word size (32 or 64). Ranges can be partially defined. The higher bound can
1581 be omitted. In such case, it is replaced by the corresponding maximum
1582 value. A better option consists in using the "process" setting of the "stats
1583 socket" line to force the process on each line.
Willy Tarreau35b7b162012-10-22 23:17:18 +02001584
Baptiste Assmann5626f482015-08-23 10:00:10 +02001585server-state-base <directory>
1586 Specifies the directory prefix to be prepended in front of all servers state
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02001587 file names which do not start with a '/'. See also "server-state-file",
1588 "load-server-state-from-file" and "server-state-file-name".
Baptiste Assmannef1f0fc2015-08-23 10:06:39 +02001589
1590server-state-file <file>
1591 Specifies the path to the file containing state of servers. If the path starts
1592 with a slash ('/'), it is considered absolute, otherwise it is considered
1593 relative to the directory specified using "server-state-base" (if set) or to
1594 the current directory. Before reloading HAProxy, it is possible to save the
1595 servers' current state using the stats command "show servers state". The
1596 output of this command must be written in the file pointed by <file>. When
1597 starting up, before handling traffic, HAProxy will read, load and apply state
1598 for each server found in the file and available in its current running
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02001599 configuration. See also "server-state-base" and "show servers state",
1600 "load-server-state-from-file" and "server-state-file-name"
Baptiste Assmann5626f482015-08-23 10:00:10 +02001601
Willy Tarreau13d2ba22021-03-26 11:38:08 +01001602set-var <var-name> <expr>
1603 Sets the process-wide variable '<var-name>' to the result of the evaluation
1604 of the sample expression <expr>. The variable '<var-name>' may only be a
1605 process-wide variable (using the 'proc.' prefix). It works exactly like the
1606 'set-var' action in TCP or HTTP rules except that the expression is evaluated
1607 at configuration parsing time and that the variable is instantly set. The
1608 sample fetch functions and converters permitted in the expression are only
1609 those using internal data, typically 'int(value)' or 'str(value)'. It's is
1610 possible to reference previously allocated variables as well. These variables
1611 will then be readable (and modifiable) from the regular rule sets.
1612
1613 Example:
1614 global
1615 set-var proc.current_state str(primary)
1616 set-var proc.prio int(100)
1617 set-var proc.threshold int(200),sub(proc.prio)
1618
Willy Tarreau1d549722016-02-16 12:41:57 +01001619setenv <name> <value>
1620 Sets environment variable <name> to value <value>. If the variable exists, it
1621 is overwritten. The changes immediately take effect so that the next line in
1622 the configuration file sees the new value. See also "presetenv", "resetenv",
1623 and "unsetenv".
1624
Willy Tarreau636848a2019-04-15 19:38:50 +02001625set-dumpable
1626 This option is better left disabled by default and enabled only upon a
William Dauchyec730982019-10-27 20:08:10 +01001627 developer's request. If it has been enabled, it may still be forcibly
1628 disabled by prefixing it with the "no" keyword. It has no impact on
1629 performance nor stability but will try hard to re-enable core dumps that were
1630 possibly disabled by file size limitations (ulimit -f), core size limitations
1631 (ulimit -c), or "dumpability" of a process after changing its UID/GID (such
1632 as /proc/sys/fs/suid_dumpable on Linux). Core dumps might still be limited by
1633 the current directory's permissions (check what directory the file is started
1634 from), the chroot directory's permission (it may be needed to temporarily
1635 disable the chroot directive or to move it to a dedicated writable location),
1636 or any other system-specific constraint. For example, some Linux flavours are
1637 notorious for replacing the default core file with a path to an executable
1638 not even installed on the system (check /proc/sys/kernel/core_pattern). Often,
1639 simply writing "core", "core.%p" or "/var/log/core/core.%p" addresses the
1640 issue. When trying to enable this option waiting for a rare issue to
1641 re-appear, it's often a good idea to first try to obtain such a dump by
1642 issuing, for example, "kill -11" to the haproxy process and verify that it
1643 leaves a core where expected when dying.
Willy Tarreau636848a2019-04-15 19:38:50 +02001644
Willy Tarreau610f04b2014-02-13 11:36:41 +01001645ssl-default-bind-ciphers <ciphers>
1646 This setting is only available when support for OpenSSL was built in. It sets
1647 the default string describing the list of cipher algorithms ("cipher suite")
Bertrand Jacquin8cf7c1e2019-02-03 18:35:25 +00001648 that are negotiated during the SSL/TLS handshake up to TLSv1.2 for all
Dirkjan Bussink415150f2018-09-14 11:14:21 +02001649 "bind" lines which do not explicitly define theirs. The format of the string
Bertrand Jacquin4f03ab02019-02-03 18:48:49 +00001650 is defined in "man 1 ciphers" from OpenSSL man pages. For background
1651 information and recommendations see e.g.
1652 (https://wiki.mozilla.org/Security/Server_Side_TLS) and
1653 (https://mozilla.github.io/server-side-tls/ssl-config-generator/). For TLSv1.3
1654 cipher configuration, please check the "ssl-default-bind-ciphersuites" keyword.
1655 Please check the "bind" keyword for more information.
Dirkjan Bussink415150f2018-09-14 11:14:21 +02001656
1657ssl-default-bind-ciphersuites <ciphersuites>
1658 This setting is only available when support for OpenSSL was built in and
1659 OpenSSL 1.1.1 or later was used to build HAProxy. It sets the default string
1660 describing the list of cipher algorithms ("cipher suite") that are negotiated
1661 during the TLSv1.3 handshake for all "bind" lines which do not explicitly define
1662 theirs. The format of the string is defined in
Bertrand Jacquin4f03ab02019-02-03 18:48:49 +00001663 "man 1 ciphers" from OpenSSL man pages under the section "ciphersuites". For
1664 cipher configuration for TLSv1.2 and earlier, please check the
1665 "ssl-default-bind-ciphers" keyword. Please check the "bind" keyword for more
Dirkjan Bussink415150f2018-09-14 11:14:21 +02001666 information.
Willy Tarreau610f04b2014-02-13 11:36:41 +01001667
Jerome Magninb203ff62020-04-03 15:28:22 +02001668ssl-default-bind-curves <curves>
1669 This setting is only available when support for OpenSSL was built in. It sets
1670 the default string describing the list of elliptic curves algorithms ("curve
1671 suite") that are negotiated during the SSL/TLS handshake with ECDHE. The format
1672 of the string is a colon-delimited list of curve name.
1673 Please check the "bind" keyword for more information.
1674
Emeric Brun2c86cbf2014-10-30 15:56:50 +01001675ssl-default-bind-options [<option>]...
1676 This setting is only available when support for OpenSSL was built in. It sets
1677 default ssl-options to force on all "bind" lines. Please check the "bind"
1678 keyword to see available options.
1679
1680 Example:
1681 global
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +02001682 ssl-default-bind-options ssl-min-ver TLSv1.0 no-tls-tickets
Emeric Brun2c86cbf2014-10-30 15:56:50 +01001683
Willy Tarreau610f04b2014-02-13 11:36:41 +01001684ssl-default-server-ciphers <ciphers>
1685 This setting is only available when support for OpenSSL was built in. It
1686 sets the default string describing the list of cipher algorithms that are
Bertrand Jacquin8cf7c1e2019-02-03 18:35:25 +00001687 negotiated during the SSL/TLS handshake up to TLSv1.2 with the server,
Dirkjan Bussink415150f2018-09-14 11:14:21 +02001688 for all "server" lines which do not explicitly define theirs. The format of
Bertrand Jacquin4f03ab02019-02-03 18:48:49 +00001689 the string is defined in "man 1 ciphers" from OpenSSL man pages. For background
1690 information and recommendations see e.g.
1691 (https://wiki.mozilla.org/Security/Server_Side_TLS) and
1692 (https://mozilla.github.io/server-side-tls/ssl-config-generator/).
1693 For TLSv1.3 cipher configuration, please check the
1694 "ssl-default-server-ciphersuites" keyword. Please check the "server" keyword
1695 for more information.
Dirkjan Bussink415150f2018-09-14 11:14:21 +02001696
1697ssl-default-server-ciphersuites <ciphersuites>
1698 This setting is only available when support for OpenSSL was built in and
1699 OpenSSL 1.1.1 or later was used to build HAProxy. It sets the default
1700 string describing the list of cipher algorithms that are negotiated during
1701 the TLSv1.3 handshake with the server, for all "server" lines which do not
1702 explicitly define theirs. The format of the string is defined in
Bertrand Jacquin4f03ab02019-02-03 18:48:49 +00001703 "man 1 ciphers" from OpenSSL man pages under the section "ciphersuites". For
1704 cipher configuration for TLSv1.2 and earlier, please check the
1705 "ssl-default-server-ciphers" keyword. Please check the "server" keyword for
1706 more information.
Willy Tarreau610f04b2014-02-13 11:36:41 +01001707
Emeric Brun2c86cbf2014-10-30 15:56:50 +01001708ssl-default-server-options [<option>]...
1709 This setting is only available when support for OpenSSL was built in. It sets
1710 default ssl-options to force on all "server" lines. Please check the "server"
1711 keyword to see available options.
1712
Remi Gacogne47783ef2015-05-29 15:53:22 +02001713ssl-dh-param-file <file>
1714 This setting is only available when support for OpenSSL was built in. It sets
1715 the default DH parameters that are used during the SSL/TLS handshake when
1716 ephemeral Diffie-Hellman (DHE) key exchange is used, for all "bind" lines
Davor Ocelice9ed2812017-12-25 17:49:28 +01001717 which do not explicitly define theirs. It will be overridden by custom DH
Remi Gacogne47783ef2015-05-29 15:53:22 +02001718 parameters found in a bind certificate file if any. If custom DH parameters
Cyril Bonté307ee1e2015-09-28 23:16:06 +02001719 are not specified either by using ssl-dh-param-file or by setting them
1720 directly in the certificate file, pre-generated DH parameters of the size
1721 specified by tune.ssl.default-dh-param will be used. Custom parameters are
1722 known to be more secure and therefore their use is recommended.
Remi Gacogne47783ef2015-05-29 15:53:22 +02001723 Custom DH parameters may be generated by using the OpenSSL command
1724 "openssl dhparam <size>", where size should be at least 2048, as 1024-bit DH
1725 parameters should not be considered secure anymore.
1726
William Lallemand8e8581e2020-10-20 17:36:46 +02001727ssl-load-extra-del-ext
1728 This setting allows to configure the way HAProxy does the lookup for the
1729 extra SSL files. By default HAProxy adds a new extension to the filename.
William Lallemand089c1382020-10-23 17:35:12 +02001730 (ex: with "foobar.crt" load "foobar.crt.key"). With this option enabled,
William Lallemand8e8581e2020-10-20 17:36:46 +02001731 HAProxy removes the extension before adding the new one (ex: with
William Lallemand089c1382020-10-23 17:35:12 +02001732 "foobar.crt" load "foobar.key").
1733
1734 Your crt file must have a ".crt" extension for this option to work.
William Lallemand8e8581e2020-10-20 17:36:46 +02001735
1736 This option is not compatible with bundle extensions (.ecdsa, .rsa. .dsa)
1737 and won't try to remove them.
1738
1739 This option is disabled by default. See also "ssl-load-extra-files".
1740
William Lallemand4c5adbf2020-02-24 14:23:22 +01001741ssl-load-extra-files <none|all|bundle|sctl|ocsp|issuer|key>*
William Lallemand3af48e72020-02-03 17:15:52 +01001742 This setting alters the way HAProxy will look for unspecified files during
Jerome Magnin587be9c2020-09-07 11:55:57 +02001743 the loading of the SSL certificates associated to "bind" lines. It does not
1744 apply to certificates used for client authentication on "server" lines.
William Lallemand3af48e72020-02-03 17:15:52 +01001745
1746 By default, HAProxy discovers automatically a lot of files not specified in
1747 the configuration, and you may want to disable this behavior if you want to
1748 optimize the startup time.
1749
1750 "none": Only load the files specified in the configuration. Don't try to load
1751 a certificate bundle if the file does not exist. In the case of a directory,
1752 it won't try to bundle the certificates if they have the same basename.
1753
1754 "all": This is the default behavior, it will try to load everything,
William Lallemand4c5adbf2020-02-24 14:23:22 +01001755 bundles, sctl, ocsp, issuer, key.
William Lallemand3af48e72020-02-03 17:15:52 +01001756
1757 "bundle": When a file specified in the configuration does not exist, HAProxy
William Lallemandf9ff3ec2020-10-02 17:57:44 +02001758 will try to load a "cert bundle".
1759
1760 Starting from HAProxy 2.3, the bundles are not loaded in the same OpenSSL
1761 certificate store, instead it will loads each certificate in a separate
1762 store which is equivalent to declaring multiple "crt". OpenSSL 1.1.1 is
1763 required to achieve this. Which means that bundles are now used only for
1764 backward compatibility and are not mandatory anymore to do an hybrid RSA/ECC
1765 bind configuration..
1766
1767 To associate these PEM files into a "cert bundle" that is recognized by
1768 haproxy, they must be named in the following way: All PEM files that are to
1769 be bundled must have the same base name, with a suffix indicating the key
1770 type. Currently, three suffixes are supported: rsa, dsa and ecdsa. For
1771 example, if www.example.com has two PEM files, an RSA file and an ECDSA
1772 file, they must be named: "example.pem.rsa" and "example.pem.ecdsa". The
1773 first part of the filename is arbitrary; only the suffix matters. To load
1774 this bundle into haproxy, specify the base name only:
1775
1776 Example : bind :8443 ssl crt example.pem
1777
1778 Note that the suffix is not given to haproxy; this tells haproxy to look for
1779 a cert bundle.
1780
1781 HAProxy will load all PEM files in the bundle as if they were configured
1782 separately in several "crt".
1783
1784 The bundle loading does not have an impact anymore on the directory loading
1785 since files are loading separately.
1786
1787 On the CLI, bundles are seen as separate files, and the bundle extension is
1788 required to commit them.
1789
William Dauchy57dd6f12020-10-06 15:22:37 +02001790 OCSP files (.ocsp), issuer files (.issuer), Certificate Transparency (.sctl)
William Lallemandf9ff3ec2020-10-02 17:57:44 +02001791 as well as private keys (.key) are supported with multi-cert bundling.
William Lallemand3af48e72020-02-03 17:15:52 +01001792
1793 "sctl": Try to load "<basename>.sctl" for each crt keyword.
1794
1795 "ocsp": Try to load "<basename>.ocsp" for each crt keyword.
1796
1797 "issuer": Try to load "<basename>.issuer" if the issuer of the OCSP file is
1798 not provided in the PEM file.
1799
William Lallemand4c5adbf2020-02-24 14:23:22 +01001800 "key": If the private key was not provided by the PEM file, try to load a
1801 file "<basename>.key" containing a private key.
1802
William Lallemand3af48e72020-02-03 17:15:52 +01001803 The default behavior is "all".
1804
1805 Example:
1806 ssl-load-extra-files bundle sctl
1807 ssl-load-extra-files sctl ocsp issuer
1808 ssl-load-extra-files none
1809
1810 See also: "crt", section 5.1 about bind options.
1811
Emeric Brun850efd52014-01-29 12:24:34 +01001812ssl-server-verify [none|required]
1813 The default behavior for SSL verify on servers side. If specified to 'none',
1814 servers certificates are not verified. The default is 'required' except if
1815 forced using cmdline option '-dV'.
1816
Emmanuel Hocdetc3b7e742020-04-22 11:06:19 +02001817ssl-skip-self-issued-ca
Daniel Corbett67a82712020-07-06 23:01:19 -04001818 Self issued CA, aka x509 root CA, is the anchor for chain validation: as a
Emmanuel Hocdetc3b7e742020-04-22 11:06:19 +02001819 server is useless to send it, client must have it. Standard configuration
1820 need to not include such CA in PEM file. This option allows you to keep such
1821 CA in PEM file without sending it to the client. Use case is to provide
1822 issuer for ocsp without the need for '.issuer' file and be able to share it
1823 with 'issuers-chain-path'. This concerns all certificates without intermediate
1824 certificates. It's useless for BoringSSL, .issuer is ignored because ocsp
William Lallemand9a1d8392020-08-10 17:28:23 +02001825 bits does not need it. Requires at least OpenSSL 1.0.2.
Emmanuel Hocdetc3b7e742020-04-22 11:06:19 +02001826
Willy Tarreauabb175f2012-09-24 12:43:26 +02001827stats socket [<address:port>|<path>] [param*]
1828 Binds a UNIX socket to <path> or a TCPv4/v6 address to <address:port>.
1829 Connections to this socket will return various statistics outputs and even
1830 allow some commands to be issued to change some runtime settings. Please
Willy Tarreau1af20c72017-06-23 16:01:14 +02001831 consult section 9.3 "Unix Socket commands" of Management Guide for more
Kevin Decherf949c7202015-10-13 23:26:44 +02001832 details.
Willy Tarreau6162db22009-10-10 17:13:00 +02001833
Willy Tarreauabb175f2012-09-24 12:43:26 +02001834 All parameters supported by "bind" lines are supported, for instance to
1835 restrict access to some users or their access rights. Please consult
1836 section 5.1 for more information.
Willy Tarreaufbee7132007-10-18 13:53:22 +02001837
1838stats timeout <timeout, in milliseconds>
1839 The default timeout on the stats socket is set to 10 seconds. It is possible
1840 to change this value with "stats timeout". The value must be passed in
Willy Tarreaubefdff12007-12-02 22:27:38 +01001841 milliseconds, or be suffixed by a time unit among { us, ms, s, m, h, d }.
Willy Tarreaufbee7132007-10-18 13:53:22 +02001842
1843stats maxconn <connections>
1844 By default, the stats socket is limited to 10 concurrent connections. It is
1845 possible to change this value with "stats maxconn".
1846
Willy Tarreau6a06a402007-07-15 20:15:28 +02001847uid <number>
Thayne McCombscdbcca92021-01-07 21:24:41 -07001848 Changes the process's user ID to <number>. It is recommended that the user ID
Willy Tarreau6a06a402007-07-15 20:15:28 +02001849 is dedicated to HAProxy or to a small set of similar daemons. HAProxy must
1850 be started with superuser privileges in order to be able to switch to another
1851 one. See also "gid" and "user".
1852
1853ulimit-n <number>
1854 Sets the maximum number of per-process file-descriptors to <number>. By
1855 default, it is automatically computed, so it is recommended not to use this
1856 option.
1857
Willy Tarreauceb24bc2010-11-09 12:46:41 +01001858unix-bind [ prefix <prefix> ] [ mode <mode> ] [ user <user> ] [ uid <uid> ]
1859 [ group <group> ] [ gid <gid> ]
1860
1861 Fixes common settings to UNIX listening sockets declared in "bind" statements.
1862 This is mainly used to simplify declaration of those UNIX sockets and reduce
1863 the risk of errors, since those settings are most commonly required but are
1864 also process-specific. The <prefix> setting can be used to force all socket
1865 path to be relative to that directory. This might be needed to access another
1866 component's chroot. Note that those paths are resolved before haproxy chroots
1867 itself, so they are absolute. The <mode>, <user>, <uid>, <group> and <gid>
1868 all have the same meaning as their homonyms used by the "bind" statement. If
1869 both are specified, the "bind" statement has priority, meaning that the
1870 "unix-bind" settings may be seen as process-wide default settings.
1871
Willy Tarreau1d549722016-02-16 12:41:57 +01001872unsetenv [<name> ...]
1873 Removes environment variables specified in arguments. This can be useful to
1874 hide some sensitive information that are occasionally inherited from the
1875 user's environment during some operations. Variables which did not exist are
1876 silently ignored so that after the operation, it is certain that none of
1877 these variables remain. The changes immediately take effect so that the next
1878 line in the configuration file will not see these variables. See also
1879 "setenv", "presetenv", and "resetenv".
1880
Willy Tarreau6a06a402007-07-15 20:15:28 +02001881user <user name>
1882 Similar to "uid" but uses the UID of user name <user name> from /etc/passwd.
1883 See also "uid" and "group".
1884
Krzysztof Piotr Oledzki48cb2ae2009-10-02 22:51:14 +02001885node <name>
1886 Only letters, digits, hyphen and underscore are allowed, like in DNS names.
1887
1888 This statement is useful in HA configurations where two or more processes or
1889 servers share the same IP address. By setting a different node-name on all
1890 nodes, it becomes easy to immediately spot what server is handling the
1891 traffic.
1892
1893description <text>
1894 Add a text that describes the instance.
1895
1896 Please note that it is required to escape certain characters (# for example)
1897 and this text is inserted into a html page so you should avoid using
1898 "<" and ">" characters.
1899
Thomas Holmesdb04f192015-05-18 13:21:39 +0100190051degrees-data-file <file path>
1901 The path of the 51Degrees data file to provide device detection services. The
Davor Ocelice9ed2812017-12-25 17:49:28 +01001902 file should be unzipped and accessible by HAProxy with relevant permissions.
Thomas Holmesdb04f192015-05-18 13:21:39 +01001903
Dragan Dosenae6d39a2015-06-29 16:43:27 +02001904 Please note that this option is only available when haproxy has been
Thomas Holmesdb04f192015-05-18 13:21:39 +01001905 compiled with USE_51DEGREES.
1906
Ben Shillitof25e8e52016-12-02 14:25:37 +0000190751degrees-property-name-list [<string> ...]
Thomas Holmesdb04f192015-05-18 13:21:39 +01001908 A list of 51Degrees property names to be load from the dataset. A full list
1909 of names is available on the 51Degrees website:
1910 https://51degrees.com/resources/property-dictionary
1911
Dragan Dosenae6d39a2015-06-29 16:43:27 +02001912 Please note that this option is only available when haproxy has been
Thomas Holmesdb04f192015-05-18 13:21:39 +01001913 compiled with USE_51DEGREES.
1914
Dragan Dosen93b38d92015-06-29 16:43:25 +0200191551degrees-property-separator <char>
Thomas Holmesdb04f192015-05-18 13:21:39 +01001916 A char that will be appended to every property value in a response header
1917 containing 51Degrees results. If not set that will be set as ','.
1918
Dragan Dosenae6d39a2015-06-29 16:43:27 +02001919 Please note that this option is only available when haproxy has been
1920 compiled with USE_51DEGREES.
1921
192251degrees-cache-size <number>
1923 Sets the size of the 51Degrees converter cache to <number> entries. This
1924 is an LRU cache which reminds previous device detections and their results.
1925 By default, this cache is disabled.
1926
1927 Please note that this option is only available when haproxy has been
Thomas Holmesdb04f192015-05-18 13:21:39 +01001928 compiled with USE_51DEGREES.
1929
Willy Tarreaub3cc9f22019-04-19 16:03:32 +02001930wurfl-data-file <file path>
1931 The path of the WURFL data file to provide device detection services. The
1932 file should be accessible by HAProxy with relevant permissions.
1933
1934 Please note that this option is only available when haproxy has been compiled
1935 with USE_WURFL=1.
1936
1937wurfl-information-list [<capability>]*
1938 A space-delimited list of WURFL capabilities, virtual capabilities, property
1939 names we plan to use in injected headers. A full list of capability and
1940 virtual capability names is available on the Scientiamobile website :
1941
1942 https://www.scientiamobile.com/wurflCapability
1943
1944 Valid WURFL properties are:
1945 - wurfl_id Contains the device ID of the matched device.
1946
1947 - wurfl_root_id Contains the device root ID of the matched
1948 device.
1949
1950 - wurfl_isdevroot Tells if the matched device is a root device.
1951 Possible values are "TRUE" or "FALSE".
1952
1953 - wurfl_useragent The original useragent coming with this
1954 particular web request.
1955
1956 - wurfl_api_version Contains a string representing the currently
1957 used Libwurfl API version.
1958
Willy Tarreaub3cc9f22019-04-19 16:03:32 +02001959 - wurfl_info A string containing information on the parsed
1960 wurfl.xml and its full path.
1961
1962 - wurfl_last_load_time Contains the UNIX timestamp of the last time
1963 WURFL has been loaded successfully.
1964
1965 - wurfl_normalized_useragent The normalized useragent.
1966
Willy Tarreaub3cc9f22019-04-19 16:03:32 +02001967 Please note that this option is only available when haproxy has been compiled
1968 with USE_WURFL=1.
1969
1970wurfl-information-list-separator <char>
1971 A char that will be used to separate values in a response header containing
1972 WURFL results. If not set that a comma (',') will be used by default.
1973
1974 Please note that this option is only available when haproxy has been compiled
1975 with USE_WURFL=1.
1976
1977wurfl-patch-file [<file path>]
1978 A list of WURFL patch file paths. Note that patches are loaded during startup
1979 thus before the chroot.
1980
1981 Please note that this option is only available when haproxy has been compiled
1982 with USE_WURFL=1.
1983
paulborilebad132c2019-04-18 11:57:04 +02001984wurfl-cache-size <size>
1985 Sets the WURFL Useragent cache size. For faster lookups, already processed user
1986 agents are kept in a LRU cache :
Willy Tarreaub3cc9f22019-04-19 16:03:32 +02001987 - "0" : no cache is used.
paulborilebad132c2019-04-18 11:57:04 +02001988 - <size> : size of lru cache in elements.
Willy Tarreaub3cc9f22019-04-19 16:03:32 +02001989
1990 Please note that this option is only available when haproxy has been compiled
1991 with USE_WURFL=1.
1992
William Dauchy0fec3ab2019-10-27 20:08:11 +01001993strict-limits
William Dauchya5194602020-03-28 19:29:58 +01001994 Makes process fail at startup when a setrlimit fails. Haproxy tries to set the
1995 best setrlimit according to what has been calculated. If it fails, it will
1996 emit a warning. This option is here to guarantee an explicit failure of
1997 haproxy when those limits fail. It is enabled by default. It may still be
1998 forcibly disabled by prefixing it with the "no" keyword.
William Dauchy0fec3ab2019-10-27 20:08:11 +01001999
Willy Tarreauc57f0e22009-05-10 13:12:33 +020020003.2. Performance tuning
Willy Tarreau6a06a402007-07-15 20:15:28 +02002001-----------------------
2002
Willy Tarreaubeb859a2018-11-22 18:07:59 +01002003busy-polling
2004 In some situations, especially when dealing with low latency on processors
2005 supporting a variable frequency or when running inside virtual machines, each
2006 time the process waits for an I/O using the poller, the processor goes back
2007 to sleep or is offered to another VM for a long time, and it causes
2008 excessively high latencies. This option provides a solution preventing the
2009 processor from sleeping by always using a null timeout on the pollers. This
2010 results in a significant latency reduction (30 to 100 microseconds observed)
2011 at the expense of a risk to overheat the processor. It may even be used with
2012 threads, in which case improperly bound threads may heavily conflict,
2013 resulting in a worse performance and high values for the CPU stolen fields
2014 in "show info" output, indicating which threads are misconfigured. It is
2015 important not to let the process run on the same processor as the network
2016 interrupts when this option is used. It is also better to avoid using it on
2017 multiple CPU threads sharing the same core. This option is disabled by
2018 default. If it has been enabled, it may still be forcibly disabled by
2019 prefixing it with the "no" keyword. It is ignored by the "select" and
2020 "poll" pollers.
2021
William Dauchy3894d972019-12-28 15:36:02 +01002022 This option is automatically disabled on old processes in the context of
2023 seamless reload; it avoids too much cpu conflicts when multiple processes
2024 stay around for some time waiting for the end of their current connections.
2025
Willy Tarreau1746eec2014-04-25 10:46:47 +02002026max-spread-checks <delay in milliseconds>
2027 By default, haproxy tries to spread the start of health checks across the
2028 smallest health check interval of all the servers in a farm. The principle is
2029 to avoid hammering services running on the same server. But when using large
2030 check intervals (10 seconds or more), the last servers in the farm take some
2031 time before starting to be tested, which can be a problem. This parameter is
2032 used to enforce an upper bound on delay between the first and the last check,
2033 even if the servers' check intervals are larger. When servers run with
2034 shorter intervals, their intervals will be respected though.
2035
Willy Tarreau6a06a402007-07-15 20:15:28 +02002036maxconn <number>
2037 Sets the maximum per-process number of concurrent connections to <number>. It
2038 is equivalent to the command-line argument "-n". Proxies will stop accepting
2039 connections when this limit is reached. The "ulimit-n" parameter is
Willy Tarreau8274e102014-06-19 15:31:25 +02002040 automatically adjusted according to this value. See also "ulimit-n". Note:
2041 the "select" poller cannot reliably use more than 1024 file descriptors on
2042 some platforms. If your platform only supports select and reports "select
2043 FAILED" on startup, you need to reduce maxconn until it works (slightly
Willy Tarreaub28f3442019-03-04 08:13:43 +01002044 below 500 in general). If this value is not set, it will automatically be
2045 calculated based on the current file descriptors limit reported by the
2046 "ulimit -n" command, possibly reduced to a lower value if a memory limit
2047 is enforced, based on the buffer size, memory allocated to compression, SSL
2048 cache size, and use or not of SSL and the associated maxsslconn (which can
2049 also be automatic).
Willy Tarreau6a06a402007-07-15 20:15:28 +02002050
Willy Tarreau81c25d02011-09-07 15:17:21 +02002051maxconnrate <number>
2052 Sets the maximum per-process number of connections per second to <number>.
2053 Proxies will stop accepting connections when this limit is reached. It can be
2054 used to limit the global capacity regardless of each frontend capacity. It is
2055 important to note that this can only be used as a service protection measure,
2056 as there will not necessarily be a fair share between frontends when the
2057 limit is reached, so it's a good idea to also limit each frontend to some
2058 value close to its expected share. Also, lowering tune.maxaccept can improve
2059 fairness.
2060
William Lallemandd85f9172012-11-09 17:05:39 +01002061maxcomprate <number>
2062 Sets the maximum per-process input compression rate to <number> kilobytes
Davor Ocelice9ed2812017-12-25 17:49:28 +01002063 per second. For each session, if the maximum is reached, the compression
William Lallemandd85f9172012-11-09 17:05:39 +01002064 level will be decreased during the session. If the maximum is reached at the
2065 beginning of a session, the session will not compress at all. If the maximum
2066 is not reached, the compression level will be increased up to
Davor Ocelice9ed2812017-12-25 17:49:28 +01002067 tune.comp.maxlevel. A value of zero means there is no limit, this is the
William Lallemandd85f9172012-11-09 17:05:39 +01002068 default value.
2069
William Lallemand072a2bf2012-11-20 17:01:01 +01002070maxcompcpuusage <number>
2071 Sets the maximum CPU usage HAProxy can reach before stopping the compression
2072 for new requests or decreasing the compression level of current requests.
2073 It works like 'maxcomprate' but measures CPU usage instead of incoming data
2074 bandwidth. The value is expressed in percent of the CPU used by haproxy. In
2075 case of multiple processes (nbproc > 1), each process manages its individual
2076 usage. A value of 100 disable the limit. The default value is 100. Setting
2077 a lower value will prevent the compression work from slowing the whole
2078 process down and from introducing high latencies.
2079
Willy Tarreauff4f82d2009-02-06 11:28:13 +01002080maxpipes <number>
2081 Sets the maximum per-process number of pipes to <number>. Currently, pipes
2082 are only used by kernel-based tcp splicing. Since a pipe contains two file
2083 descriptors, the "ulimit-n" value will be increased accordingly. The default
2084 value is maxconn/4, which seems to be more than enough for most heavy usages.
2085 The splice code dynamically allocates and releases pipes, and can fall back
2086 to standard copy, so setting this value too low may only impact performance.
2087
Willy Tarreau93e7c002013-10-07 18:51:07 +02002088maxsessrate <number>
2089 Sets the maximum per-process number of sessions per second to <number>.
2090 Proxies will stop accepting connections when this limit is reached. It can be
2091 used to limit the global capacity regardless of each frontend capacity. It is
2092 important to note that this can only be used as a service protection measure,
2093 as there will not necessarily be a fair share between frontends when the
2094 limit is reached, so it's a good idea to also limit each frontend to some
2095 value close to its expected share. Also, lowering tune.maxaccept can improve
2096 fairness.
2097
Willy Tarreau403edff2012-09-06 11:58:37 +02002098maxsslconn <number>
2099 Sets the maximum per-process number of concurrent SSL connections to
2100 <number>. By default there is no SSL-specific limit, which means that the
2101 global maxconn setting will apply to all connections. Setting this limit
2102 avoids having openssl use too much memory and crash when malloc returns NULL
2103 (since it unfortunately does not reliably check for such conditions). Note
2104 that the limit applies both to incoming and outgoing connections, so one
2105 connection which is deciphered then ciphered accounts for 2 SSL connections.
Willy Tarreaud0256482015-01-15 21:45:22 +01002106 If this value is not set, but a memory limit is enforced, this value will be
2107 automatically computed based on the memory limit, maxconn, the buffer size,
2108 memory allocated to compression, SSL cache size, and use of SSL in either
2109 frontends, backends or both. If neither maxconn nor maxsslconn are specified
2110 when there is a memory limit, haproxy will automatically adjust these values
2111 so that 100% of the connections can be made over SSL with no risk, and will
2112 consider the sides where it is enabled (frontend, backend, both).
Willy Tarreau403edff2012-09-06 11:58:37 +02002113
Willy Tarreaue43d5322013-10-07 20:01:52 +02002114maxsslrate <number>
2115 Sets the maximum per-process number of SSL sessions per second to <number>.
2116 SSL listeners will stop accepting connections when this limit is reached. It
2117 can be used to limit the global SSL CPU usage regardless of each frontend
2118 capacity. It is important to note that this can only be used as a service
2119 protection measure, as there will not necessarily be a fair share between
2120 frontends when the limit is reached, so it's a good idea to also limit each
2121 frontend to some value close to its expected share. It is also important to
2122 note that the sessions are accounted before they enter the SSL stack and not
2123 after, which also protects the stack against bad handshakes. Also, lowering
2124 tune.maxaccept can improve fairness.
2125
William Lallemand9d5f5482012-11-07 16:12:57 +01002126maxzlibmem <number>
2127 Sets the maximum amount of RAM in megabytes per process usable by the zlib.
2128 When the maximum amount is reached, future sessions will not compress as long
2129 as RAM is unavailable. When sets to 0, there is no limit.
William Lallemande3a7d992012-11-20 11:25:20 +01002130 The default value is 0. The value is available in bytes on the UNIX socket
2131 with "show info" on the line "MaxZlibMemUsage", the memory used by zlib is
2132 "ZlibMemUsage" in bytes.
2133
Willy Tarreau6a06a402007-07-15 20:15:28 +02002134noepoll
2135 Disables the use of the "epoll" event polling system on Linux. It is
2136 equivalent to the command-line argument "-de". The next polling system
Willy Tarreaue9f49e72012-11-11 17:42:00 +01002137 used will generally be "poll". See also "nopoll".
Willy Tarreau6a06a402007-07-15 20:15:28 +02002138
2139nokqueue
2140 Disables the use of the "kqueue" event polling system on BSD. It is
2141 equivalent to the command-line argument "-dk". The next polling system
2142 used will generally be "poll". See also "nopoll".
2143
Emmanuel Hocdet0ba4f482019-04-08 16:53:32 +00002144noevports
2145 Disables the use of the event ports event polling system on SunOS systems
2146 derived from Solaris 10 and later. It is equivalent to the command-line
2147 argument "-dv". The next polling system used will generally be "poll". See
2148 also "nopoll".
2149
Willy Tarreau6a06a402007-07-15 20:15:28 +02002150nopoll
2151 Disables the use of the "poll" event polling system. It is equivalent to the
2152 command-line argument "-dp". The next polling system used will be "select".
Willy Tarreau0ba27502007-12-24 16:55:16 +01002153 It should never be needed to disable "poll" since it's available on all
Emmanuel Hocdet0ba4f482019-04-08 16:53:32 +00002154 platforms supported by HAProxy. See also "nokqueue", "noepoll" and
2155 "noevports".
Willy Tarreau6a06a402007-07-15 20:15:28 +02002156
Willy Tarreauff4f82d2009-02-06 11:28:13 +01002157nosplice
2158 Disables the use of kernel tcp splicing between sockets on Linux. It is
Davor Ocelice9ed2812017-12-25 17:49:28 +01002159 equivalent to the command line argument "-dS". Data will then be copied
Willy Tarreauff4f82d2009-02-06 11:28:13 +01002160 using conventional and more portable recv/send calls. Kernel tcp splicing is
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01002161 limited to some very recent instances of kernel 2.6. Most versions between
Willy Tarreauff4f82d2009-02-06 11:28:13 +01002162 2.6.25 and 2.6.28 are buggy and will forward corrupted data, so they must not
2163 be used. This option makes it easier to globally disable kernel splicing in
2164 case of doubt. See also "option splice-auto", "option splice-request" and
2165 "option splice-response".
2166
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03002167nogetaddrinfo
2168 Disables the use of getaddrinfo(3) for name resolving. It is equivalent to
2169 the command line argument "-dG". Deprecated gethostbyname(3) will be used.
2170
Lukas Tribusa0bcbdc2016-09-12 21:42:20 +00002171noreuseport
2172 Disables the use of SO_REUSEPORT - see socket(7). It is equivalent to the
2173 command line argument "-dR".
2174
Willy Tarreaud2d33482019-04-25 17:09:07 +02002175profiling.tasks { auto | on | off }
2176 Enables ('on') or disables ('off') per-task CPU profiling. When set to 'auto'
2177 the profiling automatically turns on a thread when it starts to suffer from
2178 an average latency of 1000 microseconds or higher as reported in the
2179 "avg_loop_us" activity field, and automatically turns off when the latency
John Roeslerfb2fce12019-07-10 15:45:51 -05002180 returns below 990 microseconds (this value is an average over the last 1024
Willy Tarreaud2d33482019-04-25 17:09:07 +02002181 loops so it does not vary quickly and tends to significantly smooth short
2182 spikes). It may also spontaneously trigger from time to time on overloaded
2183 systems, containers, or virtual machines, or when the system swaps (which
2184 must absolutely never happen on a load balancer).
2185
2186 CPU profiling per task can be very convenient to report where the time is
2187 spent and which requests have what effect on which other request. Enabling
2188 it will typically affect the overall's performance by less than 1%, thus it
2189 is recommended to leave it to the default 'auto' value so that it only
2190 operates when a problem is identified. This feature requires a system
Willy Tarreau75c62c22018-11-22 11:02:09 +01002191 supporting the clock_gettime(2) syscall with clock identifiers
2192 CLOCK_MONOTONIC and CLOCK_THREAD_CPUTIME_ID, otherwise the reported time will
2193 be zero. This option may be changed at run time using "set profiling" on the
2194 CLI.
2195
Willy Tarreaufe255b72007-10-14 23:09:26 +02002196spread-checks <0..50, in percent>
Simon Hormand60d6912013-11-25 10:46:36 +09002197 Sometimes it is desirable to avoid sending agent and health checks to
2198 servers at exact intervals, for instance when many logical servers are
2199 located on the same physical server. With the help of this parameter, it
2200 becomes possible to add some randomness in the check interval between 0
2201 and +/- 50%. A value between 2 and 5 seems to show good results. The
2202 default value remains at 0.
Willy Tarreaufe255b72007-10-14 23:09:26 +02002203
Davor Ocelice9ed2812017-12-25 17:49:28 +01002204ssl-engine <name> [algo <comma-separated list of algorithms>]
Grant Zhang872f9c22017-01-21 01:10:18 +00002205 Sets the OpenSSL engine to <name>. List of valid values for <name> may be
Davor Ocelice9ed2812017-12-25 17:49:28 +01002206 obtained using the command "openssl engine". This statement may be used
Grant Zhang872f9c22017-01-21 01:10:18 +00002207 multiple times, it will simply enable multiple crypto engines. Referencing an
2208 unsupported engine will prevent haproxy from starting. Note that many engines
2209 will lead to lower HTTPS performance than pure software with recent
2210 processors. The optional command "algo" sets the default algorithms an ENGINE
2211 will supply using the OPENSSL function ENGINE_set_default_string(). A value
Davor Ocelice9ed2812017-12-25 17:49:28 +01002212 of "ALL" uses the engine for all cryptographic operations. If no list of
2213 algo is specified then the value of "ALL" is used. A comma-separated list
Grant Zhang872f9c22017-01-21 01:10:18 +00002214 of different algorithms may be specified, including: RSA, DSA, DH, EC, RAND,
2215 CIPHERS, DIGESTS, PKEY, PKEY_CRYPTO, PKEY_ASN1. This is the same format that
2216 openssl configuration file uses:
2217 https://www.openssl.org/docs/man1.0.2/apps/config.html
2218
Grant Zhangfa6c7ee2017-01-14 01:42:15 +00002219ssl-mode-async
2220 Adds SSL_MODE_ASYNC mode to the SSL context. This enables asynchronous TLS
Emeric Brun3854e012017-05-17 20:42:48 +02002221 I/O operations if asynchronous capable SSL engines are used. The current
Emeric Brunb5e42a82017-06-06 12:35:14 +00002222 implementation supports a maximum of 32 engines. The Openssl ASYNC API
2223 doesn't support moving read/write buffers and is not compliant with
2224 haproxy's buffer management. So the asynchronous mode is disabled on
John Roeslerfb2fce12019-07-10 15:45:51 -05002225 read/write operations (it is only enabled during initial and renegotiation
Emeric Brunb5e42a82017-06-06 12:35:14 +00002226 handshakes).
Grant Zhangfa6c7ee2017-01-14 01:42:15 +00002227
Willy Tarreau33cb0652014-12-23 22:52:37 +01002228tune.buffers.limit <number>
2229 Sets a hard limit on the number of buffers which may be allocated per process.
2230 The default value is zero which means unlimited. The minimum non-zero value
2231 will always be greater than "tune.buffers.reserve" and should ideally always
2232 be about twice as large. Forcing this value can be particularly useful to
2233 limit the amount of memory a process may take, while retaining a sane
Davor Ocelice9ed2812017-12-25 17:49:28 +01002234 behavior. When this limit is reached, sessions which need a buffer wait for
Willy Tarreau33cb0652014-12-23 22:52:37 +01002235 another one to be released by another session. Since buffers are dynamically
2236 allocated and released, the waiting time is very short and not perceptible
2237 provided that limits remain reasonable. In fact sometimes reducing the limit
2238 may even increase performance by increasing the CPU cache's efficiency. Tests
2239 have shown good results on average HTTP traffic with a limit to 1/10 of the
2240 expected global maxconn setting, which also significantly reduces memory
2241 usage. The memory savings come from the fact that a number of connections
2242 will not allocate 2*tune.bufsize. It is best not to touch this value unless
2243 advised to do so by an haproxy core developer.
2244
Willy Tarreau1058ae72014-12-23 22:40:40 +01002245tune.buffers.reserve <number>
2246 Sets the number of buffers which are pre-allocated and reserved for use only
2247 during memory shortage conditions resulting in failed memory allocations. The
2248 minimum value is 2 and is also the default. There is no reason a user would
2249 want to change this value, it's mostly aimed at haproxy core developers.
2250
Willy Tarreau27a674e2009-08-17 07:23:33 +02002251tune.bufsize <number>
2252 Sets the buffer size to this size (in bytes). Lower values allow more
2253 sessions to coexist in the same amount of RAM, and higher values allow some
2254 applications with very large cookies to work. The default value is 16384 and
2255 can be changed at build time. It is strongly recommended not to change this
2256 from the default value, as very low values will break some services such as
2257 statistics, and values larger than default size will increase memory usage,
2258 possibly causing the system to run out of memory. At least the global maxconn
Willy Tarreau45a66cc2017-11-24 11:28:00 +01002259 parameter should be decreased by the same factor as this one is increased. In
2260 addition, use of HTTP/2 mandates that this value must be 16384 or more. If an
2261 HTTP request is larger than (tune.bufsize - tune.maxrewrite), haproxy will
Dmitry Sivachenkof6f4f7b2012-10-21 18:10:25 +04002262 return HTTP 400 (Bad Request) error. Similarly if an HTTP response is larger
Willy Tarreauc77d3642018-12-12 06:19:42 +01002263 than this size, haproxy will return HTTP 502 (Bad Gateway). Note that the
2264 value set using this parameter will automatically be rounded up to the next
2265 multiple of 8 on 32-bit machines and 16 on 64-bit machines.
Willy Tarreau27a674e2009-08-17 07:23:33 +02002266
Christopher Fauletbb9fb8b2020-11-25 17:20:57 +01002267tune.chksize <number> (deprecated)
2268 This option is deprecated and ignored.
Willy Tarreau43961d52010-10-04 20:39:20 +02002269
William Lallemandf3747832012-11-09 12:33:10 +01002270tune.comp.maxlevel <number>
2271 Sets the maximum compression level. The compression level affects CPU
2272 usage during compression. This value affects CPU usage during compression.
2273 Each session using compression initializes the compression algorithm with
2274 this value. The default value is 1.
2275
Willy Tarreauc299e1e2019-02-27 11:35:12 +01002276tune.fail-alloc
2277 If compiled with DEBUG_FAIL_ALLOC, gives the percentage of chances an
2278 allocation attempt fails. Must be between 0 (no failure) and 100 (no
2279 success). This is useful to debug and make sure memory failures are handled
2280 gracefully.
2281
Willy Tarreaubc52bec2020-06-18 08:58:47 +02002282tune.fd.edge-triggered { on | off } [ EXPERIMENTAL ]
2283 Enables ('on') or disables ('off') the edge-triggered polling mode for FDs
2284 that support it. This is currently only support with epoll. It may noticeably
2285 reduce the number of epoll_ctl() calls and slightly improve performance in
2286 certain scenarios. This is still experimental, it may result in frozen
2287 connections if bugs are still present, and is disabled by default.
2288
Willy Tarreaufe20e5b2017-07-27 11:42:14 +02002289tune.h2.header-table-size <number>
2290 Sets the HTTP/2 dynamic header table size. It defaults to 4096 bytes and
2291 cannot be larger than 65536 bytes. A larger value may help certain clients
2292 send more compact requests, depending on their capabilities. This amount of
2293 memory is consumed for each HTTP/2 connection. It is recommended not to
2294 change it.
2295
Willy Tarreaue6baec02017-07-27 11:45:11 +02002296tune.h2.initial-window-size <number>
2297 Sets the HTTP/2 initial window size, which is the number of bytes the client
Davor Ocelice9ed2812017-12-25 17:49:28 +01002298 can upload before waiting for an acknowledgment from haproxy. This setting
2299 only affects payload contents (i.e. the body of POST requests), not headers.
Willy Tarreaue6baec02017-07-27 11:45:11 +02002300 The default value is 65535, which roughly allows up to 5 Mbps of upload
2301 bandwidth per client over a network showing a 100 ms ping time, or 500 Mbps
2302 over a 1-ms local network. It can make sense to increase this value to allow
2303 faster uploads, or to reduce it to increase fairness when dealing with many
2304 clients. It doesn't affect resource usage.
2305
Willy Tarreau5242ef82017-07-27 11:47:28 +02002306tune.h2.max-concurrent-streams <number>
2307 Sets the HTTP/2 maximum number of concurrent streams per connection (ie the
2308 number of outstanding requests on a single connection). The default value is
2309 100. A larger one may slightly improve page load time for complex sites when
2310 visited over high latency networks, but increases the amount of resources a
2311 single client may allocate. A value of zero disables the limit so a single
2312 client may create as many streams as allocatable by haproxy. It is highly
2313 recommended not to change this value.
2314
Willy Tarreaua24b35c2019-02-21 13:24:36 +01002315tune.h2.max-frame-size <number>
2316 Sets the HTTP/2 maximum frame size that haproxy announces it is willing to
2317 receive to its peers. The default value is the largest between 16384 and the
2318 buffer size (tune.bufsize). In any case, haproxy will not announce support
2319 for frame sizes larger than buffers. The main purpose of this setting is to
2320 allow to limit the maximum frame size setting when using large buffers. Too
2321 large frame sizes might have performance impact or cause some peers to
2322 misbehave. It is highly recommended not to change this value.
2323
Willy Tarreau193b8c62012-11-22 00:17:38 +01002324tune.http.cookielen <number>
2325 Sets the maximum length of captured cookies. This is the maximum value that
2326 the "capture cookie xxx len yyy" will be allowed to take, and any upper value
2327 will automatically be truncated to this one. It is important not to set too
2328 high a value because all cookie captures still allocate this size whatever
2329 their configured value (they share a same pool). This value is per request
2330 per response, so the memory allocated is twice this value per connection.
2331 When not specified, the limit is set to 63 characters. It is recommended not
2332 to change this value.
2333
Stéphane Cottin23e9e932017-05-18 08:58:41 +02002334tune.http.logurilen <number>
Davor Ocelice9ed2812017-12-25 17:49:28 +01002335 Sets the maximum length of request URI in logs. This prevents truncating long
2336 request URIs with valuable query strings in log lines. This is not related
Stéphane Cottin23e9e932017-05-18 08:58:41 +02002337 to syslog limits. If you increase this limit, you may also increase the
Davor Ocelice9ed2812017-12-25 17:49:28 +01002338 'log ... len yyy' parameter. Your syslog daemon may also need specific
Stéphane Cottin23e9e932017-05-18 08:58:41 +02002339 configuration directives too.
2340 The default value is 1024.
2341
Willy Tarreauac1932d2011-10-24 19:14:41 +02002342tune.http.maxhdr <number>
2343 Sets the maximum number of headers in a request. When a request comes with a
2344 number of headers greater than this value (including the first line), it is
2345 rejected with a "400 Bad Request" status code. Similarly, too large responses
2346 are blocked with "502 Bad Gateway". The default value is 101, which is enough
2347 for all usages, considering that the widely deployed Apache server uses the
2348 same limit. It can be useful to push this limit further to temporarily allow
Christopher Faulet50174f32017-06-21 16:31:35 +02002349 a buggy application to work by the time it gets fixed. The accepted range is
2350 1..32767. Keep in mind that each new header consumes 32bits of memory for
2351 each session, so don't push this limit too high.
Willy Tarreauac1932d2011-10-24 19:14:41 +02002352
Willy Tarreau76cc6992020-07-01 18:49:24 +02002353tune.idle-pool.shared { on | off }
2354 Enables ('on') or disables ('off') sharing of idle connection pools between
2355 threads for a same server. The default is to share them between threads in
2356 order to minimize the number of persistent connections to a server, and to
2357 optimize the connection reuse rate. But to help with debugging or when
2358 suspecting a bug in HAProxy around connection reuse, it can be convenient to
2359 forcefully disable this idle pool sharing between multiple threads, and force
Willy Tarreau0784db82021-02-19 11:45:22 +01002360 this option to "off". The default is on. It is strongly recommended against
2361 disabling this option without setting a conservative value on "pool-low-conn"
2362 for all servers relying on connection reuse to achieve a high performance
2363 level, otherwise connections might be closed very often as the thread count
2364 increases.
Willy Tarreau76cc6992020-07-01 18:49:24 +02002365
Willy Tarreau7e312732014-02-12 16:35:14 +01002366tune.idletimer <timeout>
2367 Sets the duration after which haproxy will consider that an empty buffer is
2368 probably associated with an idle stream. This is used to optimally adjust
2369 some packet sizes while forwarding large and small data alternatively. The
2370 decision to use splice() or to send large buffers in SSL is modulated by this
2371 parameter. The value is in milliseconds between 0 and 65535. A value of zero
2372 means that haproxy will not try to detect idle streams. The default is 1000,
Davor Ocelice9ed2812017-12-25 17:49:28 +01002373 which seems to correctly detect end user pauses (e.g. read a page before
John Roeslerfb2fce12019-07-10 15:45:51 -05002374 clicking). There should be no reason for changing this value. Please check
Willy Tarreau7e312732014-02-12 16:35:14 +01002375 tune.ssl.maxrecord below.
2376
Willy Tarreau7ac908b2019-02-27 12:02:18 +01002377tune.listener.multi-queue { on | off }
2378 Enables ('on') or disables ('off') the listener's multi-queue accept which
2379 spreads the incoming traffic to all threads a "bind" line is allowed to run
2380 on instead of taking them for itself. This provides a smoother traffic
2381 distribution and scales much better, especially in environments where threads
2382 may be unevenly loaded due to external activity (network interrupts colliding
2383 with one thread for example). This option is enabled by default, but it may
2384 be forcefully disabled for troubleshooting or for situations where it is
2385 estimated that the operating system already provides a good enough
2386 distribution and connections are extremely short-lived.
2387
Thierry FOURNIER90da1912015-03-05 11:17:06 +01002388tune.lua.forced-yield <number>
2389 This directive forces the Lua engine to execute a yield each <number> of
Tim Düsterhus4896c442016-11-29 02:15:19 +01002390 instructions executed. This permits interrupting a long script and allows the
Thierry FOURNIER90da1912015-03-05 11:17:06 +01002391 HAProxy scheduler to process other tasks like accepting connections or
2392 forwarding traffic. The default value is 10000 instructions. If HAProxy often
Davor Ocelice9ed2812017-12-25 17:49:28 +01002393 executes some Lua code but more responsiveness is required, this value can be
Thierry FOURNIER90da1912015-03-05 11:17:06 +01002394 lowered. If the Lua code is quite long and its result is absolutely required
2395 to process the data, the <number> can be increased.
2396
Willy Tarreau32f61e22015-03-18 17:54:59 +01002397tune.lua.maxmem
2398 Sets the maximum amount of RAM in megabytes per process usable by Lua. By
2399 default it is zero which means unlimited. It is important to set a limit to
2400 ensure that a bug in a script will not result in the system running out of
2401 memory.
2402
Thierry FOURNIER90da1912015-03-05 11:17:06 +01002403tune.lua.session-timeout <timeout>
2404 This is the execution timeout for the Lua sessions. This is useful for
Thierry FOURNIER7dd784b2015-10-01 14:49:33 +02002405 preventing infinite loops or spending too much time in Lua. This timeout
2406 counts only the pure Lua runtime. If the Lua does a sleep, the sleep is
Davor Ocelice9ed2812017-12-25 17:49:28 +01002407 not taken in account. The default timeout is 4s.
Thierry FOURNIER90da1912015-03-05 11:17:06 +01002408
2409tune.lua.task-timeout <timeout>
2410 Purpose is the same as "tune.lua.session-timeout", but this timeout is
2411 dedicated to the tasks. By default, this timeout isn't set because a task may
2412 remain alive during of the lifetime of HAProxy. For example, a task used to
2413 check servers.
2414
Thierry FOURNIER7dd784b2015-10-01 14:49:33 +02002415tune.lua.service-timeout <timeout>
2416 This is the execution timeout for the Lua services. This is useful for
2417 preventing infinite loops or spending too much time in Lua. This timeout
2418 counts only the pure Lua runtime. If the Lua does a sleep, the sleep is
Davor Ocelice9ed2812017-12-25 17:49:28 +01002419 not taken in account. The default timeout is 4s.
Thierry FOURNIER7dd784b2015-10-01 14:49:33 +02002420
Willy Tarreaua0250ba2008-01-06 11:22:57 +01002421tune.maxaccept <number>
Willy Tarreau16a21472012-11-19 12:39:59 +01002422 Sets the maximum number of consecutive connections a process may accept in a
2423 row before switching to other work. In single process mode, higher numbers
Willy Tarreau66161322021-02-19 15:50:27 +01002424 used to give better performance at high connection rates, though this is not
2425 the case anymore with the multi-queue. This value applies individually to
2426 each listener, so that the number of processes a listener is bound to is
2427 taken into account. This value defaults to 4 which showed best results. If a
2428 significantly higher value was inherited from an ancient config, it might be
2429 worth removing it as it will both increase performance and lower response
2430 time. In multi-process mode, it is divided by twice the number of processes
2431 the listener is bound to. Setting this value to -1 completely disables the
2432 limitation. It should normally not be needed to tweak this value.
Willy Tarreaua0250ba2008-01-06 11:22:57 +01002433
2434tune.maxpollevents <number>
2435 Sets the maximum amount of events that can be processed at once in a call to
2436 the polling system. The default value is adapted to the operating system. It
2437 has been noticed that reducing it below 200 tends to slightly decrease
2438 latency at the expense of network bandwidth, and increasing it above 200
2439 tends to trade latency for slightly increased bandwidth.
2440
Willy Tarreau27a674e2009-08-17 07:23:33 +02002441tune.maxrewrite <number>
2442 Sets the reserved buffer space to this size in bytes. The reserved space is
2443 used for header rewriting or appending. The first reads on sockets will never
2444 fill more than bufsize-maxrewrite. Historically it has defaulted to half of
2445 bufsize, though that does not make much sense since there are rarely large
2446 numbers of headers to add. Setting it too high prevents processing of large
2447 requests or responses. Setting it too low prevents addition of new headers
2448 to already large requests or to POST requests. It is generally wise to set it
2449 to about 1024. It is automatically readjusted to half of bufsize if it is
2450 larger than that. This means you don't have to worry about it when changing
2451 bufsize.
2452
Willy Tarreauf3045d22015-04-29 16:24:50 +02002453tune.pattern.cache-size <number>
2454 Sets the size of the pattern lookup cache to <number> entries. This is an LRU
2455 cache which reminds previous lookups and their results. It is used by ACLs
2456 and maps on slow pattern lookups, namely the ones using the "sub", "reg",
2457 "dir", "dom", "end", "bin" match methods as well as the case-insensitive
2458 strings. It applies to pattern expressions which means that it will be able
2459 to memorize the result of a lookup among all the patterns specified on a
2460 configuration line (including all those loaded from files). It automatically
2461 invalidates entries which are updated using HTTP actions or on the CLI. The
2462 default cache size is set to 10000 entries, which limits its footprint to
Willy Tarreau403bfbb2019-10-23 06:59:31 +02002463 about 5 MB per process/thread on 32-bit systems and 8 MB per process/thread
2464 on 64-bit systems, as caches are thread/process local. There is a very low
Willy Tarreauf3045d22015-04-29 16:24:50 +02002465 risk of collision in this cache, which is in the order of the size of the
2466 cache divided by 2^64. Typically, at 10000 requests per second with the
2467 default cache size of 10000 entries, there's 1% chance that a brute force
2468 attack could cause a single collision after 60 years, or 0.1% after 6 years.
2469 This is considered much lower than the risk of a memory corruption caused by
2470 aging components. If this is not acceptable, the cache can be disabled by
2471 setting this parameter to 0.
2472
Willy Tarreaubd9a0a72011-10-23 21:14:29 +02002473tune.pipesize <number>
2474 Sets the kernel pipe buffer size to this size (in bytes). By default, pipes
2475 are the default size for the system. But sometimes when using TCP splicing,
2476 it can improve performance to increase pipe sizes, especially if it is
2477 suspected that pipes are not filled and that many calls to splice() are
2478 performed. This has an impact on the kernel's memory footprint, so this must
2479 not be changed if impacts are not understood.
2480
Olivier Houchard88698d92019-04-16 19:07:22 +02002481tune.pool-high-fd-ratio <number>
2482 This setting sets the max number of file descriptors (in percentage) used by
2483 haproxy globally against the maximum number of file descriptors haproxy can
2484 use before we start killing idle connections when we can't reuse a connection
2485 and we have to create a new one. The default is 25 (one quarter of the file
2486 descriptor will mean that roughly half of the maximum front connections can
2487 keep an idle connection behind, anything beyond this probably doesn't make
John Roeslerfb2fce12019-07-10 15:45:51 -05002488 much sense in the general case when targeting connection reuse).
Olivier Houchard88698d92019-04-16 19:07:22 +02002489
Willy Tarreau83ca3052020-07-01 18:30:16 +02002490tune.pool-low-fd-ratio <number>
2491 This setting sets the max number of file descriptors (in percentage) used by
2492 haproxy globally against the maximum number of file descriptors haproxy can
2493 use before we stop putting connection into the idle pool for reuse. The
2494 default is 20.
2495
Willy Tarreaue803de22010-01-21 17:43:04 +01002496tune.rcvbuf.client <number>
2497tune.rcvbuf.server <number>
2498 Forces the kernel socket receive buffer size on the client or the server side
2499 to the specified value in bytes. This value applies to all TCP/HTTP frontends
2500 and backends. It should normally never be set, and the default size (0) lets
John Roeslerfb2fce12019-07-10 15:45:51 -05002501 the kernel auto-tune this value depending on the amount of available memory.
Davor Ocelice9ed2812017-12-25 17:49:28 +01002502 However it can sometimes help to set it to very low values (e.g. 4096) in
Willy Tarreaue803de22010-01-21 17:43:04 +01002503 order to save kernel memory by preventing it from buffering too large amounts
2504 of received data. Lower values will significantly increase CPU usage though.
2505
Willy Tarreaub22fc302015-12-14 12:04:35 +01002506tune.recv_enough <number>
Davor Ocelice9ed2812017-12-25 17:49:28 +01002507 HAProxy uses some hints to detect that a short read indicates the end of the
Willy Tarreaub22fc302015-12-14 12:04:35 +01002508 socket buffers. One of them is that a read returns more than <recv_enough>
2509 bytes, which defaults to 10136 (7 segments of 1448 each). This default value
2510 may be changed by this setting to better deal with workloads involving lots
2511 of short messages such as telnet or SSH sessions.
2512
Olivier Houchard1599b802018-05-24 18:59:04 +02002513tune.runqueue-depth <number>
John Roeslerfb2fce12019-07-10 15:45:51 -05002514 Sets the maximum amount of task that can be processed at once when running
Willy Tarreau060a7612021-03-10 11:06:26 +01002515 tasks. The default value depends on the number of threads but sits between 35
2516 and 280, which tend to show the highest request rates and lowest latencies.
2517 Increasing it may incur latency when dealing with I/Os, making it too small
2518 can incur extra overhead. Higher thread counts benefit from lower values.
2519 When experimenting with much larger values, it may be useful to also enable
2520 tune.sched.low-latency and possibly tune.fd.edge-triggered to limit the
2521 maximum latency to the lowest possible.
Willy Tarreaue7723bd2020-06-24 11:11:02 +02002522
2523tune.sched.low-latency { on | off }
2524 Enables ('on') or disables ('off') the low-latency task scheduler. By default
2525 haproxy processes tasks from several classes one class at a time as this is
2526 the most efficient. But when running with large values of tune.runqueue-depth
2527 this can have a measurable effect on request or connection latency. When this
2528 low-latency setting is enabled, tasks of lower priority classes will always
2529 be executed before other ones if they exist. This will permit to lower the
2530 maximum latency experienced by new requests or connections in the middle of
2531 massive traffic, at the expense of a higher impact on this large traffic.
2532 For regular usage it is better to leave this off. The default value is off.
Olivier Houchard1599b802018-05-24 18:59:04 +02002533
Willy Tarreaue803de22010-01-21 17:43:04 +01002534tune.sndbuf.client <number>
2535tune.sndbuf.server <number>
2536 Forces the kernel socket send buffer size on the client or the server side to
2537 the specified value in bytes. This value applies to all TCP/HTTP frontends
2538 and backends. It should normally never be set, and the default size (0) lets
John Roeslerfb2fce12019-07-10 15:45:51 -05002539 the kernel auto-tune this value depending on the amount of available memory.
Davor Ocelice9ed2812017-12-25 17:49:28 +01002540 However it can sometimes help to set it to very low values (e.g. 4096) in
Willy Tarreaue803de22010-01-21 17:43:04 +01002541 order to save kernel memory by preventing it from buffering too large amounts
2542 of received data. Lower values will significantly increase CPU usage though.
2543 Another use case is to prevent write timeouts with extremely slow clients due
2544 to the kernel waiting for a large part of the buffer to be read before
2545 notifying haproxy again.
2546
Willy Tarreau6ec58db2012-11-16 16:32:15 +01002547tune.ssl.cachesize <number>
Emeric Brunaf9619d2012-11-28 18:47:52 +01002548 Sets the size of the global SSL session cache, in a number of blocks. A block
William Dauchy9a4bbfe2021-02-12 15:58:46 +01002549 is large enough to contain an encoded session without peer certificate. An
2550 encoded session with peer certificate is stored in multiple blocks depending
2551 on the size of the peer certificate. A block uses approximately 200 bytes of
2552 memory (based on `sizeof(struct sh_ssl_sess_hdr) + SHSESS_BLOCK_MIN_SIZE`
2553 calculation used for `shctx_init` function). The default value may be forced
2554 at build time, otherwise defaults to 20000. When the cache is full, the most
2555 idle entries are purged and reassigned. Higher values reduce the occurrence
2556 of such a purge, hence the number of CPU-intensive SSL handshakes by ensuring
2557 that all users keep their session as long as possible. All entries are
2558 pre-allocated upon startup and are shared between all processes if "nbproc"
2559 is greater than 1. Setting this value to 0 disables the SSL session cache.
Willy Tarreau6ec58db2012-11-16 16:32:15 +01002560
Emeric Brun8dc60392014-05-09 13:52:00 +02002561tune.ssl.force-private-cache
Lukas Tribus27935782018-10-01 02:00:16 +02002562 This option disables SSL session cache sharing between all processes. It
Emeric Brun8dc60392014-05-09 13:52:00 +02002563 should normally not be used since it will force many renegotiations due to
2564 clients hitting a random process. But it may be required on some operating
2565 systems where none of the SSL cache synchronization method may be used. In
2566 this case, adding a first layer of hash-based load balancing before the SSL
2567 layer might limit the impact of the lack of session sharing.
2568
William Lallemand7d42ef52020-07-06 11:41:30 +02002569tune.ssl.keylog { on | off }
2570 This option activates the logging of the TLS keys. It should be used with
2571 care as it will consume more memory per SSL session and could decrease
2572 performances. This is disabled by default.
2573
2574 These sample fetches should be used to generate the SSLKEYLOGFILE that is
2575 required to decipher traffic with wireshark.
2576
2577 https://developer.mozilla.org/en-US/docs/Mozilla/Projects/NSS/Key_Log_Format
2578
2579 The SSLKEYLOG is a series of lines which are formatted this way:
2580
2581 <Label> <space> <ClientRandom> <space> <Secret>
2582
2583 The ClientRandom is provided by the %[ssl_fc_client_random,hex] sample
2584 fetch, the secret and the Label could be find in the array below. You need
2585 to generate a SSLKEYLOGFILE with all the labels in this array.
2586
2587 The following sample fetches are hexadecimal strings and does not need to be
2588 converted.
2589
2590 SSLKEYLOGFILE Label | Sample fetches for the Secrets
2591 --------------------------------|-----------------------------------------
2592 CLIENT_EARLY_TRAFFIC_SECRET | %[ssl_fc_client_early_traffic_secret]
2593 CLIENT_HANDSHAKE_TRAFFIC_SECRET | %[ssl_fc_client_handshake_traffic_secret]
2594 SERVER_HANDSHAKE_TRAFFIC_SECRET | %[ssl_fc_server_handshake_traffic_secret]
2595 CLIENT_TRAFFIC_SECRET_0 | %[ssl_fc_client_traffic_secret_0]
2596 SERVER_TRAFFIC_SECRET_0 | %[ssl_fc_server_traffic_secret_0]
William Lallemandd742b6c2020-07-07 10:14:56 +02002597 EXPORTER_SECRET | %[ssl_fc_exporter_secret]
2598 EARLY_EXPORTER_SECRET | %[ssl_fc_early_exporter_secret]
William Lallemand7d42ef52020-07-06 11:41:30 +02002599
2600 This is only available with OpenSSL 1.1.1, and useful with TLS1.3 session.
2601
2602 If you want to generate the content of a SSLKEYLOGFILE with TLS < 1.3, you
2603 only need this line:
2604
2605 "CLIENT_RANDOM %[ssl_fc_client_random,hex] %[ssl_fc_session_key,hex]"
2606
Emeric Brun4f65bff2012-11-16 15:11:00 +01002607tune.ssl.lifetime <timeout>
2608 Sets how long a cached SSL session may remain valid. This time is expressed
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03002609 in seconds and defaults to 300 (5 min). It is important to understand that it
Emeric Brun4f65bff2012-11-16 15:11:00 +01002610 does not guarantee that sessions will last that long, because if the cache is
2611 full, the longest idle sessions will be purged despite their configured
2612 lifetime. The real usefulness of this setting is to prevent sessions from
2613 being used for too long.
2614
Willy Tarreaubfd59462013-02-21 07:46:09 +01002615tune.ssl.maxrecord <number>
2616 Sets the maximum amount of bytes passed to SSL_write() at a time. Default
2617 value 0 means there is no limit. Over SSL/TLS, the client can decipher the
2618 data only once it has received a full record. With large records, it means
2619 that clients might have to download up to 16kB of data before starting to
2620 process them. Limiting the value can improve page load times on browsers
2621 located over high latency or low bandwidth networks. It is suggested to find
2622 optimal values which fit into 1 or 2 TCP segments (generally 1448 bytes over
2623 Ethernet with TCP timestamps enabled, or 1460 when timestamps are disabled),
2624 keeping in mind that SSL/TLS add some overhead. Typical values of 1419 and
2625 2859 gave good results during tests. Use "strace -e trace=write" to find the
Davor Ocelice9ed2812017-12-25 17:49:28 +01002626 best value. HAProxy will automatically switch to this setting after an idle
Willy Tarreau7e312732014-02-12 16:35:14 +01002627 stream has been detected (see tune.idletimer above).
Willy Tarreaubfd59462013-02-21 07:46:09 +01002628
Remi Gacognef46cd6e2014-06-12 14:58:40 +02002629tune.ssl.default-dh-param <number>
2630 Sets the maximum size of the Diffie-Hellman parameters used for generating
2631 the ephemeral/temporary Diffie-Hellman key in case of DHE key exchange. The
2632 final size will try to match the size of the server's RSA (or DSA) key (e.g,
2633 a 2048 bits temporary DH key for a 2048 bits RSA key), but will not exceed
Willy Tarreau3ba77d22020-05-08 09:31:18 +02002634 this maximum value. Default value if 2048. Only 1024 or higher values are
Remi Gacognef46cd6e2014-06-12 14:58:40 +02002635 allowed. Higher values will increase the CPU load, and values greater than
2636 1024 bits are not supported by Java 7 and earlier clients. This value is not
Remi Gacogne47783ef2015-05-29 15:53:22 +02002637 used if static Diffie-Hellman parameters are supplied either directly
2638 in the certificate file or by using the ssl-dh-param-file parameter.
Remi Gacognef46cd6e2014-06-12 14:58:40 +02002639
Christopher Faulet31af49d2015-06-09 17:29:50 +02002640tune.ssl.ssl-ctx-cache-size <number>
2641 Sets the size of the cache used to store generated certificates to <number>
2642 entries. This is a LRU cache. Because generating a SSL certificate
2643 dynamically is expensive, they are cached. The default cache size is set to
2644 1000 entries.
2645
Thierry FOURNIER5bf77322017-02-25 12:45:22 +01002646tune.ssl.capture-cipherlist-size <number>
2647 Sets the maximum size of the buffer used for capturing client-hello cipher
2648 list. If the value is 0 (default value) the capture is disabled, otherwise
2649 a buffer is allocated for each SSL/TLS connection.
2650
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02002651tune.vars.global-max-size <size>
Christopher Fauletff2613e2016-11-09 11:36:17 +01002652tune.vars.proc-max-size <size>
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02002653tune.vars.reqres-max-size <size>
2654tune.vars.sess-max-size <size>
2655tune.vars.txn-max-size <size>
Christopher Fauletff2613e2016-11-09 11:36:17 +01002656 These five tunes help to manage the maximum amount of memory used by the
2657 variables system. "global" limits the overall amount of memory available for
2658 all scopes. "proc" limits the memory for the process scope, "sess" limits the
2659 memory for the session scope, "txn" for the transaction scope, and "reqres"
2660 limits the memory for each request or response processing.
2661 Memory accounting is hierarchical, meaning more coarse grained limits include
2662 the finer grained ones: "proc" includes "sess", "sess" includes "txn", and
2663 "txn" includes "reqres".
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02002664
Daniel Schneller0b547052016-03-21 20:46:57 +01002665 For example, when "tune.vars.sess-max-size" is limited to 100,
2666 "tune.vars.txn-max-size" and "tune.vars.reqres-max-size" cannot exceed
2667 100 either. If we create a variable "txn.var" that contains 100 bytes,
2668 all available space is consumed.
2669 Notice that exceeding the limits at runtime will not result in an error
2670 message, but values might be cut off or corrupted. So make sure to accurately
2671 plan for the amount of space needed to store all your variables.
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02002672
William Lallemanda509e4c2012-11-07 16:54:34 +01002673tune.zlib.memlevel <number>
2674 Sets the memLevel parameter in zlib initialization for each session. It
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03002675 defines how much memory should be allocated for the internal compression
William Lallemanda509e4c2012-11-07 16:54:34 +01002676 state. A value of 1 uses minimum memory but is slow and reduces compression
Davor Ocelice9ed2812017-12-25 17:49:28 +01002677 ratio, a value of 9 uses maximum memory for optimal speed. Can be a value
William Lallemanda509e4c2012-11-07 16:54:34 +01002678 between 1 and 9. The default value is 8.
2679
2680tune.zlib.windowsize <number>
2681 Sets the window size (the size of the history buffer) as a parameter of the
2682 zlib initialization for each session. Larger values of this parameter result
Davor Ocelice9ed2812017-12-25 17:49:28 +01002683 in better compression at the expense of memory usage. Can be a value between
2684 8 and 15. The default value is 15.
Willy Tarreau6a06a402007-07-15 20:15:28 +02002685
Willy Tarreauc57f0e22009-05-10 13:12:33 +020026863.3. Debugging
2687--------------
Willy Tarreau6a06a402007-07-15 20:15:28 +02002688
Willy Tarreau6a06a402007-07-15 20:15:28 +02002689quiet
2690 Do not display any message during startup. It is equivalent to the command-
2691 line argument "-q".
2692
Willy Tarreau3eb10b82020-04-15 16:42:39 +02002693zero-warning
2694 When this option is set, haproxy will refuse to start if any warning was
2695 emitted while processing the configuration. It is highly recommended to set
2696 this option on configurations that are not changed often, as it helps detect
2697 subtle mistakes and keep the configuration clean and forward-compatible. Note
2698 that "haproxy -c" will also report errors in such a case. This option is
2699 equivalent to command line argument "-dW".
2700
Emeric Brunf099e792010-09-27 12:05:28 +02002701
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +010027023.4. Userlists
2703--------------
2704It is possible to control access to frontend/backend/listen sections or to
2705http stats by allowing only authenticated and authorized users. To do this,
2706it is required to create at least one userlist and to define users.
2707
2708userlist <listname>
Cyril Bonté78caf842010-03-10 22:41:43 +01002709 Creates new userlist with name <listname>. Many independent userlists can be
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01002710 used to store authentication & authorization data for independent customers.
2711
2712group <groupname> [users <user>,<user>,(...)]
Cyril Bonté78caf842010-03-10 22:41:43 +01002713 Adds group <groupname> to the current userlist. It is also possible to
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01002714 attach users to this group by using a comma separated list of names
2715 proceeded by "users" keyword.
2716
Cyril Bontéf0c60612010-02-06 14:44:47 +01002717user <username> [password|insecure-password <password>]
2718 [groups <group>,<group>,(...)]
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01002719 Adds user <username> to the current userlist. Both secure (encrypted) and
2720 insecure (unencrypted) passwords can be used. Encrypted passwords are
Daniel Schnellerd06f31c2017-11-06 16:51:04 +01002721 evaluated using the crypt(3) function, so depending on the system's
2722 capabilities, different algorithms are supported. For example, modern Glibc
2723 based Linux systems support MD5, SHA-256, SHA-512, and, of course, the
2724 classic DES-based method of encrypting passwords.
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01002725
Daniel Schnellerd06f31c2017-11-06 16:51:04 +01002726 Attention: Be aware that using encrypted passwords might cause significantly
2727 increased CPU usage, depending on the number of requests, and the algorithm
2728 used. For any of the hashed variants, the password for each request must
2729 be processed through the chosen algorithm, before it can be compared to the
2730 value specified in the config file. Most current algorithms are deliberately
2731 designed to be expensive to compute to achieve resistance against brute
2732 force attacks. They do not simply salt/hash the clear text password once,
2733 but thousands of times. This can quickly become a major factor in haproxy's
2734 overall CPU consumption!
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01002735
2736 Example:
Cyril Bontéf0c60612010-02-06 14:44:47 +01002737 userlist L1
2738 group G1 users tiger,scott
2739 group G2 users xdb,scott
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01002740
Cyril Bontéf0c60612010-02-06 14:44:47 +01002741 user tiger password $6$k6y3o.eP$JlKBx9za9667qe4(...)xHSwRv6J.C0/D7cV91
2742 user scott insecure-password elgato
2743 user xdb insecure-password hello
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01002744
Cyril Bontéf0c60612010-02-06 14:44:47 +01002745 userlist L2
2746 group G1
2747 group G2
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01002748
Cyril Bontéf0c60612010-02-06 14:44:47 +01002749 user tiger password $6$k6y3o.eP$JlKBx(...)xHSwRv6J.C0/D7cV91 groups G1
2750 user scott insecure-password elgato groups G1,G2
2751 user xdb insecure-password hello groups G2
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01002752
2753 Please note that both lists are functionally identical.
Willy Tarreau6a06a402007-07-15 20:15:28 +02002754
Emeric Brunf099e792010-09-27 12:05:28 +02002755
27563.5. Peers
Cyril Bontédc4d9032012-04-08 21:57:39 +02002757----------
Emeric Brun94900952015-06-11 18:25:54 +02002758It is possible to propagate entries of any data-types in stick-tables between
2759several haproxy instances over TCP connections in a multi-master fashion. Each
2760instance pushes its local updates and insertions to remote peers. The pushed
2761values overwrite remote ones without aggregation. Interrupted exchanges are
2762automatically detected and recovered from the last known point.
2763In addition, during a soft restart, the old process connects to the new one
2764using such a TCP connection to push all its entries before the new process
2765tries to connect to other peers. That ensures very fast replication during a
2766reload, it typically takes a fraction of a second even for large tables.
2767Note that Server IDs are used to identify servers remotely, so it is important
2768that configurations look similar or at least that the same IDs are forced on
2769each server on all participants.
Emeric Brunf099e792010-09-27 12:05:28 +02002770
2771peers <peersect>
Jamie Gloudon801a0a32012-08-25 00:18:33 -04002772 Creates a new peer list with name <peersect>. It is an independent section,
Emeric Brunf099e792010-09-27 12:05:28 +02002773 which is referenced by one or more stick-tables.
2774
Frédéric Lécaille2f167b32019-01-11 14:13:54 +01002775bind [<address>]:<port_range> [, ...] [param*]
2776 Defines the binding parameters of the local peer of this "peers" section.
2777 Such lines are not supported with "peer" line in the same "peers" section.
2778
Willy Tarreau77e4bd12015-05-01 20:02:17 +02002779disabled
2780 Disables a peers section. It disables both listening and any synchronization
2781 related to this section. This is provided to disable synchronization of stick
2782 tables without having to comment out all "peers" references.
2783
Frédéric Lécaille2f167b32019-01-11 14:13:54 +01002784default-bind [param*]
2785 Defines the binding parameters for the local peer, excepted its address.
2786
2787default-server [param*]
2788 Change default options for a server in a "peers" section.
2789
2790 Arguments:
2791 <param*> is a list of parameters for this server. The "default-server"
2792 keyword accepts an important number of options and has a complete
2793 section dedicated to it. Please refer to section 5 for more
2794 details.
2795
2796
2797 See also: "server" and section 5 about server options
2798
Willy Tarreau77e4bd12015-05-01 20:02:17 +02002799enable
2800 This re-enables a disabled peers section which was previously disabled.
2801
Jan Wagner3e678602020-12-17 22:22:32 +01002802log <address> [len <length>] [format <format>] [sample <ranges>:<sample_size>]
Frédéric Lécailleb6f759b2019-11-05 09:57:45 +01002803 <facility> [<level> [<minlevel>]]
2804 "peers" sections support the same "log" keyword as for the proxies to
2805 log information about the "peers" listener. See "log" option for proxies for
2806 more details.
2807
Frédéric Lécaille2f167b32019-01-11 14:13:54 +01002808peer <peername> <ip>:<port> [param*]
Emeric Brunf099e792010-09-27 12:05:28 +02002809 Defines a peer inside a peers section.
2810 If <peername> is set to the local peer name (by default hostname, or forced
Dragan Dosen13cd54c2020-06-18 18:24:05 +02002811 using "-L" command line option or "localpeer" global configuration setting),
2812 haproxy will listen for incoming remote peer connection on <ip>:<port>.
2813 Otherwise, <ip>:<port> defines where to connect to in order to join the
2814 remote peer, and <peername> is used at the protocol level to identify and
2815 validate the remote peer on the server side.
Emeric Brunf099e792010-09-27 12:05:28 +02002816
2817 During a soft restart, local peer <ip>:<port> is used by the old instance to
2818 connect the new one and initiate a complete replication (teaching process).
2819
2820 It is strongly recommended to have the exact same peers declaration on all
Dragan Dosen13cd54c2020-06-18 18:24:05 +02002821 peers and to only rely on the "-L" command line argument or the "localpeer"
2822 global configuration setting to change the local peer name. This makes it
2823 easier to maintain coherent configuration files across all peers.
Emeric Brunf099e792010-09-27 12:05:28 +02002824
William Lallemandb2f07452015-05-12 14:27:13 +02002825 You may want to reference some environment variables in the address
2826 parameter, see section 2.3 about environment variables.
Willy Tarreaudad36a32013-03-11 01:20:04 +01002827
Frédéric Lécaille2f167b32019-01-11 14:13:54 +01002828 Note: "peer" keyword may transparently be replaced by "server" keyword (see
2829 "server" keyword explanation below).
2830
2831server <peername> [<ip>:<port>] [param*]
Michael Prokop4438c602019-05-24 10:25:45 +02002832 As previously mentioned, "peer" keyword may be replaced by "server" keyword
Frédéric Lécaille2f167b32019-01-11 14:13:54 +01002833 with a support for all "server" parameters found in 5.2 paragraph.
2834 If the underlying peer is local, <ip>:<port> parameters must not be present.
2835 These parameters must be provided on a "bind" line (see "bind" keyword
2836 of this "peers" section).
2837 Some of these parameters are irrelevant for "peers" sections.
2838
2839
Cyril Bontédc4d9032012-04-08 21:57:39 +02002840 Example:
Frédéric Lécaille2f167b32019-01-11 14:13:54 +01002841 # The old way.
Emeric Brunf099e792010-09-27 12:05:28 +02002842 peers mypeers
Willy Tarreauf7b30a92010-12-06 22:59:17 +01002843 peer haproxy1 192.168.0.1:1024
2844 peer haproxy2 192.168.0.2:1024
2845 peer haproxy3 10.2.0.1:1024
Emeric Brunf099e792010-09-27 12:05:28 +02002846
2847 backend mybackend
2848 mode tcp
2849 balance roundrobin
2850 stick-table type ip size 20k peers mypeers
2851 stick on src
2852
Willy Tarreauf7b30a92010-12-06 22:59:17 +01002853 server srv1 192.168.0.30:80
2854 server srv2 192.168.0.31:80
Emeric Brunf099e792010-09-27 12:05:28 +02002855
Frédéric Lécaille2f167b32019-01-11 14:13:54 +01002856 Example:
2857 peers mypeers
2858 bind 127.0.0.11:10001 ssl crt mycerts/pem
2859 default-server ssl verify none
2860 server hostA 127.0.0.10:10000
2861 server hostB #local peer
Emeric Brunf099e792010-09-27 12:05:28 +02002862
Frédéric Lécaille4f5b77c2019-03-18 14:05:58 +01002863
2864table <tablename> type {ip | integer | string [len <length>] | binary [len <length>]}
2865 size <size> [expire <expire>] [nopurge] [store <data_type>]*
2866
2867 Configure a stickiness table for the current section. This line is parsed
2868 exactly the same way as the "stick-table" keyword in others section, except
John Roeslerfb2fce12019-07-10 15:45:51 -05002869 for the "peers" argument which is not required here and with an additional
Frédéric Lécaille4f5b77c2019-03-18 14:05:58 +01002870 mandatory first parameter to designate the stick-table. Contrary to others
2871 sections, there may be several "table" lines in "peers" sections (see also
2872 "stick-table" keyword).
2873
2874 Also be aware of the fact that "peers" sections have their own stick-table
2875 namespaces to avoid collisions between stick-table names identical in
2876 different "peers" section. This is internally handled prepending the "peers"
2877 sections names to the name of the stick-tables followed by a '/' character.
2878 If somewhere else in the configuration file you have to refer to such
2879 stick-tables declared in "peers" sections you must use the prefixed version
2880 of the stick-table name as follows:
2881
2882 peers mypeers
2883 peer A ...
2884 peer B ...
2885 table t1 ...
2886
2887 frontend fe1
2888 tcp-request content track-sc0 src table mypeers/t1
2889
2890 This is also this prefixed version of the stick-table names which must be
2891 used to refer to stick-tables through the CLI.
2892
2893 About "peers" protocol, as only "peers" belonging to the same section may
2894 communicate with each others, there is no need to do such a distinction.
2895 Several "peers" sections may declare stick-tables with the same name.
2896 This is shorter version of the stick-table name which is sent over the network.
2897 There is only a '/' character as prefix to avoid stick-table name collisions between
2898 stick-tables declared as backends and stick-table declared in "peers" sections
2899 as follows in this weird but supported configuration:
2900
2901 peers mypeers
2902 peer A ...
2903 peer B ...
2904 table t1 type string size 10m store gpc0
2905
2906 backend t1
2907 stick-table type string size 10m store gpc0 peers mypeers
2908
Daniel Corbett67a82712020-07-06 23:01:19 -04002909 Here "t1" table declared in "mypeers" section has "mypeers/t1" as global name.
Frédéric Lécaille4f5b77c2019-03-18 14:05:58 +01002910 "t1" table declared as a backend as "t1" as global name. But at peer protocol
2911 level the former table is named "/t1", the latter is again named "t1".
2912
Simon Horman51a1cf62015-02-03 13:00:44 +090029133.6. Mailers
2914------------
2915It is possible to send email alerts when the state of servers changes.
2916If configured email alerts are sent to each mailer that is configured
2917in a mailers section. Email is sent to mailers using SMTP.
2918
Pieter Baauw386a1272015-08-16 15:26:24 +02002919mailers <mailersect>
Simon Horman51a1cf62015-02-03 13:00:44 +09002920 Creates a new mailer list with the name <mailersect>. It is an
2921 independent section which is referenced by one or more proxies.
2922
2923mailer <mailername> <ip>:<port>
2924 Defines a mailer inside a mailers section.
2925
2926 Example:
2927 mailers mymailers
2928 mailer smtp1 192.168.0.1:587
2929 mailer smtp2 192.168.0.2:587
2930
2931 backend mybackend
2932 mode tcp
2933 balance roundrobin
2934
2935 email-alert mailers mymailers
2936 email-alert from test1@horms.org
2937 email-alert to test2@horms.org
2938
2939 server srv1 192.168.0.30:80
2940 server srv2 192.168.0.31:80
2941
Pieter Baauw235fcfc2016-02-13 15:33:40 +01002942timeout mail <time>
2943 Defines the time available for a mail/connection to be made and send to
2944 the mail-server. If not defined the default value is 10 seconds. To allow
2945 for at least two SYN-ACK packets to be send during initial TCP handshake it
2946 is advised to keep this value above 4 seconds.
2947
2948 Example:
2949 mailers mymailers
2950 timeout mail 20s
2951 mailer smtp1 192.168.0.1:587
Simon Horman51a1cf62015-02-03 13:00:44 +09002952
William Lallemandc9515522019-06-12 16:32:11 +020029533.7. Programs
2954-------------
2955In master-worker mode, it is possible to launch external binaries with the
2956master, these processes are called programs. These programs are launched and
2957managed the same way as the workers.
2958
2959During a reload of HAProxy, those processes are dealing with the same
2960sequence as a worker:
2961
2962 - the master is re-executed
2963 - the master sends a SIGUSR1 signal to the program
2964 - if "option start-on-reload" is not disabled, the master launches a new
2965 instance of the program
2966
2967During a stop, or restart, a SIGTERM is sent to the programs.
2968
2969program <name>
2970 This is a new program section, this section will create an instance <name>
2971 which is visible in "show proc" on the master CLI. (See "9.4. Master CLI" in
2972 the management guide).
2973
2974command <command> [arguments*]
2975 Define the command to start with optional arguments. The command is looked
2976 up in the current PATH if it does not include an absolute path. This is a
2977 mandatory option of the program section. Arguments containing spaces must
2978 be enclosed in quotes or double quotes or be prefixed by a backslash.
2979
Andrew Heberle97236962019-07-12 11:50:26 +08002980user <user name>
2981 Changes the executed command user ID to the <user name> from /etc/passwd.
2982 See also "group".
2983
2984group <group name>
2985 Changes the executed command group ID to the <group name> from /etc/group.
2986 See also "user".
2987
William Lallemandc9515522019-06-12 16:32:11 +02002988option start-on-reload
2989no option start-on-reload
2990 Start (or not) a new instance of the program upon a reload of the master.
2991 The default is to start a new instance. This option may only be used in a
2992 program section.
2993
2994
Christopher Faulet76edc0f2020-01-13 15:52:01 +010029953.8. HTTP-errors
2996----------------
2997
2998It is possible to globally declare several groups of HTTP errors, to be
2999imported afterwards in any proxy section. Same group may be referenced at
3000several places and can be fully or partially imported.
3001
3002http-errors <name>
3003 Create a new http-errors group with the name <name>. It is an independent
3004 section that may be referenced by one or more proxies using its name.
3005
3006errorfile <code> <file>
3007 Associate a file contents to an HTTP error code
3008
3009 Arguments :
3010 <code> is the HTTP status code. Currently, HAProxy is capable of
Christopher Faulet612f2ea2020-05-27 09:57:28 +02003011 generating codes 200, 400, 401, 403, 404, 405, 407, 408, 410,
Christopher Faulete095f312020-12-07 11:22:24 +01003012 425, 429, 500, 501, 502, 503, and 504.
Christopher Faulet76edc0f2020-01-13 15:52:01 +01003013
3014 <file> designates a file containing the full HTTP response. It is
3015 recommended to follow the common practice of appending ".http" to
3016 the filename so that people do not confuse the response with HTML
3017 error pages, and to use absolute paths, since files are read
3018 before any chroot is performed.
3019
3020 Please referrers to "errorfile" keyword in section 4 for details.
3021
3022 Example:
3023 http-errors website-1
3024 errorfile 400 /etc/haproxy/errorfiles/site1/400.http
3025 errorfile 404 /etc/haproxy/errorfiles/site1/404.http
3026 errorfile 408 /dev/null # work around Chrome pre-connect bug
3027
3028 http-errors website-2
3029 errorfile 400 /etc/haproxy/errorfiles/site2/400.http
3030 errorfile 404 /etc/haproxy/errorfiles/site2/404.http
3031 errorfile 408 /dev/null # work around Chrome pre-connect bug
3032
Emeric Brun99c453d2020-05-25 15:01:04 +020030333.9. Rings
3034----------
3035
3036It is possible to globally declare ring-buffers, to be used as target for log
3037servers or traces.
3038
3039ring <ringname>
3040 Creates a new ring-buffer with name <ringname>.
3041
3042description <text>
Daniel Corbett67a82712020-07-06 23:01:19 -04003043 The description is an optional description string of the ring. It will
Emeric Brun99c453d2020-05-25 15:01:04 +02003044 appear on CLI. By default, <name> is reused to fill this field.
3045
3046format <format>
3047 Format used to store events into the ring buffer.
3048
3049 Arguments:
3050 <format> is the log format used when generating syslog messages. It may be
3051 one of the following :
3052
3053 iso A message containing only the ISO date, followed by the text.
3054 The PID, process name and system name are omitted. This is
3055 designed to be used with a local log server.
3056
Emeric Brun0237c4e2020-11-27 16:24:34 +01003057 local Analog to rfc3164 syslog message format except that hostname
3058 field is stripped. This is the default.
3059 Note: option "log-send-hostname" switches the default to
3060 rfc3164.
3061
Emeric Brun99c453d2020-05-25 15:01:04 +02003062 raw A message containing only the text. The level, PID, date, time,
3063 process name and system name are omitted. This is designed to be
3064 used in containers or during development, where the severity
3065 only depends on the file descriptor used (stdout/stderr). This
3066 is the default.
3067
Emeric Brun0237c4e2020-11-27 16:24:34 +01003068 rfc3164 The RFC3164 syslog message format.
Emeric Brun99c453d2020-05-25 15:01:04 +02003069 (https://tools.ietf.org/html/rfc3164)
3070
3071 rfc5424 The RFC5424 syslog message format.
3072 (https://tools.ietf.org/html/rfc5424)
3073
3074 short A message containing only a level between angle brackets such as
3075 '<3>', followed by the text. The PID, date, time, process name
3076 and system name are omitted. This is designed to be used with a
3077 local log server. This format is compatible with what the systemd
3078 logger consumes.
3079
Emeric Brun54648852020-07-06 15:54:06 +02003080 priority A message containing only a level plus syslog facility between angle
3081 brackets such as '<63>', followed by the text. The PID, date, time,
3082 process name and system name are omitted. This is designed to be used
3083 with a local log server.
3084
Emeric Brun99c453d2020-05-25 15:01:04 +02003085 timed A message containing only a level between angle brackets such as
3086 '<3>', followed by ISO date and by the text. The PID, process
3087 name and system name are omitted. This is designed to be
3088 used with a local log server.
3089
3090maxlen <length>
3091 The maximum length of an event message stored into the ring,
3092 including formatted header. If an event message is longer than
3093 <length>, it will be truncated to this length.
3094
Emeric Brun494c5052020-05-28 11:13:15 +02003095server <name> <address> [param*]
3096 Used to configure a syslog tcp server to forward messages from ring buffer.
3097 This supports for all "server" parameters found in 5.2 paragraph. Some of
3098 these parameters are irrelevant for "ring" sections. Important point: there
3099 is little reason to add more than one server to a ring, because all servers
3100 will receive the exact same copy of the ring contents, and as such the ring
3101 will progress at the speed of the slowest server. If one server does not
3102 respond, it will prevent old messages from being purged and may block new
3103 messages from being inserted into the ring. The proper way to send messages
3104 to multiple servers is to use one distinct ring per log server, not to
Emeric Brun97556472020-05-30 01:42:45 +02003105 attach multiple servers to the same ring. Note that specific server directive
3106 "log-proto" is used to set the protocol used to send messages.
Emeric Brun494c5052020-05-28 11:13:15 +02003107
Emeric Brun99c453d2020-05-25 15:01:04 +02003108size <size>
3109 This is the optional size in bytes for the ring-buffer. Default value is
3110 set to BUFSIZE.
3111
Emeric Brun494c5052020-05-28 11:13:15 +02003112timeout connect <timeout>
3113 Set the maximum time to wait for a connection attempt to a server to succeed.
3114
3115 Arguments :
3116 <timeout> is the timeout value specified in milliseconds by default, but
3117 can be in any other unit if the number is suffixed by the unit,
3118 as explained at the top of this document.
3119
3120timeout server <timeout>
3121 Set the maximum time for pending data staying into output buffer.
3122
3123 Arguments :
3124 <timeout> is the timeout value specified in milliseconds by default, but
3125 can be in any other unit if the number is suffixed by the unit,
3126 as explained at the top of this document.
3127
Emeric Brun99c453d2020-05-25 15:01:04 +02003128 Example:
3129 global
3130 log ring@myring local7
3131
3132 ring myring
3133 description "My local buffer"
3134 format rfc3164
3135 maxlen 1200
3136 size 32764
Emeric Brun494c5052020-05-28 11:13:15 +02003137 timeout connect 5s
3138 timeout server 10s
Emeric Brun97556472020-05-30 01:42:45 +02003139 server mysyslogsrv 127.0.0.1:6514 log-proto octet-count
Emeric Brun99c453d2020-05-25 15:01:04 +02003140
Emeric Brun12941c82020-07-07 14:19:42 +020031413.10. Log forwarding
3142-------------------
3143
3144It is possible to declare one or multiple log forwarding section,
3145haproxy will forward all received log messages to a log servers list.
3146
3147log-forward <name>
3148 Creates a new log forwarder proxy identified as <name>.
3149
Emeric Bruncbb7bf72020-10-05 14:39:35 +02003150backlog <conns>
3151 Give hints to the system about the approximate listen backlog desired size
3152 on connections accept.
3153
3154bind <addr> [param*]
3155 Used to configure a stream log listener to receive messages to forward.
Emeric Brunda46c1c2020-10-08 08:39:02 +02003156 This supports the "bind" parameters found in 5.1 paragraph including
3157 those about ssl but some statements such as "alpn" may be irrelevant for
3158 syslog protocol over TCP.
3159 Those listeners support both "Octet Counting" and "Non-Transparent-Framing"
Emeric Bruncbb7bf72020-10-05 14:39:35 +02003160 modes as defined in rfc-6587.
3161
Willy Tarreau76aaa7f2020-09-16 15:07:22 +02003162dgram-bind <addr> [param*]
Emeric Bruncbb7bf72020-10-05 14:39:35 +02003163 Used to configure a datagram log listener to receive messages to forward.
3164 Addresses must be in IPv4 or IPv6 form,followed by a port. This supports
3165 for some of the "bind" parameters found in 5.1 paragraph among which
3166 "interface", "namespace" or "transparent", the other ones being
Willy Tarreau26ff5da2020-09-16 15:22:19 +02003167 silently ignored as irrelevant for UDP/syslog case.
Emeric Brun12941c82020-07-07 14:19:42 +02003168
3169log global
Jan Wagner3e678602020-12-17 22:22:32 +01003170log <address> [len <length>] [format <format>] [sample <ranges>:<sample_size>]
Emeric Brun12941c82020-07-07 14:19:42 +02003171 <facility> [<level> [<minlevel>]]
3172 Used to configure target log servers. See more details on proxies
3173 documentation.
3174 If no format specified, haproxy tries to keep the incoming log format.
3175 Configured facility is ignored, except if incoming message does not
3176 present a facility but one is mandatory on the outgoing format.
3177 If there is no timestamp available in the input format, but the field
3178 exists in output format, haproxy will use the local date.
3179
3180 Example:
3181 global
3182 log stderr format iso local7
3183
3184 ring myring
3185 description "My local buffer"
3186 format rfc5424
3187 maxlen 1200
3188 size 32764
3189 timeout connect 5s
3190 timeout server 10s
3191 # syslog tcp server
3192 server mysyslogsrv 127.0.0.1:514 log-proto octet-count
3193
3194 log-forward sylog-loadb
Emeric Bruncbb7bf72020-10-05 14:39:35 +02003195 dgram-bind 127.0.0.1:1514
3196 bind 127.0.0.1:1514
Emeric Brun12941c82020-07-07 14:19:42 +02003197 # all messages on stderr
3198 log global
3199 # all messages on local tcp syslog server
3200 log ring@myring local0
3201 # load balance messages on 4 udp syslog servers
3202 log 127.0.0.1:10001 sample 1:4 local0
3203 log 127.0.0.1:10002 sample 2:4 local0
3204 log 127.0.0.1:10003 sample 3:4 local0
3205 log 127.0.0.1:10004 sample 4:4 local0
Christopher Faulet76edc0f2020-01-13 15:52:01 +01003206
Emeric Bruncbb7bf72020-10-05 14:39:35 +02003207maxconn <conns>
3208 Fix the maximum number of concurrent connections on a log forwarder.
3209 10 is the default.
3210
3211timeout client <timeout>
3212 Set the maximum inactivity time on the client side.
3213
Willy Tarreauc57f0e22009-05-10 13:12:33 +020032144. Proxies
Willy Tarreau6a06a402007-07-15 20:15:28 +02003215----------
Willy Tarreau0ba27502007-12-24 16:55:16 +01003216
Willy Tarreau6a06a402007-07-15 20:15:28 +02003217Proxy configuration can be located in a set of sections :
Willy Tarreau7c0b4d82021-02-12 14:58:08 +01003218 - defaults [<name>] [ from <defaults_name> ]
3219 - frontend <name> [ from <defaults_name> ]
3220 - backend <name> [ from <defaults_name> ]
3221 - listen <name> [ from <defaults_name> ]
Willy Tarreau6a06a402007-07-15 20:15:28 +02003222
3223A "frontend" section describes a set of listening sockets accepting client
3224connections.
3225
3226A "backend" section describes a set of servers to which the proxy will connect
3227to forward incoming connections.
3228
3229A "listen" section defines a complete proxy with its frontend and backend
3230parts combined in one section. It is generally useful for TCP-only traffic.
3231
Willy Tarreau7c0b4d82021-02-12 14:58:08 +01003232A "defaults" section resets all settings to the documented ones and presets new
3233ones for use by subsequent sections. All of "frontend", "backend" and "listen"
3234sections always take their initial settings from a defaults section, by default
3235the latest one that appears before the newly created section. It is possible to
3236explicitly designate a specific "defaults" section to load the initial settings
3237from by indicating its name on the section line after the optional keyword
3238"from". While "defaults" section do not impose a name, this use is encouraged
3239for better readability. It is also the only way to designate a specific section
3240to use instead of the default previous one. Since "defaults" section names are
3241optional, by default a very permissive check is applied on their name and these
3242are even permitted to overlap. However if a "defaults" section is referenced by
3243any other section, its name must comply with the syntax imposed on all proxy
3244names, and this name must be unique among the defaults sections. Please note
3245that regardless of what is currently permitted, it is recommended to avoid
3246duplicate section names in general and to respect the same syntax as for proxy
3247names. This rule might be enforced in a future version.
3248
3249Note that it is even possible for a defaults section to take its initial
3250settings from another one, and as such, inherit settings across multiple levels
3251of defaults sections. This can be convenient to establish certain configuration
3252profiles to carry groups of default settings (e.g. TCP vs HTTP or short vs long
3253timeouts) but can quickly become confusing to follow.
3254
Willy Tarreau0ba27502007-12-24 16:55:16 +01003255All proxy names must be formed from upper and lower case letters, digits,
3256'-' (dash), '_' (underscore) , '.' (dot) and ':' (colon). ACL names are
3257case-sensitive, which means that "www" and "WWW" are two different proxies.
3258
3259Historically, all proxy names could overlap, it just caused troubles in the
3260logs. Since the introduction of content switching, it is mandatory that two
3261proxies with overlapping capabilities (frontend/backend) have different names.
3262However, it is still permitted that a frontend and a backend share the same
3263name, as this configuration seems to be commonly encountered.
3264
3265Right now, two major proxy modes are supported : "tcp", also known as layer 4,
3266and "http", also known as layer 7. In layer 4 mode, HAProxy simply forwards
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01003267bidirectional traffic between two sides. In layer 7 mode, HAProxy analyzes the
Willy Tarreau0ba27502007-12-24 16:55:16 +01003268protocol, and can interact with it by allowing, blocking, switching, adding,
3269modifying, or removing arbitrary contents in requests or responses, based on
3270arbitrary criteria.
3271
Willy Tarreau70dffda2014-01-30 03:07:23 +01003272In HTTP mode, the processing applied to requests and responses flowing over
3273a connection depends in the combination of the frontend's HTTP options and
Julien Pivotto21ad3152019-12-10 13:11:17 +01003274the backend's. HAProxy supports 3 connection modes :
Willy Tarreau70dffda2014-01-30 03:07:23 +01003275
3276 - KAL : keep alive ("option http-keep-alive") which is the default mode : all
3277 requests and responses are processed, and connections remain open but idle
3278 between responses and new requests.
3279
Willy Tarreau70dffda2014-01-30 03:07:23 +01003280 - SCL: server close ("option http-server-close") : the server-facing
3281 connection is closed after the end of the response is received, but the
3282 client-facing connection remains open.
3283
Christopher Faulet315b39c2018-09-21 16:26:19 +02003284 - CLO: close ("option httpclose"): the connection is closed after the end of
3285 the response and "Connection: close" appended in both directions.
Willy Tarreau70dffda2014-01-30 03:07:23 +01003286
3287The effective mode that will be applied to a connection passing through a
3288frontend and a backend can be determined by both proxy modes according to the
3289following matrix, but in short, the modes are symmetric, keep-alive is the
Christopher Faulet315b39c2018-09-21 16:26:19 +02003290weakest option and close is the strongest.
Willy Tarreau70dffda2014-01-30 03:07:23 +01003291
Christopher Faulet315b39c2018-09-21 16:26:19 +02003292 Backend mode
Willy Tarreau70dffda2014-01-30 03:07:23 +01003293
Christopher Faulet315b39c2018-09-21 16:26:19 +02003294 | KAL | SCL | CLO
3295 ----+-----+-----+----
3296 KAL | KAL | SCL | CLO
3297 ----+-----+-----+----
Christopher Faulet315b39c2018-09-21 16:26:19 +02003298 mode SCL | SCL | SCL | CLO
3299 ----+-----+-----+----
3300 CLO | CLO | CLO | CLO
Willy Tarreau70dffda2014-01-30 03:07:23 +01003301
Willy Tarreau0ba27502007-12-24 16:55:16 +01003302
Willy Tarreau70dffda2014-01-30 03:07:23 +01003303
Willy Tarreauc57f0e22009-05-10 13:12:33 +020033044.1. Proxy keywords matrix
3305--------------------------
Willy Tarreau0ba27502007-12-24 16:55:16 +01003306
Willy Tarreauc57f0e22009-05-10 13:12:33 +02003307The following list of keywords is supported. Most of them may only be used in a
3308limited set of section types. Some of them are marked as "deprecated" because
3309they are inherited from an old syntax which may be confusing or functionally
3310limited, and there are new recommended keywords to replace them. Keywords
Davor Ocelice9ed2812017-12-25 17:49:28 +01003311marked with "(*)" can be optionally inverted using the "no" prefix, e.g. "no
Willy Tarreauc57f0e22009-05-10 13:12:33 +02003312option contstats". This makes sense when the option has been enabled by default
Willy Tarreau3842f002009-06-14 11:39:52 +02003313and must be disabled for a specific instance. Such options may also be prefixed
3314with "default" in order to restore default settings regardless of what has been
3315specified in a previous "defaults" section.
Willy Tarreau0ba27502007-12-24 16:55:16 +01003316
Willy Tarreau6a06a402007-07-15 20:15:28 +02003317
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01003318 keyword defaults frontend listen backend
3319------------------------------------+----------+----------+---------+---------
3320acl - X X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01003321backlog X X X -
3322balance X - X X
3323bind - X X -
3324bind-process X X X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01003325capture cookie - X X -
3326capture request header - X X -
3327capture response header - X X -
MIZUTA Takeshib24bc0d2020-07-09 11:13:20 +09003328clitcpka-cnt X X X -
3329clitcpka-idle X X X -
3330clitcpka-intvl X X X -
William Lallemand82fe75c2012-10-23 10:25:10 +02003331compression X X X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01003332cookie X - X X
Thierry FOURNIERa0a1b752015-05-26 17:44:32 +02003333declare capture - X X -
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01003334default-server X - X X
3335default_backend X X X -
3336description - X X X
3337disabled X X X X
3338dispatch - - X X
Simon Horman51a1cf62015-02-03 13:00:44 +09003339email-alert from X X X X
Simon Horman64e34162015-02-06 11:11:57 +09003340email-alert level X X X X
Simon Horman51a1cf62015-02-03 13:00:44 +09003341email-alert mailers X X X X
3342email-alert myhostname X X X X
3343email-alert to X X X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01003344enabled X X X X
3345errorfile X X X X
Christopher Faulet76edc0f2020-01-13 15:52:01 +01003346errorfiles X X X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01003347errorloc X X X X
3348errorloc302 X X X X
3349-- keyword -------------------------- defaults - frontend - listen -- backend -
3350errorloc303 X X X X
Cyril Bonté4288c5a2018-03-12 22:02:59 +01003351force-persist - - X X
Christopher Fauletc3fe5332016-04-07 15:30:10 +02003352filter - X X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01003353fullconn X - X X
3354grace X X X X
3355hash-type X - X X
Christopher Faulet6d0c3df2020-01-22 09:26:35 +01003356http-after-response - X X X
Christopher Faulet4f5c2e22020-04-23 15:22:33 +02003357http-check comment X - X X
Christopher Faulete5870d82020-04-15 11:32:03 +02003358http-check connect X - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01003359http-check disable-on-404 X - X X
Christopher Faulete5870d82020-04-15 11:32:03 +02003360http-check expect X - X X
Peter Gervai8912ae62020-06-11 18:26:36 +02003361http-check send X - X X
Willy Tarreau7ab6aff2010-10-12 06:30:16 +02003362http-check send-state X - X X
Christopher Faulete5870d82020-04-15 11:32:03 +02003363http-check set-var X - X X
3364http-check unset-var X - X X
Christopher Faulet3b967c12020-05-15 15:47:44 +02003365http-error X X X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01003366http-request - X X X
Willy Tarreaue365c0b2013-06-11 16:06:12 +02003367http-response - X X X
Willy Tarreau30631952015-08-06 15:05:24 +02003368http-reuse X - X X
Baptiste Assmann2c42ef52013-10-09 21:57:02 +02003369http-send-name-header - - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01003370id - X X X
Cyril Bonté4288c5a2018-03-12 22:02:59 +01003371ignore-persist - - X X
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02003372load-server-state-from-file X - X X
William Lallemand0f99e342011-10-12 17:50:54 +02003373log (*) X X X X
Willy Tarreaufb4e7ea2015-01-07 14:55:17 +01003374log-format X X X -
Dragan Dosen7ad31542015-09-28 17:16:47 +02003375log-format-sd X X X -
Willy Tarreau094af4e2015-01-07 15:03:42 +01003376log-tag X X X X
Willy Tarreauc35362a2014-04-25 13:58:37 +02003377max-keep-alive-queue X - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01003378maxconn X X X -
3379mode X X X X
3380monitor fail - X X -
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01003381monitor-uri X X X -
3382option abortonclose (*) X - X X
3383option accept-invalid-http-request (*) X X X -
3384option accept-invalid-http-response (*) X - X X
3385option allbackups (*) X - X X
3386option checkcache (*) X - X X
3387option clitcpka (*) X X X -
3388option contstats (*) X X X -
Christopher Faulet89aed322020-06-02 17:33:56 +02003389option disable-h2-upgrade (*) X X X -
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01003390option dontlog-normal (*) X X X -
3391option dontlognull (*) X X X -
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01003392-- keyword -------------------------- defaults - frontend - listen -- backend -
3393option forwardfor X X X X
Christopher Faulet98fbe952019-07-22 16:18:24 +02003394option h1-case-adjust-bogus-client (*) X X X -
3395option h1-case-adjust-bogus-server (*) X - X X
Willy Tarreau9fbe18e2015-05-01 22:42:08 +02003396option http-buffer-request (*) X X X X
Willy Tarreau82649f92015-05-01 22:40:51 +02003397option http-ignore-probes (*) X X X -
Willy Tarreau16bfb022010-01-16 19:48:41 +01003398option http-keep-alive (*) X X X X
Willy Tarreau96e31212011-05-30 18:10:30 +02003399option http-no-delay (*) X X X X
Christopher Faulet98db9762018-09-21 10:25:19 +02003400option http-pretend-keepalive (*) X - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01003401option http-server-close (*) X X X X
3402option http-use-proxy-header (*) X X X -
3403option httpchk X - X X
3404option httpclose (*) X X X X
Freddy Spierenburge88b7732019-03-25 14:35:17 +01003405option httplog X X X -
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01003406option http_proxy (*) X X X X
Jamie Gloudon801a0a32012-08-25 00:18:33 -04003407option independent-streams (*) X X X X
Gabor Lekenyb4c81e42010-09-29 18:17:05 +02003408option ldap-check X - X X
Simon Horman98637e52014-06-20 12:30:16 +09003409option external-check X - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01003410option log-health-checks (*) X - X X
3411option log-separate-errors (*) X X X -
3412option logasap (*) X X X -
3413option mysql-check X - X X
3414option nolinger (*) X X X X
3415option originalto X X X X
3416option persist (*) X - X X
Baptiste Assmann809e22a2015-10-12 20:22:55 +02003417option pgsql-check X - X X
3418option prefer-last-server (*) X - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01003419option redispatch (*) X - X X
Hervé COMMOWICKec032d62011-08-05 16:23:48 +02003420option redis-check X - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01003421option smtpchk X - X X
3422option socket-stats (*) X X X -
3423option splice-auto (*) X X X X
3424option splice-request (*) X X X X
3425option splice-response (*) X X X X
Christopher Fauletba7bc162016-11-07 21:07:38 +01003426option spop-check - - - X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01003427option srvtcpka (*) X - X X
3428option ssl-hello-chk X - X X
3429-- keyword -------------------------- defaults - frontend - listen -- backend -
Willy Tarreaued179852013-12-16 01:07:00 +01003430option tcp-check X - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01003431option tcp-smart-accept (*) X X X -
3432option tcp-smart-connect (*) X - X X
3433option tcpka X X X X
3434option tcplog X X X X
3435option transparent (*) X - X X
Simon Horman98637e52014-06-20 12:30:16 +09003436external-check command X - X X
3437external-check path X - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01003438persist rdp-cookie X - X X
3439rate-limit sessions X X X -
3440redirect - X X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01003441-- keyword -------------------------- defaults - frontend - listen -- backend -
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01003442retries X - X X
Olivier Houcharda254a372019-04-05 15:30:12 +02003443retry-on X - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01003444server - - X X
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02003445server-state-file-name X - X X
Frédéric Lécaillecb4502e2017-04-20 13:36:25 +02003446server-template - - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01003447source X - X X
MIZUTA Takeshib24bc0d2020-07-09 11:13:20 +09003448srvtcpka-cnt X - X X
3449srvtcpka-idle X - X X
3450srvtcpka-intvl X - X X
Baptiste Assmann5a549212015-10-12 20:30:24 +02003451stats admin - X X X
3452stats auth X X X X
3453stats enable X X X X
3454stats hide-version X X X X
3455stats http-request - X X X
3456stats realm X X X X
3457stats refresh X X X X
3458stats scope X X X X
3459stats show-desc X X X X
3460stats show-legends X X X X
3461stats show-node X X X X
3462stats uri X X X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01003463-- keyword -------------------------- defaults - frontend - listen -- backend -
3464stick match - - X X
3465stick on - - X X
3466stick store-request - - X X
Willy Tarreaud8dc99f2011-07-01 11:33:25 +02003467stick store-response - - X X
Adam Spiers68af3c12017-04-06 16:31:39 +01003468stick-table - X X X
Christopher Faulet4f5c2e22020-04-23 15:22:33 +02003469tcp-check comment X - X X
Christopher Faulet404f9192020-04-09 23:13:54 +02003470tcp-check connect X - X X
3471tcp-check expect X - X X
3472tcp-check send X - X X
Christopher Fauletb50b3e62020-05-05 18:43:43 +02003473tcp-check send-lf X - X X
Christopher Faulet404f9192020-04-09 23:13:54 +02003474tcp-check send-binary X - X X
Christopher Fauletb50b3e62020-05-05 18:43:43 +02003475tcp-check send-binary-lf X - X X
Christopher Faulet404f9192020-04-09 23:13:54 +02003476tcp-check set-var X - X X
3477tcp-check unset-var X - X X
Willy Tarreaue9656522010-08-17 15:40:09 +02003478tcp-request connection - X X -
3479tcp-request content - X X X
Willy Tarreaua56235c2010-09-14 11:31:36 +02003480tcp-request inspect-delay - X X X
Willy Tarreau4f614292016-10-21 17:49:36 +02003481tcp-request session - X X -
Emeric Brun0a3b67f2010-09-24 15:34:53 +02003482tcp-response content - - X X
3483tcp-response inspect-delay - - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01003484timeout check X - X X
3485timeout client X X X -
Willy Tarreau05cdd962014-05-10 14:30:07 +02003486timeout client-fin X X X -
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01003487timeout connect X - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01003488timeout http-keep-alive X X X X
3489timeout http-request X X X X
3490timeout queue X - X X
3491timeout server X - X X
Willy Tarreau05cdd962014-05-10 14:30:07 +02003492timeout server-fin X - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01003493timeout tarpit X X X X
Willy Tarreauce887fd2012-05-12 12:50:00 +02003494timeout tunnel X - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01003495transparent (deprecated) X - X X
William Lallemanda73203e2012-03-12 12:48:57 +01003496unique-id-format X X X -
3497unique-id-header X X X -
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01003498use_backend - X X -
Christopher Fauletb30b3102019-09-12 23:03:09 +02003499use-fcgi-app - - X X
Willy Tarreau4a5cade2012-04-05 21:09:48 +02003500use-server - - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01003501------------------------------------+----------+----------+---------+---------
3502 keyword defaults frontend listen backend
Willy Tarreau6a06a402007-07-15 20:15:28 +02003503
Willy Tarreau0ba27502007-12-24 16:55:16 +01003504
Willy Tarreauc57f0e22009-05-10 13:12:33 +020035054.2. Alphabetically sorted keywords reference
3506---------------------------------------------
Willy Tarreau0ba27502007-12-24 16:55:16 +01003507
3508This section provides a description of each keyword and its usage.
3509
3510
3511acl <aclname> <criterion> [flags] [operator] <value> ...
3512 Declare or complete an access list.
3513 May be used in sections : defaults | frontend | listen | backend
3514 no | yes | yes | yes
3515 Example:
3516 acl invalid_src src 0.0.0.0/7 224.0.0.0/3
3517 acl invalid_src src_port 0:1023
3518 acl local_dst hdr(host) -i localhost
3519
Willy Tarreauc57f0e22009-05-10 13:12:33 +02003520 See section 7 about ACL usage.
Willy Tarreau0ba27502007-12-24 16:55:16 +01003521
3522
Willy Tarreauc73ce2b2008-01-06 10:55:10 +01003523backlog <conns>
3524 Give hints to the system about the approximate listen backlog desired size
3525 May be used in sections : defaults | frontend | listen | backend
3526 yes | yes | yes | no
3527 Arguments :
3528 <conns> is the number of pending connections. Depending on the operating
3529 system, it may represent the number of already acknowledged
Cyril Bontédc4d9032012-04-08 21:57:39 +02003530 connections, of non-acknowledged ones, or both.
Willy Tarreauc73ce2b2008-01-06 10:55:10 +01003531
3532 In order to protect against SYN flood attacks, one solution is to increase
3533 the system's SYN backlog size. Depending on the system, sometimes it is just
3534 tunable via a system parameter, sometimes it is not adjustable at all, and
3535 sometimes the system relies on hints given by the application at the time of
3536 the listen() syscall. By default, HAProxy passes the frontend's maxconn value
3537 to the listen() syscall. On systems which can make use of this value, it can
3538 sometimes be useful to be able to specify a different value, hence this
3539 backlog parameter.
3540
3541 On Linux 2.4, the parameter is ignored by the system. On Linux 2.6, it is
3542 used as a hint and the system accepts up to the smallest greater power of
3543 two, and never more than some limits (usually 32768).
3544
3545 See also : "maxconn" and the target operating system's tuning guide.
3546
3547
Willy Tarreau0ba27502007-12-24 16:55:16 +01003548balance <algorithm> [ <arguments> ]
Willy Tarreau226071e2014-04-10 11:55:45 +02003549balance url_param <param> [check_post]
Willy Tarreau0ba27502007-12-24 16:55:16 +01003550 Define the load balancing algorithm to be used in a backend.
3551 May be used in sections : defaults | frontend | listen | backend
3552 yes | no | yes | yes
3553 Arguments :
3554 <algorithm> is the algorithm used to select a server when doing load
3555 balancing. This only applies when no persistence information
3556 is available, or when a connection is redispatched to another
3557 server. <algorithm> may be one of the following :
3558
3559 roundrobin Each server is used in turns, according to their weights.
3560 This is the smoothest and fairest algorithm when the server's
3561 processing time remains equally distributed. This algorithm
3562 is dynamic, which means that server weights may be adjusted
Willy Tarreau9757a382009-10-03 12:56:50 +02003563 on the fly for slow starts for instance. It is limited by
Godbacha34bdc02013-07-22 07:44:53 +08003564 design to 4095 active servers per backend. Note that in some
Willy Tarreau9757a382009-10-03 12:56:50 +02003565 large farms, when a server becomes up after having been down
3566 for a very short time, it may sometimes take a few hundreds
3567 requests for it to be re-integrated into the farm and start
3568 receiving traffic. This is normal, though very rare. It is
3569 indicated here in case you would have the chance to observe
3570 it, so that you don't worry.
3571
3572 static-rr Each server is used in turns, according to their weights.
3573 This algorithm is as similar to roundrobin except that it is
3574 static, which means that changing a server's weight on the
3575 fly will have no effect. On the other hand, it has no design
3576 limitation on the number of servers, and when a server goes
3577 up, it is always immediately reintroduced into the farm, once
3578 the full map is recomputed. It also uses slightly less CPU to
3579 run (around -1%).
Willy Tarreau0ba27502007-12-24 16:55:16 +01003580
Willy Tarreau2d2a7f82008-03-17 12:07:56 +01003581 leastconn The server with the lowest number of connections receives the
3582 connection. Round-robin is performed within groups of servers
3583 of the same load to ensure that all servers will be used. Use
3584 of this algorithm is recommended where very long sessions are
3585 expected, such as LDAP, SQL, TSE, etc... but is not very well
3586 suited for protocols using short sessions such as HTTP. This
3587 algorithm is dynamic, which means that server weights may be
Willy Tarreau8c855f62020-10-22 17:41:45 +02003588 adjusted on the fly for slow starts for instance. It will
3589 also consider the number of queued connections in addition to
3590 the established ones in order to minimize queuing.
Willy Tarreau2d2a7f82008-03-17 12:07:56 +01003591
Willy Tarreauf09c6602012-02-13 17:12:08 +01003592 first The first server with available connection slots receives the
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03003593 connection. The servers are chosen from the lowest numeric
Willy Tarreauf09c6602012-02-13 17:12:08 +01003594 identifier to the highest (see server parameter "id"), which
3595 defaults to the server's position in the farm. Once a server
Willy Tarreau64559c52012-04-07 09:08:45 +02003596 reaches its maxconn value, the next server is used. It does
Willy Tarreauf09c6602012-02-13 17:12:08 +01003597 not make sense to use this algorithm without setting maxconn.
3598 The purpose of this algorithm is to always use the smallest
3599 number of servers so that extra servers can be powered off
3600 during non-intensive hours. This algorithm ignores the server
3601 weight, and brings more benefit to long session such as RDP
Willy Tarreau64559c52012-04-07 09:08:45 +02003602 or IMAP than HTTP, though it can be useful there too. In
3603 order to use this algorithm efficiently, it is recommended
3604 that a cloud controller regularly checks server usage to turn
3605 them off when unused, and regularly checks backend queue to
3606 turn new servers on when the queue inflates. Alternatively,
3607 using "http-check send-state" may inform servers on the load.
Willy Tarreauf09c6602012-02-13 17:12:08 +01003608
Willy Tarreau0ba27502007-12-24 16:55:16 +01003609 source The source IP address is hashed and divided by the total
3610 weight of the running servers to designate which server will
3611 receive the request. This ensures that the same client IP
3612 address will always reach the same server as long as no
3613 server goes down or up. If the hash result changes due to the
3614 number of running servers changing, many clients will be
3615 directed to a different server. This algorithm is generally
3616 used in TCP mode where no cookie may be inserted. It may also
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01003617 be used on the Internet to provide a best-effort stickiness
Willy Tarreau0ba27502007-12-24 16:55:16 +01003618 to clients which refuse session cookies. This algorithm is
Willy Tarreau6b2e11b2009-10-01 07:52:15 +02003619 static by default, which means that changing a server's
3620 weight on the fly will have no effect, but this can be
3621 changed using "hash-type".
Willy Tarreau0ba27502007-12-24 16:55:16 +01003622
Oskar Stolc8dc41842012-05-19 10:19:54 +01003623 uri This algorithm hashes either the left part of the URI (before
3624 the question mark) or the whole URI (if the "whole" parameter
3625 is present) and divides the hash value by the total weight of
3626 the running servers. The result designates which server will
3627 receive the request. This ensures that the same URI will
3628 always be directed to the same server as long as no server
3629 goes up or down. This is used with proxy caches and
3630 anti-virus proxies in order to maximize the cache hit rate.
3631 Note that this algorithm may only be used in an HTTP backend.
3632 This algorithm is static by default, which means that
3633 changing a server's weight on the fly will have no effect,
3634 but this can be changed using "hash-type".
Willy Tarreau0ba27502007-12-24 16:55:16 +01003635
Oskar Stolc8dc41842012-05-19 10:19:54 +01003636 This algorithm supports two optional parameters "len" and
Marek Majkowski9c30fc12008-04-27 23:25:55 +02003637 "depth", both followed by a positive integer number. These
3638 options may be helpful when it is needed to balance servers
3639 based on the beginning of the URI only. The "len" parameter
3640 indicates that the algorithm should only consider that many
3641 characters at the beginning of the URI to compute the hash.
3642 Note that having "len" set to 1 rarely makes sense since most
3643 URIs start with a leading "/".
3644
3645 The "depth" parameter indicates the maximum directory depth
3646 to be used to compute the hash. One level is counted for each
3647 slash in the request. If both parameters are specified, the
3648 evaluation stops when either is reached.
3649
Willy Tarreau57a37412020-09-23 08:56:29 +02003650 A "path-only" parameter indicates that the hashing key starts
3651 at the first '/' of the path. This can be used to ignore the
3652 authority part of absolute URIs, and to make sure that HTTP/1
3653 and HTTP/2 URIs will provide the same hash.
3654
Willy Tarreau0ba27502007-12-24 16:55:16 +01003655 url_param The URL parameter specified in argument will be looked up in
matt.farnsworth@nokia.com1c2ab962008-04-14 20:47:37 +02003656 the query string of each HTTP GET request.
3657
3658 If the modifier "check_post" is used, then an HTTP POST
Cyril Bontédc4d9032012-04-08 21:57:39 +02003659 request entity will be searched for the parameter argument,
3660 when it is not found in a query string after a question mark
Willy Tarreau226071e2014-04-10 11:55:45 +02003661 ('?') in the URL. The message body will only start to be
3662 analyzed once either the advertised amount of data has been
3663 received or the request buffer is full. In the unlikely event
3664 that chunked encoding is used, only the first chunk is
Cyril Bontédc4d9032012-04-08 21:57:39 +02003665 scanned. Parameter values separated by a chunk boundary, may
Willy Tarreau226071e2014-04-10 11:55:45 +02003666 be randomly balanced if at all. This keyword used to support
3667 an optional <max_wait> parameter which is now ignored.
matt.farnsworth@nokia.com1c2ab962008-04-14 20:47:37 +02003668
3669 If the parameter is found followed by an equal sign ('=') and
3670 a value, then the value is hashed and divided by the total
3671 weight of the running servers. The result designates which
3672 server will receive the request.
3673
3674 This is used to track user identifiers in requests and ensure
3675 that a same user ID will always be sent to the same server as
3676 long as no server goes up or down. If no value is found or if
3677 the parameter is not found, then a round robin algorithm is
3678 applied. Note that this algorithm may only be used in an HTTP
Willy Tarreau6b2e11b2009-10-01 07:52:15 +02003679 backend. This algorithm is static by default, which means
3680 that changing a server's weight on the fly will have no
3681 effect, but this can be changed using "hash-type".
Willy Tarreau0ba27502007-12-24 16:55:16 +01003682
Cyril Bontédc4d9032012-04-08 21:57:39 +02003683 hdr(<name>) The HTTP header <name> will be looked up in each HTTP
3684 request. Just as with the equivalent ACL 'hdr()' function,
3685 the header name in parenthesis is not case sensitive. If the
3686 header is absent or if it does not contain any value, the
3687 roundrobin algorithm is applied instead.
Benoitaffb4812009-03-25 13:02:10 +01003688
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01003689 An optional 'use_domain_only' parameter is available, for
Benoitaffb4812009-03-25 13:02:10 +01003690 reducing the hash algorithm to the main domain part with some
3691 specific headers such as 'Host'. For instance, in the Host
3692 value "haproxy.1wt.eu", only "1wt" will be considered.
3693
Willy Tarreau6b2e11b2009-10-01 07:52:15 +02003694 This algorithm is static by default, which means that
3695 changing a server's weight on the fly will have no effect,
3696 but this can be changed using "hash-type".
3697
Willy Tarreau21c741a2019-01-14 18:14:27 +01003698 random
3699 random(<draws>)
3700 A random number will be used as the key for the consistent
Willy Tarreau760e81d2018-05-03 07:20:40 +02003701 hashing function. This means that the servers' weights are
3702 respected, dynamic weight changes immediately take effect, as
3703 well as new server additions. Random load balancing can be
3704 useful with large farms or when servers are frequently added
Willy Tarreau21c741a2019-01-14 18:14:27 +01003705 or removed as it may avoid the hammering effect that could
3706 result from roundrobin or leastconn in this situation. The
3707 hash-balance-factor directive can be used to further improve
3708 fairness of the load balancing, especially in situations
3709 where servers show highly variable response times. When an
3710 argument <draws> is present, it must be an integer value one
3711 or greater, indicating the number of draws before selecting
3712 the least loaded of these servers. It was indeed demonstrated
3713 that picking the least loaded of two servers is enough to
3714 significantly improve the fairness of the algorithm, by
3715 always avoiding to pick the most loaded server within a farm
3716 and getting rid of any bias that could be induced by the
3717 unfair distribution of the consistent list. Higher values N
3718 will take away N-1 of the highest loaded servers at the
3719 expense of performance. With very high values, the algorithm
3720 will converge towards the leastconn's result but much slower.
3721 The default value is 2, which generally shows very good
3722 distribution and performance. This algorithm is also known as
3723 the Power of Two Random Choices and is described here :
3724 http://www.eecs.harvard.edu/~michaelm/postscripts/handbook2001.pdf
Willy Tarreau760e81d2018-05-03 07:20:40 +02003725
Emeric Brun736aa232009-06-30 17:56:00 +02003726 rdp-cookie
Hervé COMMOWICKa3eb39c2011-08-05 18:48:51 +02003727 rdp-cookie(<name>)
Emeric Brun736aa232009-06-30 17:56:00 +02003728 The RDP cookie <name> (or "mstshash" if omitted) will be
3729 looked up and hashed for each incoming TCP request. Just as
3730 with the equivalent ACL 'req_rdp_cookie()' function, the name
3731 is not case-sensitive. This mechanism is useful as a degraded
3732 persistence mode, as it makes it possible to always send the
3733 same user (or the same session ID) to the same server. If the
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01003734 cookie is not found, the normal roundrobin algorithm is
Emeric Brun736aa232009-06-30 17:56:00 +02003735 used instead.
3736
3737 Note that for this to work, the frontend must ensure that an
3738 RDP cookie is already present in the request buffer. For this
3739 you must use 'tcp-request content accept' rule combined with
3740 a 'req_rdp_cookie_cnt' ACL.
3741
Willy Tarreau6b2e11b2009-10-01 07:52:15 +02003742 This algorithm is static by default, which means that
3743 changing a server's weight on the fly will have no effect,
3744 but this can be changed using "hash-type".
3745
Cyril Bontédc4d9032012-04-08 21:57:39 +02003746 See also the rdp_cookie pattern fetch function.
Simon Hormanab814e02011-06-24 14:50:20 +09003747
Willy Tarreau0ba27502007-12-24 16:55:16 +01003748 <arguments> is an optional list of arguments which may be needed by some
Marek Majkowski9c30fc12008-04-27 23:25:55 +02003749 algorithms. Right now, only "url_param" and "uri" support an
3750 optional argument.
matt.farnsworth@nokia.com1c2ab962008-04-14 20:47:37 +02003751
Willy Tarreau3cd9af22009-03-15 14:06:41 +01003752 The load balancing algorithm of a backend is set to roundrobin when no other
3753 algorithm, mode nor option have been set. The algorithm may only be set once
3754 for each backend.
Willy Tarreau0ba27502007-12-24 16:55:16 +01003755
Lukas Tribus80512b12018-10-27 20:07:40 +02003756 With authentication schemes that require the same connection like NTLM, URI
John Roeslerfb2fce12019-07-10 15:45:51 -05003757 based algorithms must not be used, as they would cause subsequent requests
Lukas Tribus80512b12018-10-27 20:07:40 +02003758 to be routed to different backend servers, breaking the invalid assumptions
3759 NTLM relies on.
3760
Willy Tarreau0ba27502007-12-24 16:55:16 +01003761 Examples :
3762 balance roundrobin
3763 balance url_param userid
matt.farnsworth@nokia.com1c2ab962008-04-14 20:47:37 +02003764 balance url_param session_id check_post 64
Benoitaffb4812009-03-25 13:02:10 +01003765 balance hdr(User-Agent)
3766 balance hdr(host)
3767 balance hdr(Host) use_domain_only
matt.farnsworth@nokia.com1c2ab962008-04-14 20:47:37 +02003768
3769 Note: the following caveats and limitations on using the "check_post"
3770 extension with "url_param" must be considered :
3771
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01003772 - all POST requests are eligible for consideration, because there is no way
matt.farnsworth@nokia.com1c2ab962008-04-14 20:47:37 +02003773 to determine if the parameters will be found in the body or entity which
3774 may contain binary data. Therefore another method may be required to
3775 restrict consideration of POST requests that have no URL parameters in
Christopher Faulet87f1f3d2019-07-18 14:51:20 +02003776 the body. (see acl http_end)
matt.farnsworth@nokia.com1c2ab962008-04-14 20:47:37 +02003777
3778 - using a <max_wait> value larger than the request buffer size does not
3779 make sense and is useless. The buffer size is set at build time, and
3780 defaults to 16 kB.
3781
3782 - Content-Encoding is not supported, the parameter search will probably
3783 fail; and load balancing will fall back to Round Robin.
3784
3785 - Expect: 100-continue is not supported, load balancing will fall back to
3786 Round Robin.
3787
Lukas Tribus23953682017-04-28 13:24:30 +00003788 - Transfer-Encoding (RFC7230 3.3.1) is only supported in the first chunk.
matt.farnsworth@nokia.com1c2ab962008-04-14 20:47:37 +02003789 If the entire parameter value is not present in the first chunk, the
3790 selection of server is undefined (actually, defined by how little
3791 actually appeared in the first chunk).
3792
3793 - This feature does not support generation of a 100, 411 or 501 response.
3794
3795 - In some cases, requesting "check_post" MAY attempt to scan the entire
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01003796 contents of a message body. Scanning normally terminates when linear
matt.farnsworth@nokia.com1c2ab962008-04-14 20:47:37 +02003797 white space or control characters are found, indicating the end of what
3798 might be a URL parameter list. This is probably not a concern with SGML
3799 type message bodies.
Willy Tarreau0ba27502007-12-24 16:55:16 +01003800
Willy Tarreau294d0f02015-08-10 19:40:12 +02003801 See also : "dispatch", "cookie", "transparent", "hash-type" and "http_proxy".
Willy Tarreau0ba27502007-12-24 16:55:16 +01003802
3803
Willy Tarreaub6205fd2012-09-24 12:27:33 +02003804bind [<address>]:<port_range> [, ...] [param*]
3805bind /<path> [, ...] [param*]
Willy Tarreau0ba27502007-12-24 16:55:16 +01003806 Define one or several listening addresses and/or ports in a frontend.
3807 May be used in sections : defaults | frontend | listen | backend
3808 no | yes | yes | no
3809 Arguments :
Willy Tarreaub1e52e82008-01-13 14:49:51 +01003810 <address> is optional and can be a host name, an IPv4 address, an IPv6
3811 address, or '*'. It designates the address the frontend will
3812 listen on. If unset, all IPv4 addresses of the system will be
3813 listened on. The same will apply for '*' or the system's
David du Colombier9c938da2011-03-17 10:40:27 +01003814 special address "0.0.0.0". The IPv6 equivalent is '::'.
Willy Tarreau24709282013-03-10 21:32:12 +01003815 Optionally, an address family prefix may be used before the
3816 address to force the family regardless of the address format,
3817 which can be useful to specify a path to a unix socket with
3818 no slash ('/'). Currently supported prefixes are :
3819 - 'ipv4@' -> address is always IPv4
3820 - 'ipv6@' -> address is always IPv6
Emeric Brun3835c0d2020-07-07 09:46:09 +02003821 - 'udp@' -> address is resolved as IPv4 or IPv6 and
Emeric Brun12941c82020-07-07 14:19:42 +02003822 protocol UDP is used. Currently those listeners are
3823 supported only in log-forward sections.
Emeric Brun3835c0d2020-07-07 09:46:09 +02003824 - 'udp4@' -> address is always IPv4 and protocol UDP
Emeric Brun12941c82020-07-07 14:19:42 +02003825 is used. Currently those listeners are supported
3826 only in log-forward sections.
Emeric Brun3835c0d2020-07-07 09:46:09 +02003827 - 'udp6@' -> address is always IPv6 and protocol UDP
Emeric Brun12941c82020-07-07 14:19:42 +02003828 is used. Currently those listeners are supported
3829 only in log-forward sections.
Willy Tarreau24709282013-03-10 21:32:12 +01003830 - 'unix@' -> address is a path to a local unix socket
Willy Tarreau70f72e02014-07-08 00:37:50 +02003831 - 'abns@' -> address is in abstract namespace (Linux only).
3832 Note: since abstract sockets are not "rebindable", they
3833 do not cope well with multi-process mode during
3834 soft-restart, so it is better to avoid them if
3835 nbproc is greater than 1. The effect is that if the
3836 new process fails to start, only one of the old ones
3837 will be able to rebind to the socket.
Willy Tarreau40aa0702013-03-10 23:51:38 +01003838 - 'fd@<n>' -> use file descriptor <n> inherited from the
3839 parent. The fd must be bound and may or may not already
3840 be listening.
William Lallemand2fe7dd02018-09-11 16:51:29 +02003841 - 'sockpair@<n>'-> like fd@ but you must use the fd of a
3842 connected unix socket or of a socketpair. The bind waits
3843 to receive a FD over the unix socket and uses it as if it
3844 was the FD of an accept(). Should be used carefully.
William Lallemandb2f07452015-05-12 14:27:13 +02003845 You may want to reference some environment variables in the
3846 address parameter, see section 2.3 about environment
3847 variables.
Willy Tarreaub1e52e82008-01-13 14:49:51 +01003848
Willy Tarreauc5011ca2010-03-22 11:53:56 +01003849 <port_range> is either a unique TCP port, or a port range for which the
3850 proxy will accept connections for the IP address specified
Willy Tarreauceb24bc2010-11-09 12:46:41 +01003851 above. The port is mandatory for TCP listeners. Note that in
3852 the case of an IPv6 address, the port is always the number
3853 after the last colon (':'). A range can either be :
Willy Tarreauc5011ca2010-03-22 11:53:56 +01003854 - a numerical port (ex: '80')
3855 - a dash-delimited ports range explicitly stating the lower
3856 and upper bounds (ex: '2000-2100') which are included in
3857 the range.
3858
3859 Particular care must be taken against port ranges, because
3860 every <address:port> couple consumes one socket (= a file
3861 descriptor), so it's easy to consume lots of descriptors
3862 with a simple range, and to run out of sockets. Also, each
3863 <address:port> couple must be used only once among all
3864 instances running on a same system. Please note that binding
3865 to ports lower than 1024 generally require particular
Jamie Gloudon801a0a32012-08-25 00:18:33 -04003866 privileges to start the program, which are independent of
Willy Tarreauc5011ca2010-03-22 11:53:56 +01003867 the 'uid' parameter.
Willy Tarreau0ba27502007-12-24 16:55:16 +01003868
Willy Tarreauceb24bc2010-11-09 12:46:41 +01003869 <path> is a UNIX socket path beginning with a slash ('/'). This is
Davor Ocelice9ed2812017-12-25 17:49:28 +01003870 alternative to the TCP listening port. HAProxy will then
Willy Tarreauceb24bc2010-11-09 12:46:41 +01003871 receive UNIX connections on the socket located at this place.
3872 The path must begin with a slash and by default is absolute.
3873 It can be relative to the prefix defined by "unix-bind" in
3874 the global section. Note that the total length of the prefix
3875 followed by the socket path cannot exceed some system limits
3876 for UNIX sockets, which commonly are set to 107 characters.
3877
Willy Tarreaub6205fd2012-09-24 12:27:33 +02003878 <param*> is a list of parameters common to all sockets declared on the
3879 same line. These numerous parameters depend on OS and build
3880 options and have a complete section dedicated to them. Please
3881 refer to section 5 to for more details.
Willy Tarreaua0ee1d02012-09-10 09:01:23 +02003882
Willy Tarreau0ba27502007-12-24 16:55:16 +01003883 It is possible to specify a list of address:port combinations delimited by
3884 commas. The frontend will then listen on all of these addresses. There is no
3885 fixed limit to the number of addresses and ports which can be listened on in
3886 a frontend, as well as there is no limit to the number of "bind" statements
3887 in a frontend.
3888
3889 Example :
3890 listen http_proxy
3891 bind :80,:443
3892 bind 10.0.0.1:10080,10.0.0.1:10443
Willy Tarreauceb24bc2010-11-09 12:46:41 +01003893 bind /var/run/ssl-frontend.sock user root mode 600 accept-proxy
Willy Tarreau0ba27502007-12-24 16:55:16 +01003894
Willy Tarreaua0ee1d02012-09-10 09:01:23 +02003895 listen http_https_proxy
3896 bind :80
Cyril Bonté0d44fc62012-10-09 22:45:33 +02003897 bind :443 ssl crt /etc/haproxy/site.pem
Willy Tarreaua0ee1d02012-09-10 09:01:23 +02003898
Willy Tarreau24709282013-03-10 21:32:12 +01003899 listen http_https_proxy_explicit
3900 bind ipv6@:80
3901 bind ipv4@public_ssl:443 ssl crt /etc/haproxy/site.pem
3902 bind unix@ssl-frontend.sock user root mode 600 accept-proxy
3903
Willy Tarreaudad36a32013-03-11 01:20:04 +01003904 listen external_bind_app1
William Lallemandb2f07452015-05-12 14:27:13 +02003905 bind "fd@${FD_APP1}"
Willy Tarreaudad36a32013-03-11 01:20:04 +01003906
Willy Tarreau55dcaf62015-09-27 15:03:15 +02003907 Note: regarding Linux's abstract namespace sockets, HAProxy uses the whole
3908 sun_path length is used for the address length. Some other programs
3909 such as socat use the string length only by default. Pass the option
3910 ",unix-tightsocklen=0" to any abstract socket definition in socat to
3911 make it compatible with HAProxy's.
3912
Willy Tarreauceb24bc2010-11-09 12:46:41 +01003913 See also : "source", "option forwardfor", "unix-bind" and the PROXY protocol
Willy Tarreaub6205fd2012-09-24 12:27:33 +02003914 documentation, and section 5 about bind options.
Willy Tarreau0ba27502007-12-24 16:55:16 +01003915
3916
Christopher Fauletff4121f2017-11-22 16:38:49 +01003917bind-process [ all | odd | even | <process_num>[-[<process_num>]] ] ...
Willy Tarreau0b9c02c2009-02-04 22:05:05 +01003918 Limit visibility of an instance to a certain set of processes numbers.
3919 May be used in sections : defaults | frontend | listen | backend
3920 yes | yes | yes | yes
3921 Arguments :
3922 all All process will see this instance. This is the default. It
3923 may be used to override a default value.
3924
Willy Tarreaua9db57e2013-01-18 11:29:29 +01003925 odd This instance will be enabled on processes 1,3,5,...63. This
Willy Tarreau0b9c02c2009-02-04 22:05:05 +01003926 option may be combined with other numbers.
3927
Willy Tarreaua9db57e2013-01-18 11:29:29 +01003928 even This instance will be enabled on processes 2,4,6,...64. This
Willy Tarreau0b9c02c2009-02-04 22:05:05 +01003929 option may be combined with other numbers. Do not use it
3930 with less than 2 processes otherwise some instances might be
3931 missing from all processes.
3932
Christopher Fauletff4121f2017-11-22 16:38:49 +01003933 process_num The instance will be enabled on this process number or range,
Willy Tarreaua9db57e2013-01-18 11:29:29 +01003934 whose values must all be between 1 and 32 or 64 depending on
Christopher Fauletff4121f2017-11-22 16:38:49 +01003935 the machine's word size. Ranges can be partially defined. The
3936 higher bound can be omitted. In such case, it is replaced by
3937 the corresponding maximum value. If a proxy is bound to
3938 process numbers greater than the configured global.nbproc, it
3939 will either be forced to process #1 if a single process was
Willy Tarreau102df612014-05-07 23:56:38 +02003940 specified, or to all processes otherwise.
Willy Tarreau0b9c02c2009-02-04 22:05:05 +01003941
3942 This keyword limits binding of certain instances to certain processes. This
3943 is useful in order not to have too many processes listening to the same
3944 ports. For instance, on a dual-core machine, it might make sense to set
3945 'nbproc 2' in the global section, then distributes the listeners among 'odd'
3946 and 'even' instances.
3947
Willy Tarreaua9db57e2013-01-18 11:29:29 +01003948 At the moment, it is not possible to reference more than 32 or 64 processes
3949 using this keyword, but this should be more than enough for most setups.
3950 Please note that 'all' really means all processes regardless of the machine's
3951 word size, and is not limited to the first 32 or 64.
Willy Tarreau0b9c02c2009-02-04 22:05:05 +01003952
Willy Tarreau6ae1ba62014-05-07 19:01:58 +02003953 Each "bind" line may further be limited to a subset of the proxy's processes,
3954 please consult the "process" bind keyword in section 5.1.
3955
Willy Tarreaub369a042014-09-16 13:21:03 +02003956 When a frontend has no explicit "bind-process" line, it tries to bind to all
3957 the processes referenced by its "bind" lines. That means that frontends can
3958 easily adapt to their listeners' processes.
3959
Willy Tarreau0b9c02c2009-02-04 22:05:05 +01003960 If some backends are referenced by frontends bound to other processes, the
3961 backend automatically inherits the frontend's processes.
3962
3963 Example :
3964 listen app_ip1
3965 bind 10.0.0.1:80
Willy Tarreaubfcd3112010-10-23 11:22:08 +02003966 bind-process odd
Willy Tarreau0b9c02c2009-02-04 22:05:05 +01003967
3968 listen app_ip2
3969 bind 10.0.0.2:80
Willy Tarreaubfcd3112010-10-23 11:22:08 +02003970 bind-process even
Willy Tarreau0b9c02c2009-02-04 22:05:05 +01003971
3972 listen management
3973 bind 10.0.0.3:80
Willy Tarreaubfcd3112010-10-23 11:22:08 +02003974 bind-process 1 2 3 4
Willy Tarreau0b9c02c2009-02-04 22:05:05 +01003975
Willy Tarreau110ecc12012-11-15 17:50:01 +01003976 listen management
3977 bind 10.0.0.4:80
3978 bind-process 1-4
3979
Willy Tarreau6ae1ba62014-05-07 19:01:58 +02003980 See also : "nbproc" in global section, and "process" in section 5.1.
Willy Tarreau0b9c02c2009-02-04 22:05:05 +01003981
3982
Willy Tarreau0ba27502007-12-24 16:55:16 +01003983capture cookie <name> len <length>
3984 Capture and log a cookie in the request and in the response.
3985 May be used in sections : defaults | frontend | listen | backend
3986 no | yes | yes | no
3987 Arguments :
3988 <name> is the beginning of the name of the cookie to capture. In order
3989 to match the exact name, simply suffix the name with an equal
3990 sign ('='). The full name will appear in the logs, which is
3991 useful with application servers which adjust both the cookie name
Davor Ocelice9ed2812017-12-25 17:49:28 +01003992 and value (e.g. ASPSESSIONXXX).
Willy Tarreau0ba27502007-12-24 16:55:16 +01003993
3994 <length> is the maximum number of characters to report in the logs, which
3995 include the cookie name, the equal sign and the value, all in the
3996 standard "name=value" form. The string will be truncated on the
3997 right if it exceeds <length>.
3998
3999 Only the first cookie is captured. Both the "cookie" request headers and the
4000 "set-cookie" response headers are monitored. This is particularly useful to
4001 check for application bugs causing session crossing or stealing between
4002 users, because generally the user's cookies can only change on a login page.
4003
4004 When the cookie was not presented by the client, the associated log column
4005 will report "-". When a request does not cause a cookie to be assigned by the
4006 server, a "-" is reported in the response column.
4007
4008 The capture is performed in the frontend only because it is necessary that
4009 the log format does not change for a given frontend depending on the
4010 backends. This may change in the future. Note that there can be only one
Willy Tarreau193b8c62012-11-22 00:17:38 +01004011 "capture cookie" statement in a frontend. The maximum capture length is set
4012 by the global "tune.http.cookielen" setting and defaults to 63 characters. It
4013 is not possible to specify a capture in a "defaults" section.
Willy Tarreau0ba27502007-12-24 16:55:16 +01004014
4015 Example:
4016 capture cookie ASPSESSION len 32
4017
4018 See also : "capture request header", "capture response header" as well as
Willy Tarreauc57f0e22009-05-10 13:12:33 +02004019 section 8 about logging.
Willy Tarreau0ba27502007-12-24 16:55:16 +01004020
4021
4022capture request header <name> len <length>
Willy Tarreau4460d032012-11-21 23:37:37 +01004023 Capture and log the last occurrence of the specified request header.
Willy Tarreau0ba27502007-12-24 16:55:16 +01004024 May be used in sections : defaults | frontend | listen | backend
4025 no | yes | yes | no
4026 Arguments :
4027 <name> is the name of the header to capture. The header names are not
Willy Tarreaud2a4aa22008-01-31 15:28:22 +01004028 case-sensitive, but it is a common practice to write them as they
Willy Tarreau0ba27502007-12-24 16:55:16 +01004029 appear in the requests, with the first letter of each word in
4030 upper case. The header name will not appear in the logs, only the
4031 value is reported, but the position in the logs is respected.
4032
4033 <length> is the maximum number of characters to extract from the value and
4034 report in the logs. The string will be truncated on the right if
4035 it exceeds <length>.
4036
Willy Tarreau4460d032012-11-21 23:37:37 +01004037 The complete value of the last occurrence of the header is captured. The
Willy Tarreau0ba27502007-12-24 16:55:16 +01004038 value will be added to the logs between braces ('{}'). If multiple headers
4039 are captured, they will be delimited by a vertical bar ('|') and will appear
Willy Tarreaucc6c8912009-02-22 10:53:55 +01004040 in the same order they were declared in the configuration. Non-existent
4041 headers will be logged just as an empty string. Common uses for request
4042 header captures include the "Host" field in virtual hosting environments, the
4043 "Content-length" when uploads are supported, "User-agent" to quickly
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01004044 differentiate between real users and robots, and "X-Forwarded-For" in proxied
Willy Tarreaucc6c8912009-02-22 10:53:55 +01004045 environments to find where the request came from.
4046
4047 Note that when capturing headers such as "User-agent", some spaces may be
4048 logged, making the log analysis more difficult. Thus be careful about what
4049 you log if you know your log parser is not smart enough to rely on the
4050 braces.
Willy Tarreau0ba27502007-12-24 16:55:16 +01004051
Willy Tarreau0900abb2012-11-22 00:21:46 +01004052 There is no limit to the number of captured request headers nor to their
4053 length, though it is wise to keep them low to limit memory usage per session.
4054 In order to keep log format consistent for a same frontend, header captures
4055 can only be declared in a frontend. It is not possible to specify a capture
4056 in a "defaults" section.
Willy Tarreau0ba27502007-12-24 16:55:16 +01004057
4058 Example:
4059 capture request header Host len 15
4060 capture request header X-Forwarded-For len 15
Cyril Bontéd1b0f7c2015-10-26 22:37:39 +01004061 capture request header Referer len 15
Willy Tarreau0ba27502007-12-24 16:55:16 +01004062
Willy Tarreauc57f0e22009-05-10 13:12:33 +02004063 See also : "capture cookie", "capture response header" as well as section 8
Willy Tarreau0ba27502007-12-24 16:55:16 +01004064 about logging.
4065
4066
4067capture response header <name> len <length>
Willy Tarreau4460d032012-11-21 23:37:37 +01004068 Capture and log the last occurrence of the specified response header.
Willy Tarreau0ba27502007-12-24 16:55:16 +01004069 May be used in sections : defaults | frontend | listen | backend
4070 no | yes | yes | no
4071 Arguments :
4072 <name> is the name of the header to capture. The header names are not
Willy Tarreaud2a4aa22008-01-31 15:28:22 +01004073 case-sensitive, but it is a common practice to write them as they
Willy Tarreau0ba27502007-12-24 16:55:16 +01004074 appear in the response, with the first letter of each word in
4075 upper case. The header name will not appear in the logs, only the
4076 value is reported, but the position in the logs is respected.
4077
4078 <length> is the maximum number of characters to extract from the value and
4079 report in the logs. The string will be truncated on the right if
4080 it exceeds <length>.
4081
Willy Tarreau4460d032012-11-21 23:37:37 +01004082 The complete value of the last occurrence of the header is captured. The
Willy Tarreau0ba27502007-12-24 16:55:16 +01004083 result will be added to the logs between braces ('{}') after the captured
4084 request headers. If multiple headers are captured, they will be delimited by
4085 a vertical bar ('|') and will appear in the same order they were declared in
Willy Tarreaucc6c8912009-02-22 10:53:55 +01004086 the configuration. Non-existent headers will be logged just as an empty
4087 string. Common uses for response header captures include the "Content-length"
4088 header which indicates how many bytes are expected to be returned, the
4089 "Location" header to track redirections.
Willy Tarreau0ba27502007-12-24 16:55:16 +01004090
Willy Tarreau0900abb2012-11-22 00:21:46 +01004091 There is no limit to the number of captured response headers nor to their
4092 length, though it is wise to keep them low to limit memory usage per session.
4093 In order to keep log format consistent for a same frontend, header captures
4094 can only be declared in a frontend. It is not possible to specify a capture
4095 in a "defaults" section.
Willy Tarreau0ba27502007-12-24 16:55:16 +01004096
4097 Example:
4098 capture response header Content-length len 9
4099 capture response header Location len 15
4100
Willy Tarreauc57f0e22009-05-10 13:12:33 +02004101 See also : "capture cookie", "capture request header" as well as section 8
Willy Tarreau0ba27502007-12-24 16:55:16 +01004102 about logging.
4103
4104
MIZUTA Takeshib24bc0d2020-07-09 11:13:20 +09004105clitcpka-cnt <count>
4106 Sets the maximum number of keepalive probes TCP should send before dropping
4107 the connection on the client side.
4108 May be used in sections : defaults | frontend | listen | backend
4109 yes | yes | yes | no
4110 Arguments :
4111 <count> is the maximum number of keepalive probes.
4112
4113 This keyword corresponds to the socket option TCP_KEEPCNT. If this keyword
4114 is not specified, system-wide TCP parameter (tcp_keepalive_probes) is used.
Willy Tarreau52543212020-07-09 05:58:51 +02004115 The availability of this setting depends on the operating system. It is
4116 known to work on Linux.
MIZUTA Takeshib24bc0d2020-07-09 11:13:20 +09004117
4118 See also : "option clitcpka", "clitcpka-idle", "clitcpka-intvl".
4119
4120
4121clitcpka-idle <timeout>
4122 Sets the time the connection needs to remain idle before TCP starts sending
4123 keepalive probes, if enabled the sending of TCP keepalive packets on the
4124 client side.
4125 May be used in sections : defaults | frontend | listen | backend
4126 yes | yes | yes | no
4127 Arguments :
4128 <timeout> is the time the connection needs to remain idle before TCP starts
4129 sending keepalive probes. It is specified in seconds by default,
4130 but can be in any other unit if the number is suffixed by the
4131 unit, as explained at the top of this document.
4132
4133 This keyword corresponds to the socket option TCP_KEEPIDLE. If this keyword
4134 is not specified, system-wide TCP parameter (tcp_keepalive_time) is used.
Willy Tarreau52543212020-07-09 05:58:51 +02004135 The availability of this setting depends on the operating system. It is
4136 known to work on Linux.
MIZUTA Takeshib24bc0d2020-07-09 11:13:20 +09004137
4138 See also : "option clitcpka", "clitcpka-cnt", "clitcpka-intvl".
4139
4140
4141clitcpka-intvl <timeout>
4142 Sets the time between individual keepalive probes on the client side.
4143 May be used in sections : defaults | frontend | listen | backend
4144 yes | yes | yes | no
4145 Arguments :
4146 <timeout> is the time between individual keepalive probes. It is specified
4147 in seconds by default, but can be in any other unit if the number
4148 is suffixed by the unit, as explained at the top of this
4149 document.
4150
4151 This keyword corresponds to the socket option TCP_KEEPINTVL. If this keyword
4152 is not specified, system-wide TCP parameter (tcp_keepalive_intvl) is used.
Willy Tarreau52543212020-07-09 05:58:51 +02004153 The availability of this setting depends on the operating system. It is
4154 known to work on Linux.
MIZUTA Takeshib24bc0d2020-07-09 11:13:20 +09004155
4156 See also : "option clitcpka", "clitcpka-cnt", "clitcpka-idle".
4157
4158
Cyril Bonté316a8cf2012-11-11 13:38:27 +01004159compression algo <algorithm> ...
4160compression type <mime type> ...
Willy Tarreau70737d12012-10-27 00:34:28 +02004161compression offload
William Lallemand82fe75c2012-10-23 10:25:10 +02004162 Enable HTTP compression.
4163 May be used in sections : defaults | frontend | listen | backend
4164 yes | yes | yes | yes
4165 Arguments :
Cyril Bonté316a8cf2012-11-11 13:38:27 +01004166 algo is followed by the list of supported compression algorithms.
4167 type is followed by the list of MIME types that will be compressed.
4168 offload makes haproxy work as a compression offloader only (see notes).
4169
4170 The currently supported algorithms are :
Willy Tarreauc91840a2015-03-28 17:00:39 +01004171 identity this is mostly for debugging, and it was useful for developing
4172 the compression feature. Identity does not apply any change on
4173 data.
Cyril Bonté316a8cf2012-11-11 13:38:27 +01004174
Willy Tarreauc91840a2015-03-28 17:00:39 +01004175 gzip applies gzip compression. This setting is only available when
Baptiste Assmannf085d632015-12-21 17:57:32 +01004176 support for zlib or libslz was built in.
Willy Tarreauc91840a2015-03-28 17:00:39 +01004177
4178 deflate same as "gzip", but with deflate algorithm and zlib format.
4179 Note that this algorithm has ambiguous support on many
4180 browsers and no support at all from recent ones. It is
4181 strongly recommended not to use it for anything else than
4182 experimentation. This setting is only available when support
Baptiste Assmannf085d632015-12-21 17:57:32 +01004183 for zlib or libslz was built in.
Cyril Bonté316a8cf2012-11-11 13:38:27 +01004184
Willy Tarreauc91840a2015-03-28 17:00:39 +01004185 raw-deflate same as "deflate" without the zlib wrapper, and used as an
4186 alternative when the browser wants "deflate". All major
4187 browsers understand it and despite violating the standards,
4188 it is known to work better than "deflate", at least on MSIE
4189 and some versions of Safari. Do not use it in conjunction
4190 with "deflate", use either one or the other since both react
4191 to the same Accept-Encoding token. This setting is only
Baptiste Assmannf085d632015-12-21 17:57:32 +01004192 available when support for zlib or libslz was built in.
Cyril Bonté316a8cf2012-11-11 13:38:27 +01004193
Dmitry Sivachenko87c208b2012-11-22 20:03:26 +04004194 Compression will be activated depending on the Accept-Encoding request
Cyril Bonté316a8cf2012-11-11 13:38:27 +01004195 header. With identity, it does not take care of that header.
Dmitry Sivachenkoc9f3b452012-11-28 17:47:11 +04004196 If backend servers support HTTP compression, these directives
4197 will be no-op: haproxy will see the compressed response and will not
4198 compress again. If backend servers do not support HTTP compression and
4199 there is Accept-Encoding header in request, haproxy will compress the
4200 matching response.
Willy Tarreau70737d12012-10-27 00:34:28 +02004201
4202 The "offload" setting makes haproxy remove the Accept-Encoding header to
4203 prevent backend servers from compressing responses. It is strongly
4204 recommended not to do this because this means that all the compression work
4205 will be done on the single point where haproxy is located. However in some
4206 deployment scenarios, haproxy may be installed in front of a buggy gateway
Dmitry Sivachenkoc9f3b452012-11-28 17:47:11 +04004207 with broken HTTP compression implementation which can't be turned off.
4208 In that case haproxy can be used to prevent that gateway from emitting
4209 invalid payloads. In this case, simply removing the header in the
4210 configuration does not work because it applies before the header is parsed,
4211 so that prevents haproxy from compressing. The "offload" setting should
Willy Tarreauffea9fd2014-07-12 16:37:02 +02004212 then be used for such scenarios. Note: for now, the "offload" setting is
4213 ignored when set in a defaults section.
William Lallemand82fe75c2012-10-23 10:25:10 +02004214
William Lallemand05097442012-11-20 12:14:28 +01004215 Compression is disabled when:
Baptiste Assmann650d53d2013-01-05 15:44:44 +01004216 * the request does not advertise a supported compression algorithm in the
4217 "Accept-Encoding" header
4218 * the response message is not HTTP/1.1
Tim Duesterhusbb48c9a2019-01-30 23:46:04 +01004219 * HTTP status code is not one of 200, 201, 202, or 203
Baptiste Assmann650d53d2013-01-05 15:44:44 +01004220 * response contain neither a "Content-Length" header nor a
4221 "Transfer-Encoding" whose last value is "chunked"
4222 * response contains a "Content-Type" header whose first value starts with
4223 "multipart"
4224 * the response contains the "no-transform" value in the "Cache-control"
4225 header
4226 * User-Agent matches "Mozilla/4" unless it is MSIE 6 with XP SP2, or MSIE 7
4227 and later
4228 * The response contains a "Content-Encoding" header, indicating that the
4229 response is already compressed (see compression offload)
Tim Duesterhusbb48c9a2019-01-30 23:46:04 +01004230 * The response contains an invalid "ETag" header or multiple ETag headers
William Lallemand05097442012-11-20 12:14:28 +01004231
Tim Duesterhusb229f012019-01-29 16:38:56 +01004232 Note: The compression does not emit the Warning header.
William Lallemand05097442012-11-20 12:14:28 +01004233
William Lallemand82fe75c2012-10-23 10:25:10 +02004234 Examples :
4235 compression algo gzip
4236 compression type text/html text/plain
Willy Tarreau0ba27502007-12-24 16:55:16 +01004237
Christopher Fauletc3fe5332016-04-07 15:30:10 +02004238
Willy Tarreau55165fe2009-05-10 12:02:55 +02004239cookie <name> [ rewrite | insert | prefix ] [ indirect ] [ nocache ]
Willy Tarreau4992dd22012-05-31 21:02:17 +02004240 [ postonly ] [ preserve ] [ httponly ] [ secure ]
4241 [ domain <domain> ]* [ maxidle <idle> ] [ maxlife <life> ]
Christopher Faulet2f533902020-01-21 11:06:48 +01004242 [ dynamic ] [ attr <value> ]*
Willy Tarreau0ba27502007-12-24 16:55:16 +01004243 Enable cookie-based persistence in a backend.
4244 May be used in sections : defaults | frontend | listen | backend
4245 yes | no | yes | yes
4246 Arguments :
4247 <name> is the name of the cookie which will be monitored, modified or
4248 inserted in order to bring persistence. This cookie is sent to
4249 the client via a "Set-Cookie" header in the response, and is
4250 brought back by the client in a "Cookie" header in all requests.
4251 Special care should be taken to choose a name which does not
4252 conflict with any likely application cookie. Also, if the same
Davor Ocelice9ed2812017-12-25 17:49:28 +01004253 backends are subject to be used by the same clients (e.g.
Willy Tarreau0ba27502007-12-24 16:55:16 +01004254 HTTP/HTTPS), care should be taken to use different cookie names
4255 between all backends if persistence between them is not desired.
4256
4257 rewrite This keyword indicates that the cookie will be provided by the
4258 server and that haproxy will have to modify its value to set the
4259 server's identifier in it. This mode is handy when the management
4260 of complex combinations of "Set-cookie" and "Cache-control"
4261 headers is left to the application. The application can then
4262 decide whether or not it is appropriate to emit a persistence
Lukas Tribusf01a9cd2016-02-03 18:09:37 +01004263 cookie. Since all responses should be monitored, this mode
4264 doesn't work in HTTP tunnel mode. Unless the application
Davor Ocelice9ed2812017-12-25 17:49:28 +01004265 behavior is very complex and/or broken, it is advised not to
Lukas Tribusf01a9cd2016-02-03 18:09:37 +01004266 start with this mode for new deployments. This keyword is
4267 incompatible with "insert" and "prefix".
Willy Tarreau0ba27502007-12-24 16:55:16 +01004268
4269 insert This keyword indicates that the persistence cookie will have to
Willy Tarreaua79094d2010-08-31 22:54:15 +02004270 be inserted by haproxy in server responses if the client did not
Willy Tarreauba4c5be2010-10-23 12:46:42 +02004271
Willy Tarreaua79094d2010-08-31 22:54:15 +02004272 already have a cookie that would have permitted it to access this
Willy Tarreauba4c5be2010-10-23 12:46:42 +02004273 server. When used without the "preserve" option, if the server
Michael Prokop4438c602019-05-24 10:25:45 +02004274 emits a cookie with the same name, it will be removed before
Davor Ocelice9ed2812017-12-25 17:49:28 +01004275 processing. For this reason, this mode can be used to upgrade
Willy Tarreauba4c5be2010-10-23 12:46:42 +02004276 existing configurations running in the "rewrite" mode. The cookie
4277 will only be a session cookie and will not be stored on the
4278 client's disk. By default, unless the "indirect" option is added,
4279 the server will see the cookies emitted by the client. Due to
4280 caching effects, it is generally wise to add the "nocache" or
4281 "postonly" keywords (see below). The "insert" keyword is not
4282 compatible with "rewrite" and "prefix".
Willy Tarreau0ba27502007-12-24 16:55:16 +01004283
4284 prefix This keyword indicates that instead of relying on a dedicated
4285 cookie for the persistence, an existing one will be completed.
4286 This may be needed in some specific environments where the client
4287 does not support more than one single cookie and the application
4288 already needs it. In this case, whenever the server sets a cookie
4289 named <name>, it will be prefixed with the server's identifier
4290 and a delimiter. The prefix will be removed from all client
4291 requests so that the server still finds the cookie it emitted.
4292 Since all requests and responses are subject to being modified,
Lukas Tribusf01a9cd2016-02-03 18:09:37 +01004293 this mode doesn't work with tunnel mode. The "prefix" keyword is
Willy Tarreau37229df2011-10-17 12:24:55 +02004294 not compatible with "rewrite" and "insert". Note: it is highly
4295 recommended not to use "indirect" with "prefix", otherwise server
4296 cookie updates would not be sent to clients.
Willy Tarreau0ba27502007-12-24 16:55:16 +01004297
Willy Tarreaua79094d2010-08-31 22:54:15 +02004298 indirect When this option is specified, no cookie will be emitted to a
4299 client which already has a valid one for the server which has
4300 processed the request. If the server sets such a cookie itself,
Willy Tarreauba4c5be2010-10-23 12:46:42 +02004301 it will be removed, unless the "preserve" option is also set. In
4302 "insert" mode, this will additionally remove cookies from the
4303 requests transmitted to the server, making the persistence
4304 mechanism totally transparent from an application point of view.
Willy Tarreau37229df2011-10-17 12:24:55 +02004305 Note: it is highly recommended not to use "indirect" with
4306 "prefix", otherwise server cookie updates would not be sent to
4307 clients.
Willy Tarreau0ba27502007-12-24 16:55:16 +01004308
4309 nocache This option is recommended in conjunction with the insert mode
4310 when there is a cache between the client and HAProxy, as it
4311 ensures that a cacheable response will be tagged non-cacheable if
4312 a cookie needs to be inserted. This is important because if all
4313 persistence cookies are added on a cacheable home page for
4314 instance, then all customers will then fetch the page from an
4315 outer cache and will all share the same persistence cookie,
4316 leading to one server receiving much more traffic than others.
4317 See also the "insert" and "postonly" options.
4318
4319 postonly This option ensures that cookie insertion will only be performed
4320 on responses to POST requests. It is an alternative to the
4321 "nocache" option, because POST responses are not cacheable, so
4322 this ensures that the persistence cookie will never get cached.
4323 Since most sites do not need any sort of persistence before the
4324 first POST which generally is a login request, this is a very
4325 efficient method to optimize caching without risking to find a
4326 persistence cookie in the cache.
4327 See also the "insert" and "nocache" options.
4328
Willy Tarreauba4c5be2010-10-23 12:46:42 +02004329 preserve This option may only be used with "insert" and/or "indirect". It
4330 allows the server to emit the persistence cookie itself. In this
4331 case, if a cookie is found in the response, haproxy will leave it
4332 untouched. This is useful in order to end persistence after a
4333 logout request for instance. For this, the server just has to
Davor Ocelice9ed2812017-12-25 17:49:28 +01004334 emit a cookie with an invalid value (e.g. empty) or with a date in
Willy Tarreauba4c5be2010-10-23 12:46:42 +02004335 the past. By combining this mechanism with the "disable-on-404"
4336 check option, it is possible to perform a completely graceful
4337 shutdown because users will definitely leave the server after
4338 they logout.
4339
Willy Tarreau4992dd22012-05-31 21:02:17 +02004340 httponly This option tells haproxy to add an "HttpOnly" cookie attribute
4341 when a cookie is inserted. This attribute is used so that a
4342 user agent doesn't share the cookie with non-HTTP components.
4343 Please check RFC6265 for more information on this attribute.
4344
4345 secure This option tells haproxy to add a "Secure" cookie attribute when
4346 a cookie is inserted. This attribute is used so that a user agent
4347 never emits this cookie over non-secure channels, which means
4348 that a cookie learned with this flag will be presented only over
4349 SSL/TLS connections. Please check RFC6265 for more information on
4350 this attribute.
4351
Krzysztof Piotr Oledzkiefe3b6f2008-05-23 23:49:32 +02004352 domain This option allows to specify the domain at which a cookie is
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01004353 inserted. It requires exactly one parameter: a valid domain
Willy Tarreau68a897b2009-12-03 23:28:34 +01004354 name. If the domain begins with a dot, the browser is allowed to
4355 use it for any host ending with that name. It is also possible to
4356 specify several domain names by invoking this option multiple
4357 times. Some browsers might have small limits on the number of
4358 domains, so be careful when doing that. For the record, sending
4359 10 domains to MSIE 6 or Firefox 2 works as expected.
Krzysztof Piotr Oledzkiefe3b6f2008-05-23 23:49:32 +02004360
Willy Tarreau996a92c2010-10-13 19:30:47 +02004361 maxidle This option allows inserted cookies to be ignored after some idle
4362 time. It only works with insert-mode cookies. When a cookie is
4363 sent to the client, the date this cookie was emitted is sent too.
4364 Upon further presentations of this cookie, if the date is older
4365 than the delay indicated by the parameter (in seconds), it will
4366 be ignored. Otherwise, it will be refreshed if needed when the
4367 response is sent to the client. This is particularly useful to
4368 prevent users who never close their browsers from remaining for
Davor Ocelice9ed2812017-12-25 17:49:28 +01004369 too long on the same server (e.g. after a farm size change). When
Willy Tarreau996a92c2010-10-13 19:30:47 +02004370 this option is set and a cookie has no date, it is always
4371 accepted, but gets refreshed in the response. This maintains the
4372 ability for admins to access their sites. Cookies that have a
4373 date in the future further than 24 hours are ignored. Doing so
4374 lets admins fix timezone issues without risking kicking users off
4375 the site.
4376
4377 maxlife This option allows inserted cookies to be ignored after some life
4378 time, whether they're in use or not. It only works with insert
4379 mode cookies. When a cookie is first sent to the client, the date
4380 this cookie was emitted is sent too. Upon further presentations
4381 of this cookie, if the date is older than the delay indicated by
4382 the parameter (in seconds), it will be ignored. If the cookie in
4383 the request has no date, it is accepted and a date will be set.
4384 Cookies that have a date in the future further than 24 hours are
4385 ignored. Doing so lets admins fix timezone issues without risking
4386 kicking users off the site. Contrary to maxidle, this value is
4387 not refreshed, only the first visit date counts. Both maxidle and
4388 maxlife may be used at the time. This is particularly useful to
4389 prevent users who never close their browsers from remaining for
Davor Ocelice9ed2812017-12-25 17:49:28 +01004390 too long on the same server (e.g. after a farm size change). This
Willy Tarreau996a92c2010-10-13 19:30:47 +02004391 is stronger than the maxidle method in that it forces a
4392 redispatch after some absolute delay.
4393
Olivier Houchard4e694042017-03-14 20:01:29 +01004394 dynamic Activate dynamic cookies. When used, a session cookie is
4395 dynamically created for each server, based on the IP and port
4396 of the server, and a secret key, specified in the
4397 "dynamic-cookie-key" backend directive.
4398 The cookie will be regenerated each time the IP address change,
4399 and is only generated for IPv4/IPv6.
4400
Christopher Faulet2f533902020-01-21 11:06:48 +01004401 attr This option tells haproxy to add an extra attribute when a
4402 cookie is inserted. The attribute value can contain any
4403 characters except control ones or ";". This option may be
4404 repeated.
4405
Willy Tarreau0ba27502007-12-24 16:55:16 +01004406 There can be only one persistence cookie per HTTP backend, and it can be
4407 declared in a defaults section. The value of the cookie will be the value
4408 indicated after the "cookie" keyword in a "server" statement. If no cookie
4409 is declared for a given server, the cookie is not set.
Willy Tarreau6a06a402007-07-15 20:15:28 +02004410
Willy Tarreau0ba27502007-12-24 16:55:16 +01004411 Examples :
4412 cookie JSESSIONID prefix
4413 cookie SRV insert indirect nocache
4414 cookie SRV insert postonly indirect
Willy Tarreau996a92c2010-10-13 19:30:47 +02004415 cookie SRV insert indirect nocache maxidle 30m maxlife 8h
Willy Tarreau0ba27502007-12-24 16:55:16 +01004416
Willy Tarreau294d0f02015-08-10 19:40:12 +02004417 See also : "balance source", "capture cookie", "server" and "ignore-persist".
Willy Tarreau0ba27502007-12-24 16:55:16 +01004418
Willy Tarreau983e01e2010-01-11 18:42:06 +01004419
Thierry FOURNIERa0a1b752015-05-26 17:44:32 +02004420declare capture [ request | response ] len <length>
4421 Declares a capture slot.
4422 May be used in sections : defaults | frontend | listen | backend
4423 no | yes | yes | no
4424 Arguments:
4425 <length> is the length allowed for the capture.
4426
4427 This declaration is only available in the frontend or listen section, but the
4428 reserved slot can be used in the backends. The "request" keyword allocates a
4429 capture slot for use in the request, and "response" allocates a capture slot
4430 for use in the response.
4431
4432 See also: "capture-req", "capture-res" (sample converters),
Baptiste Assmann5ac425c2015-10-21 23:13:46 +02004433 "capture.req.hdr", "capture.res.hdr" (sample fetches),
Thierry FOURNIERa0a1b752015-05-26 17:44:32 +02004434 "http-request capture" and "http-response capture".
4435
4436
Krzysztof Piotr Oledzkic6df0662010-01-05 16:38:49 +01004437default-server [param*]
4438 Change default options for a server in a backend
4439 May be used in sections : defaults | frontend | listen | backend
4440 yes | no | yes | yes
4441 Arguments:
Willy Tarreau983e01e2010-01-11 18:42:06 +01004442 <param*> is a list of parameters for this server. The "default-server"
4443 keyword accepts an important number of options and has a complete
4444 section dedicated to it. Please refer to section 5 for more
4445 details.
Krzysztof Piotr Oledzkic6df0662010-01-05 16:38:49 +01004446
Willy Tarreau983e01e2010-01-11 18:42:06 +01004447 Example :
Krzysztof Piotr Oledzkic6df0662010-01-05 16:38:49 +01004448 default-server inter 1000 weight 13
4449
4450 See also: "server" and section 5 about server options
Willy Tarreau0ba27502007-12-24 16:55:16 +01004451
Willy Tarreau983e01e2010-01-11 18:42:06 +01004452
Willy Tarreau0ba27502007-12-24 16:55:16 +01004453default_backend <backend>
4454 Specify the backend to use when no "use_backend" rule has been matched.
4455 May be used in sections : defaults | frontend | listen | backend
4456 yes | yes | yes | no
4457 Arguments :
4458 <backend> is the name of the backend to use.
4459
4460 When doing content-switching between frontend and backends using the
4461 "use_backend" keyword, it is often useful to indicate which backend will be
4462 used when no rule has matched. It generally is the dynamic backend which
4463 will catch all undetermined requests.
4464
Willy Tarreau0ba27502007-12-24 16:55:16 +01004465 Example :
4466
4467 use_backend dynamic if url_dyn
4468 use_backend static if url_css url_img extension_img
4469 default_backend dynamic
4470
Willy Tarreau98d04852015-05-26 12:18:29 +02004471 See also : "use_backend"
Willy Tarreau2769aa02007-12-27 18:26:09 +01004472
Willy Tarreau0ba27502007-12-24 16:55:16 +01004473
Baptiste Assmann27f51342013-10-09 06:51:49 +02004474description <string>
4475 Describe a listen, frontend or backend.
4476 May be used in sections : defaults | frontend | listen | backend
4477 no | yes | yes | yes
4478 Arguments : string
4479
4480 Allows to add a sentence to describe the related object in the HAProxy HTML
4481 stats page. The description will be printed on the right of the object name
4482 it describes.
4483 No need to backslash spaces in the <string> arguments.
4484
4485
Willy Tarreau0ba27502007-12-24 16:55:16 +01004486disabled
4487 Disable a proxy, frontend or backend.
4488 May be used in sections : defaults | frontend | listen | backend
4489 yes | yes | yes | yes
4490 Arguments : none
4491
4492 The "disabled" keyword is used to disable an instance, mainly in order to
4493 liberate a listening port or to temporarily disable a service. The instance
4494 will still be created and its configuration will be checked, but it will be
4495 created in the "stopped" state and will appear as such in the statistics. It
4496 will not receive any traffic nor will it send any health-checks or logs. It
4497 is possible to disable many instances at once by adding the "disabled"
4498 keyword in a "defaults" section.
4499
4500 See also : "enabled"
4501
4502
Willy Tarreau5ce94572010-06-07 14:35:41 +02004503dispatch <address>:<port>
4504 Set a default server address
4505 May be used in sections : defaults | frontend | listen | backend
4506 no | no | yes | yes
Cyril Bonté108cf6e2012-04-21 23:30:29 +02004507 Arguments :
Willy Tarreau5ce94572010-06-07 14:35:41 +02004508
4509 <address> is the IPv4 address of the default server. Alternatively, a
4510 resolvable hostname is supported, but this name will be resolved
4511 during start-up.
4512
4513 <ports> is a mandatory port specification. All connections will be sent
4514 to this port, and it is not permitted to use port offsets as is
4515 possible with normal servers.
4516
Willy Tarreau787aed52011-04-15 06:45:37 +02004517 The "dispatch" keyword designates a default server for use when no other
Willy Tarreau5ce94572010-06-07 14:35:41 +02004518 server can take the connection. In the past it was used to forward non
4519 persistent connections to an auxiliary load balancer. Due to its simple
4520 syntax, it has also been used for simple TCP relays. It is recommended not to
4521 use it for more clarity, and to use the "server" directive instead.
4522
4523 See also : "server"
4524
Olivier Houchard4e694042017-03-14 20:01:29 +01004525
4526dynamic-cookie-key <string>
4527 Set the dynamic cookie secret key for a backend.
4528 May be used in sections : defaults | frontend | listen | backend
4529 yes | no | yes | yes
4530 Arguments : The secret key to be used.
4531
4532 When dynamic cookies are enabled (see the "dynamic" directive for cookie),
Davor Ocelice9ed2812017-12-25 17:49:28 +01004533 a dynamic cookie is created for each server (unless one is explicitly
Olivier Houchard4e694042017-03-14 20:01:29 +01004534 specified on the "server" line), using a hash of the IP address of the
4535 server, the TCP port, and the secret key.
Davor Ocelice9ed2812017-12-25 17:49:28 +01004536 That way, we can ensure session persistence across multiple load-balancers,
Olivier Houchard4e694042017-03-14 20:01:29 +01004537 even if servers are dynamically added or removed.
Willy Tarreau5ce94572010-06-07 14:35:41 +02004538
Willy Tarreau0ba27502007-12-24 16:55:16 +01004539enabled
4540 Enable a proxy, frontend or backend.
4541 May be used in sections : defaults | frontend | listen | backend
4542 yes | yes | yes | yes
4543 Arguments : none
4544
4545 The "enabled" keyword is used to explicitly enable an instance, when the
4546 defaults has been set to "disabled". This is very rarely used.
4547
4548 See also : "disabled"
4549
4550
4551errorfile <code> <file>
4552 Return a file contents instead of errors generated by HAProxy
4553 May be used in sections : defaults | frontend | listen | backend
4554 yes | yes | yes | yes
4555 Arguments :
4556 <code> is the HTTP status code. Currently, HAProxy is capable of
Christopher Faulet612f2ea2020-05-27 09:57:28 +02004557 generating codes 200, 400, 401, 403, 404, 405, 407, 408, 410,
Christopher Faulete095f312020-12-07 11:22:24 +01004558 413, 425, 429, 500, 501, 502, 503, and 504.
Willy Tarreau0ba27502007-12-24 16:55:16 +01004559
4560 <file> designates a file containing the full HTTP response. It is
Willy Tarreaud2a4aa22008-01-31 15:28:22 +01004561 recommended to follow the common practice of appending ".http" to
Willy Tarreau0ba27502007-12-24 16:55:16 +01004562 the filename so that people do not confuse the response with HTML
Willy Tarreau59140a22009-02-22 12:02:30 +01004563 error pages, and to use absolute paths, since files are read
4564 before any chroot is performed.
Willy Tarreau0ba27502007-12-24 16:55:16 +01004565
4566 It is important to understand that this keyword is not meant to rewrite
4567 errors returned by the server, but errors detected and returned by HAProxy.
4568 This is why the list of supported errors is limited to a small set.
4569
Willy Tarreauae94d4d2011-05-11 16:28:49 +02004570 Code 200 is emitted in response to requests matching a "monitor-uri" rule.
4571
Christopher Faulet70170672020-05-18 17:42:48 +02004572 The files are parsed when HAProxy starts and must be valid according to the
4573 HTTP specification. They should not exceed the configured buffer size
4574 (BUFSIZE), which generally is 16 kB, otherwise an internal error will be
4575 returned. It is also wise not to put any reference to local contents
4576 (e.g. images) in order to avoid loops between the client and HAProxy when all
4577 servers are down, causing an error to be returned instead of an
4578 image. Finally, The response cannot exceed (tune.bufsize - tune.maxrewrite)
4579 so that "http-after-response" rules still have room to operate (see
4580 "tune.maxrewrite").
Willy Tarreau59140a22009-02-22 12:02:30 +01004581
Willy Tarreau0ba27502007-12-24 16:55:16 +01004582 The files are read at the same time as the configuration and kept in memory.
4583 For this reason, the errors continue to be returned even when the process is
4584 chrooted, and no file change is considered while the process is running. A
Willy Tarreauc27debf2008-01-06 08:57:02 +01004585 simple method for developing those files consists in associating them to the
Willy Tarreau0ba27502007-12-24 16:55:16 +01004586 403 status code and interrogating a blocked URL.
4587
Christopher Faulet3b967c12020-05-15 15:47:44 +02004588 See also : "http-error", "errorloc", "errorloc302", "errorloc303"
Willy Tarreau0ba27502007-12-24 16:55:16 +01004589
Willy Tarreau59140a22009-02-22 12:02:30 +01004590 Example :
4591 errorfile 400 /etc/haproxy/errorfiles/400badreq.http
Willy Tarreau989222a2016-01-15 10:26:26 +01004592 errorfile 408 /dev/null # work around Chrome pre-connect bug
Willy Tarreau59140a22009-02-22 12:02:30 +01004593 errorfile 403 /etc/haproxy/errorfiles/403forbid.http
4594 errorfile 503 /etc/haproxy/errorfiles/503sorry.http
4595
Willy Tarreau2769aa02007-12-27 18:26:09 +01004596
Christopher Faulet76edc0f2020-01-13 15:52:01 +01004597errorfiles <name> [<code> ...]
4598 Import, fully or partially, the error files defined in the <name> http-errors
4599 section.
4600 May be used in sections : defaults | frontend | listen | backend
4601 yes | yes | yes | yes
4602 Arguments :
4603 <name> is the name of an existing http-errors section.
4604
4605 <code> is a HTTP status code. Several status code may be listed.
Christopher Faulet612f2ea2020-05-27 09:57:28 +02004606 Currently, HAProxy is capable of generating codes 200, 400, 401,
Christopher Faulete095f312020-12-07 11:22:24 +01004607 403, 404, 405, 407, 408, 410, 413, 425, 429, 500, 501, 502, 503,
4608 and 504.
Christopher Faulet76edc0f2020-01-13 15:52:01 +01004609
4610 Errors defined in the http-errors section with the name <name> are imported
4611 in the current proxy. If no status code is specified, all error files of the
4612 http-errors section are imported. Otherwise, only error files associated to
4613 the listed status code are imported. Those error files override the already
4614 defined custom errors for the proxy. And they may be overridden by following
Daniel Corbett67a82712020-07-06 23:01:19 -04004615 ones. Functionally, it is exactly the same as declaring all error files by
Christopher Faulet76edc0f2020-01-13 15:52:01 +01004616 hand using "errorfile" directives.
4617
Christopher Faulet3b967c12020-05-15 15:47:44 +02004618 See also : "http-error", "errorfile", "errorloc", "errorloc302" ,
4619 "errorloc303" and section 3.8 about http-errors.
Christopher Faulet76edc0f2020-01-13 15:52:01 +01004620
4621 Example :
4622 errorfiles generic
4623 errorfiles site-1 403 404
4624
4625
Willy Tarreau2769aa02007-12-27 18:26:09 +01004626errorloc <code> <url>
4627errorloc302 <code> <url>
4628 Return an HTTP redirection to a URL instead of errors generated by HAProxy
4629 May be used in sections : defaults | frontend | listen | backend
4630 yes | yes | yes | yes
4631 Arguments :
4632 <code> is the HTTP status code. Currently, HAProxy is capable of
Christopher Faulet612f2ea2020-05-27 09:57:28 +02004633 generating codes 200, 400, 401, 403, 404, 405, 407, 408, 410,
Christopher Faulete095f312020-12-07 11:22:24 +01004634 413, 425, 429, 500, 501, 502, 503, and 504.
Willy Tarreau2769aa02007-12-27 18:26:09 +01004635
4636 <url> it is the exact contents of the "Location" header. It may contain
4637 either a relative URI to an error page hosted on the same site,
4638 or an absolute URI designating an error page on another site.
4639 Special care should be given to relative URIs to avoid redirect
Davor Ocelice9ed2812017-12-25 17:49:28 +01004640 loops if the URI itself may generate the same error (e.g. 500).
Willy Tarreau2769aa02007-12-27 18:26:09 +01004641
4642 It is important to understand that this keyword is not meant to rewrite
4643 errors returned by the server, but errors detected and returned by HAProxy.
4644 This is why the list of supported errors is limited to a small set.
4645
Willy Tarreauae94d4d2011-05-11 16:28:49 +02004646 Code 200 is emitted in response to requests matching a "monitor-uri" rule.
4647
Willy Tarreau2769aa02007-12-27 18:26:09 +01004648 Note that both keyword return the HTTP 302 status code, which tells the
4649 client to fetch the designated URL using the same HTTP method. This can be
4650 quite problematic in case of non-GET methods such as POST, because the URL
4651 sent to the client might not be allowed for something other than GET. To
Willy Tarreau989222a2016-01-15 10:26:26 +01004652 work around this problem, please use "errorloc303" which send the HTTP 303
Willy Tarreau2769aa02007-12-27 18:26:09 +01004653 status code, indicating to the client that the URL must be fetched with a GET
4654 request.
4655
Christopher Faulet3b967c12020-05-15 15:47:44 +02004656 See also : "http-error", "errorfile", "errorloc303"
Willy Tarreau2769aa02007-12-27 18:26:09 +01004657
4658
4659errorloc303 <code> <url>
4660 Return an HTTP redirection to a URL instead of errors generated by HAProxy
4661 May be used in sections : defaults | frontend | listen | backend
4662 yes | yes | yes | yes
4663 Arguments :
4664 <code> is the HTTP status code. Currently, HAProxy is capable of
Christopher Faulet612f2ea2020-05-27 09:57:28 +02004665 generating codes 200, 400, 401, 403, 404, 405, 407, 408, 410,
Christopher Faulete095f312020-12-07 11:22:24 +01004666 413, 425, 429, 500, 501, 502, 503, and 504.
Willy Tarreau2769aa02007-12-27 18:26:09 +01004667
4668 <url> it is the exact contents of the "Location" header. It may contain
4669 either a relative URI to an error page hosted on the same site,
4670 or an absolute URI designating an error page on another site.
4671 Special care should be given to relative URIs to avoid redirect
Davor Ocelice9ed2812017-12-25 17:49:28 +01004672 loops if the URI itself may generate the same error (e.g. 500).
Willy Tarreau2769aa02007-12-27 18:26:09 +01004673
4674 It is important to understand that this keyword is not meant to rewrite
4675 errors returned by the server, but errors detected and returned by HAProxy.
4676 This is why the list of supported errors is limited to a small set.
4677
Willy Tarreauae94d4d2011-05-11 16:28:49 +02004678 Code 200 is emitted in response to requests matching a "monitor-uri" rule.
4679
Willy Tarreau2769aa02007-12-27 18:26:09 +01004680 Note that both keyword return the HTTP 303 status code, which tells the
4681 client to fetch the designated URL using the same HTTP GET method. This
4682 solves the usual problems associated with "errorloc" and the 302 code. It is
4683 possible that some very old browsers designed before HTTP/1.1 do not support
Willy Tarreaud2a4aa22008-01-31 15:28:22 +01004684 it, but no such problem has been reported till now.
Willy Tarreau2769aa02007-12-27 18:26:09 +01004685
Christopher Faulet3b967c12020-05-15 15:47:44 +02004686 See also : "http-error", "errorfile", "errorloc", "errorloc302"
Willy Tarreau2769aa02007-12-27 18:26:09 +01004687
4688
Simon Horman51a1cf62015-02-03 13:00:44 +09004689email-alert from <emailaddr>
4690 Declare the from email address to be used in both the envelope and header
Davor Ocelice9ed2812017-12-25 17:49:28 +01004691 of email alerts. This is the address that email alerts are sent from.
Simon Horman51a1cf62015-02-03 13:00:44 +09004692 May be used in sections: defaults | frontend | listen | backend
4693 yes | yes | yes | yes
4694
4695 Arguments :
4696
4697 <emailaddr> is the from email address to use when sending email alerts
4698
4699 Also requires "email-alert mailers" and "email-alert to" to be set
4700 and if so sending email alerts is enabled for the proxy.
4701
Simon Horman64e34162015-02-06 11:11:57 +09004702 See also : "email-alert level", "email-alert mailers",
Cyril Bonté307ee1e2015-09-28 23:16:06 +02004703 "email-alert myhostname", "email-alert to", section 3.6 about
4704 mailers.
Simon Horman64e34162015-02-06 11:11:57 +09004705
4706
4707email-alert level <level>
4708 Declare the maximum log level of messages for which email alerts will be
4709 sent. This acts as a filter on the sending of email alerts.
4710 May be used in sections: defaults | frontend | listen | backend
4711 yes | yes | yes | yes
4712
4713 Arguments :
4714
4715 <level> One of the 8 syslog levels:
4716 emerg alert crit err warning notice info debug
4717 The above syslog levels are ordered from lowest to highest.
4718
4719 By default level is alert
4720
4721 Also requires "email-alert from", "email-alert mailers" and
4722 "email-alert to" to be set and if so sending email alerts is enabled
4723 for the proxy.
4724
Simon Horman1421e212015-04-30 13:10:35 +09004725 Alerts are sent when :
4726
4727 * An un-paused server is marked as down and <level> is alert or lower
4728 * A paused server is marked as down and <level> is notice or lower
4729 * A server is marked as up or enters the drain state and <level>
4730 is notice or lower
4731 * "option log-health-checks" is enabled, <level> is info or lower,
4732 and a health check status update occurs
4733
Simon Horman64e34162015-02-06 11:11:57 +09004734 See also : "email-alert from", "email-alert mailers",
4735 "email-alert myhostname", "email-alert to",
Simon Horman51a1cf62015-02-03 13:00:44 +09004736 section 3.6 about mailers.
4737
4738
4739email-alert mailers <mailersect>
4740 Declare the mailers to be used when sending email alerts
4741 May be used in sections: defaults | frontend | listen | backend
4742 yes | yes | yes | yes
4743
4744 Arguments :
4745
4746 <mailersect> is the name of the mailers section to send email alerts.
4747
4748 Also requires "email-alert from" and "email-alert to" to be set
4749 and if so sending email alerts is enabled for the proxy.
4750
Simon Horman64e34162015-02-06 11:11:57 +09004751 See also : "email-alert from", "email-alert level", "email-alert myhostname",
4752 "email-alert to", section 3.6 about mailers.
Simon Horman51a1cf62015-02-03 13:00:44 +09004753
4754
4755email-alert myhostname <hostname>
4756 Declare the to hostname address to be used when communicating with
4757 mailers.
4758 May be used in sections: defaults | frontend | listen | backend
4759 yes | yes | yes | yes
4760
4761 Arguments :
4762
Baptiste Assmann738bad92015-12-21 15:27:53 +01004763 <hostname> is the hostname to use when communicating with mailers
Simon Horman51a1cf62015-02-03 13:00:44 +09004764
4765 By default the systems hostname is used.
4766
4767 Also requires "email-alert from", "email-alert mailers" and
4768 "email-alert to" to be set and if so sending email alerts is enabled
4769 for the proxy.
4770
Simon Horman64e34162015-02-06 11:11:57 +09004771 See also : "email-alert from", "email-alert level", "email-alert mailers",
4772 "email-alert to", section 3.6 about mailers.
Simon Horman51a1cf62015-02-03 13:00:44 +09004773
4774
4775email-alert to <emailaddr>
Davor Ocelice9ed2812017-12-25 17:49:28 +01004776 Declare both the recipient address in the envelope and to address in the
Simon Horman51a1cf62015-02-03 13:00:44 +09004777 header of email alerts. This is the address that email alerts are sent to.
4778 May be used in sections: defaults | frontend | listen | backend
4779 yes | yes | yes | yes
4780
4781 Arguments :
4782
4783 <emailaddr> is the to email address to use when sending email alerts
4784
4785 Also requires "email-alert mailers" and "email-alert to" to be set
4786 and if so sending email alerts is enabled for the proxy.
4787
Simon Horman64e34162015-02-06 11:11:57 +09004788 See also : "email-alert from", "email-alert level", "email-alert mailers",
Simon Horman51a1cf62015-02-03 13:00:44 +09004789 "email-alert myhostname", section 3.6 about mailers.
4790
4791
Willy Tarreau4de91492010-01-22 19:10:05 +01004792force-persist { if | unless } <condition>
4793 Declare a condition to force persistence on down servers
4794 May be used in sections: defaults | frontend | listen | backend
Cyril Bonté4288c5a2018-03-12 22:02:59 +01004795 no | no | yes | yes
Willy Tarreau4de91492010-01-22 19:10:05 +01004796
4797 By default, requests are not dispatched to down servers. It is possible to
4798 force this using "option persist", but it is unconditional and redispatches
4799 to a valid server if "option redispatch" is set. That leaves with very little
4800 possibilities to force some requests to reach a server which is artificially
4801 marked down for maintenance operations.
4802
4803 The "force-persist" statement allows one to declare various ACL-based
4804 conditions which, when met, will cause a request to ignore the down status of
4805 a server and still try to connect to it. That makes it possible to start a
4806 server, still replying an error to the health checks, and run a specially
4807 configured browser to test the service. Among the handy methods, one could
4808 use a specific source IP address, or a specific cookie. The cookie also has
4809 the advantage that it can easily be added/removed on the browser from a test
4810 page. Once the service is validated, it is then possible to open the service
4811 to the world by returning a valid response to health checks.
4812
4813 The forced persistence is enabled when an "if" condition is met, or unless an
4814 "unless" condition is met. The final redispatch is always disabled when this
4815 is used.
4816
Cyril Bonté0d4bf012010-04-25 23:21:46 +02004817 See also : "option redispatch", "ignore-persist", "persist",
Cyril Bontéa8e7bbc2010-04-25 22:29:29 +02004818 and section 7 about ACL usage.
Willy Tarreau4de91492010-01-22 19:10:05 +01004819
Christopher Fauletc3fe5332016-04-07 15:30:10 +02004820
4821filter <name> [param*]
4822 Add the filter <name> in the filter list attached to the proxy.
4823 May be used in sections : defaults | frontend | listen | backend
4824 no | yes | yes | yes
4825 Arguments :
4826 <name> is the name of the filter. Officially supported filters are
4827 referenced in section 9.
4828
Tim Düsterhus4896c442016-11-29 02:15:19 +01004829 <param*> is a list of parameters accepted by the filter <name>. The
Christopher Fauletc3fe5332016-04-07 15:30:10 +02004830 parsing of these parameters are the responsibility of the
Tim Düsterhus4896c442016-11-29 02:15:19 +01004831 filter. Please refer to the documentation of the corresponding
4832 filter (section 9) for all details on the supported parameters.
Christopher Fauletc3fe5332016-04-07 15:30:10 +02004833
4834 Multiple occurrences of the filter line can be used for the same proxy. The
4835 same filter can be referenced many times if needed.
4836
4837 Example:
4838 listen
4839 bind *:80
4840
4841 filter trace name BEFORE-HTTP-COMP
4842 filter compression
4843 filter trace name AFTER-HTTP-COMP
4844
4845 compression algo gzip
4846 compression offload
4847
4848 server srv1 192.168.0.1:80
4849
4850 See also : section 9.
4851
Willy Tarreau4de91492010-01-22 19:10:05 +01004852
Willy Tarreau2769aa02007-12-27 18:26:09 +01004853fullconn <conns>
4854 Specify at what backend load the servers will reach their maxconn
4855 May be used in sections : defaults | frontend | listen | backend
4856 yes | no | yes | yes
4857 Arguments :
4858 <conns> is the number of connections on the backend which will make the
4859 servers use the maximal number of connections.
4860
Willy Tarreau198a7442008-01-17 12:05:32 +01004861 When a server has a "maxconn" parameter specified, it means that its number
Willy Tarreau2769aa02007-12-27 18:26:09 +01004862 of concurrent connections will never go higher. Additionally, if it has a
Willy Tarreau198a7442008-01-17 12:05:32 +01004863 "minconn" parameter, it indicates a dynamic limit following the backend's
Willy Tarreau2769aa02007-12-27 18:26:09 +01004864 load. The server will then always accept at least <minconn> connections,
4865 never more than <maxconn>, and the limit will be on the ramp between both
4866 values when the backend has less than <conns> concurrent connections. This
4867 makes it possible to limit the load on the servers during normal loads, but
4868 push it further for important loads without overloading the servers during
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01004869 exceptional loads.
Willy Tarreau2769aa02007-12-27 18:26:09 +01004870
Willy Tarreaufbb78422011-06-05 15:38:35 +02004871 Since it's hard to get this value right, haproxy automatically sets it to
4872 10% of the sum of the maxconns of all frontends that may branch to this
Bertrand Jacquin702d44f2013-11-19 11:43:06 +01004873 backend (based on "use_backend" and "default_backend" rules). That way it's
4874 safe to leave it unset. However, "use_backend" involving dynamic names are
4875 not counted since there is no way to know if they could match or not.
Willy Tarreaufbb78422011-06-05 15:38:35 +02004876
Willy Tarreau2769aa02007-12-27 18:26:09 +01004877 Example :
4878 # The servers will accept between 100 and 1000 concurrent connections each
4879 # and the maximum of 1000 will be reached when the backend reaches 10000
4880 # connections.
4881 backend dynamic
4882 fullconn 10000
4883 server srv1 dyn1:80 minconn 100 maxconn 1000
4884 server srv2 dyn2:80 minconn 100 maxconn 1000
4885
4886 See also : "maxconn", "server"
4887
4888
Willy Tarreauab0a5192020-10-09 19:07:01 +02004889grace <time> (deprecated)
Willy Tarreau2769aa02007-12-27 18:26:09 +01004890 Maintain a proxy operational for some time after a soft stop
4891 May be used in sections : defaults | frontend | listen | backend
Cyril Bonté99ed3272010-01-24 23:29:44 +01004892 yes | yes | yes | yes
Willy Tarreau2769aa02007-12-27 18:26:09 +01004893 Arguments :
4894 <time> is the time (by default in milliseconds) for which the instance
4895 will remain operational with the frontend sockets still listening
4896 when a soft-stop is received via the SIGUSR1 signal.
4897
4898 This may be used to ensure that the services disappear in a certain order.
4899 This was designed so that frontends which are dedicated to monitoring by an
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01004900 external equipment fail immediately while other ones remain up for the time
Willy Tarreau2769aa02007-12-27 18:26:09 +01004901 needed by the equipment to detect the failure.
4902
4903 Note that currently, there is very little benefit in using this parameter,
4904 and it may in fact complicate the soft-reconfiguration process more than
4905 simplify it.
4906
Willy Tarreau0ba27502007-12-24 16:55:16 +01004907
Andrew Rodland17be45e2016-10-25 17:04:12 -04004908hash-balance-factor <factor>
4909 Specify the balancing factor for bounded-load consistent hashing
4910 May be used in sections : defaults | frontend | listen | backend
4911 yes | no | no | yes
4912 Arguments :
4913 <factor> is the control for the maximum number of concurrent requests to
4914 send to a server, expressed as a percentage of the average number
Frédéric Lécaille93d33162019-03-06 09:35:59 +01004915 of concurrent requests across all of the active servers.
Andrew Rodland17be45e2016-10-25 17:04:12 -04004916
4917 Specifying a "hash-balance-factor" for a server with "hash-type consistent"
4918 enables an algorithm that prevents any one server from getting too many
4919 requests at once, even if some hash buckets receive many more requests than
4920 others. Setting <factor> to 0 (the default) disables the feature. Otherwise,
4921 <factor> is a percentage greater than 100. For example, if <factor> is 150,
4922 then no server will be allowed to have a load more than 1.5 times the average.
4923 If server weights are used, they will be respected.
4924
4925 If the first-choice server is disqualified, the algorithm will choose another
4926 server based on the request hash, until a server with additional capacity is
4927 found. A higher <factor> allows more imbalance between the servers, while a
4928 lower <factor> means that more servers will be checked on average, affecting
4929 performance. Reasonable values are from 125 to 200.
4930
Willy Tarreau760e81d2018-05-03 07:20:40 +02004931 This setting is also used by "balance random" which internally relies on the
4932 consistent hashing mechanism.
4933
Andrew Rodland17be45e2016-10-25 17:04:12 -04004934 See also : "balance" and "hash-type".
4935
4936
Bhaskar Maddalab6c0ac92013-11-05 11:54:02 -05004937hash-type <method> <function> <modifier>
Willy Tarreau6b2e11b2009-10-01 07:52:15 +02004938 Specify a method to use for mapping hashes to servers
4939 May be used in sections : defaults | frontend | listen | backend
4940 yes | no | yes | yes
4941 Arguments :
Bhaskar98634f02013-10-29 23:30:51 -04004942 <method> is the method used to select a server from the hash computed by
4943 the <function> :
Willy Tarreau6b2e11b2009-10-01 07:52:15 +02004944
Bhaskar98634f02013-10-29 23:30:51 -04004945 map-based the hash table is a static array containing all alive servers.
4946 The hashes will be very smooth, will consider weights, but
4947 will be static in that weight changes while a server is up
4948 will be ignored. This means that there will be no slow start.
4949 Also, since a server is selected by its position in the array,
4950 most mappings are changed when the server count changes. This
4951 means that when a server goes up or down, or when a server is
4952 added to a farm, most connections will be redistributed to
4953 different servers. This can be inconvenient with caches for
4954 instance.
Willy Tarreau798a39c2010-11-24 15:04:29 +01004955
Bhaskar98634f02013-10-29 23:30:51 -04004956 consistent the hash table is a tree filled with many occurrences of each
4957 server. The hash key is looked up in the tree and the closest
4958 server is chosen. This hash is dynamic, it supports changing
4959 weights while the servers are up, so it is compatible with the
4960 slow start feature. It has the advantage that when a server
4961 goes up or down, only its associations are moved. When a
4962 server is added to the farm, only a few part of the mappings
4963 are redistributed, making it an ideal method for caches.
4964 However, due to its principle, the distribution will never be
4965 very smooth and it may sometimes be necessary to adjust a
4966 server's weight or its ID to get a more balanced distribution.
4967 In order to get the same distribution on multiple load
4968 balancers, it is important that all servers have the exact
Bhaskar Maddalab6c0ac92013-11-05 11:54:02 -05004969 same IDs. Note: consistent hash uses sdbm and avalanche if no
4970 hash function is specified.
Bhaskar98634f02013-10-29 23:30:51 -04004971
4972 <function> is the hash function to be used :
4973
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03004974 sdbm this function was created initially for sdbm (a public-domain
Bhaskar98634f02013-10-29 23:30:51 -04004975 reimplementation of ndbm) database library. It was found to do
4976 well in scrambling bits, causing better distribution of the keys
4977 and fewer splits. It also happens to be a good general hashing
Bhaskar Maddalab6c0ac92013-11-05 11:54:02 -05004978 function with good distribution, unless the total server weight
4979 is a multiple of 64, in which case applying the avalanche
4980 modifier may help.
Bhaskar98634f02013-10-29 23:30:51 -04004981
4982 djb2 this function was first proposed by Dan Bernstein many years ago
4983 on comp.lang.c. Studies have shown that for certain workload this
Bhaskar Maddalab6c0ac92013-11-05 11:54:02 -05004984 function provides a better distribution than sdbm. It generally
4985 works well with text-based inputs though it can perform extremely
4986 poorly with numeric-only input or when the total server weight is
4987 a multiple of 33, unless the avalanche modifier is also used.
4988
Willy Tarreaua0f42712013-11-14 14:30:35 +01004989 wt6 this function was designed for haproxy while testing other
4990 functions in the past. It is not as smooth as the other ones, but
4991 is much less sensible to the input data set or to the number of
4992 servers. It can make sense as an alternative to sdbm+avalanche or
4993 djb2+avalanche for consistent hashing or when hashing on numeric
4994 data such as a source IP address or a visitor identifier in a URL
4995 parameter.
4996
Willy Tarreau324f07f2015-01-20 19:44:50 +01004997 crc32 this is the most common CRC32 implementation as used in Ethernet,
4998 gzip, PNG, etc. It is slower than the other ones but may provide
4999 a better distribution or less predictable results especially when
5000 used on strings.
5001
Bhaskar Maddalab6c0ac92013-11-05 11:54:02 -05005002 <modifier> indicates an optional method applied after hashing the key :
5003
5004 avalanche This directive indicates that the result from the hash
5005 function above should not be used in its raw form but that
5006 a 4-byte full avalanche hash must be applied first. The
5007 purpose of this step is to mix the resulting bits from the
5008 previous hash in order to avoid any undesired effect when
5009 the input contains some limited values or when the number of
5010 servers is a multiple of one of the hash's components (64
5011 for SDBM, 33 for DJB2). Enabling avalanche tends to make the
5012 result less predictable, but it's also not as smooth as when
5013 using the original function. Some testing might be needed
5014 with some workloads. This hash is one of the many proposed
5015 by Bob Jenkins.
Willy Tarreau6b2e11b2009-10-01 07:52:15 +02005016
Bhaskar98634f02013-10-29 23:30:51 -04005017 The default hash type is "map-based" and is recommended for most usages. The
5018 default function is "sdbm", the selection of a function should be based on
5019 the range of the values being hashed.
Willy Tarreau6b2e11b2009-10-01 07:52:15 +02005020
Andrew Rodland17be45e2016-10-25 17:04:12 -04005021 See also : "balance", "hash-balance-factor", "server"
Willy Tarreau6b2e11b2009-10-01 07:52:15 +02005022
5023
Christopher Faulet6d0c3df2020-01-22 09:26:35 +01005024http-after-response <action> <options...> [ { if | unless } <condition> ]
5025 Access control for all Layer 7 responses (server, applet/service and internal
5026 ones).
5027
5028 May be used in sections: defaults | frontend | listen | backend
5029 no | yes | yes | yes
5030
5031 The http-after-response statement defines a set of rules which apply to layer
5032 7 processing. The rules are evaluated in their declaration order when they
5033 are met in a frontend, listen or backend section. Any rule may optionally be
5034 followed by an ACL-based condition, in which case it will only be evaluated
5035 if the condition is true. Since these rules apply on responses, the backend
5036 rules are applied first, followed by the frontend's rules.
5037
5038 Unlike http-response rules, these ones are applied on all responses, the
5039 server ones but also to all responses generated by HAProxy. These rules are
5040 evaluated at the end of the responses analysis, before the data forwarding.
5041
5042 The first keyword is the rule's action. The supported actions are described
5043 below.
5044
5045 There is no limit to the number of http-after-response statements per
5046 instance.
5047
Christopher Fauletd5ac6de2020-12-02 08:40:14 +01005048 Note: Errors emitted in early stage of the request parsing are handled by the
5049 multiplexer at a lower level, before any http analysis. Thus no
5050 http-after-response ruleset is evaluated on these errors.
5051
Christopher Faulet6d0c3df2020-01-22 09:26:35 +01005052 Example:
5053 http-after-response set-header Strict-Transport-Security "max-age=31536000"
5054 http-after-response set-header Cache-Control "no-store,no-cache,private"
5055 http-after-response set-header Pragma "no-cache"
5056
5057http-after-response add-header <name> <fmt> [ { if | unless } <condition> ]
5058
5059 This appends an HTTP header field whose name is specified in <name> and whose
5060 value is defined by <fmt> which follows the log-format rules (see Custom Log
5061 Format in section 8.2.4). This may be used to send a cookie to a client for
5062 example, or to pass some internal information.
5063 This rule is not final, so it is possible to add other similar rules.
5064 Note that header addition is performed immediately, so one rule might reuse
5065 the resulting header from a previous rule.
5066
5067http-after-response allow [ { if | unless } <condition> ]
5068
5069 This stops the evaluation of the rules and lets the response pass the check.
5070 No further "http-after-response" rules are evaluated.
5071
Maciej Zdebebdd4c52020-11-20 13:58:48 +00005072http-after-response del-header <name> [ -m <meth> ] [ { if | unless } <condition> ]
Christopher Faulet6d0c3df2020-01-22 09:26:35 +01005073
Maciej Zdebebdd4c52020-11-20 13:58:48 +00005074 This removes all HTTP header fields whose name is specified in <name>. <meth>
5075 is the matching method, applied on the header name. Supported matching methods
5076 are "str" (exact match), "beg" (prefix match), "end" (suffix match), "sub"
5077 (substring match) and "reg" (regex match). If not specified, exact matching
5078 method is used.
Christopher Faulet6d0c3df2020-01-22 09:26:35 +01005079
5080http-after-response replace-header <name> <regex-match> <replace-fmt>
5081 [ { if | unless } <condition> ]
5082
5083 This works like "http-response replace-header".
5084
5085 Example:
5086 http-after-response replace-header Set-Cookie (C=[^;]*);(.*) \1;ip=%bi;\2
5087
5088 # applied to:
5089 Set-Cookie: C=1; expires=Tue, 14-Jun-2016 01:40:45 GMT
5090
5091 # outputs:
5092 Set-Cookie: C=1;ip=192.168.1.20; expires=Tue, 14-Jun-2016 01:40:45 GMT
5093
5094 # assuming the backend IP is 192.168.1.20.
5095
5096http-after-response replace-value <name> <regex-match> <replace-fmt>
5097 [ { if | unless } <condition> ]
5098
5099 This works like "http-response replace-value".
5100
5101 Example:
5102 http-after-response replace-value Cache-control ^public$ private
5103
5104 # applied to:
5105 Cache-Control: max-age=3600, public
5106
5107 # outputs:
5108 Cache-Control: max-age=3600, private
5109
5110http-after-response set-header <name> <fmt> [ { if | unless } <condition> ]
5111
5112 This does the same as "add-header" except that the header name is first
5113 removed if it existed. This is useful when passing security information to
5114 the server, where the header must not be manipulated by external users.
5115
5116http-after-response set-status <status> [reason <str>]
5117 [ { if | unless } <condition> ]
5118
5119 This replaces the response status code with <status> which must be an integer
5120 between 100 and 999. Optionally, a custom reason text can be provided defined
5121 by <str>, or the default reason for the specified code will be used as a
5122 fallback.
5123
5124 Example:
5125 # return "431 Request Header Fields Too Large"
5126 http-response set-status 431
5127 # return "503 Slow Down", custom reason
5128 http-response set-status 503 reason "Slow Down"
5129
5130http-after-response set-var(<var-name>) <expr> [ { if | unless } <condition> ]
5131
5132 This is used to set the contents of a variable. The variable is declared
5133 inline.
5134
5135 Arguments:
5136 <var-name> The name of the variable starts with an indication about its
5137 scope. The scopes allowed are:
5138 "proc" : the variable is shared with the whole process
5139 "sess" : the variable is shared with the whole session
5140 "txn" : the variable is shared with the transaction
5141 (request and response)
5142 "req" : the variable is shared only during request
5143 processing
5144 "res" : the variable is shared only during response
5145 processing
5146 This prefix is followed by a name. The separator is a '.'.
5147 The name may only contain characters 'a-z', 'A-Z', '0-9', '.'
5148 and '_'.
5149
5150 <expr> Is a standard HAProxy expression formed by a sample-fetch
5151 followed by some converters.
5152
5153 Example:
5154 http-after-response set-var(sess.last_redir) res.hdr(location)
5155
5156http-after-response strict-mode { on | off }
5157
5158 This enables or disables the strict rewriting mode for following rules. It
5159 does not affect rules declared before it and it is only applicable on rules
5160 performing a rewrite on the responses. When the strict mode is enabled, any
5161 rewrite failure triggers an internal error. Otherwise, such errors are
5162 silently ignored. The purpose of the strict rewriting mode is to make some
Ilya Shipitsin8525fd92020-02-29 12:34:59 +05005163 rewrites optional while others must be performed to continue the response
Christopher Faulet6d0c3df2020-01-22 09:26:35 +01005164 processing.
5165
5166 By default, the strict rewriting mode is enabled. Its value is also reset
5167 when a ruleset evaluation ends. So, for instance, if you change the mode on
Daniel Corbett67a82712020-07-06 23:01:19 -04005168 the backend, the default mode is restored when HAProxy starts the frontend
Christopher Faulet6d0c3df2020-01-22 09:26:35 +01005169 rules evaluation.
5170
5171http-after-response unset-var(<var-name>) [ { if | unless } <condition> ]
5172
5173 This is used to unset a variable. See "http-after-response set-var" for
5174 details about <var-name>.
5175
5176 Example:
5177 http-after-response unset-var(sess.last_redir)
5178
Christopher Faulet4f5c2e22020-04-23 15:22:33 +02005179
5180http-check comment <string>
5181 Defines a comment for the following the http-check rule, reported in logs if
5182 it fails.
5183 May be used in sections : defaults | frontend | listen | backend
5184 yes | no | yes | yes
5185
Christopher Faulet4f5c2e22020-04-23 15:22:33 +02005186 Arguments :
5187 <string> is the comment message to add in logs if the following http-check
5188 rule fails.
5189
Christopher Fauletc52ea4d2020-04-23 15:43:35 +02005190 It only works for connect, send and expect rules. It is useful to make
5191 user-friendly error reporting.
5192
Daniel Corbett67a82712020-07-06 23:01:19 -04005193 See also : "option httpchk", "http-check connect", "http-check send" and
Christopher Faulet4f5c2e22020-04-23 15:22:33 +02005194 "http-check expect".
5195
5196
Christopher Fauletc52ea4d2020-04-23 15:43:35 +02005197http-check connect [default] [port <expr>] [addr <ip>] [send-proxy]
5198 [via-socks4] [ssl] [sni <sni>] [alpn <alpn>] [linger]
Christopher Fauletedc6ed92020-04-23 16:27:59 +02005199 [proto <name>] [comment <msg>]
Christopher Faulete5870d82020-04-15 11:32:03 +02005200 Opens a new connection to perform an HTTP health check
5201 May be used in sections : defaults | frontend | listen | backend
5202 yes | no | yes | yes
5203
Christopher Fauletc52ea4d2020-04-23 15:43:35 +02005204 Arguments :
Christopher Faulet4f5c2e22020-04-23 15:22:33 +02005205 comment <msg> defines a message to report if the rule evaluation fails.
5206
Christopher Faulete5870d82020-04-15 11:32:03 +02005207 default Use default options of the server line to do the health
Daniel Corbett67a82712020-07-06 23:01:19 -04005208 checks. The server options are used only if not redefined.
Christopher Faulete5870d82020-04-15 11:32:03 +02005209
5210 port <expr> if not set, check port or server port is used.
5211 It tells HAProxy where to open the connection to.
5212 <port> must be a valid TCP port source integer, from 1 to
5213 65535 or an sample-fetch expression.
5214
5215 addr <ip> defines the IP address to do the health check.
5216
5217 send-proxy send a PROXY protocol string
5218
5219 via-socks4 enables outgoing health checks using upstream socks4 proxy.
5220
5221 ssl opens a ciphered connection
5222
5223 sni <sni> specifies the SNI to use to do health checks over SSL.
5224
5225 alpn <alpn> defines which protocols to advertise with ALPN. The protocol
5226 list consists in a comma-delimited list of protocol names,
5227 for instance: "h2,http/1.1". If it is not set, the server ALPN
5228 is used.
5229
Christopher Fauletedc6ed92020-04-23 16:27:59 +02005230 proto <name> forces the multiplexer's protocol to use for this connection.
5231 It must be an HTTP mux protocol and it must be usable on the
5232 backend side. The list of available protocols is reported in
5233 haproxy -vv.
5234
Christopher Faulete5870d82020-04-15 11:32:03 +02005235 linger cleanly close the connection instead of using a single RST.
5236
Christopher Fauletc52ea4d2020-04-23 15:43:35 +02005237 Just like tcp-check health checks, it is possible to configure the connection
5238 to use to perform HTTP health check. This directive should also be used to
5239 describe a scenario involving several request/response exchanges, possibly on
5240 different ports or with different servers.
5241
5242 When there are no TCP port configured on the server line neither server port
5243 directive, then the first step of the http-check sequence must be to specify
5244 the port with a "http-check connect".
5245
5246 In an http-check ruleset a 'connect' is required, it is also mandatory to start
5247 the ruleset with a 'connect' rule. Purpose is to ensure admin know what they
5248 do.
5249
5250 When a connect must start the ruleset, if may still be preceded by set-var,
5251 unset-var or comment rules.
5252
5253 Examples :
Christopher Faulete5870d82020-04-15 11:32:03 +02005254 # check HTTP and HTTPs services on a server.
5255 # first open port 80 thanks to server line port directive, then
5256 # tcp-check opens port 443, ciphered and run a request on it:
5257 option httpchk
5258
5259 http-check connect
Christopher Fauleta5c14ef2020-04-29 14:19:13 +02005260 http-check send meth GET uri / ver HTTP/1.1 hdr host haproxy.1wt.eu
Christopher Faulet8021a5f2020-04-24 13:53:12 +02005261 http-check expect status 200-399
Christopher Faulete5870d82020-04-15 11:32:03 +02005262 http-check connect port 443 ssl sni haproxy.1wt.eu
Christopher Fauleta5c14ef2020-04-29 14:19:13 +02005263 http-check send meth GET uri / ver HTTP/1.1 hdr host haproxy.1wt.eu
Christopher Faulet8021a5f2020-04-24 13:53:12 +02005264 http-check expect status 200-399
Christopher Faulete5870d82020-04-15 11:32:03 +02005265
5266 server www 10.0.0.1 check port 80
5267
5268 See also : "option httpchk", "http-check send", "http-check expect"
Christopher Faulet6d0c3df2020-01-22 09:26:35 +01005269
Christopher Fauletc52ea4d2020-04-23 15:43:35 +02005270
Willy Tarreau0ba27502007-12-24 16:55:16 +01005271http-check disable-on-404
5272 Enable a maintenance mode upon HTTP/404 response to health-checks
5273 May be used in sections : defaults | frontend | listen | backend
Willy Tarreau2769aa02007-12-27 18:26:09 +01005274 yes | no | yes | yes
Willy Tarreau0ba27502007-12-24 16:55:16 +01005275 Arguments : none
5276
5277 When this option is set, a server which returns an HTTP code 404 will be
5278 excluded from further load-balancing, but will still receive persistent
5279 connections. This provides a very convenient method for Web administrators
5280 to perform a graceful shutdown of their servers. It is also important to note
5281 that a server which is detected as failed while it was in this mode will not
5282 generate an alert, just a notice. If the server responds 2xx or 3xx again, it
5283 will immediately be reinserted into the farm. The status on the stats page
5284 reports "NOLB" for a server in this mode. It is important to note that this
Willy Tarreaubd741542010-03-16 18:46:54 +01005285 option only works in conjunction with the "httpchk" option. If this option
5286 is used with "http-check expect", then it has precedence over it so that 404
Christopher Fauletfa8b89a2020-11-20 18:54:13 +01005287 responses will still be considered as soft-stop. Note also that a stopped
5288 server will stay stopped even if it replies 404s. This option is only
5289 evaluated for running servers.
Willy Tarreaubd741542010-03-16 18:46:54 +01005290
Christopher Fauletc52ea4d2020-04-23 15:43:35 +02005291 See also : "option httpchk" and "http-check expect".
Willy Tarreaubd741542010-03-16 18:46:54 +01005292
5293
Christopher Faulet4f5c2e22020-04-23 15:22:33 +02005294http-check expect [min-recv <int>] [comment <msg>]
Christopher Faulete5870d82020-04-15 11:32:03 +02005295 [ok-status <st>] [error-status <st>] [tout-status <st>]
5296 [on-success <fmt>] [on-error <fmt>] [status-code <expr>]
5297 [!] <match> <pattern>
Jamie Gloudonaaa21002012-08-25 00:18:33 -04005298 Make HTTP health checks consider response contents or specific status codes
Willy Tarreaubd741542010-03-16 18:46:54 +01005299 May be used in sections : defaults | frontend | listen | backend
Willy Tarreau1ee51a62011-08-19 20:04:17 +02005300 yes | no | yes | yes
Christopher Faulete5870d82020-04-15 11:32:03 +02005301
Willy Tarreaubd741542010-03-16 18:46:54 +01005302 Arguments :
Christopher Faulet4f5c2e22020-04-23 15:22:33 +02005303 comment <msg> defines a message to report if the rule evaluation fails.
5304
Christopher Faulete5870d82020-04-15 11:32:03 +02005305 min-recv is optional and can define the minimum amount of data required to
5306 evaluate the current expect rule. If the number of received bytes
5307 is under this limit, the check will wait for more data. This
5308 option can be used to resolve some ambiguous matching rules or to
5309 avoid executing costly regex matches on content known to be still
5310 incomplete. If an exact string is used, the minimum between the
5311 string length and this parameter is used. This parameter is
5312 ignored if it is set to -1. If the expect rule does not match,
5313 the check will wait for more data. If set to 0, the evaluation
5314 result is always conclusive.
5315
5316 ok-status <st> is optional and can be used to set the check status if
5317 the expect rule is successfully evaluated and if it is
5318 the last rule in the tcp-check ruleset. "L7OK", "L7OKC",
Christopher Fauletd888f0f2020-05-07 07:40:17 +02005319 "L6OK" and "L4OK" are supported :
5320 - L7OK : check passed on layer 7
Christopher Faulet83662b52020-11-20 17:47:47 +01005321 - L7OKC : check conditionally passed on layer 7, set
5322 server to NOLB state.
Christopher Fauletd888f0f2020-05-07 07:40:17 +02005323 - L6OK : check passed on layer 6
5324 - L4OK : check passed on layer 4
5325 By default "L7OK" is used.
Christopher Faulete5870d82020-04-15 11:32:03 +02005326
5327 error-status <st> is optional and can be used to set the check status if
5328 an error occurred during the expect rule evaluation.
Christopher Faulet83662b52020-11-20 17:47:47 +01005329 "L7OKC", "L7RSP", "L7STS", "L6RSP" and "L4CON" are
5330 supported :
5331 - L7OKC : check conditionally passed on layer 7, set
5332 server to NOLB state.
Christopher Fauletd888f0f2020-05-07 07:40:17 +02005333 - L7RSP : layer 7 invalid response - protocol error
5334 - L7STS : layer 7 response error, for example HTTP 5xx
5335 - L6RSP : layer 6 invalid response - protocol error
5336 - L4CON : layer 1-4 connection problem
5337 By default "L7RSP" is used.
Christopher Faulete5870d82020-04-15 11:32:03 +02005338
5339 tout-status <st> is optional and can be used to set the check status if
5340 a timeout occurred during the expect rule evaluation.
Christopher Fauletd888f0f2020-05-07 07:40:17 +02005341 "L7TOUT", "L6TOUT", and "L4TOUT" are supported :
5342 - L7TOUT : layer 7 (HTTP/SMTP) timeout
5343 - L6TOUT : layer 6 (SSL) timeout
5344 - L4TOUT : layer 1-4 timeout
Christopher Faulete5870d82020-04-15 11:32:03 +02005345 By default "L7TOUT" is used.
5346
5347 on-success <fmt> is optional and can be used to customize the
5348 informational message reported in logs if the expect
5349 rule is successfully evaluated and if it is the last rule
5350 in the tcp-check ruleset. <fmt> is a log-format string.
5351
5352 on-error <fmt> is optional and can be used to customize the
5353 informational message reported in logs if an error
5354 occurred during the expect rule evaluation. <fmt> is a
5355 log-format string.
5356
Willy Tarreaubd741542010-03-16 18:46:54 +01005357 <match> is a keyword indicating how to look for a specific pattern in the
Christopher Fauletb5594262020-05-05 20:23:13 +02005358 response. The keyword may be one of "status", "rstatus", "hdr",
5359 "fhdr", "string", or "rstring". The keyword may be preceded by an
Willy Tarreaubd741542010-03-16 18:46:54 +01005360 exclamation mark ("!") to negate the match. Spaces are allowed
5361 between the exclamation mark and the keyword. See below for more
5362 details on the supported keywords.
5363
Christopher Faulet39708192020-05-05 10:47:36 +02005364 <pattern> is the pattern to look for. It may be a string, a regular
5365 expression or a more complex pattern with several arguments. If
5366 the string pattern contains spaces, they must be escaped with the
5367 usual backslash ('\').
Willy Tarreaubd741542010-03-16 18:46:54 +01005368
5369 By default, "option httpchk" considers that response statuses 2xx and 3xx
5370 are valid, and that others are invalid. When "http-check expect" is used,
5371 it defines what is considered valid or invalid. Only one "http-check"
5372 statement is supported in a backend. If a server fails to respond or times
5373 out, the check obviously fails. The available matches are :
5374
Christopher Faulet8021a5f2020-04-24 13:53:12 +02005375 status <codes> : test the status codes found parsing <codes> string. it
5376 must be a comma-separated list of status codes or range
5377 codes. A health check response will be considered as
5378 valid if the response's status code matches any status
5379 code or is inside any range of the list. If the "status"
5380 keyword is prefixed with "!", then the response will be
5381 considered invalid if the status code matches.
Willy Tarreaubd741542010-03-16 18:46:54 +01005382
5383 rstatus <regex> : test a regular expression for the HTTP status code.
Jamie Gloudonaaa21002012-08-25 00:18:33 -04005384 A health check response will be considered valid if the
Willy Tarreaubd741542010-03-16 18:46:54 +01005385 response's status code matches the expression. If the
5386 "rstatus" keyword is prefixed with "!", then the response
5387 will be considered invalid if the status code matches.
5388 This is mostly used to check for multiple codes.
5389
Christopher Fauletb5594262020-05-05 20:23:13 +02005390 hdr { name | name-lf } [ -m <meth> ] <name>
5391 [ { value | value-lf } [ -m <meth> ] <value> :
Christopher Faulet39708192020-05-05 10:47:36 +02005392 test the specified header pattern on the HTTP response
5393 headers. The name pattern is mandatory but the value
5394 pattern is optional. If not specified, only the header
5395 presence is verified. <meth> is the matching method,
5396 applied on the header name or the header value. Supported
5397 matching methods are "str" (exact match), "beg" (prefix
5398 match), "end" (suffix match), "sub" (substring match) or
5399 "reg" (regex match). If not specified, exact matching
Christopher Fauletb5594262020-05-05 20:23:13 +02005400 method is used. If the "name-lf" parameter is used,
5401 <name> is evaluated as a log-format string. If "value-lf"
5402 parameter is used, <value> is evaluated as a log-format
5403 string. These parameters cannot be used with the regex
5404 matching method. Finally, the header value is considered
5405 as comma-separated list. Note that matchings are case
5406 insensitive on the header names.
5407
5408 fhdr { name | name-lf } [ -m <meth> ] <name>
5409 [ { value | value-lf } [ -m <meth> ] <value> :
5410 test the specified full header pattern on the HTTP
5411 response headers. It does exactly the same than "hdr"
5412 keyword, except the full header value is tested, commas
5413 are not considered as delimiters.
Christopher Faulet39708192020-05-05 10:47:36 +02005414
Willy Tarreaubd741542010-03-16 18:46:54 +01005415 string <string> : test the exact string match in the HTTP response body.
Jamie Gloudonaaa21002012-08-25 00:18:33 -04005416 A health check response will be considered valid if the
Willy Tarreaubd741542010-03-16 18:46:54 +01005417 response's body contains this exact string. If the
5418 "string" keyword is prefixed with "!", then the response
5419 will be considered invalid if the body contains this
5420 string. This can be used to look for a mandatory word at
5421 the end of a dynamic page, or to detect a failure when a
Davor Ocelice9ed2812017-12-25 17:49:28 +01005422 specific error appears on the check page (e.g. a stack
Willy Tarreaubd741542010-03-16 18:46:54 +01005423 trace).
5424
5425 rstring <regex> : test a regular expression on the HTTP response body.
Jamie Gloudonaaa21002012-08-25 00:18:33 -04005426 A health check response will be considered valid if the
Willy Tarreaubd741542010-03-16 18:46:54 +01005427 response's body matches this expression. If the "rstring"
5428 keyword is prefixed with "!", then the response will be
5429 considered invalid if the body matches the expression.
5430 This can be used to look for a mandatory word at the end
5431 of a dynamic page, or to detect a failure when a specific
Davor Ocelice9ed2812017-12-25 17:49:28 +01005432 error appears on the check page (e.g. a stack trace).
Willy Tarreaubd741542010-03-16 18:46:54 +01005433
Christopher Fauletaaab0832020-05-05 15:54:22 +02005434 string-lf <fmt> : test a log-format string match in the HTTP response body.
5435 A health check response will be considered valid if the
5436 response's body contains the string resulting of the
5437 evaluation of <fmt>, which follows the log-format rules.
5438 If prefixed with "!", then the response will be
5439 considered invalid if the body contains the string.
5440
Willy Tarreaubd741542010-03-16 18:46:54 +01005441 It is important to note that the responses will be limited to a certain size
Christopher Fauletbb9fb8b2020-11-25 17:20:57 +01005442 defined by the global "tune.bufsize" option, which defaults to 16384 bytes.
Willy Tarreaubd741542010-03-16 18:46:54 +01005443 Thus, too large responses may not contain the mandatory pattern when using
5444 "string" or "rstring". If a large response is absolutely required, it is
5445 possible to change the default max size by setting the global variable.
5446 However, it is worth keeping in mind that parsing very large responses can
5447 waste some CPU cycles, especially when regular expressions are used, and that
5448 it is always better to focus the checks on smaller resources.
5449
Christopher Faulete5870d82020-04-15 11:32:03 +02005450 In an http-check ruleset, the last expect rule may be implicit. If no expect
5451 rule is specified after the last "http-check send", an implicit expect rule
5452 is defined to match on 2xx or 3xx status codes. It means this rule is also
5453 defined if there is no "http-check" rule at all, when only "option httpchk"
5454 is set.
Cyril Bonté32602d22015-01-30 00:07:07 +01005455
Willy Tarreaubd741542010-03-16 18:46:54 +01005456 Last, if "http-check expect" is combined with "http-check disable-on-404",
5457 then this last one has precedence when the server responds with 404.
5458
5459 Examples :
5460 # only accept status 200 as valid
Christopher Faulet8021a5f2020-04-24 13:53:12 +02005461 http-check expect status 200,201,300-310
Willy Tarreaubd741542010-03-16 18:46:54 +01005462
Christopher Faulet39708192020-05-05 10:47:36 +02005463 # be sure a sessid coookie is set
5464 http-check expect header name "set-cookie" value -m beg "sessid="
5465
Willy Tarreaubd741542010-03-16 18:46:54 +01005466 # consider SQL errors as errors
Willy Tarreau8f2a1e72011-01-06 16:36:10 +01005467 http-check expect ! string SQL\ Error
Willy Tarreaubd741542010-03-16 18:46:54 +01005468
5469 # consider status 5xx only as errors
Willy Tarreau8f2a1e72011-01-06 16:36:10 +01005470 http-check expect ! rstatus ^5
Willy Tarreaubd741542010-03-16 18:46:54 +01005471
5472 # check that we have a correct hexadecimal tag before /html
Jarno Huuskonene5ae7022017-04-03 14:36:21 +03005473 http-check expect rstring <!--tag:[0-9a-f]*--></html>
Willy Tarreau0ba27502007-12-24 16:55:16 +01005474
Christopher Faulete5870d82020-04-15 11:32:03 +02005475 See also : "option httpchk", "http-check connect", "http-check disable-on-404"
Christopher Fauletc52ea4d2020-04-23 15:43:35 +02005476 and "http-check send".
Willy Tarreau2769aa02007-12-27 18:26:09 +01005477
5478
Christopher Faulet7c95f5f2020-05-06 15:06:34 +02005479http-check send [meth <method>] [{ uri <uri> | uri-lf <fmt> }>] [ver <version>]
Christopher Faulet574e7bd2020-05-06 15:38:58 +02005480 [hdr <name> <fmt>]* [{ body <string> | body-lf <fmt> }]
5481 [comment <msg>]
Christopher Faulet8acb1282020-04-09 08:44:06 +02005482 Add a possible list of headers and/or a body to the request sent during HTTP
5483 health checks.
5484 May be used in sections : defaults | frontend | listen | backend
5485 yes | no | yes | yes
5486 Arguments :
Christopher Faulet4f5c2e22020-04-23 15:22:33 +02005487 comment <msg> defines a message to report if the rule evaluation fails.
5488
Christopher Faulete5870d82020-04-15 11:32:03 +02005489 meth <method> is the optional HTTP method used with the requests. When not
5490 set, the "OPTIONS" method is used, as it generally requires
5491 low server processing and is easy to filter out from the
5492 logs. Any method may be used, though it is not recommended
5493 to invent non-standard ones.
5494
Christopher Faulet7c95f5f2020-05-06 15:06:34 +02005495 uri <uri> is optional and set the URI referenced in the HTTP requests
5496 to the string <uri>. It defaults to "/" which is accessible
5497 by default on almost any server, but may be changed to any
5498 other URI. Query strings are permitted.
5499
5500 uri-lf <fmt> is optional and set the URI referenced in the HTTP requests
5501 using the log-format string <fmt>. It defaults to "/" which
5502 is accessible by default on almost any server, but may be
5503 changed to any other URI. Query strings are permitted.
Christopher Faulet8acb1282020-04-09 08:44:06 +02005504
Christopher Faulet907701b2020-04-28 09:37:00 +02005505 ver <version> is the optional HTTP version string. It defaults to
Christopher Faulete5870d82020-04-15 11:32:03 +02005506 "HTTP/1.0" but some servers might behave incorrectly in HTTP
Daniel Corbett67a82712020-07-06 23:01:19 -04005507 1.0, so turning it to HTTP/1.1 may sometimes help. Note that
Christopher Faulete5870d82020-04-15 11:32:03 +02005508 the Host field is mandatory in HTTP/1.1, use "hdr" argument
5509 to add it.
5510
5511 hdr <name> <fmt> adds the HTTP header field whose name is specified in
5512 <name> and whose value is defined by <fmt>, which follows
5513 to the log-format rules.
5514
5515 body <string> add the body defined by <string> to the request sent during
5516 HTTP health checks. If defined, the "Content-Length" header
5517 is thus automatically added to the request.
Christopher Faulet8acb1282020-04-09 08:44:06 +02005518
Christopher Faulet574e7bd2020-05-06 15:38:58 +02005519 body-lf <fmt> add the body defined by the log-format string <fmt> to the
5520 request sent during HTTP health checks. If defined, the
5521 "Content-Length" header is thus automatically added to the
5522 request.
5523
Christopher Faulet8acb1282020-04-09 08:44:06 +02005524 In addition to the request line defined by the "option httpchk" directive,
5525 this one is the valid way to add some headers and optionally a body to the
5526 request sent during HTTP health checks. If a body is defined, the associate
Christopher Faulet9df910c2020-04-29 14:20:47 +02005527 "Content-Length" header is automatically added. Thus, this header or
5528 "Transfer-encoding" header should not be present in the request provided by
5529 "http-check send". If so, it will be ignored. The old trick consisting to add
5530 headers after the version string on the "option httpchk" line is now
Amaury Denoyelle6d975f02020-12-22 14:08:52 +01005531 deprecated.
Christopher Faulet8acb1282020-04-09 08:44:06 +02005532
Christopher Faulete5870d82020-04-15 11:32:03 +02005533 Also "http-check send" doesn't support HTTP keep-alive. Keep in mind that it
Amaury Denoyelle6d975f02020-12-22 14:08:52 +01005534 will automatically append a "Connection: close" header, unless a Connection
5535 header has already already been configured via a hdr entry.
Christopher Faulet9df910c2020-04-29 14:20:47 +02005536
5537 Note that the Host header and the request authority, when both defined, are
5538 automatically synchronized. It means when the HTTP request is sent, when a
5539 Host is inserted in the request, the request authority is accordingly
5540 updated. Thus, don't be surprised if the Host header value overwrites the
5541 configured request authority.
5542
5543 Note also for now, no Host header is automatically added in HTTP/1.1 or above
5544 requests. You should add it explicitly.
Christopher Faulete5870d82020-04-15 11:32:03 +02005545
Christopher Fauletc52ea4d2020-04-23 15:43:35 +02005546 See also : "option httpchk", "http-check send-state" and "http-check expect".
Christopher Faulet8acb1282020-04-09 08:44:06 +02005547
5548
Willy Tarreauef781042010-01-27 11:53:01 +01005549http-check send-state
5550 Enable emission of a state header with HTTP health checks
5551 May be used in sections : defaults | frontend | listen | backend
5552 yes | no | yes | yes
5553 Arguments : none
5554
5555 When this option is set, haproxy will systematically send a special header
5556 "X-Haproxy-Server-State" with a list of parameters indicating to each server
5557 how they are seen by haproxy. This can be used for instance when a server is
5558 manipulated without access to haproxy and the operator needs to know whether
5559 haproxy still sees it up or not, or if the server is the last one in a farm.
5560
5561 The header is composed of fields delimited by semi-colons, the first of which
5562 is a word ("UP", "DOWN", "NOLB"), possibly followed by a number of valid
5563 checks on the total number before transition, just as appears in the stats
5564 interface. Next headers are in the form "<variable>=<value>", indicating in
5565 no specific order some values available in the stats interface :
Joseph Lynch514061c2015-01-15 17:52:59 -08005566 - a variable "address", containing the address of the backend server.
5567 This corresponds to the <address> field in the server declaration. For
5568 unix domain sockets, it will read "unix".
5569
5570 - a variable "port", containing the port of the backend server. This
5571 corresponds to the <port> field in the server declaration. For unix
5572 domain sockets, it will read "unix".
5573
Willy Tarreauef781042010-01-27 11:53:01 +01005574 - a variable "name", containing the name of the backend followed by a slash
5575 ("/") then the name of the server. This can be used when a server is
5576 checked in multiple backends.
5577
5578 - a variable "node" containing the name of the haproxy node, as set in the
5579 global "node" variable, otherwise the system's hostname if unspecified.
5580
5581 - a variable "weight" indicating the weight of the server, a slash ("/")
5582 and the total weight of the farm (just counting usable servers). This
5583 helps to know if other servers are available to handle the load when this
5584 one fails.
5585
5586 - a variable "scur" indicating the current number of concurrent connections
5587 on the server, followed by a slash ("/") then the total number of
5588 connections on all servers of the same backend.
5589
5590 - a variable "qcur" indicating the current number of requests in the
5591 server's queue.
5592
5593 Example of a header received by the application server :
5594 >>> X-Haproxy-Server-State: UP 2/3; name=bck/srv2; node=lb1; weight=1/2; \
5595 scur=13/22; qcur=0
5596
Christopher Fauletc52ea4d2020-04-23 15:43:35 +02005597 See also : "option httpchk", "http-check disable-on-404" and
5598 "http-check send".
Willy Tarreauef781042010-01-27 11:53:01 +01005599
Christopher Faulete5870d82020-04-15 11:32:03 +02005600
5601http-check set-var(<var-name>) <expr>
Christopher Faulete5870d82020-04-15 11:32:03 +02005602 This operation sets the content of a variable. The variable is declared inline.
Christopher Faulete5870d82020-04-15 11:32:03 +02005603 May be used in sections: defaults | frontend | listen | backend
5604 yes | no | yes | yes
5605
Christopher Fauletc52ea4d2020-04-23 15:43:35 +02005606 Arguments :
Christopher Faulete5870d82020-04-15 11:32:03 +02005607 <var-name> The name of the variable starts with an indication about its
5608 scope. The scopes allowed for http-check are:
5609 "proc" : the variable is shared with the whole process.
5610 "sess" : the variable is shared with the tcp-check session.
5611 "check": the variable is declared for the lifetime of the tcp-check.
5612 This prefix is followed by a name. The separator is a '.'.
5613 The name may only contain characters 'a-z', 'A-Z', '0-9', '.',
5614 and '-'.
5615
5616 <expr> Is a sample-fetch expression potentially followed by converters.
5617
Christopher Fauletc52ea4d2020-04-23 15:43:35 +02005618 Examples :
5619 http-check set-var(check.port) int(1234)
Christopher Faulete5870d82020-04-15 11:32:03 +02005620
5621
5622http-check unset-var(<var-name>)
Christopher Faulete5870d82020-04-15 11:32:03 +02005623 Free a reference to a variable within its scope.
Christopher Faulete5870d82020-04-15 11:32:03 +02005624 May be used in sections: defaults | frontend | listen | backend
5625 yes | no | yes | yes
5626
Christopher Fauletc52ea4d2020-04-23 15:43:35 +02005627 Arguments :
Christopher Faulete5870d82020-04-15 11:32:03 +02005628 <var-name> The name of the variable starts with an indication about its
5629 scope. The scopes allowed for http-check are:
5630 "proc" : the variable is shared with the whole process.
5631 "sess" : the variable is shared with the tcp-check session.
5632 "check": the variable is declared for the lifetime of the tcp-check.
5633 This prefix is followed by a name. The separator is a '.'.
5634 The name may only contain characters 'a-z', 'A-Z', '0-9', '.',
5635 and '-'.
5636
Christopher Fauletc52ea4d2020-04-23 15:43:35 +02005637 Examples :
5638 http-check unset-var(check.port)
Christopher Faulete5870d82020-04-15 11:32:03 +02005639
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02005640
Christopher Faulet3b967c12020-05-15 15:47:44 +02005641http-error status <code> [content-type <type>]
5642 [ { default-errorfiles | errorfile <file> | errorfiles <name> |
5643 file <file> | lf-file <file> | string <str> | lf-string <fmt> } ]
5644 [ hdr <name> <fmt> ]*
5645 Defines a custom error message to use instead of errors generated by HAProxy.
5646 May be used in sections : defaults | frontend | listen | backend
5647 yes | yes | yes | yes
5648 Arguments :
Ilya Shipitsin11057a32020-06-21 21:18:27 +05005649 status <code> is the HTTP status code. It must be specified.
Christopher Faulet3b967c12020-05-15 15:47:44 +02005650 Currently, HAProxy is capable of generating codes
Anthonin Bonnefoy85048f82020-06-22 09:17:01 +02005651 200, 400, 401, 403, 404, 405, 407, 408, 410, 413, 425,
Christopher Faulete095f312020-12-07 11:22:24 +01005652 429, 500, 501, 502, 503, and 504.
Christopher Faulet3b967c12020-05-15 15:47:44 +02005653
5654 content-type <type> is the response content type, for instance
5655 "text/plain". This parameter is ignored and should be
5656 omitted when an errorfile is configured or when the
5657 payload is empty. Otherwise, it must be defined.
5658
5659 default-errorfiles Reset the previously defined error message for current
5660 proxy for the status <code>. If used on a backend, the
5661 frontend error message is used, if defined. If used on
5662 a frontend, the default error message is used.
5663
5664 errorfile <file> designates a file containing the full HTTP response.
5665 It is recommended to follow the common practice of
5666 appending ".http" to the filename so that people do
5667 not confuse the response with HTML error pages, and to
5668 use absolute paths, since files are read before any
5669 chroot is performed.
5670
5671 errorfiles <name> designates the http-errors section to use to import
5672 the error message with the status code <code>. If no
5673 such message is found, the proxy's error messages are
5674 considered.
5675
5676 file <file> specifies the file to use as response payload. If the
5677 file is not empty, its content-type must be set as
5678 argument to "content-type", otherwise, any
5679 "content-type" argument is ignored. <file> is
5680 considered as a raw string.
5681
5682 string <str> specifies the raw string to use as response payload.
5683 The content-type must always be set as argument to
5684 "content-type".
5685
5686 lf-file <file> specifies the file to use as response payload. If the
5687 file is not empty, its content-type must be set as
5688 argument to "content-type", otherwise, any
5689 "content-type" argument is ignored. <file> is
5690 evaluated as a log-format string.
5691
5692 lf-string <str> specifies the log-format string to use as response
5693 payload. The content-type must always be set as
5694 argument to "content-type".
5695
5696 hdr <name> <fmt> adds to the response the HTTP header field whose name
5697 is specified in <name> and whose value is defined by
5698 <fmt>, which follows to the log-format rules.
5699 This parameter is ignored if an errorfile is used.
5700
5701 This directive may be used instead of "errorfile", to define a custom error
5702 message. As "errorfile" directive, it is used for errors detected and
5703 returned by HAProxy. If an errorfile is defined, it is parsed when HAProxy
5704 starts and must be valid according to the HTTP standards. The generated
5705 response must not exceed the configured buffer size (BUFFSIZE), otherwise an
5706 internal error will be returned. Finally, if you consider to use some
5707 http-after-response rules to rewrite these errors, the reserved buffer space
5708 should be available (see "tune.maxrewrite").
5709
5710 The files are read at the same time as the configuration and kept in memory.
5711 For this reason, the errors continue to be returned even when the process is
5712 chrooted, and no file change is considered while the process is running.
5713
Christopher Fauletd5ac6de2020-12-02 08:40:14 +01005714 Note: 400/408/500 errors emitted in early stage of the request parsing are
5715 handled by the multiplexer at a lower level. No custom formatting is
5716 supported at this level. Thus only static error messages, defined with
5717 "errorfile" directive, are supported. However, this limitation only
5718 exists during the request headers parsing or between two transactions.
5719
Christopher Faulet3b967c12020-05-15 15:47:44 +02005720 See also : "errorfile", "errorfiles", "errorloc", "errorloc302",
5721 "errorloc303" and section 3.8 about http-errors.
5722
5723
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02005724http-request <action> [options...] [ { if | unless } <condition> ]
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01005725 Access control for Layer 7 requests
5726
5727 May be used in sections: defaults | frontend | listen | backend
5728 no | yes | yes | yes
5729
Willy Tarreau20b0de52012-12-24 15:45:22 +01005730 The http-request statement defines a set of rules which apply to layer 7
5731 processing. The rules are evaluated in their declaration order when they are
5732 met in a frontend, listen or backend section. Any rule may optionally be
5733 followed by an ACL-based condition, in which case it will only be evaluated
5734 if the condition is true.
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01005735
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02005736 The first keyword is the rule's action. The supported actions are described
5737 below.
Willy Tarreau20b0de52012-12-24 15:45:22 +01005738
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02005739 There is no limit to the number of http-request statements per instance.
Willy Tarreau20b0de52012-12-24 15:45:22 +01005740
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02005741 Example:
5742 acl nagios src 192.168.129.3
5743 acl local_net src 192.168.0.0/16
5744 acl auth_ok http_auth(L1)
Willy Tarreau20b0de52012-12-24 15:45:22 +01005745
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02005746 http-request allow if nagios
5747 http-request allow if local_net auth_ok
5748 http-request auth realm Gimme if local_net auth_ok
5749 http-request deny
Willy Tarreau81499eb2012-12-27 12:19:02 +01005750
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02005751 Example:
5752 acl key req.hdr(X-Add-Acl-Key) -m found
5753 acl add path /addacl
5754 acl del path /delacl
Willy Tarreau20b0de52012-12-24 15:45:22 +01005755
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02005756 acl myhost hdr(Host) -f myhost.lst
Willy Tarreau20b0de52012-12-24 15:45:22 +01005757
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02005758 http-request add-acl(myhost.lst) %[req.hdr(X-Add-Acl-Key)] if key add
5759 http-request del-acl(myhost.lst) %[req.hdr(X-Add-Acl-Key)] if key del
Thierry FOURNIERdad3d1d2014-04-22 18:07:25 +02005760
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02005761 Example:
5762 acl value req.hdr(X-Value) -m found
5763 acl setmap path /setmap
5764 acl delmap path /delmap
Sasha Pachev218f0642014-06-16 12:05:59 -06005765
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02005766 use_backend bk_appli if { hdr(Host),map_str(map.lst) -m found }
Sasha Pachev218f0642014-06-16 12:05:59 -06005767
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02005768 http-request set-map(map.lst) %[src] %[req.hdr(X-Value)] if setmap value
5769 http-request del-map(map.lst) %[src] if delmap
Sasha Pachev218f0642014-06-16 12:05:59 -06005770
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02005771 See also : "stats http-request", section 3.4 about userlists and section 7
5772 about ACL usage.
Sasha Pachev218f0642014-06-16 12:05:59 -06005773
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02005774http-request add-acl(<file-name>) <key fmt> [ { if | unless } <condition> ]
Sasha Pachev218f0642014-06-16 12:05:59 -06005775
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02005776 This is used to add a new entry into an ACL. The ACL must be loaded from a
5777 file (even a dummy empty file). The file name of the ACL to be updated is
5778 passed between parentheses. It takes one argument: <key fmt>, which follows
5779 log-format rules, to collect content of the new entry. It performs a lookup
5780 in the ACL before insertion, to avoid duplicated (or more) values. This
5781 lookup is done by a linear search and can be expensive with large lists!
5782 It is the equivalent of the "add acl" command from the stats socket, but can
5783 be triggered by an HTTP request.
Sasha Pachev218f0642014-06-16 12:05:59 -06005784
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02005785http-request add-header <name> <fmt> [ { if | unless } <condition> ]
Sasha Pachev218f0642014-06-16 12:05:59 -06005786
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02005787 This appends an HTTP header field whose name is specified in <name> and
5788 whose value is defined by <fmt> which follows the log-format rules (see
5789 Custom Log Format in section 8.2.4). This is particularly useful to pass
5790 connection-specific information to the server (e.g. the client's SSL
5791 certificate), or to combine several headers into one. This rule is not
5792 final, so it is possible to add other similar rules. Note that header
5793 addition is performed immediately, so one rule might reuse the resulting
5794 header from a previous rule.
Sasha Pachev218f0642014-06-16 12:05:59 -06005795
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02005796http-request allow [ { if | unless } <condition> ]
Sasha Pachev218f0642014-06-16 12:05:59 -06005797
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02005798 This stops the evaluation of the rules and lets the request pass the check.
5799 No further "http-request" rules are evaluated.
Sasha Pachev218f0642014-06-16 12:05:59 -06005800
Sasha Pachev218f0642014-06-16 12:05:59 -06005801
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02005802http-request auth [realm <realm>] [ { if | unless } <condition> ]
Sasha Pachev218f0642014-06-16 12:05:59 -06005803
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02005804 This stops the evaluation of the rules and immediately responds with an
5805 HTTP 401 or 407 error code to invite the user to present a valid user name
5806 and password. No further "http-request" rules are evaluated. An optional
5807 "realm" parameter is supported, it sets the authentication realm that is
5808 returned with the response (typically the application's name).
Sasha Pachev218f0642014-06-16 12:05:59 -06005809
Christopher Faulet612f2ea2020-05-27 09:57:28 +02005810 The corresponding proxy's error message is used. It may be customized using
5811 an "errorfile" or an "http-error" directive. For 401 responses, all
5812 occurrences of the WWW-Authenticate header are removed and replaced by a new
5813 one with a basic authentication challenge for realm "<realm>". For 407
5814 responses, the same is done on the Proxy-Authenticate header. If the error
5815 message must not be altered, consider to use "http-request return" rule
5816 instead.
5817
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02005818 Example:
5819 acl auth_ok http_auth_group(L1) G1
5820 http-request auth unless auth_ok
Sasha Pachev218f0642014-06-16 12:05:59 -06005821
Jarno Huuskonen251a6b72019-01-04 14:05:02 +02005822http-request cache-use <name> [ { if | unless } <condition> ]
Sasha Pachev218f0642014-06-16 12:05:59 -06005823
Christopher Faulet87f1f3d2019-07-18 14:51:20 +02005824 See section 6.2 about cache setup.
Willy Tarreaua0dc23f2015-01-22 20:46:11 +01005825
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02005826http-request capture <sample> [ len <length> | id <id> ]
5827 [ { if | unless } <condition> ]
Willy Tarreaua0dc23f2015-01-22 20:46:11 +01005828
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02005829 This captures sample expression <sample> from the request buffer, and
5830 converts it to a string of at most <len> characters. The resulting string is
5831 stored into the next request "capture" slot, so it will possibly appear next
5832 to some captured HTTP headers. It will then automatically appear in the logs,
5833 and it will be possible to extract it using sample fetch rules to feed it
5834 into headers or anything. The length should be limited given that this size
5835 will be allocated for each capture during the whole session life.
5836 Please check section 7.3 (Fetching samples) and "capture request header" for
5837 more information.
Willy Tarreaua0dc23f2015-01-22 20:46:11 +01005838
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02005839 If the keyword "id" is used instead of "len", the action tries to store the
5840 captured string in a previously declared capture slot. This is useful to run
5841 captures in backends. The slot id can be declared by a previous directive
Baptiste Assmann19a69b32020-01-16 14:34:22 +01005842 "http-request capture" or with the "declare capture" keyword.
5843
5844 When using this action in a backend, double check that the relevant
5845 frontend(s) have the required capture slots otherwise, this rule will be
5846 ignored at run time. This can't be detected at configuration parsing time
5847 due to HAProxy's ability to dynamically resolve backend name at runtime.
Willy Tarreaua0dc23f2015-01-22 20:46:11 +01005848
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02005849http-request del-acl(<file-name>) <key fmt> [ { if | unless } <condition> ]
Willy Tarreaua0dc23f2015-01-22 20:46:11 +01005850
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02005851 This is used to delete an entry from an ACL. The ACL must be loaded from a
5852 file (even a dummy empty file). The file name of the ACL to be updated is
5853 passed between parentheses. It takes one argument: <key fmt>, which follows
5854 log-format rules, to collect content of the entry to delete.
5855 It is the equivalent of the "del acl" command from the stats socket, but can
5856 be triggered by an HTTP request.
Willy Tarreaua0dc23f2015-01-22 20:46:11 +01005857
Maciej Zdebebdd4c52020-11-20 13:58:48 +00005858http-request del-header <name> [ -m <meth> ] [ { if | unless } <condition> ]
Willy Tarreauf4c43c12013-06-11 17:01:13 +02005859
Maciej Zdebebdd4c52020-11-20 13:58:48 +00005860 This removes all HTTP header fields whose name is specified in <name>. <meth>
5861 is the matching method, applied on the header name. Supported matching methods
5862 are "str" (exact match), "beg" (prefix match), "end" (suffix match), "sub"
5863 (substring match) and "reg" (regex match). If not specified, exact matching
5864 method is used.
Willy Tarreau9a355ec2013-06-11 17:45:46 +02005865
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02005866http-request del-map(<file-name>) <key fmt> [ { if | unless } <condition> ]
Willy Tarreau42cf39e2013-06-11 18:51:32 +02005867
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02005868 This is used to delete an entry from a MAP. The MAP must be loaded from a
5869 file (even a dummy empty file). The file name of the MAP to be updated is
5870 passed between parentheses. It takes one argument: <key fmt>, which follows
5871 log-format rules, to collect content of the entry to delete.
5872 It takes one argument: "file name" It is the equivalent of the "del map"
5873 command from the stats socket, but can be triggered by an HTTP request.
Willy Tarreau51347ed2013-06-11 19:34:13 +02005874
Christopher Faulet5cb513a2020-05-13 17:56:56 +02005875http-request deny [deny_status <status>] [ { if | unless } <condition> ]
5876http-request deny [ { status | deny_status } <code>] [content-type <type>]
5877 [ { default-errorfiles | errorfile <file> | errorfiles <name> |
5878 file <file> | lf-file <file> | string <str> | lf-string <fmt> } ]
5879 [ hdr <name> <fmt> ]*
5880 [ { if | unless } <condition> ]
Patrick Hemmer268a7072018-05-11 12:52:31 -04005881
Christopher Faulet5cb513a2020-05-13 17:56:56 +02005882 This stops the evaluation of the rules and immediately rejects the request.
5883 By default an HTTP 403 error is returned. But the response may be customized
5884 using same syntax than "http-request return" rules. Thus, see "http-request
Ilya Shipitsin11057a32020-06-21 21:18:27 +05005885 return" for details. For compatibility purpose, when no argument is defined,
Christopher Faulet5cb513a2020-05-13 17:56:56 +02005886 or only "deny_status", the argument "default-errorfiles" is implied. It means
5887 "http-request deny [deny_status <status>]" is an alias of
5888 "http-request deny [status <status>] default-errorfiles".
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02005889 No further "http-request" rules are evaluated.
Christopher Faulet5cb513a2020-05-13 17:56:56 +02005890 See also "http-request return".
Patrick Hemmer268a7072018-05-11 12:52:31 -04005891
Olivier Houchard602bf7d2019-05-10 13:59:15 +02005892http-request disable-l7-retry [ { if | unless } <condition> ]
5893 This disables any attempt to retry the request if it fails for any other
5894 reason than a connection failure. This can be useful for example to make
5895 sure POST requests aren't retried on failure.
5896
Baptiste Assmann333939c2019-01-21 08:34:50 +01005897http-request do-resolve(<var>,<resolvers>,[ipv4,ipv6]) <expr> :
5898
5899 This action performs a DNS resolution of the output of <expr> and stores
5900 the result in the variable <var>. It uses the DNS resolvers section
5901 pointed by <resolvers>.
5902 It is possible to choose a resolution preference using the optional
5903 arguments 'ipv4' or 'ipv6'.
5904 When performing the DNS resolution, the client side connection is on
5905 pause waiting till the end of the resolution.
5906 If an IP address can be found, it is stored into <var>. If any kind of
5907 error occurs, then <var> is not set.
5908 One can use this action to discover a server IP address at run time and
5909 based on information found in the request (IE a Host header).
5910 If this action is used to find the server's IP address (using the
5911 "set-dst" action), then the server IP address in the backend must be set
5912 to 0.0.0.0.
5913
5914 Example:
5915 resolvers mydns
5916 nameserver local 127.0.0.53:53
5917 nameserver google 8.8.8.8:53
5918 timeout retry 1s
5919 hold valid 10s
5920 hold nx 3s
5921 hold other 3s
5922 hold obsolete 0s
5923 accepted_payload_size 8192
5924
5925 frontend fe
5926 bind 10.42.0.1:80
5927 http-request do-resolve(txn.myip,mydns,ipv4) hdr(Host),lower
5928 http-request capture var(txn.myip) len 40
5929
5930 # return 503 when the variable is not set,
5931 # which mean DNS resolution error
5932 use_backend b_503 unless { var(txn.myip) -m found }
5933
5934 default_backend be
5935
5936 backend b_503
5937 # dummy backend used to return 503.
5938 # one can use the errorfile directive to send a nice
5939 # 503 error page to end users
5940
5941 backend be
5942 # rule to prevent HAProxy from reconnecting to services
5943 # on the local network (forged DNS name used to scan the network)
5944 http-request deny if { var(txn.myip) -m ip 127.0.0.0/8 10.0.0.0/8 }
5945 http-request set-dst var(txn.myip)
5946 server clear 0.0.0.0:0
5947
5948 NOTE: Don't forget to set the "protection" rules to ensure HAProxy won't
5949 be used to scan the network or worst won't loop over itself...
5950
Frédéric Lécaille06f5b642018-11-12 11:01:10 +01005951http-request early-hint <name> <fmt> [ { if | unless } <condition> ]
5952
5953 This is used to build an HTTP 103 Early Hints response prior to any other one.
5954 This appends an HTTP header field to this response whose name is specified in
5955 <name> and whose value is defined by <fmt> which follows the log-format rules
5956 (see Custom Log Format in section 8.2.4). This is particularly useful to pass
Frédéric Lécaille3aac1062018-11-13 09:42:13 +01005957 to the client some Link headers to preload resources required to render the
5958 HTML documents.
Frédéric Lécaille06f5b642018-11-12 11:01:10 +01005959
5960 See RFC 8297 for more information.
5961
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02005962http-request redirect <rule> [ { if | unless } <condition> ]
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02005963
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02005964 This performs an HTTP redirection based on a redirect rule. This is exactly
5965 the same as the "redirect" statement except that it inserts a redirect rule
5966 which can be processed in the middle of other "http-request" rules and that
5967 these rules use the "log-format" strings. See the "redirect" keyword for the
5968 rule's syntax.
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02005969
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02005970http-request reject [ { if | unless } <condition> ]
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02005971
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02005972 This stops the evaluation of the rules and immediately closes the connection
5973 without sending any response. It acts similarly to the
5974 "tcp-request content reject" rules. It can be useful to force an immediate
5975 connection closure on HTTP/2 connections.
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02005976
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02005977http-request replace-header <name> <match-regex> <replace-fmt>
5978 [ { if | unless } <condition> ]
Willy Tarreaua9083d02015-05-08 15:27:59 +02005979
Ilya Shipitsin8525fd92020-02-29 12:34:59 +05005980 This matches the value of all occurrences of header field <name> against
Tim Duesterhus2252beb2019-10-29 00:05:13 +01005981 <match-regex>. Matching is performed case-sensitively. Matching values are
5982 completely replaced by <replace-fmt>. Format characters are allowed in
5983 <replace-fmt> and work like <fmt> arguments in "http-request add-header".
5984 Standard back-references using the backslash ('\') followed by a number are
5985 supported.
Thierry FOURNIER82bf70d2015-05-26 17:58:29 +02005986
Tim Duesterhus2252beb2019-10-29 00:05:13 +01005987 This action acts on whole header lines, regardless of the number of values
5988 they may contain. Thus it is well-suited to process headers naturally
5989 containing commas in their value, such as If-Modified-Since. Headers that
5990 contain a comma-separated list of values, such as Accept, should be processed
5991 using "http-request replace-value".
William Lallemand86d0df02017-11-24 21:36:45 +01005992
Tim Duesterhus2252beb2019-10-29 00:05:13 +01005993 Example:
5994 http-request replace-header Cookie foo=([^;]*);(.*) foo=\1;ip=%bi;\2
5995
5996 # applied to:
5997 Cookie: foo=foobar; expires=Tue, 14-Jun-2016 01:40:45 GMT;
5998
5999 # outputs:
6000 Cookie: foo=foobar;ip=192.168.1.20; expires=Tue, 14-Jun-2016 01:40:45 GMT;
6001
6002 # assuming the backend IP is 192.168.1.20
Willy Tarreau09448f72014-06-25 18:12:15 +02006003
Tim Duesterhus2252beb2019-10-29 00:05:13 +01006004 http-request replace-header User-Agent curl foo
6005
6006 # applied to:
6007 User-Agent: curl/7.47.0
Willy Tarreau09448f72014-06-25 18:12:15 +02006008
Tim Duesterhus2252beb2019-10-29 00:05:13 +01006009 # outputs:
6010 User-Agent: foo
Willy Tarreau09448f72014-06-25 18:12:15 +02006011
Willy Tarreau262c3f12019-12-17 06:52:51 +01006012http-request replace-path <match-regex> <replace-fmt>
6013 [ { if | unless } <condition> ]
6014
6015 This works like "replace-header" except that it works on the request's path
6016 component instead of a header. The path component starts at the first '/'
Christopher Faulet82c83322020-09-02 14:16:59 +02006017 after an optional scheme+authority and ends before the question mark. Thus,
6018 the replacement does not modify the scheme, the authority and the
6019 query-string.
Willy Tarreau262c3f12019-12-17 06:52:51 +01006020
6021 It is worth noting that regular expressions may be more expensive to evaluate
6022 than certain ACLs, so rare replacements may benefit from a condition to avoid
6023 performing the evaluation at all if it does not match.
6024
6025 Example:
6026 # prefix /foo : turn /bar?q=1 into /foo/bar?q=1 :
6027 http-request replace-path (.*) /foo\1
6028
Willy Tarreau262c3f12019-12-17 06:52:51 +01006029 # strip /foo : turn /foo/bar?q=1 into /bar?q=1
6030 http-request replace-path /foo/(.*) /\1
6031 # or more efficient if only some requests match :
6032 http-request replace-path /foo/(.*) /\1 if { url_beg /foo/ }
6033
Christopher Faulet312294f2020-09-02 17:17:44 +02006034http-request replace-pathq <match-regex> <replace-fmt>
6035 [ { if | unless } <condition> ]
6036
6037 This does the same as "http-request replace-path" except that the path
6038 contains the query-string if any is present. Thus, the path and the
6039 query-string are replaced.
6040
6041 Example:
6042 # suffix /foo : turn /bar?q=1 into /bar/foo?q=1 :
6043 http-request replace-pathq ([^?]*)(\?(.*))? \1/foo\2
6044
Willy Tarreau33810222019-06-12 17:44:02 +02006045http-request replace-uri <match-regex> <replace-fmt>
6046 [ { if | unless } <condition> ]
6047
Tim Duesterhus2252beb2019-10-29 00:05:13 +01006048 This works like "replace-header" except that it works on the request's URI part
6049 instead of a header. The URI part may contain an optional scheme, authority or
6050 query string. These are considered to be part of the value that is matched
6051 against.
6052
6053 It is worth noting that regular expressions may be more expensive to evaluate
6054 than certain ACLs, so rare replacements may benefit from a condition to avoid
6055 performing the evaluation at all if it does not match.
Willy Tarreau33810222019-06-12 17:44:02 +02006056
Willy Tarreau62b59132019-12-17 06:51:20 +01006057 IMPORTANT NOTE: historically in HTTP/1.x, the vast majority of requests sent
6058 by browsers use the "origin form", which differs from the "absolute form" in
6059 that they do not contain a scheme nor authority in the URI portion. Mostly
6060 only requests sent to proxies, those forged by hand and some emitted by
6061 certain applications use the absolute form. As such, "replace-uri" usually
6062 works fine most of the time in HTTP/1.x with rules starting with a "/". But
6063 with HTTP/2, clients are encouraged to send absolute URIs only, which look
6064 like the ones HTTP/1 clients use to talk to proxies. Such partial replace-uri
6065 rules may then fail in HTTP/2 when they work in HTTP/1. Either the rules need
Willy Tarreau262c3f12019-12-17 06:52:51 +01006066 to be adapted to optionally match a scheme and authority, or replace-path
6067 should be used.
Willy Tarreau33810222019-06-12 17:44:02 +02006068
Willy Tarreau62b59132019-12-17 06:51:20 +01006069 Example:
6070 # rewrite all "http" absolute requests to "https":
6071 http-request replace-uri ^http://(.*) https://\1
Willy Tarreau33810222019-06-12 17:44:02 +02006072
Willy Tarreau62b59132019-12-17 06:51:20 +01006073 # prefix /foo : turn /bar?q=1 into /foo/bar?q=1 :
6074 http-request replace-uri ([^/:]*://[^/]*)?(.*) \1/foo\2
Willy Tarreau33810222019-06-12 17:44:02 +02006075
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006076http-request replace-value <name> <match-regex> <replace-fmt>
6077 [ { if | unless } <condition> ]
Willy Tarreau09448f72014-06-25 18:12:15 +02006078
Tim Duesterhus2252beb2019-10-29 00:05:13 +01006079 This works like "replace-header" except that it matches the regex against
6080 every comma-delimited value of the header field <name> instead of the
6081 entire header. This is suited for all headers which are allowed to carry
6082 more than one value. An example could be the Accept header.
Willy Tarreau09448f72014-06-25 18:12:15 +02006083
Tim Duesterhus2252beb2019-10-29 00:05:13 +01006084 Example:
6085 http-request replace-value X-Forwarded-For ^192\.168\.(.*)$ 172.16.\1
Thierry FOURNIER236657b2015-08-19 08:25:14 +02006086
Tim Duesterhus2252beb2019-10-29 00:05:13 +01006087 # applied to:
6088 X-Forwarded-For: 192.168.10.1, 192.168.13.24, 10.0.0.37
Thierry FOURNIERe0627bd2015-08-04 08:20:33 +02006089
Tim Duesterhus2252beb2019-10-29 00:05:13 +01006090 # outputs:
6091 X-Forwarded-For: 172.16.10.1, 172.16.13.24, 10.0.0.37
Frédéric Lécaille6778b272018-01-29 15:22:53 +01006092
Christopher Faulet24231ab2020-01-24 17:44:23 +01006093http-request return [status <code>] [content-type <type>]
6094 [ { default-errorfiles | errorfile <file> | errorfiles <name> |
6095 file <file> | lf-file <file> | string <str> | lf-string <fmt> } ]
Christopher Faulet4a2c1422020-01-31 17:36:01 +01006096 [ hdr <name> <fmt> ]*
Christopher Faulet24231ab2020-01-24 17:44:23 +01006097 [ { if | unless } <condition> ]
6098
Ilya Shipitsin8525fd92020-02-29 12:34:59 +05006099 This stops the evaluation of the rules and immediately returns a response. The
Christopher Faulet24231ab2020-01-24 17:44:23 +01006100 default status code used for the response is 200. It can be optionally
6101 specified as an arguments to "status". The response content-type may also be
Daniel Corbett67a82712020-07-06 23:01:19 -04006102 specified as an argument to "content-type". Finally the response itself may
Sébastien Grossab877122020-10-08 10:06:03 +02006103 be defined. It can be a full HTTP response specifying the errorfile to use,
Ilya Shipitsin8525fd92020-02-29 12:34:59 +05006104 or the response payload specifying the file or the string to use. These rules
Christopher Faulet24231ab2020-01-24 17:44:23 +01006105 are followed to create the response :
6106
6107 * If neither the errorfile nor the payload to use is defined, a dummy
6108 response is returned. Only the "status" argument is considered. It can be
6109 any code in the range [200, 599]. The "content-type" argument, if any, is
6110 ignored.
6111
6112 * If "default-errorfiles" argument is set, the proxy's errorfiles are
6113 considered. If the "status" argument is defined, it must be one of the
Daniel Corbett67a82712020-07-06 23:01:19 -04006114 status code handled by haproxy (200, 400, 403, 404, 405, 408, 410, 413,
Christopher Faulete095f312020-12-07 11:22:24 +01006115 425, 429, 500, 501, 502, 503, and 504). The "content-type" argument, if
6116 any, is ignored.
Christopher Faulet24231ab2020-01-24 17:44:23 +01006117
6118 * If a specific errorfile is defined, with an "errorfile" argument, the
6119 corresponding file, containing a full HTTP response, is returned. Only the
6120 "status" argument is considered. It must be one of the status code handled
Christopher Faulete095f312020-12-07 11:22:24 +01006121 by haproxy (200, 400, 403, 404, 405, 408, 410, 413, 425, 429, 500, 501,
6122 502, 503, and 504). The "content-type" argument, if any, is ignored.
Christopher Faulet24231ab2020-01-24 17:44:23 +01006123
6124 * If an http-errors section is defined, with an "errorfiles" argument, the
6125 corresponding file in the specified http-errors section, containing a full
6126 HTTP response, is returned. Only the "status" argument is considered. It
Daniel Corbett67a82712020-07-06 23:01:19 -04006127 must be one of the status code handled by haproxy (200, 400, 403, 404, 405,
Christopher Faulete095f312020-12-07 11:22:24 +01006128 408, 410, 413, 425, 429, 500, 501, 502, 503, and 504). The "content-type"
Anthonin Bonnefoy85048f82020-06-22 09:17:01 +02006129 argument, if any, is ignored.
Christopher Faulet24231ab2020-01-24 17:44:23 +01006130
6131 * If a "file" or a "lf-file" argument is specified, the file's content is
6132 used as the response payload. If the file is not empty, its content-type
6133 must be set as argument to "content-type". Otherwise, any "content-type"
6134 argument is ignored. With a "lf-file" argument, the file's content is
6135 evaluated as a log-format string. With a "file" argument, it is considered
6136 as a raw content.
6137
6138 * If a "string" or "lf-string" argument is specified, the defined string is
6139 used as the response payload. The content-type must always be set as
6140 argument to "content-type". With a "lf-string" argument, the string is
6141 evaluated as a log-format string. With a "string" argument, it is
6142 considered as a raw string.
6143
Sébastien Grossab877122020-10-08 10:06:03 +02006144 When the response is not based on an errorfile, it is possible to append HTTP
Christopher Faulet4a2c1422020-01-31 17:36:01 +01006145 header fields to the response using "hdr" arguments. Otherwise, all "hdr"
6146 arguments are ignored. For each one, the header name is specified in <name>
6147 and its value is defined by <fmt> which follows the log-format rules.
6148
Christopher Faulet24231ab2020-01-24 17:44:23 +01006149 Note that the generated response must be smaller than a buffer. And to avoid
6150 any warning, when an errorfile or a raw file is loaded, the buffer space
Sébastien Grossab877122020-10-08 10:06:03 +02006151 reserved for the headers rewriting should also be free.
Christopher Faulet24231ab2020-01-24 17:44:23 +01006152
6153 No further "http-request" rules are evaluated.
6154
6155 Example:
Daniel Corbett67a82712020-07-06 23:01:19 -04006156 http-request return errorfile /etc/haproxy/errorfiles/200.http \
Christopher Faulet24231ab2020-01-24 17:44:23 +01006157 if { path /ping }
6158
6159 http-request return content-type image/x-icon file /var/www/favicon.ico \
6160 if { path /favicon.ico }
6161
6162 http-request return status 403 content-type text/plain \
6163 lf-string "Access denied. IP %[src] is blacklisted." \
6164 if { src -f /etc/haproxy/blacklist.lst }
6165
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006166http-request sc-inc-gpc0(<sc-id>) [ { if | unless } <condition> ]
6167http-request sc-inc-gpc1(<sc-id>) [ { if | unless } <condition> ]
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02006168
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006169 This actions increments the GPC0 or GPC1 counter according with the sticky
6170 counter designated by <sc-id>. If an error occurs, this action silently fails
6171 and the actions evaluation continues.
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02006172
Cédric Dufour0d7712d2019-11-06 18:38:53 +01006173http-request sc-set-gpt0(<sc-id>) { <int> | <expr> }
6174 [ { if | unless } <condition> ]
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02006175
Cédric Dufour0d7712d2019-11-06 18:38:53 +01006176 This action sets the 32-bit unsigned GPT0 tag according to the sticky counter
6177 designated by <sc-id> and the value of <int>/<expr>. The expected result is a
6178 boolean. If an error occurs, this action silently fails and the actions
6179 evaluation continues.
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02006180
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006181http-request set-dst <expr> [ { if | unless } <condition> ]
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02006182
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006183 This is used to set the destination IP address to the value of specified
6184 expression. Useful when a proxy in front of HAProxy rewrites destination IP,
6185 but provides the correct IP in a HTTP header; or you want to mask the IP for
6186 privacy. If you want to connect to the new address/port, use '0.0.0.0:0' as a
6187 server address in the backend.
Christopher Faulet85d79c92016-11-09 16:54:56 +01006188
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006189 Arguments:
6190 <expr> Is a standard HAProxy expression formed by a sample-fetch followed
6191 by some converters.
Christopher Faulet85d79c92016-11-09 16:54:56 +01006192
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006193 Example:
6194 http-request set-dst hdr(x-dst)
6195 http-request set-dst dst,ipmask(24)
Christopher Faulet85d79c92016-11-09 16:54:56 +01006196
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006197 When possible, set-dst preserves the original destination port as long as the
6198 address family allows it, otherwise the destination port is set to 0.
Adis Nezirovic2fbcafc2015-07-06 15:44:30 +02006199
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006200http-request set-dst-port <expr> [ { if | unless } <condition> ]
Adis Nezirovic2fbcafc2015-07-06 15:44:30 +02006201
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006202 This is used to set the destination port address to the value of specified
6203 expression. If you want to connect to the new address/port, use '0.0.0.0:0'
6204 as a server address in the backend.
Adis Nezirovic2fbcafc2015-07-06 15:44:30 +02006205
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006206 Arguments:
6207 <expr> Is a standard HAProxy expression formed by a sample-fetch
6208 followed by some converters.
Adis Nezirovic2fbcafc2015-07-06 15:44:30 +02006209
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006210 Example:
6211 http-request set-dst-port hdr(x-port)
6212 http-request set-dst-port int(4000)
Adis Nezirovic2fbcafc2015-07-06 15:44:30 +02006213
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006214 When possible, set-dst-port preserves the original destination address as
6215 long as the address family supports a port, otherwise it forces the
6216 destination address to IPv4 "0.0.0.0" before rewriting the port.
William Lallemand44be6402016-05-25 01:51:35 +02006217
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006218http-request set-header <name> <fmt> [ { if | unless } <condition> ]
William Lallemand44be6402016-05-25 01:51:35 +02006219
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006220 This does the same as "http-request add-header" except that the header name
6221 is first removed if it existed. This is useful when passing security
6222 information to the server, where the header must not be manipulated by
6223 external users. Note that the new value is computed before the removal so it
6224 is possible to concatenate a value to an existing header.
William Lallemand44be6402016-05-25 01:51:35 +02006225
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006226 Example:
6227 http-request set-header X-Haproxy-Current-Date %T
6228 http-request set-header X-SSL %[ssl_fc]
6229 http-request set-header X-SSL-Session_ID %[ssl_fc_session_id,hex]
6230 http-request set-header X-SSL-Client-Verify %[ssl_c_verify]
6231 http-request set-header X-SSL-Client-DN %{+Q}[ssl_c_s_dn]
6232 http-request set-header X-SSL-Client-CN %{+Q}[ssl_c_s_dn(cn)]
6233 http-request set-header X-SSL-Issuer %{+Q}[ssl_c_i_dn]
6234 http-request set-header X-SSL-Client-NotBefore %{+Q}[ssl_c_notbefore]
6235 http-request set-header X-SSL-Client-NotAfter %{+Q}[ssl_c_notafter]
William Lallemand44be6402016-05-25 01:51:35 +02006236
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006237http-request set-log-level <level> [ { if | unless } <condition> ]
William Lallemand44be6402016-05-25 01:51:35 +02006238
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006239 This is used to change the log level of the current request when a certain
6240 condition is met. Valid levels are the 8 syslog levels (see the "log"
6241 keyword) plus the special level "silent" which disables logging for this
6242 request. This rule is not final so the last matching rule wins. This rule
6243 can be useful to disable health checks coming from another equipment.
William Lallemand13e9b0c2016-05-25 02:34:07 +02006244
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006245http-request set-map(<file-name>) <key fmt> <value fmt>
6246 [ { if | unless } <condition> ]
William Lallemand13e9b0c2016-05-25 02:34:07 +02006247
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006248 This is used to add a new entry into a MAP. The MAP must be loaded from a
6249 file (even a dummy empty file). The file name of the MAP to be updated is
6250 passed between parentheses. It takes 2 arguments: <key fmt>, which follows
6251 log-format rules, used to collect MAP key, and <value fmt>, which follows
6252 log-format rules, used to collect content for the new entry.
6253 It performs a lookup in the MAP before insertion, to avoid duplicated (or
6254 more) values. This lookup is done by a linear search and can be expensive
6255 with large lists! It is the equivalent of the "set map" command from the
6256 stats socket, but can be triggered by an HTTP request.
William Lallemand13e9b0c2016-05-25 02:34:07 +02006257
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006258http-request set-mark <mark> [ { if | unless } <condition> ]
William Lallemand13e9b0c2016-05-25 02:34:07 +02006259
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006260 This is used to set the Netfilter MARK on all packets sent to the client to
6261 the value passed in <mark> on platforms which support it. This value is an
6262 unsigned 32 bit value which can be matched by netfilter and by the routing
6263 table. It can be expressed both in decimal or hexadecimal format (prefixed by
6264 "0x"). This can be useful to force certain packets to take a different route
6265 (for example a cheaper network path for bulk downloads). This works on Linux
6266 kernels 2.6.32 and above and requires admin privileges.
Willy Tarreau00005ce2016-10-21 15:07:45 +02006267
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006268http-request set-method <fmt> [ { if | unless } <condition> ]
William Lallemand13e9b0c2016-05-25 02:34:07 +02006269
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006270 This rewrites the request method with the result of the evaluation of format
6271 string <fmt>. There should be very few valid reasons for having to do so as
6272 this is more likely to break something than to fix it.
William Lallemand13e9b0c2016-05-25 02:34:07 +02006273
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006274http-request set-nice <nice> [ { if | unless } <condition> ]
William Lallemand13e9b0c2016-05-25 02:34:07 +02006275
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006276 This sets the "nice" factor of the current request being processed. It only
6277 has effect against the other requests being processed at the same time.
6278 The default value is 0, unless altered by the "nice" setting on the "bind"
6279 line. The accepted range is -1024..1024. The higher the value, the nicest
6280 the request will be. Lower values will make the request more important than
6281 other ones. This can be useful to improve the speed of some requests, or
6282 lower the priority of non-important requests. Using this setting without
6283 prior experimentation can cause some major slowdown.
William Lallemand13e9b0c2016-05-25 02:34:07 +02006284
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006285http-request set-path <fmt> [ { if | unless } <condition> ]
Willy Tarreau00005ce2016-10-21 15:07:45 +02006286
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006287 This rewrites the request path with the result of the evaluation of format
6288 string <fmt>. The query string, if any, is left intact. If a scheme and
6289 authority is found before the path, they are left intact as well. If the
6290 request doesn't have a path ("*"), this one is replaced with the format.
6291 This can be used to prepend a directory component in front of a path for
6292 example. See also "http-request set-query" and "http-request set-uri".
Willy Tarreau2d392c22015-08-24 01:43:45 +02006293
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006294 Example :
6295 # prepend the host name before the path
6296 http-request set-path /%[hdr(host)]%[path]
Christopher Faulet76c09ef2017-09-21 11:03:52 +02006297
Christopher Faulet312294f2020-09-02 17:17:44 +02006298http-request set-pathq <fmt> [ { if | unless } <condition> ]
6299
6300 This does the same as "http-request set-path" except that the query-string is
6301 also rewritten. It may be used to remove the query-string, including the
6302 question mark (it is not possible using "http-request set-query").
6303
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006304http-request set-priority-class <expr> [ { if | unless } <condition> ]
Olivier Houchardccaa7de2017-10-02 11:51:03 +02006305
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006306 This is used to set the queue priority class of the current request.
6307 The value must be a sample expression which converts to an integer in the
6308 range -2047..2047. Results outside this range will be truncated.
6309 The priority class determines the order in which queued requests are
6310 processed. Lower values have higher priority.
Christopher Faulet76c09ef2017-09-21 11:03:52 +02006311
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006312http-request set-priority-offset <expr> [ { if | unless } <condition> ]
Christopher Faulet76c09ef2017-09-21 11:03:52 +02006313
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006314 This is used to set the queue priority timestamp offset of the current
6315 request. The value must be a sample expression which converts to an integer
6316 in the range -524287..524287. Results outside this range will be truncated.
6317 When a request is queued, it is ordered first by the priority class, then by
6318 the current timestamp adjusted by the given offset in milliseconds. Lower
6319 values have higher priority.
6320 Note that the resulting timestamp is is only tracked with enough precision
6321 for 524,287ms (8m44s287ms). If the request is queued long enough to where the
6322 adjusted timestamp exceeds this value, it will be misidentified as highest
6323 priority. Thus it is important to set "timeout queue" to a value, where when
6324 combined with the offset, does not exceed this limit.
Christopher Faulet76c09ef2017-09-21 11:03:52 +02006325
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006326http-request set-query <fmt> [ { if | unless } <condition> ]
Willy Tarreau20b0de52012-12-24 15:45:22 +01006327
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006328 This rewrites the request's query string which appears after the first
6329 question mark ("?") with the result of the evaluation of format string <fmt>.
6330 The part prior to the question mark is left intact. If the request doesn't
6331 contain a question mark and the new value is not empty, then one is added at
6332 the end of the URI, followed by the new value. If a question mark was
6333 present, it will never be removed even if the value is empty. This can be
6334 used to add or remove parameters from the query string.
Ruoshan Huangeb5a3632015-12-08 21:00:23 +08006335
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006336 See also "http-request set-query" and "http-request set-uri".
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01006337
6338 Example:
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006339 # replace "%3D" with "=" in the query string
6340 http-request set-query %[query,regsub(%3D,=,g)]
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01006341
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006342http-request set-src <expr> [ { if | unless } <condition> ]
6343 This is used to set the source IP address to the value of specified
6344 expression. Useful when a proxy in front of HAProxy rewrites source IP, but
6345 provides the correct IP in a HTTP header; or you want to mask source IP for
Olivier Doucet56e31202020-04-21 09:32:56 +02006346 privacy. All subsequent calls to "src" fetch will return this value
6347 (see example).
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006348
6349 Arguments :
6350 <expr> Is a standard HAProxy expression formed by a sample-fetch followed
6351 by some converters.
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01006352
Olivier Doucet56e31202020-04-21 09:32:56 +02006353 See also "option forwardfor".
6354
Cyril Bonté78caf842010-03-10 22:41:43 +01006355 Example:
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006356 http-request set-src hdr(x-forwarded-for)
6357 http-request set-src src,ipmask(24)
6358
Olivier Doucet56e31202020-04-21 09:32:56 +02006359 # After the masking this will track connections
6360 # based on the IP address with the last byte zeroed out.
6361 http-request track-sc0 src
6362
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006363 When possible, set-src preserves the original source port as long as the
6364 address family allows it, otherwise the source port is set to 0.
6365
6366http-request set-src-port <expr> [ { if | unless } <condition> ]
6367
6368 This is used to set the source port address to the value of specified
6369 expression.
6370
6371 Arguments:
6372 <expr> Is a standard HAProxy expression formed by a sample-fetch followed
6373 by some converters.
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01006374
Willy Tarreau20b0de52012-12-24 15:45:22 +01006375 Example:
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006376 http-request set-src-port hdr(x-port)
6377 http-request set-src-port int(4000)
6378
6379 When possible, set-src-port preserves the original source address as long as
6380 the address family supports a port, otherwise it forces the source address to
6381 IPv4 "0.0.0.0" before rewriting the port.
6382
Amaury Denoyelle8d228232020-12-10 13:43:54 +01006383http-request set-timeout server|tunnel { <timeout> | <expr> }
6384 [ { if | unless } <condition> ]
6385
6386 This action overrides the specified "server" or "tunnel" timeout for the
6387 current stream only. The timeout can be specified in millisecond or with any
6388 other unit if the number is suffixed by the unit as explained at the top of
6389 this document. It is also possible to write an expression which must returns
6390 a number interpreted as a timeout in millisecond.
6391
6392 Note that the server/tunnel timeouts are only relevant on the backend side
6393 and thus this rule is only available for the proxies with backend
6394 capabilities. Also the timeout value must be non-null to obtain the expected
6395 results.
6396
6397 Example:
6398 http-request set-timeout server 5s
6399 http-request set-timeout hdr(host),map_int(host.lst)
6400
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006401http-request set-tos <tos> [ { if | unless } <condition> ]
6402
6403 This is used to set the TOS or DSCP field value of packets sent to the client
6404 to the value passed in <tos> on platforms which support this. This value
6405 represents the whole 8 bits of the IP TOS field, and can be expressed both in
6406 decimal or hexadecimal format (prefixed by "0x"). Note that only the 6 higher
6407 bits are used in DSCP or TOS, and the two lower bits are always 0. This can
6408 be used to adjust some routing behavior on border routers based on some
6409 information from the request.
6410
6411 See RFC 2474, 2597, 3260 and 4594 for more information.
6412
6413http-request set-uri <fmt> [ { if | unless } <condition> ]
6414
6415 This rewrites the request URI with the result of the evaluation of format
6416 string <fmt>. The scheme, authority, path and query string are all replaced
6417 at once. This can be used to rewrite hosts in front of proxies, or to
6418 perform complex modifications to the URI such as moving parts between the
6419 path and the query string.
6420 See also "http-request set-path" and "http-request set-query".
6421
6422http-request set-var(<var-name>) <expr> [ { if | unless } <condition> ]
6423
6424 This is used to set the contents of a variable. The variable is declared
6425 inline.
6426
6427 Arguments:
6428 <var-name> The name of the variable starts with an indication about its
6429 scope. The scopes allowed are:
6430 "proc" : the variable is shared with the whole process
6431 "sess" : the variable is shared with the whole session
6432 "txn" : the variable is shared with the transaction
6433 (request and response)
6434 "req" : the variable is shared only during request
6435 processing
6436 "res" : the variable is shared only during response
6437 processing
6438 This prefix is followed by a name. The separator is a '.'.
6439 The name may only contain characters 'a-z', 'A-Z', '0-9'
6440 and '_'.
6441
6442 <expr> Is a standard HAProxy expression formed by a sample-fetch
6443 followed by some converters.
Willy Tarreau20b0de52012-12-24 15:45:22 +01006444
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02006445 Example:
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006446 http-request set-var(req.my_var) req.fhdr(user-agent),lower
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02006447
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006448http-request send-spoe-group <engine-name> <group-name>
6449 [ { if | unless } <condition> ]
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02006450
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006451 This action is used to trigger sending of a group of SPOE messages. To do so,
6452 the SPOE engine used to send messages must be defined, as well as the SPOE
6453 group to send. Of course, the SPOE engine must refer to an existing SPOE
6454 filter. If not engine name is provided on the SPOE filter line, the SPOE
6455 agent name must be used.
6456
6457 Arguments:
6458 <engine-name> The SPOE engine name.
6459
6460 <group-name> The SPOE group name as specified in the engine
6461 configuration.
6462
6463http-request silent-drop [ { if | unless } <condition> ]
6464
6465 This stops the evaluation of the rules and makes the client-facing connection
6466 suddenly disappear using a system-dependent way that tries to prevent the
6467 client from being notified. The effect it then that the client still sees an
6468 established connection while there's none on HAProxy. The purpose is to
6469 achieve a comparable effect to "tarpit" except that it doesn't use any local
6470 resource at all on the machine running HAProxy. It can resist much higher
6471 loads than "tarpit", and slow down stronger attackers. It is important to
6472 understand the impact of using this mechanism. All stateful equipment placed
6473 between the client and HAProxy (firewalls, proxies, load balancers) will also
6474 keep the established connection for a long time and may suffer from this
6475 action.
6476 On modern Linux systems running with enough privileges, the TCP_REPAIR socket
6477 option is used to block the emission of a TCP reset. On other systems, the
6478 socket's TTL is reduced to 1 so that the TCP reset doesn't pass the first
6479 router, though it's still delivered to local networks. Do not use it unless
6480 you fully understand how it works.
6481
Christopher Faulet46f95542019-12-20 10:07:22 +01006482http-request strict-mode { on | off }
6483
6484 This enables or disables the strict rewriting mode for following rules. It
6485 does not affect rules declared before it and it is only applicable on rules
6486 performing a rewrite on the requests. When the strict mode is enabled, any
6487 rewrite failure triggers an internal error. Otherwise, such errors are
6488 silently ignored. The purpose of the strict rewriting mode is to make some
Ilya Shipitsin8525fd92020-02-29 12:34:59 +05006489 rewrites optional while others must be performed to continue the request
Christopher Faulet46f95542019-12-20 10:07:22 +01006490 processing.
6491
Christopher Faulet1aea50e2020-01-17 16:03:53 +01006492 By default, the strict rewriting mode is enabled. Its value is also reset
Christopher Faulet46f95542019-12-20 10:07:22 +01006493 when a ruleset evaluation ends. So, for instance, if you change the mode on
6494 the frontend, the default mode is restored when HAProxy starts the backend
6495 rules evaluation.
6496
Christopher Faulet5cb513a2020-05-13 17:56:56 +02006497http-request tarpit [deny_status <status>] [ { if | unless } <condition> ]
6498http-request tarpit [ { status | deny_status } <code>] [content-type <type>]
6499 [ { default-errorfiles | errorfile <file> | errorfiles <name> |
6500 file <file> | lf-file <file> | string <str> | lf-string <fmt> } ]
6501 [ hdr <name> <fmt> ]*
6502 [ { if | unless } <condition> ]
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006503
6504 This stops the evaluation of the rules and immediately blocks the request
6505 without responding for a delay specified by "timeout tarpit" or
6506 "timeout connect" if the former is not set. After that delay, if the client
Christopher Faulet5cb513a2020-05-13 17:56:56 +02006507 is still connected, a response is returned so that the client does not
6508 suspect it has been tarpitted. Logs will report the flags "PT". The goal of
6509 the tarpit rule is to slow down robots during an attack when they're limited
6510 on the number of concurrent requests. It can be very efficient against very
6511 dumb robots, and will significantly reduce the load on firewalls compared to
6512 a "deny" rule. But when facing "correctly" developed robots, it can make
6513 things worse by forcing haproxy and the front firewall to support insane
6514 number of concurrent connections. By default an HTTP error 500 is returned.
6515 But the response may be customized using same syntax than
6516 "http-request return" rules. Thus, see "http-request return" for details.
Ilya Shipitsin11057a32020-06-21 21:18:27 +05006517 For compatibility purpose, when no argument is defined, or only "deny_status",
Christopher Faulet5cb513a2020-05-13 17:56:56 +02006518 the argument "default-errorfiles" is implied. It means
6519 "http-request tarpit [deny_status <status>]" is an alias of
6520 "http-request tarpit [status <status>] default-errorfiles".
6521 No further "http-request" rules are evaluated.
6522 See also "http-request return" and "http-request silent-drop".
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006523
6524http-request track-sc0 <key> [table <table>] [ { if | unless } <condition> ]
6525http-request track-sc1 <key> [table <table>] [ { if | unless } <condition> ]
6526http-request track-sc2 <key> [table <table>] [ { if | unless } <condition> ]
6527
6528 This enables tracking of sticky counters from current request. These rules do
6529 not stop evaluation and do not change default action. The number of counters
6530 that may be simultaneously tracked by the same connection is set in
6531 MAX_SESS_STKCTR at build time (reported in haproxy -vv) which defaults to 3,
Matteo Contrini1857b8c2020-10-16 17:35:54 +02006532 so the track-sc number is between 0 and (MAX_SESS_STKCTR-1). The first
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006533 "track-sc0" rule executed enables tracking of the counters of the specified
6534 table as the first set. The first "track-sc1" rule executed enables tracking
6535 of the counters of the specified table as the second set. The first
6536 "track-sc2" rule executed enables tracking of the counters of the specified
6537 table as the third set. It is a recommended practice to use the first set of
6538 counters for the per-frontend counters and the second set for the per-backend
6539 ones. But this is just a guideline, all may be used everywhere.
6540
6541 Arguments :
6542 <key> is mandatory, and is a sample expression rule as described in
6543 section 7.3. It describes what elements of the incoming request or
6544 connection will be analyzed, extracted, combined, and used to
6545 select which table entry to update the counters.
6546
6547 <table> is an optional table to be used instead of the default one, which
6548 is the stick-table declared in the current proxy. All the counters
6549 for the matches and updates for the key will then be performed in
6550 that table until the session ends.
6551
6552 Once a "track-sc*" rule is executed, the key is looked up in the table and if
6553 it is not found, an entry is allocated for it. Then a pointer to that entry
6554 is kept during all the session's life, and this entry's counters are updated
6555 as often as possible, every time the session's counters are updated, and also
6556 systematically when the session ends. Counters are only updated for events
6557 that happen after the tracking has been started. As an exception, connection
6558 counters and request counters are systematically updated so that they reflect
6559 useful information.
6560
6561 If the entry tracks concurrent connection counters, one connection is counted
6562 for as long as the entry is tracked, and the entry will not expire during
6563 that time. Tracking counters also provides a performance advantage over just
6564 checking the keys, because only one table lookup is performed for all ACL
6565 checks that make use of it.
6566
6567http-request unset-var(<var-name>) [ { if | unless } <condition> ]
6568
6569 This is used to unset a variable. See above for details about <var-name>.
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02006570
6571 Example:
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006572 http-request unset-var(req.my_var)
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02006573
Christopher Faulet579d83b2019-11-22 15:34:17 +01006574http-request use-service <service-name> [ { if | unless } <condition> ]
6575
6576 This directive executes the configured HTTP service to reply to the request
6577 and stops the evaluation of the rules. An HTTP service may choose to reply by
6578 sending any valid HTTP response or it may immediately close the connection
6579 without sending any response. Outside natives services, for instance the
6580 Prometheus exporter, it is possible to write your own services in Lua. No
6581 further "http-request" rules are evaluated.
6582
6583 Arguments :
6584 <service-name> is mandatory. It is the service to call
6585
6586 Example:
6587 http-request use-service prometheus-exporter if { path /metrics }
6588
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006589http-request wait-for-handshake [ { if | unless } <condition> ]
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02006590
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02006591 This will delay the processing of the request until the SSL handshake
6592 happened. This is mostly useful to delay processing early data until we're
6593 sure they are valid.
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02006594
Willy Tarreauef781042010-01-27 11:53:01 +01006595
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02006596http-response <action> <options...> [ { if | unless } <condition> ]
Willy Tarreaue365c0b2013-06-11 16:06:12 +02006597 Access control for Layer 7 responses
6598
6599 May be used in sections: defaults | frontend | listen | backend
6600 no | yes | yes | yes
6601
6602 The http-response statement defines a set of rules which apply to layer 7
6603 processing. The rules are evaluated in their declaration order when they are
6604 met in a frontend, listen or backend section. Any rule may optionally be
6605 followed by an ACL-based condition, in which case it will only be evaluated
6606 if the condition is true. Since these rules apply on responses, the backend
6607 rules are applied first, followed by the frontend's rules.
6608
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02006609 The first keyword is the rule's action. The supported actions are described
6610 below.
Willy Tarreaue365c0b2013-06-11 16:06:12 +02006611
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02006612 There is no limit to the number of http-response statements per instance.
Willy Tarreaue365c0b2013-06-11 16:06:12 +02006613
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02006614 Example:
6615 acl key_acl res.hdr(X-Acl-Key) -m found
Thierry FOURNIERdad3d1d2014-04-22 18:07:25 +02006616
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02006617 acl myhost hdr(Host) -f myhost.lst
Sasha Pachev218f0642014-06-16 12:05:59 -06006618
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02006619 http-response add-acl(myhost.lst) %[res.hdr(X-Acl-Key)] if key_acl
6620 http-response del-acl(myhost.lst) %[res.hdr(X-Acl-Key)] if key_acl
Sasha Pachev218f0642014-06-16 12:05:59 -06006621
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02006622 Example:
6623 acl value res.hdr(X-Value) -m found
Sasha Pachev218f0642014-06-16 12:05:59 -06006624
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02006625 use_backend bk_appli if { hdr(Host),map_str(map.lst) -m found }
Sasha Pachev218f0642014-06-16 12:05:59 -06006626
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02006627 http-response set-map(map.lst) %[src] %[res.hdr(X-Value)] if value
6628 http-response del-map(map.lst) %[src] if ! value
Sasha Pachev218f0642014-06-16 12:05:59 -06006629
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02006630 See also : "http-request", section 3.4 about userlists and section 7 about
6631 ACL usage.
Sasha Pachev218f0642014-06-16 12:05:59 -06006632
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02006633http-response add-acl(<file-name>) <key fmt> [ { if | unless } <condition> ]
Sasha Pachev218f0642014-06-16 12:05:59 -06006634
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02006635 This is used to add a new entry into an ACL. The ACL must be loaded from a
6636 file (even a dummy empty file). The file name of the ACL to be updated is
6637 passed between parentheses. It takes one argument: <key fmt>, which follows
6638 log-format rules, to collect content of the new entry. It performs a lookup
6639 in the ACL before insertion, to avoid duplicated (or more) values.
6640 This lookup is done by a linear search and can be expensive with large lists!
6641 It is the equivalent of the "add acl" command from the stats socket, but can
6642 be triggered by an HTTP response.
Sasha Pachev218f0642014-06-16 12:05:59 -06006643
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02006644http-response add-header <name> <fmt> [ { if | unless } <condition> ]
Sasha Pachev218f0642014-06-16 12:05:59 -06006645
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02006646 This appends an HTTP header field whose name is specified in <name> and whose
6647 value is defined by <fmt> which follows the log-format rules (see Custom Log
6648 Format in section 8.2.4). This may be used to send a cookie to a client for
6649 example, or to pass some internal information.
6650 This rule is not final, so it is possible to add other similar rules.
6651 Note that header addition is performed immediately, so one rule might reuse
6652 the resulting header from a previous rule.
Sasha Pachev218f0642014-06-16 12:05:59 -06006653
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02006654http-response allow [ { if | unless } <condition> ]
Sasha Pachev218f0642014-06-16 12:05:59 -06006655
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02006656 This stops the evaluation of the rules and lets the response pass the check.
6657 No further "http-response" rules are evaluated for the current section.
Sasha Pachev218f0642014-06-16 12:05:59 -06006658
Jarno Huuskonen251a6b72019-01-04 14:05:02 +02006659http-response cache-store <name> [ { if | unless } <condition> ]
Sasha Pachev218f0642014-06-16 12:05:59 -06006660
Christopher Faulet87f1f3d2019-07-18 14:51:20 +02006661 See section 6.2 about cache setup.
Sasha Pachev218f0642014-06-16 12:05:59 -06006662
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02006663http-response capture <sample> id <id> [ { if | unless } <condition> ]
Sasha Pachev218f0642014-06-16 12:05:59 -06006664
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02006665 This captures sample expression <sample> from the response buffer, and
6666 converts it to a string. The resulting string is stored into the next request
6667 "capture" slot, so it will possibly appear next to some captured HTTP
6668 headers. It will then automatically appear in the logs, and it will be
6669 possible to extract it using sample fetch rules to feed it into headers or
6670 anything. Please check section 7.3 (Fetching samples) and
6671 "capture response header" for more information.
Thierry FOURNIER35d70ef2015-08-26 16:21:56 +02006672
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02006673 The keyword "id" is the id of the capture slot which is used for storing the
6674 string. The capture slot must be defined in an associated frontend.
6675 This is useful to run captures in backends. The slot id can be declared by a
6676 previous directive "http-response capture" or with the "declare capture"
6677 keyword.
Baptiste Assmann19a69b32020-01-16 14:34:22 +01006678
6679 When using this action in a backend, double check that the relevant
6680 frontend(s) have the required capture slots otherwise, this rule will be
6681 ignored at run time. This can't be detected at configuration parsing time
6682 due to HAProxy's ability to dynamically resolve backend name at runtime.
Thierry FOURNIER35d70ef2015-08-26 16:21:56 +02006683
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02006684http-response del-acl(<file-name>) <key fmt> [ { if | unless } <condition> ]
Thierry FOURNIER35d70ef2015-08-26 16:21:56 +02006685
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02006686 This is used to delete an entry from an ACL. The ACL must be loaded from a
6687 file (even a dummy empty file). The file name of the ACL to be updated is
6688 passed between parentheses. It takes one argument: <key fmt>, which follows
6689 log-format rules, to collect content of the entry to delete.
6690 It is the equivalent of the "del acl" command from the stats socket, but can
6691 be triggered by an HTTP response.
Willy Tarreauf4c43c12013-06-11 17:01:13 +02006692
Maciej Zdebebdd4c52020-11-20 13:58:48 +00006693http-response del-header <name> [ -m <meth> ] [ { if | unless } <condition> ]
Willy Tarreau9a355ec2013-06-11 17:45:46 +02006694
Maciej Zdebebdd4c52020-11-20 13:58:48 +00006695 This removes all HTTP header fields whose name is specified in <name>. <meth>
6696 is the matching method, applied on the header name. Supported matching methods
6697 are "str" (exact match), "beg" (prefix match), "end" (suffix match), "sub"
6698 (substring match) and "reg" (regex match). If not specified, exact matching
6699 method is used.
Willy Tarreau42cf39e2013-06-11 18:51:32 +02006700
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02006701http-response del-map(<file-name>) <key fmt> [ { if | unless } <condition> ]
Willy Tarreau51347ed2013-06-11 19:34:13 +02006702
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02006703 This is used to delete an entry from a MAP. The MAP must be loaded from a
6704 file (even a dummy empty file). The file name of the MAP to be updated is
6705 passed between parentheses. It takes one argument: <key fmt>, which follows
6706 log-format rules, to collect content of the entry to delete.
6707 It takes one argument: "file name" It is the equivalent of the "del map"
6708 command from the stats socket, but can be triggered by an HTTP response.
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02006709
Christopher Faulet5cb513a2020-05-13 17:56:56 +02006710http-response deny [deny_status <status>] [ { if | unless } <condition> ]
6711http-response deny [ { status | deny_status } <code>] [content-type <type>]
6712 [ { default-errorfiles | errorfile <file> | errorfiles <name> |
6713 file <file> | lf-file <file> | string <str> | lf-string <fmt> } ]
6714 [ hdr <name> <fmt> ]*
6715 [ { if | unless } <condition> ]
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02006716
Christopher Faulet5cb513a2020-05-13 17:56:56 +02006717 This stops the evaluation of the rules and immediately rejects the response.
6718 By default an HTTP 502 error is returned. But the response may be customized
6719 using same syntax than "http-response return" rules. Thus, see
Ilya Shipitsin11057a32020-06-21 21:18:27 +05006720 "http-response return" for details. For compatibility purpose, when no
Christopher Faulet5cb513a2020-05-13 17:56:56 +02006721 argument is defined, or only "deny_status", the argument "default-errorfiles"
6722 is implied. It means "http-response deny [deny_status <status>]" is an alias
6723 of "http-response deny [status <status>] default-errorfiles".
Christopher Faulet040c8cd2020-01-13 16:43:45 +01006724 No further "http-response" rules are evaluated.
Christopher Faulet5cb513a2020-05-13 17:56:56 +02006725 See also "http-response return".
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02006726
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02006727http-response redirect <rule> [ { if | unless } <condition> ]
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02006728
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02006729 This performs an HTTP redirection based on a redirect rule.
6730 This supports a format string similarly to "http-request redirect" rules,
6731 with the exception that only the "location" type of redirect is possible on
6732 the response. See the "redirect" keyword for the rule's syntax. When a
6733 redirect rule is applied during a response, connections to the server are
6734 closed so that no data can be forwarded from the server to the client.
Thierry FOURNIERe80fada2015-05-26 18:06:31 +02006735
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02006736http-response replace-header <name> <regex-match> <replace-fmt>
6737 [ { if | unless } <condition> ]
Thierry FOURNIERe80fada2015-05-26 18:06:31 +02006738
Tim Duesterhus2252beb2019-10-29 00:05:13 +01006739 This works like "http-request replace-header" except that it works on the
6740 server's response instead of the client's request.
William Lallemand86d0df02017-11-24 21:36:45 +01006741
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02006742 Example:
6743 http-response replace-header Set-Cookie (C=[^;]*);(.*) \1;ip=%bi;\2
Willy Tarreau51d861a2015-05-22 17:30:48 +02006744
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02006745 # applied to:
6746 Set-Cookie: C=1; expires=Tue, 14-Jun-2016 01:40:45 GMT
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02006747
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02006748 # outputs:
6749 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 +02006750
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02006751 # assuming the backend IP is 192.168.1.20.
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02006752
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02006753http-response replace-value <name> <regex-match> <replace-fmt>
6754 [ { if | unless } <condition> ]
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02006755
Tim Duesterhus6bd909b2020-01-17 15:53:18 +01006756 This works like "http-request replace-value" except that it works on the
Tim Duesterhus2252beb2019-10-29 00:05:13 +01006757 server's response instead of the client's request.
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02006758
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02006759 Example:
6760 http-response replace-value Cache-control ^public$ private
Christopher Faulet85d79c92016-11-09 16:54:56 +01006761
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02006762 # applied to:
6763 Cache-Control: max-age=3600, public
Christopher Faulet85d79c92016-11-09 16:54:56 +01006764
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02006765 # outputs:
6766 Cache-Control: max-age=3600, private
Christopher Faulet85d79c92016-11-09 16:54:56 +01006767
Christopher Faulet24231ab2020-01-24 17:44:23 +01006768http-response return [status <code>] [content-type <type>]
6769 [ { default-errorfiles | errorfile <file> | errorfiles <name> |
6770 file <file> | lf-file <file> | string <str> | lf-string <fmt> } ]
Christopher Faulet4a2c1422020-01-31 17:36:01 +01006771 [ hdr <name> <value> ]*
Christopher Faulet24231ab2020-01-24 17:44:23 +01006772 [ { if | unless } <condition> ]
6773
Ilya Shipitsin8525fd92020-02-29 12:34:59 +05006774 This stops the evaluation of the rules and immediately returns a response. The
Christopher Faulet24231ab2020-01-24 17:44:23 +01006775 default status code used for the response is 200. It can be optionally
6776 specified as an arguments to "status". The response content-type may also be
Daniel Corbett67a82712020-07-06 23:01:19 -04006777 specified as an argument to "content-type". Finally the response itself may
Christopher Faulet24231ab2020-01-24 17:44:23 +01006778 be defined. If can be a full HTTP response specifying the errorfile to use,
Ilya Shipitsin8525fd92020-02-29 12:34:59 +05006779 or the response payload specifying the file or the string to use. These rules
Christopher Faulet24231ab2020-01-24 17:44:23 +01006780 are followed to create the response :
6781
6782 * If neither the errorfile nor the payload to use is defined, a dummy
6783 response is returned. Only the "status" argument is considered. It can be
6784 any code in the range [200, 599]. The "content-type" argument, if any, is
6785 ignored.
6786
6787 * If "default-errorfiles" argument is set, the proxy's errorfiles are
6788 considered. If the "status" argument is defined, it must be one of the
Daniel Corbett67a82712020-07-06 23:01:19 -04006789 status code handled by haproxy (200, 400, 403, 404, 405, 408, 410, 413,
Christopher Faulete095f312020-12-07 11:22:24 +01006790 425, 429, 500, 501, 502, 503, and 504). The "content-type" argument, if
6791 any, is ignored.
Christopher Faulet24231ab2020-01-24 17:44:23 +01006792
6793 * If a specific errorfile is defined, with an "errorfile" argument, the
6794 corresponding file, containing a full HTTP response, is returned. Only the
6795 "status" argument is considered. It must be one of the status code handled
Christopher Faulete095f312020-12-07 11:22:24 +01006796 by haproxy (200, 400, 403, 404, 405, 408, 410, 413, 425, 429, 500, 501,
6797 502, 503, and 504). The "content-type" argument, if any, is ignored.
Christopher Faulet24231ab2020-01-24 17:44:23 +01006798
6799 * If an http-errors section is defined, with an "errorfiles" argument, the
6800 corresponding file in the specified http-errors section, containing a full
6801 HTTP response, is returned. Only the "status" argument is considered. It
Daniel Corbett67a82712020-07-06 23:01:19 -04006802 must be one of the status code handled by haproxy (200, 400, 403, 404, 405,
Christopher Faulete095f312020-12-07 11:22:24 +01006803 408, 410, 413, 425, 429, 500, 501, 502, 503, and 504). The "content-type"
Anthonin Bonnefoy85048f82020-06-22 09:17:01 +02006804 argument, if any, is ignored.
Christopher Faulet24231ab2020-01-24 17:44:23 +01006805
6806 * If a "file" or a "lf-file" argument is specified, the file's content is
6807 used as the response payload. If the file is not empty, its content-type
6808 must be set as argument to "content-type". Otherwise, any "content-type"
6809 argument is ignored. With a "lf-file" argument, the file's content is
6810 evaluated as a log-format string. With a "file" argument, it is considered
6811 as a raw content.
6812
6813 * If a "string" or "lf-string" argument is specified, the defined string is
6814 used as the response payload. The content-type must always be set as
6815 argument to "content-type". With a "lf-string" argument, the string is
6816 evaluated as a log-format string. With a "string" argument, it is
6817 considered as a raw string.
6818
Christopher Faulet4a2c1422020-01-31 17:36:01 +01006819 When the response is not based an errorfile, it is possible to appends HTTP
6820 header fields to the response using "hdr" arguments. Otherwise, all "hdr"
6821 arguments are ignored. For each one, the header name is specified in <name>
6822 and its value is defined by <fmt> which follows the log-format rules.
6823
Christopher Faulet24231ab2020-01-24 17:44:23 +01006824 Note that the generated response must be smaller than a buffer. And to avoid
6825 any warning, when an errorfile or a raw file is loaded, the buffer space
Ilya Shipitsin11057a32020-06-21 21:18:27 +05006826 reserved to the headers rewriting should also be free.
Christopher Faulet24231ab2020-01-24 17:44:23 +01006827
6828 No further "http-response" rules are evaluated.
6829
6830 Example:
Daniel Corbett67a82712020-07-06 23:01:19 -04006831 http-response return errorfile /etc/haproxy/errorfiles/200.http \
Christopher Faulet24231ab2020-01-24 17:44:23 +01006832 if { status eq 404 }
6833
6834 http-response return content-type text/plain \
6835 string "This is the end !" \
6836 if { status eq 500 }
6837
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02006838http-response sc-inc-gpc0(<sc-id>) [ { if | unless } <condition> ]
6839http-response sc-inc-gpc1(<sc-id>) [ { if | unless } <condition> ]
Ruoshan Huange4edc6b2016-07-14 15:07:45 +08006840
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02006841 This action increments the GPC0 or GPC1 counter according with the sticky
6842 counter designated by <sc-id>. If an error occurs, this action silently fails
6843 and the actions evaluation continues.
Thierry FOURNIER236657b2015-08-19 08:25:14 +02006844
Cédric Dufour0d7712d2019-11-06 18:38:53 +01006845http-response sc-set-gpt0(<sc-id>) { <int> | <expr> }
6846 [ { if | unless } <condition> ]
Thierry FOURNIERe0627bd2015-08-04 08:20:33 +02006847
Cédric Dufour0d7712d2019-11-06 18:38:53 +01006848 This action sets the 32-bit unsigned GPT0 tag according to the sticky counter
6849 designated by <sc-id> and the value of <int>/<expr>. The expected result is a
6850 boolean. If an error occurs, this action silently fails and the actions
6851 evaluation continues.
Frédéric Lécaille6778b272018-01-29 15:22:53 +01006852
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02006853http-response send-spoe-group [ { if | unless } <condition> ]
Willy Tarreau2d392c22015-08-24 01:43:45 +02006854
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02006855 This action is used to trigger sending of a group of SPOE messages. To do so,
6856 the SPOE engine used to send messages must be defined, as well as the SPOE
6857 group to send. Of course, the SPOE engine must refer to an existing SPOE
6858 filter. If not engine name is provided on the SPOE filter line, the SPOE
6859 agent name must be used.
Christopher Faulet76c09ef2017-09-21 11:03:52 +02006860
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02006861 Arguments:
6862 <engine-name> The SPOE engine name.
Christopher Faulet76c09ef2017-09-21 11:03:52 +02006863
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02006864 <group-name> The SPOE group name as specified in the engine
6865 configuration.
Christopher Faulet76c09ef2017-09-21 11:03:52 +02006866
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02006867http-response set-header <name> <fmt> [ { if | unless } <condition> ]
Willy Tarreaue365c0b2013-06-11 16:06:12 +02006868
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02006869 This does the same as "add-header" except that the header name is first
6870 removed if it existed. This is useful when passing security information to
6871 the server, where the header must not be manipulated by external users.
Willy Tarreaue365c0b2013-06-11 16:06:12 +02006872
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02006873http-response set-log-level <level> [ { if | unless } <condition> ]
6874
6875 This is used to change the log level of the current request when a certain
6876 condition is met. Valid levels are the 8 syslog levels (see the "log"
6877 keyword) plus the special level "silent" which disables logging for this
6878 request. This rule is not final so the last matching rule wins. This rule can
6879 be useful to disable health checks coming from another equipment.
6880
6881http-response set-map(<file-name>) <key fmt> <value fmt>
6882
6883 This is used to add a new entry into a MAP. The MAP must be loaded from a
6884 file (even a dummy empty file). The file name of the MAP to be updated is
6885 passed between parentheses. It takes 2 arguments: <key fmt>, which follows
6886 log-format rules, used to collect MAP key, and <value fmt>, which follows
6887 log-format rules, used to collect content for the new entry. It performs a
6888 lookup in the MAP before insertion, to avoid duplicated (or more) values.
6889 This lookup is done by a linear search and can be expensive with large lists!
6890 It is the equivalent of the "set map" command from the stats socket, but can
6891 be triggered by an HTTP response.
6892
6893http-response set-mark <mark> [ { if | unless } <condition> ]
6894
6895 This is used to set the Netfilter MARK on all packets sent to the client to
6896 the value passed in <mark> on platforms which support it. This value is an
6897 unsigned 32 bit value which can be matched by netfilter and by the routing
6898 table. It can be expressed both in decimal or hexadecimal format (prefixed
6899 by "0x"). This can be useful to force certain packets to take a different
6900 route (for example a cheaper network path for bulk downloads). This works on
6901 Linux kernels 2.6.32 and above and requires admin privileges.
6902
6903http-response set-nice <nice> [ { if | unless } <condition> ]
6904
6905 This sets the "nice" factor of the current request being processed.
6906 It only has effect against the other requests being processed at the same
6907 time. The default value is 0, unless altered by the "nice" setting on the
6908 "bind" line. The accepted range is -1024..1024. The higher the value, the
6909 nicest the request will be. Lower values will make the request more important
6910 than other ones. This can be useful to improve the speed of some requests, or
6911 lower the priority of non-important requests. Using this setting without
6912 prior experimentation can cause some major slowdown.
6913
6914http-response set-status <status> [reason <str>]
6915 [ { if | unless } <condition> ]
6916
6917 This replaces the response status code with <status> which must be an integer
6918 between 100 and 999. Optionally, a custom reason text can be provided defined
6919 by <str>, or the default reason for the specified code will be used as a
6920 fallback.
Ruoshan Huangeb5a3632015-12-08 21:00:23 +08006921
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02006922 Example:
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02006923 # return "431 Request Header Fields Too Large"
6924 http-response set-status 431
6925 # return "503 Slow Down", custom reason
6926 http-response set-status 503 reason "Slow Down".
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02006927
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02006928http-response set-tos <tos> [ { if | unless } <condition> ]
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02006929
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02006930 This is used to set the TOS or DSCP field value of packets sent to the client
6931 to the value passed in <tos> on platforms which support this.
6932 This value represents the whole 8 bits of the IP TOS field, and can be
6933 expressed both in decimal or hexadecimal format (prefixed by "0x"). Note that
6934 only the 6 higher bits are used in DSCP or TOS, and the two lower bits are
6935 always 0. This can be used to adjust some routing behavior on border routers
6936 based on some information from the request.
6937
6938 See RFC 2474, 2597, 3260 and 4594 for more information.
6939
6940http-response set-var(<var-name>) <expr> [ { if | unless } <condition> ]
6941
6942 This is used to set the contents of a variable. The variable is declared
6943 inline.
6944
6945 Arguments:
6946 <var-name> The name of the variable starts with an indication about its
6947 scope. The scopes allowed are:
6948 "proc" : the variable is shared with the whole process
6949 "sess" : the variable is shared with the whole session
6950 "txn" : the variable is shared with the transaction
6951 (request and response)
6952 "req" : the variable is shared only during request
6953 processing
6954 "res" : the variable is shared only during response
6955 processing
6956 This prefix is followed by a name. The separator is a '.'.
6957 The name may only contain characters 'a-z', 'A-Z', '0-9', '.'
6958 and '_'.
6959
6960 <expr> Is a standard HAProxy expression formed by a sample-fetch
6961 followed by some converters.
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02006962
6963 Example:
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02006964 http-response set-var(sess.last_redir) res.hdr(location)
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02006965
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02006966http-response silent-drop [ { if | unless } <condition> ]
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02006967
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02006968 This stops the evaluation of the rules and makes the client-facing connection
6969 suddenly disappear using a system-dependent way that tries to prevent the
6970 client from being notified. The effect it then that the client still sees an
6971 established connection while there's none on HAProxy. The purpose is to
6972 achieve a comparable effect to "tarpit" except that it doesn't use any local
6973 resource at all on the machine running HAProxy. It can resist much higher
6974 loads than "tarpit", and slow down stronger attackers. It is important to
6975 understand the impact of using this mechanism. All stateful equipment placed
6976 between the client and HAProxy (firewalls, proxies, load balancers) will also
6977 keep the established connection for a long time and may suffer from this
6978 action.
6979 On modern Linux systems running with enough privileges, the TCP_REPAIR socket
6980 option is used to block the emission of a TCP reset. On other systems, the
6981 socket's TTL is reduced to 1 so that the TCP reset doesn't pass the first
6982 router, though it's still delivered to local networks. Do not use it unless
6983 you fully understand how it works.
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02006984
Christopher Faulet46f95542019-12-20 10:07:22 +01006985http-response strict-mode { on | off }
6986
6987 This enables or disables the strict rewriting mode for following rules. It
6988 does not affect rules declared before it and it is only applicable on rules
6989 performing a rewrite on the responses. When the strict mode is enabled, any
6990 rewrite failure triggers an internal error. Otherwise, such errors are
6991 silently ignored. The purpose of the strict rewriting mode is to make some
Ilya Shipitsin8525fd92020-02-29 12:34:59 +05006992 rewrites optional while others must be performed to continue the response
Christopher Faulet46f95542019-12-20 10:07:22 +01006993 processing.
6994
Christopher Faulet1aea50e2020-01-17 16:03:53 +01006995 By default, the strict rewriting mode is enabled. Its value is also reset
Christopher Faulet46f95542019-12-20 10:07:22 +01006996 when a ruleset evaluation ends. So, for instance, if you change the mode on
Daniel Corbett67a82712020-07-06 23:01:19 -04006997 the backend, the default mode is restored when HAProxy starts the frontend
Christopher Faulet46f95542019-12-20 10:07:22 +01006998 rules evaluation.
6999
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02007000http-response track-sc0 <key> [table <table>] [ { if | unless } <condition> ]
7001http-response track-sc1 <key> [table <table>] [ { if | unless } <condition> ]
7002http-response track-sc2 <key> [table <table>] [ { if | unless } <condition> ]
Willy Tarreaue365c0b2013-06-11 16:06:12 +02007003
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02007004 This enables tracking of sticky counters from current response. Please refer
7005 to "http-request track-sc" for a complete description. The only difference
7006 from "http-request track-sc" is the <key> sample expression can only make use
7007 of samples in response (e.g. res.*, status etc.) and samples below Layer 6
7008 (e.g. SSL-related samples, see section 7.3.4). If the sample is not
7009 supported, haproxy will fail and warn while parsing the config.
7010
7011http-response unset-var(<var-name>) [ { if | unless } <condition> ]
7012
7013 This is used to unset a variable. See "http-response set-var" for details
7014 about <var-name>.
7015
7016 Example:
7017 http-response unset-var(sess.last_redir)
7018
Baptiste Assmann5ecb77f2013-10-06 23:24:13 +02007019
Willy Tarreau30631952015-08-06 15:05:24 +02007020http-reuse { never | safe | aggressive | always }
7021 Declare how idle HTTP connections may be shared between requests
7022
7023 May be used in sections: defaults | frontend | listen | backend
7024 yes | no | yes | yes
7025
7026 By default, a connection established between haproxy and the backend server
Olivier Houchard86006a52018-12-14 19:37:49 +01007027 which is considered safe for reuse is moved back to the server's idle
7028 connections pool so that any other request can make use of it. This is the
7029 "safe" strategy below.
Willy Tarreau30631952015-08-06 15:05:24 +02007030
7031 The argument indicates the desired connection reuse strategy :
7032
Olivier Houchard86006a52018-12-14 19:37:49 +01007033 - "never" : idle connections are never shared between sessions. This mode
7034 may be enforced to cancel a different strategy inherited from
7035 a defaults section or for troubleshooting. For example, if an
7036 old bogus application considers that multiple requests over
7037 the same connection come from the same client and it is not
7038 possible to fix the application, it may be desirable to
7039 disable connection sharing in a single backend. An example of
7040 such an application could be an old haproxy using cookie
7041 insertion in tunnel mode and not checking any request past the
7042 first one.
Willy Tarreau30631952015-08-06 15:05:24 +02007043
Olivier Houchard86006a52018-12-14 19:37:49 +01007044 - "safe" : this is the default and the recommended strategy. The first
7045 request of a session is always sent over its own connection,
7046 and only subsequent requests may be dispatched over other
7047 existing connections. This ensures that in case the server
7048 closes the connection when the request is being sent, the
7049 browser can decide to silently retry it. Since it is exactly
7050 equivalent to regular keep-alive, there should be no side
Amaury Denoyelle27179652020-10-14 18:17:12 +02007051 effects. There is also a special handling for the connections
7052 using protocols subject to Head-of-line blocking (backend with
7053 h2 or fcgi). In this case, when at least one stream is
7054 processed, the used connection is reserved to handle streams
7055 of the same session. When no more streams are processed, the
7056 connection is released and can be reused.
Willy Tarreau30631952015-08-06 15:05:24 +02007057
7058 - "aggressive" : this mode may be useful in webservices environments where
7059 all servers are not necessarily known and where it would be
7060 appreciable to deliver most first requests over existing
7061 connections. In this case, first requests are only delivered
7062 over existing connections that have been reused at least once,
7063 proving that the server correctly supports connection reuse.
7064 It should only be used when it's sure that the client can
7065 retry a failed request once in a while and where the benefit
Michael Prokop4438c602019-05-24 10:25:45 +02007066 of aggressive connection reuse significantly outweighs the
Willy Tarreau30631952015-08-06 15:05:24 +02007067 downsides of rare connection failures.
7068
7069 - "always" : this mode is only recommended when the path to the server is
7070 known for never breaking existing connections quickly after
7071 releasing them. It allows the first request of a session to be
7072 sent to an existing connection. This can provide a significant
7073 performance increase over the "safe" strategy when the backend
7074 is a cache farm, since such components tend to show a
Davor Ocelice9ed2812017-12-25 17:49:28 +01007075 consistent behavior and will benefit from the connection
Willy Tarreau30631952015-08-06 15:05:24 +02007076 sharing. It is recommended that the "http-keep-alive" timeout
7077 remains low in this mode so that no dead connections remain
7078 usable. In most cases, this will lead to the same performance
7079 gains as "aggressive" but with more risks. It should only be
7080 used when it improves the situation over "aggressive".
7081
7082 When http connection sharing is enabled, a great care is taken to respect the
Amaury Denoyelled773a4e2021-01-29 15:18:49 +01007083 connection properties and compatibility. Indeed, some properties are specific
7084 and it is not possibly to reuse it blindly. Those are the SSL SNI, source
7085 and destination address and proxy protocol block. A connection is reused only
7086 if it shares the same set of properties with the request.
Willy Tarreau30631952015-08-06 15:05:24 +02007087
Amaury Denoyelled773a4e2021-01-29 15:18:49 +01007088 Also note that connections with certain bogus authentication schemes (relying
7089 on the connection) like NTLM are marked private and never shared.
Willy Tarreau30631952015-08-06 15:05:24 +02007090
Lukas Tribuse8adfeb2019-11-06 11:50:25 +01007091 A connection pool is involved and configurable with "pool-max-conn".
Willy Tarreau30631952015-08-06 15:05:24 +02007092
7093 Note: connection reuse improves the accuracy of the "server maxconn" setting,
7094 because almost no new connection will be established while idle connections
7095 remain available. This is particularly true with the "always" strategy.
7096
7097 See also : "option http-keep-alive", "server maxconn"
7098
7099
Mark Lamourinec2247f02012-01-04 13:02:01 -05007100http-send-name-header [<header>]
7101 Add the server name to a request. Use the header string given by <header>
Mark Lamourinec2247f02012-01-04 13:02:01 -05007102 May be used in sections: defaults | frontend | listen | backend
7103 yes | no | yes | yes
Mark Lamourinec2247f02012-01-04 13:02:01 -05007104 Arguments :
Mark Lamourinec2247f02012-01-04 13:02:01 -05007105 <header> The header string to use to send the server name
7106
Willy Tarreau81bef7e2019-10-07 14:58:02 +02007107 The "http-send-name-header" statement causes the header field named <header>
7108 to be set to the name of the target server at the moment the request is about
7109 to be sent on the wire. Any existing occurrences of this header are removed.
7110 Upon retries and redispatches, the header field is updated to always reflect
7111 the server being attempted to connect to. Given that this header is modified
7112 very late in the connection setup, it may have unexpected effects on already
7113 modified headers. For example using it with transport-level header such as
7114 connection, content-length, transfer-encoding and so on will likely result in
7115 invalid requests being sent to the server. Additionally it has been reported
7116 that this directive is currently being used as a way to overwrite the Host
7117 header field in outgoing requests; while this trick has been known to work
7118 as a side effect of the feature for some time, it is not officially supported
7119 and might possibly not work anymore in a future version depending on the
7120 technical difficulties this feature induces. A long-term solution instead
7121 consists in fixing the application which required this trick so that it binds
7122 to the correct host name.
Mark Lamourinec2247f02012-01-04 13:02:01 -05007123
7124 See also : "server"
7125
Krzysztof Piotr Oledzkif58a9622008-02-23 01:19:10 +01007126id <value>
Willy Tarreau53fb4ae2009-10-04 23:04:08 +02007127 Set a persistent ID to a proxy.
7128 May be used in sections : defaults | frontend | listen | backend
7129 no | yes | yes | yes
7130 Arguments : none
7131
7132 Set a persistent ID for the proxy. This ID must be unique and positive.
7133 An unused ID will automatically be assigned if unset. The first assigned
7134 value will be 1. This ID is currently only returned in statistics.
Krzysztof Piotr Oledzkif58a9622008-02-23 01:19:10 +01007135
7136
Cyril Bonté0d4bf012010-04-25 23:21:46 +02007137ignore-persist { if | unless } <condition>
7138 Declare a condition to ignore persistence
7139 May be used in sections: defaults | frontend | listen | backend
Cyril Bonté4288c5a2018-03-12 22:02:59 +01007140 no | no | yes | yes
Cyril Bonté0d4bf012010-04-25 23:21:46 +02007141
7142 By default, when cookie persistence is enabled, every requests containing
7143 the cookie are unconditionally persistent (assuming the target server is up
7144 and running).
7145
7146 The "ignore-persist" statement allows one to declare various ACL-based
7147 conditions which, when met, will cause a request to ignore persistence.
7148 This is sometimes useful to load balance requests for static files, which
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03007149 often don't require persistence. This can also be used to fully disable
Cyril Bonté0d4bf012010-04-25 23:21:46 +02007150 persistence for a specific User-Agent (for example, some web crawler bots).
7151
Cyril Bonté0d4bf012010-04-25 23:21:46 +02007152 The persistence is ignored when an "if" condition is met, or unless an
7153 "unless" condition is met.
7154
Jarno Huuskonene5ae7022017-04-03 14:36:21 +03007155 Example:
7156 acl url_static path_beg /static /images /img /css
7157 acl url_static path_end .gif .png .jpg .css .js
7158 ignore-persist if url_static
7159
Cyril Bonté0d4bf012010-04-25 23:21:46 +02007160 See also : "force-persist", "cookie", and section 7 about ACL usage.
7161
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02007162load-server-state-from-file { global | local | none }
7163 Allow seamless reload of HAProxy
7164 May be used in sections: defaults | frontend | listen | backend
7165 yes | no | yes | yes
7166
7167 This directive points HAProxy to a file where server state from previous
7168 running process has been saved. That way, when starting up, before handling
7169 traffic, the new process can apply old states to servers exactly has if no
Davor Ocelice9ed2812017-12-25 17:49:28 +01007170 reload occurred. The purpose of the "load-server-state-from-file" directive is
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02007171 to tell haproxy which file to use. For now, only 2 arguments to either prevent
7172 loading state or load states from a file containing all backends and servers.
7173 The state file can be generated by running the command "show servers state"
7174 over the stats socket and redirect output.
7175
Davor Ocelice9ed2812017-12-25 17:49:28 +01007176 The format of the file is versioned and is very specific. To understand it,
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02007177 please read the documentation of the "show servers state" command (chapter
Willy Tarreau1af20c72017-06-23 16:01:14 +02007178 9.3 of Management Guide).
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02007179
7180 Arguments:
7181 global load the content of the file pointed by the global directive
7182 named "server-state-file".
7183
7184 local load the content of the file pointed by the directive
7185 "server-state-file-name" if set. If not set, then the backend
7186 name is used as a file name.
7187
7188 none don't load any stat for this backend
7189
7190 Notes:
Willy Tarreaue5a60682016-11-09 14:54:53 +01007191 - server's IP address is preserved across reloads by default, but the
7192 order can be changed thanks to the server's "init-addr" setting. This
7193 means that an IP address change performed on the CLI at run time will
Davor Ocelice9ed2812017-12-25 17:49:28 +01007194 be preserved, and that any change to the local resolver (e.g. /etc/hosts)
Willy Tarreaue5a60682016-11-09 14:54:53 +01007195 will possibly not have any effect if the state file is in use.
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02007196
7197 - server's weight is applied from previous running process unless it has
7198 has changed between previous and new configuration files.
7199
Olivier Doucetaa1ea8a2016-08-05 17:15:20 +02007200 Example: Minimal configuration
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02007201
Olivier Doucetaa1ea8a2016-08-05 17:15:20 +02007202 global
7203 stats socket /tmp/socket
7204 server-state-file /tmp/server_state
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02007205
Olivier Doucetaa1ea8a2016-08-05 17:15:20 +02007206 defaults
7207 load-server-state-from-file global
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02007208
Olivier Doucetaa1ea8a2016-08-05 17:15:20 +02007209 backend bk
7210 server s1 127.0.0.1:22 check weight 11
7211 server s2 127.0.0.1:22 check weight 12
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02007212
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02007213
7214 Then one can run :
7215
7216 socat /tmp/socket - <<< "show servers state" > /tmp/server_state
7217
7218 Content of the file /tmp/server_state would be like this:
7219
7220 1
7221 # <field names skipped for the doc example>
7222 1 bk 1 s1 127.0.0.1 2 0 11 11 4 6 3 4 6 0 0
7223 1 bk 2 s2 127.0.0.1 2 0 12 12 4 6 3 4 6 0 0
7224
Olivier Doucetaa1ea8a2016-08-05 17:15:20 +02007225 Example: Minimal configuration
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02007226
7227 global
7228 stats socket /tmp/socket
7229 server-state-base /etc/haproxy/states
7230
7231 defaults
7232 load-server-state-from-file local
7233
7234 backend bk
7235 server s1 127.0.0.1:22 check weight 11
7236 server s2 127.0.0.1:22 check weight 12
7237
Olivier Doucetaa1ea8a2016-08-05 17:15:20 +02007238
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02007239 Then one can run :
7240
7241 socat /tmp/socket - <<< "show servers state bk" > /etc/haproxy/states/bk
7242
7243 Content of the file /etc/haproxy/states/bk would be like this:
7244
7245 1
7246 # <field names skipped for the doc example>
7247 1 bk 1 s1 127.0.0.1 2 0 11 11 4 6 3 4 6 0 0
7248 1 bk 2 s2 127.0.0.1 2 0 12 12 4 6 3 4 6 0 0
7249
7250 See also: "server-state-file", "server-state-file-name", and
7251 "show servers state"
7252
Cyril Bonté0d4bf012010-04-25 23:21:46 +02007253
Willy Tarreau2769aa02007-12-27 18:26:09 +01007254log global
Jan Wagner3e678602020-12-17 22:22:32 +01007255log <address> [len <length>] [format <format>] [sample <ranges>:<sample_size>]
Frédéric Lécailled690dfa2019-04-25 10:52:17 +02007256 <facility> [<level> [<minlevel>]]
William Lallemand0f99e342011-10-12 17:50:54 +02007257no log
Willy Tarreau2769aa02007-12-27 18:26:09 +01007258 Enable per-instance logging of events and traffic.
7259 May be used in sections : defaults | frontend | listen | backend
7260 yes | yes | yes | yes
William Lallemand0f99e342011-10-12 17:50:54 +02007261
7262 Prefix :
7263 no should be used when the logger list must be flushed. For example,
7264 if you don't want to inherit from the default logger list. This
7265 prefix does not allow arguments.
7266
Willy Tarreau2769aa02007-12-27 18:26:09 +01007267 Arguments :
7268 global should be used when the instance's logging parameters are the
7269 same as the global ones. This is the most common usage. "global"
7270 replaces <address>, <facility> and <level> with those of the log
7271 entries found in the "global" section. Only one "log global"
7272 statement may be used per instance, and this form takes no other
7273 parameter.
7274
7275 <address> indicates where to send the logs. It takes the same format as
7276 for the "global" section's logs, and can be one of :
7277
7278 - An IPv4 address optionally followed by a colon (':') and a UDP
7279 port. If no port is specified, 514 is used by default (the
7280 standard syslog port).
7281
David du Colombier24bb5f52011-03-17 10:40:23 +01007282 - An IPv6 address followed by a colon (':') and optionally a UDP
7283 port. If no port is specified, 514 is used by default (the
7284 standard syslog port).
7285
Willy Tarreau2769aa02007-12-27 18:26:09 +01007286 - A filesystem path to a UNIX domain socket, keeping in mind
7287 considerations for chroot (be sure the path is accessible
7288 inside the chroot) and uid/gid (be sure the path is
Davor Ocelice9ed2812017-12-25 17:49:28 +01007289 appropriately writable).
Willy Tarreau2769aa02007-12-27 18:26:09 +01007290
Willy Tarreau5a32ecc2018-11-12 07:34:59 +01007291 - A file descriptor number in the form "fd@<number>", which may
7292 point to a pipe, terminal, or socket. In this case unbuffered
7293 logs are used and one writev() call per log is performed. This
7294 is a bit expensive but acceptable for most workloads. Messages
7295 sent this way will not be truncated but may be dropped, in
7296 which case the DroppedLogs counter will be incremented. The
7297 writev() call is atomic even on pipes for messages up to
7298 PIPE_BUF size, which POSIX recommends to be at least 512 and
7299 which is 4096 bytes on most modern operating systems. Any
7300 larger message may be interleaved with messages from other
7301 processes. Exceptionally for debugging purposes the file
7302 descriptor may also be directed to a file, but doing so will
7303 significantly slow haproxy down as non-blocking calls will be
7304 ignored. Also there will be no way to purge nor rotate this
7305 file without restarting the process. Note that the configured
7306 syslog format is preserved, so the output is suitable for use
Willy Tarreauc1b06452018-11-12 11:57:56 +01007307 with a TCP syslog server. See also the "short" and "raw"
7308 formats below.
Willy Tarreau5a32ecc2018-11-12 07:34:59 +01007309
7310 - "stdout" / "stderr", which are respectively aliases for "fd@1"
7311 and "fd@2", see above.
7312
Willy Tarreauc046d162019-08-30 15:24:59 +02007313 - A ring buffer in the form "ring@<name>", which will correspond
7314 to an in-memory ring buffer accessible over the CLI using the
7315 "show events" command, which will also list existing rings and
7316 their sizes. Such buffers are lost on reload or restart but
7317 when used as a complement this can help troubleshooting by
7318 having the logs instantly available.
7319
Willy Tarreau5a32ecc2018-11-12 07:34:59 +01007320 You may want to reference some environment variables in the
7321 address parameter, see section 2.3 about environment variables.
Willy Tarreaudad36a32013-03-11 01:20:04 +01007322
Willy Tarreau18324f52014-06-27 18:10:07 +02007323 <length> is an optional maximum line length. Log lines larger than this
7324 value will be truncated before being sent. The reason is that
7325 syslog servers act differently on log line length. All servers
7326 support the default value of 1024, but some servers simply drop
7327 larger lines while others do log them. If a server supports long
7328 lines, it may make sense to set this value here in order to avoid
7329 truncating long lines. Similarly, if a server drops long lines,
7330 it is preferable to truncate them before sending them. Accepted
7331 values are 80 to 65535 inclusive. The default value of 1024 is
7332 generally fine for all standard usages. Some specific cases of
Davor Ocelice9ed2812017-12-25 17:49:28 +01007333 long captures or JSON-formatted logs may require larger values.
Willy Tarreau18324f52014-06-27 18:10:07 +02007334
Frédéric Lécailled690dfa2019-04-25 10:52:17 +02007335 <ranges> A list of comma-separated ranges to identify the logs to sample.
7336 This is used to balance the load of the logs to send to the log
7337 server. The limits of the ranges cannot be null. They are numbered
7338 from 1. The size or period (in number of logs) of the sample must
7339 be set with <sample_size> parameter.
7340
7341 <sample_size>
7342 The size of the sample in number of logs to consider when balancing
7343 their logging loads. It is used to balance the load of the logs to
7344 send to the syslog server. This size must be greater or equal to the
7345 maximum of the high limits of the ranges.
7346 (see also <ranges> parameter).
7347
Willy Tarreauadb345d2018-11-12 07:56:13 +01007348 <format> is the log format used when generating syslog messages. It may be
7349 one of the following :
7350
Emeric Brun0237c4e2020-11-27 16:24:34 +01007351 local Analog to rfc3164 syslog message format except that hostname
7352 field is stripped. This is the default.
7353 Note: option "log-send-hostname" switches the default to
7354 rfc3164.
7355
7356 rfc3164 The RFC3164 syslog message format.
Willy Tarreauadb345d2018-11-12 07:56:13 +01007357 (https://tools.ietf.org/html/rfc3164)
7358
7359 rfc5424 The RFC5424 syslog message format.
7360 (https://tools.ietf.org/html/rfc5424)
7361
Emeric Brun54648852020-07-06 15:54:06 +02007362 priority A message containing only a level plus syslog facility between
7363 angle brackets such as '<63>', followed by the text. The PID,
7364 date, time, process name and system name are omitted. This is
7365 designed to be used with a local log server.
7366
Willy Tarreaue8746a02018-11-12 08:45:00 +01007367 short A message containing only a level between angle brackets such as
7368 '<3>', followed by the text. The PID, date, time, process name
7369 and system name are omitted. This is designed to be used with a
7370 local log server. This format is compatible with what the
7371 systemd logger consumes.
7372
Emeric Brun54648852020-07-06 15:54:06 +02007373 timed A message containing only a level between angle brackets such as
7374 '<3>', followed by ISO date and by the text. The PID, process
7375 name and system name are omitted. This is designed to be
7376 used with a local log server.
7377
7378 iso A message containing only the ISO date, followed by the text.
7379 The PID, process name and system name are omitted. This is
7380 designed to be used with a local log server.
7381
Willy Tarreauc1b06452018-11-12 11:57:56 +01007382 raw A message containing only the text. The level, PID, date, time,
7383 process name and system name are omitted. This is designed to
7384 be used in containers or during development, where the severity
7385 only depends on the file descriptor used (stdout/stderr).
7386
Willy Tarreau2769aa02007-12-27 18:26:09 +01007387 <facility> must be one of the 24 standard syslog facilities :
7388
Willy Tarreaue8746a02018-11-12 08:45:00 +01007389 kern user mail daemon auth syslog lpr news
7390 uucp cron auth2 ftp ntp audit alert cron2
7391 local0 local1 local2 local3 local4 local5 local6 local7
7392
Willy Tarreauc1b06452018-11-12 11:57:56 +01007393 Note that the facility is ignored for the "short" and "raw"
7394 formats, but still required as a positional field. It is
7395 recommended to use "daemon" in this case to make it clear that
7396 it's only supposed to be used locally.
Willy Tarreau2769aa02007-12-27 18:26:09 +01007397
7398 <level> is optional and can be specified to filter outgoing messages. By
7399 default, all messages are sent. If a level is specified, only
7400 messages with a severity at least as important as this level
Willy Tarreauf7edefa2009-05-10 17:20:05 +02007401 will be sent. An optional minimum level can be specified. If it
7402 is set, logs emitted with a more severe level than this one will
7403 be capped to this level. This is used to avoid sending "emerg"
7404 messages on all terminals on some default syslog configurations.
7405 Eight levels are known :
Willy Tarreau2769aa02007-12-27 18:26:09 +01007406
7407 emerg alert crit err warning notice info debug
7408
William Lallemand0f99e342011-10-12 17:50:54 +02007409 It is important to keep in mind that it is the frontend which decides what to
7410 log from a connection, and that in case of content switching, the log entries
7411 from the backend will be ignored. Connections are logged at level "info".
Willy Tarreaucc6c8912009-02-22 10:53:55 +01007412
7413 However, backend log declaration define how and where servers status changes
7414 will be logged. Level "notice" will be used to indicate a server going up,
7415 "warning" will be used for termination signals and definitive service
7416 termination, and "alert" will be used for when a server goes down.
7417
7418 Note : According to RFC3164, messages are truncated to 1024 bytes before
7419 being emitted.
Willy Tarreau2769aa02007-12-27 18:26:09 +01007420
7421 Example :
7422 log global
Willy Tarreauc1b06452018-11-12 11:57:56 +01007423 log stdout format short daemon # send log to systemd
7424 log stdout format raw daemon # send everything to stdout
7425 log stderr format raw daemon notice # send important events to stderr
Willy Tarreauf7edefa2009-05-10 17:20:05 +02007426 log 127.0.0.1:514 local0 notice # only send important events
7427 log 127.0.0.1:514 local0 notice notice # same but limit output level
William Lallemandb2f07452015-05-12 14:27:13 +02007428 log "${LOCAL_SYSLOG}:514" local0 notice # send to local server
Willy Tarreaudad36a32013-03-11 01:20:04 +01007429
Willy Tarreau2769aa02007-12-27 18:26:09 +01007430
William Lallemand48940402012-01-30 16:47:22 +01007431log-format <string>
Willy Tarreaufb4e7ea2015-01-07 14:55:17 +01007432 Specifies the log format string to use for traffic logs
7433 May be used in sections: defaults | frontend | listen | backend
7434 yes | yes | yes | no
William Lallemand48940402012-01-30 16:47:22 +01007435
Willy Tarreaufb4e7ea2015-01-07 14:55:17 +01007436 This directive specifies the log format string that will be used for all logs
7437 resulting from traffic passing through the frontend using this line. If the
7438 directive is used in a defaults section, all subsequent frontends will use
7439 the same log format. Please see section 8.2.4 which covers the log format
7440 string in depth.
William Lallemand48940402012-01-30 16:47:22 +01007441
Guillaume de Lafond29f45602017-03-31 19:52:15 +02007442 "log-format" directive overrides previous "option tcplog", "log-format" and
7443 "option httplog" directives.
7444
Dragan Dosen7ad31542015-09-28 17:16:47 +02007445log-format-sd <string>
7446 Specifies the RFC5424 structured-data log format string
7447 May be used in sections: defaults | frontend | listen | backend
7448 yes | yes | yes | no
7449
7450 This directive specifies the RFC5424 structured-data log format string that
7451 will be used for all logs resulting from traffic passing through the frontend
7452 using this line. If the directive is used in a defaults section, all
7453 subsequent frontends will use the same log format. Please see section 8.2.4
7454 which covers the log format string in depth.
7455
7456 See https://tools.ietf.org/html/rfc5424#section-6.3 for more information
7457 about the RFC5424 structured-data part.
7458
7459 Note : This log format string will be used only for loggers that have set
7460 log format to "rfc5424".
7461
7462 Example :
7463 log-format-sd [exampleSDID@1234\ bytes=\"%B\"\ status=\"%ST\"]
7464
7465
Willy Tarreau094af4e2015-01-07 15:03:42 +01007466log-tag <string>
7467 Specifies the log tag to use for all outgoing logs
7468 May be used in sections: defaults | frontend | listen | backend
7469 yes | yes | yes | yes
7470
7471 Sets the tag field in the syslog header to this string. It defaults to the
7472 log-tag set in the global section, otherwise the program name as launched
7473 from the command line, which usually is "haproxy". Sometimes it can be useful
7474 to differentiate between multiple processes running on the same host, or to
7475 differentiate customer instances running in the same process. In the backend,
7476 logs about servers up/down will use this tag. As a hint, it can be convenient
7477 to set a log-tag related to a hosted customer in a defaults section then put
7478 all the frontends and backends for that customer, then start another customer
7479 in a new defaults section. See also the global "log-tag" directive.
Willy Tarreau2769aa02007-12-27 18:26:09 +01007480
Willy Tarreauc35362a2014-04-25 13:58:37 +02007481max-keep-alive-queue <value>
7482 Set the maximum server queue size for maintaining keep-alive connections
7483 May be used in sections: defaults | frontend | listen | backend
7484 yes | no | yes | yes
7485
7486 HTTP keep-alive tries to reuse the same server connection whenever possible,
7487 but sometimes it can be counter-productive, for example if a server has a lot
7488 of connections while other ones are idle. This is especially true for static
7489 servers.
7490
7491 The purpose of this setting is to set a threshold on the number of queued
7492 connections at which haproxy stops trying to reuse the same server and prefers
7493 to find another one. The default value, -1, means there is no limit. A value
7494 of zero means that keep-alive requests will never be queued. For very close
7495 servers which can be reached with a low latency and which are not sensible to
Davor Ocelice9ed2812017-12-25 17:49:28 +01007496 breaking keep-alive, a low value is recommended (e.g. local static server can
Willy Tarreauc35362a2014-04-25 13:58:37 +02007497 use a value of 10 or less). For remote servers suffering from a high latency,
7498 higher values might be needed to cover for the latency and/or the cost of
7499 picking a different server.
7500
7501 Note that this has no impact on responses which are maintained to the same
7502 server consecutively to a 401 response. They will still go to the same server
7503 even if they have to be queued.
7504
7505 See also : "option http-server-close", "option prefer-last-server", server
7506 "maxconn" and cookie persistence.
7507
Olivier Houcharda4d4fdf2018-12-14 19:27:06 +01007508max-session-srv-conns <nb>
7509 Set the maximum number of outgoing connections we can keep idling for a given
7510 client session. The default is 5 (it precisely equals MAX_SRV_LIST which is
7511 defined at build time).
Willy Tarreauc35362a2014-04-25 13:58:37 +02007512
Willy Tarreau2769aa02007-12-27 18:26:09 +01007513maxconn <conns>
7514 Fix the maximum number of concurrent connections on a frontend
7515 May be used in sections : defaults | frontend | listen | backend
7516 yes | yes | yes | no
7517 Arguments :
7518 <conns> is the maximum number of concurrent connections the frontend will
7519 accept to serve. Excess connections will be queued by the system
7520 in the socket's listen queue and will be served once a connection
7521 closes.
7522
7523 If the system supports it, it can be useful on big sites to raise this limit
7524 very high so that haproxy manages connection queues, instead of leaving the
7525 clients with unanswered connection attempts. This value should not exceed the
7526 global maxconn. Also, keep in mind that a connection contains two buffers
Baptiste Assmann79fb45d2016-03-06 23:34:31 +01007527 of tune.bufsize (16kB by default) each, as well as some other data resulting
7528 in about 33 kB of RAM being consumed per established connection. That means
7529 that a medium system equipped with 1GB of RAM can withstand around
7530 20000-25000 concurrent connections if properly tuned.
Willy Tarreau2769aa02007-12-27 18:26:09 +01007531
7532 Also, when <conns> is set to large values, it is possible that the servers
7533 are not sized to accept such loads, and for this reason it is generally wise
7534 to assign them some reasonable connection limits.
7535
Willy Tarreauc8d5b952019-02-27 17:25:52 +01007536 When this value is set to zero, which is the default, the global "maxconn"
7537 value is used.
Vincent Bernat6341be52012-06-27 17:18:30 +02007538
Willy Tarreau2769aa02007-12-27 18:26:09 +01007539 See also : "server", global section's "maxconn", "fullconn"
7540
7541
Willy Tarreau77e0dae2020-10-14 15:44:27 +02007542mode { tcp|http }
Willy Tarreau2769aa02007-12-27 18:26:09 +01007543 Set the running mode or protocol of the instance
7544 May be used in sections : defaults | frontend | listen | backend
7545 yes | yes | yes | yes
7546 Arguments :
7547 tcp The instance will work in pure TCP mode. A full-duplex connection
7548 will be established between clients and servers, and no layer 7
7549 examination will be performed. This is the default mode. It
7550 should be used for SSL, SSH, SMTP, ...
7551
7552 http The instance will work in HTTP mode. The client request will be
7553 analyzed in depth before connecting to any server. Any request
7554 which is not RFC-compliant will be rejected. Layer 7 filtering,
7555 processing and switching will be possible. This is the mode which
7556 brings HAProxy most of its value.
7557
Cyril Bonté108cf6e2012-04-21 23:30:29 +02007558 When doing content switching, it is mandatory that the frontend and the
7559 backend are in the same mode (generally HTTP), otherwise the configuration
7560 will be refused.
Willy Tarreau2769aa02007-12-27 18:26:09 +01007561
Cyril Bonté108cf6e2012-04-21 23:30:29 +02007562 Example :
Willy Tarreau2769aa02007-12-27 18:26:09 +01007563 defaults http_instances
7564 mode http
7565
Willy Tarreau0ba27502007-12-24 16:55:16 +01007566
Cyril Bontéf0c60612010-02-06 14:44:47 +01007567monitor fail { if | unless } <condition>
Willy Tarreau2769aa02007-12-27 18:26:09 +01007568 Add a condition to report a failure to a monitor HTTP request.
Willy Tarreau0ba27502007-12-24 16:55:16 +01007569 May be used in sections : defaults | frontend | listen | backend
7570 no | yes | yes | no
Willy Tarreau0ba27502007-12-24 16:55:16 +01007571 Arguments :
7572 if <cond> the monitor request will fail if the condition is satisfied,
7573 and will succeed otherwise. The condition should describe a
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01007574 combined test which must induce a failure if all conditions
Willy Tarreau0ba27502007-12-24 16:55:16 +01007575 are met, for instance a low number of servers both in a
7576 backend and its backup.
7577
7578 unless <cond> the monitor request will succeed only if the condition is
7579 satisfied, and will fail otherwise. Such a condition may be
7580 based on a test on the presence of a minimum number of active
7581 servers in a list of backends.
7582
7583 This statement adds a condition which can force the response to a monitor
7584 request to report a failure. By default, when an external component queries
7585 the URI dedicated to monitoring, a 200 response is returned. When one of the
7586 conditions above is met, haproxy will return 503 instead of 200. This is
7587 very useful to report a site failure to an external component which may base
7588 routing advertisements between multiple sites on the availability reported by
7589 haproxy. In this case, one would rely on an ACL involving the "nbsrv"
Willy Tarreauae94d4d2011-05-11 16:28:49 +02007590 criterion. Note that "monitor fail" only works in HTTP mode. Both status
7591 messages may be tweaked using "errorfile" or "errorloc" if needed.
Willy Tarreau0ba27502007-12-24 16:55:16 +01007592
7593 Example:
7594 frontend www
Willy Tarreau2769aa02007-12-27 18:26:09 +01007595 mode http
Willy Tarreau0ba27502007-12-24 16:55:16 +01007596 acl site_dead nbsrv(dynamic) lt 2
7597 acl site_dead nbsrv(static) lt 2
7598 monitor-uri /site_alive
7599 monitor fail if site_dead
7600
Willy Tarreau9e9919d2020-10-14 15:55:23 +02007601 See also : "monitor-uri", "errorfile", "errorloc"
Willy Tarreau2769aa02007-12-27 18:26:09 +01007602
7603
Willy Tarreau2769aa02007-12-27 18:26:09 +01007604monitor-uri <uri>
7605 Intercept a URI used by external components' monitor requests
7606 May be used in sections : defaults | frontend | listen | backend
7607 yes | yes | yes | no
7608 Arguments :
7609 <uri> is the exact URI which we want to intercept to return HAProxy's
7610 health status instead of forwarding the request.
7611
7612 When an HTTP request referencing <uri> will be received on a frontend,
7613 HAProxy will not forward it nor log it, but instead will return either
7614 "HTTP/1.0 200 OK" or "HTTP/1.0 503 Service unavailable", depending on failure
7615 conditions defined with "monitor fail". This is normally enough for any
7616 front-end HTTP probe to detect that the service is UP and running without
7617 forwarding the request to a backend server. Note that the HTTP method, the
7618 version and all headers are ignored, but the request must at least be valid
7619 at the HTTP level. This keyword may only be used with an HTTP-mode frontend.
7620
Willy Tarreau721d8e02017-12-01 18:25:08 +01007621 Monitor requests are processed very early, just after the request is parsed
Christopher Faulet87f1f3d2019-07-18 14:51:20 +02007622 and even before any "http-request". The only rulesets applied before are the
7623 tcp-request ones. They cannot be logged either, and it is the intended
7624 purpose. They are only used to report HAProxy's health to an upper component,
7625 nothing more. However, it is possible to add any number of conditions using
7626 "monitor fail" and ACLs so that the result can be adjusted to whatever check
7627 can be imagined (most often the number of available servers in a backend).
Willy Tarreau2769aa02007-12-27 18:26:09 +01007628
Christopher Faulet6072beb2020-02-18 15:34:58 +01007629 Note: if <uri> starts by a slash ('/'), the matching is performed against the
7630 request's path instead of the request's uri. It is a workaround to let
7631 the HTTP/2 requests match the monitor-uri. Indeed, in HTTP/2, clients
7632 are encouraged to send absolute URIs only.
7633
Willy Tarreau2769aa02007-12-27 18:26:09 +01007634 Example :
7635 # Use /haproxy_test to report haproxy's status
7636 frontend www
7637 mode http
7638 monitor-uri /haproxy_test
7639
Willy Tarreau9e9919d2020-10-14 15:55:23 +02007640 See also : "monitor fail"
Willy Tarreau2769aa02007-12-27 18:26:09 +01007641
Willy Tarreau0ba27502007-12-24 16:55:16 +01007642
Willy Tarreaubf1f8162007-12-28 17:42:56 +01007643option abortonclose
7644no option abortonclose
7645 Enable or disable early dropping of aborted requests pending in queues.
7646 May be used in sections : defaults | frontend | listen | backend
7647 yes | no | yes | yes
7648 Arguments : none
7649
7650 In presence of very high loads, the servers will take some time to respond.
7651 The per-instance connection queue will inflate, and the response time will
7652 increase respective to the size of the queue times the average per-session
7653 response time. When clients will wait for more than a few seconds, they will
Willy Tarreau198a7442008-01-17 12:05:32 +01007654 often hit the "STOP" button on their browser, leaving a useless request in
Willy Tarreaubf1f8162007-12-28 17:42:56 +01007655 the queue, and slowing down other users, and the servers as well, because the
7656 request will eventually be served, then aborted at the first error
7657 encountered while delivering the response.
7658
7659 As there is no way to distinguish between a full STOP and a simple output
7660 close on the client side, HTTP agents should be conservative and consider
7661 that the client might only have closed its output channel while waiting for
7662 the response. However, this introduces risks of congestion when lots of users
7663 do the same, and is completely useless nowadays because probably no client at
7664 all will close the session while waiting for the response. Some HTTP agents
Davor Ocelice9ed2812017-12-25 17:49:28 +01007665 support this behavior (Squid, Apache, HAProxy), and others do not (TUX, most
Willy Tarreaubf1f8162007-12-28 17:42:56 +01007666 hardware-based load balancers). So the probability for a closed input channel
Willy Tarreau198a7442008-01-17 12:05:32 +01007667 to represent a user hitting the "STOP" button is close to 100%, and the risk
Willy Tarreaubf1f8162007-12-28 17:42:56 +01007668 of being the single component to break rare but valid traffic is extremely
7669 low, which adds to the temptation to be able to abort a session early while
7670 still not served and not pollute the servers.
7671
Davor Ocelice9ed2812017-12-25 17:49:28 +01007672 In HAProxy, the user can choose the desired behavior using the option
7673 "abortonclose". By default (without the option) the behavior is HTTP
Willy Tarreaubf1f8162007-12-28 17:42:56 +01007674 compliant and aborted requests will be served. But when the option is
7675 specified, a session with an incoming channel closed will be aborted while
7676 it is still possible, either pending in the queue for a connection slot, or
7677 during the connection establishment if the server has not yet acknowledged
7678 the connection request. This considerably reduces the queue size and the load
7679 on saturated servers when users are tempted to click on STOP, which in turn
Willy Tarreaud72758d2010-01-12 10:42:19 +01007680 reduces the response time for other users.
Willy Tarreaubf1f8162007-12-28 17:42:56 +01007681
7682 If this option has been enabled in a "defaults" section, it can be disabled
7683 in a specific instance by prepending the "no" keyword before it.
7684
7685 See also : "timeout queue" and server's "maxconn" and "maxqueue" parameters
7686
7687
Willy Tarreau4076a152009-04-02 15:18:36 +02007688option accept-invalid-http-request
7689no option accept-invalid-http-request
7690 Enable or disable relaxing of HTTP request parsing
7691 May be used in sections : defaults | frontend | listen | backend
7692 yes | yes | yes | no
7693 Arguments : none
7694
Willy Tarreau91852eb2015-05-01 13:26:00 +02007695 By default, HAProxy complies with RFC7230 in terms of message parsing. This
Willy Tarreau4076a152009-04-02 15:18:36 +02007696 means that invalid characters in header names are not permitted and cause an
Davor Ocelice9ed2812017-12-25 17:49:28 +01007697 error to be returned to the client. This is the desired behavior as such
Willy Tarreau4076a152009-04-02 15:18:36 +02007698 forbidden characters are essentially used to build attacks exploiting server
7699 weaknesses, and bypass security filtering. Sometimes, a buggy browser or
7700 server will emit invalid header names for whatever reason (configuration,
7701 implementation) and the issue will not be immediately fixed. In such a case,
7702 it is possible to relax HAProxy's header name parser to accept any character
Willy Tarreau422246e2012-01-07 23:54:13 +01007703 even if that does not make sense, by specifying this option. Similarly, the
7704 list of characters allowed to appear in a URI is well defined by RFC3986, and
7705 chars 0-31, 32 (space), 34 ('"'), 60 ('<'), 62 ('>'), 92 ('\'), 94 ('^'), 96
7706 ('`'), 123 ('{'), 124 ('|'), 125 ('}'), 127 (delete) and anything above are
Davor Ocelice9ed2812017-12-25 17:49:28 +01007707 not allowed at all. HAProxy always blocks a number of them (0..32, 127). The
Willy Tarreau91852eb2015-05-01 13:26:00 +02007708 remaining ones are blocked by default unless this option is enabled. This
Willy Tarreau13317662015-05-01 13:47:08 +02007709 option also relaxes the test on the HTTP version, it allows HTTP/0.9 requests
7710 to pass through (no version specified) and multiple digits for both the major
7711 and the minor version.
Willy Tarreau4076a152009-04-02 15:18:36 +02007712
7713 This option should never be enabled by default as it hides application bugs
7714 and open security breaches. It should only be deployed after a problem has
7715 been confirmed.
7716
7717 When this option is enabled, erroneous header names will still be accepted in
7718 requests, but the complete request will be captured in order to permit later
Willy Tarreau422246e2012-01-07 23:54:13 +01007719 analysis using the "show errors" request on the UNIX stats socket. Similarly,
7720 requests containing invalid chars in the URI part will be logged. Doing this
Willy Tarreau4076a152009-04-02 15:18:36 +02007721 also helps confirming that the issue has been solved.
7722
7723 If this option has been enabled in a "defaults" section, it can be disabled
7724 in a specific instance by prepending the "no" keyword before it.
7725
7726 See also : "option accept-invalid-http-response" and "show errors" on the
7727 stats socket.
7728
7729
7730option accept-invalid-http-response
7731no option accept-invalid-http-response
7732 Enable or disable relaxing of HTTP response parsing
7733 May be used in sections : defaults | frontend | listen | backend
7734 yes | no | yes | yes
7735 Arguments : none
7736
Willy Tarreau91852eb2015-05-01 13:26:00 +02007737 By default, HAProxy complies with RFC7230 in terms of message parsing. This
Willy Tarreau4076a152009-04-02 15:18:36 +02007738 means that invalid characters in header names are not permitted and cause an
Davor Ocelice9ed2812017-12-25 17:49:28 +01007739 error to be returned to the client. This is the desired behavior as such
Willy Tarreau4076a152009-04-02 15:18:36 +02007740 forbidden characters are essentially used to build attacks exploiting server
7741 weaknesses, and bypass security filtering. Sometimes, a buggy browser or
7742 server will emit invalid header names for whatever reason (configuration,
7743 implementation) and the issue will not be immediately fixed. In such a case,
7744 it is possible to relax HAProxy's header name parser to accept any character
Willy Tarreau91852eb2015-05-01 13:26:00 +02007745 even if that does not make sense, by specifying this option. This option also
7746 relaxes the test on the HTTP version format, it allows multiple digits for
7747 both the major and the minor version.
Willy Tarreau4076a152009-04-02 15:18:36 +02007748
7749 This option should never be enabled by default as it hides application bugs
7750 and open security breaches. It should only be deployed after a problem has
7751 been confirmed.
7752
7753 When this option is enabled, erroneous header names will still be accepted in
7754 responses, but the complete response will be captured in order to permit
7755 later analysis using the "show errors" request on the UNIX stats socket.
7756 Doing this also helps confirming that the issue has been solved.
7757
7758 If this option has been enabled in a "defaults" section, it can be disabled
7759 in a specific instance by prepending the "no" keyword before it.
7760
7761 See also : "option accept-invalid-http-request" and "show errors" on the
7762 stats socket.
7763
7764
Willy Tarreaubf1f8162007-12-28 17:42:56 +01007765option allbackups
7766no option allbackups
7767 Use either all backup servers at a time or only the first one
7768 May be used in sections : defaults | frontend | listen | backend
7769 yes | no | yes | yes
7770 Arguments : none
7771
7772 By default, the first operational backup server gets all traffic when normal
7773 servers are all down. Sometimes, it may be preferred to use multiple backups
7774 at once, because one will not be enough. When "option allbackups" is enabled,
7775 the load balancing will be performed among all backup servers when all normal
7776 ones are unavailable. The same load balancing algorithm will be used and the
7777 servers' weights will be respected. Thus, there will not be any priority
7778 order between the backup servers anymore.
7779
7780 This option is mostly used with static server farms dedicated to return a
7781 "sorry" page when an application is completely offline.
7782
7783 If this option has been enabled in a "defaults" section, it can be disabled
7784 in a specific instance by prepending the "no" keyword before it.
7785
7786
7787option checkcache
7788no option checkcache
Godbach7056a352013-12-11 20:01:07 +08007789 Analyze all server responses and block responses with cacheable cookies
Willy Tarreaubf1f8162007-12-28 17:42:56 +01007790 May be used in sections : defaults | frontend | listen | backend
7791 yes | no | yes | yes
7792 Arguments : none
7793
7794 Some high-level frameworks set application cookies everywhere and do not
7795 always let enough control to the developer to manage how the responses should
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01007796 be cached. When a session cookie is returned on a cacheable object, there is a
Willy Tarreaubf1f8162007-12-28 17:42:56 +01007797 high risk of session crossing or stealing between users traversing the same
7798 caches. In some situations, it is better to block the response than to let
Willy Tarreau3c92c5f2011-08-28 09:45:47 +02007799 some sensitive session information go in the wild.
Willy Tarreaubf1f8162007-12-28 17:42:56 +01007800
7801 The option "checkcache" enables deep inspection of all server responses for
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01007802 strict compliance with HTTP specification in terms of cacheability. It
Willy Tarreau198a7442008-01-17 12:05:32 +01007803 carefully checks "Cache-control", "Pragma" and "Set-cookie" headers in server
Willy Tarreaubf1f8162007-12-28 17:42:56 +01007804 response to check if there's a risk of caching a cookie on a client-side
7805 proxy. When this option is enabled, the only responses which can be delivered
Willy Tarreau198a7442008-01-17 12:05:32 +01007806 to the client are :
Davor Ocelice9ed2812017-12-25 17:49:28 +01007807 - all those without "Set-Cookie" header;
Willy Tarreauc55ddce2017-12-21 11:41:38 +01007808 - all those with a return code other than 200, 203, 204, 206, 300, 301,
7809 404, 405, 410, 414, 501, provided that the server has not set a
Davor Ocelice9ed2812017-12-25 17:49:28 +01007810 "Cache-control: public" header field;
Willy Tarreau24ea0bc2017-12-21 11:32:55 +01007811 - all those that result from a request using a method other than GET, HEAD,
7812 OPTIONS, TRACE, provided that the server has not set a 'Cache-Control:
Davor Ocelice9ed2812017-12-25 17:49:28 +01007813 public' header field;
Willy Tarreaubf1f8162007-12-28 17:42:56 +01007814 - those with a 'Pragma: no-cache' header
7815 - those with a 'Cache-control: private' header
7816 - those with a 'Cache-control: no-store' header
7817 - those with a 'Cache-control: max-age=0' header
7818 - those with a 'Cache-control: s-maxage=0' header
7819 - those with a 'Cache-control: no-cache' header
7820 - those with a 'Cache-control: no-cache="set-cookie"' header
7821 - those with a 'Cache-control: no-cache="set-cookie,' header
7822 (allowing other fields after set-cookie)
7823
7824 If a response doesn't respect these requirements, then it will be blocked
Christopher Faulet87f1f3d2019-07-18 14:51:20 +02007825 just as if it was from an "http-response deny" rule, with an "HTTP 502 bad
7826 gateway". The session state shows "PH--" meaning that the proxy blocked the
7827 response during headers processing. Additionally, an alert will be sent in
7828 the logs so that admins are informed that there's something to be fixed.
Willy Tarreaubf1f8162007-12-28 17:42:56 +01007829
7830 Due to the high impact on the application, the application should be tested
7831 in depth with the option enabled before going to production. It is also a
Willy Tarreaud2a4aa22008-01-31 15:28:22 +01007832 good practice to always activate it during tests, even if it is not used in
Davor Ocelice9ed2812017-12-25 17:49:28 +01007833 production, as it will report potentially dangerous application behaviors.
Willy Tarreaubf1f8162007-12-28 17:42:56 +01007834
7835 If this option has been enabled in a "defaults" section, it can be disabled
7836 in a specific instance by prepending the "no" keyword before it.
7837
7838
7839option clitcpka
7840no option clitcpka
7841 Enable or disable the sending of TCP keepalive packets on the client side
7842 May be used in sections : defaults | frontend | listen | backend
7843 yes | yes | yes | no
7844 Arguments : none
7845
7846 When there is a firewall or any session-aware component between a client and
7847 a server, and when the protocol involves very long sessions with long idle
Davor Ocelice9ed2812017-12-25 17:49:28 +01007848 periods (e.g. remote desktops), there is a risk that one of the intermediate
Willy Tarreaubf1f8162007-12-28 17:42:56 +01007849 components decides to expire a session which has remained idle for too long.
7850
7851 Enabling socket-level TCP keep-alives makes the system regularly send packets
7852 to the other end of the connection, leaving it active. The delay between
7853 keep-alive probes is controlled by the system only and depends both on the
7854 operating system and its tuning parameters.
7855
7856 It is important to understand that keep-alive packets are neither emitted nor
7857 received at the application level. It is only the network stacks which sees
7858 them. For this reason, even if one side of the proxy already uses keep-alives
7859 to maintain its connection alive, those keep-alive packets will not be
7860 forwarded to the other side of the proxy.
7861
7862 Please note that this has nothing to do with HTTP keep-alive.
7863
7864 Using option "clitcpka" enables the emission of TCP keep-alive probes on the
7865 client side of a connection, which should help when session expirations are
7866 noticed between HAProxy and a client.
7867
7868 If this option has been enabled in a "defaults" section, it can be disabled
7869 in a specific instance by prepending the "no" keyword before it.
7870
7871 See also : "option srvtcpka", "option tcpka"
7872
7873
Willy Tarreau0ba27502007-12-24 16:55:16 +01007874option contstats
7875 Enable continuous traffic statistics updates
7876 May be used in sections : defaults | frontend | listen | backend
7877 yes | yes | yes | no
7878 Arguments : none
7879
7880 By default, counters used for statistics calculation are incremented
7881 only when a session finishes. It works quite well when serving small
7882 objects, but with big ones (for example large images or archives) or
7883 with A/V streaming, a graph generated from haproxy counters looks like
Willy Tarreaudef0d222016-11-08 22:03:00 +01007884 a hedgehog. With this option enabled counters get incremented frequently
7885 along the session, typically every 5 seconds, which is often enough to
7886 produce clean graphs. Recounting touches a hotpath directly so it is not
7887 not enabled by default, as it can cause a lot of wakeups for very large
7888 session counts and cause a small performance drop.
Willy Tarreau0ba27502007-12-24 16:55:16 +01007889
Christopher Faulet89aed322020-06-02 17:33:56 +02007890option disable-h2-upgrade
7891no option disable-h2-upgrade
7892 Enable or disable the implicit HTTP/2 upgrade from an HTTP/1.x client
7893 connection.
7894 May be used in sections : defaults | frontend | listen | backend
7895 yes | yes | yes | no
7896 Arguments : none
7897
7898 By default, HAProxy is able to implicitly upgrade an HTTP/1.x client
7899 connection to an HTTP/2 connection if the first request it receives from a
7900 given HTTP connection matches the HTTP/2 connection preface (i.e. the string
7901 "PRI * HTTP/2.0\r\n\r\nSM\r\n\r\n"). This way, it is possible to support
7902 HTTP/1.x and HTTP/2 clients on a non-SSL connections. This option must be used to
7903 disable the implicit upgrade. Note this implicit upgrade is only supported
7904 for HTTP proxies, thus this option too. Note also it is possible to force the
7905 HTTP/2 on clear connections by specifying "proto h2" on the bind line.
7906
7907 If this option has been enabled in a "defaults" section, it can be disabled
7908 in a specific instance by prepending the "no" keyword before it.
Willy Tarreau0ba27502007-12-24 16:55:16 +01007909
Willy Tarreauc9bd0cc2009-05-10 11:57:02 +02007910option dontlog-normal
7911no option dontlog-normal
7912 Enable or disable logging of normal, successful connections
7913 May be used in sections : defaults | frontend | listen | backend
7914 yes | yes | yes | no
7915 Arguments : none
7916
7917 There are large sites dealing with several thousand connections per second
7918 and for which logging is a major pain. Some of them are even forced to turn
7919 logs off and cannot debug production issues. Setting this option ensures that
7920 normal connections, those which experience no error, no timeout, no retry nor
7921 redispatch, will not be logged. This leaves disk space for anomalies. In HTTP
7922 mode, the response status code is checked and return codes 5xx will still be
7923 logged.
7924
7925 It is strongly discouraged to use this option as most of the time, the key to
7926 complex issues is in the normal logs which will not be logged here. If you
7927 need to separate logs, see the "log-separate-errors" option instead.
7928
Willy Tarreauc57f0e22009-05-10 13:12:33 +02007929 See also : "log", "dontlognull", "log-separate-errors" and section 8 about
Willy Tarreauc9bd0cc2009-05-10 11:57:02 +02007930 logging.
7931
7932
Willy Tarreaubf1f8162007-12-28 17:42:56 +01007933option dontlognull
7934no option dontlognull
7935 Enable or disable logging of null connections
7936 May be used in sections : defaults | frontend | listen | backend
7937 yes | yes | yes | no
7938 Arguments : none
7939
7940 In certain environments, there are components which will regularly connect to
7941 various systems to ensure that they are still alive. It can be the case from
7942 another load balancer as well as from monitoring systems. By default, even a
7943 simple port probe or scan will produce a log. If those connections pollute
7944 the logs too much, it is possible to enable option "dontlognull" to indicate
7945 that a connection on which no data has been transferred will not be logged,
Willy Tarreau0f228a02015-05-01 15:37:53 +02007946 which typically corresponds to those probes. Note that errors will still be
7947 returned to the client and accounted for in the stats. If this is not what is
7948 desired, option http-ignore-probes can be used instead.
Willy Tarreaubf1f8162007-12-28 17:42:56 +01007949
7950 It is generally recommended not to use this option in uncontrolled
Davor Ocelice9ed2812017-12-25 17:49:28 +01007951 environments (e.g. internet), otherwise scans and other malicious activities
Willy Tarreaubf1f8162007-12-28 17:42:56 +01007952 would not be logged.
7953
7954 If this option has been enabled in a "defaults" section, it can be disabled
7955 in a specific instance by prepending the "no" keyword before it.
7956
Willy Tarreau9e9919d2020-10-14 15:55:23 +02007957 See also : "log", "http-ignore-probes", "monitor-uri", and
Willy Tarreau0f228a02015-05-01 15:37:53 +02007958 section 8 about logging.
Willy Tarreaubf1f8162007-12-28 17:42:56 +01007959
Willy Tarreaubf1f8162007-12-28 17:42:56 +01007960
Willy Tarreau87cf5142011-08-19 22:57:24 +02007961option forwardfor [ except <network> ] [ header <name> ] [ if-none ]
Willy Tarreauc27debf2008-01-06 08:57:02 +01007962 Enable insertion of the X-Forwarded-For header to requests sent to servers
7963 May be used in sections : defaults | frontend | listen | backend
7964 yes | yes | yes | yes
7965 Arguments :
7966 <network> is an optional argument used to disable this option for sources
7967 matching <network>
Ross Westaf72a1d2008-08-03 10:51:45 +02007968 <name> an optional argument to specify a different "X-Forwarded-For"
Willy Tarreaud72758d2010-01-12 10:42:19 +01007969 header name.
Willy Tarreauc27debf2008-01-06 08:57:02 +01007970
7971 Since HAProxy works in reverse-proxy mode, the servers see its IP address as
7972 their client address. This is sometimes annoying when the client's IP address
7973 is expected in server logs. To solve this problem, the well-known HTTP header
7974 "X-Forwarded-For" may be added by HAProxy to all requests sent to the server.
7975 This header contains a value representing the client's IP address. Since this
7976 header is always appended at the end of the existing header list, the server
7977 must be configured to always use the last occurrence of this header only. See
Ross Westaf72a1d2008-08-03 10:51:45 +02007978 the server's manual to find how to enable use of this standard header. Note
7979 that only the last occurrence of the header must be used, since it is really
7980 possible that the client has already brought one.
7981
Willy Tarreaud72758d2010-01-12 10:42:19 +01007982 The keyword "header" may be used to supply a different header name to replace
Ross Westaf72a1d2008-08-03 10:51:45 +02007983 the default "X-Forwarded-For". This can be useful where you might already
Davor Ocelice9ed2812017-12-25 17:49:28 +01007984 have a "X-Forwarded-For" header from a different application (e.g. stunnel),
Willy Tarreaud72758d2010-01-12 10:42:19 +01007985 and you need preserve it. Also if your backend server doesn't use the
Davor Ocelice9ed2812017-12-25 17:49:28 +01007986 "X-Forwarded-For" header and requires different one (e.g. Zeus Web Servers
Ross Westaf72a1d2008-08-03 10:51:45 +02007987 require "X-Cluster-Client-IP").
Willy Tarreauc27debf2008-01-06 08:57:02 +01007988
7989 Sometimes, a same HAProxy instance may be shared between a direct client
7990 access and a reverse-proxy access (for instance when an SSL reverse-proxy is
7991 used to decrypt HTTPS traffic). It is possible to disable the addition of the
7992 header for a known source address or network by adding the "except" keyword
7993 followed by the network address. In this case, any source IP matching the
7994 network will not cause an addition of this header. Most common uses are with
Christopher Faulet5d1def62021-02-26 09:19:15 +01007995 private networks or 127.0.0.1. IPv4 and IPv6 are both supported.
Willy Tarreauc27debf2008-01-06 08:57:02 +01007996
Willy Tarreau87cf5142011-08-19 22:57:24 +02007997 Alternatively, the keyword "if-none" states that the header will only be
7998 added if it is not present. This should only be used in perfectly trusted
7999 environment, as this might cause a security issue if headers reaching haproxy
8000 are under the control of the end-user.
8001
Willy Tarreauc27debf2008-01-06 08:57:02 +01008002 This option may be specified either in the frontend or in the backend. If at
Ross Westaf72a1d2008-08-03 10:51:45 +02008003 least one of them uses it, the header will be added. Note that the backend's
8004 setting of the header subargument takes precedence over the frontend's if
Willy Tarreau87cf5142011-08-19 22:57:24 +02008005 both are defined. In the case of the "if-none" argument, if at least one of
8006 the frontend or the backend does not specify it, it wants the addition to be
8007 mandatory, so it wins.
Willy Tarreauc27debf2008-01-06 08:57:02 +01008008
Olivier Doucetaa1ea8a2016-08-05 17:15:20 +02008009 Example :
Willy Tarreauc27debf2008-01-06 08:57:02 +01008010 # Public HTTP address also used by stunnel on the same machine
8011 frontend www
8012 mode http
8013 option forwardfor except 127.0.0.1 # stunnel already adds the header
8014
Ross Westaf72a1d2008-08-03 10:51:45 +02008015 # Those servers want the IP Address in X-Client
8016 backend www
8017 mode http
8018 option forwardfor header X-Client
8019
Willy Tarreau87cf5142011-08-19 22:57:24 +02008020 See also : "option httpclose", "option http-server-close",
Christopher Faulet315b39c2018-09-21 16:26:19 +02008021 "option http-keep-alive"
Willy Tarreauc27debf2008-01-06 08:57:02 +01008022
Willy Tarreau8a8e1d92010-04-05 16:15:16 +02008023
Christopher Faulet98fbe952019-07-22 16:18:24 +02008024option h1-case-adjust-bogus-client
8025no option h1-case-adjust-bogus-client
8026 Enable or disable the case adjustment of HTTP/1 headers sent to bogus clients
8027 May be used in sections : defaults | frontend | listen | backend
8028 yes | yes | yes | no
8029 Arguments : none
8030
8031 There is no standard case for header names because, as stated in RFC7230,
8032 they are case-insensitive. So applications must handle them in a case-
8033 insensitive manner. But some bogus applications violate the standards and
8034 erroneously rely on the cases most commonly used by browsers. This problem
8035 becomes critical with HTTP/2 because all header names must be exchanged in
8036 lower case, and HAProxy follows the same convention. All header names are
8037 sent in lower case to clients and servers, regardless of the HTTP version.
8038
8039 When HAProxy receives an HTTP/1 response, its header names are converted to
8040 lower case and manipulated and sent this way to the clients. If a client is
8041 known to violate the HTTP standards and to fail to process a response coming
8042 from HAProxy, it is possible to transform the lower case header names to a
8043 different format when the response is formatted and sent to the client, by
8044 enabling this option and specifying the list of headers to be reformatted
8045 using the global directives "h1-case-adjust" or "h1-case-adjust-file". This
8046 must only be a temporary workaround for the time it takes the client to be
8047 fixed, because clients which require such workarounds might be vulnerable to
8048 content smuggling attacks and must absolutely be fixed.
8049
8050 Please note that this option will not affect standards-compliant clients.
8051
8052 If this option has been enabled in a "defaults" section, it can be disabled
8053 in a specific instance by prepending the "no" keyword before it.
8054
8055 See also: "option h1-case-adjust-bogus-server", "h1-case-adjust",
8056 "h1-case-adjust-file".
8057
8058
8059option h1-case-adjust-bogus-server
8060no option h1-case-adjust-bogus-server
8061 Enable or disable the case adjustment of HTTP/1 headers sent to bogus servers
8062 May be used in sections : defaults | frontend | listen | backend
8063 yes | no | yes | yes
8064 Arguments : none
8065
8066 There is no standard case for header names because, as stated in RFC7230,
8067 they are case-insensitive. So applications must handle them in a case-
8068 insensitive manner. But some bogus applications violate the standards and
8069 erroneously rely on the cases most commonly used by browsers. This problem
8070 becomes critical with HTTP/2 because all header names must be exchanged in
8071 lower case, and HAProxy follows the same convention. All header names are
8072 sent in lower case to clients and servers, regardless of the HTTP version.
8073
8074 When HAProxy receives an HTTP/1 request, its header names are converted to
8075 lower case and manipulated and sent this way to the servers. If a server is
8076 known to violate the HTTP standards and to fail to process a request coming
8077 from HAProxy, it is possible to transform the lower case header names to a
8078 different format when the request is formatted and sent to the server, by
8079 enabling this option and specifying the list of headers to be reformatted
8080 using the global directives "h1-case-adjust" or "h1-case-adjust-file". This
8081 must only be a temporary workaround for the time it takes the server to be
8082 fixed, because servers which require such workarounds might be vulnerable to
8083 content smuggling attacks and must absolutely be fixed.
8084
8085 Please note that this option will not affect standards-compliant servers.
8086
8087 If this option has been enabled in a "defaults" section, it can be disabled
8088 in a specific instance by prepending the "no" keyword before it.
8089
8090 See also: "option h1-case-adjust-bogus-client", "h1-case-adjust",
8091 "h1-case-adjust-file".
8092
8093
Willy Tarreau9fbe18e2015-05-01 22:42:08 +02008094option http-buffer-request
8095no option http-buffer-request
8096 Enable or disable waiting for whole HTTP request body before proceeding
8097 May be used in sections : defaults | frontend | listen | backend
8098 yes | yes | yes | yes
8099 Arguments : none
8100
8101 It is sometimes desirable to wait for the body of an HTTP request before
8102 taking a decision. This is what is being done by "balance url_param" for
8103 example. The first use case is to buffer requests from slow clients before
8104 connecting to the server. Another use case consists in taking the routing
8105 decision based on the request body's contents. This option placed in a
8106 frontend or backend forces the HTTP processing to wait until either the whole
Christopher Faulet6db8a2e2019-11-19 16:27:25 +01008107 body is received or the request buffer is full. It can have undesired side
8108 effects with some applications abusing HTTP by expecting unbuffered
8109 transmissions between the frontend and the backend, so this should definitely
8110 not be used by default.
Willy Tarreau9fbe18e2015-05-01 22:42:08 +02008111
Baptiste Assmanneccdf432015-10-28 13:49:01 +01008112 See also : "option http-no-delay", "timeout http-request"
Willy Tarreau9fbe18e2015-05-01 22:42:08 +02008113
8114
Willy Tarreau0f228a02015-05-01 15:37:53 +02008115option http-ignore-probes
8116no option http-ignore-probes
8117 Enable or disable logging of null connections and request timeouts
8118 May be used in sections : defaults | frontend | listen | backend
8119 yes | yes | yes | no
8120 Arguments : none
8121
8122 Recently some browsers started to implement a "pre-connect" feature
8123 consisting in speculatively connecting to some recently visited web sites
8124 just in case the user would like to visit them. This results in many
8125 connections being established to web sites, which end up in 408 Request
8126 Timeout if the timeout strikes first, or 400 Bad Request when the browser
8127 decides to close them first. These ones pollute the log and feed the error
8128 counters. There was already "option dontlognull" but it's insufficient in
8129 this case. Instead, this option does the following things :
8130 - prevent any 400/408 message from being sent to the client if nothing
Davor Ocelice9ed2812017-12-25 17:49:28 +01008131 was received over a connection before it was closed;
8132 - prevent any log from being emitted in this situation;
Willy Tarreau0f228a02015-05-01 15:37:53 +02008133 - prevent any error counter from being incremented
8134
8135 That way the empty connection is silently ignored. Note that it is better
8136 not to use this unless it is clear that it is needed, because it will hide
8137 real problems. The most common reason for not receiving a request and seeing
8138 a 408 is due to an MTU inconsistency between the client and an intermediary
8139 element such as a VPN, which blocks too large packets. These issues are
8140 generally seen with POST requests as well as GET with large cookies. The logs
8141 are often the only way to detect them.
8142
8143 If this option has been enabled in a "defaults" section, it can be disabled
8144 in a specific instance by prepending the "no" keyword before it.
8145
8146 See also : "log", "dontlognull", "errorfile", and section 8 about logging.
8147
8148
Willy Tarreau16bfb022010-01-16 19:48:41 +01008149option http-keep-alive
8150no option http-keep-alive
8151 Enable or disable HTTP keep-alive from client to server
8152 May be used in sections : defaults | frontend | listen | backend
8153 yes | yes | yes | yes
8154 Arguments : none
8155
Willy Tarreau70dffda2014-01-30 03:07:23 +01008156 By default HAProxy operates in keep-alive mode with regards to persistent
8157 connections: for each connection it processes each request and response, and
Christopher Faulet315b39c2018-09-21 16:26:19 +02008158 leaves the connection idle on both sides between the end of a response and
8159 the start of a new request. This mode may be changed by several options such
Christopher Faulet159e6672019-07-16 15:09:52 +02008160 as "option http-server-close" or "option httpclose". This option allows to
8161 set back the keep-alive mode, which can be useful when another mode was used
8162 in a defaults section.
Willy Tarreau70dffda2014-01-30 03:07:23 +01008163
8164 Setting "option http-keep-alive" enables HTTP keep-alive mode on the client-
8165 and server- sides. This provides the lowest latency on the client side (slow
Willy Tarreau16bfb022010-01-16 19:48:41 +01008166 network) and the fastest session reuse on the server side at the expense
8167 of maintaining idle connections to the servers. In general, it is possible
8168 with this option to achieve approximately twice the request rate that the
8169 "http-server-close" option achieves on small objects. There are mainly two
8170 situations where this option may be useful :
8171
8172 - when the server is non-HTTP compliant and authenticates the connection
Davor Ocelice9ed2812017-12-25 17:49:28 +01008173 instead of requests (e.g. NTLM authentication)
Willy Tarreau16bfb022010-01-16 19:48:41 +01008174
8175 - when the cost of establishing the connection to the server is significant
8176 compared to the cost of retrieving the associated object from the server.
8177
8178 This last case can happen when the server is a fast static server of cache.
8179 In this case, the server will need to be properly tuned to support high enough
8180 connection counts because connections will last until the client sends another
8181 request.
8182
8183 If the client request has to go to another backend or another server due to
8184 content switching or the load balancing algorithm, the idle connection will
Willy Tarreau9420b122013-12-15 18:58:25 +01008185 immediately be closed and a new one re-opened. Option "prefer-last-server" is
8186 available to try optimize server selection so that if the server currently
8187 attached to an idle connection is usable, it will be used.
Willy Tarreau16bfb022010-01-16 19:48:41 +01008188
Willy Tarreau16bfb022010-01-16 19:48:41 +01008189 At the moment, logs will not indicate whether requests came from the same
8190 session or not. The accept date reported in the logs corresponds to the end
8191 of the previous request, and the request time corresponds to the time spent
8192 waiting for a new request. The keep-alive request time is still bound to the
8193 timeout defined by "timeout http-keep-alive" or "timeout http-request" if
8194 not set.
8195
Christopher Faulet159e6672019-07-16 15:09:52 +02008196 This option disables and replaces any previous "option httpclose" or "option
8197 http-server-close". When backend and frontend options differ, all of these 4
8198 options have precedence over "option http-keep-alive".
Willy Tarreau16bfb022010-01-16 19:48:41 +01008199
Christopher Faulet315b39c2018-09-21 16:26:19 +02008200 See also : "option httpclose",, "option http-server-close",
Willy Tarreau9420b122013-12-15 18:58:25 +01008201 "option prefer-last-server", "option http-pretend-keepalive",
Frédéric Lécaille93d33162019-03-06 09:35:59 +01008202 and "1.1. The HTTP transaction model".
Willy Tarreau16bfb022010-01-16 19:48:41 +01008203
8204
Willy Tarreau96e31212011-05-30 18:10:30 +02008205option http-no-delay
8206no option http-no-delay
8207 Instruct the system to favor low interactive delays over performance in HTTP
8208 May be used in sections : defaults | frontend | listen | backend
8209 yes | yes | yes | yes
8210 Arguments : none
8211
8212 In HTTP, each payload is unidirectional and has no notion of interactivity.
8213 Any agent is expected to queue data somewhat for a reasonably low delay.
8214 There are some very rare server-to-server applications that abuse the HTTP
8215 protocol and expect the payload phase to be highly interactive, with many
8216 interleaved data chunks in both directions within a single request. This is
8217 absolutely not supported by the HTTP specification and will not work across
8218 most proxies or servers. When such applications attempt to do this through
8219 haproxy, it works but they will experience high delays due to the network
8220 optimizations which favor performance by instructing the system to wait for
8221 enough data to be available in order to only send full packets. Typical
8222 delays are around 200 ms per round trip. Note that this only happens with
8223 abnormal uses. Normal uses such as CONNECT requests nor WebSockets are not
8224 affected.
8225
8226 When "option http-no-delay" is present in either the frontend or the backend
8227 used by a connection, all such optimizations will be disabled in order to
8228 make the exchanges as fast as possible. Of course this offers no guarantee on
8229 the functionality, as it may break at any other place. But if it works via
8230 HAProxy, it will work as fast as possible. This option should never be used
8231 by default, and should never be used at all unless such a buggy application
8232 is discovered. The impact of using this option is an increase of bandwidth
8233 usage and CPU usage, which may significantly lower performance in high
8234 latency environments.
8235
Willy Tarreau9fbe18e2015-05-01 22:42:08 +02008236 See also : "option http-buffer-request"
8237
Willy Tarreau96e31212011-05-30 18:10:30 +02008238
Willy Tarreau8a8e1d92010-04-05 16:15:16 +02008239option http-pretend-keepalive
8240no option http-pretend-keepalive
8241 Define whether haproxy will announce keepalive to the server or not
8242 May be used in sections : defaults | frontend | listen | backend
Christopher Faulet98db9762018-09-21 10:25:19 +02008243 yes | no | yes | yes
Willy Tarreau8a8e1d92010-04-05 16:15:16 +02008244 Arguments : none
8245
Christopher Faulet315b39c2018-09-21 16:26:19 +02008246 When running with "option http-server-close" or "option httpclose", haproxy
Willy Tarreau8a8e1d92010-04-05 16:15:16 +02008247 adds a "Connection: close" header to the request forwarded to the server.
8248 Unfortunately, when some servers see this header, they automatically refrain
8249 from using the chunked encoding for responses of unknown length, while this
8250 is totally unrelated. The immediate effect is that this prevents haproxy from
8251 maintaining the client connection alive. A second effect is that a client or
8252 a cache could receive an incomplete response without being aware of it, and
8253 consider the response complete.
8254
8255 By setting "option http-pretend-keepalive", haproxy will make the server
8256 believe it will keep the connection alive. The server will then not fall back
8257 to the abnormal undesired above. When haproxy gets the whole response, it
8258 will close the connection with the server just as it would do with the
Christopher Faulet315b39c2018-09-21 16:26:19 +02008259 "option httpclose". That way the client gets a normal response and the
Willy Tarreau8a8e1d92010-04-05 16:15:16 +02008260 connection is correctly closed on the server side.
8261
8262 It is recommended not to enable this option by default, because most servers
8263 will more efficiently close the connection themselves after the last packet,
8264 and release its buffers slightly earlier. Also, the added packet on the
8265 network could slightly reduce the overall peak performance. However it is
8266 worth noting that when this option is enabled, haproxy will have slightly
8267 less work to do. So if haproxy is the bottleneck on the whole architecture,
8268 enabling this option might save a few CPU cycles.
8269
Christopher Faulet98db9762018-09-21 10:25:19 +02008270 This option may be set in backend and listen sections. Using it in a frontend
8271 section will be ignored and a warning will be reported during startup. It is
8272 a backend related option, so there is no real reason to set it on a
8273 frontend. This option may be combined with "option httpclose", which will
8274 cause keepalive to be announced to the server and close to be announced to
8275 the client. This practice is discouraged though.
Willy Tarreau8a8e1d92010-04-05 16:15:16 +02008276
8277 If this option has been enabled in a "defaults" section, it can be disabled
8278 in a specific instance by prepending the "no" keyword before it.
8279
Christopher Faulet315b39c2018-09-21 16:26:19 +02008280 See also : "option httpclose", "option http-server-close", and
Willy Tarreau16bfb022010-01-16 19:48:41 +01008281 "option http-keep-alive"
Willy Tarreau8a8e1d92010-04-05 16:15:16 +02008282
Willy Tarreauc27debf2008-01-06 08:57:02 +01008283
Willy Tarreaub608feb2010-01-02 22:47:18 +01008284option http-server-close
8285no option http-server-close
8286 Enable or disable HTTP connection closing on the server side
8287 May be used in sections : defaults | frontend | listen | backend
8288 yes | yes | yes | yes
8289 Arguments : none
8290
Willy Tarreau70dffda2014-01-30 03:07:23 +01008291 By default HAProxy operates in keep-alive mode with regards to persistent
8292 connections: for each connection it processes each request and response, and
8293 leaves the connection idle on both sides between the end of a response and
8294 the start of a new request. This mode may be changed by several options such
Christopher Faulet159e6672019-07-16 15:09:52 +02008295 as "option http-server-close" or "option httpclose". Setting "option
8296 http-server-close" enables HTTP connection-close mode on the server side
8297 while keeping the ability to support HTTP keep-alive and pipelining on the
8298 client side. This provides the lowest latency on the client side (slow
8299 network) and the fastest session reuse on the server side to save server
8300 resources, similarly to "option httpclose". It also permits non-keepalive
8301 capable servers to be served in keep-alive mode to the clients if they
8302 conform to the requirements of RFC7230. Please note that some servers do not
8303 always conform to those requirements when they see "Connection: close" in the
8304 request. The effect will be that keep-alive will never be used. A workaround
8305 consists in enabling "option http-pretend-keepalive".
Willy Tarreaub608feb2010-01-02 22:47:18 +01008306
8307 At the moment, logs will not indicate whether requests came from the same
8308 session or not. The accept date reported in the logs corresponds to the end
8309 of the previous request, and the request time corresponds to the time spent
8310 waiting for a new request. The keep-alive request time is still bound to the
Willy Tarreaub16a5742010-01-10 14:46:16 +01008311 timeout defined by "timeout http-keep-alive" or "timeout http-request" if
8312 not set.
Willy Tarreaub608feb2010-01-02 22:47:18 +01008313
8314 This option may be set both in a frontend and in a backend. It is enabled if
8315 at least one of the frontend or backend holding a connection has it enabled.
Christopher Faulet159e6672019-07-16 15:09:52 +02008316 It disables and replaces any previous "option httpclose" or "option
8317 http-keep-alive". Please check section 4 ("Proxies") to see how this option
8318 combines with others when frontend and backend options differ.
Willy Tarreaub608feb2010-01-02 22:47:18 +01008319
8320 If this option has been enabled in a "defaults" section, it can be disabled
8321 in a specific instance by prepending the "no" keyword before it.
8322
Christopher Faulet315b39c2018-09-21 16:26:19 +02008323 See also : "option httpclose", "option http-pretend-keepalive",
8324 "option http-keep-alive", and "1.1. The HTTP transaction model".
Willy Tarreaub608feb2010-01-02 22:47:18 +01008325
Willy Tarreau88d349d2010-01-25 12:15:43 +01008326option http-use-proxy-header
Cyril Bontéf0c60612010-02-06 14:44:47 +01008327no option http-use-proxy-header
Willy Tarreau88d349d2010-01-25 12:15:43 +01008328 Make use of non-standard Proxy-Connection header instead of Connection
8329 May be used in sections : defaults | frontend | listen | backend
8330 yes | yes | yes | no
8331 Arguments : none
8332
Lukas Tribus23953682017-04-28 13:24:30 +00008333 While RFC7230 explicitly states that HTTP/1.1 agents must use the
Willy Tarreau88d349d2010-01-25 12:15:43 +01008334 Connection header to indicate their wish of persistent or non-persistent
8335 connections, both browsers and proxies ignore this header for proxied
8336 connections and make use of the undocumented, non-standard Proxy-Connection
8337 header instead. The issue begins when trying to put a load balancer between
8338 browsers and such proxies, because there will be a difference between what
8339 haproxy understands and what the client and the proxy agree on.
8340
8341 By setting this option in a frontend, haproxy can automatically switch to use
8342 that non-standard header if it sees proxied requests. A proxied request is
Lukas Tribusf01a9cd2016-02-03 18:09:37 +01008343 defined here as one where the URI begins with neither a '/' nor a '*'. This
8344 is incompatible with the HTTP tunnel mode. Note that this option can only be
8345 specified in a frontend and will affect the request along its whole life.
Willy Tarreau88d349d2010-01-25 12:15:43 +01008346
Willy Tarreau844a7e72010-01-31 21:46:18 +01008347 Also, when this option is set, a request which requires authentication will
8348 automatically switch to use proxy authentication headers if it is itself a
8349 proxied request. That makes it possible to check or enforce authentication in
8350 front of an existing proxy.
8351
Willy Tarreau88d349d2010-01-25 12:15:43 +01008352 This option should normally never be used, except in front of a proxy.
8353
Christopher Faulet315b39c2018-09-21 16:26:19 +02008354 See also : "option httpclose", and "option http-server-close".
Willy Tarreau88d349d2010-01-25 12:15:43 +01008355
Willy Tarreaud63335a2010-02-26 12:56:52 +01008356option httpchk
8357option httpchk <uri>
8358option httpchk <method> <uri>
8359option httpchk <method> <uri> <version>
Christopher Fauletc52ea4d2020-04-23 15:43:35 +02008360 Enables HTTP protocol to check on the servers health
Willy Tarreaud63335a2010-02-26 12:56:52 +01008361 May be used in sections : defaults | frontend | listen | backend
8362 yes | no | yes | yes
8363 Arguments :
8364 <method> is the optional HTTP method used with the requests. When not set,
8365 the "OPTIONS" method is used, as it generally requires low server
8366 processing and is easy to filter out from the logs. Any method
8367 may be used, though it is not recommended to invent non-standard
8368 ones.
8369
8370 <uri> is the URI referenced in the HTTP requests. It defaults to " / "
8371 which is accessible by default on almost any server, but may be
8372 changed to any other URI. Query strings are permitted.
8373
8374 <version> is the optional HTTP version string. It defaults to "HTTP/1.0"
8375 but some servers might behave incorrectly in HTTP 1.0, so turning
8376 it to HTTP/1.1 may sometimes help. Note that the Host field is
Christopher Faulet8acb1282020-04-09 08:44:06 +02008377 mandatory in HTTP/1.1, use "http-check send" directive to add it.
Willy Tarreaud63335a2010-02-26 12:56:52 +01008378
8379 By default, server health checks only consist in trying to establish a TCP
8380 connection. When "option httpchk" is specified, a complete HTTP request is
8381 sent once the TCP connection is established, and responses 2xx and 3xx are
8382 considered valid, while all other ones indicate a server failure, including
8383 the lack of any response.
8384
Christopher Faulete5870d82020-04-15 11:32:03 +02008385 Combined with "http-check" directives, it is possible to customize the
8386 request sent during the HTTP health checks or the matching rules on the
8387 response. It is also possible to configure a send/expect sequence, just like
8388 with the directive "tcp-check" for TCP health checks.
8389
8390 The server configuration is used by default to open connections to perform
8391 HTTP health checks. By it is also possible to overwrite server parameters
8392 using "http-check connect" rules.
Willy Tarreaud63335a2010-02-26 12:56:52 +01008393
Christopher Faulete5870d82020-04-15 11:32:03 +02008394 "httpchk" option does not necessarily require an HTTP backend, it also works
8395 with plain TCP backends. This is particularly useful to check simple scripts
Christopher Faulet14cd3162020-04-16 14:50:06 +02008396 bound to some dedicated ports using the inetd daemon. However, it will always
Daniel Corbett67a82712020-07-06 23:01:19 -04008397 internally relies on an HTX multiplexer. Thus, it means the request
Christopher Faulet14cd3162020-04-16 14:50:06 +02008398 formatting and the response parsing will be strict.
Willy Tarreaud63335a2010-02-26 12:56:52 +01008399
Christopher Faulet8acb1282020-04-09 08:44:06 +02008400 Note : For a while, there was no way to add headers or body in the request
8401 used for HTTP health checks. So a workaround was to hide it at the end
8402 of the version string with a "\r\n" after the version. It is now
8403 deprecated. The directive "http-check send" must be used instead.
8404
Willy Tarreaud63335a2010-02-26 12:56:52 +01008405 Examples :
Christopher Fauletc52ea4d2020-04-23 15:43:35 +02008406 # Relay HTTPS traffic to Apache instance and check service availability
8407 # using HTTP request "OPTIONS * HTTP/1.1" on port 80.
8408 backend https_relay
8409 mode tcp
8410 option httpchk OPTIONS * HTTP/1.1
8411 http-check send hdr Host www
8412 server apache1 192.168.1.1:443 check port 80
Willy Tarreaud63335a2010-02-26 12:56:52 +01008413
Simon Hormanafc47ee2013-11-25 10:46:35 +09008414 See also : "option ssl-hello-chk", "option smtpchk", "option mysql-check",
8415 "option pgsql-check", "http-check" and the "check", "port" and
8416 "inter" server options.
Willy Tarreaud63335a2010-02-26 12:56:52 +01008417
8418
Willy Tarreauc27debf2008-01-06 08:57:02 +01008419option httpclose
8420no option httpclose
Christopher Faulet315b39c2018-09-21 16:26:19 +02008421 Enable or disable HTTP connection closing
Willy Tarreauc27debf2008-01-06 08:57:02 +01008422 May be used in sections : defaults | frontend | listen | backend
8423 yes | yes | yes | yes
8424 Arguments : none
8425
Willy Tarreau70dffda2014-01-30 03:07:23 +01008426 By default HAProxy operates in keep-alive mode with regards to persistent
8427 connections: for each connection it processes each request and response, and
8428 leaves the connection idle on both sides between the end of a response and
8429 the start of a new request. This mode may be changed by several options such
Christopher Faulet159e6672019-07-16 15:09:52 +02008430 as "option http-server-close" or "option httpclose".
Willy Tarreau70dffda2014-01-30 03:07:23 +01008431
Christopher Faulet315b39c2018-09-21 16:26:19 +02008432 If "option httpclose" is set, HAProxy will close connections with the server
8433 and the client as soon as the request and the response are received. It will
John Roeslerfb2fce12019-07-10 15:45:51 -05008434 also check if a "Connection: close" header is already set in each direction,
Christopher Faulet315b39c2018-09-21 16:26:19 +02008435 and will add one if missing. Any "Connection" header different from "close"
8436 will also be removed.
Willy Tarreauc27debf2008-01-06 08:57:02 +01008437
Christopher Faulet315b39c2018-09-21 16:26:19 +02008438 This option may also be combined with "option http-pretend-keepalive", which
8439 will disable sending of the "Connection: close" header, but will still cause
8440 the connection to be closed once the whole response is received.
Willy Tarreauc27debf2008-01-06 08:57:02 +01008441
8442 This option may be set both in a frontend and in a backend. It is enabled if
8443 at least one of the frontend or backend holding a connection has it enabled.
Christopher Faulet159e6672019-07-16 15:09:52 +02008444 It disables and replaces any previous "option http-server-close" or "option
8445 http-keep-alive". Please check section 4 ("Proxies") to see how this option
8446 combines with others when frontend and backend options differ.
Willy Tarreauc27debf2008-01-06 08:57:02 +01008447
8448 If this option has been enabled in a "defaults" section, it can be disabled
8449 in a specific instance by prepending the "no" keyword before it.
8450
Christopher Faulet315b39c2018-09-21 16:26:19 +02008451 See also : "option http-server-close" and "1.1. The HTTP transaction model".
Willy Tarreauc27debf2008-01-06 08:57:02 +01008452
8453
Emeric Brun3a058f32009-06-30 18:26:00 +02008454option httplog [ clf ]
Willy Tarreauc27debf2008-01-06 08:57:02 +01008455 Enable logging of HTTP request, session state and timers
8456 May be used in sections : defaults | frontend | listen | backend
Tim Duesterhus9ad9f352018-02-05 20:52:27 +01008457 yes | yes | yes | no
Emeric Brun3a058f32009-06-30 18:26:00 +02008458 Arguments :
8459 clf if the "clf" argument is added, then the output format will be
8460 the CLF format instead of HAProxy's default HTTP format. You can
8461 use this when you need to feed HAProxy's logs through a specific
Davor Ocelice9ed2812017-12-25 17:49:28 +01008462 log analyzer which only support the CLF format and which is not
Emeric Brun3a058f32009-06-30 18:26:00 +02008463 extensible.
Willy Tarreauc27debf2008-01-06 08:57:02 +01008464
8465 By default, the log output format is very poor, as it only contains the
8466 source and destination addresses, and the instance name. By specifying
8467 "option httplog", each log line turns into a much richer format including,
8468 but not limited to, the HTTP request, the connection timers, the session
8469 status, the connections numbers, the captured headers and cookies, the
8470 frontend, backend and server name, and of course the source address and
8471 ports.
8472
PiBa-NLbd556bf2014-12-11 21:31:54 +01008473 Specifying only "option httplog" will automatically clear the 'clf' mode
8474 if it was set by default.
Emeric Brun3a058f32009-06-30 18:26:00 +02008475
Guillaume de Lafond29f45602017-03-31 19:52:15 +02008476 "option httplog" overrides any previous "log-format" directive.
8477
Willy Tarreauc57f0e22009-05-10 13:12:33 +02008478 See also : section 8 about logging.
Willy Tarreauc27debf2008-01-06 08:57:02 +01008479
Willy Tarreau55165fe2009-05-10 12:02:55 +02008480
8481option http_proxy
8482no option http_proxy
8483 Enable or disable plain HTTP proxy mode
8484 May be used in sections : defaults | frontend | listen | backend
8485 yes | yes | yes | yes
8486 Arguments : none
8487
8488 It sometimes happens that people need a pure HTTP proxy which understands
8489 basic proxy requests without caching nor any fancy feature. In this case,
8490 it may be worth setting up an HAProxy instance with the "option http_proxy"
8491 set. In this mode, no server is declared, and the connection is forwarded to
8492 the IP address and port found in the URL after the "http://" scheme.
8493
8494 No host address resolution is performed, so this only works when pure IP
8495 addresses are passed. Since this option's usage perimeter is rather limited,
Lukas Tribusf01a9cd2016-02-03 18:09:37 +01008496 it will probably be used only by experts who know they need exactly it. This
8497 is incompatible with the HTTP tunnel mode.
Willy Tarreau55165fe2009-05-10 12:02:55 +02008498
8499 If this option has been enabled in a "defaults" section, it can be disabled
8500 in a specific instance by prepending the "no" keyword before it.
8501
8502 Example :
8503 # this backend understands HTTP proxy requests and forwards them directly.
8504 backend direct_forward
8505 option httpclose
8506 option http_proxy
8507
8508 See also : "option httpclose"
8509
Willy Tarreau211ad242009-10-03 21:45:07 +02008510
Jamie Gloudon801a0a32012-08-25 00:18:33 -04008511option independent-streams
8512no option independent-streams
8513 Enable or disable independent timeout processing for both directions
Willy Tarreauf27b5ea2009-10-03 22:01:18 +02008514 May be used in sections : defaults | frontend | listen | backend
8515 yes | yes | yes | yes
8516 Arguments : none
8517
8518 By default, when data is sent over a socket, both the write timeout and the
8519 read timeout for that socket are refreshed, because we consider that there is
8520 activity on that socket, and we have no other means of guessing if we should
8521 receive data or not.
8522
Davor Ocelice9ed2812017-12-25 17:49:28 +01008523 While this default behavior is desirable for almost all applications, there
Willy Tarreauf27b5ea2009-10-03 22:01:18 +02008524 exists a situation where it is desirable to disable it, and only refresh the
8525 read timeout if there are incoming data. This happens on sessions with large
8526 timeouts and low amounts of exchanged data such as telnet session. If the
8527 server suddenly disappears, the output data accumulates in the system's
8528 socket buffers, both timeouts are correctly refreshed, and there is no way
8529 to know the server does not receive them, so we don't timeout. However, when
8530 the underlying protocol always echoes sent data, it would be enough by itself
8531 to detect the issue using the read timeout. Note that this problem does not
8532 happen with more verbose protocols because data won't accumulate long in the
8533 socket buffers.
8534
8535 When this option is set on the frontend, it will disable read timeout updates
8536 on data sent to the client. There probably is little use of this case. When
8537 the option is set on the backend, it will disable read timeout updates on
8538 data sent to the server. Doing so will typically break large HTTP posts from
8539 slow lines, so use it with caution.
8540
Willy Tarreauce887fd2012-05-12 12:50:00 +02008541 See also : "timeout client", "timeout server" and "timeout tunnel"
Willy Tarreauf27b5ea2009-10-03 22:01:18 +02008542
8543
Gabor Lekenyb4c81e42010-09-29 18:17:05 +02008544option ldap-check
8545 Use LDAPv3 health checks for server testing
8546 May be used in sections : defaults | frontend | listen | backend
8547 yes | no | yes | yes
8548 Arguments : none
8549
8550 It is possible to test that the server correctly talks LDAPv3 instead of just
8551 testing that it accepts the TCP connection. When this option is set, an
8552 LDAPv3 anonymous simple bind message is sent to the server, and the response
8553 is analyzed to find an LDAPv3 bind response message.
8554
8555 The server is considered valid only when the LDAP response contains success
8556 resultCode (http://tools.ietf.org/html/rfc4511#section-4.1.9).
8557
8558 Logging of bind requests is server dependent see your documentation how to
8559 configure it.
8560
8561 Example :
8562 option ldap-check
8563
8564 See also : "option httpchk"
8565
8566
Simon Horman98637e52014-06-20 12:30:16 +09008567option external-check
8568 Use external processes for server health checks
8569 May be used in sections : defaults | frontend | listen | backend
8570 yes | no | yes | yes
8571
8572 It is possible to test the health of a server using an external command.
8573 This is achieved by running the executable set using "external-check
8574 command".
8575
8576 Requires the "external-check" global to be set.
8577
8578 See also : "external-check", "external-check command", "external-check path"
8579
8580
Willy Tarreau211ad242009-10-03 21:45:07 +02008581option log-health-checks
8582no option log-health-checks
Willy Tarreaubef1b322014-05-13 21:01:39 +02008583 Enable or disable logging of health checks status updates
Willy Tarreau211ad242009-10-03 21:45:07 +02008584 May be used in sections : defaults | frontend | listen | backend
8585 yes | no | yes | yes
8586 Arguments : none
8587
Willy Tarreaubef1b322014-05-13 21:01:39 +02008588 By default, failed health check are logged if server is UP and successful
8589 health checks are logged if server is DOWN, so the amount of additional
8590 information is limited.
Willy Tarreau211ad242009-10-03 21:45:07 +02008591
Willy Tarreaubef1b322014-05-13 21:01:39 +02008592 When this option is enabled, any change of the health check status or to
8593 the server's health will be logged, so that it becomes possible to know
8594 that a server was failing occasional checks before crashing, or exactly when
8595 it failed to respond a valid HTTP status, then when the port started to
8596 reject connections, then when the server stopped responding at all.
8597
Davor Ocelice9ed2812017-12-25 17:49:28 +01008598 Note that status changes not caused by health checks (e.g. enable/disable on
Willy Tarreaubef1b322014-05-13 21:01:39 +02008599 the CLI) are intentionally not logged by this option.
Willy Tarreau211ad242009-10-03 21:45:07 +02008600
Willy Tarreaubef1b322014-05-13 21:01:39 +02008601 See also: "option httpchk", "option ldap-check", "option mysql-check",
8602 "option pgsql-check", "option redis-check", "option smtpchk",
8603 "option tcp-check", "log" and section 8 about logging.
Willy Tarreau211ad242009-10-03 21:45:07 +02008604
Willy Tarreauc9bd0cc2009-05-10 11:57:02 +02008605
8606option log-separate-errors
8607no option log-separate-errors
8608 Change log level for non-completely successful connections
8609 May be used in sections : defaults | frontend | listen | backend
8610 yes | yes | yes | no
8611 Arguments : none
8612
8613 Sometimes looking for errors in logs is not easy. This option makes haproxy
8614 raise the level of logs containing potentially interesting information such
8615 as errors, timeouts, retries, redispatches, or HTTP status codes 5xx. The
8616 level changes from "info" to "err". This makes it possible to log them
8617 separately to a different file with most syslog daemons. Be careful not to
8618 remove them from the original file, otherwise you would lose ordering which
8619 provides very important information.
8620
8621 Using this option, large sites dealing with several thousand connections per
8622 second may log normal traffic to a rotating buffer and only archive smaller
8623 error logs.
8624
Willy Tarreauc57f0e22009-05-10 13:12:33 +02008625 See also : "log", "dontlognull", "dontlog-normal" and section 8 about
Willy Tarreauc9bd0cc2009-05-10 11:57:02 +02008626 logging.
8627
Willy Tarreauc27debf2008-01-06 08:57:02 +01008628
8629option logasap
8630no option logasap
Jerome Magnin95fb57b2020-04-23 19:01:17 +02008631 Enable or disable early logging.
Willy Tarreauc27debf2008-01-06 08:57:02 +01008632 May be used in sections : defaults | frontend | listen | backend
8633 yes | yes | yes | no
8634 Arguments : none
8635
Jerome Magnin95fb57b2020-04-23 19:01:17 +02008636 By default, logs are emitted when all the log format variables and sample
8637 fetches used in the definition of the log-format string return a value, or
8638 when the session is terminated. This allows the built in log-format strings
8639 to account for the transfer time, or the number of bytes in log messages.
8640
8641 When handling long lived connections such as large file transfers or RDP,
8642 it may take a while for the request or connection to appear in the logs.
8643 Using "option logasap", the log message is created as soon as the server
8644 connection is established in mode tcp, or as soon as the server sends the
8645 complete headers in mode http. Missing information in the logs will be the
Ilya Shipitsin4329a9a2020-05-05 21:17:10 +05008646 total number of bytes which will only indicate the amount of data transferred
Jerome Magnin95fb57b2020-04-23 19:01:17 +02008647 before the message was created and the total time which will not take the
8648 remainder of the connection life or transfer time into account. For the case
8649 of HTTP, it is good practice to capture the Content-Length response header
8650 so that the logs at least indicate how many bytes are expected to be
Ilya Shipitsin4329a9a2020-05-05 21:17:10 +05008651 transferred.
Willy Tarreauc27debf2008-01-06 08:57:02 +01008652
Willy Tarreaucc6c8912009-02-22 10:53:55 +01008653 Examples :
8654 listen http_proxy 0.0.0.0:80
8655 mode http
8656 option httplog
8657 option logasap
8658 log 192.168.2.200 local3
8659
8660 >>> Feb 6 12:14:14 localhost \
8661 haproxy[14389]: 10.0.1.2:33317 [06/Feb/2009:12:14:14.655] http-in \
8662 static/srv1 9/10/7/14/+30 200 +243 - - ---- 3/1/1/1/0 1/0 \
8663 "GET /image.iso HTTP/1.0"
8664
Willy Tarreauc57f0e22009-05-10 13:12:33 +02008665 See also : "option httplog", "capture response header", and section 8 about
Willy Tarreauc27debf2008-01-06 08:57:02 +01008666 logging.
8667
8668
Christopher Faulet62f79fe2020-05-18 18:13:03 +02008669option mysql-check [ user <username> [ { post-41 | pre-41 } ] ]
Hervé COMMOWICK8776f1b2010-10-18 15:58:36 +02008670 Use MySQL health checks for server testing
Hervé COMMOWICK698ae002010-01-12 09:25:13 +01008671 May be used in sections : defaults | frontend | listen | backend
8672 yes | no | yes | yes
Hervé COMMOWICK8776f1b2010-10-18 15:58:36 +02008673 Arguments :
Cyril Bonté108cf6e2012-04-21 23:30:29 +02008674 <username> This is the username which will be used when connecting to MySQL
8675 server.
Christopher Faulet62f79fe2020-05-18 18:13:03 +02008676 post-41 Send post v4.1 client compatible checks (the default)
8677 pre-41 Send pre v4.1 client compatible checks
Hervé COMMOWICK8776f1b2010-10-18 15:58:36 +02008678
8679 If you specify a username, the check consists of sending two MySQL packet,
8680 one Client Authentication packet, and one QUIT packet, to correctly close
Davor Ocelice9ed2812017-12-25 17:49:28 +01008681 MySQL session. We then parse the MySQL Handshake Initialization packet and/or
Hervé COMMOWICK8776f1b2010-10-18 15:58:36 +02008682 Error packet. It is a basic but useful test which does not produce error nor
8683 aborted connect on the server. However, it requires adding an authorization
8684 in the MySQL table, like this :
8685
8686 USE mysql;
8687 INSERT INTO user (Host,User) values ('<ip_of_haproxy>','<username>');
8688 FLUSH PRIVILEGES;
8689
8690 If you don't specify a username (it is deprecated and not recommended), the
Davor Ocelice9ed2812017-12-25 17:49:28 +01008691 check only consists in parsing the Mysql Handshake Initialization packet or
Hervé COMMOWICK8776f1b2010-10-18 15:58:36 +02008692 Error packet, we don't send anything in this mode. It was reported that it
8693 can generate lockout if check is too frequent and/or if there is not enough
8694 traffic. In fact, you need in this case to check MySQL "max_connect_errors"
8695 value as if a connection is established successfully within fewer than MySQL
8696 "max_connect_errors" attempts after a previous connection was interrupted,
8697 the error count for the host is cleared to zero. If HAProxy's server get
8698 blocked, the "FLUSH HOSTS" statement is the only way to unblock it.
8699
8700 Remember that this does not check database presence nor database consistency.
8701 To do this, you can use an external check with xinetd for example.
Hervé COMMOWICK698ae002010-01-12 09:25:13 +01008702
Hervé COMMOWICK212f7782011-06-10 14:05:59 +02008703 The check requires MySQL >=3.22, for older version, please use TCP check.
Hervé COMMOWICK698ae002010-01-12 09:25:13 +01008704
8705 Most often, an incoming MySQL server needs to see the client's IP address for
8706 various purposes, including IP privilege matching and connection logging.
8707 When possible, it is often wise to masquerade the client's IP address when
8708 connecting to the server using the "usesrc" argument of the "source" keyword,
Willy Tarreau29fbe512015-08-20 19:35:14 +02008709 which requires the transparent proxy feature to be compiled in, and the MySQL
8710 server to route the client via the machine hosting haproxy.
Hervé COMMOWICK698ae002010-01-12 09:25:13 +01008711
8712 See also: "option httpchk"
8713
8714
Willy Tarreaua453bdd2008-01-08 19:50:52 +01008715option nolinger
8716no option nolinger
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01008717 Enable or disable immediate session resource cleaning after close
Willy Tarreaua453bdd2008-01-08 19:50:52 +01008718 May be used in sections: defaults | frontend | listen | backend
8719 yes | yes | yes | yes
Willy Tarreaueabeafa2008-01-16 16:17:06 +01008720 Arguments : none
Willy Tarreaua453bdd2008-01-08 19:50:52 +01008721
Davor Ocelice9ed2812017-12-25 17:49:28 +01008722 When clients or servers abort connections in a dirty way (e.g. they are
Willy Tarreaua453bdd2008-01-08 19:50:52 +01008723 physically disconnected), the session timeouts triggers and the session is
8724 closed. But it will remain in FIN_WAIT1 state for some time in the system,
8725 using some resources and possibly limiting the ability to establish newer
8726 connections.
8727
8728 When this happens, it is possible to activate "option nolinger" which forces
8729 the system to immediately remove any socket's pending data on close. Thus,
Willy Tarreau4a321032020-10-16 04:55:19 +02008730 a TCP RST is emitted, any pending data are truncated, and the session is
8731 instantly purged from the system's tables. The generally visible effect for
8732 a client is that responses are truncated if the close happens with a last
8733 block of data (e.g. on a redirect or error response). On the server side,
8734 it may help release the source ports immediately when forwarding a client
8735 aborts in tunnels. In both cases, TCP resets are emitted and given that
8736 the session is instantly destroyed, there will be no retransmit. On a lossy
8737 network this can increase problems, especially when there is a firewall on
8738 the lossy side, because the firewall might see and process the reset (hence
8739 purge its session) and block any further traffic for this session,, including
8740 retransmits from the other side. So if the other side doesn't receive it,
8741 it will never receive any RST again, and the firewall might log many blocked
8742 packets.
Willy Tarreaua453bdd2008-01-08 19:50:52 +01008743
Willy Tarreau4a321032020-10-16 04:55:19 +02008744 For all these reasons, it is strongly recommended NOT to use this option,
8745 unless absolutely needed as a last resort. In most situations, using the
8746 "client-fin" or "server-fin" timeouts achieves similar results with a more
8747 reliable behavior. On Linux it's also possible to use the "tcp-ut" bind or
8748 server setting.
Willy Tarreaua453bdd2008-01-08 19:50:52 +01008749
8750 This option may be used both on frontends and backends, depending on the side
8751 where it is required. Use it on the frontend for clients, and on the backend
Willy Tarreau4a321032020-10-16 04:55:19 +02008752 for servers. While this option is technically supported in "defaults"
Ilya Shipitsin2272d8a2020-12-21 01:22:40 +05008753 sections, it must really not be used there as it risks to accidentally
Willy Tarreau4a321032020-10-16 04:55:19 +02008754 propagate to sections that must no use it and to cause problems there.
Willy Tarreaua453bdd2008-01-08 19:50:52 +01008755
8756 If this option has been enabled in a "defaults" section, it can be disabled
8757 in a specific instance by prepending the "no" keyword before it.
8758
Willy Tarreau4a321032020-10-16 04:55:19 +02008759 See also: "timeout client-fin", "timeout server-fin", "tcp-ut" bind or server
8760 keywords.
Willy Tarreaua453bdd2008-01-08 19:50:52 +01008761
Willy Tarreau55165fe2009-05-10 12:02:55 +02008762option originalto [ except <network> ] [ header <name> ]
8763 Enable insertion of the X-Original-To header to requests sent to servers
8764 May be used in sections : defaults | frontend | listen | backend
8765 yes | yes | yes | yes
8766 Arguments :
8767 <network> is an optional argument used to disable this option for sources
8768 matching <network>
8769 <name> an optional argument to specify a different "X-Original-To"
8770 header name.
8771
8772 Since HAProxy can work in transparent mode, every request from a client can
8773 be redirected to the proxy and HAProxy itself can proxy every request to a
8774 complex SQUID environment and the destination host from SO_ORIGINAL_DST will
8775 be lost. This is annoying when you want access rules based on destination ip
8776 addresses. To solve this problem, a new HTTP header "X-Original-To" may be
8777 added by HAProxy to all requests sent to the server. This header contains a
8778 value representing the original destination IP address. Since this must be
8779 configured to always use the last occurrence of this header only. Note that
8780 only the last occurrence of the header must be used, since it is really
8781 possible that the client has already brought one.
8782
8783 The keyword "header" may be used to supply a different header name to replace
8784 the default "X-Original-To". This can be useful where you might already
8785 have a "X-Original-To" header from a different application, and you need
8786 preserve it. Also if your backend server doesn't use the "X-Original-To"
8787 header and requires different one.
8788
8789 Sometimes, a same HAProxy instance may be shared between a direct client
8790 access and a reverse-proxy access (for instance when an SSL reverse-proxy is
8791 used to decrypt HTTPS traffic). It is possible to disable the addition of the
Amaury Denoyellef8b42922021-03-04 18:41:14 +01008792 header for a known destination address or network by adding the "except"
8793 keyword followed by the network address. In this case, any destination IP
8794 matching the network will not cause an addition of this header. Most common
8795 uses are with private networks or 127.0.0.1. IPv4 and IPv6 are both
8796 supported.
Willy Tarreau55165fe2009-05-10 12:02:55 +02008797
8798 This option may be specified either in the frontend or in the backend. If at
8799 least one of them uses it, the header will be added. Note that the backend's
8800 setting of the header subargument takes precedence over the frontend's if
8801 both are defined.
8802
Willy Tarreau55165fe2009-05-10 12:02:55 +02008803 Examples :
8804 # Original Destination address
8805 frontend www
8806 mode http
8807 option originalto except 127.0.0.1
8808
8809 # Those servers want the IP Address in X-Client-Dst
8810 backend www
8811 mode http
8812 option originalto header X-Client-Dst
8813
Christopher Faulet315b39c2018-09-21 16:26:19 +02008814 See also : "option httpclose", "option http-server-close".
Willy Tarreau55165fe2009-05-10 12:02:55 +02008815
8816
Willy Tarreaua453bdd2008-01-08 19:50:52 +01008817option persist
8818no option persist
8819 Enable or disable forced persistence on down servers
8820 May be used in sections: defaults | frontend | listen | backend
8821 yes | no | yes | yes
Willy Tarreaueabeafa2008-01-16 16:17:06 +01008822 Arguments : none
Willy Tarreaua453bdd2008-01-08 19:50:52 +01008823
8824 When an HTTP request reaches a backend with a cookie which references a dead
8825 server, by default it is redispatched to another server. It is possible to
8826 force the request to be sent to the dead server first using "option persist"
8827 if absolutely needed. A common use case is when servers are under extreme
8828 load and spend their time flapping. In this case, the users would still be
8829 directed to the server they opened the session on, in the hope they would be
8830 correctly served. It is recommended to use "option redispatch" in conjunction
8831 with this option so that in the event it would not be possible to connect to
8832 the server at all (server definitely dead), the client would finally be
8833 redirected to another valid server.
8834
8835 If this option has been enabled in a "defaults" section, it can be disabled
8836 in a specific instance by prepending the "no" keyword before it.
8837
Willy Tarreau4de91492010-01-22 19:10:05 +01008838 See also : "option redispatch", "retries", "force-persist"
Willy Tarreaua453bdd2008-01-08 19:50:52 +01008839
8840
Willy Tarreau0c122822013-12-15 18:49:01 +01008841option pgsql-check [ user <username> ]
8842 Use PostgreSQL health checks for server testing
8843 May be used in sections : defaults | frontend | listen | backend
8844 yes | no | yes | yes
8845 Arguments :
8846 <username> This is the username which will be used when connecting to
8847 PostgreSQL server.
8848
8849 The check sends a PostgreSQL StartupMessage and waits for either
8850 Authentication request or ErrorResponse message. It is a basic but useful
8851 test which does not produce error nor aborted connect on the server.
8852 This check is identical with the "mysql-check".
8853
8854 See also: "option httpchk"
8855
8856
Willy Tarreau9420b122013-12-15 18:58:25 +01008857option prefer-last-server
8858no option prefer-last-server
8859 Allow multiple load balanced requests to remain on the same server
8860 May be used in sections: defaults | frontend | listen | backend
8861 yes | no | yes | yes
8862 Arguments : none
8863
8864 When the load balancing algorithm in use is not deterministic, and a previous
8865 request was sent to a server to which haproxy still holds a connection, it is
8866 sometimes desirable that subsequent requests on a same session go to the same
8867 server as much as possible. Note that this is different from persistence, as
8868 we only indicate a preference which haproxy tries to apply without any form
8869 of warranty. The real use is for keep-alive connections sent to servers. When
8870 this option is used, haproxy will try to reuse the same connection that is
8871 attached to the server instead of rebalancing to another server, causing a
8872 close of the connection. This can make sense for static file servers. It does
Willy Tarreau068621e2013-12-23 15:11:25 +01008873 not make much sense to use this in combination with hashing algorithms. Note,
8874 haproxy already automatically tries to stick to a server which sends a 401 or
Lukas Tribus80512b12018-10-27 20:07:40 +02008875 to a proxy which sends a 407 (authentication required), when the load
8876 balancing algorithm is not deterministic. This is mandatory for use with the
8877 broken NTLM authentication challenge, and significantly helps in
Willy Tarreau068621e2013-12-23 15:11:25 +01008878 troubleshooting some faulty applications. Option prefer-last-server might be
8879 desirable in these environments as well, to avoid redistributing the traffic
8880 after every other response.
Willy Tarreau9420b122013-12-15 18:58:25 +01008881
8882 If this option has been enabled in a "defaults" section, it can be disabled
8883 in a specific instance by prepending the "no" keyword before it.
8884
8885 See also: "option http-keep-alive"
8886
8887
Krzysztof Piotr Oledzki25b501a2008-01-06 16:36:16 +01008888option redispatch
Joseph Lynch726ab712015-05-11 23:25:34 -07008889option redispatch <interval>
Krzysztof Piotr Oledzki25b501a2008-01-06 16:36:16 +01008890no option redispatch
8891 Enable or disable session redistribution in case of connection failure
8892 May be used in sections: defaults | frontend | listen | backend
8893 yes | no | yes | yes
Joseph Lynch726ab712015-05-11 23:25:34 -07008894 Arguments :
8895 <interval> The optional integer value that controls how often redispatches
8896 occur when retrying connections. Positive value P indicates a
8897 redispatch is desired on every Pth retry, and negative value
Davor Ocelice9ed2812017-12-25 17:49:28 +01008898 N indicate a redispatch is desired on the Nth retry prior to the
Joseph Lynch726ab712015-05-11 23:25:34 -07008899 last retry. For example, the default of -1 preserves the
Davor Ocelice9ed2812017-12-25 17:49:28 +01008900 historical behavior of redispatching on the last retry, a
Joseph Lynch726ab712015-05-11 23:25:34 -07008901 positive value of 1 would indicate a redispatch on every retry,
8902 and a positive value of 3 would indicate a redispatch on every
8903 third retry. You can disable redispatches with a value of 0.
8904
Krzysztof Piotr Oledzki25b501a2008-01-06 16:36:16 +01008905
8906 In HTTP mode, if a server designated by a cookie is down, clients may
8907 definitely stick to it because they cannot flush the cookie, so they will not
8908 be able to access the service anymore.
8909
Willy Tarreau59884a62019-01-02 14:48:31 +01008910 Specifying "option redispatch" will allow the proxy to break cookie or
8911 consistent hash based persistence and redistribute them to a working server.
Krzysztof Piotr Oledzki25b501a2008-01-06 16:36:16 +01008912
Olivier Carrère6e6f59b2020-04-15 11:30:18 +02008913 Active servers are selected from a subset of the list of available
8914 servers. Active servers that are not down or in maintenance (i.e., whose
8915 health is not checked or that have been checked as "up"), are selected in the
8916 following order:
8917
8918 1. Any active, non-backup server, if any, or,
8919
8920 2. If the "allbackups" option is not set, the first backup server in the
8921 list, or
8922
8923 3. If the "allbackups" option is set, any backup server.
8924
8925 When a retry occurs, HAProxy tries to select another server than the last
8926 one. The new server is selected from the current list of servers.
8927
8928 Sometimes, if the list is updated between retries (e.g., if numerous retries
8929 occur and last longer than the time needed to check that a server is down,
8930 remove it from the list and fall back on the list of backup servers),
8931 connections may be redirected to a backup server, though.
8932
Joseph Lynch726ab712015-05-11 23:25:34 -07008933 It also allows to retry connections to another server in case of multiple
Krzysztof Piotr Oledzki25b501a2008-01-06 16:36:16 +01008934 connection failures. Of course, it requires having "retries" set to a nonzero
8935 value.
Willy Tarreaud72758d2010-01-12 10:42:19 +01008936
Krzysztof Piotr Oledzki25b501a2008-01-06 16:36:16 +01008937 If this option has been enabled in a "defaults" section, it can be disabled
8938 in a specific instance by prepending the "no" keyword before it.
8939
Christopher Faulet87f1f3d2019-07-18 14:51:20 +02008940 See also : "retries", "force-persist"
Krzysztof Piotr Oledzki25b501a2008-01-06 16:36:16 +01008941
Willy Tarreaua453bdd2008-01-08 19:50:52 +01008942
Hervé COMMOWICKec032d62011-08-05 16:23:48 +02008943option redis-check
8944 Use redis health checks for server testing
8945 May be used in sections : defaults | frontend | listen | backend
8946 yes | no | yes | yes
8947 Arguments : none
8948
8949 It is possible to test that the server correctly talks REDIS protocol instead
8950 of just testing that it accepts the TCP connection. When this option is set,
8951 a PING redis command is sent to the server, and the response is analyzed to
8952 find the "+PONG" response message.
8953
8954 Example :
8955 option redis-check
8956
Jarno Huuskonene5ae7022017-04-03 14:36:21 +03008957 See also : "option httpchk", "option tcp-check", "tcp-check expect"
Hervé COMMOWICKec032d62011-08-05 16:23:48 +02008958
8959
Willy Tarreaua453bdd2008-01-08 19:50:52 +01008960option smtpchk
8961option smtpchk <hello> <domain>
8962 Use SMTP health checks for server testing
8963 May be used in sections : defaults | frontend | listen | backend
8964 yes | no | yes | yes
Willy Tarreaud72758d2010-01-12 10:42:19 +01008965 Arguments :
Willy Tarreaua453bdd2008-01-08 19:50:52 +01008966 <hello> is an optional argument. It is the "hello" command to use. It can
Lukas Tribus27935782018-10-01 02:00:16 +02008967 be either "HELO" (for SMTP) or "EHLO" (for ESMTP). All other
Willy Tarreaua453bdd2008-01-08 19:50:52 +01008968 values will be turned into the default command ("HELO").
8969
8970 <domain> is the domain name to present to the server. It may only be
8971 specified (and is mandatory) if the hello command has been
8972 specified. By default, "localhost" is used.
8973
8974 When "option smtpchk" is set, the health checks will consist in TCP
8975 connections followed by an SMTP command. By default, this command is
8976 "HELO localhost". The server's return code is analyzed and only return codes
8977 starting with a "2" will be considered as valid. All other responses,
8978 including a lack of response will constitute an error and will indicate a
8979 dead server.
8980
8981 This test is meant to be used with SMTP servers or relays. Depending on the
8982 request, it is possible that some servers do not log each connection attempt,
Davor Ocelice9ed2812017-12-25 17:49:28 +01008983 so you may want to experiment to improve the behavior. Using telnet on port
Willy Tarreaua453bdd2008-01-08 19:50:52 +01008984 25 is often easier than adjusting the configuration.
8985
8986 Most often, an incoming SMTP server needs to see the client's IP address for
8987 various purposes, including spam filtering, anti-spoofing and logging. When
8988 possible, it is often wise to masquerade the client's IP address when
8989 connecting to the server using the "usesrc" argument of the "source" keyword,
Willy Tarreau29fbe512015-08-20 19:35:14 +02008990 which requires the transparent proxy feature to be compiled in.
Willy Tarreaua453bdd2008-01-08 19:50:52 +01008991
8992 Example :
8993 option smtpchk HELO mydomain.org
8994
8995 See also : "option httpchk", "source"
8996
Krzysztof Piotr Oledzki25b501a2008-01-06 16:36:16 +01008997
Krzysztof Piotr Oledzkiaeebf9b2009-10-04 15:43:17 +02008998option socket-stats
8999no option socket-stats
9000
9001 Enable or disable collecting & providing separate statistics for each socket.
9002 May be used in sections : defaults | frontend | listen | backend
9003 yes | yes | yes | no
9004
9005 Arguments : none
9006
9007
Willy Tarreauff4f82d2009-02-06 11:28:13 +01009008option splice-auto
9009no option splice-auto
9010 Enable or disable automatic kernel acceleration on sockets in both directions
9011 May be used in sections : defaults | frontend | listen | backend
9012 yes | yes | yes | yes
9013 Arguments : none
9014
9015 When this option is enabled either on a frontend or on a backend, haproxy
9016 will automatically evaluate the opportunity to use kernel tcp splicing to
Davor Ocelice9ed2812017-12-25 17:49:28 +01009017 forward data between the client and the server, in either direction. HAProxy
Willy Tarreauff4f82d2009-02-06 11:28:13 +01009018 uses heuristics to estimate if kernel splicing might improve performance or
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01009019 not. Both directions are handled independently. Note that the heuristics used
Willy Tarreauff4f82d2009-02-06 11:28:13 +01009020 are not much aggressive in order to limit excessive use of splicing. This
9021 option requires splicing to be enabled at compile time, and may be globally
9022 disabled with the global option "nosplice". Since splice uses pipes, using it
9023 requires that there are enough spare pipes.
9024
9025 Important note: kernel-based TCP splicing is a Linux-specific feature which
9026 first appeared in kernel 2.6.25. It offers kernel-based acceleration to
9027 transfer data between sockets without copying these data to user-space, thus
9028 providing noticeable performance gains and CPU cycles savings. Since many
9029 early implementations are buggy, corrupt data and/or are inefficient, this
9030 feature is not enabled by default, and it should be used with extreme care.
9031 While it is not possible to detect the correctness of an implementation,
9032 2.6.29 is the first version offering a properly working implementation. In
9033 case of doubt, splicing may be globally disabled using the global "nosplice"
9034 keyword.
9035
9036 Example :
9037 option splice-auto
9038
9039 If this option has been enabled in a "defaults" section, it can be disabled
9040 in a specific instance by prepending the "no" keyword before it.
9041
9042 See also : "option splice-request", "option splice-response", and global
9043 options "nosplice" and "maxpipes"
9044
9045
9046option splice-request
9047no option splice-request
9048 Enable or disable automatic kernel acceleration on sockets for requests
9049 May be used in sections : defaults | frontend | listen | backend
9050 yes | yes | yes | yes
9051 Arguments : none
9052
9053 When this option is enabled either on a frontend or on a backend, haproxy
Jamie Gloudonaaa21002012-08-25 00:18:33 -04009054 will use kernel tcp splicing whenever possible to forward data going from
Willy Tarreauff4f82d2009-02-06 11:28:13 +01009055 the client to the server. It might still use the recv/send scheme if there
9056 are no spare pipes left. This option requires splicing to be enabled at
9057 compile time, and may be globally disabled with the global option "nosplice".
9058 Since splice uses pipes, using it requires that there are enough spare pipes.
9059
9060 Important note: see "option splice-auto" for usage limitations.
9061
9062 Example :
9063 option splice-request
9064
9065 If this option has been enabled in a "defaults" section, it can be disabled
9066 in a specific instance by prepending the "no" keyword before it.
9067
9068 See also : "option splice-auto", "option splice-response", and global options
9069 "nosplice" and "maxpipes"
9070
9071
9072option splice-response
9073no option splice-response
9074 Enable or disable automatic kernel acceleration on sockets for responses
9075 May be used in sections : defaults | frontend | listen | backend
9076 yes | yes | yes | yes
9077 Arguments : none
9078
9079 When this option is enabled either on a frontend or on a backend, haproxy
Jamie Gloudonaaa21002012-08-25 00:18:33 -04009080 will use kernel tcp splicing whenever possible to forward data going from
Willy Tarreauff4f82d2009-02-06 11:28:13 +01009081 the server to the client. It might still use the recv/send scheme if there
9082 are no spare pipes left. This option requires splicing to be enabled at
9083 compile time, and may be globally disabled with the global option "nosplice".
9084 Since splice uses pipes, using it requires that there are enough spare pipes.
9085
9086 Important note: see "option splice-auto" for usage limitations.
9087
9088 Example :
9089 option splice-response
9090
9091 If this option has been enabled in a "defaults" section, it can be disabled
9092 in a specific instance by prepending the "no" keyword before it.
9093
9094 See also : "option splice-auto", "option splice-request", and global options
9095 "nosplice" and "maxpipes"
9096
9097
Christopher Fauletba7bc162016-11-07 21:07:38 +01009098option spop-check
9099 Use SPOP health checks for server testing
9100 May be used in sections : defaults | frontend | listen | backend
9101 no | no | no | yes
9102 Arguments : none
9103
9104 It is possible to test that the server correctly talks SPOP protocol instead
9105 of just testing that it accepts the TCP connection. When this option is set,
9106 a HELLO handshake is performed between HAProxy and the server, and the
9107 response is analyzed to check no error is reported.
9108
9109 Example :
9110 option spop-check
9111
9112 See also : "option httpchk"
9113
9114
Willy Tarreaubf1f8162007-12-28 17:42:56 +01009115option srvtcpka
9116no option srvtcpka
9117 Enable or disable the sending of TCP keepalive packets on the server side
9118 May be used in sections : defaults | frontend | listen | backend
9119 yes | no | yes | yes
9120 Arguments : none
9121
9122 When there is a firewall or any session-aware component between a client and
9123 a server, and when the protocol involves very long sessions with long idle
Davor Ocelice9ed2812017-12-25 17:49:28 +01009124 periods (e.g. remote desktops), there is a risk that one of the intermediate
Willy Tarreaubf1f8162007-12-28 17:42:56 +01009125 components decides to expire a session which has remained idle for too long.
9126
9127 Enabling socket-level TCP keep-alives makes the system regularly send packets
9128 to the other end of the connection, leaving it active. The delay between
9129 keep-alive probes is controlled by the system only and depends both on the
9130 operating system and its tuning parameters.
9131
9132 It is important to understand that keep-alive packets are neither emitted nor
9133 received at the application level. It is only the network stacks which sees
9134 them. For this reason, even if one side of the proxy already uses keep-alives
9135 to maintain its connection alive, those keep-alive packets will not be
9136 forwarded to the other side of the proxy.
9137
9138 Please note that this has nothing to do with HTTP keep-alive.
9139
9140 Using option "srvtcpka" enables the emission of TCP keep-alive probes on the
9141 server side of a connection, which should help when session expirations are
9142 noticed between HAProxy and a server.
9143
9144 If this option has been enabled in a "defaults" section, it can be disabled
9145 in a specific instance by prepending the "no" keyword before it.
9146
9147 See also : "option clitcpka", "option tcpka"
9148
9149
Willy Tarreaua453bdd2008-01-08 19:50:52 +01009150option ssl-hello-chk
9151 Use SSLv3 client hello health checks for server testing
9152 May be used in sections : defaults | frontend | listen | backend
9153 yes | no | yes | yes
9154 Arguments : none
9155
9156 When some SSL-based protocols are relayed in TCP mode through HAProxy, it is
9157 possible to test that the server correctly talks SSL instead of just testing
9158 that it accepts the TCP connection. When "option ssl-hello-chk" is set, pure
9159 SSLv3 client hello messages are sent once the connection is established to
9160 the server, and the response is analyzed to find an SSL server hello message.
9161 The server is considered valid only when the response contains this server
9162 hello message.
9163
9164 All servers tested till there correctly reply to SSLv3 client hello messages,
9165 and most servers tested do not even log the requests containing only hello
9166 messages, which is appreciable.
9167
Willy Tarreau763a95b2012-10-04 23:15:39 +02009168 Note that this check works even when SSL support was not built into haproxy
9169 because it forges the SSL message. When SSL support is available, it is best
9170 to use native SSL health checks instead of this one.
Willy Tarreaua453bdd2008-01-08 19:50:52 +01009171
Willy Tarreau763a95b2012-10-04 23:15:39 +02009172 See also: "option httpchk", "check-ssl"
9173
Willy Tarreaua453bdd2008-01-08 19:50:52 +01009174
Willy Tarreaued179852013-12-16 01:07:00 +01009175option tcp-check
9176 Perform health checks using tcp-check send/expect sequences
9177 May be used in sections: defaults | frontend | listen | backend
9178 yes | no | yes | yes
9179
9180 This health check method is intended to be combined with "tcp-check" command
9181 lists in order to support send/expect types of health check sequences.
9182
9183 TCP checks currently support 4 modes of operations :
9184 - no "tcp-check" directive : the health check only consists in a connection
9185 attempt, which remains the default mode.
9186
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03009187 - "tcp-check send" or "tcp-check send-binary" only is mentioned : this is
Willy Tarreaued179852013-12-16 01:07:00 +01009188 used to send a string along with a connection opening. With some
9189 protocols, it helps sending a "QUIT" message for example that prevents
9190 the server from logging a connection error for each health check. The
9191 check result will still be based on the ability to open the connection
9192 only.
9193
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03009194 - "tcp-check expect" only is mentioned : this is used to test a banner.
Willy Tarreaued179852013-12-16 01:07:00 +01009195 The connection is opened and haproxy waits for the server to present some
9196 contents which must validate some rules. The check result will be based
9197 on the matching between the contents and the rules. This is suited for
9198 POP, IMAP, SMTP, FTP, SSH, TELNET.
9199
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03009200 - both "tcp-check send" and "tcp-check expect" are mentioned : this is
Davor Ocelice9ed2812017-12-25 17:49:28 +01009201 used to test a hello-type protocol. HAProxy sends a message, the server
9202 responds and its response is analyzed. the check result will be based on
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03009203 the matching between the response contents and the rules. This is often
Willy Tarreaued179852013-12-16 01:07:00 +01009204 suited for protocols which require a binding or a request/response model.
9205 LDAP, MySQL, Redis and SSL are example of such protocols, though they
9206 already all have their dedicated checks with a deeper understanding of
9207 the respective protocols.
9208 In this mode, many questions may be sent and many answers may be
Davor Ocelice9ed2812017-12-25 17:49:28 +01009209 analyzed.
Willy Tarreaued179852013-12-16 01:07:00 +01009210
Christopher Fauletc52ea4d2020-04-23 15:43:35 +02009211 A fifth mode can be used to insert comments in different steps of the script.
Baptiste Assmannd60a9e52015-04-25 16:27:23 +02009212
Christopher Fauletc52ea4d2020-04-23 15:43:35 +02009213 For each tcp-check rule you create, you can add a "comment" directive,
9214 followed by a string. This string will be reported in the log and stderr in
9215 debug mode. It is useful to make user-friendly error reporting. The
9216 "comment" is of course optional.
Baptiste Assmannd60a9e52015-04-25 16:27:23 +02009217
Christopher Fauletc52ea4d2020-04-23 15:43:35 +02009218 During the execution of a health check, a variable scope is made available to
9219 store data samples, using the "tcp-check set-var" operation. Freeing those
9220 variable is possible using "tcp-check unset-var".
Gaetan Rivet0c39ecc2020-02-24 17:34:11 +01009221
Baptiste Assmannd60a9e52015-04-25 16:27:23 +02009222
Willy Tarreaued179852013-12-16 01:07:00 +01009223 Examples :
Davor Ocelice9ed2812017-12-25 17:49:28 +01009224 # perform a POP check (analyze only server's banner)
Willy Tarreaued179852013-12-16 01:07:00 +01009225 option tcp-check
Baptiste Assmannd60a9e52015-04-25 16:27:23 +02009226 tcp-check expect string +OK\ POP3\ ready comment POP\ protocol
Willy Tarreaued179852013-12-16 01:07:00 +01009227
Davor Ocelice9ed2812017-12-25 17:49:28 +01009228 # perform an IMAP check (analyze only server's banner)
Willy Tarreaued179852013-12-16 01:07:00 +01009229 option tcp-check
Baptiste Assmannd60a9e52015-04-25 16:27:23 +02009230 tcp-check expect string *\ OK\ IMAP4\ ready comment IMAP\ protocol
Willy Tarreaued179852013-12-16 01:07:00 +01009231
9232 # look for the redis master server after ensuring it speaks well
9233 # redis protocol, then it exits properly.
Davor Ocelice9ed2812017-12-25 17:49:28 +01009234 # (send a command then analyze the response 3 times)
Willy Tarreaued179852013-12-16 01:07:00 +01009235 option tcp-check
Baptiste Assmannd60a9e52015-04-25 16:27:23 +02009236 tcp-check comment PING\ phase
Willy Tarreaued179852013-12-16 01:07:00 +01009237 tcp-check send PING\r\n
Baptiste Assmanna3322992015-08-04 10:12:18 +02009238 tcp-check expect string +PONG
Baptiste Assmannd60a9e52015-04-25 16:27:23 +02009239 tcp-check comment role\ check
Willy Tarreaued179852013-12-16 01:07:00 +01009240 tcp-check send info\ replication\r\n
9241 tcp-check expect string role:master
Baptiste Assmannd60a9e52015-04-25 16:27:23 +02009242 tcp-check comment QUIT\ phase
Willy Tarreaued179852013-12-16 01:07:00 +01009243 tcp-check send QUIT\r\n
9244 tcp-check expect string +OK
9245
Davor Ocelice9ed2812017-12-25 17:49:28 +01009246 forge a HTTP request, then analyze the response
Willy Tarreaued179852013-12-16 01:07:00 +01009247 (send many headers before analyzing)
9248 option tcp-check
Baptiste Assmannd60a9e52015-04-25 16:27:23 +02009249 tcp-check comment forge\ and\ send\ HTTP\ request
Willy Tarreaued179852013-12-16 01:07:00 +01009250 tcp-check send HEAD\ /\ HTTP/1.1\r\n
9251 tcp-check send Host:\ www.mydomain.com\r\n
9252 tcp-check send User-Agent:\ HAProxy\ tcpcheck\r\n
9253 tcp-check send \r\n
Baptiste Assmannd60a9e52015-04-25 16:27:23 +02009254 tcp-check expect rstring HTTP/1\..\ (2..|3..) comment check\ HTTP\ response
Willy Tarreaued179852013-12-16 01:07:00 +01009255
9256
Christopher Fauletc52ea4d2020-04-23 15:43:35 +02009257 See also : "tcp-check connect", "tcp-check expect" and "tcp-check send".
Willy Tarreaued179852013-12-16 01:07:00 +01009258
9259
Willy Tarreau9ea05a72009-06-14 12:07:01 +02009260option tcp-smart-accept
9261no option tcp-smart-accept
9262 Enable or disable the saving of one ACK packet during the accept sequence
9263 May be used in sections : defaults | frontend | listen | backend
9264 yes | yes | yes | no
9265 Arguments : none
9266
9267 When an HTTP connection request comes in, the system acknowledges it on
9268 behalf of HAProxy, then the client immediately sends its request, and the
9269 system acknowledges it too while it is notifying HAProxy about the new
9270 connection. HAProxy then reads the request and responds. This means that we
9271 have one TCP ACK sent by the system for nothing, because the request could
9272 very well be acknowledged by HAProxy when it sends its response.
9273
9274 For this reason, in HTTP mode, HAProxy automatically asks the system to avoid
9275 sending this useless ACK on platforms which support it (currently at least
9276 Linux). It must not cause any problem, because the system will send it anyway
9277 after 40 ms if the response takes more time than expected to come.
9278
9279 During complex network debugging sessions, it may be desirable to disable
9280 this optimization because delayed ACKs can make troubleshooting more complex
9281 when trying to identify where packets are delayed. It is then possible to
Davor Ocelice9ed2812017-12-25 17:49:28 +01009282 fall back to normal behavior by specifying "no option tcp-smart-accept".
Willy Tarreau9ea05a72009-06-14 12:07:01 +02009283
9284 It is also possible to force it for non-HTTP proxies by simply specifying
9285 "option tcp-smart-accept". For instance, it can make sense with some services
9286 such as SMTP where the server speaks first.
9287
9288 It is recommended to avoid forcing this option in a defaults section. In case
9289 of doubt, consider setting it back to automatic values by prepending the
9290 "default" keyword before it, or disabling it using the "no" keyword.
9291
Willy Tarreaud88edf22009-06-14 15:48:17 +02009292 See also : "option tcp-smart-connect"
9293
9294
9295option tcp-smart-connect
9296no option tcp-smart-connect
9297 Enable or disable the saving of one ACK packet during the connect sequence
9298 May be used in sections : defaults | frontend | listen | backend
9299 yes | no | yes | yes
9300 Arguments : none
9301
9302 On certain systems (at least Linux), HAProxy can ask the kernel not to
9303 immediately send an empty ACK upon a connection request, but to directly
9304 send the buffer request instead. This saves one packet on the network and
9305 thus boosts performance. It can also be useful for some servers, because they
9306 immediately get the request along with the incoming connection.
9307
9308 This feature is enabled when "option tcp-smart-connect" is set in a backend.
9309 It is not enabled by default because it makes network troubleshooting more
9310 complex.
9311
9312 It only makes sense to enable it with protocols where the client speaks first
9313 such as HTTP. In other situations, if there is no data to send in place of
9314 the ACK, a normal ACK is sent.
9315
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 tcp-smart-accept"
9320
Willy Tarreau9ea05a72009-06-14 12:07:01 +02009321
Willy Tarreaubf1f8162007-12-28 17:42:56 +01009322option tcpka
9323 Enable or disable the sending of TCP keepalive packets on both sides
9324 May be used in sections : defaults | frontend | listen | backend
9325 yes | yes | yes | yes
9326 Arguments : none
9327
9328 When there is a firewall or any session-aware component between a client and
9329 a server, and when the protocol involves very long sessions with long idle
Davor Ocelice9ed2812017-12-25 17:49:28 +01009330 periods (e.g. remote desktops), there is a risk that one of the intermediate
Willy Tarreaubf1f8162007-12-28 17:42:56 +01009331 components decides to expire a session which has remained idle for too long.
9332
9333 Enabling socket-level TCP keep-alives makes the system regularly send packets
9334 to the other end of the connection, leaving it active. The delay between
9335 keep-alive probes is controlled by the system only and depends both on the
9336 operating system and its tuning parameters.
9337
9338 It is important to understand that keep-alive packets are neither emitted nor
9339 received at the application level. It is only the network stacks which sees
9340 them. For this reason, even if one side of the proxy already uses keep-alives
9341 to maintain its connection alive, those keep-alive packets will not be
9342 forwarded to the other side of the proxy.
9343
9344 Please note that this has nothing to do with HTTP keep-alive.
9345
9346 Using option "tcpka" enables the emission of TCP keep-alive probes on both
9347 the client and server sides of a connection. Note that this is meaningful
9348 only in "defaults" or "listen" sections. If this option is used in a
9349 frontend, only the client side will get keep-alives, and if this option is
9350 used in a backend, only the server side will get keep-alives. For this
9351 reason, it is strongly recommended to explicitly use "option clitcpka" and
9352 "option srvtcpka" when the configuration is split between frontends and
9353 backends.
9354
9355 See also : "option clitcpka", "option srvtcpka"
9356
Willy Tarreau844e3c52008-01-11 16:28:18 +01009357
9358option tcplog
9359 Enable advanced logging of TCP connections with session state and timers
9360 May be used in sections : defaults | frontend | listen | backend
Tim Duesterhus9ad9f352018-02-05 20:52:27 +01009361 yes | yes | yes | no
Willy Tarreau844e3c52008-01-11 16:28:18 +01009362 Arguments : none
9363
9364 By default, the log output format is very poor, as it only contains the
9365 source and destination addresses, and the instance name. By specifying
9366 "option tcplog", each log line turns into a much richer format including, but
9367 not limited to, the connection timers, the session status, the connections
9368 numbers, the frontend, backend and server name, and of course the source
9369 address and ports. This option is useful for pure TCP proxies in order to
9370 find which of the client or server disconnects or times out. For normal HTTP
9371 proxies, it's better to use "option httplog" which is even more complete.
9372
Guillaume de Lafond29f45602017-03-31 19:52:15 +02009373 "option tcplog" overrides any previous "log-format" directive.
9374
Willy Tarreauc57f0e22009-05-10 13:12:33 +02009375 See also : "option httplog", and section 8 about logging.
Willy Tarreau844e3c52008-01-11 16:28:18 +01009376
9377
Willy Tarreau844e3c52008-01-11 16:28:18 +01009378option transparent
9379no option transparent
9380 Enable client-side transparent proxying
9381 May be used in sections : defaults | frontend | listen | backend
Willy Tarreau4b1f8592008-12-23 23:13:55 +01009382 yes | no | yes | yes
Willy Tarreau844e3c52008-01-11 16:28:18 +01009383 Arguments : none
9384
9385 This option was introduced in order to provide layer 7 persistence to layer 3
9386 load balancers. The idea is to use the OS's ability to redirect an incoming
9387 connection for a remote address to a local process (here HAProxy), and let
9388 this process know what address was initially requested. When this option is
9389 used, sessions without cookies will be forwarded to the original destination
9390 IP address of the incoming request (which should match that of another
9391 equipment), while requests with cookies will still be forwarded to the
9392 appropriate server.
9393
9394 Note that contrary to a common belief, this option does NOT make HAProxy
9395 present the client's IP to the server when establishing the connection.
9396
Willy Tarreaua1146052011-03-01 09:51:54 +01009397 See also: the "usesrc" argument of the "source" keyword, and the
Willy Tarreaueabeafa2008-01-16 16:17:06 +01009398 "transparent" option of the "bind" keyword.
Willy Tarreau844e3c52008-01-11 16:28:18 +01009399
Willy Tarreaubf1f8162007-12-28 17:42:56 +01009400
Simon Horman98637e52014-06-20 12:30:16 +09009401external-check command <command>
9402 Executable to run when performing an external-check
9403 May be used in sections : defaults | frontend | listen | backend
9404 yes | no | yes | yes
9405
9406 Arguments :
9407 <command> is the external command to run
9408
Simon Horman98637e52014-06-20 12:30:16 +09009409 The arguments passed to the to the command are:
9410
Cyril Bonté777be862014-12-02 21:21:35 +01009411 <proxy_address> <proxy_port> <server_address> <server_port>
Simon Horman98637e52014-06-20 12:30:16 +09009412
Cyril Bonté777be862014-12-02 21:21:35 +01009413 The <proxy_address> and <proxy_port> are derived from the first listener
9414 that is either IPv4, IPv6 or a UNIX socket. In the case of a UNIX socket
9415 listener the proxy_address will be the path of the socket and the
9416 <proxy_port> will be the string "NOT_USED". In a backend section, it's not
9417 possible to determine a listener, and both <proxy_address> and <proxy_port>
9418 will have the string value "NOT_USED".
Simon Horman98637e52014-06-20 12:30:16 +09009419
Cyril Bonté72cda2a2014-12-27 22:28:39 +01009420 Some values are also provided through environment variables.
9421
9422 Environment variables :
9423 HAPROXY_PROXY_ADDR The first bind address if available (or empty if not
9424 applicable, for example in a "backend" section).
9425
9426 HAPROXY_PROXY_ID The backend id.
9427
9428 HAPROXY_PROXY_NAME The backend name.
9429
9430 HAPROXY_PROXY_PORT The first bind port if available (or empty if not
9431 applicable, for example in a "backend" section or
9432 for a UNIX socket).
9433
9434 HAPROXY_SERVER_ADDR The server address.
9435
9436 HAPROXY_SERVER_CURCONN The current number of connections on the server.
9437
9438 HAPROXY_SERVER_ID The server id.
9439
9440 HAPROXY_SERVER_MAXCONN The server max connections.
9441
9442 HAPROXY_SERVER_NAME The server name.
9443
9444 HAPROXY_SERVER_PORT The server port if available (or empty for a UNIX
9445 socket).
9446
9447 PATH The PATH environment variable used when executing
9448 the command may be set using "external-check path".
9449
William Lallemand4d03e432019-06-14 15:35:37 +02009450 See also "2.3. Environment variables" for other variables.
9451
Simon Horman98637e52014-06-20 12:30:16 +09009452 If the command executed and exits with a zero status then the check is
9453 considered to have passed, otherwise the check is considered to have
9454 failed.
9455
9456 Example :
9457 external-check command /bin/true
9458
9459 See also : "external-check", "option external-check", "external-check path"
9460
9461
9462external-check path <path>
9463 The value of the PATH environment variable used when running an external-check
9464 May be used in sections : defaults | frontend | listen | backend
9465 yes | no | yes | yes
9466
9467 Arguments :
9468 <path> is the path used when executing external command to run
9469
9470 The default path is "".
9471
9472 Example :
9473 external-check path "/usr/bin:/bin"
9474
9475 See also : "external-check", "option external-check",
9476 "external-check command"
9477
9478
Emeric Brun647caf12009-06-30 17:57:00 +02009479persist rdp-cookie
Hervé COMMOWICKa3eb39c2011-08-05 18:48:51 +02009480persist rdp-cookie(<name>)
Emeric Brun647caf12009-06-30 17:57:00 +02009481 Enable RDP cookie-based persistence
9482 May be used in sections : defaults | frontend | listen | backend
9483 yes | no | yes | yes
9484 Arguments :
9485 <name> is the optional name of the RDP cookie to check. If omitted, the
Willy Tarreau61e28f22010-05-16 22:31:05 +02009486 default cookie name "msts" will be used. There currently is no
9487 valid reason to change this name.
Emeric Brun647caf12009-06-30 17:57:00 +02009488
9489 This statement enables persistence based on an RDP cookie. The RDP cookie
9490 contains all information required to find the server in the list of known
Davor Ocelice9ed2812017-12-25 17:49:28 +01009491 servers. So when this option is set in the backend, the request is analyzed
Emeric Brun647caf12009-06-30 17:57:00 +02009492 and if an RDP cookie is found, it is decoded. If it matches a known server
9493 which is still UP (or if "option persist" is set), then the connection is
9494 forwarded to this server.
9495
9496 Note that this only makes sense in a TCP backend, but for this to work, the
9497 frontend must have waited long enough to ensure that an RDP cookie is present
9498 in the request buffer. This is the same requirement as with the "rdp-cookie"
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01009499 load-balancing method. Thus it is highly recommended to put all statements in
Emeric Brun647caf12009-06-30 17:57:00 +02009500 a single "listen" section.
9501
Willy Tarreau61e28f22010-05-16 22:31:05 +02009502 Also, it is important to understand that the terminal server will emit this
9503 RDP cookie only if it is configured for "token redirection mode", which means
9504 that the "IP address redirection" option is disabled.
9505
Emeric Brun647caf12009-06-30 17:57:00 +02009506 Example :
9507 listen tse-farm
9508 bind :3389
9509 # wait up to 5s for an RDP cookie in the request
9510 tcp-request inspect-delay 5s
9511 tcp-request content accept if RDP_COOKIE
9512 # apply RDP cookie persistence
9513 persist rdp-cookie
9514 # if server is unknown, let's balance on the same cookie.
Cyril Bontédc4d9032012-04-08 21:57:39 +02009515 # alternatively, "balance leastconn" may be useful too.
Emeric Brun647caf12009-06-30 17:57:00 +02009516 balance rdp-cookie
9517 server srv1 1.1.1.1:3389
9518 server srv2 1.1.1.2:3389
9519
Simon Hormanab814e02011-06-24 14:50:20 +09009520 See also : "balance rdp-cookie", "tcp-request", the "req_rdp_cookie" ACL and
9521 the rdp_cookie pattern fetch function.
Emeric Brun647caf12009-06-30 17:57:00 +02009522
9523
Willy Tarreau3a7d2072009-03-05 23:48:25 +01009524rate-limit sessions <rate>
9525 Set a limit on the number of new sessions accepted per second on a frontend
9526 May be used in sections : defaults | frontend | listen | backend
9527 yes | yes | yes | no
9528 Arguments :
9529 <rate> The <rate> parameter is an integer designating the maximum number
9530 of new sessions per second to accept on the frontend.
9531
9532 When the frontend reaches the specified number of new sessions per second, it
9533 stops accepting new connections until the rate drops below the limit again.
9534 During this time, the pending sessions will be kept in the socket's backlog
9535 (in system buffers) and haproxy will not even be aware that sessions are
9536 pending. When applying very low limit on a highly loaded service, it may make
9537 sense to increase the socket's backlog using the "backlog" keyword.
9538
9539 This feature is particularly efficient at blocking connection-based attacks
9540 or service abuse on fragile servers. Since the session rate is measured every
9541 millisecond, it is extremely accurate. Also, the limit applies immediately,
9542 no delay is needed at all to detect the threshold.
9543
9544 Example : limit the connection rate on SMTP to 10 per second max
9545 listen smtp
9546 mode tcp
9547 bind :25
9548 rate-limit sessions 10
Panagiotis Panagiotopoulos7282d8e2016-02-11 16:37:15 +02009549 server smtp1 127.0.0.1:1025
Willy Tarreau3a7d2072009-03-05 23:48:25 +01009550
Willy Tarreaua17c2d92011-07-25 08:16:20 +02009551 Note : when the maximum rate is reached, the frontend's status is not changed
9552 but its sockets appear as "WAITING" in the statistics if the
9553 "socket-stats" option is enabled.
Willy Tarreau3a7d2072009-03-05 23:48:25 +01009554
9555 See also : the "backlog" keyword and the "fe_sess_rate" ACL criterion.
9556
9557
Willy Tarreau2e1dca82012-09-12 08:43:15 +02009558redirect location <loc> [code <code>] <option> [{if | unless} <condition>]
9559redirect prefix <pfx> [code <code>] <option> [{if | unless} <condition>]
9560redirect scheme <sch> [code <code>] <option> [{if | unless} <condition>]
Willy Tarreaub463dfb2008-06-07 23:08:56 +02009561 Return an HTTP redirection if/unless a condition is matched
9562 May be used in sections : defaults | frontend | listen | backend
9563 no | yes | yes | yes
9564
9565 If/unless the condition is matched, the HTTP request will lead to a redirect
Willy Tarreauf285f542010-01-03 20:03:03 +01009566 response. If no condition is specified, the redirect applies unconditionally.
Willy Tarreaub463dfb2008-06-07 23:08:56 +02009567
Willy Tarreau0140f252008-11-19 21:07:09 +01009568 Arguments :
Willy Tarreau2e1dca82012-09-12 08:43:15 +02009569 <loc> With "redirect location", the exact value in <loc> is placed into
Thierry FOURNIERd18cd0f2013-11-29 12:15:45 +01009570 the HTTP "Location" header. When used in an "http-request" rule,
9571 <loc> value follows the log-format rules and can include some
9572 dynamic values (see Custom Log Format in section 8.2.4).
Willy Tarreau2e1dca82012-09-12 08:43:15 +02009573
9574 <pfx> With "redirect prefix", the "Location" header is built from the
9575 concatenation of <pfx> and the complete URI path, including the
9576 query string, unless the "drop-query" option is specified (see
9577 below). As a special case, if <pfx> equals exactly "/", then
9578 nothing is inserted before the original URI. It allows one to
Thierry FOURNIERd18cd0f2013-11-29 12:15:45 +01009579 redirect to the same URL (for instance, to insert a cookie). When
9580 used in an "http-request" rule, <pfx> value follows the log-format
9581 rules and can include some dynamic values (see Custom Log Format
9582 in section 8.2.4).
Willy Tarreau2e1dca82012-09-12 08:43:15 +02009583
9584 <sch> With "redirect scheme", then the "Location" header is built by
9585 concatenating <sch> with "://" then the first occurrence of the
9586 "Host" header, and then the URI path, including the query string
9587 unless the "drop-query" option is specified (see below). If no
9588 path is found or if the path is "*", then "/" is used instead. If
9589 no "Host" header is found, then an empty host component will be
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03009590 returned, which most recent browsers interpret as redirecting to
Willy Tarreau2e1dca82012-09-12 08:43:15 +02009591 the same host. This directive is mostly used to redirect HTTP to
Thierry FOURNIERd18cd0f2013-11-29 12:15:45 +01009592 HTTPS. When used in an "http-request" rule, <sch> value follows
9593 the log-format rules and can include some dynamic values (see
9594 Custom Log Format in section 8.2.4).
Willy Tarreau0140f252008-11-19 21:07:09 +01009595
9596 <code> The code is optional. It indicates which type of HTTP redirection
Willy Tarreaub67fdc42013-03-29 19:28:11 +01009597 is desired. Only codes 301, 302, 303, 307 and 308 are supported,
9598 with 302 used by default if no code is specified. 301 means
9599 "Moved permanently", and a browser may cache the Location. 302
Baptiste Assmannea849c02015-08-03 11:42:50 +02009600 means "Moved temporarily" and means that the browser should not
Willy Tarreaub67fdc42013-03-29 19:28:11 +01009601 cache the redirection. 303 is equivalent to 302 except that the
9602 browser will fetch the location with a GET method. 307 is just
9603 like 302 but makes it clear that the same method must be reused.
9604 Likewise, 308 replaces 301 if the same method must be used.
Willy Tarreau0140f252008-11-19 21:07:09 +01009605
9606 <option> There are several options which can be specified to adjust the
Davor Ocelice9ed2812017-12-25 17:49:28 +01009607 expected behavior of a redirection :
Willy Tarreau0140f252008-11-19 21:07:09 +01009608
9609 - "drop-query"
9610 When this keyword is used in a prefix-based redirection, then the
9611 location will be set without any possible query-string, which is useful
9612 for directing users to a non-secure page for instance. It has no effect
9613 with a location-type redirect.
9614
Willy Tarreau81e3b4f2010-01-10 00:42:19 +01009615 - "append-slash"
9616 This keyword may be used in conjunction with "drop-query" to redirect
9617 users who use a URL not ending with a '/' to the same one with the '/'.
9618 It can be useful to ensure that search engines will only see one URL.
9619 For this, a return code 301 is preferred.
9620
Willy Tarreau0140f252008-11-19 21:07:09 +01009621 - "set-cookie NAME[=value]"
9622 A "Set-Cookie" header will be added with NAME (and optionally "=value")
9623 to the response. This is sometimes used to indicate that a user has
9624 been seen, for instance to protect against some types of DoS. No other
9625 cookie option is added, so the cookie will be a session cookie. Note
9626 that for a browser, a sole cookie name without an equal sign is
9627 different from a cookie with an equal sign.
9628
9629 - "clear-cookie NAME[=]"
9630 A "Set-Cookie" header will be added with NAME (and optionally "="), but
9631 with the "Max-Age" attribute set to zero. This will tell the browser to
9632 delete this cookie. It is useful for instance on logout pages. It is
9633 important to note that clearing the cookie "NAME" will not remove a
9634 cookie set with "NAME=value". You have to clear the cookie "NAME=" for
9635 that, because the browser makes the difference.
Willy Tarreaub463dfb2008-06-07 23:08:56 +02009636
9637 Example: move the login URL only to HTTPS.
9638 acl clear dst_port 80
9639 acl secure dst_port 8080
9640 acl login_page url_beg /login
Willy Tarreau0140f252008-11-19 21:07:09 +01009641 acl logout url_beg /logout
Willy Tarreau79da4692008-11-19 20:03:04 +01009642 acl uid_given url_reg /login?userid=[^&]+
Willy Tarreau0140f252008-11-19 21:07:09 +01009643 acl cookie_set hdr_sub(cookie) SEEN=1
9644
9645 redirect prefix https://mysite.com set-cookie SEEN=1 if !cookie_set
Willy Tarreau79da4692008-11-19 20:03:04 +01009646 redirect prefix https://mysite.com if login_page !secure
9647 redirect prefix http://mysite.com drop-query if login_page !uid_given
9648 redirect location http://mysite.com/ if !login_page secure
Willy Tarreau0140f252008-11-19 21:07:09 +01009649 redirect location / clear-cookie USERID= if logout
Willy Tarreaub463dfb2008-06-07 23:08:56 +02009650
Willy Tarreau81e3b4f2010-01-10 00:42:19 +01009651 Example: send redirects for request for articles without a '/'.
9652 acl missing_slash path_reg ^/article/[^/]*$
9653 redirect code 301 prefix / drop-query append-slash if missing_slash
9654
Willy Tarreau2e1dca82012-09-12 08:43:15 +02009655 Example: redirect all HTTP traffic to HTTPS when SSL is handled by haproxy.
David BERARDe7153042012-11-03 00:11:31 +01009656 redirect scheme https if !{ ssl_fc }
Willy Tarreau2e1dca82012-09-12 08:43:15 +02009657
Thierry FOURNIERd18cd0f2013-11-29 12:15:45 +01009658 Example: append 'www.' prefix in front of all hosts not having it
Coen Rosdorff596659b2016-04-11 11:33:49 +02009659 http-request redirect code 301 location \
9660 http://www.%[hdr(host)]%[capture.req.uri] \
9661 unless { hdr_beg(host) -i www }
Thierry FOURNIERd18cd0f2013-11-29 12:15:45 +01009662
Willy Tarreauc57f0e22009-05-10 13:12:33 +02009663 See section 7 about ACL usage.
Willy Tarreaub463dfb2008-06-07 23:08:56 +02009664
Willy Tarreau303c0352008-01-17 19:01:39 +01009665
Willy Tarreaue5c5ce92008-06-20 17:27:19 +02009666retries <value>
9667 Set the number of retries to perform on a server after a connection failure
9668 May be used in sections: defaults | frontend | listen | backend
9669 yes | no | yes | yes
9670 Arguments :
9671 <value> is the number of times a connection attempt should be retried on
9672 a server when a connection either is refused or times out. The
9673 default value is 3.
9674
9675 It is important to understand that this value applies to the number of
9676 connection attempts, not full requests. When a connection has effectively
9677 been established to a server, there will be no more retry.
9678
9679 In order to avoid immediate reconnections to a server which is restarting,
Joseph Lynch726ab712015-05-11 23:25:34 -07009680 a turn-around timer of min("timeout connect", one second) is applied before
9681 a retry occurs.
Willy Tarreaue5c5ce92008-06-20 17:27:19 +02009682
9683 When "option redispatch" is set, the last retry may be performed on another
9684 server even if a cookie references a different server.
9685
9686 See also : "option redispatch"
9687
9688
Olivier Houcharda254a372019-04-05 15:30:12 +02009689retry-on [list of keywords]
Jerome Magnin5ce3c142020-05-13 20:09:57 +02009690 Specify when to attempt to automatically retry a failed request.
9691 This setting is only valid when "mode" is set to http and is silently ignored
9692 otherwise.
Olivier Houcharda254a372019-04-05 15:30:12 +02009693 May be used in sections: defaults | frontend | listen | backend
9694 yes | no | yes | yes
9695 Arguments :
9696 <keywords> is a list of keywords or HTTP status codes, each representing a
9697 type of failure event on which an attempt to retry the request
9698 is desired. Please read the notes at the bottom before changing
9699 this setting. The following keywords are supported :
9700
9701 none never retry
9702
9703 conn-failure retry when the connection or the SSL handshake failed
9704 and the request could not be sent. This is the default.
9705
9706 empty-response retry when the server connection was closed after part
9707 of the request was sent, and nothing was received from
9708 the server. This type of failure may be caused by the
9709 request timeout on the server side, poor network
9710 condition, or a server crash or restart while
9711 processing the request.
9712
Olivier Houcharde3249a92019-05-03 23:01:47 +02009713 junk-response retry when the server returned something not looking
9714 like a complete HTTP response. This includes partial
9715 responses headers as well as non-HTTP contents. It
9716 usually is a bad idea to retry on such events, which
9717 may be caused a configuration issue (wrong server port)
9718 or by the request being harmful to the server (buffer
9719 overflow attack for example).
9720
Olivier Houcharda254a372019-04-05 15:30:12 +02009721 response-timeout the server timeout stroke while waiting for the server
9722 to respond to the request. This may be caused by poor
9723 network condition, the reuse of an idle connection
9724 which has expired on the path, or by the request being
9725 extremely expensive to process. It generally is a bad
9726 idea to retry on such events on servers dealing with
9727 heavy database processing (full scans, etc) as it may
9728 amplify denial of service attacks.
9729
Olivier Houchard865d8392019-05-03 22:46:27 +02009730 0rtt-rejected retry requests which were sent over early data and were
9731 rejected by the server. These requests are generally
9732 considered to be safe to retry.
9733
Julien Pivotto2de240a2020-11-12 11:14:05 +01009734 <status> any HTTP status code among "401" (Unauthorized), "403"
9735 (Forbidden), "404" (Not Found), "408" (Request Timeout),
9736 "425" (Too Early), "500" (Server Error), "501" (Not
9737 Implemented), "502" (Bad Gateway), "503" (Service
9738 Unavailable), "504" (Gateway Timeout).
Olivier Houcharda254a372019-04-05 15:30:12 +02009739
Olivier Houchardddf0e032019-05-10 18:05:40 +02009740 all-retryable-errors
9741 retry request for any error that are considered
9742 retryable. This currently activates "conn-failure",
9743 "empty-response", "junk-response", "response-timeout",
9744 "0rtt-rejected", "500", "502", "503", and "504".
9745
Olivier Houcharda254a372019-04-05 15:30:12 +02009746 Using this directive replaces any previous settings with the new ones; it is
9747 not cumulative.
9748
9749 Please note that using anything other than "none" and "conn-failure" requires
9750 to allocate a buffer and copy the whole request into it, so it has memory and
9751 performance impacts. Requests not fitting in a single buffer will never be
9752 retried (see the global tune.bufsize setting).
9753
9754 You have to make sure the application has a replay protection mechanism built
9755 in such as a unique transaction IDs passed in requests, or that replaying the
9756 same request has no consequence, or it is very dangerous to use any retry-on
9757 value beside "conn-failure" and "none". Static file servers and caches are
9758 generally considered safe against any type of retry. Using a status code can
9759 be useful to quickly leave a server showing an abnormal behavior (out of
9760 memory, file system issues, etc), but in this case it may be a good idea to
9761 immediately redispatch the connection to another server (please see "option
9762 redispatch" for this). Last, it is important to understand that most causes
9763 of failures are the requests themselves and that retrying a request causing a
9764 server to misbehave will often make the situation even worse for this server,
9765 or for the whole service in case of redispatch.
9766
9767 Unless you know exactly how the application deals with replayed requests, you
9768 should not use this directive.
9769
9770 The default is "conn-failure".
9771
9772 See also: "retries", "option redispatch", "tune.bufsize"
9773
David du Colombier486df472011-03-17 10:40:26 +01009774server <name> <address>[:[port]] [param*]
Willy Tarreaueabeafa2008-01-16 16:17:06 +01009775 Declare a server in a backend
9776 May be used in sections : defaults | frontend | listen | backend
9777 no | no | yes | yes
9778 Arguments :
9779 <name> is the internal name assigned to this server. This name will
Davor Ocelice9ed2812017-12-25 17:49:28 +01009780 appear in logs and alerts. If "http-send-name-header" is
Mark Lamourinec2247f02012-01-04 13:02:01 -05009781 set, it will be added to the request header sent to the server.
Willy Tarreaueabeafa2008-01-16 16:17:06 +01009782
David du Colombier486df472011-03-17 10:40:26 +01009783 <address> is the IPv4 or IPv6 address of the server. Alternatively, a
9784 resolvable hostname is supported, but this name will be resolved
9785 during start-up. Address "0.0.0.0" or "*" has a special meaning.
9786 It indicates that the connection will be forwarded to the same IP
Willy Tarreaud669a4f2010-07-13 14:49:50 +02009787 address as the one from the client connection. This is useful in
9788 transparent proxy architectures where the client's connection is
9789 intercepted and haproxy must forward to the original destination
9790 address. This is more or less what the "transparent" keyword does
9791 except that with a server it's possible to limit concurrency and
Willy Tarreau24709282013-03-10 21:32:12 +01009792 to report statistics. Optionally, an address family prefix may be
9793 used before the address to force the family regardless of the
9794 address format, which can be useful to specify a path to a unix
9795 socket with no slash ('/'). Currently supported prefixes are :
9796 - 'ipv4@' -> address is always IPv4
9797 - 'ipv6@' -> address is always IPv6
9798 - 'unix@' -> address is a path to a local unix socket
Willy Tarreauccfccef2014-05-10 01:49:15 +02009799 - 'abns@' -> address is in abstract namespace (Linux only)
William Lallemand2fe7dd02018-09-11 16:51:29 +02009800 - 'sockpair@' -> address is the FD of a connected unix
9801 socket or of a socketpair. During a connection, the
9802 backend creates a pair of connected sockets, and passes
9803 one of them over the FD. The bind part will use the
9804 received socket as the client FD. Should be used
9805 carefully.
William Lallemandb2f07452015-05-12 14:27:13 +02009806 You may want to reference some environment variables in the
9807 address parameter, see section 2.3 about environment
Willy Tarreau6a031d12016-11-07 19:42:35 +01009808 variables. The "init-addr" setting can be used to modify the way
9809 IP addresses should be resolved upon startup.
Willy Tarreaueabeafa2008-01-16 16:17:06 +01009810
Willy Tarreaub6205fd2012-09-24 12:27:33 +02009811 <port> is an optional port specification. If set, all connections will
Willy Tarreaueabeafa2008-01-16 16:17:06 +01009812 be sent to this port. If unset, the same port the client
9813 connected to will be used. The port may also be prefixed by a "+"
9814 or a "-". In this case, the server's port will be determined by
9815 adding this value to the client's port.
9816
9817 <param*> is a list of parameters for this server. The "server" keywords
9818 accepts an important number of options and has a complete section
Willy Tarreauc57f0e22009-05-10 13:12:33 +02009819 dedicated to it. Please refer to section 5 for more details.
Willy Tarreaueabeafa2008-01-16 16:17:06 +01009820
9821 Examples :
9822 server first 10.1.1.1:1080 cookie first check inter 1000
9823 server second 10.1.1.2:1080 cookie second check inter 1000
Willy Tarreau24709282013-03-10 21:32:12 +01009824 server transp ipv4@
William Lallemandb2f07452015-05-12 14:27:13 +02009825 server backup "${SRV_BACKUP}:1080" backup
9826 server www1_dc1 "${LAN_DC1}.101:80"
9827 server www1_dc2 "${LAN_DC2}.101:80"
Willy Tarreaueabeafa2008-01-16 16:17:06 +01009828
Willy Tarreau55dcaf62015-09-27 15:03:15 +02009829 Note: regarding Linux's abstract namespace sockets, HAProxy uses the whole
9830 sun_path length is used for the address length. Some other programs
9831 such as socat use the string length only by default. Pass the option
9832 ",unix-tightsocklen=0" to any abstract socket definition in socat to
9833 make it compatible with HAProxy's.
9834
Mark Lamourinec2247f02012-01-04 13:02:01 -05009835 See also: "default-server", "http-send-name-header" and section 5 about
9836 server options
Willy Tarreaueabeafa2008-01-16 16:17:06 +01009837
Christopher Faulet583b6de2021-02-12 09:27:10 +01009838server-state-file-name [ { use-backend-name | <file> } ]
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02009839 Set the server state file to read, load and apply to servers available in
Christopher Faulet583b6de2021-02-12 09:27:10 +01009840 this backend.
9841 May be used in sections: defaults | frontend | listen | backend
9842 no | no | yes | yes
9843
9844 It only applies when the directive "load-server-state-from-file" is set to
9845 "local". When <file> is not provided, if "use-backend-name" is used or if
9846 this directive is not set, then backend name is used. If <file> starts with a
9847 slash '/', then it is considered as an absolute path. Otherwise, <file> is
9848 concatenated to the global directive "server-state-base".
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02009849
9850 Example: the minimal configuration below would make HAProxy look for the
9851 state server file '/etc/haproxy/states/bk':
9852
9853 global
9854 server-state-file-base /etc/haproxy/states
9855
Willy Tarreauc9c6cdb2020-03-05 16:03:58 +01009856 backend bk
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02009857 load-server-state-from-file
9858
Christopher Faulet583b6de2021-02-12 09:27:10 +01009859 See also: "server-state-base", "load-server-state-from-file", and
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02009860 "show servers state"
Willy Tarreaueabeafa2008-01-16 16:17:06 +01009861
Frédéric Lécaillecb4502e2017-04-20 13:36:25 +02009862server-template <prefix> <num | range> <fqdn>[:<port>] [params*]
9863 Set a template to initialize servers with shared parameters.
9864 The names of these servers are built from <prefix> and <num | range> parameters.
9865 May be used in sections : defaults | frontend | listen | backend
9866 no | no | yes | yes
9867
9868 Arguments:
9869 <prefix> A prefix for the server names to be built.
9870
9871 <num | range>
9872 If <num> is provided, this template initializes <num> servers
9873 with 1 up to <num> as server name suffixes. A range of numbers
9874 <num_low>-<num_high> may also be used to use <num_low> up to
9875 <num_high> as server name suffixes.
9876
9877 <fqdn> A FQDN for all the servers this template initializes.
9878
9879 <port> Same meaning as "server" <port> argument (see "server" keyword).
9880
9881 <params*>
9882 Remaining server parameters among all those supported by "server"
9883 keyword.
9884
9885 Examples:
9886 # Initializes 3 servers with srv1, srv2 and srv3 as names,
9887 # google.com as FQDN, and health-check enabled.
9888 server-template srv 1-3 google.com:80 check
9889
9890 # or
9891 server-template srv 3 google.com:80 check
9892
9893 # would be equivalent to:
9894 server srv1 google.com:80 check
9895 server srv2 google.com:80 check
9896 server srv3 google.com:80 check
9897
9898
9899
Willy Tarreaueabeafa2008-01-16 16:17:06 +01009900source <addr>[:<port>] [usesrc { <addr2>[:<port2>] | client | clientip } ]
Willy Tarreaubce70882009-09-07 11:51:47 +02009901source <addr>[:<port>] [usesrc { <addr2>[:<port2>] | hdr_ip(<hdr>[,<occ>]) } ]
Willy Tarreaud53f96b2009-02-04 18:46:54 +01009902source <addr>[:<port>] [interface <name>]
Willy Tarreaueabeafa2008-01-16 16:17:06 +01009903 Set the source address for outgoing connections
9904 May be used in sections : defaults | frontend | listen | backend
9905 yes | no | yes | yes
9906 Arguments :
9907 <addr> is the IPv4 address HAProxy will bind to before connecting to a
9908 server. This address is also used as a source for health checks.
Willy Tarreau24709282013-03-10 21:32:12 +01009909
Willy Tarreaueabeafa2008-01-16 16:17:06 +01009910 The default value of 0.0.0.0 means that the system will select
Willy Tarreau24709282013-03-10 21:32:12 +01009911 the most appropriate address to reach its destination. Optionally
9912 an address family prefix may be used before the address to force
9913 the family regardless of the address format, which can be useful
9914 to specify a path to a unix socket with no slash ('/'). Currently
9915 supported prefixes are :
9916 - 'ipv4@' -> address is always IPv4
9917 - 'ipv6@' -> address is always IPv6
9918 - 'unix@' -> address is a path to a local unix socket
Willy Tarreauccfccef2014-05-10 01:49:15 +02009919 - 'abns@' -> address is in abstract namespace (Linux only)
Cyril Bonté307ee1e2015-09-28 23:16:06 +02009920 You may want to reference some environment variables in the
9921 address parameter, see section 2.3 about environment variables.
Willy Tarreaueabeafa2008-01-16 16:17:06 +01009922
9923 <port> is an optional port. It is normally not needed but may be useful
9924 in some very specific contexts. The default value of zero means
Willy Tarreauc6f4ce82009-06-10 11:09:37 +02009925 the system will select a free port. Note that port ranges are not
9926 supported in the backend. If you want to force port ranges, you
9927 have to specify them on each "server" line.
Willy Tarreaueabeafa2008-01-16 16:17:06 +01009928
9929 <addr2> is the IP address to present to the server when connections are
9930 forwarded in full transparent proxy mode. This is currently only
9931 supported on some patched Linux kernels. When this address is
9932 specified, clients connecting to the server will be presented
9933 with this address, while health checks will still use the address
9934 <addr>.
9935
9936 <port2> is the optional port to present to the server when connections
9937 are forwarded in full transparent proxy mode (see <addr2> above).
9938 The default value of zero means the system will select a free
9939 port.
9940
Willy Tarreaubce70882009-09-07 11:51:47 +02009941 <hdr> is the name of a HTTP header in which to fetch the IP to bind to.
9942 This is the name of a comma-separated header list which can
9943 contain multiple IP addresses. By default, the last occurrence is
9944 used. This is designed to work with the X-Forwarded-For header
Baptiste Assmannea3e73b2013-02-02 23:47:49 +01009945 and to automatically bind to the client's IP address as seen
Willy Tarreaubce70882009-09-07 11:51:47 +02009946 by previous proxy, typically Stunnel. In order to use another
9947 occurrence from the last one, please see the <occ> parameter
9948 below. When the header (or occurrence) is not found, no binding
9949 is performed so that the proxy's default IP address is used. Also
9950 keep in mind that the header name is case insensitive, as for any
9951 HTTP header.
9952
9953 <occ> is the occurrence number of a value to be used in a multi-value
9954 header. This is to be used in conjunction with "hdr_ip(<hdr>)",
Jamie Gloudonaaa21002012-08-25 00:18:33 -04009955 in order to specify which occurrence to use for the source IP
Willy Tarreaubce70882009-09-07 11:51:47 +02009956 address. Positive values indicate a position from the first
9957 occurrence, 1 being the first one. Negative values indicate
9958 positions relative to the last one, -1 being the last one. This
9959 is helpful for situations where an X-Forwarded-For header is set
9960 at the entry point of an infrastructure and must be used several
9961 proxy layers away. When this value is not specified, -1 is
9962 assumed. Passing a zero here disables the feature.
9963
Willy Tarreaud53f96b2009-02-04 18:46:54 +01009964 <name> is an optional interface name to which to bind to for outgoing
9965 traffic. On systems supporting this features (currently, only
9966 Linux), this allows one to bind all traffic to the server to
9967 this interface even if it is not the one the system would select
9968 based on routing tables. This should be used with extreme care.
9969 Note that using this option requires root privileges.
9970
Willy Tarreaueabeafa2008-01-16 16:17:06 +01009971 The "source" keyword is useful in complex environments where a specific
9972 address only is allowed to connect to the servers. It may be needed when a
9973 private address must be used through a public gateway for instance, and it is
9974 known that the system cannot determine the adequate source address by itself.
9975
9976 An extension which is available on certain patched Linux kernels may be used
9977 through the "usesrc" optional keyword. It makes it possible to connect to the
9978 servers with an IP address which does not belong to the system itself. This
9979 is called "full transparent proxy mode". For this to work, the destination
9980 servers have to route their traffic back to this address through the machine
9981 running HAProxy, and IP forwarding must generally be enabled on this machine.
9982
9983 In this "full transparent proxy" mode, it is possible to force a specific IP
9984 address to be presented to the servers. This is not much used in fact. A more
9985 common use is to tell HAProxy to present the client's IP address. For this,
9986 there are two methods :
9987
9988 - present the client's IP and port addresses. This is the most transparent
9989 mode, but it can cause problems when IP connection tracking is enabled on
9990 the machine, because a same connection may be seen twice with different
9991 states. However, this solution presents the huge advantage of not
9992 limiting the system to the 64k outgoing address+port couples, because all
9993 of the client ranges may be used.
9994
9995 - present only the client's IP address and select a spare port. This
9996 solution is still quite elegant but slightly less transparent (downstream
9997 firewalls logs will not match upstream's). It also presents the downside
9998 of limiting the number of concurrent connections to the usual 64k ports.
9999 However, since the upstream and downstream ports are different, local IP
10000 connection tracking on the machine will not be upset by the reuse of the
10001 same session.
10002
Willy Tarreaueabeafa2008-01-16 16:17:06 +010010003 This option sets the default source for all servers in the backend. It may
10004 also be specified in a "defaults" section. Finer source address specification
10005 is possible at the server level using the "source" server option. Refer to
Willy Tarreauc57f0e22009-05-10 13:12:33 +020010006 section 5 for more information.
Willy Tarreaueabeafa2008-01-16 16:17:06 +010010007
Baptiste Assmann91bd3372015-07-17 21:59:42 +020010008 In order to work, "usesrc" requires root privileges.
10009
Willy Tarreaueabeafa2008-01-16 16:17:06 +010010010 Examples :
10011 backend private
10012 # Connect to the servers using our 192.168.1.200 source address
10013 source 192.168.1.200
10014
10015 backend transparent_ssl1
10016 # Connect to the SSL farm from the client's source address
10017 source 192.168.1.200 usesrc clientip
10018
10019 backend transparent_ssl2
10020 # Connect to the SSL farm from the client's source address and port
10021 # not recommended if IP conntrack is present on the local machine.
10022 source 192.168.1.200 usesrc client
10023
10024 backend transparent_ssl3
10025 # Connect to the SSL farm from the client's source address. It
10026 # is more conntrack-friendly.
10027 source 192.168.1.200 usesrc clientip
10028
10029 backend transparent_smtp
10030 # Connect to the SMTP farm from the client's source address/port
10031 # with Tproxy version 4.
10032 source 0.0.0.0 usesrc clientip
10033
Willy Tarreaubce70882009-09-07 11:51:47 +020010034 backend transparent_http
10035 # Connect to the servers using the client's IP as seen by previous
10036 # proxy.
10037 source 0.0.0.0 usesrc hdr_ip(x-forwarded-for,-1)
10038
Willy Tarreauc57f0e22009-05-10 13:12:33 +020010039 See also : the "source" server option in section 5, the Tproxy patches for
Willy Tarreaueabeafa2008-01-16 16:17:06 +010010040 the Linux kernel on www.balabit.com, the "bind" keyword.
10041
Willy Tarreau844e3c52008-01-11 16:28:18 +010010042
MIZUTA Takeshib24bc0d2020-07-09 11:13:20 +090010043srvtcpka-cnt <count>
10044 Sets the maximum number of keepalive probes TCP should send before dropping
10045 the connection on the server side.
10046 May be used in sections : defaults | frontend | listen | backend
10047 yes | no | yes | yes
10048 Arguments :
10049 <count> is the maximum number of keepalive probes.
10050
10051 This keyword corresponds to the socket option TCP_KEEPCNT. If this keyword
10052 is not specified, system-wide TCP parameter (tcp_keepalive_probes) is used.
Willy Tarreau52543212020-07-09 05:58:51 +020010053 The availability of this setting depends on the operating system. It is
10054 known to work on Linux.
MIZUTA Takeshib24bc0d2020-07-09 11:13:20 +090010055
10056 See also : "option srvtcpka", "srvtcpka-idle", "srvtcpka-intvl".
10057
10058
10059srvtcpka-idle <timeout>
10060 Sets the time the connection needs to remain idle before TCP starts sending
10061 keepalive probes, if enabled the sending of TCP keepalive packets on the
10062 server side.
10063 May be used in sections : defaults | frontend | listen | backend
10064 yes | no | yes | yes
10065 Arguments :
10066 <timeout> is the time the connection needs to remain idle before TCP starts
10067 sending keepalive probes. It is specified in seconds by default,
10068 but can be in any other unit if the number is suffixed by the
10069 unit, as explained at the top of this document.
10070
10071 This keyword corresponds to the socket option TCP_KEEPIDLE. If this keyword
10072 is not specified, system-wide TCP parameter (tcp_keepalive_time) is used.
Willy Tarreau52543212020-07-09 05:58:51 +020010073 The availability of this setting depends on the operating system. It is
10074 known to work on Linux.
MIZUTA Takeshib24bc0d2020-07-09 11:13:20 +090010075
10076 See also : "option srvtcpka", "srvtcpka-cnt", "srvtcpka-intvl".
10077
10078
10079srvtcpka-intvl <timeout>
10080 Sets the time between individual keepalive probes on the server side.
10081 May be used in sections : defaults | frontend | listen | backend
10082 yes | no | yes | yes
10083 Arguments :
10084 <timeout> is the time between individual keepalive probes. It is specified
10085 in seconds by default, but can be in any other unit if the number
10086 is suffixed by the unit, as explained at the top of this
10087 document.
10088
10089 This keyword corresponds to the socket option TCP_KEEPINTVL. If this keyword
10090 is not specified, system-wide TCP parameter (tcp_keepalive_intvl) is used.
Willy Tarreau52543212020-07-09 05:58:51 +020010091 The availability of this setting depends on the operating system. It is
10092 known to work on Linux.
MIZUTA Takeshib24bc0d2020-07-09 11:13:20 +090010093
10094 See also : "option srvtcpka", "srvtcpka-cnt", "srvtcpka-idle".
10095
10096
Cyril Bonté66c327d2010-10-12 00:14:37 +020010097stats admin { if | unless } <cond>
10098 Enable statistics admin level if/unless a condition is matched
10099 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaued2119c2014-04-24 22:10:39 +020010100 no | yes | yes | yes
Cyril Bonté66c327d2010-10-12 00:14:37 +020010101
10102 This statement enables the statistics admin level if/unless a condition is
10103 matched.
10104
10105 The admin level allows to enable/disable servers from the web interface. By
10106 default, statistics page is read-only for security reasons.
10107
Cyril Bonté02ff8ef2010-12-14 22:48:49 +010010108 Note : Consider not using this feature in multi-process mode (nbproc > 1)
10109 unless you know what you do : memory is not shared between the
Davor Ocelice9ed2812017-12-25 17:49:28 +010010110 processes, which can result in random behaviors.
Cyril Bonté02ff8ef2010-12-14 22:48:49 +010010111
Cyril Bonté23b39d92011-02-10 22:54:44 +010010112 Currently, the POST request is limited to the buffer size minus the reserved
10113 buffer space, which means that if the list of servers is too long, the
10114 request won't be processed. It is recommended to alter few servers at a
10115 time.
Cyril Bonté66c327d2010-10-12 00:14:37 +020010116
10117 Example :
10118 # statistics admin level only for localhost
10119 backend stats_localhost
10120 stats enable
10121 stats admin if LOCALHOST
10122
10123 Example :
10124 # statistics admin level always enabled because of the authentication
10125 backend stats_auth
10126 stats enable
10127 stats auth admin:AdMiN123
10128 stats admin if TRUE
10129
10130 Example :
10131 # statistics admin level depends on the authenticated user
10132 userlist stats-auth
10133 group admin users admin
10134 user admin insecure-password AdMiN123
10135 group readonly users haproxy
10136 user haproxy insecure-password haproxy
10137
10138 backend stats_auth
10139 stats enable
10140 acl AUTH http_auth(stats-auth)
10141 acl AUTH_ADMIN http_auth_group(stats-auth) admin
10142 stats http-request auth unless AUTH
10143 stats admin if AUTH_ADMIN
10144
Cyril Bonté02ff8ef2010-12-14 22:48:49 +010010145 See also : "stats enable", "stats auth", "stats http-request", "nbproc",
10146 "bind-process", section 3.4 about userlists and section 7 about
10147 ACL usage.
Cyril Bonté66c327d2010-10-12 00:14:37 +020010148
10149
Willy Tarreaueabeafa2008-01-16 16:17:06 +010010150stats auth <user>:<passwd>
10151 Enable statistics with authentication and grant access to an account
10152 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaued2119c2014-04-24 22:10:39 +020010153 yes | yes | yes | yes
Willy Tarreaueabeafa2008-01-16 16:17:06 +010010154 Arguments :
10155 <user> is a user name to grant access to
10156
10157 <passwd> is the cleartext password associated to this user
10158
10159 This statement enables statistics with default settings, and restricts access
10160 to declared users only. It may be repeated as many times as necessary to
10161 allow as many users as desired. When a user tries to access the statistics
10162 without a valid account, a "401 Forbidden" response will be returned so that
10163 the browser asks the user to provide a valid user and password. The real
10164 which will be returned to the browser is configurable using "stats realm".
10165
10166 Since the authentication method is HTTP Basic Authentication, the passwords
10167 circulate in cleartext on the network. Thus, it was decided that the
10168 configuration file would also use cleartext passwords to remind the users
Willy Tarreau3c92c5f2011-08-28 09:45:47 +020010169 that those ones should not be sensitive and not shared with any other account.
Willy Tarreaueabeafa2008-01-16 16:17:06 +010010170
10171 It is also possible to reduce the scope of the proxies which appear in the
10172 report using "stats scope".
10173
10174 Though this statement alone is enough to enable statistics reporting, it is
10175 recommended to set all other settings in order to avoid relying on default
10176 unobvious parameters.
10177
10178 Example :
10179 # public access (limited to this backend only)
10180 backend public_www
10181 server srv1 192.168.0.1:80
10182 stats enable
10183 stats hide-version
10184 stats scope .
10185 stats uri /admin?stats
Davor Ocelice9ed2812017-12-25 17:49:28 +010010186 stats realm HAProxy\ Statistics
Willy Tarreaueabeafa2008-01-16 16:17:06 +010010187 stats auth admin1:AdMiN123
10188 stats auth admin2:AdMiN321
10189
10190 # internal monitoring access (unlimited)
10191 backend private_monitoring
10192 stats enable
10193 stats uri /admin?stats
10194 stats refresh 5s
10195
10196 See also : "stats enable", "stats realm", "stats scope", "stats uri"
10197
10198
10199stats enable
10200 Enable statistics reporting with default settings
10201 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaued2119c2014-04-24 22:10:39 +020010202 yes | yes | yes | yes
Willy Tarreaueabeafa2008-01-16 16:17:06 +010010203 Arguments : none
10204
10205 This statement enables statistics reporting with default settings defined
10206 at build time. Unless stated otherwise, these settings are used :
10207 - stats uri : /haproxy?stats
10208 - stats realm : "HAProxy Statistics"
10209 - stats auth : no authentication
10210 - stats scope : no restriction
10211
10212 Though this statement alone is enough to enable statistics reporting, it is
10213 recommended to set all other settings in order to avoid relying on default
10214 unobvious parameters.
10215
10216 Example :
10217 # public access (limited to this backend only)
10218 backend public_www
10219 server srv1 192.168.0.1:80
10220 stats enable
10221 stats hide-version
10222 stats scope .
10223 stats uri /admin?stats
Davor Ocelice9ed2812017-12-25 17:49:28 +010010224 stats realm HAProxy\ Statistics
Willy Tarreaueabeafa2008-01-16 16:17:06 +010010225 stats auth admin1:AdMiN123
10226 stats auth admin2:AdMiN321
10227
10228 # internal monitoring access (unlimited)
10229 backend private_monitoring
10230 stats enable
10231 stats uri /admin?stats
10232 stats refresh 5s
10233
10234 See also : "stats auth", "stats realm", "stats uri"
10235
10236
Willy Tarreaud63335a2010-02-26 12:56:52 +010010237stats hide-version
10238 Enable statistics and hide HAProxy version reporting
Willy Tarreau1d45b7c2009-08-16 10:29:18 +020010239 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaued2119c2014-04-24 22:10:39 +020010240 yes | yes | yes | yes
Willy Tarreaud63335a2010-02-26 12:56:52 +010010241 Arguments : none
Willy Tarreau1d45b7c2009-08-16 10:29:18 +020010242
Willy Tarreaud63335a2010-02-26 12:56:52 +010010243 By default, the stats page reports some useful status information along with
10244 the statistics. Among them is HAProxy's version. However, it is generally
10245 considered dangerous to report precise version to anyone, as it can help them
10246 target known weaknesses with specific attacks. The "stats hide-version"
10247 statement removes the version from the statistics report. This is recommended
10248 for public sites or any site with a weak login/password.
Willy Tarreau1d45b7c2009-08-16 10:29:18 +020010249
Krzysztof Piotr Oledzki48cb2ae2009-10-02 22:51:14 +020010250 Though this statement alone is enough to enable statistics reporting, it is
10251 recommended to set all other settings in order to avoid relying on default
10252 unobvious parameters.
10253
Willy Tarreaud63335a2010-02-26 12:56:52 +010010254 Example :
10255 # public access (limited to this backend only)
10256 backend public_www
10257 server srv1 192.168.0.1:80
Krzysztof Piotr Oledzki48cb2ae2009-10-02 22:51:14 +020010258 stats enable
Willy Tarreaud63335a2010-02-26 12:56:52 +010010259 stats hide-version
10260 stats scope .
10261 stats uri /admin?stats
Davor Ocelice9ed2812017-12-25 17:49:28 +010010262 stats realm HAProxy\ Statistics
Willy Tarreaud63335a2010-02-26 12:56:52 +010010263 stats auth admin1:AdMiN123
10264 stats auth admin2:AdMiN321
Willy Tarreau1d45b7c2009-08-16 10:29:18 +020010265
Willy Tarreau1d45b7c2009-08-16 10:29:18 +020010266 # internal monitoring access (unlimited)
10267 backend private_monitoring
10268 stats enable
Willy Tarreaud63335a2010-02-26 12:56:52 +010010269 stats uri /admin?stats
10270 stats refresh 5s
Krzysztof Piotr Oledzki15514c22010-01-04 16:03:09 +010010271
Willy Tarreaud63335a2010-02-26 12:56:52 +010010272 See also : "stats auth", "stats enable", "stats realm", "stats uri"
Willy Tarreau1d45b7c2009-08-16 10:29:18 +020010273
Willy Tarreau983e01e2010-01-11 18:42:06 +010010274
Cyril Bonté2be1b3f2010-09-30 23:46:30 +020010275stats http-request { allow | deny | auth [realm <realm>] }
10276 [ { if | unless } <condition> ]
10277 Access control for statistics
10278
10279 May be used in sections: defaults | frontend | listen | backend
10280 no | no | yes | yes
10281
10282 As "http-request", these set of options allow to fine control access to
10283 statistics. Each option may be followed by if/unless and acl.
10284 First option with matched condition (or option without condition) is final.
10285 For "deny" a 403 error will be returned, for "allow" normal processing is
10286 performed, for "auth" a 401/407 error code is returned so the client
10287 should be asked to enter a username and password.
10288
10289 There is no fixed limit to the number of http-request statements per
10290 instance.
10291
10292 See also : "http-request", section 3.4 about userlists and section 7
10293 about ACL usage.
10294
10295
Willy Tarreaueabeafa2008-01-16 16:17:06 +010010296stats realm <realm>
10297 Enable statistics and set authentication realm
10298 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaued2119c2014-04-24 22:10:39 +020010299 yes | yes | yes | yes
Willy Tarreaueabeafa2008-01-16 16:17:06 +010010300 Arguments :
10301 <realm> is the name of the HTTP Basic Authentication realm reported to
10302 the browser. The browser uses it to display it in the pop-up
10303 inviting the user to enter a valid username and password.
10304
10305 The realm is read as a single word, so any spaces in it should be escaped
10306 using a backslash ('\').
10307
10308 This statement is useful only in conjunction with "stats auth" since it is
10309 only related to authentication.
10310
10311 Though this statement alone is enough to enable statistics reporting, it is
10312 recommended to set all other settings in order to avoid relying on default
10313 unobvious parameters.
10314
10315 Example :
10316 # public access (limited to this backend only)
10317 backend public_www
10318 server srv1 192.168.0.1:80
10319 stats enable
10320 stats hide-version
10321 stats scope .
10322 stats uri /admin?stats
Davor Ocelice9ed2812017-12-25 17:49:28 +010010323 stats realm HAProxy\ Statistics
Willy Tarreaueabeafa2008-01-16 16:17:06 +010010324 stats auth admin1:AdMiN123
10325 stats auth admin2:AdMiN321
10326
10327 # internal monitoring access (unlimited)
10328 backend private_monitoring
10329 stats enable
10330 stats uri /admin?stats
10331 stats refresh 5s
10332
10333 See also : "stats auth", "stats enable", "stats uri"
10334
10335
10336stats refresh <delay>
10337 Enable statistics with automatic refresh
10338 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaued2119c2014-04-24 22:10:39 +020010339 yes | yes | yes | yes
Willy Tarreaueabeafa2008-01-16 16:17:06 +010010340 Arguments :
10341 <delay> is the suggested refresh delay, specified in seconds, which will
10342 be returned to the browser consulting the report page. While the
10343 browser is free to apply any delay, it will generally respect it
10344 and refresh the page this every seconds. The refresh interval may
10345 be specified in any other non-default time unit, by suffixing the
10346 unit after the value, as explained at the top of this document.
10347
10348 This statement is useful on monitoring displays with a permanent page
10349 reporting the load balancer's activity. When set, the HTML report page will
10350 include a link "refresh"/"stop refresh" so that the user can select whether
Jackie Tapia749f74c2020-07-22 18:59:40 -050010351 they want automatic refresh of the page or not.
Willy Tarreaueabeafa2008-01-16 16:17:06 +010010352
10353 Though this statement alone is enough to enable statistics reporting, it is
10354 recommended to set all other settings in order to avoid relying on default
10355 unobvious parameters.
10356
10357 Example :
10358 # public access (limited to this backend only)
10359 backend public_www
10360 server srv1 192.168.0.1:80
10361 stats enable
10362 stats hide-version
10363 stats scope .
10364 stats uri /admin?stats
Davor Ocelice9ed2812017-12-25 17:49:28 +010010365 stats realm HAProxy\ Statistics
Willy Tarreaueabeafa2008-01-16 16:17:06 +010010366 stats auth admin1:AdMiN123
10367 stats auth admin2:AdMiN321
10368
10369 # internal monitoring access (unlimited)
10370 backend private_monitoring
10371 stats enable
10372 stats uri /admin?stats
10373 stats refresh 5s
10374
10375 See also : "stats auth", "stats enable", "stats realm", "stats uri"
10376
10377
10378stats scope { <name> | "." }
10379 Enable statistics and limit access scope
10380 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaued2119c2014-04-24 22:10:39 +020010381 yes | yes | yes | yes
Willy Tarreaueabeafa2008-01-16 16:17:06 +010010382 Arguments :
10383 <name> is the name of a listen, frontend or backend section to be
10384 reported. The special name "." (a single dot) designates the
10385 section in which the statement appears.
10386
10387 When this statement is specified, only the sections enumerated with this
10388 statement will appear in the report. All other ones will be hidden. This
10389 statement may appear as many times as needed if multiple sections need to be
10390 reported. Please note that the name checking is performed as simple string
10391 comparisons, and that it is never checked that a give section name really
10392 exists.
10393
10394 Though this statement alone is enough to enable statistics reporting, it is
10395 recommended to set all other settings in order to avoid relying on default
10396 unobvious parameters.
10397
10398 Example :
10399 # public access (limited to this backend only)
10400 backend public_www
10401 server srv1 192.168.0.1:80
10402 stats enable
10403 stats hide-version
10404 stats scope .
10405 stats uri /admin?stats
Davor Ocelice9ed2812017-12-25 17:49:28 +010010406 stats realm HAProxy\ Statistics
Willy Tarreaueabeafa2008-01-16 16:17:06 +010010407 stats auth admin1:AdMiN123
10408 stats auth admin2:AdMiN321
10409
10410 # internal monitoring access (unlimited)
10411 backend private_monitoring
10412 stats enable
10413 stats uri /admin?stats
10414 stats refresh 5s
10415
10416 See also : "stats auth", "stats enable", "stats realm", "stats uri"
10417
Willy Tarreaud63335a2010-02-26 12:56:52 +010010418
Willy Tarreauc9705a12010-07-27 20:05:50 +020010419stats show-desc [ <desc> ]
Willy Tarreaud63335a2010-02-26 12:56:52 +010010420 Enable reporting of a description on the statistics page.
10421 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaued2119c2014-04-24 22:10:39 +020010422 yes | yes | yes | yes
Willy Tarreaud63335a2010-02-26 12:56:52 +010010423
Willy Tarreauc9705a12010-07-27 20:05:50 +020010424 <desc> is an optional description to be reported. If unspecified, the
Willy Tarreaud63335a2010-02-26 12:56:52 +010010425 description from global section is automatically used instead.
10426
10427 This statement is useful for users that offer shared services to their
10428 customers, where node or description should be different for each customer.
10429
10430 Though this statement alone is enough to enable statistics reporting, it is
10431 recommended to set all other settings in order to avoid relying on default
Davor Ocelice9ed2812017-12-25 17:49:28 +010010432 unobvious parameters. By default description is not shown.
Willy Tarreaud63335a2010-02-26 12:56:52 +010010433
10434 Example :
10435 # internal monitoring access (unlimited)
10436 backend private_monitoring
10437 stats enable
10438 stats show-desc Master node for Europe, Asia, Africa
10439 stats uri /admin?stats
10440 stats refresh 5s
10441
10442 See also: "show-node", "stats enable", "stats uri" and "description" in
10443 global section.
10444
10445
10446stats show-legends
Willy Tarreaued2119c2014-04-24 22:10:39 +020010447 Enable reporting additional information on the statistics page
10448 May be used in sections : defaults | frontend | listen | backend
10449 yes | yes | yes | yes
10450 Arguments : none
10451
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030010452 Enable reporting additional information on the statistics page :
Willy Tarreaud63335a2010-02-26 12:56:52 +010010453 - cap: capabilities (proxy)
10454 - mode: one of tcp, http or health (proxy)
10455 - id: SNMP ID (proxy, socket, server)
10456 - IP (socket, server)
10457 - cookie (backend, server)
10458
10459 Though this statement alone is enough to enable statistics reporting, it is
10460 recommended to set all other settings in order to avoid relying on default
Davor Ocelice9ed2812017-12-25 17:49:28 +010010461 unobvious parameters. Default behavior is not to show this information.
Willy Tarreaud63335a2010-02-26 12:56:52 +010010462
10463 See also: "stats enable", "stats uri".
10464
10465
Amaury Denoyelle0b70a8a2020-10-05 11:49:45 +020010466stats show-modules
10467 Enable display of extra statistics module on the statistics page
10468 May be used in sections : defaults | frontend | listen | backend
10469 yes | yes | yes | yes
10470 Arguments : none
10471
10472 New columns are added at the end of the line containing the extra statistics
10473 values as a tooltip.
10474
10475 Though this statement alone is enough to enable statistics reporting, it is
10476 recommended to set all other settings in order to avoid relying on default
10477 unobvious parameters. Default behavior is not to show this information.
10478
10479 See also: "stats enable", "stats uri".
10480
10481
Willy Tarreaud63335a2010-02-26 12:56:52 +010010482stats show-node [ <name> ]
10483 Enable reporting of a host name on the statistics page.
10484 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaued2119c2014-04-24 22:10:39 +020010485 yes | yes | yes | yes
Willy Tarreaud63335a2010-02-26 12:56:52 +010010486 Arguments:
10487 <name> is an optional name to be reported. If unspecified, the
10488 node name from global section is automatically used instead.
10489
10490 This statement is useful for users that offer shared services to their
10491 customers, where node or description might be different on a stats page
Davor Ocelice9ed2812017-12-25 17:49:28 +010010492 provided for each customer. Default behavior is not to show host name.
Willy Tarreaud63335a2010-02-26 12:56:52 +010010493
10494 Though this statement alone is enough to enable statistics reporting, it is
10495 recommended to set all other settings in order to avoid relying on default
10496 unobvious parameters.
10497
10498 Example:
10499 # internal monitoring access (unlimited)
10500 backend private_monitoring
10501 stats enable
10502 stats show-node Europe-1
10503 stats uri /admin?stats
10504 stats refresh 5s
10505
10506 See also: "show-desc", "stats enable", "stats uri", and "node" in global
10507 section.
10508
Willy Tarreaueabeafa2008-01-16 16:17:06 +010010509
10510stats uri <prefix>
10511 Enable statistics and define the URI prefix to access them
10512 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaued2119c2014-04-24 22:10:39 +020010513 yes | yes | yes | yes
Willy Tarreaueabeafa2008-01-16 16:17:06 +010010514 Arguments :
10515 <prefix> is the prefix of any URI which will be redirected to stats. This
10516 prefix may contain a question mark ('?') to indicate part of a
10517 query string.
10518
10519 The statistics URI is intercepted on the relayed traffic, so it appears as a
10520 page within the normal application. It is strongly advised to ensure that the
10521 selected URI will never appear in the application, otherwise it will never be
10522 possible to reach it in the application.
10523
10524 The default URI compiled in haproxy is "/haproxy?stats", but this may be
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +010010525 changed at build time, so it's better to always explicitly specify it here.
Willy Tarreaueabeafa2008-01-16 16:17:06 +010010526 It is generally a good idea to include a question mark in the URI so that
10527 intermediate proxies refrain from caching the results. Also, since any string
10528 beginning with the prefix will be accepted as a stats request, the question
10529 mark helps ensuring that no valid URI will begin with the same words.
10530
10531 It is sometimes very convenient to use "/" as the URI prefix, and put that
10532 statement in a "listen" instance of its own. That makes it easy to dedicate
10533 an address or a port to statistics only.
10534
10535 Though this statement alone is enough to enable statistics reporting, it is
10536 recommended to set all other settings in order to avoid relying on default
10537 unobvious parameters.
10538
10539 Example :
10540 # public access (limited to this backend only)
10541 backend public_www
10542 server srv1 192.168.0.1:80
10543 stats enable
10544 stats hide-version
10545 stats scope .
10546 stats uri /admin?stats
Davor Ocelice9ed2812017-12-25 17:49:28 +010010547 stats realm HAProxy\ Statistics
Willy Tarreaueabeafa2008-01-16 16:17:06 +010010548 stats auth admin1:AdMiN123
10549 stats auth admin2:AdMiN321
10550
10551 # internal monitoring access (unlimited)
10552 backend private_monitoring
10553 stats enable
10554 stats uri /admin?stats
10555 stats refresh 5s
10556
10557 See also : "stats auth", "stats enable", "stats realm"
10558
10559
Willy Tarreaud63335a2010-02-26 12:56:52 +010010560stick match <pattern> [table <table>] [{if | unless} <cond>]
10561 Define a request pattern matching condition to stick a user to a server
Willy Tarreaueabeafa2008-01-16 16:17:06 +010010562 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaud63335a2010-02-26 12:56:52 +010010563 no | no | yes | yes
Willy Tarreaub937b7e2010-01-12 15:27:54 +010010564
10565 Arguments :
Willy Tarreaube722a22014-06-13 16:31:59 +020010566 <pattern> is a sample expression rule as described in section 7.3. It
Willy Tarreaub937b7e2010-01-12 15:27:54 +010010567 describes what elements of the incoming request or connection
Davor Ocelice9ed2812017-12-25 17:49:28 +010010568 will be analyzed in the hope to find a matching entry in a
Willy Tarreaub937b7e2010-01-12 15:27:54 +010010569 stickiness table. This rule is mandatory.
10570
10571 <table> is an optional stickiness table name. If unspecified, the same
10572 backend's table is used. A stickiness table is declared using
10573 the "stick-table" statement.
10574
10575 <cond> is an optional matching condition. It makes it possible to match
10576 on a certain criterion only when other conditions are met (or
10577 not met). For instance, it could be used to match on a source IP
10578 address except when a request passes through a known proxy, in
10579 which case we'd match on a header containing that IP address.
10580
10581 Some protocols or applications require complex stickiness rules and cannot
10582 always simply rely on cookies nor hashing. The "stick match" statement
10583 describes a rule to extract the stickiness criterion from an incoming request
10584 or connection. See section 7 for a complete list of possible patterns and
10585 transformation rules.
10586
10587 The table has to be declared using the "stick-table" statement. It must be of
10588 a type compatible with the pattern. By default it is the one which is present
10589 in the same backend. It is possible to share a table with other backends by
10590 referencing it using the "table" keyword. If another table is referenced,
10591 the server's ID inside the backends are used. By default, all server IDs
10592 start at 1 in each backend, so the server ordering is enough. But in case of
10593 doubt, it is highly recommended to force server IDs using their "id" setting.
10594
10595 It is possible to restrict the conditions where a "stick match" statement
10596 will apply, using "if" or "unless" followed by a condition. See section 7 for
10597 ACL based conditions.
10598
10599 There is no limit on the number of "stick match" statements. The first that
10600 applies and matches will cause the request to be directed to the same server
10601 as was used for the request which created the entry. That way, multiple
10602 matches can be used as fallbacks.
10603
10604 The stick rules are checked after the persistence cookies, so they will not
10605 affect stickiness if a cookie has already been used to select a server. That
10606 way, it becomes very easy to insert cookies and match on IP addresses in
10607 order to maintain stickiness between HTTP and HTTPS.
10608
Cyril Bonté02ff8ef2010-12-14 22:48:49 +010010609 Note : Consider not using this feature in multi-process mode (nbproc > 1)
10610 unless you know what you do : memory is not shared between the
Davor Ocelice9ed2812017-12-25 17:49:28 +010010611 processes, which can result in random behaviors.
Cyril Bonté02ff8ef2010-12-14 22:48:49 +010010612
Willy Tarreaub937b7e2010-01-12 15:27:54 +010010613 Example :
10614 # forward SMTP users to the same server they just used for POP in the
10615 # last 30 minutes
10616 backend pop
10617 mode tcp
10618 balance roundrobin
10619 stick store-request src
10620 stick-table type ip size 200k expire 30m
10621 server s1 192.168.1.1:110
10622 server s2 192.168.1.1:110
10623
10624 backend smtp
10625 mode tcp
10626 balance roundrobin
10627 stick match src table pop
10628 server s1 192.168.1.1:25
10629 server s2 192.168.1.1:25
10630
Cyril Bonté02ff8ef2010-12-14 22:48:49 +010010631 See also : "stick-table", "stick on", "nbproc", "bind-process" and section 7
Willy Tarreaube722a22014-06-13 16:31:59 +020010632 about ACLs and samples fetching.
Willy Tarreaub937b7e2010-01-12 15:27:54 +010010633
10634
10635stick on <pattern> [table <table>] [{if | unless} <condition>]
10636 Define a request pattern to associate a user to a server
10637 May be used in sections : defaults | frontend | listen | backend
10638 no | no | yes | yes
10639
10640 Note : This form is exactly equivalent to "stick match" followed by
10641 "stick store-request", all with the same arguments. Please refer
10642 to both keywords for details. It is only provided as a convenience
10643 for writing more maintainable configurations.
10644
Cyril Bonté02ff8ef2010-12-14 22:48:49 +010010645 Note : Consider not using this feature in multi-process mode (nbproc > 1)
10646 unless you know what you do : memory is not shared between the
Davor Ocelice9ed2812017-12-25 17:49:28 +010010647 processes, which can result in random behaviors.
Cyril Bonté02ff8ef2010-12-14 22:48:49 +010010648
Willy Tarreaub937b7e2010-01-12 15:27:54 +010010649 Examples :
10650 # The following form ...
Willy Tarreauec579d82010-02-26 19:15:04 +010010651 stick on src table pop if !localhost
Willy Tarreaub937b7e2010-01-12 15:27:54 +010010652
10653 # ...is strictly equivalent to this one :
10654 stick match src table pop if !localhost
10655 stick store-request src table pop if !localhost
10656
10657
10658 # Use cookie persistence for HTTP, and stick on source address for HTTPS as
10659 # well as HTTP without cookie. Share the same table between both accesses.
10660 backend http
10661 mode http
10662 balance roundrobin
10663 stick on src table https
10664 cookie SRV insert indirect nocache
10665 server s1 192.168.1.1:80 cookie s1
10666 server s2 192.168.1.1:80 cookie s2
10667
10668 backend https
10669 mode tcp
10670 balance roundrobin
10671 stick-table type ip size 200k expire 30m
10672 stick on src
10673 server s1 192.168.1.1:443
10674 server s2 192.168.1.1:443
10675
Cyril Bonté02ff8ef2010-12-14 22:48:49 +010010676 See also : "stick match", "stick store-request", "nbproc" and "bind-process".
Willy Tarreaub937b7e2010-01-12 15:27:54 +010010677
10678
10679stick store-request <pattern> [table <table>] [{if | unless} <condition>]
10680 Define a request pattern used to create an entry in a stickiness table
10681 May be used in sections : defaults | frontend | listen | backend
10682 no | no | yes | yes
10683
10684 Arguments :
Willy Tarreaube722a22014-06-13 16:31:59 +020010685 <pattern> is a sample expression rule as described in section 7.3. It
Willy Tarreaub937b7e2010-01-12 15:27:54 +010010686 describes what elements of the incoming request or connection
Davor Ocelice9ed2812017-12-25 17:49:28 +010010687 will be analyzed, extracted and stored in the table once a
Willy Tarreaub937b7e2010-01-12 15:27:54 +010010688 server is selected.
10689
10690 <table> is an optional stickiness table name. If unspecified, the same
10691 backend's table is used. A stickiness table is declared using
10692 the "stick-table" statement.
10693
10694 <cond> is an optional storage condition. It makes it possible to store
10695 certain criteria only when some conditions are met (or not met).
10696 For instance, it could be used to store the source IP address
10697 except when the request passes through a known proxy, in which
10698 case we'd store a converted form of a header containing that IP
10699 address.
10700
10701 Some protocols or applications require complex stickiness rules and cannot
10702 always simply rely on cookies nor hashing. The "stick store-request" statement
10703 describes a rule to decide what to extract from the request and when to do
10704 it, in order to store it into a stickiness table for further requests to
10705 match it using the "stick match" statement. Obviously the extracted part must
10706 make sense and have a chance to be matched in a further request. Storing a
10707 client's IP address for instance often makes sense. Storing an ID found in a
10708 URL parameter also makes sense. Storing a source port will almost never make
10709 any sense because it will be randomly matched. See section 7 for a complete
10710 list of possible patterns and transformation rules.
10711
10712 The table has to be declared using the "stick-table" statement. It must be of
10713 a type compatible with the pattern. By default it is the one which is present
10714 in the same backend. It is possible to share a table with other backends by
10715 referencing it using the "table" keyword. If another table is referenced,
10716 the server's ID inside the backends are used. By default, all server IDs
10717 start at 1 in each backend, so the server ordering is enough. But in case of
10718 doubt, it is highly recommended to force server IDs using their "id" setting.
10719
10720 It is possible to restrict the conditions where a "stick store-request"
10721 statement will apply, using "if" or "unless" followed by a condition. This
10722 condition will be evaluated while parsing the request, so any criteria can be
10723 used. See section 7 for ACL based conditions.
10724
10725 There is no limit on the number of "stick store-request" statements, but
10726 there is a limit of 8 simultaneous stores per request or response. This
10727 makes it possible to store up to 8 criteria, all extracted from either the
10728 request or the response, regardless of the number of rules. Only the 8 first
10729 ones which match will be kept. Using this, it is possible to feed multiple
10730 tables at once in the hope to increase the chance to recognize a user on
Willy Tarreau9667a802013-12-09 12:52:13 +010010731 another protocol or access method. Using multiple store-request rules with
10732 the same table is possible and may be used to find the best criterion to rely
10733 on, by arranging the rules by decreasing preference order. Only the first
10734 extracted criterion for a given table will be stored. All subsequent store-
10735 request rules referencing the same table will be skipped and their ACLs will
10736 not be evaluated.
Willy Tarreaub937b7e2010-01-12 15:27:54 +010010737
10738 The "store-request" rules are evaluated once the server connection has been
10739 established, so that the table will contain the real server that processed
10740 the request.
10741
Cyril Bonté02ff8ef2010-12-14 22:48:49 +010010742 Note : Consider not using this feature in multi-process mode (nbproc > 1)
10743 unless you know what you do : memory is not shared between the
Davor Ocelice9ed2812017-12-25 17:49:28 +010010744 processes, which can result in random behaviors.
Cyril Bonté02ff8ef2010-12-14 22:48:49 +010010745
Willy Tarreaub937b7e2010-01-12 15:27:54 +010010746 Example :
10747 # forward SMTP users to the same server they just used for POP in the
10748 # last 30 minutes
10749 backend pop
10750 mode tcp
10751 balance roundrobin
10752 stick store-request src
10753 stick-table type ip size 200k expire 30m
10754 server s1 192.168.1.1:110
10755 server s2 192.168.1.1:110
10756
10757 backend smtp
10758 mode tcp
10759 balance roundrobin
10760 stick match src table pop
10761 server s1 192.168.1.1:25
10762 server s2 192.168.1.1:25
10763
Cyril Bonté02ff8ef2010-12-14 22:48:49 +010010764 See also : "stick-table", "stick on", "nbproc", "bind-process" and section 7
Willy Tarreaube722a22014-06-13 16:31:59 +020010765 about ACLs and sample fetching.
Willy Tarreaub937b7e2010-01-12 15:27:54 +010010766
10767
Emeric Brun7c6b82e2010-09-24 16:34:28 +020010768stick-table type {ip | integer | string [len <length>] | binary [len <length>]}
Thayne McCombs92149f92020-11-20 01:28:26 -070010769 size <size> [expire <expire>] [nopurge] [peers <peersect>] [srvkey <srvkey>]
Emeric Brunf099e792010-09-27 12:05:28 +020010770 [store <data_type>]*
Godbach64cef792013-12-04 16:08:22 +080010771 Configure the stickiness table for the current section
Willy Tarreaub937b7e2010-01-12 15:27:54 +010010772 May be used in sections : defaults | frontend | listen | backend
Willy Tarreauc00cdc22010-06-06 16:48:26 +020010773 no | yes | yes | yes
Willy Tarreaub937b7e2010-01-12 15:27:54 +010010774
10775 Arguments :
10776 ip a table declared with "type ip" will only store IPv4 addresses.
10777 This form is very compact (about 50 bytes per entry) and allows
10778 very fast entry lookup and stores with almost no overhead. This
10779 is mainly used to store client source IP addresses.
10780
David du Colombier9a6d3c92011-03-17 10:40:24 +010010781 ipv6 a table declared with "type ipv6" will only store IPv6 addresses.
10782 This form is very compact (about 60 bytes per entry) and allows
10783 very fast entry lookup and stores with almost no overhead. This
10784 is mainly used to store client source IP addresses.
10785
Willy Tarreaub937b7e2010-01-12 15:27:54 +010010786 integer a table declared with "type integer" will store 32bit integers
10787 which can represent a client identifier found in a request for
10788 instance.
10789
10790 string a table declared with "type string" will store substrings of up
10791 to <len> characters. If the string provided by the pattern
10792 extractor is larger than <len>, it will be truncated before
10793 being stored. During matching, at most <len> characters will be
10794 compared between the string in the table and the extracted
10795 pattern. When not specified, the string is automatically limited
Emeric Brun7c6b82e2010-09-24 16:34:28 +020010796 to 32 characters.
10797
10798 binary a table declared with "type binary" will store binary blocks
10799 of <len> bytes. If the block provided by the pattern
10800 extractor is larger than <len>, it will be truncated before
Willy Tarreaube722a22014-06-13 16:31:59 +020010801 being stored. If the block provided by the sample expression
Emeric Brun7c6b82e2010-09-24 16:34:28 +020010802 is shorter than <len>, it will be padded by 0. When not
10803 specified, the block is automatically limited to 32 bytes.
Willy Tarreaub937b7e2010-01-12 15:27:54 +010010804
10805 <length> is the maximum number of characters that will be stored in a
Emeric Brun7c6b82e2010-09-24 16:34:28 +020010806 "string" type table (See type "string" above). Or the number
10807 of bytes of the block in "binary" type table. Be careful when
Willy Tarreaub937b7e2010-01-12 15:27:54 +010010808 changing this parameter as memory usage will proportionally
10809 increase.
10810
10811 <size> is the maximum number of entries that can fit in the table. This
Cyril Bonté78caf842010-03-10 22:41:43 +010010812 value directly impacts memory usage. Count approximately
10813 50 bytes per entry, plus the size of a string if any. The size
10814 supports suffixes "k", "m", "g" for 2^10, 2^20 and 2^30 factors.
Willy Tarreaub937b7e2010-01-12 15:27:54 +010010815
10816 [nopurge] indicates that we refuse to purge older entries when the table
10817 is full. When not specified and the table is full when haproxy
10818 wants to store an entry in it, it will flush a few of the oldest
10819 entries in order to release some space for the new ones. This is
Davor Ocelice9ed2812017-12-25 17:49:28 +010010820 most often the desired behavior. In some specific cases, it
Willy Tarreaub937b7e2010-01-12 15:27:54 +010010821 be desirable to refuse new entries instead of purging the older
10822 ones. That may be the case when the amount of data to store is
10823 far above the hardware limits and we prefer not to offer access
10824 to new clients than to reject the ones already connected. When
10825 using this parameter, be sure to properly set the "expire"
10826 parameter (see below).
10827
Emeric Brunf099e792010-09-27 12:05:28 +020010828 <peersect> is the name of the peers section to use for replication. Entries
10829 which associate keys to server IDs are kept synchronized with
10830 the remote peers declared in this section. All entries are also
10831 automatically learned from the local peer (old process) during a
10832 soft restart.
10833
Willy Tarreau1abc6732015-05-01 19:21:02 +020010834 NOTE : each peers section may be referenced only by tables
10835 belonging to the same unique process.
Cyril Bonté02ff8ef2010-12-14 22:48:49 +010010836
Willy Tarreaub937b7e2010-01-12 15:27:54 +010010837 <expire> defines the maximum duration of an entry in the table since it
10838 was last created, refreshed or matched. The expiration delay is
10839 defined using the standard time format, similarly as the various
10840 timeouts. The maximum duration is slightly above 24 days. See
Willy Tarreau4b103022021-02-12 17:59:10 +010010841 section 2.5 for more information. If this delay is not specified,
Cyril Bontédc4d9032012-04-08 21:57:39 +020010842 the session won't automatically expire, but older entries will
Willy Tarreaub937b7e2010-01-12 15:27:54 +010010843 be removed once full. Be sure not to use the "nopurge" parameter
10844 if not expiration delay is specified.
10845
Thayne McCombs92149f92020-11-20 01:28:26 -070010846 <srvkey> specifies how each server is identified for the purposes of the
10847 stick table. The valid values are "name" and "addr". If "name" is
10848 given, then <name> argument for the server (may be generated by
10849 a template). If "addr" is given, then the server is identified
10850 by its current network address, including the port. "addr" is
10851 especially useful if you are using service discovery to generate
10852 the addresses for servers with peered stick-tables and want
10853 to consistently use the same host across peers for a stickiness
10854 token.
10855
Willy Tarreau08d5f982010-06-06 13:34:54 +020010856 <data_type> is used to store additional information in the stick-table. This
10857 may be used by ACLs in order to control various criteria related
10858 to the activity of the client matching the stick-table. For each
10859 item specified here, the size of each entry will be inflated so
Willy Tarreauc9705a12010-07-27 20:05:50 +020010860 that the additional data can fit. Several data types may be
10861 stored with an entry. Multiple data types may be specified after
10862 the "store" keyword, as a comma-separated list. Alternatively,
10863 it is possible to repeat the "store" keyword followed by one or
10864 several data types. Except for the "server_id" type which is
10865 automatically detected and enabled, all data types must be
10866 explicitly declared to be stored. If an ACL references a data
10867 type which is not stored, the ACL will simply not match. Some
10868 data types require an argument which must be passed just after
10869 the type between parenthesis. See below for the supported data
10870 types and their arguments.
10871
10872 The data types that can be stored with an entry are the following :
10873 - server_id : this is an integer which holds the numeric ID of the server a
10874 request was assigned to. It is used by the "stick match", "stick store",
10875 and "stick on" rules. It is automatically enabled when referenced.
10876
10877 - gpc0 : first General Purpose Counter. It is a positive 32-bit integer
10878 integer which may be used for anything. Most of the time it will be used
10879 to put a special tag on some entries, for instance to note that a
Davor Ocelice9ed2812017-12-25 17:49:28 +010010880 specific behavior was detected and must be known for future matches.
Willy Tarreauc9705a12010-07-27 20:05:50 +020010881
Willy Tarreauba2ffd12013-05-29 15:54:14 +020010882 - gpc0_rate(<period>) : increment rate of the first General Purpose Counter
10883 over a period. It is a positive 32-bit integer integer which may be used
10884 for anything. Just like <gpc0>, it counts events, but instead of keeping
Davor Ocelice9ed2812017-12-25 17:49:28 +010010885 a cumulative number, it maintains the rate at which the counter is
Willy Tarreauba2ffd12013-05-29 15:54:14 +020010886 incremented. Most of the time it will be used to measure the frequency of
Davor Ocelice9ed2812017-12-25 17:49:28 +010010887 occurrence of certain events (e.g. requests to a specific URL).
Willy Tarreauba2ffd12013-05-29 15:54:14 +020010888
Frédéric Lécaille6778b272018-01-29 15:22:53 +010010889 - gpc1 : second General Purpose Counter. It is a positive 32-bit integer
10890 integer which may be used for anything. Most of the time it will be used
10891 to put a special tag on some entries, for instance to note that a
10892 specific behavior was detected and must be known for future matches.
10893
10894 - gpc1_rate(<period>) : increment rate of the second General Purpose Counter
10895 over a period. It is a positive 32-bit integer integer which may be used
10896 for anything. Just like <gpc1>, it counts events, but instead of keeping
10897 a cumulative number, it maintains the rate at which the counter is
10898 incremented. Most of the time it will be used to measure the frequency of
10899 occurrence of certain events (e.g. requests to a specific URL).
10900
Willy Tarreauc9705a12010-07-27 20:05:50 +020010901 - conn_cnt : Connection Count. It is a positive 32-bit integer which counts
10902 the absolute number of connections received from clients which matched
10903 this entry. It does not mean the connections were accepted, just that
10904 they were received.
10905
10906 - conn_cur : Current Connections. It is a positive 32-bit integer which
10907 stores the concurrent connection counts for the entry. It is incremented
10908 once an incoming connection matches the entry, and decremented once the
10909 connection leaves. That way it is possible to know at any time the exact
10910 number of concurrent connections for an entry.
10911
10912 - conn_rate(<period>) : frequency counter (takes 12 bytes). It takes an
10913 integer parameter <period> which indicates in milliseconds the length
10914 of the period over which the average is measured. It reports the average
10915 incoming connection rate over that period, in connections per period. The
10916 result is an integer which can be matched using ACLs.
10917
10918 - sess_cnt : Session Count. It is a positive 32-bit integer which counts
10919 the absolute number of sessions received from clients which matched this
10920 entry. A session is a connection that was accepted by the layer 4 rules.
10921
10922 - sess_rate(<period>) : frequency counter (takes 12 bytes). It takes an
10923 integer parameter <period> which indicates in milliseconds the length
10924 of the period over which the average is measured. It reports the average
10925 incoming session rate over that period, in sessions per period. The
10926 result is an integer which can be matched using ACLs.
10927
10928 - http_req_cnt : HTTP request Count. It is a positive 32-bit integer which
10929 counts the absolute number of HTTP requests received from clients which
10930 matched this entry. It does not matter whether they are valid requests or
10931 not. Note that this is different from sessions when keep-alive is used on
10932 the client side.
10933
10934 - http_req_rate(<period>) : frequency counter (takes 12 bytes). It takes an
10935 integer parameter <period> which indicates in milliseconds the length
10936 of the period over which the average is measured. It reports the average
10937 HTTP request rate over that period, in requests per period. The result is
10938 an integer which can be matched using ACLs. It does not matter whether
10939 they are valid requests or not. Note that this is different from sessions
10940 when keep-alive is used on the client side.
10941
10942 - http_err_cnt : HTTP Error Count. It is a positive 32-bit integer which
10943 counts the absolute number of HTTP requests errors induced by clients
10944 which matched this entry. Errors are counted on invalid and truncated
10945 requests, as well as on denied or tarpitted requests, and on failed
10946 authentications. If the server responds with 4xx, then the request is
10947 also counted as an error since it's an error triggered by the client
Davor Ocelice9ed2812017-12-25 17:49:28 +010010948 (e.g. vulnerability scan).
Willy Tarreauc9705a12010-07-27 20:05:50 +020010949
10950 - http_err_rate(<period>) : frequency counter (takes 12 bytes). It takes an
10951 integer parameter <period> which indicates in milliseconds the length
10952 of the period over which the average is measured. It reports the average
10953 HTTP request error rate over that period, in requests per period (see
10954 http_err_cnt above for what is accounted as an error). The result is an
10955 integer which can be matched using ACLs.
10956
Willy Tarreau826f3ab2021-02-10 12:07:15 +010010957 - http_fail_cnt : HTTP Failure Count. It is a positive 32-bit integer which
10958 counts the absolute number of HTTP response failures induced by servers
10959 which matched this entry. Errors are counted on invalid and truncated
10960 responses, as well as any 5xx response other than 501 or 505. It aims at
10961 being used combined with path or URI to detect service failures.
10962
10963 - http_fail_rate(<period>) : frequency counter (takes 12 bytes). It takes
10964 an integer parameter <period> which indicates in milliseconds the length
10965 of the period over which the average is measured. It reports the average
10966 HTTP response failure rate over that period, in requests per period (see
10967 http_fail_cnt above for what is accounted as a failure). The result is an
10968 integer which can be matched using ACLs.
10969
Willy Tarreauc9705a12010-07-27 20:05:50 +020010970 - bytes_in_cnt : client to server byte count. It is a positive 64-bit
Davor Ocelice9ed2812017-12-25 17:49:28 +010010971 integer which counts the cumulative number of bytes received from clients
Willy Tarreauc9705a12010-07-27 20:05:50 +020010972 which matched this entry. Headers are included in the count. This may be
10973 used to limit abuse of upload features on photo or video servers.
10974
10975 - bytes_in_rate(<period>) : frequency counter (takes 12 bytes). It takes an
10976 integer parameter <period> which indicates in milliseconds the length
10977 of the period over which the average is measured. It reports the average
10978 incoming bytes rate over that period, in bytes per period. It may be used
10979 to detect users which upload too much and too fast. Warning: with large
10980 uploads, it is possible that the amount of uploaded data will be counted
10981 once upon termination, thus causing spikes in the average transfer speed
10982 instead of having a smooth one. This may partially be smoothed with
10983 "option contstats" though this is not perfect yet. Use of byte_in_cnt is
10984 recommended for better fairness.
10985
10986 - bytes_out_cnt : server to client byte count. It is a positive 64-bit
Davor Ocelice9ed2812017-12-25 17:49:28 +010010987 integer which counts the cumulative number of bytes sent to clients which
Willy Tarreauc9705a12010-07-27 20:05:50 +020010988 matched this entry. Headers are included in the count. This may be used
10989 to limit abuse of bots sucking the whole site.
10990
10991 - bytes_out_rate(<period>) : frequency counter (takes 12 bytes). It takes
10992 an integer parameter <period> which indicates in milliseconds the length
10993 of the period over which the average is measured. It reports the average
10994 outgoing bytes rate over that period, in bytes per period. It may be used
10995 to detect users which download too much and too fast. Warning: with large
10996 transfers, it is possible that the amount of transferred data will be
10997 counted once upon termination, thus causing spikes in the average
10998 transfer speed instead of having a smooth one. This may partially be
10999 smoothed with "option contstats" though this is not perfect yet. Use of
11000 byte_out_cnt is recommended for better fairness.
Willy Tarreau08d5f982010-06-06 13:34:54 +020011001
Willy Tarreauc00cdc22010-06-06 16:48:26 +020011002 There is only one stick-table per proxy. At the moment of writing this doc,
11003 it does not seem useful to have multiple tables per proxy. If this happens
Willy Tarreaub937b7e2010-01-12 15:27:54 +010011004 to be required, simply create a dummy backend with a stick-table in it and
11005 reference it.
11006
11007 It is important to understand that stickiness based on learning information
11008 has some limitations, including the fact that all learned associations are
Baptiste Assmann123ff042016-03-06 23:29:28 +010011009 lost upon restart unless peers are properly configured to transfer such
11010 information upon restart (recommended). In general it can be good as a
11011 complement but not always as an exclusive stickiness.
Willy Tarreaub937b7e2010-01-12 15:27:54 +010011012
Willy Tarreauc9705a12010-07-27 20:05:50 +020011013 Last, memory requirements may be important when storing many data types.
11014 Indeed, storing all indicators above at once in each entry requires 116 bytes
11015 per entry, or 116 MB for a 1-million entries table. This is definitely not
11016 something that can be ignored.
11017
11018 Example:
11019 # Keep track of counters of up to 1 million IP addresses over 5 minutes
11020 # and store a general purpose counter and the average connection rate
11021 # computed over a sliding window of 30 seconds.
11022 stick-table type ip size 1m expire 5m store gpc0,conn_rate(30s)
11023
Willy Tarreau4b103022021-02-12 17:59:10 +010011024 See also : "stick match", "stick on", "stick store-request", section 2.5
David du Colombiera13d1b92011-03-17 10:40:22 +010011025 about time format and section 7 about ACLs.
Willy Tarreaub937b7e2010-01-12 15:27:54 +010011026
11027
Emeric Brun6a1cefa2010-09-24 18:15:17 +020011028stick store-response <pattern> [table <table>] [{if | unless} <condition>]
Baptiste Assmann2f2d2ec2016-03-06 23:27:24 +010011029 Define a response pattern used to create an entry in a stickiness table
Emeric Brun6a1cefa2010-09-24 18:15:17 +020011030 May be used in sections : defaults | frontend | listen | backend
11031 no | no | yes | yes
11032
11033 Arguments :
Willy Tarreaube722a22014-06-13 16:31:59 +020011034 <pattern> is a sample expression rule as described in section 7.3. It
Emeric Brun6a1cefa2010-09-24 18:15:17 +020011035 describes what elements of the response or connection will
Davor Ocelice9ed2812017-12-25 17:49:28 +010011036 be analyzed, extracted and stored in the table once a
Emeric Brun6a1cefa2010-09-24 18:15:17 +020011037 server is selected.
11038
11039 <table> is an optional stickiness table name. If unspecified, the same
11040 backend's table is used. A stickiness table is declared using
11041 the "stick-table" statement.
11042
11043 <cond> is an optional storage condition. It makes it possible to store
11044 certain criteria only when some conditions are met (or not met).
11045 For instance, it could be used to store the SSL session ID only
11046 when the response is a SSL server hello.
11047
11048 Some protocols or applications require complex stickiness rules and cannot
11049 always simply rely on cookies nor hashing. The "stick store-response"
11050 statement describes a rule to decide what to extract from the response and
11051 when to do it, in order to store it into a stickiness table for further
11052 requests to match it using the "stick match" statement. Obviously the
11053 extracted part must make sense and have a chance to be matched in a further
Cyril Bonté108cf6e2012-04-21 23:30:29 +020011054 request. Storing an ID found in a header of a response makes sense.
Emeric Brun6a1cefa2010-09-24 18:15:17 +020011055 See section 7 for a complete list of possible patterns and transformation
11056 rules.
11057
11058 The table has to be declared using the "stick-table" statement. It must be of
11059 a type compatible with the pattern. By default it is the one which is present
11060 in the same backend. It is possible to share a table with other backends by
11061 referencing it using the "table" keyword. If another table is referenced,
11062 the server's ID inside the backends are used. By default, all server IDs
11063 start at 1 in each backend, so the server ordering is enough. But in case of
11064 doubt, it is highly recommended to force server IDs using their "id" setting.
11065
11066 It is possible to restrict the conditions where a "stick store-response"
11067 statement will apply, using "if" or "unless" followed by a condition. This
11068 condition will be evaluated while parsing the response, so any criteria can
11069 be used. See section 7 for ACL based conditions.
11070
11071 There is no limit on the number of "stick store-response" statements, but
11072 there is a limit of 8 simultaneous stores per request or response. This
11073 makes it possible to store up to 8 criteria, all extracted from either the
11074 request or the response, regardless of the number of rules. Only the 8 first
11075 ones which match will be kept. Using this, it is possible to feed multiple
11076 tables at once in the hope to increase the chance to recognize a user on
Willy Tarreau9667a802013-12-09 12:52:13 +010011077 another protocol or access method. Using multiple store-response rules with
11078 the same table is possible and may be used to find the best criterion to rely
11079 on, by arranging the rules by decreasing preference order. Only the first
11080 extracted criterion for a given table will be stored. All subsequent store-
11081 response rules referencing the same table will be skipped and their ACLs will
11082 not be evaluated. However, even if a store-request rule references a table, a
11083 store-response rule may also use the same table. This means that each table
11084 may learn exactly one element from the request and one element from the
11085 response at once.
Emeric Brun6a1cefa2010-09-24 18:15:17 +020011086
11087 The table will contain the real server that processed the request.
11088
11089 Example :
11090 # Learn SSL session ID from both request and response and create affinity.
11091 backend https
11092 mode tcp
11093 balance roundrobin
Cyril Bontédc4d9032012-04-08 21:57:39 +020011094 # maximum SSL session ID length is 32 bytes.
Emeric Brun6a1cefa2010-09-24 18:15:17 +020011095 stick-table type binary len 32 size 30k expire 30m
Cyril Bonté108cf6e2012-04-21 23:30:29 +020011096
Emeric Brun6a1cefa2010-09-24 18:15:17 +020011097 acl clienthello req_ssl_hello_type 1
11098 acl serverhello rep_ssl_hello_type 2
11099
11100 # use tcp content accepts to detects ssl client and server hello.
11101 tcp-request inspect-delay 5s
11102 tcp-request content accept if clienthello
11103
11104 # no timeout on response inspect delay by default.
11105 tcp-response content accept if serverhello
Cyril Bonté108cf6e2012-04-21 23:30:29 +020011106
Emeric Brun6a1cefa2010-09-24 18:15:17 +020011107 # SSL session ID (SSLID) may be present on a client or server hello.
11108 # Its length is coded on 1 byte at offset 43 and its value starts
11109 # at offset 44.
11110
11111 # Match and learn on request if client hello.
11112 stick on payload_lv(43,1) if clienthello
11113
11114 # Learn on response if server hello.
11115 stick store-response payload_lv(43,1) if serverhello
Cyril Bontédc4d9032012-04-08 21:57:39 +020011116
Emeric Brun6a1cefa2010-09-24 18:15:17 +020011117 server s1 192.168.1.1:443
11118 server s2 192.168.1.1:443
11119
11120 See also : "stick-table", "stick on", and section 7 about ACLs and pattern
11121 extraction.
11122
11123
Christopher Faulet4f5c2e22020-04-23 15:22:33 +020011124tcp-check comment <string>
11125 Defines a comment for the following the tcp-check rule, reported in logs if
11126 it fails.
11127 May be used in sections : defaults | frontend | listen | backend
11128 yes | no | yes | yes
11129
Christopher Faulet4f5c2e22020-04-23 15:22:33 +020011130 Arguments :
11131 <string> is the comment message to add in logs if the following tcp-check
11132 rule fails.
11133
Christopher Fauletc52ea4d2020-04-23 15:43:35 +020011134 It only works for connect, send and expect rules. It is useful to make
11135 user-friendly error reporting.
11136
Christopher Faulet4f5c2e22020-04-23 15:22:33 +020011137 See also : "option tcp-check", "tcp-check connect", "tcp-check send" and
11138 "tcp-check expect".
11139
11140
Christopher Fauletc52ea4d2020-04-23 15:43:35 +020011141tcp-check connect [default] [port <expr>] [addr <ip>] [send-proxy] [via-socks4]
11142 [ssl] [sni <sni>] [alpn <alpn>] [linger]
Christopher Fauletedc6ed92020-04-23 16:27:59 +020011143 [proto <name>] [comment <msg>]
Willy Tarreau938c7fe2014-04-25 14:21:39 +020011144 Opens a new connection
11145 May be used in sections: defaults | frontend | listen | backend
Christopher Faulet404f9192020-04-09 23:13:54 +020011146 yes | no | yes | yes
Willy Tarreau938c7fe2014-04-25 14:21:39 +020011147
Christopher Fauletc52ea4d2020-04-23 15:43:35 +020011148 Arguments :
Christopher Faulet4f5c2e22020-04-23 15:22:33 +020011149 comment <msg> defines a message to report if the rule evaluation fails.
11150
Christopher Faulet4dce5922020-03-30 13:54:42 +020011151 default Use default options of the server line to do the health
Daniel Corbett67a82712020-07-06 23:01:19 -040011152 checks. The server options are used only if not redefined.
Christopher Faulet4dce5922020-03-30 13:54:42 +020011153
Christopher Fauletb7d30092020-03-30 15:19:03 +020011154 port <expr> if not set, check port or server port is used.
Christopher Faulet5c288742020-03-31 08:15:58 +020011155 It tells HAProxy where to open the connection to.
11156 <port> must be a valid TCP port source integer, from 1 to
Christopher Fauletb7d30092020-03-30 15:19:03 +020011157 65535 or an sample-fetch expression.
Christopher Faulet5c288742020-03-31 08:15:58 +020011158
11159 addr <ip> defines the IP address to do the health check.
Willy Tarreau938c7fe2014-04-25 14:21:39 +020011160
11161 send-proxy send a PROXY protocol string
11162
Christopher Faulet085426a2020-03-30 13:07:02 +020011163 via-socks4 enables outgoing health checks using upstream socks4 proxy.
11164
Willy Tarreau938c7fe2014-04-25 14:21:39 +020011165 ssl opens a ciphered connection
11166
Christopher Faulet79b31d42020-03-30 13:00:05 +020011167 sni <sni> specifies the SNI to use to do health checks over SSL.
11168
Christopher Faulet98572322020-03-30 13:16:44 +020011169 alpn <alpn> defines which protocols to advertise with ALPN. The protocol
11170 list consists in a comma-delimited list of protocol names,
11171 for instance: "http/1.1,http/1.0" (without quotes).
11172 If it is not set, the server ALPN is used.
11173
Christopher Fauletedc6ed92020-04-23 16:27:59 +020011174 proto <name> forces the multiplexer's protocol to use for this connection.
11175 It must be a TCP mux protocol and it must be usable on the
11176 backend side. The list of available protocols is reported in
11177 haproxy -vv.
11178
Christopher Faulet5c288742020-03-31 08:15:58 +020011179 linger cleanly close the connection instead of using a single RST.
Gaetan Rivetf8ba6772020-02-07 15:37:17 +010011180
Christopher Fauletc52ea4d2020-04-23 15:43:35 +020011181 When an application lies on more than a single TCP port or when HAProxy
11182 load-balance many services in a single backend, it makes sense to probe all
11183 the services individually before considering a server as operational.
11184
11185 When there are no TCP port configured on the server line neither server port
11186 directive, then the 'tcp-check connect port <port>' must be the first step
11187 of the sequence.
11188
11189 In a tcp-check ruleset a 'connect' is required, it is also mandatory to start
11190 the ruleset with a 'connect' rule. Purpose is to ensure admin know what they
11191 do.
11192
11193 When a connect must start the ruleset, if may still be preceded by set-var,
11194 unset-var or comment rules.
11195
11196 Examples :
Willy Tarreau938c7fe2014-04-25 14:21:39 +020011197 # check HTTP and HTTPs services on a server.
11198 # first open port 80 thanks to server line port directive, then
11199 # tcp-check opens port 443, ciphered and run a request on it:
11200 option tcp-check
11201 tcp-check connect
11202 tcp-check send GET\ /\ HTTP/1.0\r\n
11203 tcp-check send Host:\ haproxy.1wt.eu\r\n
11204 tcp-check send \r\n
11205 tcp-check expect rstring (2..|3..)
11206 tcp-check connect port 443 ssl
11207 tcp-check send GET\ /\ HTTP/1.0\r\n
11208 tcp-check send Host:\ haproxy.1wt.eu\r\n
11209 tcp-check send \r\n
11210 tcp-check expect rstring (2..|3..)
11211 server www 10.0.0.1 check port 80
11212
11213 # check both POP and IMAP from a single server:
11214 option tcp-check
Gaetan Rivetf8ba6772020-02-07 15:37:17 +010011215 tcp-check connect port 110 linger
Willy Tarreau938c7fe2014-04-25 14:21:39 +020011216 tcp-check expect string +OK\ POP3\ ready
11217 tcp-check connect port 143
11218 tcp-check expect string *\ OK\ IMAP4\ ready
11219 server mail 10.0.0.1 check
11220
11221 See also : "option tcp-check", "tcp-check send", "tcp-check expect"
11222
11223
Christopher Faulet4f5c2e22020-04-23 15:22:33 +020011224tcp-check expect [min-recv <int>] [comment <msg>]
Christopher Fauletec07e382020-04-07 14:56:26 +020011225 [ok-status <st>] [error-status <st>] [tout-status <st>]
Christopher Faulet98cc57c2020-04-01 20:52:31 +020011226 [on-success <fmt>] [on-error <fmt>] [status-code <expr>]
Christopher Fauletcf80f2f2020-04-01 11:04:52 +020011227 [!] <match> <pattern>
Davor Ocelice9ed2812017-12-25 17:49:28 +010011228 Specify data to be collected and analyzed during a generic health check
Willy Tarreau938c7fe2014-04-25 14:21:39 +020011229 May be used in sections: defaults | frontend | listen | backend
Christopher Faulet404f9192020-04-09 23:13:54 +020011230 yes | no | yes | yes
Willy Tarreau938c7fe2014-04-25 14:21:39 +020011231
11232 Arguments :
Christopher Faulet4f5c2e22020-04-23 15:22:33 +020011233 comment <msg> defines a message to report if the rule evaluation fails.
11234
Gaetan Rivet1afd8262020-02-07 15:37:17 +010011235 min-recv is optional and can define the minimum amount of data required to
11236 evaluate the current expect rule. If the number of received bytes
11237 is under this limit, the check will wait for more data. This
11238 option can be used to resolve some ambiguous matching rules or to
11239 avoid executing costly regex matches on content known to be still
11240 incomplete. If an exact string (string or binary) is used, the
11241 minimum between the string length and this parameter is used.
11242 This parameter is ignored if it is set to -1. If the expect rule
11243 does not match, the check will wait for more data. If set to 0,
11244 the evaluation result is always conclusive.
11245
Willy Tarreau938c7fe2014-04-25 14:21:39 +020011246 <match> is a keyword indicating how to look for a specific pattern in the
Gaetan Rivetefab6c62020-02-07 15:37:17 +010011247 response. The keyword may be one of "string", "rstring", "binary" or
11248 "rbinary".
Willy Tarreau938c7fe2014-04-25 14:21:39 +020011249 The keyword may be preceded by an exclamation mark ("!") to negate
11250 the match. Spaces are allowed between the exclamation mark and the
11251 keyword. See below for more details on the supported keywords.
11252
Christopher Fauletec07e382020-04-07 14:56:26 +020011253 ok-status <st> is optional and can be used to set the check status if
11254 the expect rule is successfully evaluated and if it is
11255 the last rule in the tcp-check ruleset. "L7OK", "L7OKC",
Christopher Fauletd888f0f2020-05-07 07:40:17 +020011256 "L6OK" and "L4OK" are supported :
11257 - L7OK : check passed on layer 7
Christopher Faulet83662b52020-11-20 17:47:47 +010011258 - L7OKC : check conditionally passed on layer 7, set
11259 server to NOLB state.
Christopher Fauletd888f0f2020-05-07 07:40:17 +020011260 - L6OK : check passed on layer 6
11261 - L4OK : check passed on layer 4
Christopher Fauletec07e382020-04-07 14:56:26 +020011262 By default "L7OK" is used.
11263
Christopher Fauletcf80f2f2020-04-01 11:04:52 +020011264 error-status <st> is optional and can be used to set the check status if
11265 an error occurred during the expect rule evaluation.
Christopher Faulet83662b52020-11-20 17:47:47 +010011266 "L7OKC", "L7RSP", "L7STS", "L6RSP" and "L4CON" are
11267 supported :
11268 - L7OKC : check conditionally passed on layer 7, set
11269 server to NOLB state.
Christopher Fauletd888f0f2020-05-07 07:40:17 +020011270 - L7RSP : layer 7 invalid response - protocol error
11271 - L7STS : layer 7 response error, for example HTTP 5xx
11272 - L6RSP : layer 6 invalid response - protocol error
11273 - L4CON : layer 1-4 connection problem
11274 By default "L7RSP" is used.
Christopher Fauletcf80f2f2020-04-01 11:04:52 +020011275
Christopher Fauletec07e382020-04-07 14:56:26 +020011276 tout-status <st> is optional and can be used to set the check status if
Christopher Fauletcf80f2f2020-04-01 11:04:52 +020011277 a timeout occurred during the expect rule evaluation.
Christopher Fauletd888f0f2020-05-07 07:40:17 +020011278 "L7TOUT", "L6TOUT", and "L4TOUT" are supported :
11279 - L7TOUT : layer 7 (HTTP/SMTP) timeout
11280 - L6TOUT : layer 6 (SSL) timeout
11281 - L4TOUT : layer 1-4 timeout
Christopher Fauletcf80f2f2020-04-01 11:04:52 +020011282 By default "L7TOUT" is used.
11283
Christopher Fauletbe52b4d2020-04-01 16:30:22 +020011284 on-success <fmt> is optional and can be used to customize the
11285 informational message reported in logs if the expect
11286 rule is successfully evaluated and if it is the last rule
11287 in the tcp-check ruleset. <fmt> is a log-format string.
11288
11289 on-error <fmt> is optional and can be used to customize the
11290 informational message reported in logs if an error
11291 occurred during the expect rule evaluation. <fmt> is a
11292 log-format string.
11293
Christopher Faulet98cc57c2020-04-01 20:52:31 +020011294 status-code <expr> is optional and can be used to set the check status code
11295 reported in logs, on success or on error. <expr> is a
11296 standard HAProxy expression formed by a sample-fetch
11297 followed by some converters.
11298
Willy Tarreau938c7fe2014-04-25 14:21:39 +020011299 <pattern> is the pattern to look for. It may be a string or a regular
11300 expression. If the pattern contains spaces, they must be escaped
11301 with the usual backslash ('\').
11302 If the match is set to binary, then the pattern must be passed as
Davor Ocelice9ed2812017-12-25 17:49:28 +010011303 a series of hexadecimal digits in an even number. Each sequence of
Willy Tarreau938c7fe2014-04-25 14:21:39 +020011304 two digits will represent a byte. The hexadecimal digits may be
11305 used upper or lower case.
11306
Willy Tarreau938c7fe2014-04-25 14:21:39 +020011307 The available matches are intentionally similar to their http-check cousins :
11308
11309 string <string> : test the exact string matches in the response buffer.
11310 A health check response will be considered valid if the
11311 response's buffer contains this exact string. If the
11312 "string" keyword is prefixed with "!", then the response
11313 will be considered invalid if the body contains this
11314 string. This can be used to look for a mandatory pattern
11315 in a protocol response, or to detect a failure when a
11316 specific error appears in a protocol banner.
11317
11318 rstring <regex> : test a regular expression on the response buffer.
11319 A health check response will be considered valid if the
11320 response's buffer matches this expression. If the
11321 "rstring" keyword is prefixed with "!", then the response
11322 will be considered invalid if the body matches the
11323 expression.
11324
Christopher Fauletaaab0832020-05-05 15:54:22 +020011325 string-lf <fmt> : test a log-format string match in the response's buffer.
11326 A health check response will be considered valid if the
11327 response's buffer contains the string resulting of the
11328 evaluation of <fmt>, which follows the log-format rules.
11329 If prefixed with "!", then the response will be
11330 considered invalid if the buffer contains the string.
11331
Willy Tarreau938c7fe2014-04-25 14:21:39 +020011332 binary <hexstring> : test the exact string in its hexadecimal form matches
11333 in the response buffer. A health check response will
11334 be considered valid if the response's buffer contains
11335 this exact hexadecimal string.
11336 Purpose is to match data on binary protocols.
11337
Gaetan Rivetefab6c62020-02-07 15:37:17 +010011338 rbinary <regex> : test a regular expression on the response buffer, like
11339 "rstring". However, the response buffer is transformed
11340 into its hexadecimal form, including NUL-bytes. This
11341 allows using all regex engines to match any binary
11342 content. The hexadecimal transformation takes twice the
11343 size of the original response. As such, the expected
11344 pattern should work on at-most half the response buffer
11345 size.
11346
Christopher Fauletaaab0832020-05-05 15:54:22 +020011347 binary-lf <hexfmt> : test a log-format string in its hexadecimal form
11348 match in the response's buffer. A health check response
11349 will be considered valid if the response's buffer
11350 contains the hexadecimal string resulting of the
11351 evaluation of <fmt>, which follows the log-format
11352 rules. If prefixed with "!", then the response will be
11353 considered invalid if the buffer contains the
11354 hexadecimal string. The hexadecimal string is converted
11355 in a binary string before matching the response's
11356 buffer.
11357
Willy Tarreau938c7fe2014-04-25 14:21:39 +020011358 It is important to note that the responses will be limited to a certain size
Christopher Fauletbb9fb8b2020-11-25 17:20:57 +010011359 defined by the global "tune.bufsize" option, which defaults to 16384 bytes.
Willy Tarreau938c7fe2014-04-25 14:21:39 +020011360 Thus, too large responses may not contain the mandatory pattern when using
11361 "string", "rstring" or binary. If a large response is absolutely required, it
11362 is possible to change the default max size by setting the global variable.
11363 However, it is worth keeping in mind that parsing very large responses can
11364 waste some CPU cycles, especially when regular expressions are used, and that
11365 it is always better to focus the checks on smaller resources. Also, in its
11366 current state, the check will not find any string nor regex past a null
11367 character in the response. Similarly it is not possible to request matching
11368 the null character.
11369
11370 Examples :
11371 # perform a POP check
11372 option tcp-check
11373 tcp-check expect string +OK\ POP3\ ready
11374
11375 # perform an IMAP check
11376 option tcp-check
11377 tcp-check expect string *\ OK\ IMAP4\ ready
11378
11379 # look for the redis master server
11380 option tcp-check
11381 tcp-check send PING\r\n
Baptiste Assmanna3322992015-08-04 10:12:18 +020011382 tcp-check expect string +PONG
Willy Tarreau938c7fe2014-04-25 14:21:39 +020011383 tcp-check send info\ replication\r\n
11384 tcp-check expect string role:master
11385 tcp-check send QUIT\r\n
11386 tcp-check expect string +OK
11387
11388
11389 See also : "option tcp-check", "tcp-check connect", "tcp-check send",
Christopher Fauletbb9fb8b2020-11-25 17:20:57 +010011390 "tcp-check send-binary", "http-check expect", tune.bufsize
Willy Tarreau938c7fe2014-04-25 14:21:39 +020011391
11392
Christopher Fauletb50b3e62020-05-05 18:43:43 +020011393tcp-check send <data> [comment <msg>]
11394tcp-check send-lf <fmt> [comment <msg>]
11395 Specify a string or a log-format string to be sent as a question during a
11396 generic health check
Willy Tarreau938c7fe2014-04-25 14:21:39 +020011397 May be used in sections: defaults | frontend | listen | backend
Christopher Faulet404f9192020-04-09 23:13:54 +020011398 yes | no | yes | yes
Willy Tarreau938c7fe2014-04-25 14:21:39 +020011399
Christopher Faulet4f5c2e22020-04-23 15:22:33 +020011400 Arguments :
11401 comment <msg> defines a message to report if the rule evaluation fails.
11402
Christopher Fauletb50b3e62020-05-05 18:43:43 +020011403 <data> is the string that will be sent during a generic health
11404 check session.
Christopher Faulet16fff672020-04-30 07:50:54 +020011405
Christopher Fauletb50b3e62020-05-05 18:43:43 +020011406 <fmt> is the log-format string that will be sent, once evaluated,
11407 during a generic health check session.
Willy Tarreau938c7fe2014-04-25 14:21:39 +020011408
11409 Examples :
11410 # look for the redis master server
11411 option tcp-check
11412 tcp-check send info\ replication\r\n
11413 tcp-check expect string role:master
11414
11415 See also : "option tcp-check", "tcp-check connect", "tcp-check expect",
Christopher Fauletbb9fb8b2020-11-25 17:20:57 +010011416 "tcp-check send-binary", tune.bufsize
Willy Tarreau938c7fe2014-04-25 14:21:39 +020011417
11418
Christopher Fauletb50b3e62020-05-05 18:43:43 +020011419tcp-check send-binary <hexstring> [comment <msg>]
11420tcp-check send-binary-lf <hexfmt> [comment <msg>]
11421 Specify an hex digits string or an hex digits log-format string to be sent as
11422 a binary question during a raw tcp health check
Willy Tarreau938c7fe2014-04-25 14:21:39 +020011423 May be used in sections: defaults | frontend | listen | backend
Christopher Faulet404f9192020-04-09 23:13:54 +020011424 yes | no | yes | yes
Willy Tarreau938c7fe2014-04-25 14:21:39 +020011425
Christopher Faulet4f5c2e22020-04-23 15:22:33 +020011426 Arguments :
11427 comment <msg> defines a message to report if the rule evaluation fails.
Christopher Fauletc52ea4d2020-04-23 15:43:35 +020011428
Christopher Fauletb50b3e62020-05-05 18:43:43 +020011429 <hexstring> is the hexadecimal string that will be send, once converted
11430 to binary, during a generic health check session.
Christopher Faulet16fff672020-04-30 07:50:54 +020011431
Christopher Fauletb50b3e62020-05-05 18:43:43 +020011432 <hexfmt> is the hexadecimal log-format string that will be send, once
11433 evaluated and converted to binary, during a generic health
11434 check session.
Willy Tarreau938c7fe2014-04-25 14:21:39 +020011435
11436 Examples :
11437 # redis check in binary
11438 option tcp-check
11439 tcp-check send-binary 50494e470d0a # PING\r\n
11440 tcp-check expect binary 2b504F4e47 # +PONG
11441
11442
11443 See also : "option tcp-check", "tcp-check connect", "tcp-check expect",
Christopher Fauletbb9fb8b2020-11-25 17:20:57 +010011444 "tcp-check send", tune.bufsize
Willy Tarreau938c7fe2014-04-25 14:21:39 +020011445
11446
Gaetan Rivet0c39ecc2020-02-24 17:34:11 +010011447tcp-check set-var(<var-name>) <expr>
Gaetan Rivet0c39ecc2020-02-24 17:34:11 +010011448 This operation sets the content of a variable. The variable is declared inline.
Gaetan Rivet0c39ecc2020-02-24 17:34:11 +010011449 May be used in sections: defaults | frontend | listen | backend
Christopher Faulet404f9192020-04-09 23:13:54 +020011450 yes | no | yes | yes
Gaetan Rivet0c39ecc2020-02-24 17:34:11 +010011451
Christopher Fauletc52ea4d2020-04-23 15:43:35 +020011452 Arguments :
Gaetan Rivet0c39ecc2020-02-24 17:34:11 +010011453 <var-name> The name of the variable starts with an indication about its
11454 scope. The scopes allowed for tcp-check are:
11455 "proc" : the variable is shared with the whole process.
11456 "sess" : the variable is shared with the tcp-check session.
11457 "check": the variable is declared for the lifetime of the tcp-check.
11458 This prefix is followed by a name. The separator is a '.'.
11459 The name may only contain characters 'a-z', 'A-Z', '0-9', '.',
11460 and '-'.
11461
11462 <expr> Is a sample-fetch expression potentially followed by converters.
11463
Christopher Fauletc52ea4d2020-04-23 15:43:35 +020011464 Examples :
Gaetan Rivet0c39ecc2020-02-24 17:34:11 +010011465 tcp-check set-var(check.port) int(1234)
11466
11467
11468tcp-check unset-var(<var-name>)
Gaetan Rivet0c39ecc2020-02-24 17:34:11 +010011469 Free a reference to a variable within its scope.
Gaetan Rivet0c39ecc2020-02-24 17:34:11 +010011470 May be used in sections: defaults | frontend | listen | backend
Christopher Faulet404f9192020-04-09 23:13:54 +020011471 yes | no | yes | yes
Gaetan Rivet0c39ecc2020-02-24 17:34:11 +010011472
Christopher Fauletc52ea4d2020-04-23 15:43:35 +020011473 Arguments :
Gaetan Rivet0c39ecc2020-02-24 17:34:11 +010011474 <var-name> The name of the variable starts with an indication about its
11475 scope. The scopes allowed for tcp-check are:
11476 "proc" : the variable is shared with the whole process.
11477 "sess" : the variable is shared with the tcp-check session.
11478 "check": the variable is declared for the lifetime of the tcp-check.
11479 This prefix is followed by a name. The separator is a '.'.
11480 The name may only contain characters 'a-z', 'A-Z', '0-9', '.',
11481 and '-'.
11482
Christopher Fauletc52ea4d2020-04-23 15:43:35 +020011483 Examples :
Gaetan Rivet0c39ecc2020-02-24 17:34:11 +010011484 tcp-check unset-var(check.port)
11485
11486
Willy Tarreaue9656522010-08-17 15:40:09 +020011487tcp-request connection <action> [{if | unless} <condition>]
11488 Perform an action on an incoming connection depending on a layer 4 condition
Willy Tarreau1a687942010-05-23 22:40:30 +020011489 May be used in sections : defaults | frontend | listen | backend
11490 no | yes | yes | no
Willy Tarreaue9656522010-08-17 15:40:09 +020011491 Arguments :
Willy Tarreauc870bfd2015-09-28 18:47:38 +020011492 <action> defines the action to perform if the condition applies. See
11493 below.
Willy Tarreau1a687942010-05-23 22:40:30 +020011494
Willy Tarreaue9656522010-08-17 15:40:09 +020011495 <condition> is a standard layer4-only ACL-based condition (see section 7).
Willy Tarreau68c03ab2010-08-06 15:08:45 +020011496
11497 Immediately after acceptance of a new incoming connection, it is possible to
11498 evaluate some conditions to decide whether this connection must be accepted
Willy Tarreaue9656522010-08-17 15:40:09 +020011499 or dropped or have its counters tracked. Those conditions cannot make use of
11500 any data contents because the connection has not been read from yet, and the
11501 buffers are not yet allocated. This is used to selectively and very quickly
11502 accept or drop connections from various sources with a very low overhead. If
11503 some contents need to be inspected in order to take the decision, the
11504 "tcp-request content" statements must be used instead.
Willy Tarreau68c03ab2010-08-06 15:08:45 +020011505
Willy Tarreaue9656522010-08-17 15:40:09 +020011506 The "tcp-request connection" rules are evaluated in their exact declaration
11507 order. If no rule matches or if there is no rule, the default action is to
11508 accept the incoming connection. There is no specific limit to the number of
11509 rules which may be inserted.
Willy Tarreau68c03ab2010-08-06 15:08:45 +020011510
Willy Tarreaua9083d02015-05-08 15:27:59 +020011511 Four types of actions are supported :
Willy Tarreaue9656522010-08-17 15:40:09 +020011512 - accept :
11513 accepts the connection if the condition is true (when used with "if")
11514 or false (when used with "unless"). The first such rule executed ends
11515 the rules evaluation.
Willy Tarreau68c03ab2010-08-06 15:08:45 +020011516
Willy Tarreaue9656522010-08-17 15:40:09 +020011517 - reject :
11518 rejects the connection if the condition is true (when used with "if")
11519 or false (when used with "unless"). The first such rule executed ends
11520 the rules evaluation. Rejected connections do not even become a
11521 session, which is why they are accounted separately for in the stats,
11522 as "denied connections". They are not considered for the session
11523 rate-limit and are not logged either. The reason is that these rules
11524 should only be used to filter extremely high connection rates such as
11525 the ones encountered during a massive DDoS attack. Under these extreme
11526 conditions, the simple action of logging each event would make the
11527 system collapse and would considerably lower the filtering capacity. If
11528 logging is absolutely desired, then "tcp-request content" rules should
Willy Tarreau4f614292016-10-21 17:49:36 +020011529 be used instead, as "tcp-request session" rules will not log either.
Willy Tarreau68c03ab2010-08-06 15:08:45 +020011530
Willy Tarreau4f0d9192013-06-11 20:40:55 +020011531 - expect-proxy layer4 :
11532 configures the client-facing connection to receive a PROXY protocol
11533 header before any byte is read from the socket. This is equivalent to
11534 having the "accept-proxy" keyword on the "bind" line, except that using
11535 the TCP rule allows the PROXY protocol to be accepted only for certain
11536 IP address ranges using an ACL. This is convenient when multiple layers
11537 of load balancers are passed through by traffic coming from public
11538 hosts.
11539
Bertrand Jacquin90759682016-06-06 15:35:39 +010011540 - expect-netscaler-cip layer4 :
11541 configures the client-facing connection to receive a NetScaler Client
11542 IP insertion protocol header before any byte is read from the socket.
11543 This is equivalent to having the "accept-netscaler-cip" keyword on the
11544 "bind" line, except that using the TCP rule allows the PROXY protocol
11545 to be accepted only for certain IP address ranges using an ACL. This
11546 is convenient when multiple layers of load balancers are passed
11547 through by traffic coming from public hosts.
11548
Willy Tarreau18bf01e2014-06-13 16:18:52 +020011549 - capture <sample> len <length> :
11550 This only applies to "tcp-request content" rules. It captures sample
11551 expression <sample> from the request buffer, and converts it to a
11552 string of at most <len> characters. The resulting string is stored into
11553 the next request "capture" slot, so it will possibly appear next to
11554 some captured HTTP headers. It will then automatically appear in the
11555 logs, and it will be possible to extract it using sample fetch rules to
11556 feed it into headers or anything. The length should be limited given
11557 that this size will be allocated for each capture during the whole
Willy Tarreaua9083d02015-05-08 15:27:59 +020011558 session life. Please check section 7.3 (Fetching samples) and "capture
11559 request header" for more information.
Willy Tarreau18bf01e2014-06-13 16:18:52 +020011560
Willy Tarreaube4a3ef2013-06-17 15:04:07 +020011561 - { track-sc0 | track-sc1 | track-sc2 } <key> [table <table>] :
Willy Tarreaue9656522010-08-17 15:40:09 +020011562 enables tracking of sticky counters from current connection. These
Moemen MHEDHBI9cf46342018-09-25 17:50:53 +020011563 rules do not stop evaluation and do not change default action. The
11564 number of counters that may be simultaneously tracked by the same
11565 connection is set in MAX_SESS_STKCTR at build time (reported in
John Roeslerfb2fce12019-07-10 15:45:51 -050011566 haproxy -vv) which defaults to 3, so the track-sc number is between 0
Matteo Contrini1857b8c2020-10-16 17:35:54 +020011567 and (MAX_SESS_STKCTR-1). The first "track-sc0" rule executed enables
Moemen MHEDHBI9cf46342018-09-25 17:50:53 +020011568 tracking of the counters of the specified table as the first set. The
11569 first "track-sc1" rule executed enables tracking of the counters of the
11570 specified table as the second set. The first "track-sc2" rule executed
11571 enables tracking of the counters of the specified table as the third
11572 set. It is a recommended practice to use the first set of counters for
11573 the per-frontend counters and the second set for the per-backend ones.
11574 But this is just a guideline, all may be used everywhere.
Willy Tarreau68c03ab2010-08-06 15:08:45 +020011575
Willy Tarreaue9656522010-08-17 15:40:09 +020011576 These actions take one or two arguments :
Willy Tarreaube722a22014-06-13 16:31:59 +020011577 <key> is mandatory, and is a sample expression rule as described
Willy Tarreau74ca5042013-06-11 23:12:07 +020011578 in section 7.3. It describes what elements of the incoming
Davor Ocelice9ed2812017-12-25 17:49:28 +010011579 request or connection will be analyzed, extracted, combined,
Willy Tarreau5d5b5d82012-12-09 12:00:04 +010011580 and used to select which table entry to update the counters.
11581 Note that "tcp-request connection" cannot use content-based
11582 fetches.
Willy Tarreau68c03ab2010-08-06 15:08:45 +020011583
Willy Tarreaue9656522010-08-17 15:40:09 +020011584 <table> is an optional table to be used instead of the default one,
11585 which is the stick-table declared in the current proxy. All
11586 the counters for the matches and updates for the key will
11587 then be performed in that table until the session ends.
Willy Tarreau68c03ab2010-08-06 15:08:45 +020011588
Willy Tarreaue9656522010-08-17 15:40:09 +020011589 Once a "track-sc*" rule is executed, the key is looked up in the table
11590 and if it is not found, an entry is allocated for it. Then a pointer to
11591 that entry is kept during all the session's life, and this entry's
11592 counters are updated as often as possible, every time the session's
11593 counters are updated, and also systematically when the session ends.
Willy Tarreau5d5b5d82012-12-09 12:00:04 +010011594 Counters are only updated for events that happen after the tracking has
11595 been started. For example, connection counters will not be updated when
11596 tracking layer 7 information, since the connection event happens before
11597 layer7 information is extracted.
11598
Willy Tarreaue9656522010-08-17 15:40:09 +020011599 If the entry tracks concurrent connection counters, one connection is
11600 counted for as long as the entry is tracked, and the entry will not
11601 expire during that time. Tracking counters also provides a performance
11602 advantage over just checking the keys, because only one table lookup is
11603 performed for all ACL checks that make use of it.
Willy Tarreau68c03ab2010-08-06 15:08:45 +020011604
Thierry FOURNIERe0627bd2015-08-04 08:20:33 +020011605 - sc-inc-gpc0(<sc-id>):
11606 The "sc-inc-gpc0" increments the GPC0 counter according to the sticky
11607 counter designated by <sc-id>. If an error occurs, this action silently
11608 fails and the actions evaluation continues.
11609
Frédéric Lécaille6778b272018-01-29 15:22:53 +010011610 - sc-inc-gpc1(<sc-id>):
11611 The "sc-inc-gpc1" increments the GPC1 counter according to the sticky
11612 counter designated by <sc-id>. If an error occurs, this action silently
11613 fails and the actions evaluation continues.
11614
Cédric Dufour0d7712d2019-11-06 18:38:53 +010011615 - sc-set-gpt0(<sc-id>) { <int> | <expr> }:
11616 This action sets the 32-bit unsigned GPT0 tag according to the sticky
11617 counter designated by <sc-id> and the value of <int>/<expr>. The
11618 expected result is a boolean. If an error occurs, this action silently
11619 fails and the actions evaluation continues.
Thierry FOURNIER236657b2015-08-19 08:25:14 +020011620
William Lallemand2e785f22016-05-25 01:48:42 +020011621 - set-src <expr> :
11622 Is used to set the source IP address to the value of specified
11623 expression. Useful if you want to mask source IP for privacy.
11624 If you want to provide an IP from a HTTP header use "http-request
Cyril Bonté6c81d5f2018-10-17 00:14:51 +020011625 set-src".
William Lallemand2e785f22016-05-25 01:48:42 +020011626
Cyril Bonté6c81d5f2018-10-17 00:14:51 +020011627 Arguments:
11628 <expr> Is a standard HAProxy expression formed by a sample-fetch
11629 followed by some converters.
William Lallemand2e785f22016-05-25 01:48:42 +020011630
11631 Example:
William Lallemand2e785f22016-05-25 01:48:42 +020011632 tcp-request connection set-src src,ipmask(24)
11633
Willy Tarreau0c630532016-10-21 17:52:58 +020011634 When possible, set-src preserves the original source port as long as the
11635 address family allows it, otherwise the source port is set to 0.
William Lallemand2e785f22016-05-25 01:48:42 +020011636
William Lallemand44be6402016-05-25 01:51:35 +020011637 - set-src-port <expr> :
11638 Is used to set the source port address to the value of specified
11639 expression.
11640
Cyril Bonté6c81d5f2018-10-17 00:14:51 +020011641 Arguments:
11642 <expr> Is a standard HAProxy expression formed by a sample-fetch
11643 followed by some converters.
William Lallemand44be6402016-05-25 01:51:35 +020011644
11645 Example:
William Lallemand44be6402016-05-25 01:51:35 +020011646 tcp-request connection set-src-port int(4000)
11647
Willy Tarreau0c630532016-10-21 17:52:58 +020011648 When possible, set-src-port preserves the original source address as long
11649 as the address family supports a port, otherwise it forces the source
11650 address to IPv4 "0.0.0.0" before rewriting the port.
William Lallemand44be6402016-05-25 01:51:35 +020011651
William Lallemand13e9b0c2016-05-25 02:34:07 +020011652 - set-dst <expr> :
11653 Is used to set the destination IP address to the value of specified
11654 expression. Useful if you want to mask IP for privacy in log.
11655 If you want to provide an IP from a HTTP header use "http-request
11656 set-dst". If you want to connect to the new address/port, use
11657 '0.0.0.0:0' as a server address in the backend.
11658
11659 <expr> Is a standard HAProxy expression formed by a sample-fetch
11660 followed by some converters.
11661
11662 Example:
11663
11664 tcp-request connection set-dst dst,ipmask(24)
11665 tcp-request connection set-dst ipv4(10.0.0.1)
11666
Willy Tarreau0c630532016-10-21 17:52:58 +020011667 When possible, set-dst preserves the original destination port as long as
11668 the address family allows it, otherwise the destination port is set to 0.
11669
William Lallemand13e9b0c2016-05-25 02:34:07 +020011670 - set-dst-port <expr> :
11671 Is used to set the destination port address to the value of specified
11672 expression. If you want to connect to the new address/port, use
11673 '0.0.0.0:0' as a server address in the backend.
11674
11675
11676 <expr> Is a standard HAProxy expression formed by a sample-fetch
11677 followed by some converters.
11678
11679 Example:
11680
11681 tcp-request connection set-dst-port int(4000)
11682
Willy Tarreau0c630532016-10-21 17:52:58 +020011683 When possible, set-dst-port preserves the original destination address as
11684 long as the address family supports a port, otherwise it forces the
11685 destination address to IPv4 "0.0.0.0" before rewriting the port.
11686
Willy Tarreau2d392c22015-08-24 01:43:45 +020011687 - "silent-drop" :
11688 This stops the evaluation of the rules and makes the client-facing
Davor Ocelice9ed2812017-12-25 17:49:28 +010011689 connection suddenly disappear using a system-dependent way that tries
Willy Tarreau2d392c22015-08-24 01:43:45 +020011690 to prevent the client from being notified. The effect it then that the
11691 client still sees an established connection while there's none on
11692 HAProxy. The purpose is to achieve a comparable effect to "tarpit"
11693 except that it doesn't use any local resource at all on the machine
11694 running HAProxy. It can resist much higher loads than "tarpit", and
Davor Ocelice9ed2812017-12-25 17:49:28 +010011695 slow down stronger attackers. It is important to understand the impact
11696 of using this mechanism. All stateful equipment placed between the
Willy Tarreau2d392c22015-08-24 01:43:45 +020011697 client and HAProxy (firewalls, proxies, load balancers) will also keep
11698 the established connection for a long time and may suffer from this
Davor Ocelice9ed2812017-12-25 17:49:28 +010011699 action. On modern Linux systems running with enough privileges, the
Willy Tarreau2d392c22015-08-24 01:43:45 +020011700 TCP_REPAIR socket option is used to block the emission of a TCP
11701 reset. On other systems, the socket's TTL is reduced to 1 so that the
11702 TCP reset doesn't pass the first router, though it's still delivered to
11703 local networks. Do not use it unless you fully understand how it works.
11704
Willy Tarreaue9656522010-08-17 15:40:09 +020011705 Note that the "if/unless" condition is optional. If no condition is set on
11706 the action, it is simply performed unconditionally. That can be useful for
11707 "track-sc*" actions as well as for changing the default action to a reject.
Willy Tarreau68c03ab2010-08-06 15:08:45 +020011708
Willy Tarreaue9656522010-08-17 15:40:09 +020011709 Example: accept all connections from white-listed hosts, reject too fast
11710 connection without counting them, and track accepted connections.
11711 This results in connection rate being capped from abusive sources.
Willy Tarreau68c03ab2010-08-06 15:08:45 +020011712
Willy Tarreaue9656522010-08-17 15:40:09 +020011713 tcp-request connection accept if { src -f /etc/haproxy/whitelist.lst }
Willy Tarreau68c03ab2010-08-06 15:08:45 +020011714 tcp-request connection reject if { src_conn_rate gt 10 }
Willy Tarreaube4a3ef2013-06-17 15:04:07 +020011715 tcp-request connection track-sc0 src
Willy Tarreau68c03ab2010-08-06 15:08:45 +020011716
Willy Tarreaue9656522010-08-17 15:40:09 +020011717 Example: accept all connections from white-listed hosts, count all other
11718 connections and reject too fast ones. This results in abusive ones
11719 being blocked as long as they don't slow down.
Willy Tarreau68c03ab2010-08-06 15:08:45 +020011720
Willy Tarreaue9656522010-08-17 15:40:09 +020011721 tcp-request connection accept if { src -f /etc/haproxy/whitelist.lst }
Willy Tarreaube4a3ef2013-06-17 15:04:07 +020011722 tcp-request connection track-sc0 src
11723 tcp-request connection reject if { sc0_conn_rate gt 10 }
Willy Tarreau68c03ab2010-08-06 15:08:45 +020011724
Willy Tarreau4f0d9192013-06-11 20:40:55 +020011725 Example: enable the PROXY protocol for traffic coming from all known proxies.
11726
11727 tcp-request connection expect-proxy layer4 if { src -f proxies.lst }
11728
Willy Tarreau68c03ab2010-08-06 15:08:45 +020011729 See section 7 about ACL usage.
11730
Willy Tarreau4f614292016-10-21 17:49:36 +020011731 See also : "tcp-request session", "tcp-request content", "stick-table"
Willy Tarreau68c03ab2010-08-06 15:08:45 +020011732
11733
Willy Tarreaue9656522010-08-17 15:40:09 +020011734tcp-request content <action> [{if | unless} <condition>]
11735 Perform an action on a new session depending on a layer 4-7 condition
Willy Tarreau62644772008-07-16 18:36:06 +020011736 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaufb356202010-08-03 14:02:05 +020011737 no | yes | yes | yes
Willy Tarreaue9656522010-08-17 15:40:09 +020011738 Arguments :
Willy Tarreauc870bfd2015-09-28 18:47:38 +020011739 <action> defines the action to perform if the condition applies. See
11740 below.
Willy Tarreau62644772008-07-16 18:36:06 +020011741
Willy Tarreaue9656522010-08-17 15:40:09 +020011742 <condition> is a standard layer 4-7 ACL-based condition (see section 7).
Willy Tarreau62644772008-07-16 18:36:06 +020011743
Davor Ocelice9ed2812017-12-25 17:49:28 +010011744 A request's contents can be analyzed at an early stage of request processing
Willy Tarreaue9656522010-08-17 15:40:09 +020011745 called "TCP content inspection". During this stage, ACL-based rules are
11746 evaluated every time the request contents are updated, until either an
11747 "accept" or a "reject" rule matches, or the TCP request inspection delay
11748 expires with no matching rule.
Willy Tarreau62644772008-07-16 18:36:06 +020011749
Willy Tarreaue9656522010-08-17 15:40:09 +020011750 The first difference between these rules and "tcp-request connection" rules
11751 is that "tcp-request content" rules can make use of contents to take a
11752 decision. Most often, these decisions will consider a protocol recognition or
11753 validity. The second difference is that content-based rules can be used in
Willy Tarreauf3338342014-01-28 21:40:28 +010011754 both frontends and backends. In case of HTTP keep-alive with the client, all
11755 tcp-request content rules are evaluated again, so haproxy keeps a record of
11756 what sticky counters were assigned by a "tcp-request connection" versus a
11757 "tcp-request content" rule, and flushes all the content-related ones after
11758 processing an HTTP request, so that they may be evaluated again by the rules
11759 being evaluated again for the next request. This is of particular importance
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030011760 when the rule tracks some L7 information or when it is conditioned by an
Willy Tarreauf3338342014-01-28 21:40:28 +010011761 L7-based ACL, since tracking may change between requests.
Willy Tarreau62644772008-07-16 18:36:06 +020011762
Willy Tarreaue9656522010-08-17 15:40:09 +020011763 Content-based rules are evaluated in their exact declaration order. If no
11764 rule matches or if there is no rule, the default action is to accept the
11765 contents. There is no specific limit to the number of rules which may be
11766 inserted.
Willy Tarreau62644772008-07-16 18:36:06 +020011767
Thierry FOURNIER236657b2015-08-19 08:25:14 +020011768 Several types of actions are supported :
Willy Tarreau18bf01e2014-06-13 16:18:52 +020011769 - accept : the request is accepted
Baptiste Assmann333939c2019-01-21 08:34:50 +010011770 - do-resolve: perform a DNS resolution
Willy Tarreau18bf01e2014-06-13 16:18:52 +020011771 - reject : the request is rejected and the connection is closed
11772 - capture : the specified sample expression is captured
Patrick Hemmer268a7072018-05-11 12:52:31 -040011773 - set-priority-class <expr> | set-priority-offset <expr>
Willy Tarreaube4a3ef2013-06-17 15:04:07 +020011774 - { track-sc0 | track-sc1 | track-sc2 } <key> [table <table>]
Thierry FOURNIERe0627bd2015-08-04 08:20:33 +020011775 - sc-inc-gpc0(<sc-id>)
Frédéric Lécaille6778b272018-01-29 15:22:53 +010011776 - sc-inc-gpc1(<sc-id>)
Cédric Dufour0d7712d2019-11-06 18:38:53 +010011777 - sc-set-gpt0(<sc-id>) { <int> | <expr> }
Baptiste Assmanne1afd4f2019-04-18 16:21:13 +020011778 - set-dst <expr>
11779 - set-dst-port <expr>
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020011780 - set-var(<var-name>) <expr>
Christopher Faulet85d79c92016-11-09 16:54:56 +010011781 - unset-var(<var-name>)
Willy Tarreau2d392c22015-08-24 01:43:45 +020011782 - silent-drop
Davor Ocelice9ed2812017-12-25 17:49:28 +010011783 - send-spoe-group <engine-name> <group-name>
Christopher Faulet579d83b2019-11-22 15:34:17 +010011784 - use-service <service-name>
Willy Tarreau62644772008-07-16 18:36:06 +020011785
Willy Tarreaue9656522010-08-17 15:40:09 +020011786 They have the same meaning as their counter-parts in "tcp-request connection"
11787 so please refer to that section for a complete description.
Baptiste Assmann333939c2019-01-21 08:34:50 +010011788 For "do-resolve" action, please check the "http-request do-resolve"
11789 configuration section.
Willy Tarreau62644772008-07-16 18:36:06 +020011790
Willy Tarreauf3338342014-01-28 21:40:28 +010011791 While there is nothing mandatory about it, it is recommended to use the
11792 track-sc0 in "tcp-request connection" rules, track-sc1 for "tcp-request
11793 content" rules in the frontend, and track-sc2 for "tcp-request content"
11794 rules in the backend, because that makes the configuration more readable
11795 and easier to troubleshoot, but this is just a guideline and all counters
11796 may be used everywhere.
Willy Tarreau62644772008-07-16 18:36:06 +020011797
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +010011798 Note that the "if/unless" condition is optional. If no condition is set on
Willy Tarreaue9656522010-08-17 15:40:09 +020011799 the action, it is simply performed unconditionally. That can be useful for
11800 "track-sc*" actions as well as for changing the default action to a reject.
Willy Tarreau62644772008-07-16 18:36:06 +020011801
Christopher Faulet2079a4a2020-10-02 11:48:57 +020011802 Note also that it is recommended to use a "tcp-request session" rule to track
11803 information that does *not* depend on Layer 7 contents, especially for HTTP
11804 frontends. Some HTTP processing are performed at the session level and may
11805 lead to an early rejection of the requests. Thus, the tracking at the content
11806 level may be disturbed in such case. A warning is emitted during startup to
11807 prevent, as far as possible, such unreliable usage.
11808
Willy Tarreaue9656522010-08-17 15:40:09 +020011809 It is perfectly possible to match layer 7 contents with "tcp-request content"
Christopher Faulet7ea509e2020-10-02 11:38:46 +020011810 rules from a TCP proxy, since HTTP-specific ACL matches are able to
11811 preliminarily parse the contents of a buffer before extracting the required
11812 data. If the buffered contents do not parse as a valid HTTP message, then the
11813 ACL does not match. The parser which is involved there is exactly the same
11814 as for all other HTTP processing, so there is no risk of parsing something
11815 differently. In an HTTP frontend or an HTTP backend, it is guaranteed that
11816 HTTP contents will always be immediately present when the rule is evaluated
11817 first because the HTTP parsing is performed in the early stages of the
11818 connection processing, at the session level. But for such proxies, using
11819 "http-request" rules is much more natural and recommended.
Willy Tarreau62644772008-07-16 18:36:06 +020011820
Willy Tarreau5d5b5d82012-12-09 12:00:04 +010011821 Tracking layer7 information is also possible provided that the information
Willy Tarreau4d54c7c2014-09-16 15:48:15 +020011822 are present when the rule is processed. The rule processing engine is able to
11823 wait until the inspect delay expires when the data to be tracked is not yet
11824 available.
Willy Tarreau5d5b5d82012-12-09 12:00:04 +010011825
Baptiste Assmanne1afd4f2019-04-18 16:21:13 +020011826 The "set-dst" and "set-dst-port" are used to set respectively the destination
11827 IP and port. More information on how to use it at "http-request set-dst".
11828
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020011829 The "set-var" is used to set the content of a variable. The variable is
Willy Tarreau4f614292016-10-21 17:49:36 +020011830 declared inline. For "tcp-request session" rules, only session-level
11831 variables can be used, without any layer7 contents.
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020011832
Daniel Schneller0b547052016-03-21 20:46:57 +010011833 <var-name> The name of the variable starts with an indication about
11834 its scope. The scopes allowed are:
Christopher Fauletff2613e2016-11-09 11:36:17 +010011835 "proc" : the variable is shared with the whole process
Daniel Schneller0b547052016-03-21 20:46:57 +010011836 "sess" : the variable is shared with the whole session
11837 "txn" : the variable is shared with the transaction
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020011838 (request and response)
Daniel Schneller0b547052016-03-21 20:46:57 +010011839 "req" : the variable is shared only during request
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020011840 processing
Daniel Schneller0b547052016-03-21 20:46:57 +010011841 "res" : the variable is shared only during response
11842 processing
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020011843 This prefix is followed by a name. The separator is a '.'.
Christopher Fauletb71557a2016-10-31 10:49:03 +010011844 The name may only contain characters 'a-z', 'A-Z', '0-9',
11845 '.' and '_'.
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020011846
11847 <expr> Is a standard HAProxy expression formed by a sample-fetch
11848 followed by some converters.
11849
Christopher Faulet85d79c92016-11-09 16:54:56 +010011850 The "unset-var" is used to unset a variable. See above for details about
11851 <var-name>.
11852
Patrick Hemmer268a7072018-05-11 12:52:31 -040011853 The "set-priority-class" is used to set the queue priority class of the
11854 current request. The value must be a sample expression which converts to an
11855 integer in the range -2047..2047. Results outside this range will be
11856 truncated. The priority class determines the order in which queued requests
11857 are processed. Lower values have higher priority.
11858
11859 The "set-priority-offset" is used to set the queue priority timestamp offset
11860 of the current request. The value must be a sample expression which converts
11861 to an integer in the range -524287..524287. Results outside this range will be
11862 truncated. When a request is queued, it is ordered first by the priority
11863 class, then by the current timestamp adjusted by the given offset in
11864 milliseconds. Lower values have higher priority.
11865 Note that the resulting timestamp is is only tracked with enough precision for
11866 524,287ms (8m44s287ms). If the request is queued long enough to where the
11867 adjusted timestamp exceeds this value, it will be misidentified as highest
11868 priority. Thus it is important to set "timeout queue" to a value, where when
11869 combined with the offset, does not exceed this limit.
11870
Christopher Faulet76c09ef2017-09-21 11:03:52 +020011871 The "send-spoe-group" is used to trigger sending of a group of SPOE
11872 messages. To do so, the SPOE engine used to send messages must be defined, as
11873 well as the SPOE group to send. Of course, the SPOE engine must refer to an
11874 existing SPOE filter. If not engine name is provided on the SPOE filter line,
11875 the SPOE agent name must be used.
11876
11877 <engine-name> The SPOE engine name.
11878
11879 <group-name> The SPOE group name as specified in the engine configuration.
11880
Christopher Faulet579d83b2019-11-22 15:34:17 +010011881 The "use-service" is used to executes a TCP service which will reply to the
11882 request and stop the evaluation of the rules. This service may choose to
11883 reply by sending any valid response or it may immediately close the
11884 connection without sending anything. Outside natives services, it is possible
11885 to write your own services in Lua. No further "tcp-request" rules are
11886 evaluated.
11887
11888 Example:
11889 tcp-request content use-service lua.deny { src -f /etc/haproxy/blacklist.lst }
11890
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020011891 Example:
11892
11893 tcp-request content set-var(sess.my_var) src
Christopher Faulet85d79c92016-11-09 16:54:56 +010011894 tcp-request content unset-var(sess.my_var2)
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020011895
Willy Tarreau62644772008-07-16 18:36:06 +020011896 Example:
Willy Tarreaue9656522010-08-17 15:40:09 +020011897 # Accept HTTP requests containing a Host header saying "example.com"
11898 # and reject everything else.
11899 acl is_host_com hdr(Host) -i example.com
11900 tcp-request inspect-delay 30s
Willy Tarreauc0239e02012-04-16 14:42:55 +020011901 tcp-request content accept if is_host_com
Willy Tarreaue9656522010-08-17 15:40:09 +020011902 tcp-request content reject
11903
11904 Example:
Willy Tarreau62644772008-07-16 18:36:06 +020011905 # reject SMTP connection if client speaks first
11906 tcp-request inspect-delay 30s
11907 acl content_present req_len gt 0
Willy Tarreau68c03ab2010-08-06 15:08:45 +020011908 tcp-request content reject if content_present
Willy Tarreau62644772008-07-16 18:36:06 +020011909
11910 # Forward HTTPS connection only if client speaks
11911 tcp-request inspect-delay 30s
11912 acl content_present req_len gt 0
Willy Tarreau68c03ab2010-08-06 15:08:45 +020011913 tcp-request content accept if content_present
Willy Tarreaue9656522010-08-17 15:40:09 +020011914 tcp-request content reject
11915
Willy Tarreau5d5b5d82012-12-09 12:00:04 +010011916 Example:
Jarno Huuskonene5ae7022017-04-03 14:36:21 +030011917 # Track the last IP(stick-table type string) from X-Forwarded-For
Willy Tarreau5d5b5d82012-12-09 12:00:04 +010011918 tcp-request inspect-delay 10s
Willy Tarreau4d54c7c2014-09-16 15:48:15 +020011919 tcp-request content track-sc0 hdr(x-forwarded-for,-1)
Jarno Huuskonene5ae7022017-04-03 14:36:21 +030011920 # Or track the last IP(stick-table type ip|ipv6) from X-Forwarded-For
11921 tcp-request content track-sc0 req.hdr_ip(x-forwarded-for,-1)
Willy Tarreau5d5b5d82012-12-09 12:00:04 +010011922
11923 Example:
11924 # track request counts per "base" (concatenation of Host+URL)
11925 tcp-request inspect-delay 10s
Willy Tarreau4d54c7c2014-09-16 15:48:15 +020011926 tcp-request content track-sc0 base table req-rate
Willy Tarreau5d5b5d82012-12-09 12:00:04 +010011927
Willy Tarreaue9656522010-08-17 15:40:09 +020011928 Example: track per-frontend and per-backend counters, block abusers at the
Jarno Huuskonene5ae7022017-04-03 14:36:21 +030011929 frontend when the backend detects abuse(and marks gpc0).
Willy Tarreaue9656522010-08-17 15:40:09 +020011930
11931 frontend http
Davor Ocelice9ed2812017-12-25 17:49:28 +010011932 # Use General Purpose Counter 0 in SC0 as a global abuse counter
Willy Tarreaue9656522010-08-17 15:40:09 +020011933 # protecting all our sites
11934 stick-table type ip size 1m expire 5m store gpc0
Willy Tarreaube4a3ef2013-06-17 15:04:07 +020011935 tcp-request connection track-sc0 src
11936 tcp-request connection reject if { sc0_get_gpc0 gt 0 }
Willy Tarreaue9656522010-08-17 15:40:09 +020011937 ...
11938 use_backend http_dynamic if { path_end .php }
11939
11940 backend http_dynamic
11941 # if a source makes too fast requests to this dynamic site (tracked
Willy Tarreaube4a3ef2013-06-17 15:04:07 +020011942 # by SC1), block it globally in the frontend.
Willy Tarreaue9656522010-08-17 15:40:09 +020011943 stick-table type ip size 1m expire 5m store http_req_rate(10s)
Willy Tarreaube4a3ef2013-06-17 15:04:07 +020011944 acl click_too_fast sc1_http_req_rate gt 10
Jarno Huuskonene5ae7022017-04-03 14:36:21 +030011945 acl mark_as_abuser sc0_inc_gpc0(http) gt 0
Willy Tarreaube4a3ef2013-06-17 15:04:07 +020011946 tcp-request content track-sc1 src
Willy Tarreaue9656522010-08-17 15:40:09 +020011947 tcp-request content reject if click_too_fast mark_as_abuser
Willy Tarreau62644772008-07-16 18:36:06 +020011948
Willy Tarreauc57f0e22009-05-10 13:12:33 +020011949 See section 7 about ACL usage.
Willy Tarreau62644772008-07-16 18:36:06 +020011950
Jarno Huuskonen95b012b2017-04-06 13:59:14 +030011951 See also : "tcp-request connection", "tcp-request session",
11952 "tcp-request inspect-delay", and "http-request".
Willy Tarreau62644772008-07-16 18:36:06 +020011953
11954
11955tcp-request inspect-delay <timeout>
11956 Set the maximum allowed time to wait for data during content inspection
11957 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaufb356202010-08-03 14:02:05 +020011958 no | yes | yes | yes
Willy Tarreau62644772008-07-16 18:36:06 +020011959 Arguments :
11960 <timeout> is the timeout value specified in milliseconds by default, but
11961 can be in any other unit if the number is suffixed by the unit,
11962 as explained at the top of this document.
11963
11964 People using haproxy primarily as a TCP relay are often worried about the
11965 risk of passing any type of protocol to a server without any analysis. In
11966 order to be able to analyze the request contents, we must first withhold
11967 the data then analyze them. This statement simply enables withholding of
11968 data for at most the specified amount of time.
11969
Willy Tarreaufb356202010-08-03 14:02:05 +020011970 TCP content inspection applies very early when a connection reaches a
11971 frontend, then very early when the connection is forwarded to a backend. This
11972 means that a connection may experience a first delay in the frontend and a
11973 second delay in the backend if both have tcp-request rules.
11974
Willy Tarreau62644772008-07-16 18:36:06 +020011975 Note that when performing content inspection, haproxy will evaluate the whole
11976 rules for every new chunk which gets in, taking into account the fact that
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +010011977 those data are partial. If no rule matches before the aforementioned delay,
Willy Tarreau62644772008-07-16 18:36:06 +020011978 a last check is performed upon expiration, this time considering that the
Willy Tarreaud869b242009-03-15 14:43:58 +010011979 contents are definitive. If no delay is set, haproxy will not wait at all
11980 and will immediately apply a verdict based on the available information.
11981 Obviously this is unlikely to be very useful and might even be racy, so such
11982 setups are not recommended.
Willy Tarreau62644772008-07-16 18:36:06 +020011983
11984 As soon as a rule matches, the request is released and continues as usual. If
11985 the timeout is reached and no rule matches, the default policy will be to let
11986 it pass through unaffected.
11987
11988 For most protocols, it is enough to set it to a few seconds, as most clients
11989 send the full request immediately upon connection. Add 3 or more seconds to
11990 cover TCP retransmits but that's all. For some protocols, it may make sense
Willy Tarreaud72758d2010-01-12 10:42:19 +010011991 to use large values, for instance to ensure that the client never talks
Davor Ocelice9ed2812017-12-25 17:49:28 +010011992 before the server (e.g. SMTP), or to wait for a client to talk before passing
11993 data to the server (e.g. SSL). Note that the client timeout must cover at
Willy Tarreaub824b002010-09-29 16:36:16 +020011994 least the inspection delay, otherwise it will expire first. If the client
11995 closes the connection or if the buffer is full, the delay immediately expires
11996 since the contents will not be able to change anymore.
Willy Tarreau62644772008-07-16 18:36:06 +020011997
Willy Tarreau55165fe2009-05-10 12:02:55 +020011998 See also : "tcp-request content accept", "tcp-request content reject",
Willy Tarreau62644772008-07-16 18:36:06 +020011999 "timeout client".
12000
12001
Emeric Brun0a3b67f2010-09-24 15:34:53 +020012002tcp-response content <action> [{if | unless} <condition>]
12003 Perform an action on a session response depending on a layer 4-7 condition
12004 May be used in sections : defaults | frontend | listen | backend
12005 no | no | yes | yes
12006 Arguments :
Willy Tarreauc870bfd2015-09-28 18:47:38 +020012007 <action> defines the action to perform if the condition applies. See
12008 below.
Emeric Brun0a3b67f2010-09-24 15:34:53 +020012009
12010 <condition> is a standard layer 4-7 ACL-based condition (see section 7).
12011
Davor Ocelice9ed2812017-12-25 17:49:28 +010012012 Response contents can be analyzed at an early stage of response processing
Emeric Brun0a3b67f2010-09-24 15:34:53 +020012013 called "TCP content inspection". During this stage, ACL-based rules are
12014 evaluated every time the response contents are updated, until either an
Willy Tarreaucc1e04b2013-09-11 23:20:29 +020012015 "accept", "close" or a "reject" rule matches, or a TCP response inspection
12016 delay is set and expires with no matching rule.
Emeric Brun0a3b67f2010-09-24 15:34:53 +020012017
12018 Most often, these decisions will consider a protocol recognition or validity.
12019
12020 Content-based rules are evaluated in their exact declaration order. If no
12021 rule matches or if there is no rule, the default action is to accept the
12022 contents. There is no specific limit to the number of rules which may be
12023 inserted.
12024
Thierry FOURNIER236657b2015-08-19 08:25:14 +020012025 Several types of actions are supported :
Emeric Brun0a3b67f2010-09-24 15:34:53 +020012026 - accept :
12027 accepts the response if the condition is true (when used with "if")
12028 or false (when used with "unless"). The first such rule executed ends
12029 the rules evaluation.
12030
Willy Tarreaucc1e04b2013-09-11 23:20:29 +020012031 - close :
12032 immediately closes the connection with the server if the condition is
12033 true (when used with "if"), or false (when used with "unless"). The
12034 first such rule executed ends the rules evaluation. The main purpose of
12035 this action is to force a connection to be finished between a client
12036 and a server after an exchange when the application protocol expects
12037 some long time outs to elapse first. The goal is to eliminate idle
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030012038 connections which take significant resources on servers with certain
Willy Tarreaucc1e04b2013-09-11 23:20:29 +020012039 protocols.
12040
Emeric Brun0a3b67f2010-09-24 15:34:53 +020012041 - reject :
12042 rejects the response if the condition is true (when used with "if")
12043 or false (when used with "unless"). The first such rule executed ends
Jamie Gloudonaaa21002012-08-25 00:18:33 -040012044 the rules evaluation. Rejected session are immediately closed.
Emeric Brun0a3b67f2010-09-24 15:34:53 +020012045
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020012046 - set-var(<var-name>) <expr>
12047 Sets a variable.
12048
Christopher Faulet85d79c92016-11-09 16:54:56 +010012049 - unset-var(<var-name>)
12050 Unsets a variable.
12051
Thierry FOURNIERe0627bd2015-08-04 08:20:33 +020012052 - sc-inc-gpc0(<sc-id>):
12053 This action increments the GPC0 counter according to the sticky
12054 counter designated by <sc-id>. If an error occurs, this action fails
12055 silently and the actions evaluation continues.
12056
Frédéric Lécaille6778b272018-01-29 15:22:53 +010012057 - sc-inc-gpc1(<sc-id>):
12058 This action increments the GPC1 counter according to the sticky
12059 counter designated by <sc-id>. If an error occurs, this action fails
12060 silently and the actions evaluation continues.
12061
Cédric Dufour0d7712d2019-11-06 18:38:53 +010012062 - sc-set-gpt0(<sc-id>) { <int> | <expr> }
12063 This action sets the 32-bit unsigned GPT0 tag according to the sticky
12064 counter designated by <sc-id> and the value of <int>/<expr>. The
12065 expected result is a boolean. If an error occurs, this action silently
12066 fails and the actions evaluation continues.
Thierry FOURNIER236657b2015-08-19 08:25:14 +020012067
Willy Tarreau2d392c22015-08-24 01:43:45 +020012068 - "silent-drop" :
12069 This stops the evaluation of the rules and makes the client-facing
Davor Ocelice9ed2812017-12-25 17:49:28 +010012070 connection suddenly disappear using a system-dependent way that tries
Willy Tarreau2d392c22015-08-24 01:43:45 +020012071 to prevent the client from being notified. The effect it then that the
12072 client still sees an established connection while there's none on
12073 HAProxy. The purpose is to achieve a comparable effect to "tarpit"
12074 except that it doesn't use any local resource at all on the machine
12075 running HAProxy. It can resist much higher loads than "tarpit", and
Davor Ocelice9ed2812017-12-25 17:49:28 +010012076 slow down stronger attackers. It is important to understand the impact
12077 of using this mechanism. All stateful equipment placed between the
Willy Tarreau2d392c22015-08-24 01:43:45 +020012078 client and HAProxy (firewalls, proxies, load balancers) will also keep
12079 the established connection for a long time and may suffer from this
Davor Ocelice9ed2812017-12-25 17:49:28 +010012080 action. On modern Linux systems running with enough privileges, the
Willy Tarreau2d392c22015-08-24 01:43:45 +020012081 TCP_REPAIR socket option is used to block the emission of a TCP
12082 reset. On other systems, the socket's TTL is reduced to 1 so that the
12083 TCP reset doesn't pass the first router, though it's still delivered to
12084 local networks. Do not use it unless you fully understand how it works.
12085
Christopher Faulet76c09ef2017-09-21 11:03:52 +020012086 - send-spoe-group <engine-name> <group-name>
12087 Send a group of SPOE messages.
12088
Emeric Brun0a3b67f2010-09-24 15:34:53 +020012089 Note that the "if/unless" condition is optional. If no condition is set on
12090 the action, it is simply performed unconditionally. That can be useful for
12091 for changing the default action to a reject.
12092
Jamie Gloudonaaa21002012-08-25 00:18:33 -040012093 It is perfectly possible to match layer 7 contents with "tcp-response
12094 content" rules, but then it is important to ensure that a full response has
12095 been buffered, otherwise no contents will match. In order to achieve this,
12096 the best solution involves detecting the HTTP protocol during the inspection
Emeric Brun0a3b67f2010-09-24 15:34:53 +020012097 period.
12098
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020012099 The "set-var" is used to set the content of a variable. The variable is
12100 declared inline.
12101
Daniel Schneller0b547052016-03-21 20:46:57 +010012102 <var-name> The name of the variable starts with an indication about
12103 its scope. The scopes allowed are:
Christopher Fauletff2613e2016-11-09 11:36:17 +010012104 "proc" : the variable is shared with the whole process
Daniel Schneller0b547052016-03-21 20:46:57 +010012105 "sess" : the variable is shared with the whole session
12106 "txn" : the variable is shared with the transaction
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020012107 (request and response)
Daniel Schneller0b547052016-03-21 20:46:57 +010012108 "req" : the variable is shared only during request
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020012109 processing
Daniel Schneller0b547052016-03-21 20:46:57 +010012110 "res" : the variable is shared only during response
12111 processing
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020012112 This prefix is followed by a name. The separator is a '.'.
Christopher Fauletb71557a2016-10-31 10:49:03 +010012113 The name may only contain characters 'a-z', 'A-Z', '0-9',
12114 '.' and '_'.
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020012115
12116 <expr> Is a standard HAProxy expression formed by a sample-fetch
12117 followed by some converters.
12118
12119 Example:
12120
12121 tcp-request content set-var(sess.my_var) src
12122
Christopher Faulet85d79c92016-11-09 16:54:56 +010012123 The "unset-var" is used to unset a variable. See above for details about
12124 <var-name>.
12125
12126 Example:
12127
12128 tcp-request content unset-var(sess.my_var)
12129
Christopher Faulet76c09ef2017-09-21 11:03:52 +020012130 The "send-spoe-group" is used to trigger sending of a group of SPOE
12131 messages. To do so, the SPOE engine used to send messages must be defined, as
12132 well as the SPOE group to send. Of course, the SPOE engine must refer to an
12133 existing SPOE filter. If not engine name is provided on the SPOE filter line,
12134 the SPOE agent name must be used.
12135
12136 <engine-name> The SPOE engine name.
12137
12138 <group-name> The SPOE group name as specified in the engine configuration.
12139
Emeric Brun0a3b67f2010-09-24 15:34:53 +020012140 See section 7 about ACL usage.
12141
12142 See also : "tcp-request content", "tcp-response inspect-delay"
12143
12144
Willy Tarreau4f614292016-10-21 17:49:36 +020012145tcp-request session <action> [{if | unless} <condition>]
12146 Perform an action on a validated session depending on a layer 5 condition
12147 May be used in sections : defaults | frontend | listen | backend
12148 no | yes | yes | no
12149 Arguments :
12150 <action> defines the action to perform if the condition applies. See
12151 below.
12152
12153 <condition> is a standard layer5-only ACL-based condition (see section 7).
12154
Davor Ocelice9ed2812017-12-25 17:49:28 +010012155 Once a session is validated, (i.e. after all handshakes have been completed),
Willy Tarreau4f614292016-10-21 17:49:36 +020012156 it is possible to evaluate some conditions to decide whether this session
12157 must be accepted or dropped or have its counters tracked. Those conditions
12158 cannot make use of any data contents because no buffers are allocated yet and
12159 the processing cannot wait at this stage. The main use case it to copy some
12160 early information into variables (since variables are accessible in the
12161 session), or to keep track of some information collected after the handshake,
12162 such as SSL-level elements (SNI, ciphers, client cert's CN) or information
Davor Ocelice9ed2812017-12-25 17:49:28 +010012163 from the PROXY protocol header (e.g. track a source forwarded this way). The
Willy Tarreau4f614292016-10-21 17:49:36 +020012164 extracted information can thus be copied to a variable or tracked using
12165 "track-sc" rules. Of course it is also possible to decide to accept/reject as
12166 with other rulesets. Most operations performed here could also be performed
12167 in "tcp-request content" rules, except that in HTTP these rules are evaluated
12168 for each new request, and that might not always be acceptable. For example a
12169 rule might increment a counter on each evaluation. It would also be possible
12170 that a country is resolved by geolocation from the source IP address,
12171 assigned to a session-wide variable, then the source address rewritten from
12172 an HTTP header for all requests. If some contents need to be inspected in
12173 order to take the decision, the "tcp-request content" statements must be used
12174 instead.
12175
12176 The "tcp-request session" rules are evaluated in their exact declaration
12177 order. If no rule matches or if there is no rule, the default action is to
12178 accept the incoming session. There is no specific limit to the number of
12179 rules which may be inserted.
12180
12181 Several types of actions are supported :
12182 - accept : the request is accepted
12183 - reject : the request is rejected and the connection is closed
12184 - { track-sc0 | track-sc1 | track-sc2 } <key> [table <table>]
12185 - sc-inc-gpc0(<sc-id>)
Frédéric Lécaille6778b272018-01-29 15:22:53 +010012186 - sc-inc-gpc1(<sc-id>)
Cédric Dufour0d7712d2019-11-06 18:38:53 +010012187 - sc-set-gpt0(<sc-id>) { <int> | <expr> }
Willy Tarreau4f614292016-10-21 17:49:36 +020012188 - set-var(<var-name>) <expr>
Christopher Faulet85d79c92016-11-09 16:54:56 +010012189 - unset-var(<var-name>)
Willy Tarreau4f614292016-10-21 17:49:36 +020012190 - silent-drop
12191
12192 These actions have the same meaning as their respective counter-parts in
12193 "tcp-request connection" and "tcp-request content", so please refer to these
12194 sections for a complete description.
12195
12196 Note that the "if/unless" condition is optional. If no condition is set on
12197 the action, it is simply performed unconditionally. That can be useful for
12198 "track-sc*" actions as well as for changing the default action to a reject.
12199
12200 Example: track the original source address by default, or the one advertised
12201 in the PROXY protocol header for connection coming from the local
12202 proxies. The first connection-level rule enables receipt of the
12203 PROXY protocol for these ones, the second rule tracks whatever
12204 address we decide to keep after optional decoding.
12205
12206 tcp-request connection expect-proxy layer4 if { src -f proxies.lst }
12207 tcp-request session track-sc0 src
12208
12209 Example: accept all sessions from white-listed hosts, reject too fast
12210 sessions without counting them, and track accepted sessions.
12211 This results in session rate being capped from abusive sources.
12212
12213 tcp-request session accept if { src -f /etc/haproxy/whitelist.lst }
12214 tcp-request session reject if { src_sess_rate gt 10 }
12215 tcp-request session track-sc0 src
12216
12217 Example: accept all sessions from white-listed hosts, count all other
12218 sessions and reject too fast ones. This results in abusive ones
12219 being blocked as long as they don't slow down.
12220
12221 tcp-request session accept if { src -f /etc/haproxy/whitelist.lst }
12222 tcp-request session track-sc0 src
12223 tcp-request session reject if { sc0_sess_rate gt 10 }
12224
12225 See section 7 about ACL usage.
12226
12227 See also : "tcp-request connection", "tcp-request content", "stick-table"
12228
12229
Emeric Brun0a3b67f2010-09-24 15:34:53 +020012230tcp-response inspect-delay <timeout>
12231 Set the maximum allowed time to wait for a response during content inspection
12232 May be used in sections : defaults | frontend | listen | backend
12233 no | no | yes | yes
12234 Arguments :
12235 <timeout> is the timeout value specified in milliseconds by default, but
12236 can be in any other unit if the number is suffixed by the unit,
12237 as explained at the top of this document.
12238
12239 See also : "tcp-response content", "tcp-request inspect-delay".
12240
12241
Krzysztof Piotr Oledzki5259dfe2008-01-21 01:54:06 +010012242timeout check <timeout>
12243 Set additional check timeout, but only after a connection has been already
12244 established.
12245
12246 May be used in sections: defaults | frontend | listen | backend
12247 yes | no | yes | yes
12248 Arguments:
12249 <timeout> is the timeout value specified in milliseconds by default, but
12250 can be in any other unit if the number is suffixed by the unit,
12251 as explained at the top of this document.
12252
12253 If set, haproxy uses min("timeout connect", "inter") as a connect timeout
12254 for check and "timeout check" as an additional read timeout. The "min" is
Davor Ocelice9ed2812017-12-25 17:49:28 +010012255 used so that people running with *very* long "timeout connect" (e.g. those
Krzysztof Piotr Oledzki5259dfe2008-01-21 01:54:06 +010012256 who needed this due to the queue or tarpit) do not slow down their checks.
Willy Tarreaud7550a22010-02-10 05:10:19 +010012257 (Please also note that there is no valid reason to have such long connect
12258 timeouts, because "timeout queue" and "timeout tarpit" can always be used to
12259 avoid that).
Krzysztof Piotr Oledzki5259dfe2008-01-21 01:54:06 +010012260
12261 If "timeout check" is not set haproxy uses "inter" for complete check
12262 timeout (connect + read) exactly like all <1.3.15 version.
12263
12264 In most cases check request is much simpler and faster to handle than normal
12265 requests and people may want to kick out laggy servers so this timeout should
Willy Tarreau41a340d2008-01-22 12:25:31 +010012266 be smaller than "timeout server".
Krzysztof Piotr Oledzki5259dfe2008-01-21 01:54:06 +010012267
12268 This parameter is specific to backends, but can be specified once for all in
12269 "defaults" sections. This is in fact one of the easiest solutions not to
12270 forget about it.
12271
Willy Tarreau41a340d2008-01-22 12:25:31 +010012272 See also: "timeout connect", "timeout queue", "timeout server",
12273 "timeout tarpit".
Krzysztof Piotr Oledzki5259dfe2008-01-21 01:54:06 +010012274
12275
Willy Tarreau0ba27502007-12-24 16:55:16 +010012276timeout client <timeout>
Willy Tarreau0ba27502007-12-24 16:55:16 +010012277 Set the maximum inactivity time on the client side.
12278 May be used in sections : defaults | frontend | listen | backend
12279 yes | yes | yes | no
12280 Arguments :
Willy Tarreau844e3c52008-01-11 16:28:18 +010012281 <timeout> is the timeout value specified in milliseconds by default, but
Willy Tarreau0ba27502007-12-24 16:55:16 +010012282 can be in any other unit if the number is suffixed by the unit,
12283 as explained at the top of this document.
12284
12285 The inactivity timeout applies when the client is expected to acknowledge or
12286 send data. In HTTP mode, this timeout is particularly important to consider
12287 during the first phase, when the client sends the request, and during the
Baptiste Assmann2e1941e2016-03-06 23:24:12 +010012288 response while it is reading data sent by the server. That said, for the
12289 first phase, it is preferable to set the "timeout http-request" to better
12290 protect HAProxy from Slowloris like attacks. The value is specified in
12291 milliseconds by default, but can be in any other unit if the number is
Willy Tarreau0ba27502007-12-24 16:55:16 +010012292 suffixed by the unit, as specified at the top of this document. In TCP mode
12293 (and to a lesser extent, in HTTP mode), it is highly recommended that the
12294 client timeout remains equal to the server timeout in order to avoid complex
Willy Tarreaud2a4aa22008-01-31 15:28:22 +010012295 situations to debug. It is a good practice to cover one or several TCP packet
Willy Tarreau0ba27502007-12-24 16:55:16 +010012296 losses by specifying timeouts that are slightly above multiples of 3 seconds
Davor Ocelice9ed2812017-12-25 17:49:28 +010012297 (e.g. 4 or 5 seconds). If some long-lived sessions are mixed with short-lived
12298 sessions (e.g. WebSocket and HTTP), it's worth considering "timeout tunnel",
Willy Tarreau05cdd962014-05-10 14:30:07 +020012299 which overrides "timeout client" and "timeout server" for tunnels, as well as
12300 "timeout client-fin" for half-closed connections.
Willy Tarreau0ba27502007-12-24 16:55:16 +010012301
12302 This parameter is specific to frontends, but can be specified once for all in
12303 "defaults" sections. This is in fact one of the easiest solutions not to
12304 forget about it. An unspecified timeout results in an infinite timeout, which
12305 is not recommended. Such a usage is accepted and works but reports a warning
John Roeslerfb2fce12019-07-10 15:45:51 -050012306 during startup because it may result in accumulation of expired sessions in
Willy Tarreau0ba27502007-12-24 16:55:16 +010012307 the system if the system's timeouts are not configured either.
12308
Willy Tarreau95c4e142017-11-26 12:18:55 +010012309 This also applies to HTTP/2 connections, which will be closed with GOAWAY.
Lukas Tribus75df9d72017-11-24 19:05:12 +010012310
Tim Duesterhus86e6b6e2019-05-14 20:57:59 +020012311 See also : "timeout server", "timeout tunnel", "timeout http-request".
Willy Tarreau0ba27502007-12-24 16:55:16 +010012312
Willy Tarreau0ba27502007-12-24 16:55:16 +010012313
Willy Tarreau05cdd962014-05-10 14:30:07 +020012314timeout client-fin <timeout>
12315 Set the inactivity timeout on the client side for half-closed connections.
12316 May be used in sections : defaults | frontend | listen | backend
12317 yes | yes | yes | no
12318 Arguments :
12319 <timeout> is the timeout value specified in milliseconds by default, but
12320 can be in any other unit if the number is suffixed by the unit,
12321 as explained at the top of this document.
12322
12323 The inactivity timeout applies when the client is expected to acknowledge or
12324 send data while one direction is already shut down. This timeout is different
12325 from "timeout client" in that it only applies to connections which are closed
12326 in one direction. This is particularly useful to avoid keeping connections in
12327 FIN_WAIT state for too long when clients do not disconnect cleanly. This
12328 problem is particularly common long connections such as RDP or WebSocket.
12329 Note that this timeout can override "timeout tunnel" when a connection shuts
Willy Tarreau599391a2017-11-24 10:16:00 +010012330 down in one direction. It is applied to idle HTTP/2 connections once a GOAWAY
12331 frame was sent, often indicating an expectation that the connection quickly
12332 ends.
Willy Tarreau05cdd962014-05-10 14:30:07 +020012333
12334 This parameter is specific to frontends, but can be specified once for all in
12335 "defaults" sections. By default it is not set, so half-closed connections
12336 will use the other timeouts (timeout.client or timeout.tunnel).
12337
12338 See also : "timeout client", "timeout server-fin", and "timeout tunnel".
12339
12340
Willy Tarreau0ba27502007-12-24 16:55:16 +010012341timeout connect <timeout>
Willy Tarreau0ba27502007-12-24 16:55:16 +010012342 Set the maximum time to wait for a connection attempt to a server to succeed.
12343 May be used in sections : defaults | frontend | listen | backend
12344 yes | no | yes | yes
12345 Arguments :
Willy Tarreau844e3c52008-01-11 16:28:18 +010012346 <timeout> is the timeout value specified in milliseconds by default, but
Willy Tarreau0ba27502007-12-24 16:55:16 +010012347 can be in any other unit if the number is suffixed by the unit,
12348 as explained at the top of this document.
12349
12350 If the server is located on the same LAN as haproxy, the connection should be
Willy Tarreaud2a4aa22008-01-31 15:28:22 +010012351 immediate (less than a few milliseconds). Anyway, it is a good practice to
Willy Tarreaud72758d2010-01-12 10:42:19 +010012352 cover one or several TCP packet losses by specifying timeouts that are
Davor Ocelice9ed2812017-12-25 17:49:28 +010012353 slightly above multiples of 3 seconds (e.g. 4 or 5 seconds). By default, the
Krzysztof Piotr Oledzki5259dfe2008-01-21 01:54:06 +010012354 connect timeout also presets both queue and tarpit timeouts to the same value
12355 if these have not been specified.
Willy Tarreau0ba27502007-12-24 16:55:16 +010012356
12357 This parameter is specific to backends, but can be specified once for all in
12358 "defaults" sections. This is in fact one of the easiest solutions not to
12359 forget about it. An unspecified timeout results in an infinite timeout, which
12360 is not recommended. Such a usage is accepted and works but reports a warning
John Roeslerfb2fce12019-07-10 15:45:51 -050012361 during startup because it may result in accumulation of failed sessions in
Willy Tarreau0ba27502007-12-24 16:55:16 +010012362 the system if the system's timeouts are not configured either.
12363
Tim Duesterhus86e6b6e2019-05-14 20:57:59 +020012364 See also: "timeout check", "timeout queue", "timeout server", "timeout tarpit".
Willy Tarreau0ba27502007-12-24 16:55:16 +010012365
Willy Tarreau0ba27502007-12-24 16:55:16 +010012366
Willy Tarreaub16a5742010-01-10 14:46:16 +010012367timeout http-keep-alive <timeout>
12368 Set the maximum allowed time to wait for a new HTTP request to appear
12369 May be used in sections : defaults | frontend | listen | backend
12370 yes | yes | yes | yes
12371 Arguments :
12372 <timeout> is the timeout value specified in milliseconds by default, but
12373 can be in any other unit if the number is suffixed by the unit,
12374 as explained at the top of this document.
12375
12376 By default, the time to wait for a new request in case of keep-alive is set
12377 by "timeout http-request". However this is not always convenient because some
12378 people want very short keep-alive timeouts in order to release connections
12379 faster, and others prefer to have larger ones but still have short timeouts
12380 once the request has started to present itself.
12381
12382 The "http-keep-alive" timeout covers these needs. It will define how long to
12383 wait for a new HTTP request to start coming after a response was sent. Once
12384 the first byte of request has been seen, the "http-request" timeout is used
12385 to wait for the complete request to come. Note that empty lines prior to a
12386 new request do not refresh the timeout and are not counted as a new request.
12387
12388 There is also another difference between the two timeouts : when a connection
12389 expires during timeout http-keep-alive, no error is returned, the connection
12390 just closes. If the connection expires in "http-request" while waiting for a
12391 connection to complete, a HTTP 408 error is returned.
12392
12393 In general it is optimal to set this value to a few tens to hundreds of
12394 milliseconds, to allow users to fetch all objects of a page at once but
Davor Ocelice9ed2812017-12-25 17:49:28 +010012395 without waiting for further clicks. Also, if set to a very small value (e.g.
Willy Tarreaub16a5742010-01-10 14:46:16 +010012396 1 millisecond) it will probably only accept pipelined requests but not the
12397 non-pipelined ones. It may be a nice trade-off for very large sites running
Patrick Mézard2382ad62010-05-09 10:43:32 +020012398 with tens to hundreds of thousands of clients.
Willy Tarreaub16a5742010-01-10 14:46:16 +010012399
12400 If this parameter is not set, the "http-request" timeout applies, and if both
12401 are not set, "timeout client" still applies at the lower level. It should be
12402 set in the frontend to take effect, unless the frontend is in TCP mode, in
12403 which case the HTTP backend's timeout will be used.
12404
Willy Tarreau95c4e142017-11-26 12:18:55 +010012405 When using HTTP/2 "timeout client" is applied instead. This is so we can keep
12406 using short keep-alive timeouts in HTTP/1.1 while using longer ones in HTTP/2
Lukas Tribus75df9d72017-11-24 19:05:12 +010012407 (where we only have one connection per client and a connection setup).
12408
Willy Tarreaub16a5742010-01-10 14:46:16 +010012409 See also : "timeout http-request", "timeout client".
12410
12411
Willy Tarreau036fae02008-01-06 13:24:40 +010012412timeout http-request <timeout>
12413 Set the maximum allowed time to wait for a complete HTTP request
12414 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaucd7afc02009-07-12 10:03:17 +020012415 yes | yes | yes | yes
Willy Tarreau036fae02008-01-06 13:24:40 +010012416 Arguments :
Willy Tarreau844e3c52008-01-11 16:28:18 +010012417 <timeout> is the timeout value specified in milliseconds by default, but
Willy Tarreau036fae02008-01-06 13:24:40 +010012418 can be in any other unit if the number is suffixed by the unit,
12419 as explained at the top of this document.
12420
12421 In order to offer DoS protection, it may be required to lower the maximum
12422 accepted time to receive a complete HTTP request without affecting the client
12423 timeout. This helps protecting against established connections on which
12424 nothing is sent. The client timeout cannot offer a good protection against
12425 this abuse because it is an inactivity timeout, which means that if the
12426 attacker sends one character every now and then, the timeout will not
12427 trigger. With the HTTP request timeout, no matter what speed the client
Willy Tarreau2705a612014-05-23 17:38:34 +020012428 types, the request will be aborted if it does not complete in time. When the
12429 timeout expires, an HTTP 408 response is sent to the client to inform it
12430 about the problem, and the connection is closed. The logs will report
12431 termination codes "cR". Some recent browsers are having problems with this
Davor Ocelice9ed2812017-12-25 17:49:28 +010012432 standard, well-documented behavior, so it might be needed to hide the 408
Willy Tarreau0f228a02015-05-01 15:37:53 +020012433 code using "option http-ignore-probes" or "errorfile 408 /dev/null". See
12434 more details in the explanations of the "cR" termination code in section 8.5.
Willy Tarreau036fae02008-01-06 13:24:40 +010012435
Baptiste Assmanneccdf432015-10-28 13:49:01 +010012436 By default, this timeout only applies to the header part of the request,
12437 and not to any data. As soon as the empty line is received, this timeout is
12438 not used anymore. When combined with "option http-buffer-request", this
12439 timeout also applies to the body of the request..
12440 It is used again on keep-alive connections to wait for a second
Willy Tarreaub16a5742010-01-10 14:46:16 +010012441 request if "timeout http-keep-alive" is not set.
Willy Tarreau036fae02008-01-06 13:24:40 +010012442
12443 Generally it is enough to set it to a few seconds, as most clients send the
12444 full request immediately upon connection. Add 3 or more seconds to cover TCP
Davor Ocelice9ed2812017-12-25 17:49:28 +010012445 retransmits but that's all. Setting it to very low values (e.g. 50 ms) will
Willy Tarreau036fae02008-01-06 13:24:40 +010012446 generally work on local networks as long as there are no packet losses. This
12447 will prevent people from sending bare HTTP requests using telnet.
12448
12449 If this parameter is not set, the client timeout still applies between each
Willy Tarreaucd7afc02009-07-12 10:03:17 +020012450 chunk of the incoming request. It should be set in the frontend to take
12451 effect, unless the frontend is in TCP mode, in which case the HTTP backend's
12452 timeout will be used.
Willy Tarreau036fae02008-01-06 13:24:40 +010012453
Willy Tarreau0f228a02015-05-01 15:37:53 +020012454 See also : "errorfile", "http-ignore-probes", "timeout http-keep-alive", and
Baptiste Assmanneccdf432015-10-28 13:49:01 +010012455 "timeout client", "option http-buffer-request".
Willy Tarreau036fae02008-01-06 13:24:40 +010012456
Willy Tarreau844e3c52008-01-11 16:28:18 +010012457
12458timeout queue <timeout>
12459 Set the maximum time to wait in the queue for a connection slot to be free
12460 May be used in sections : defaults | frontend | listen | backend
12461 yes | no | yes | yes
12462 Arguments :
12463 <timeout> is the timeout value specified in milliseconds by default, but
12464 can be in any other unit if the number is suffixed by the unit,
12465 as explained at the top of this document.
12466
12467 When a server's maxconn is reached, connections are left pending in a queue
12468 which may be server-specific or global to the backend. In order not to wait
12469 indefinitely, a timeout is applied to requests pending in the queue. If the
12470 timeout is reached, it is considered that the request will almost never be
12471 served, so it is dropped and a 503 error is returned to the client.
12472
12473 The "timeout queue" statement allows to fix the maximum time for a request to
12474 be left pending in a queue. If unspecified, the same value as the backend's
12475 connection timeout ("timeout connect") is used, for backwards compatibility
12476 with older versions with no "timeout queue" parameter.
12477
Tim Duesterhus86e6b6e2019-05-14 20:57:59 +020012478 See also : "timeout connect".
Willy Tarreau844e3c52008-01-11 16:28:18 +010012479
12480
12481timeout server <timeout>
Willy Tarreau844e3c52008-01-11 16:28:18 +010012482 Set the maximum inactivity time on the server side.
12483 May be used in sections : defaults | frontend | listen | backend
12484 yes | no | yes | yes
12485 Arguments :
12486 <timeout> is the timeout value specified in milliseconds by default, but
12487 can be in any other unit if the number is suffixed by the unit,
12488 as explained at the top of this document.
12489
12490 The inactivity timeout applies when the server is expected to acknowledge or
12491 send data. In HTTP mode, this timeout is particularly important to consider
12492 during the first phase of the server's response, when it has to send the
12493 headers, as it directly represents the server's processing time for the
12494 request. To find out what value to put there, it's often good to start with
12495 what would be considered as unacceptable response times, then check the logs
12496 to observe the response time distribution, and adjust the value accordingly.
12497
12498 The value is specified in milliseconds by default, but can be in any other
12499 unit if the number is suffixed by the unit, as specified at the top of this
12500 document. In TCP mode (and to a lesser extent, in HTTP mode), it is highly
12501 recommended that the client timeout remains equal to the server timeout in
12502 order to avoid complex situations to debug. Whatever the expected server
Willy Tarreaud2a4aa22008-01-31 15:28:22 +010012503 response times, it is a good practice to cover at least one or several TCP
Willy Tarreau844e3c52008-01-11 16:28:18 +010012504 packet losses by specifying timeouts that are slightly above multiples of 3
Davor Ocelice9ed2812017-12-25 17:49:28 +010012505 seconds (e.g. 4 or 5 seconds minimum). If some long-lived sessions are mixed
12506 with short-lived sessions (e.g. WebSocket and HTTP), it's worth considering
Willy Tarreauce887fd2012-05-12 12:50:00 +020012507 "timeout tunnel", which overrides "timeout client" and "timeout server" for
12508 tunnels.
Willy Tarreau844e3c52008-01-11 16:28:18 +010012509
12510 This parameter is specific to backends, but can be specified once for all in
12511 "defaults" sections. This is in fact one of the easiest solutions not to
12512 forget about it. An unspecified timeout results in an infinite timeout, which
12513 is not recommended. Such a usage is accepted and works but reports a warning
John Roeslerfb2fce12019-07-10 15:45:51 -050012514 during startup because it may result in accumulation of expired sessions in
Willy Tarreau844e3c52008-01-11 16:28:18 +010012515 the system if the system's timeouts are not configured either.
12516
Tim Duesterhus86e6b6e2019-05-14 20:57:59 +020012517 See also : "timeout client" and "timeout tunnel".
Willy Tarreau844e3c52008-01-11 16:28:18 +010012518
Willy Tarreau05cdd962014-05-10 14:30:07 +020012519
12520timeout server-fin <timeout>
12521 Set the inactivity timeout on the server side for half-closed connections.
12522 May be used in sections : defaults | frontend | listen | backend
12523 yes | no | yes | yes
12524 Arguments :
12525 <timeout> is the timeout value specified in milliseconds by default, but
12526 can be in any other unit if the number is suffixed by the unit,
12527 as explained at the top of this document.
12528
12529 The inactivity timeout applies when the server is expected to acknowledge or
12530 send data while one direction is already shut down. This timeout is different
12531 from "timeout server" in that it only applies to connections which are closed
12532 in one direction. This is particularly useful to avoid keeping connections in
12533 FIN_WAIT state for too long when a remote server does not disconnect cleanly.
12534 This problem is particularly common long connections such as RDP or WebSocket.
12535 Note that this timeout can override "timeout tunnel" when a connection shuts
12536 down in one direction. This setting was provided for completeness, but in most
12537 situations, it should not be needed.
12538
12539 This parameter is specific to backends, but can be specified once for all in
12540 "defaults" sections. By default it is not set, so half-closed connections
12541 will use the other timeouts (timeout.server or timeout.tunnel).
12542
12543 See also : "timeout client-fin", "timeout server", and "timeout tunnel".
12544
Willy Tarreau844e3c52008-01-11 16:28:18 +010012545
12546timeout tarpit <timeout>
Cyril Bonté78caf842010-03-10 22:41:43 +010012547 Set the duration for which tarpitted connections will be maintained
Willy Tarreau844e3c52008-01-11 16:28:18 +010012548 May be used in sections : defaults | frontend | listen | backend
12549 yes | yes | yes | yes
12550 Arguments :
12551 <timeout> is the tarpit duration specified in milliseconds by default, but
12552 can be in any other unit if the number is suffixed by the unit,
12553 as explained at the top of this document.
12554
Christopher Faulet87f1f3d2019-07-18 14:51:20 +020012555 When a connection is tarpitted using "http-request tarpit", it is maintained
12556 open with no activity for a certain amount of time, then closed. "timeout
12557 tarpit" defines how long it will be maintained open.
Willy Tarreau844e3c52008-01-11 16:28:18 +010012558
12559 The value is specified in milliseconds by default, but can be in any other
12560 unit if the number is suffixed by the unit, as specified at the top of this
12561 document. If unspecified, the same value as the backend's connection timeout
12562 ("timeout connect") is used, for backwards compatibility with older versions
Cyril Bonté78caf842010-03-10 22:41:43 +010012563 with no "timeout tarpit" parameter.
Willy Tarreau844e3c52008-01-11 16:28:18 +010012564
Tim Duesterhus86e6b6e2019-05-14 20:57:59 +020012565 See also : "timeout connect".
Willy Tarreau844e3c52008-01-11 16:28:18 +010012566
12567
Willy Tarreauce887fd2012-05-12 12:50:00 +020012568timeout tunnel <timeout>
12569 Set the maximum inactivity time on the client and server side for tunnels.
12570 May be used in sections : defaults | frontend | listen | backend
12571 yes | no | yes | yes
12572 Arguments :
12573 <timeout> is the timeout value specified in milliseconds by default, but
12574 can be in any other unit if the number is suffixed by the unit,
12575 as explained at the top of this document.
12576
Jamie Gloudonaaa21002012-08-25 00:18:33 -040012577 The tunnel timeout applies when a bidirectional connection is established
Willy Tarreauce887fd2012-05-12 12:50:00 +020012578 between a client and a server, and the connection remains inactive in both
12579 directions. This timeout supersedes both the client and server timeouts once
12580 the connection becomes a tunnel. In TCP, this timeout is used as soon as no
Davor Ocelice9ed2812017-12-25 17:49:28 +010012581 analyzer remains attached to either connection (e.g. tcp content rules are
12582 accepted). In HTTP, this timeout is used when a connection is upgraded (e.g.
Willy Tarreauce887fd2012-05-12 12:50:00 +020012583 when switching to the WebSocket protocol, or forwarding a CONNECT request
12584 to a proxy), or after the first response when no keepalive/close option is
12585 specified.
12586
Willy Tarreau05cdd962014-05-10 14:30:07 +020012587 Since this timeout is usually used in conjunction with long-lived connections,
12588 it usually is a good idea to also set "timeout client-fin" to handle the
12589 situation where a client suddenly disappears from the net and does not
12590 acknowledge a close, or sends a shutdown and does not acknowledge pending
12591 data anymore. This can happen in lossy networks where firewalls are present,
12592 and is detected by the presence of large amounts of sessions in a FIN_WAIT
12593 state.
12594
Willy Tarreauce887fd2012-05-12 12:50:00 +020012595 The value is specified in milliseconds by default, but can be in any other
12596 unit if the number is suffixed by the unit, as specified at the top of this
12597 document. Whatever the expected normal idle time, it is a good practice to
12598 cover at least one or several TCP packet losses by specifying timeouts that
Davor Ocelice9ed2812017-12-25 17:49:28 +010012599 are slightly above multiples of 3 seconds (e.g. 4 or 5 seconds minimum).
Willy Tarreauce887fd2012-05-12 12:50:00 +020012600
12601 This parameter is specific to backends, but can be specified once for all in
12602 "defaults" sections. This is in fact one of the easiest solutions not to
12603 forget about it.
12604
12605 Example :
12606 defaults http
12607 option http-server-close
12608 timeout connect 5s
12609 timeout client 30s
Willy Tarreau05cdd962014-05-10 14:30:07 +020012610 timeout client-fin 30s
Willy Tarreauce887fd2012-05-12 12:50:00 +020012611 timeout server 30s
12612 timeout tunnel 1h # timeout to use with WebSocket and CONNECT
12613
Willy Tarreau05cdd962014-05-10 14:30:07 +020012614 See also : "timeout client", "timeout client-fin", "timeout server".
Willy Tarreauce887fd2012-05-12 12:50:00 +020012615
12616
Willy Tarreau844e3c52008-01-11 16:28:18 +010012617transparent (deprecated)
12618 Enable client-side transparent proxying
12619 May be used in sections : defaults | frontend | listen | backend
Willy Tarreau4b1f8592008-12-23 23:13:55 +010012620 yes | no | yes | yes
Willy Tarreau844e3c52008-01-11 16:28:18 +010012621 Arguments : none
12622
12623 This keyword was introduced in order to provide layer 7 persistence to layer
12624 3 load balancers. The idea is to use the OS's ability to redirect an incoming
12625 connection for a remote address to a local process (here HAProxy), and let
12626 this process know what address was initially requested. When this option is
12627 used, sessions without cookies will be forwarded to the original destination
12628 IP address of the incoming request (which should match that of another
12629 equipment), while requests with cookies will still be forwarded to the
12630 appropriate server.
12631
12632 The "transparent" keyword is deprecated, use "option transparent" instead.
12633
12634 Note that contrary to a common belief, this option does NOT make HAProxy
12635 present the client's IP to the server when establishing the connection.
12636
Willy Tarreau844e3c52008-01-11 16:28:18 +010012637 See also: "option transparent"
12638
William Lallemanda73203e2012-03-12 12:48:57 +010012639unique-id-format <string>
12640 Generate a unique ID for each request.
12641 May be used in sections : defaults | frontend | listen | backend
12642 yes | yes | yes | no
12643 Arguments :
12644 <string> is a log-format string.
12645
Cyril Bonté108cf6e2012-04-21 23:30:29 +020012646 This keyword creates a ID for each request using the custom log format. A
12647 unique ID is useful to trace a request passing through many components of
12648 a complex infrastructure. The newly created ID may also be logged using the
12649 %ID tag the log-format string.
William Lallemanda73203e2012-03-12 12:48:57 +010012650
Cyril Bonté108cf6e2012-04-21 23:30:29 +020012651 The format should be composed from elements that are guaranteed to be
12652 unique when combined together. For instance, if multiple haproxy instances
12653 are involved, it might be important to include the node name. It is often
12654 needed to log the incoming connection's source and destination addresses
12655 and ports. Note that since multiple requests may be performed over the same
12656 connection, including a request counter may help differentiate them.
12657 Similarly, a timestamp may protect against a rollover of the counter.
12658 Logging the process ID will avoid collisions after a service restart.
William Lallemanda73203e2012-03-12 12:48:57 +010012659
Cyril Bonté108cf6e2012-04-21 23:30:29 +020012660 It is recommended to use hexadecimal notation for many fields since it
12661 makes them more compact and saves space in logs.
William Lallemanda73203e2012-03-12 12:48:57 +010012662
Cyril Bonté108cf6e2012-04-21 23:30:29 +020012663 Example:
William Lallemanda73203e2012-03-12 12:48:57 +010012664
Julien Vehentf21be322014-03-07 08:27:34 -050012665 unique-id-format %{+X}o\ %ci:%cp_%fi:%fp_%Ts_%rt:%pid
William Lallemanda73203e2012-03-12 12:48:57 +010012666
12667 will generate:
12668
12669 7F000001:8296_7F00001E:1F90_4F7B0A69_0003:790A
12670
12671 See also: "unique-id-header"
12672
12673unique-id-header <name>
12674 Add a unique ID header in the HTTP request.
12675 May be used in sections : defaults | frontend | listen | backend
12676 yes | yes | yes | no
12677 Arguments :
12678 <name> is the name of the header.
12679
Cyril Bonté108cf6e2012-04-21 23:30:29 +020012680 Add a unique-id header in the HTTP request sent to the server, using the
12681 unique-id-format. It can't work if the unique-id-format doesn't exist.
William Lallemanda73203e2012-03-12 12:48:57 +010012682
Cyril Bonté108cf6e2012-04-21 23:30:29 +020012683 Example:
William Lallemanda73203e2012-03-12 12:48:57 +010012684
Julien Vehentf21be322014-03-07 08:27:34 -050012685 unique-id-format %{+X}o\ %ci:%cp_%fi:%fp_%Ts_%rt:%pid
William Lallemanda73203e2012-03-12 12:48:57 +010012686 unique-id-header X-Unique-ID
12687
12688 will generate:
12689
12690 X-Unique-ID: 7F000001:8296_7F00001E:1F90_4F7B0A69_0003:790A
12691
12692 See also: "unique-id-format"
Willy Tarreau844e3c52008-01-11 16:28:18 +010012693
Willy Tarreauf51658d2014-04-23 01:21:56 +020012694use_backend <backend> [{if | unless} <condition>]
Willy Tarreau1d0dfb12009-07-07 15:10:31 +020012695 Switch to a specific backend if/unless an ACL-based condition is matched.
Willy Tarreau844e3c52008-01-11 16:28:18 +010012696 May be used in sections : defaults | frontend | listen | backend
12697 no | yes | yes | no
12698 Arguments :
Bertrand Jacquin702d44f2013-11-19 11:43:06 +010012699 <backend> is the name of a valid backend or "listen" section, or a
12700 "log-format" string resolving to a backend name.
Willy Tarreau844e3c52008-01-11 16:28:18 +010012701
Willy Tarreauf51658d2014-04-23 01:21:56 +020012702 <condition> is a condition composed of ACLs, as described in section 7. If
12703 it is omitted, the rule is unconditionally applied.
Willy Tarreau844e3c52008-01-11 16:28:18 +010012704
12705 When doing content-switching, connections arrive on a frontend and are then
12706 dispatched to various backends depending on a number of conditions. The
12707 relation between the conditions and the backends is described with the
Willy Tarreau1d0dfb12009-07-07 15:10:31 +020012708 "use_backend" keyword. While it is normally used with HTTP processing, it can
Davor Ocelice9ed2812017-12-25 17:49:28 +010012709 also be used in pure TCP, either without content using stateless ACLs (e.g.
Willy Tarreau1d0dfb12009-07-07 15:10:31 +020012710 source address validation) or combined with a "tcp-request" rule to wait for
12711 some payload.
Willy Tarreau844e3c52008-01-11 16:28:18 +010012712
12713 There may be as many "use_backend" rules as desired. All of these rules are
12714 evaluated in their declaration order, and the first one which matches will
12715 assign the backend.
12716
12717 In the first form, the backend will be used if the condition is met. In the
12718 second form, the backend will be used if the condition is not met. If no
12719 condition is valid, the backend defined with "default_backend" will be used.
12720 If no default backend is defined, either the servers in the same section are
12721 used (in case of a "listen" section) or, in case of a frontend, no server is
12722 used and a 503 service unavailable response is returned.
12723
Willy Tarreau51aecc72009-07-12 09:47:04 +020012724 Note that it is possible to switch from a TCP frontend to an HTTP backend. In
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +010012725 this case, either the frontend has already checked that the protocol is HTTP,
Willy Tarreau51aecc72009-07-12 09:47:04 +020012726 and backend processing will immediately follow, or the backend will wait for
12727 a complete HTTP request to get in. This feature is useful when a frontend
12728 must decode several protocols on a unique port, one of them being HTTP.
12729
Bertrand Jacquin702d44f2013-11-19 11:43:06 +010012730 When <backend> is a simple name, it is resolved at configuration time, and an
12731 error is reported if the specified backend does not exist. If <backend> is
12732 a log-format string instead, no check may be done at configuration time, so
12733 the backend name is resolved dynamically at run time. If the resulting
12734 backend name does not correspond to any valid backend, no other rule is
12735 evaluated, and the default_backend directive is applied instead. Note that
12736 when using dynamic backend names, it is highly recommended to use a prefix
12737 that no other backend uses in order to ensure that an unauthorized backend
12738 cannot be forced from the request.
12739
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030012740 It is worth mentioning that "use_backend" rules with an explicit name are
Bertrand Jacquin702d44f2013-11-19 11:43:06 +010012741 used to detect the association between frontends and backends to compute the
12742 backend's "fullconn" setting. This cannot be done for dynamic names.
12743
12744 See also: "default_backend", "tcp-request", "fullconn", "log-format", and
12745 section 7 about ACLs.
Willy Tarreaud72758d2010-01-12 10:42:19 +010012746
Christopher Fauletb30b3102019-09-12 23:03:09 +020012747use-fcgi-app <name>
12748 Defines the FastCGI application to use for the backend.
12749 May be used in sections : defaults | frontend | listen | backend
12750 no | no | yes | yes
12751 Arguments :
12752 <name> is the name of the FastCGI application to use.
12753
12754 See section 10.1 about FastCGI application setup for details.
Willy Tarreau036fae02008-01-06 13:24:40 +010012755
Willy Tarreau4a5cade2012-04-05 21:09:48 +020012756use-server <server> if <condition>
12757use-server <server> unless <condition>
12758 Only use a specific server if/unless an ACL-based condition is matched.
12759 May be used in sections : defaults | frontend | listen | backend
12760 no | no | yes | yes
12761 Arguments :
Jerome Magnin824186b2020-03-29 09:37:12 +020012762 <server> is the name of a valid server in the same backend section
12763 or a "log-format" string resolving to a server name.
Willy Tarreau4a5cade2012-04-05 21:09:48 +020012764
12765 <condition> is a condition composed of ACLs, as described in section 7.
12766
12767 By default, connections which arrive to a backend are load-balanced across
12768 the available servers according to the configured algorithm, unless a
12769 persistence mechanism such as a cookie is used and found in the request.
12770
12771 Sometimes it is desirable to forward a particular request to a specific
12772 server without having to declare a dedicated backend for this server. This
12773 can be achieved using the "use-server" rules. These rules are evaluated after
12774 the "redirect" rules and before evaluating cookies, and they have precedence
12775 on them. There may be as many "use-server" rules as desired. All of these
12776 rules are evaluated in their declaration order, and the first one which
12777 matches will assign the server.
12778
12779 If a rule designates a server which is down, and "option persist" is not used
12780 and no force-persist rule was validated, it is ignored and evaluation goes on
12781 with the next rules until one matches.
12782
12783 In the first form, the server will be used if the condition is met. In the
12784 second form, the server will be used if the condition is not met. If no
12785 condition is valid, the processing continues and the server will be assigned
12786 according to other persistence mechanisms.
12787
12788 Note that even if a rule is matched, cookie processing is still performed but
12789 does not assign the server. This allows prefixed cookies to have their prefix
12790 stripped.
12791
12792 The "use-server" statement works both in HTTP and TCP mode. This makes it
12793 suitable for use with content-based inspection. For instance, a server could
Lukas Tribusa267b5d2020-07-19 00:25:06 +020012794 be selected in a farm according to the TLS SNI field when using protocols with
12795 implicit TLS (also see "req_ssl_sni"). And if these servers have their weight
12796 set to zero, they will not be used for other traffic.
Willy Tarreau4a5cade2012-04-05 21:09:48 +020012797
12798 Example :
12799 # intercept incoming TLS requests based on the SNI field
12800 use-server www if { req_ssl_sni -i www.example.com }
12801 server www 192.168.0.1:443 weight 0
12802 use-server mail if { req_ssl_sni -i mail.example.com }
Lukas Tribusa267b5d2020-07-19 00:25:06 +020012803 server mail 192.168.0.1:465 weight 0
Willy Tarreau4a5cade2012-04-05 21:09:48 +020012804 use-server imap if { req_ssl_sni -i imap.example.com }
Lukas Tribus98a3e3f2017-03-26 12:55:35 +000012805 server imap 192.168.0.1:993 weight 0
Willy Tarreau4a5cade2012-04-05 21:09:48 +020012806 # all the rest is forwarded to this server
12807 server default 192.168.0.2:443 check
12808
Jerome Magnin824186b2020-03-29 09:37:12 +020012809 When <server> is a simple name, it is checked against existing servers in the
12810 configuration and an error is reported if the specified server does not exist.
12811 If it is a log-format, no check is performed when parsing the configuration,
12812 and if we can't resolve a valid server name at runtime but the use-server rule
Ilya Shipitsin11057a32020-06-21 21:18:27 +050012813 was conditioned by an ACL returning true, no other use-server rule is applied
Jerome Magnin824186b2020-03-29 09:37:12 +020012814 and we fall back to load balancing.
12815
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030012816 See also: "use_backend", section 5 about server and section 7 about ACLs.
Willy Tarreau4a5cade2012-04-05 21:09:48 +020012817
Willy Tarreaub6205fd2012-09-24 12:27:33 +020012818
Davor Ocelice9ed2812017-12-25 17:49:28 +0100128195. Bind and server options
Willy Tarreaub6205fd2012-09-24 12:27:33 +020012820--------------------------
12821
12822The "bind", "server" and "default-server" keywords support a number of settings
12823depending on some build options and on the system HAProxy was built on. These
12824settings generally each consist in one word sometimes followed by a value,
12825written on the same line as the "bind" or "server" line. All these options are
12826described in this section.
12827
12828
128295.1. Bind options
12830-----------------
12831
12832The "bind" keyword supports a certain number of settings which are all passed
12833as arguments on the same line. The order in which those arguments appear makes
12834no importance, provided that they appear after the bind address. All of these
12835parameters are optional. Some of them consist in a single words (booleans),
12836while other ones expect a value after them. In this case, the value must be
12837provided immediately after the setting name.
12838
12839The currently supported settings are the following ones.
12840
Bertrand Jacquin93b227d2016-06-04 15:11:10 +010012841accept-netscaler-cip <magic number>
12842 Enforces the use of the NetScaler Client IP insertion protocol over any
12843 connection accepted by any of the TCP sockets declared on the same line. The
12844 NetScaler Client IP insertion protocol dictates the layer 3/4 addresses of
12845 the incoming connection to be used everywhere an address is used, with the
12846 only exception of "tcp-request connection" rules which will only see the
12847 real connection address. Logs will reflect the addresses indicated in the
12848 protocol, unless it is violated, in which case the real address will still
12849 be used. This keyword combined with support from external components can be
12850 used as an efficient and reliable alternative to the X-Forwarded-For
Bertrand Jacquin90759682016-06-06 15:35:39 +010012851 mechanism which is not always reliable and not even always usable. See also
12852 "tcp-request connection expect-netscaler-cip" for a finer-grained setting of
12853 which client is allowed to use the protocol.
Bertrand Jacquin93b227d2016-06-04 15:11:10 +010012854
Willy Tarreaub6205fd2012-09-24 12:27:33 +020012855accept-proxy
12856 Enforces the use of the PROXY protocol over any connection accepted by any of
Willy Tarreau77992672014-06-14 11:06:17 +020012857 the sockets declared on the same line. Versions 1 and 2 of the PROXY protocol
12858 are supported and correctly detected. The PROXY protocol dictates the layer
Willy Tarreaub6205fd2012-09-24 12:27:33 +020012859 3/4 addresses of the incoming connection to be used everywhere an address is
12860 used, with the only exception of "tcp-request connection" rules which will
12861 only see the real connection address. Logs will reflect the addresses
12862 indicated in the protocol, unless it is violated, in which case the real
Davor Ocelice9ed2812017-12-25 17:49:28 +010012863 address will still be used. This keyword combined with support from external
Willy Tarreaub6205fd2012-09-24 12:27:33 +020012864 components can be used as an efficient and reliable alternative to the
12865 X-Forwarded-For mechanism which is not always reliable and not even always
Willy Tarreau4f0d9192013-06-11 20:40:55 +020012866 usable. See also "tcp-request connection expect-proxy" for a finer-grained
12867 setting of which client is allowed to use the protocol.
Willy Tarreaub6205fd2012-09-24 12:27:33 +020012868
Olivier Houchardc2aae742017-09-22 18:26:28 +020012869allow-0rtt
Bertrand Jacquina25282b2018-08-14 00:56:13 +010012870 Allow receiving early data when using TLSv1.3. This is disabled by default,
Olivier Houchard69752962019-01-08 15:35:32 +010012871 due to security considerations. Because it is vulnerable to replay attacks,
John Roeslerfb2fce12019-07-10 15:45:51 -050012872 you should only allow if for requests that are safe to replay, i.e. requests
Olivier Houchard69752962019-01-08 15:35:32 +010012873 that are idempotent. You can use the "wait-for-handshake" action for any
12874 request that wouldn't be safe with early data.
Olivier Houchardc2aae742017-09-22 18:26:28 +020012875
Willy Tarreauab861d32013-04-02 02:30:41 +020012876alpn <protocols>
12877 This enables the TLS ALPN extension and advertises the specified protocol
12878 list as supported on top of ALPN. The protocol list consists in a comma-
12879 delimited list of protocol names, for instance: "http/1.1,http/1.0" (without
John Roeslerfb2fce12019-07-10 15:45:51 -050012880 quotes). This requires that the SSL library is built with support for TLS
Willy Tarreauab861d32013-04-02 02:30:41 +020012881 extensions enabled (check with haproxy -vv). The ALPN extension replaces the
Willy Tarreau95c4e142017-11-26 12:18:55 +010012882 initial NPN extension. ALPN is required to enable HTTP/2 on an HTTP frontend.
12883 Versions of OpenSSL prior to 1.0.2 didn't support ALPN and only supposed the
12884 now obsolete NPN extension. At the time of writing this, most browsers still
12885 support both ALPN and NPN for HTTP/2 so a fallback to NPN may still work for
12886 a while. But ALPN must be used whenever possible. If both HTTP/2 and HTTP/1.1
12887 are expected to be supported, both versions can be advertised, in order of
12888 preference, like below :
12889
12890 bind :443 ssl crt pub.pem alpn h2,http/1.1
Willy Tarreauab861d32013-04-02 02:30:41 +020012891
Willy Tarreaub6205fd2012-09-24 12:27:33 +020012892backlog <backlog>
Willy Tarreaue2711c72019-02-27 15:39:41 +010012893 Sets the socket's backlog to this value. If unspecified or 0, the frontend's
Willy Tarreaub6205fd2012-09-24 12:27:33 +020012894 backlog is used instead, which generally defaults to the maxconn value.
12895
Emmanuel Hocdete7f2b732017-01-09 16:15:54 +010012896curves <curves>
12897 This setting is only available when support for OpenSSL was built in. It sets
12898 the string describing the list of elliptic curves algorithms ("curve suite")
12899 that are negotiated during the SSL/TLS handshake with ECDHE. The format of the
12900 string is a colon-delimited list of curve name.
12901 Example: "X25519:P-256" (without quote)
12902 When "curves" is set, "ecdhe" parameter is ignored.
12903
Emeric Brun7fb34422012-09-28 15:26:15 +020012904ecdhe <named curve>
12905 This setting is only available when support for OpenSSL was built in. It sets
Emeric Brun6924ef82013-03-06 14:08:53 +010012906 the named curve (RFC 4492) used to generate ECDH ephemeral keys. By default,
12907 used named curve is prime256v1.
Emeric Brun7fb34422012-09-28 15:26:15 +020012908
Emeric Brunfd33a262012-10-11 16:28:27 +020012909ca-file <cafile>
Emeric Brun1a073b42012-09-28 17:07:34 +020012910 This setting is only available when support for OpenSSL was built in. It
12911 designates a PEM file from which to load CA certificates used to verify
12912 client's certificate.
12913
Emeric Brunb6dc9342012-09-28 17:55:37 +020012914ca-ignore-err [all|<errorID>,...]
12915 This setting is only available when support for OpenSSL was built in.
12916 Sets a comma separated list of errorIDs to ignore during verify at depth > 0.
12917 If set to 'all', all errors are ignored. SSL handshake is not aborted if an
12918 error is ignored.
12919
Christopher Faulet31af49d2015-06-09 17:29:50 +020012920ca-sign-file <cafile>
12921 This setting is only available when support for OpenSSL was built in. It
12922 designates a PEM file containing both the CA certificate and the CA private
12923 key used to create and sign server's certificates. This is a mandatory
12924 setting when the dynamic generation of certificates is enabled. See
12925 'generate-certificates' for details.
12926
Bertrand Jacquind4d0a232016-11-13 16:37:12 +000012927ca-sign-pass <passphrase>
Christopher Faulet31af49d2015-06-09 17:29:50 +020012928 This setting is only available when support for OpenSSL was built in. It is
12929 the CA private key passphrase. This setting is optional and used only when
12930 the dynamic generation of certificates is enabled. See
12931 'generate-certificates' for details.
12932
Emmanuel Hocdet842e94e2019-12-16 16:39:17 +010012933ca-verify-file <cafile>
12934 This setting designates a PEM file from which to load CA certificates used to
12935 verify client's certificate. It designates CA certificates which must not be
12936 included in CA names sent in server hello message. Typically, "ca-file" must
12937 be defined with intermediate certificates, and "ca-verify-file" with
12938 certificates to ending the chain, like root CA.
12939
Willy Tarreaub6205fd2012-09-24 12:27:33 +020012940ciphers <ciphers>
12941 This setting is only available when support for OpenSSL was built in. It sets
12942 the string describing the list of cipher algorithms ("cipher suite") that are
Bertrand Jacquin8cf7c1e2019-02-03 18:35:25 +000012943 negotiated during the SSL/TLS handshake up to TLSv1.2. The format of the
Bertrand Jacquin4f03ab02019-02-03 18:48:49 +000012944 string is defined in "man 1 ciphers" from OpenSSL man pages. For background
Dirkjan Bussink415150f2018-09-14 11:14:21 +020012945 information and recommendations see e.g.
12946 (https://wiki.mozilla.org/Security/Server_Side_TLS) and
12947 (https://mozilla.github.io/server-side-tls/ssl-config-generator/). For TLSv1.3
12948 cipher configuration, please check the "ciphersuites" keyword.
12949
12950ciphersuites <ciphersuites>
12951 This setting is only available when support for OpenSSL was built in and
12952 OpenSSL 1.1.1 or later was used to build HAProxy. It sets the string describing
12953 the list of cipher algorithms ("cipher suite") that are negotiated during the
12954 TLSv1.3 handshake. The format of the string is defined in "man 1 ciphers" from
Bertrand Jacquin4f03ab02019-02-03 18:48:49 +000012955 OpenSSL man pages under the "ciphersuites" section. For cipher configuration
12956 for TLSv1.2 and earlier, please check the "ciphers" keyword.
Willy Tarreaub6205fd2012-09-24 12:27:33 +020012957
Emeric Brunfd33a262012-10-11 16:28:27 +020012958crl-file <crlfile>
Emeric Brun1a073b42012-09-28 17:07:34 +020012959 This setting is only available when support for OpenSSL was built in. It
12960 designates a PEM file from which to load certificate revocation list used
12961 to verify client's certificate.
12962
Willy Tarreaub6205fd2012-09-24 12:27:33 +020012963crt <cert>
Alex Davies0fbf0162013-03-02 16:04:50 +000012964 This setting is only available when support for OpenSSL was built in. It
12965 designates a PEM file containing both the required certificates and any
12966 associated private keys. This file can be built by concatenating multiple
12967 PEM files into one (e.g. cat cert.pem key.pem > combined.pem). If your CA
12968 requires an intermediate certificate, this can also be concatenated into this
Emmanuel Hocdet70df7bf2019-01-04 11:08:20 +010012969 file. Intermediate certificate can also be shared in a directory via
12970 "issuers-chain-path" directive.
Alex Davies0fbf0162013-03-02 16:04:50 +000012971
William Lallemand4c5adbf2020-02-24 14:23:22 +010012972 If the file does not contain a private key, HAProxy will try to load
12973 the key at the same path suffixed by a ".key".
12974
Alex Davies0fbf0162013-03-02 16:04:50 +000012975 If the OpenSSL used supports Diffie-Hellman, parameters present in this file
12976 are loaded.
12977
12978 If a directory name is used instead of a PEM file, then all files found in
William Lallemand3f25ae32020-02-24 16:30:12 +010012979 that directory will be loaded in alphabetic order unless their name ends
12980 with '.key', '.issuer', '.ocsp' or '.sctl' (reserved extensions). This
12981 directive may be specified multiple times in order to load certificates from
12982 multiple files or directories. The certificates will be presented to clients
12983 who provide a valid TLS Server Name Indication field matching one of their
12984 CN or alt subjects. Wildcards are supported, where a wildcard character '*'
12985 is used instead of the first hostname component (e.g. *.example.org matches
Janusz Dziemidowicz2c701b52015-03-07 23:03:59 +010012986 www.example.org but not www.sub.example.org).
Alex Davies0fbf0162013-03-02 16:04:50 +000012987
12988 If no SNI is provided by the client or if the SSL library does not support
12989 TLS extensions, or if the client provides an SNI hostname which does not
12990 match any certificate, then the first loaded certificate will be presented.
12991 This means that when loading certificates from a directory, it is highly
Cyril Bonté3180f7b2015-01-25 00:16:08 +010012992 recommended to load the default one first as a file or to ensure that it will
12993 always be the first one in the directory.
Alex Davies0fbf0162013-03-02 16:04:50 +000012994
Emeric Brune032bfa2012-09-28 13:01:45 +020012995 Note that the same cert may be loaded multiple times without side effects.
Willy Tarreaub6205fd2012-09-24 12:27:33 +020012996
Davor Ocelice9ed2812017-12-25 17:49:28 +010012997 Some CAs (such as GoDaddy) offer a drop down list of server types that do not
Alex Davies0fbf0162013-03-02 16:04:50 +000012998 include HAProxy when obtaining a certificate. If this happens be sure to
Davor Ocelice9ed2812017-12-25 17:49:28 +010012999 choose a web server that the CA believes requires an intermediate CA (for
13000 GoDaddy, selection Apache Tomcat will get the correct bundle, but many
Alex Davies0fbf0162013-03-02 16:04:50 +000013001 others, e.g. nginx, result in a wrong bundle that will not work for some
13002 clients).
13003
Emeric Brun4147b2e2014-06-16 18:36:30 +020013004 For each PEM file, haproxy checks for the presence of file at the same path
13005 suffixed by ".ocsp". If such file is found, support for the TLS Certificate
13006 Status Request extension (also known as "OCSP stapling") is automatically
13007 enabled. The content of this file is optional. If not empty, it must contain
13008 a valid OCSP Response in DER format. In order to be valid an OCSP Response
13009 must comply with the following rules: it has to indicate a good status,
13010 it has to be a single response for the certificate of the PEM file, and it
13011 has to be valid at the moment of addition. If these rules are not respected
13012 the OCSP Response is ignored and a warning is emitted. In order to identify
13013 which certificate an OCSP Response applies to, the issuer's certificate is
13014 necessary. If the issuer's certificate is not found in the PEM file, it will
13015 be loaded from a file at the same path as the PEM file suffixed by ".issuer"
13016 if it exists otherwise it will fail with an error.
13017
Janusz Dziemidowicz2c701b52015-03-07 23:03:59 +010013018 For each PEM file, haproxy also checks for the presence of file at the same
13019 path suffixed by ".sctl". If such file is found, support for Certificate
13020 Transparency (RFC6962) TLS extension is enabled. The file must contain a
13021 valid Signed Certificate Timestamp List, as described in RFC. File is parsed
13022 to check basic syntax, but no signatures are verified.
13023
yanbzhu6c25e9e2016-01-05 12:52:02 -050013024 There are cases where it is desirable to support multiple key types, e.g. RSA
13025 and ECDSA in the cipher suites offered to the clients. This allows clients
13026 that support EC certificates to be able to use EC ciphers, while
13027 simultaneously supporting older, RSA only clients.
yanbzhud19630c2015-12-14 15:10:25 -050013028
William Lallemandf9ff3ec2020-10-02 17:57:44 +020013029 To achieve this, OpenSSL 1.1.1 is required, you can configure this behavior
13030 by providing one crt entry per certificate type, or by configuring a "cert
13031 bundle" like it was required before HAProxy 1.8. See "ssl-load-extra-files".
yanbzhud19630c2015-12-14 15:10:25 -050013032
Emeric Brunb6dc9342012-09-28 17:55:37 +020013033crt-ignore-err <errors>
Alex Davies0fbf0162013-03-02 16:04:50 +000013034 This setting is only available when support for OpenSSL was built in. Sets a
Davor Ocelice9ed2812017-12-25 17:49:28 +010013035 comma separated list of errorIDs to ignore during verify at depth == 0. If
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030013036 set to 'all', all errors are ignored. SSL handshake is not aborted if an error
Alex Davies0fbf0162013-03-02 16:04:50 +000013037 is ignored.
Emeric Brunb6dc9342012-09-28 17:55:37 +020013038
Emmanuel Hocdetfe616562013-01-22 15:31:15 +010013039crt-list <file>
13040 This setting is only available when support for OpenSSL was built in. It
Emmanuel Hocdet98263292016-12-29 18:26:15 +010013041 designates a list of PEM file with an optional ssl configuration and a SNI
13042 filter per certificate, with the following format for each line :
Emmanuel Hocdetfe616562013-01-22 15:31:15 +010013043
Emmanuel Hocdet98263292016-12-29 18:26:15 +010013044 <crtfile> [\[<sslbindconf> ...\]] [[!]<snifilter> ...]
13045
William Lallemand5d036392020-06-30 16:11:36 +020013046 sslbindconf supports "allow-0rtt", "alpn", "ca-file", "ca-verify-file",
13047 "ciphers", "ciphersuites", "crl-file", "curves", "ecdhe", "no-ca-names",
13048 "npn", "verify" configuration. With BoringSSL and Openssl >= 1.1.1
13049 "ssl-min-ver" and "ssl-max-ver" are also supported. It overrides the
13050 configuration set in bind line for the certificate.
Emmanuel Hocdetfe616562013-01-22 15:31:15 +010013051
Emmanuel Hocdet7c41a1b2013-05-07 20:20:06 +020013052 Wildcards are supported in the SNI filter. Negative filter are also supported,
Joao Moraise51fab02020-11-21 07:42:20 -030013053 useful in combination with a wildcard filter to exclude a particular SNI, or
13054 after the first certificate to exclude a pattern from its CN or Subject Alt
13055 Name (SAN). The certificates will be presented to clients who provide a valid
13056 TLS Server Name Indication field matching one of the SNI filters. If no SNI
13057 filter is specified, the CN and SAN are used. This directive may be specified
Emmanuel Hocdet7c41a1b2013-05-07 20:20:06 +020013058 multiple times. See the "crt" option for more information. The default
13059 certificate is still needed to meet OpenSSL expectations. If it is not used,
13060 the 'strict-sni' option may be used.
Emmanuel Hocdetfe616562013-01-22 15:31:15 +010013061
William Lallemandf9ff3ec2020-10-02 17:57:44 +020013062 Multi-cert bundling (see "ssl-load-extra-files") is supported with crt-list,
13063 as long as only the base name is given in the crt-list. SNI filter will do
13064 the same work on all bundled certificates.
yanbzhud19630c2015-12-14 15:10:25 -050013065
William Lallemand7c26ed72020-06-03 17:34:48 +020013066 Empty lines as well as lines beginning with a hash ('#') will be ignored.
13067
Joao Moraisaa8fcc42020-11-24 08:24:30 -030013068 The first declared certificate of a bind line is used as the default
13069 certificate, either from crt or crt-list option, which haproxy should use in
13070 the TLS handshake if no other certificate matches. This certificate will also
13071 be used if the provided SNI matches its CN or SAN, even if a matching SNI
13072 filter is found on any crt-list. The SNI filter !* can be used after the first
13073 declared certificate to not include its CN and SAN in the SNI tree, so it will
13074 never match except if no other certificate matches. This way the first
13075 declared certificate act as a fallback.
Joao Moraise51fab02020-11-21 07:42:20 -030013076
Emmanuel Hocdet98263292016-12-29 18:26:15 +010013077 crt-list file example:
Joao Moraise51fab02020-11-21 07:42:20 -030013078 cert1.pem !*
William Lallemand7c26ed72020-06-03 17:34:48 +020013079 # comment
Emmanuel Hocdet05942112017-02-20 16:11:50 +010013080 cert2.pem [alpn h2,http/1.1]
Emmanuel Hocdet98263292016-12-29 18:26:15 +010013081 certW.pem *.domain.tld !secure.domain.tld
Emmanuel Hocdet05942112017-02-20 16:11:50 +010013082 certS.pem [curves X25519:P-256 ciphers ECDHE-ECDSA-AES256-GCM-SHA384] secure.domain.tld
Emmanuel Hocdet98263292016-12-29 18:26:15 +010013083
Willy Tarreaub6205fd2012-09-24 12:27:33 +020013084defer-accept
13085 Is an optional keyword which is supported only on certain Linux kernels. It
13086 states that a connection will only be accepted once some data arrive on it,
13087 or at worst after the first retransmit. This should be used only on protocols
Davor Ocelice9ed2812017-12-25 17:49:28 +010013088 for which the client talks first (e.g. HTTP). It can slightly improve
Willy Tarreaub6205fd2012-09-24 12:27:33 +020013089 performance by ensuring that most of the request is already available when
13090 the connection is accepted. On the other hand, it will not be able to detect
13091 connections which don't talk. It is important to note that this option is
13092 broken in all kernels up to 2.6.31, as the connection is never accepted until
13093 the client talks. This can cause issues with front firewalls which would see
13094 an established connection while the proxy will only see it in SYN_RECV. This
13095 option is only supported on TCPv4/TCPv6 sockets and ignored by other ones.
13096
William Lallemandf6975e92017-05-26 17:42:10 +020013097expose-fd listeners
13098 This option is only usable with the stats socket. It gives your stats socket
13099 the capability to pass listeners FD to another HAProxy process.
William Lallemande202b1e2017-06-01 17:38:56 +020013100 During a reload with the master-worker mode, the process is automatically
13101 reexecuted adding -x and one of the stats socket with this option.
Davor Ocelice9ed2812017-12-25 17:49:28 +010013102 See also "-x" in the management guide.
William Lallemandf6975e92017-05-26 17:42:10 +020013103
Emeric Brun2cb7ae52012-10-05 14:14:21 +020013104force-sslv3
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030013105 This option enforces use of SSLv3 only on SSL connections instantiated from
Emeric Brun2cb7ae52012-10-05 14:14:21 +020013106 this listener. SSLv3 is generally less expensive than the TLS counterparts
Emeric Brun2c86cbf2014-10-30 15:56:50 +010013107 for high connection rates. This option is also available on global statement
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020013108 "ssl-default-bind-options". See also "ssl-min-ver" and "ssl-max-ver".
Emeric Brun2cb7ae52012-10-05 14:14:21 +020013109
13110force-tlsv10
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030013111 This option enforces use of TLSv1.0 only on SSL connections instantiated from
Emeric Brun2c86cbf2014-10-30 15:56:50 +010013112 this listener. This option is also available on global statement
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020013113 "ssl-default-bind-options". See also "ssl-min-ver" and "ssl-max-ver".
Emeric Brun2cb7ae52012-10-05 14:14:21 +020013114
13115force-tlsv11
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030013116 This option enforces use of TLSv1.1 only on SSL connections instantiated from
Emeric Brun2c86cbf2014-10-30 15:56:50 +010013117 this listener. This option is also available on global statement
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020013118 "ssl-default-bind-options". See also "ssl-min-ver" and "ssl-max-ver".
Emeric Brun2cb7ae52012-10-05 14:14:21 +020013119
13120force-tlsv12
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030013121 This option enforces use of TLSv1.2 only on SSL connections instantiated from
Emeric Brun2c86cbf2014-10-30 15:56:50 +010013122 this listener. This option is also available on global statement
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020013123 "ssl-default-bind-options". See also "ssl-min-ver" and "ssl-max-ver".
Emeric Brun2cb7ae52012-10-05 14:14:21 +020013124
Emmanuel Hocdet42fb9802017-03-30 19:29:39 +020013125force-tlsv13
13126 This option enforces use of TLSv1.3 only on SSL connections instantiated from
13127 this listener. This option is also available on global statement
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020013128 "ssl-default-bind-options". See also "ssl-min-ver" and "ssl-max-ver".
Emmanuel Hocdet42fb9802017-03-30 19:29:39 +020013129
Christopher Faulet31af49d2015-06-09 17:29:50 +020013130generate-certificates
13131 This setting is only available when support for OpenSSL was built in. It
13132 enables the dynamic SSL certificates generation. A CA certificate and its
13133 private key are necessary (see 'ca-sign-file'). When HAProxy is configured as
13134 a transparent forward proxy, SSL requests generate errors because of a common
13135 name mismatch on the certificate presented to the client. With this option
13136 enabled, HAProxy will try to forge a certificate using the SNI hostname
13137 indicated by the client. This is done only if no certificate matches the SNI
13138 hostname (see 'crt-list'). If an error occurs, the default certificate is
13139 used, else the 'strict-sni' option is set.
13140 It can also be used when HAProxy is configured as a reverse proxy to ease the
13141 deployment of an architecture with many backends.
13142
13143 Creating a SSL certificate is an expensive operation, so a LRU cache is used
13144 to store forged certificates (see 'tune.ssl.ssl-ctx-cache-size'). It
Davor Ocelice9ed2812017-12-25 17:49:28 +010013145 increases the HAProxy's memory footprint to reduce latency when the same
Christopher Faulet31af49d2015-06-09 17:29:50 +020013146 certificate is used many times.
13147
Willy Tarreaub6205fd2012-09-24 12:27:33 +020013148gid <gid>
13149 Sets the group of the UNIX sockets to the designated system gid. It can also
13150 be set by default in the global section's "unix-bind" statement. Note that
13151 some platforms simply ignore this. This setting is equivalent to the "group"
13152 setting except that the group ID is used instead of its name. This setting is
13153 ignored by non UNIX sockets.
13154
13155group <group>
13156 Sets the group of the UNIX sockets to the designated system group. It can
13157 also be set by default in the global section's "unix-bind" statement. Note
13158 that some platforms simply ignore this. This setting is equivalent to the
13159 "gid" setting except that the group name is used instead of its gid. This
13160 setting is ignored by non UNIX sockets.
13161
13162id <id>
13163 Fixes the socket ID. By default, socket IDs are automatically assigned, but
13164 sometimes it is more convenient to fix them to ease monitoring. This value
13165 must be strictly positive and unique within the listener/frontend. This
13166 option can only be used when defining only a single socket.
13167
13168interface <interface>
Lukas Tribusfce2e962013-02-12 22:13:19 +010013169 Restricts the socket to a specific interface. When specified, only packets
13170 received from that particular interface are processed by the socket. This is
13171 currently only supported on Linux. The interface must be a primary system
13172 interface, not an aliased interface. It is also possible to bind multiple
13173 frontends to the same address if they are bound to different interfaces. Note
13174 that binding to a network interface requires root privileges. This parameter
Jérôme Magnin61275192018-02-07 11:39:58 +010013175 is only compatible with TCPv4/TCPv6 sockets. When specified, return traffic
13176 uses the same interface as inbound traffic, and its associated routing table,
13177 even if there are explicit routes through different interfaces configured.
13178 This can prove useful to address asymmetric routing issues when the same
13179 client IP addresses need to be able to reach frontends hosted on different
13180 interfaces.
Willy Tarreaub6205fd2012-09-24 12:27:33 +020013181
Willy Tarreauabb175f2012-09-24 12:43:26 +020013182level <level>
13183 This setting is used with the stats sockets only to restrict the nature of
13184 the commands that can be issued on the socket. It is ignored by other
13185 sockets. <level> can be one of :
Davor Ocelice9ed2812017-12-25 17:49:28 +010013186 - "user" is the least privileged level; only non-sensitive stats can be
Willy Tarreauabb175f2012-09-24 12:43:26 +020013187 read, and no change is allowed. It would make sense on systems where it
13188 is not easy to restrict access to the socket.
13189 - "operator" is the default level and fits most common uses. All data can
Davor Ocelice9ed2812017-12-25 17:49:28 +010013190 be read, and only non-sensitive changes are permitted (e.g. clear max
Willy Tarreauabb175f2012-09-24 12:43:26 +020013191 counters).
Davor Ocelice9ed2812017-12-25 17:49:28 +010013192 - "admin" should be used with care, as everything is permitted (e.g. clear
Willy Tarreauabb175f2012-09-24 12:43:26 +020013193 all counters).
13194
Andjelko Iharosc4df59e2017-07-20 11:59:48 +020013195severity-output <format>
13196 This setting is used with the stats sockets only to configure severity
13197 level output prepended to informational feedback messages. Severity
13198 level of messages can range between 0 and 7, conforming to syslog
13199 rfc5424. Valid and successful socket commands requesting data
13200 (i.e. "show map", "get acl foo" etc.) will never have a severity level
13201 prepended. It is ignored by other sockets. <format> can be one of :
13202 - "none" (default) no severity level is prepended to feedback messages.
13203 - "number" severity level is prepended as a number.
13204 - "string" severity level is prepended as a string following the
13205 rfc5424 convention.
13206
Willy Tarreaub6205fd2012-09-24 12:27:33 +020013207maxconn <maxconn>
13208 Limits the sockets to this number of concurrent connections. Extraneous
13209 connections will remain in the system's backlog until a connection is
13210 released. If unspecified, the limit will be the same as the frontend's
13211 maxconn. Note that in case of port ranges or multiple addresses, the same
13212 value will be applied to each socket. This setting enables different
13213 limitations on expensive sockets, for instance SSL entries which may easily
13214 eat all memory.
13215
13216mode <mode>
13217 Sets the octal mode used to define access permissions on the UNIX socket. It
13218 can also be set by default in the global section's "unix-bind" statement.
13219 Note that some platforms simply ignore this. This setting is ignored by non
13220 UNIX sockets.
13221
13222mss <maxseg>
13223 Sets the TCP Maximum Segment Size (MSS) value to be advertised on incoming
13224 connections. This can be used to force a lower MSS for certain specific
13225 ports, for instance for connections passing through a VPN. Note that this
13226 relies on a kernel feature which is theoretically supported under Linux but
13227 was buggy in all versions prior to 2.6.28. It may or may not work on other
13228 operating systems. It may also not change the advertised value but change the
13229 effective size of outgoing segments. The commonly advertised value for TCPv4
13230 over Ethernet networks is 1460 = 1500(MTU) - 40(IP+TCP). If this value is
13231 positive, it will be used as the advertised MSS. If it is negative, it will
13232 indicate by how much to reduce the incoming connection's advertised MSS for
13233 outgoing segments. This parameter is only compatible with TCP v4/v6 sockets.
13234
13235name <name>
13236 Sets an optional name for these sockets, which will be reported on the stats
13237 page.
13238
Willy Tarreaud72f0f32015-10-13 14:50:22 +020013239namespace <name>
13240 On Linux, it is possible to specify which network namespace a socket will
13241 belong to. This directive makes it possible to explicitly bind a listener to
13242 a namespace different from the default one. Please refer to your operating
13243 system's documentation to find more details about network namespaces.
13244
Willy Tarreaub6205fd2012-09-24 12:27:33 +020013245nice <nice>
13246 Sets the 'niceness' of connections initiated from the socket. Value must be
13247 in the range -1024..1024 inclusive, and defaults to zero. Positive values
13248 means that such connections are more friendly to others and easily offer
13249 their place in the scheduler. On the opposite, negative values mean that
13250 connections want to run with a higher priority than others. The difference
13251 only happens under high loads when the system is close to saturation.
13252 Negative values are appropriate for low-latency or administration services,
13253 and high values are generally recommended for CPU intensive tasks such as SSL
13254 processing or bulk transfers which are less sensible to latency. For example,
13255 it may make sense to use a positive value for an SMTP socket and a negative
13256 one for an RDP socket.
13257
Emmanuel Hocdet174dfe52017-07-28 15:01:05 +020013258no-ca-names
13259 This setting is only available when support for OpenSSL was built in. It
13260 prevents from send CA names in server hello message when ca-file is used.
Emmanuel Hocdet842e94e2019-12-16 16:39:17 +010013261 Use "ca-verify-file" instead of "ca-file" with "no-ca-names".
Emmanuel Hocdet174dfe52017-07-28 15:01:05 +020013262
Emeric Brun9b3009b2012-10-05 11:55:06 +020013263no-sslv3
Willy Tarreaub6205fd2012-09-24 12:27:33 +020013264 This setting is only available when support for OpenSSL was built in. It
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030013265 disables support for SSLv3 on any sockets instantiated from the listener when
Willy Tarreaub6205fd2012-09-24 12:27:33 +020013266 SSL is supported. Note that SSLv2 is forced disabled in the code and cannot
Emeric Brun2c86cbf2014-10-30 15:56:50 +010013267 be enabled using any configuration option. This option is also available on
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020013268 global statement "ssl-default-bind-options". Use "ssl-min-ver" and
13269 "ssl-max-ver" instead.
Willy Tarreaub6205fd2012-09-24 12:27:33 +020013270
Emeric Brun90ad8722012-10-02 14:00:59 +020013271no-tls-tickets
13272 This setting is only available when support for OpenSSL was built in. It
13273 disables the stateless session resumption (RFC 5077 TLS Ticket
13274 extension) and force to use stateful session resumption. Stateless
Emeric Brun2c86cbf2014-10-30 15:56:50 +010013275 session resumption is more expensive in CPU usage. This option is also
13276 available on global statement "ssl-default-bind-options".
Lukas Tribusbdb386d2020-03-10 00:56:09 +010013277 The TLS ticket mechanism is only used up to TLS 1.2.
13278 Forward Secrecy is compromised with TLS tickets, unless ticket keys
13279 are periodically rotated (via reload or by using "tls-ticket-keys").
Emeric Brun90ad8722012-10-02 14:00:59 +020013280
Emeric Brun9b3009b2012-10-05 11:55:06 +020013281no-tlsv10
Willy Tarreaub6205fd2012-09-24 12:27:33 +020013282 This setting is only available when support for OpenSSL was built in. It
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030013283 disables support for TLSv1.0 on any sockets instantiated from the listener
Emeric Brun2cb7ae52012-10-05 14:14:21 +020013284 when SSL is supported. Note that SSLv2 is forced disabled in the code and
Emeric Brun2c86cbf2014-10-30 15:56:50 +010013285 cannot be enabled using any configuration option. This option is also
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020013286 available on global statement "ssl-default-bind-options". Use "ssl-min-ver"
13287 and "ssl-max-ver" instead.
Willy Tarreaub6205fd2012-09-24 12:27:33 +020013288
Emeric Brun9b3009b2012-10-05 11:55:06 +020013289no-tlsv11
Emeric Brunf5da4932012-09-28 19:42:54 +020013290 This setting is only available when support for OpenSSL was built in. It
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030013291 disables support for TLSv1.1 on any sockets instantiated from the listener
Emeric Brun2cb7ae52012-10-05 14:14:21 +020013292 when SSL is supported. Note that SSLv2 is forced disabled in the code and
Emeric Brun2c86cbf2014-10-30 15:56:50 +010013293 cannot be enabled using any configuration option. This option is also
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020013294 available on global statement "ssl-default-bind-options". Use "ssl-min-ver"
13295 and "ssl-max-ver" instead.
Emeric Brunf5da4932012-09-28 19:42:54 +020013296
Emeric Brun9b3009b2012-10-05 11:55:06 +020013297no-tlsv12
Emeric Brunf5da4932012-09-28 19:42:54 +020013298 This setting is only available when support for OpenSSL was built in. It
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030013299 disables support for TLSv1.2 on any sockets instantiated from the listener
Emeric Brun2cb7ae52012-10-05 14:14:21 +020013300 when SSL is supported. Note that SSLv2 is forced disabled in the code and
Emeric Brun2c86cbf2014-10-30 15:56:50 +010013301 cannot be enabled using any configuration option. This option is also
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020013302 available on global statement "ssl-default-bind-options". Use "ssl-min-ver"
13303 and "ssl-max-ver" instead.
Emeric Brunf5da4932012-09-28 19:42:54 +020013304
Emmanuel Hocdet42fb9802017-03-30 19:29:39 +020013305no-tlsv13
13306 This setting is only available when support for OpenSSL was built in. It
13307 disables support for TLSv1.3 on any sockets instantiated from the listener
13308 when SSL is supported. Note that SSLv2 is forced disabled in the code and
13309 cannot be enabled using any configuration option. This option is also
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020013310 available on global statement "ssl-default-bind-options". Use "ssl-min-ver"
13311 and "ssl-max-ver" instead.
Emmanuel Hocdet42fb9802017-03-30 19:29:39 +020013312
Willy Tarreau6c9a3d52012-10-18 18:57:14 +020013313npn <protocols>
13314 This enables the NPN TLS extension and advertises the specified protocol list
13315 as supported on top of NPN. The protocol list consists in a comma-delimited
13316 list of protocol names, for instance: "http/1.1,http/1.0" (without quotes).
John Roeslerfb2fce12019-07-10 15:45:51 -050013317 This requires that the SSL library is built with support for TLS extensions
Willy Tarreauab861d32013-04-02 02:30:41 +020013318 enabled (check with haproxy -vv). Note that the NPN extension has been
Willy Tarreau95c4e142017-11-26 12:18:55 +010013319 replaced with the ALPN extension (see the "alpn" keyword), though this one is
13320 only available starting with OpenSSL 1.0.2. If HTTP/2 is desired on an older
13321 version of OpenSSL, NPN might still be used as most clients still support it
13322 at the time of writing this. It is possible to enable both NPN and ALPN
13323 though it probably doesn't make any sense out of testing.
Willy Tarreau6c9a3d52012-10-18 18:57:14 +020013324
Lukas Tribus53ae85c2017-05-04 15:45:40 +000013325prefer-client-ciphers
13326 Use the client's preference when selecting the cipher suite, by default
13327 the server's preference is enforced. This option is also available on
13328 global statement "ssl-default-bind-options".
Lukas Tribus926594f2018-05-18 17:55:57 +020013329 Note that with OpenSSL >= 1.1.1 ChaCha20-Poly1305 is reprioritized anyway
13330 (without setting this option), if a ChaCha20-Poly1305 cipher is at the top of
13331 the client cipher list.
Lukas Tribus53ae85c2017-05-04 15:45:40 +000013332
Christopher Fauletc644fa92017-11-23 22:44:11 +010013333process <process-set>[/<thread-set>]
Willy Tarreaua36b3242019-02-02 13:14:34 +010013334 This restricts the list of processes or threads on which this listener is
Christopher Fauletc644fa92017-11-23 22:44:11 +010013335 allowed to run. It does not enforce any process but eliminates those which do
Davor Ocelice9ed2812017-12-25 17:49:28 +010013336 not match. If the frontend uses a "bind-process" setting, the intersection
Christopher Fauletc644fa92017-11-23 22:44:11 +010013337 between the two is applied. If in the end the listener is not allowed to run
13338 on any remaining process, a warning is emitted, and the listener will either
13339 run on the first process of the listener if a single process was specified,
13340 or on all of its processes if multiple processes were specified. If a thread
Davor Ocelice9ed2812017-12-25 17:49:28 +010013341 set is specified, it limits the threads allowed to process incoming
Willy Tarreaua36b3242019-02-02 13:14:34 +010013342 connections for this listener, for the the process set. If multiple processes
13343 and threads are configured, a warning is emitted, as it either results from a
13344 configuration error or a misunderstanding of these models. For the unlikely
13345 case where several ranges are needed, this directive may be repeated.
13346 <process-set> and <thread-set> must use the format
Christopher Fauletc644fa92017-11-23 22:44:11 +010013347
13348 all | odd | even | number[-[number]]
13349
13350 Ranges can be partially defined. The higher bound can be omitted. In such
13351 case, it is replaced by the corresponding maximum value. The main purpose of
13352 this directive is to be used with the stats sockets and have one different
13353 socket per process. The second purpose is to have multiple bind lines sharing
13354 the same IP:port but not the same process in a listener, so that the system
13355 can distribute the incoming connections into multiple queues and allow a
13356 smoother inter-process load balancing. Currently Linux 3.9 and above is known
13357 for supporting this. See also "bind-process" and "nbproc".
Willy Tarreau6ae1ba62014-05-07 19:01:58 +020013358
Christopher Fauleta717b992018-04-10 14:43:00 +020013359proto <name>
13360 Forces the multiplexer's protocol to use for the incoming connections. It
13361 must be compatible with the mode of the frontend (TCP or HTTP). It must also
13362 be usable on the frontend side. The list of available protocols is reported
13363 in haproxy -vv.
Daniel Corbett67a82712020-07-06 23:01:19 -040013364 Idea behind this option is to bypass the selection of the best multiplexer's
Christopher Fauleta717b992018-04-10 14:43:00 +020013365 protocol for all connections instantiated from this listening socket. For
Joseph Herlant71b4b152018-11-13 16:55:16 -080013366 instance, it is possible to force the http/2 on clear TCP by specifying "proto
Christopher Fauleta717b992018-04-10 14:43:00 +020013367 h2" on the bind line.
13368
Willy Tarreaub6205fd2012-09-24 12:27:33 +020013369ssl
13370 This setting is only available when support for OpenSSL was built in. It
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030013371 enables SSL deciphering on connections instantiated from this listener. A
Willy Tarreaub6205fd2012-09-24 12:27:33 +020013372 certificate is necessary (see "crt" above). All contents in the buffers will
13373 appear in clear text, so that ACLs and HTTP processing will only have access
Emmanuel Hocdetbd695fe2017-05-15 15:53:41 +020013374 to deciphered contents. SSLv3 is disabled per default, use "ssl-min-ver SSLv3"
13375 to enable it.
Willy Tarreaub6205fd2012-09-24 12:27:33 +020013376
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020013377ssl-max-ver [ SSLv3 | TLSv1.0 | TLSv1.1 | TLSv1.2 | TLSv1.3 ]
13378 This option enforces use of <version> or lower on SSL connections instantiated
William Lallemand50df1cb2020-06-02 10:52:24 +020013379 from this listener. Using this setting without "ssl-min-ver" can be
13380 ambiguous because the default ssl-min-ver value could change in future HAProxy
13381 versions. This option is also available on global statement
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020013382 "ssl-default-bind-options". See also "ssl-min-ver".
13383
13384ssl-min-ver [ SSLv3 | TLSv1.0 | TLSv1.1 | TLSv1.2 | TLSv1.3 ]
William Lallemand50df1cb2020-06-02 10:52:24 +020013385 This option enforces use of <version> or upper on SSL connections
13386 instantiated from this listener. The default value is "TLSv1.2". This option
13387 is also available on global statement "ssl-default-bind-options".
13388 See also "ssl-max-ver".
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020013389
Emmanuel Hocdet65623372013-01-24 17:17:15 +010013390strict-sni
13391 This setting is only available when support for OpenSSL was built in. The
13392 SSL/TLS negotiation is allow only if the client provided an SNI which match
13393 a certificate. The default certificate is not used.
13394 See the "crt" option for more information.
13395
Willy Tarreau2af207a2015-02-04 00:45:58 +010013396tcp-ut <delay>
Tim Düsterhus4896c442016-11-29 02:15:19 +010013397 Sets the TCP User Timeout for all incoming connections instantiated from this
Willy Tarreau2af207a2015-02-04 00:45:58 +010013398 listening socket. This option is available on Linux since version 2.6.37. It
13399 allows haproxy to configure a timeout for sockets which contain data not
Davor Ocelice9ed2812017-12-25 17:49:28 +010013400 receiving an acknowledgment for the configured delay. This is especially
Willy Tarreau2af207a2015-02-04 00:45:58 +010013401 useful on long-lived connections experiencing long idle periods such as
13402 remote terminals or database connection pools, where the client and server
13403 timeouts must remain high to allow a long period of idle, but where it is
13404 important to detect that the client has disappeared in order to release all
13405 resources associated with its connection (and the server's session). The
13406 argument is a delay expressed in milliseconds by default. This only works
13407 for regular TCP connections, and is ignored for other protocols.
13408
Willy Tarreau1c862c52012-10-05 16:21:00 +020013409tfo
Lukas Tribus0defb902013-02-13 23:35:39 +010013410 Is an optional keyword which is supported only on Linux kernels >= 3.7. It
Willy Tarreau1c862c52012-10-05 16:21:00 +020013411 enables TCP Fast Open on the listening socket, which means that clients which
13412 support this feature will be able to send a request and receive a response
13413 during the 3-way handshake starting from second connection, thus saving one
13414 round-trip after the first connection. This only makes sense with protocols
13415 that use high connection rates and where each round trip matters. This can
13416 possibly cause issues with many firewalls which do not accept data on SYN
13417 packets, so this option should only be enabled once well tested. This option
Lukas Tribus0999f762013-04-02 16:43:24 +020013418 is only supported on TCPv4/TCPv6 sockets and ignored by other ones. You may
13419 need to build HAProxy with USE_TFO=1 if your libc doesn't define
13420 TCP_FASTOPEN.
Willy Tarreau1c862c52012-10-05 16:21:00 +020013421
Nenad Merdanovic188ad3e2015-02-27 19:56:50 +010013422tls-ticket-keys <keyfile>
13423 Sets the TLS ticket keys file to load the keys from. The keys need to be 48
Emeric Brun9e754772019-01-10 17:51:55 +010013424 or 80 bytes long, depending if aes128 or aes256 is used, encoded with base64
13425 with one line per key (ex. openssl rand 80 | openssl base64 -A | xargs echo).
13426 The first key determines the key length used for next keys: you can't mix
13427 aes128 and aes256 keys. Number of keys is specified by the TLS_TICKETS_NO
13428 build option (default 3) and at least as many keys need to be present in
13429 the file. Last TLS_TICKETS_NO keys will be used for decryption and the
13430 penultimate one for encryption. This enables easy key rotation by just
13431 appending new key to the file and reloading the process. Keys must be
13432 periodically rotated (ex. every 12h) or Perfect Forward Secrecy is
13433 compromised. It is also a good idea to keep the keys off any permanent
Nenad Merdanovic188ad3e2015-02-27 19:56:50 +010013434 storage such as hard drives (hint: use tmpfs and don't swap those files).
13435 Lifetime hint can be changed using tune.ssl.timeout.
13436
Willy Tarreaub6205fd2012-09-24 12:27:33 +020013437transparent
13438 Is an optional keyword which is supported only on certain Linux kernels. It
13439 indicates that the addresses will be bound even if they do not belong to the
13440 local machine, and that packets targeting any of these addresses will be
13441 intercepted just as if the addresses were locally configured. This normally
13442 requires that IP forwarding is enabled. Caution! do not use this with the
13443 default address '*', as it would redirect any traffic for the specified port.
13444 This keyword is available only when HAProxy is built with USE_LINUX_TPROXY=1.
13445 This parameter is only compatible with TCPv4 and TCPv6 sockets, depending on
13446 kernel version. Some distribution kernels include backports of the feature,
13447 so check for support with your vendor.
13448
Willy Tarreau77e3af92012-11-24 15:07:23 +010013449v4v6
13450 Is an optional keyword which is supported only on most recent systems
13451 including Linux kernels >= 2.4.21. It is used to bind a socket to both IPv4
13452 and IPv6 when it uses the default address. Doing so is sometimes necessary
13453 on systems which bind to IPv6 only by default. It has no effect on non-IPv6
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030013454 sockets, and is overridden by the "v6only" option.
Willy Tarreau77e3af92012-11-24 15:07:23 +010013455
Willy Tarreau9b6700f2012-11-24 11:55:28 +010013456v6only
13457 Is an optional keyword which is supported only on most recent systems
13458 including Linux kernels >= 2.4.21. It is used to bind a socket to IPv6 only
13459 when it uses the default address. Doing so is sometimes preferred to doing it
Willy Tarreau77e3af92012-11-24 15:07:23 +010013460 system-wide as it is per-listener. It has no effect on non-IPv6 sockets and
13461 has precedence over the "v4v6" option.
Willy Tarreau9b6700f2012-11-24 11:55:28 +010013462
Willy Tarreaub6205fd2012-09-24 12:27:33 +020013463uid <uid>
13464 Sets the owner of the UNIX sockets to the designated system uid. It can also
13465 be set by default in the global section's "unix-bind" statement. Note that
13466 some platforms simply ignore this. This setting is equivalent to the "user"
13467 setting except that the user numeric ID is used instead of its name. This
13468 setting is ignored by non UNIX sockets.
13469
13470user <user>
13471 Sets the owner of the UNIX sockets to the designated system user. It can also
13472 be set by default in the global section's "unix-bind" statement. Note that
13473 some platforms simply ignore this. This setting is equivalent to the "uid"
13474 setting except that the user name is used instead of its uid. This setting is
13475 ignored by non UNIX sockets.
13476
Emeric Brun1a073b42012-09-28 17:07:34 +020013477verify [none|optional|required]
13478 This setting is only available when support for OpenSSL was built in. If set
13479 to 'none', client certificate is not requested. This is the default. In other
13480 cases, a client certificate is requested. If the client does not provide a
13481 certificate after the request and if 'verify' is set to 'required', then the
13482 handshake is aborted, while it would have succeeded if set to 'optional'. The
Emeric Brunfd33a262012-10-11 16:28:27 +020013483 certificate provided by the client is always verified using CAs from
13484 'ca-file' and optional CRLs from 'crl-file'. On verify failure the handshake
13485 is aborted, regardless of the 'verify' option, unless the error code exactly
13486 matches one of those listed with 'ca-ignore-err' or 'crt-ignore-err'.
Willy Tarreau4a5cade2012-04-05 21:09:48 +020013487
Willy Tarreaub6205fd2012-09-24 12:27:33 +0200134885.2. Server and default-server options
Cyril Bontéf0c60612010-02-06 14:44:47 +010013489------------------------------------
Willy Tarreau6a06a402007-07-15 20:15:28 +020013490
Krzysztof Piotr Oledzkic6df0662010-01-05 16:38:49 +010013491The "server" and "default-server" keywords support a certain number of settings
13492which are all passed as arguments on the server line. The order in which those
13493arguments appear does not count, and they are all optional. Some of those
13494settings are single words (booleans) while others expect one or several values
13495after them. In this case, the values must immediately follow the setting name.
13496Except default-server, all those settings must be specified after the server's
13497address if they are used:
Willy Tarreau6a06a402007-07-15 20:15:28 +020013498
Willy Tarreauc57f0e22009-05-10 13:12:33 +020013499 server <name> <address>[:port] [settings ...]
Krzysztof Piotr Oledzkic6df0662010-01-05 16:38:49 +010013500 default-server [settings ...]
Willy Tarreau6a06a402007-07-15 20:15:28 +020013501
Frédéric Lécailled2376272017-03-21 18:52:12 +010013502Note that all these settings are supported both by "server" and "default-server"
13503keywords, except "id" which is only supported by "server".
13504
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010013505The currently supported settings are the following ones.
Willy Tarreau0ba27502007-12-24 16:55:16 +010013506
Willy Tarreauceb4ac92012-04-28 00:41:46 +020013507addr <ipv4|ipv6>
Willy Tarreauc57f0e22009-05-10 13:12:33 +020013508 Using the "addr" parameter, it becomes possible to use a different IP address
Baptiste Assmann13f83532016-03-06 23:14:36 +010013509 to send health-checks or to probe the agent-check. On some servers, it may be
13510 desirable to dedicate an IP address to specific component able to perform
13511 complex tests which are more suitable to health-checks than the application.
13512 This parameter is ignored if the "check" parameter is not set. See also the
13513 "port" parameter.
Willy Tarreau6a06a402007-07-15 20:15:28 +020013514
Simon Hormand60d6912013-11-25 10:46:36 +090013515agent-check
13516 Enable an auxiliary agent check which is run independently of a regular
Willy Tarreau81f5d942013-12-09 20:51:51 +010013517 health check. An agent health check is performed by making a TCP connection
Willy Tarreau7a0139e2018-12-16 08:42:56 +010013518 to the port set by the "agent-port" parameter and reading an ASCII string
13519 terminated by the first '\r' or '\n' met. The string is made of a series of
13520 words delimited by spaces, tabs or commas in any order, each consisting of :
Simon Hormand60d6912013-11-25 10:46:36 +090013521
Willy Tarreau81f5d942013-12-09 20:51:51 +010013522 - An ASCII representation of a positive integer percentage, e.g. "75%".
Simon Hormand60d6912013-11-25 10:46:36 +090013523 Values in this format will set the weight proportional to the initial
Willy Tarreauc5af3a62014-10-07 15:27:33 +020013524 weight of a server as configured when haproxy starts. Note that a zero
13525 weight is reported on the stats page as "DRAIN" since it has the same
13526 effect on the server (it's removed from the LB farm).
Simon Hormand60d6912013-11-25 10:46:36 +090013527
Davor Ocelice9ed2812017-12-25 17:49:28 +010013528 - The string "maxconn:" followed by an integer (no space between). Values
13529 in this format will set the maxconn of a server. The maximum number of
13530 connections advertised needs to be multiplied by the number of load
13531 balancers and different backends that use this health check to get the
13532 total number of connections the server might receive. Example: maxconn:30
Nenad Merdanovic174dd372016-04-24 23:10:06 +020013533
Willy Tarreau81f5d942013-12-09 20:51:51 +010013534 - The word "ready". This will turn the server's administrative state to the
Davor Ocelice9ed2812017-12-25 17:49:28 +010013535 READY mode, thus canceling any DRAIN or MAINT state
Simon Hormand60d6912013-11-25 10:46:36 +090013536
Willy Tarreau81f5d942013-12-09 20:51:51 +010013537 - The word "drain". This will turn the server's administrative state to the
13538 DRAIN mode, thus it will not accept any new connections other than those
13539 that are accepted via persistence.
Simon Hormand60d6912013-11-25 10:46:36 +090013540
Willy Tarreau81f5d942013-12-09 20:51:51 +010013541 - The word "maint". This will turn the server's administrative state to the
13542 MAINT mode, thus it will not accept any new connections at all, and health
13543 checks will be stopped.
Simon Hormand60d6912013-11-25 10:46:36 +090013544
William Dauchyf8e795c2020-09-26 13:35:51 +020013545 - The words "down", "fail", or "stopped", optionally followed by a
Willy Tarreau81f5d942013-12-09 20:51:51 +010013546 description string after a sharp ('#'). All of these mark the server's
13547 operating state as DOWN, but since the word itself is reported on the stats
13548 page, the difference allows an administrator to know if the situation was
13549 expected or not : the service may intentionally be stopped, may appear up
Davor Ocelice9ed2812017-12-25 17:49:28 +010013550 but fail some validity tests, or may be seen as down (e.g. missing process,
Willy Tarreau81f5d942013-12-09 20:51:51 +010013551 or port not responding).
Simon Hormand60d6912013-11-25 10:46:36 +090013552
Willy Tarreau81f5d942013-12-09 20:51:51 +010013553 - The word "up" sets back the server's operating state as UP if health checks
13554 also report that the service is accessible.
Simon Hormand60d6912013-11-25 10:46:36 +090013555
Willy Tarreau81f5d942013-12-09 20:51:51 +010013556 Parameters which are not advertised by the agent are not changed. For
13557 example, an agent might be designed to monitor CPU usage and only report a
13558 relative weight and never interact with the operating status. Similarly, an
13559 agent could be designed as an end-user interface with 3 radio buttons
13560 allowing an administrator to change only the administrative state. However,
13561 it is important to consider that only the agent may revert its own actions,
13562 so if a server is set to DRAIN mode or to DOWN state using the agent, the
13563 agent must implement the other equivalent actions to bring the service into
13564 operations again.
Simon Hormand60d6912013-11-25 10:46:36 +090013565
Simon Horman2f1f9552013-11-25 10:46:37 +090013566 Failure to connect to the agent is not considered an error as connectivity
13567 is tested by the regular health check which is enabled by the "check"
Willy Tarreau81f5d942013-12-09 20:51:51 +010013568 parameter. Warning though, it is not a good idea to stop an agent after it
13569 reports "down", since only an agent reporting "up" will be able to turn the
13570 server up again. Note that the CLI on the Unix stats socket is also able to
Willy Tarreau989222a2016-01-15 10:26:26 +010013571 force an agent's result in order to work around a bogus agent if needed.
Simon Horman2f1f9552013-11-25 10:46:37 +090013572
Willy Tarreau81f5d942013-12-09 20:51:51 +010013573 Requires the "agent-port" parameter to be set. See also the "agent-inter"
Frédéric Lécailled2376272017-03-21 18:52:12 +010013574 and "no-agent-check" parameters.
Simon Hormand60d6912013-11-25 10:46:36 +090013575
James Brown55f9ff12015-10-21 18:19:05 -070013576agent-send <string>
13577 If this option is specified, haproxy will send the given string (verbatim)
13578 to the agent server upon connection. You could, for example, encode
13579 the backend name into this string, which would enable your agent to send
13580 different responses based on the backend. Make sure to include a '\n' if
13581 you want to terminate your request with a newline.
13582
Simon Hormand60d6912013-11-25 10:46:36 +090013583agent-inter <delay>
13584 The "agent-inter" parameter sets the interval between two agent checks
13585 to <delay> milliseconds. If left unspecified, the delay defaults to 2000 ms.
13586
13587 Just as with every other time-based parameter, it may be entered in any
13588 other explicit unit among { us, ms, s, m, h, d }. The "agent-inter"
13589 parameter also serves as a timeout for agent checks "timeout check" is
13590 not set. In order to reduce "resonance" effects when multiple servers are
13591 hosted on the same hardware, the agent and health checks of all servers
13592 are started with a small time offset between them. It is also possible to
13593 add some random noise in the agent and health checks interval using the
13594 global "spread-checks" keyword. This makes sense for instance when a lot
13595 of backends use the same servers.
13596
13597 See also the "agent-check" and "agent-port" parameters.
13598
Misiek768d8602017-01-09 09:52:43 +010013599agent-addr <addr>
13600 The "agent-addr" parameter sets address for agent check.
13601
13602 You can offload agent-check to another target, so you can make single place
13603 managing status and weights of servers defined in haproxy in case you can't
13604 make self-aware and self-managing services. You can specify both IP or
13605 hostname, it will be resolved.
13606
Simon Hormand60d6912013-11-25 10:46:36 +090013607agent-port <port>
13608 The "agent-port" parameter sets the TCP port used for agent checks.
13609
13610 See also the "agent-check" and "agent-inter" parameters.
13611
Olivier Houchard8cb2d2e2019-05-06 18:58:48 +020013612allow-0rtt
13613 Allow sending early data to the server when using TLS 1.3.
Olivier Houchard22c9b442019-05-06 19:01:04 +020013614 Note that early data will be sent only if the client used early data, or
13615 if the backend uses "retry-on" with the "0rtt-rejected" keyword.
Olivier Houchard8cb2d2e2019-05-06 18:58:48 +020013616
Olivier Houchardc7566002018-11-20 23:33:50 +010013617alpn <protocols>
13618 This enables the TLS ALPN extension and advertises the specified protocol
13619 list as supported on top of ALPN. The protocol list consists in a comma-
13620 delimited list of protocol names, for instance: "http/1.1,http/1.0" (without
John Roeslerfb2fce12019-07-10 15:45:51 -050013621 quotes). This requires that the SSL library is built with support for TLS
Olivier Houchardc7566002018-11-20 23:33:50 +010013622 extensions enabled (check with haproxy -vv). The ALPN extension replaces the
13623 initial NPN extension. ALPN is required to connect to HTTP/2 servers.
13624 Versions of OpenSSL prior to 1.0.2 didn't support ALPN and only supposed the
13625 now obsolete NPN extension.
13626 If both HTTP/2 and HTTP/1.1 are expected to be supported, both versions can
13627 be advertised, in order of preference, like below :
13628
13629 server 127.0.0.1:443 ssl crt pub.pem alpn h2,http/1.1
13630
Willy Tarreauc57f0e22009-05-10 13:12:33 +020013631backup
13632 When "backup" is present on a server line, the server is only used in load
13633 balancing when all other non-backup servers are unavailable. Requests coming
13634 with a persistence cookie referencing the server will always be served
13635 though. By default, only the first operational backup server is used, unless
Frédéric Lécailled2376272017-03-21 18:52:12 +010013636 the "allbackups" option is set in the backend. See also the "no-backup" and
13637 "allbackups" options.
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010013638
Emeric Brunef42d922012-10-11 16:11:36 +020013639ca-file <cafile>
13640 This setting is only available when support for OpenSSL was built in. It
13641 designates a PEM file from which to load CA certificates used to verify
13642 server's certificate.
13643
Willy Tarreauc57f0e22009-05-10 13:12:33 +020013644check
Jerome Magnin90702bc2020-04-26 14:23:04 +020013645 This option enables health checks on a server:
13646 - when not set, no health checking is performed, and the server is always
13647 considered available.
13648 - when set and no other check method is configured, the server is considered
13649 available when a connection can be established at the highest configured
13650 transport layer. This means TCP by default, or SSL/TLS when "ssl" or
13651 "check-ssl" are set, both possibly combined with connection prefixes such
13652 as a PROXY protocol header when "send-proxy" or "check-send-proxy" are
13653 set.
13654 - when set and an application-level health check is defined, the
13655 application-level exchanges are performed on top of the configured
13656 transport layer and the server is considered available if all of the
13657 exchanges succeed.
13658
13659 By default, health checks are performed on the same address and port as
13660 configured on the server, using the same encapsulation parameters (SSL/TLS,
13661 proxy-protocol header, etc... ). It is possible to change the destination
13662 address using "addr" and the port using "port". When done, it is assumed the
13663 server isn't checked on the service port, and configured encapsulation
Ilya Shipitsin4329a9a2020-05-05 21:17:10 +050013664 parameters are not reused. One must explicitly set "check-send-proxy" to send
Jerome Magnin90702bc2020-04-26 14:23:04 +020013665 connection headers, "check-ssl" to use SSL/TLS.
13666
13667 When "sni" or "alpn" are set on the server line, their value is not used for
13668 health checks and one must use "check-sni" or "check-alpn".
13669
13670 The default source address for health check traffic is the same as the one
13671 defined in the backend. It can be changed with the "source" keyword.
13672
13673 The interval between checks can be set using the "inter" keyword, and the
13674 "rise" and "fall" keywords can be used to define how many successful or
13675 failed health checks are required to flag a server available or not
13676 available.
13677
13678 Optional application-level health checks can be configured with "option
13679 httpchk", "option mysql-check" "option smtpchk", "option pgsql-check",
13680 "option ldap-check", or "option redis-check".
13681
13682 Example:
13683 # simple tcp check
13684 backend foo
13685 server s1 192.168.0.1:80 check
13686 # this does a tcp connect + tls handshake
13687 backend foo
13688 server s1 192.168.0.1:443 ssl check
13689 # simple tcp check is enough for check success
13690 backend foo
13691 option tcp-check
13692 tcp-check connect
13693 server s1 192.168.0.1:443 ssl check
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010013694
Willy Tarreau6c16adc2012-10-05 00:04:16 +020013695check-send-proxy
13696 This option forces emission of a PROXY protocol line with outgoing health
13697 checks, regardless of whether the server uses send-proxy or not for the
13698 normal traffic. By default, the PROXY protocol is enabled for health checks
13699 if it is already enabled for normal traffic and if no "port" nor "addr"
13700 directive is present. However, if such a directive is present, the
13701 "check-send-proxy" option needs to be used to force the use of the
13702 protocol. See also the "send-proxy" option for more information.
13703
Olivier Houchard92150142018-12-21 19:47:01 +010013704check-alpn <protocols>
13705 Defines which protocols to advertise with ALPN. The protocol list consists in
13706 a comma-delimited list of protocol names, for instance: "http/1.1,http/1.0"
13707 (without quotes). If it is not set, the server ALPN is used.
13708
Christopher Fauletedc6ed92020-04-23 16:27:59 +020013709check-proto <name>
13710 Forces the multiplexer's protocol to use for the server's health-check
13711 connections. It must be compatible with the health-check type (TCP or
13712 HTTP). It must also be usable on the backend side. The list of available
13713 protocols is reported in haproxy -vv.
Daniel Corbett67a82712020-07-06 23:01:19 -040013714 Idea behind this option is to bypass the selection of the best multiplexer's
Christopher Fauletedc6ed92020-04-23 16:27:59 +020013715 protocol for health-check connections established to this server.
13716 If not defined, the server one will be used, if set.
13717
Jérôme Magninae9bb762018-12-09 16:08:26 +010013718check-sni <sni>
Olivier Houchard9130a962017-10-17 17:33:43 +020013719 This option allows you to specify the SNI to be used when doing health checks
Jérôme Magninae9bb762018-12-09 16:08:26 +010013720 over SSL. It is only possible to use a string to set <sni>. If you want to
13721 set a SNI for proxied traffic, see "sni".
Olivier Houchard9130a962017-10-17 17:33:43 +020013722
Willy Tarreau763a95b2012-10-04 23:15:39 +020013723check-ssl
13724 This option forces encryption of all health checks over SSL, regardless of
13725 whether the server uses SSL or not for the normal traffic. This is generally
13726 used when an explicit "port" or "addr" directive is specified and SSL health
13727 checks are not inherited. It is important to understand that this option
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030013728 inserts an SSL transport layer below the checks, so that a simple TCP connect
Willy Tarreau763a95b2012-10-04 23:15:39 +020013729 check becomes an SSL connect, which replaces the old ssl-hello-chk. The most
13730 common use is to send HTTPS checks by combining "httpchk" with SSL checks.
Davor Ocelice9ed2812017-12-25 17:49:28 +010013731 All SSL settings are common to health checks and traffic (e.g. ciphers).
Frédéric Lécailled2376272017-03-21 18:52:12 +010013732 See the "ssl" option for more information and "no-check-ssl" to disable
13733 this option.
Willy Tarreau763a95b2012-10-04 23:15:39 +020013734
Alexander Liu2a54bb72019-05-22 19:44:48 +080013735check-via-socks4
John Roeslerfb2fce12019-07-10 15:45:51 -050013736 This option enables outgoing health checks using upstream socks4 proxy. By
Alexander Liu2a54bb72019-05-22 19:44:48 +080013737 default, the health checks won't go through socks tunnel even it was enabled
13738 for normal traffic.
13739
Willy Tarreaua0ee1d02012-09-10 09:01:23 +020013740ciphers <ciphers>
Dirkjan Bussink415150f2018-09-14 11:14:21 +020013741 This setting is only available when support for OpenSSL was built in. This
13742 option sets the string describing the list of cipher algorithms that is
13743 negotiated during the SSL/TLS handshake with the server. The format of the
Bertrand Jacquin4f03ab02019-02-03 18:48:49 +000013744 string is defined in "man 1 ciphers" from OpenSSL man pages. For background
13745 information and recommendations see e.g.
13746 (https://wiki.mozilla.org/Security/Server_Side_TLS) and
13747 (https://mozilla.github.io/server-side-tls/ssl-config-generator/). For TLSv1.3
13748 cipher configuration, please check the "ciphersuites" keyword.
Willy Tarreaua0ee1d02012-09-10 09:01:23 +020013749
Dirkjan Bussink415150f2018-09-14 11:14:21 +020013750ciphersuites <ciphersuites>
13751 This setting is only available when support for OpenSSL was built in and
13752 OpenSSL 1.1.1 or later was used to build HAProxy. This option sets the string
13753 describing the list of cipher algorithms that is negotiated during the TLS
13754 1.3 handshake with the server. The format of the string is defined in
Bertrand Jacquin4f03ab02019-02-03 18:48:49 +000013755 "man 1 ciphers" from OpenSSL man pages under the "ciphersuites" section.
13756 For cipher configuration for TLSv1.2 and earlier, please check the "ciphers"
13757 keyword.
Dirkjan Bussink415150f2018-09-14 11:14:21 +020013758
Willy Tarreauc57f0e22009-05-10 13:12:33 +020013759cookie <value>
13760 The "cookie" parameter sets the cookie value assigned to the server to
13761 <value>. This value will be checked in incoming requests, and the first
13762 operational server possessing the same value will be selected. In return, in
13763 cookie insertion or rewrite modes, this value will be assigned to the cookie
13764 sent to the client. There is nothing wrong in having several servers sharing
13765 the same cookie value, and it is in fact somewhat common between normal and
13766 backup servers. See also the "cookie" keyword in backend section.
13767
Emeric Brunef42d922012-10-11 16:11:36 +020013768crl-file <crlfile>
13769 This setting is only available when support for OpenSSL was built in. It
13770 designates a PEM file from which to load certificate revocation list used
13771 to verify server's certificate.
13772
Emeric Bruna7aa3092012-10-26 12:58:00 +020013773crt <cert>
13774 This setting is only available when support for OpenSSL was built in.
13775 It designates a PEM file from which to load both a certificate and the
13776 associated private key. This file can be built by concatenating both PEM
13777 files into one. This certificate will be sent if the server send a client
13778 certificate request.
13779
Willy Tarreau96839092010-03-29 10:02:24 +020013780disabled
13781 The "disabled" keyword starts the server in the "disabled" state. That means
13782 that it is marked down in maintenance mode, and no connection other than the
13783 ones allowed by persist mode will reach it. It is very well suited to setup
13784 new servers, because normal traffic will never reach them, while it is still
13785 possible to test the service by making use of the force-persist mechanism.
Frédéric Lécailled2376272017-03-21 18:52:12 +010013786 See also "enabled" setting.
Willy Tarreau96839092010-03-29 10:02:24 +020013787
Frédéric Lécailled2376272017-03-21 18:52:12 +010013788enabled
13789 This option may be used as 'server' setting to reset any 'disabled'
13790 setting which would have been inherited from 'default-server' directive as
13791 default value.
13792 It may also be used as 'default-server' setting to reset any previous
13793 'default-server' 'disabled' setting.
Willy Tarreau96839092010-03-29 10:02:24 +020013794
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010013795error-limit <count>
Willy Tarreau983e01e2010-01-11 18:42:06 +010013796 If health observing is enabled, the "error-limit" parameter specifies the
13797 number of consecutive errors that triggers event selected by the "on-error"
13798 option. By default it is set to 10 consecutive errors.
Krzysztof Piotr Oledzki97f07b82009-12-15 22:31:24 +010013799
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010013800 See also the "check", "error-limit" and "on-error".
Krzysztof Piotr Oledzki97f07b82009-12-15 22:31:24 +010013801
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010013802fall <count>
Willy Tarreauc57f0e22009-05-10 13:12:33 +020013803 The "fall" parameter states that a server will be considered as dead after
13804 <count> consecutive unsuccessful health checks. This value defaults to 3 if
13805 unspecified. See also the "check", "inter" and "rise" parameters.
13806
Emeric Brun8694b9a2012-10-05 14:39:07 +020013807force-sslv3
13808 This option enforces use of SSLv3 only when SSL is used to communicate with
13809 the server. SSLv3 is generally less expensive than the TLS counterparts for
Emeric Brun2c86cbf2014-10-30 15:56:50 +010013810 high connection rates. This option is also available on global statement
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020013811 "ssl-default-server-options". See also "ssl-min-ver" and ssl-max-ver".
Emeric Brun8694b9a2012-10-05 14:39:07 +020013812
13813force-tlsv10
13814 This option enforces use of TLSv1.0 only when SSL is used to communicate with
Emeric Brun2c86cbf2014-10-30 15:56:50 +010013815 the server. This option is also available on global statement
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020013816 "ssl-default-server-options". See also "ssl-min-ver" and ssl-max-ver".
Emeric Brun8694b9a2012-10-05 14:39:07 +020013817
13818force-tlsv11
13819 This option enforces use of TLSv1.1 only when SSL is used to communicate with
Emeric Brun2c86cbf2014-10-30 15:56:50 +010013820 the server. This option is also available on global statement
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020013821 "ssl-default-server-options". See also "ssl-min-ver" and ssl-max-ver".
Emeric Brun8694b9a2012-10-05 14:39:07 +020013822
13823force-tlsv12
13824 This option enforces use of TLSv1.2 only when SSL is used to communicate with
Emeric Brun2c86cbf2014-10-30 15:56:50 +010013825 the server. This option is also available on global statement
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020013826 "ssl-default-server-options". See also "ssl-min-ver" and ssl-max-ver".
Emeric Brun8694b9a2012-10-05 14:39:07 +020013827
Emmanuel Hocdet42fb9802017-03-30 19:29:39 +020013828force-tlsv13
13829 This option enforces use of TLSv1.3 only when SSL is used to communicate with
13830 the server. This option is also available on global statement
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020013831 "ssl-default-server-options". See also "ssl-min-ver" and ssl-max-ver".
Emmanuel Hocdet42fb9802017-03-30 19:29:39 +020013832
Willy Tarreauc57f0e22009-05-10 13:12:33 +020013833id <value>
Willy Tarreau53fb4ae2009-10-04 23:04:08 +020013834 Set a persistent ID for the server. This ID must be positive and unique for
13835 the proxy. An unused ID will automatically be assigned if unset. The first
13836 assigned value will be 1. This ID is currently only returned in statistics.
Willy Tarreauc57f0e22009-05-10 13:12:33 +020013837
Willy Tarreau6a031d12016-11-07 19:42:35 +010013838init-addr {last | libc | none | <ip>},[...]*
13839 Indicate in what order the server's address should be resolved upon startup
13840 if it uses an FQDN. Attempts are made to resolve the address by applying in
Davor Ocelice9ed2812017-12-25 17:49:28 +010013841 turn each of the methods mentioned in the comma-delimited list. The first
Willy Tarreau6a031d12016-11-07 19:42:35 +010013842 method which succeeds is used. If the end of the list is reached without
13843 finding a working method, an error is thrown. Method "last" suggests to pick
13844 the address which appears in the state file (see "server-state-file"). Method
13845 "libc" uses the libc's internal resolver (gethostbyname() or getaddrinfo()
13846 depending on the operating system and build options). Method "none"
13847 specifically indicates that the server should start without any valid IP
13848 address in a down state. It can be useful to ignore some DNS issues upon
13849 startup, waiting for the situation to get fixed later. Finally, an IP address
13850 (IPv4 or IPv6) may be provided. It can be the currently known address of the
Davor Ocelice9ed2812017-12-25 17:49:28 +010013851 server (e.g. filled by a configuration generator), or the address of a dummy
Willy Tarreau6a031d12016-11-07 19:42:35 +010013852 server used to catch old sessions and present them with a decent error
13853 message for example. When the "first" load balancing algorithm is used, this
13854 IP address could point to a fake server used to trigger the creation of new
13855 instances on the fly. This option defaults to "last,libc" indicating that the
13856 previous address found in the state file (if any) is used first, otherwise
13857 the libc's resolver is used. This ensures continued compatibility with the
Davor Ocelice9ed2812017-12-25 17:49:28 +010013858 historic behavior.
Willy Tarreau6a031d12016-11-07 19:42:35 +010013859
13860 Example:
13861 defaults
13862 # never fail on address resolution
13863 default-server init-addr last,libc,none
13864
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010013865inter <delay>
13866fastinter <delay>
13867downinter <delay>
Willy Tarreauc57f0e22009-05-10 13:12:33 +020013868 The "inter" parameter sets the interval between two consecutive health checks
13869 to <delay> milliseconds. If left unspecified, the delay defaults to 2000 ms.
13870 It is also possible to use "fastinter" and "downinter" to optimize delays
13871 between checks depending on the server state :
13872
Pieter Baauw44fc9df2015-09-17 21:30:46 +020013873 Server state | Interval used
13874 ----------------------------------------+----------------------------------
13875 UP 100% (non-transitional) | "inter"
13876 ----------------------------------------+----------------------------------
13877 Transitionally UP (going down "fall"), | "fastinter" if set,
13878 Transitionally DOWN (going up "rise"), | "inter" otherwise.
13879 or yet unchecked. |
13880 ----------------------------------------+----------------------------------
13881 DOWN 100% (non-transitional) | "downinter" if set,
13882 | "inter" otherwise.
13883 ----------------------------------------+----------------------------------
Willy Tarreaud72758d2010-01-12 10:42:19 +010013884
Willy Tarreauc57f0e22009-05-10 13:12:33 +020013885 Just as with every other time-based parameter, they can be entered in any
13886 other explicit unit among { us, ms, s, m, h, d }. The "inter" parameter also
13887 serves as a timeout for health checks sent to servers if "timeout check" is
13888 not set. In order to reduce "resonance" effects when multiple servers are
Simon Hormand60d6912013-11-25 10:46:36 +090013889 hosted on the same hardware, the agent and health checks of all servers
13890 are started with a small time offset between them. It is also possible to
13891 add some random noise in the agent and health checks interval using the
13892 global "spread-checks" keyword. This makes sense for instance when a lot
13893 of backends use the same servers.
Willy Tarreauc57f0e22009-05-10 13:12:33 +020013894
Emeric Brun97556472020-05-30 01:42:45 +020013895log-proto <logproto>
13896 The "log-proto" specifies the protocol used to forward event messages to
13897 a server configured in a ring section. Possible values are "legacy"
13898 and "octet-count" corresponding respectively to "Non-transparent-framing"
13899 and "Octet counting" in rfc6587. "legacy" is the default.
13900
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010013901maxconn <maxconn>
Willy Tarreauc57f0e22009-05-10 13:12:33 +020013902 The "maxconn" parameter specifies the maximal number of concurrent
13903 connections that will be sent to this server. If the number of incoming
Tim Duesterhuscefbbd92019-11-27 22:35:27 +010013904 concurrent connections goes higher than this value, they will be queued,
13905 waiting for a slot to be released. This parameter is very important as it can
Willy Tarreauc57f0e22009-05-10 13:12:33 +020013906 save fragile servers from going down under extreme loads. If a "minconn"
13907 parameter is specified, the limit becomes dynamic. The default value is "0"
13908 which means unlimited. See also the "minconn" and "maxqueue" parameters, and
13909 the backend's "fullconn" keyword.
13910
Tim Duesterhuscefbbd92019-11-27 22:35:27 +010013911 In HTTP mode this parameter limits the number of concurrent requests instead
13912 of the number of connections. Multiple requests might be multiplexed over a
13913 single TCP connection to the server. As an example if you specify a maxconn
13914 of 50 you might see between 1 and 50 actual server connections, but no more
13915 than 50 concurrent requests.
13916
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010013917maxqueue <maxqueue>
Willy Tarreauc57f0e22009-05-10 13:12:33 +020013918 The "maxqueue" parameter specifies the maximal number of connections which
13919 will wait in the queue for this server. If this limit is reached, next
13920 requests will be redispatched to other servers instead of indefinitely
13921 waiting to be served. This will break persistence but may allow people to
Willy Tarreau8ae8c482020-10-22 17:19:07 +020013922 quickly re-log in when the server they try to connect to is dying. Some load
13923 balancing algorithms such as leastconn take this into account and accept to
13924 add requests into a server's queue up to this value if it is explicitly set
13925 to a value greater than zero, which often allows to better smooth the load
13926 when dealing with single-digit maxconn values. The default value is "0" which
13927 means the queue is unlimited. See also the "maxconn" and "minconn" parameters
13928 and "balance leastconn".
Willy Tarreauc57f0e22009-05-10 13:12:33 +020013929
Willy Tarreau9c538e02019-01-23 10:21:49 +010013930max-reuse <count>
13931 The "max-reuse" argument indicates the HTTP connection processors that they
13932 should not reuse a server connection more than this number of times to send
13933 new requests. Permitted values are -1 (the default), which disables this
13934 limit, or any positive value. Value zero will effectively disable keep-alive.
13935 This is only used to work around certain server bugs which cause them to leak
13936 resources over time. The argument is not necessarily respected by the lower
13937 layers as there might be technical limitations making it impossible to
13938 enforce. At least HTTP/2 connections to servers will respect it.
13939
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010013940minconn <minconn>
Willy Tarreauc57f0e22009-05-10 13:12:33 +020013941 When the "minconn" parameter is set, the maxconn limit becomes a dynamic
13942 limit following the backend's load. The server will always accept at least
13943 <minconn> connections, never more than <maxconn>, and the limit will be on
13944 the ramp between both values when the backend has less than <fullconn>
13945 concurrent connections. This makes it possible to limit the load on the
13946 server during normal loads, but push it further for important loads without
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +010013947 overloading the server during exceptional loads. See also the "maxconn"
Willy Tarreauc57f0e22009-05-10 13:12:33 +020013948 and "maxqueue" parameters, as well as the "fullconn" backend keyword.
Krzysztof Piotr Oledzki97f07b82009-12-15 22:31:24 +010013949
Willy Tarreaud72f0f32015-10-13 14:50:22 +020013950namespace <name>
13951 On Linux, it is possible to specify which network namespace a socket will
13952 belong to. This directive makes it possible to explicitly bind a server to
13953 a namespace different from the default one. Please refer to your operating
13954 system's documentation to find more details about network namespaces.
13955
Frédéric Lécailled2376272017-03-21 18:52:12 +010013956no-agent-check
13957 This option may be used as "server" setting to reset any "agent-check"
13958 setting which would have been inherited from "default-server" directive as
13959 default value.
13960 It may also be used as "default-server" setting to reset any previous
13961 "default-server" "agent-check" setting.
13962
13963no-backup
13964 This option may be used as "server" setting to reset any "backup"
13965 setting which would have been inherited from "default-server" directive as
13966 default value.
13967 It may also be used as "default-server" setting to reset any previous
13968 "default-server" "backup" setting.
13969
13970no-check
13971 This option may be used as "server" setting to reset any "check"
13972 setting which would have been inherited from "default-server" directive as
13973 default value.
13974 It may also be used as "default-server" setting to reset any previous
13975 "default-server" "check" setting.
13976
13977no-check-ssl
13978 This option may be used as "server" setting to reset any "check-ssl"
13979 setting which would have been inherited from "default-server" directive as
13980 default value.
13981 It may also be used as "default-server" setting to reset any previous
13982 "default-server" "check-ssl" setting.
13983
Frédéric Lécailled2376272017-03-21 18:52:12 +010013984no-send-proxy
13985 This option may be used as "server" setting to reset any "send-proxy"
13986 setting which would have been inherited from "default-server" directive as
13987 default value.
13988 It may also be used as "default-server" setting to reset any previous
13989 "default-server" "send-proxy" setting.
13990
13991no-send-proxy-v2
13992 This option may be used as "server" setting to reset any "send-proxy-v2"
13993 setting which would have been inherited from "default-server" directive as
13994 default value.
13995 It may also be used as "default-server" setting to reset any previous
13996 "default-server" "send-proxy-v2" setting.
13997
13998no-send-proxy-v2-ssl
13999 This option may be used as "server" setting to reset any "send-proxy-v2-ssl"
14000 setting which would have been inherited from "default-server" directive as
14001 default value.
14002 It may also be used as "default-server" setting to reset any previous
14003 "default-server" "send-proxy-v2-ssl" setting.
14004
14005no-send-proxy-v2-ssl-cn
14006 This option may be used as "server" setting to reset any "send-proxy-v2-ssl-cn"
14007 setting which would have been inherited from "default-server" directive as
14008 default value.
14009 It may also be used as "default-server" setting to reset any previous
14010 "default-server" "send-proxy-v2-ssl-cn" setting.
14011
14012no-ssl
14013 This option may be used as "server" setting to reset any "ssl"
14014 setting which would have been inherited from "default-server" directive as
14015 default value.
14016 It may also be used as "default-server" setting to reset any previous
14017 "default-server" "ssl" setting.
14018
William Dauchyf6370442020-11-14 19:25:33 +010014019 Note that using `default-server ssl` setting and `no-ssl` on server will
14020 however init SSL connection, so it can be later be enabled through the
14021 runtime API: see `set server` commands in management doc.
14022
Willy Tarreau2a3fb1c2015-02-05 16:47:07 +010014023no-ssl-reuse
14024 This option disables SSL session reuse when SSL is used to communicate with
14025 the server. It will force the server to perform a full handshake for every
14026 new connection. It's probably only useful for benchmarking, troubleshooting,
14027 and for paranoid users.
14028
Emeric Brun9b3009b2012-10-05 11:55:06 +020014029no-sslv3
Willy Tarreaua0ee1d02012-09-10 09:01:23 +020014030 This option disables support for SSLv3 when SSL is used to communicate with
14031 the server. Note that SSLv2 is disabled in the code and cannot be enabled
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020014032 using any configuration option. Use "ssl-min-ver" and "ssl-max-ver" instead.
Willy Tarreaua0ee1d02012-09-10 09:01:23 +020014033
Emmanuel Hocdet6cb2d1e2017-03-30 14:43:31 +020014034 Supported in default-server: No
14035
Emeric Brunf9c5c472012-10-11 15:28:34 +020014036no-tls-tickets
14037 This setting is only available when support for OpenSSL was built in. It
14038 disables the stateless session resumption (RFC 5077 TLS Ticket
14039 extension) and force to use stateful session resumption. Stateless
Emeric Brun2c86cbf2014-10-30 15:56:50 +010014040 session resumption is more expensive in CPU usage for servers. This option
14041 is also available on global statement "ssl-default-server-options".
Lukas Tribusbdb386d2020-03-10 00:56:09 +010014042 The TLS ticket mechanism is only used up to TLS 1.2.
14043 Forward Secrecy is compromised with TLS tickets, unless ticket keys
14044 are periodically rotated (via reload or by using "tls-ticket-keys").
Frédéric Lécailled2376272017-03-21 18:52:12 +010014045 See also "tls-tickets".
Emeric Brunf9c5c472012-10-11 15:28:34 +020014046
Emeric Brun9b3009b2012-10-05 11:55:06 +020014047no-tlsv10
Emeric Brun8694b9a2012-10-05 14:39:07 +020014048 This option disables support for TLSv1.0 when SSL is used to communicate with
Emeric Brunf5da4932012-09-28 19:42:54 +020014049 the server. Note that SSLv2 is disabled in the code and cannot be enabled
14050 using any configuration option. TLSv1 is more expensive than SSLv3 so it
Emeric Brun2c86cbf2014-10-30 15:56:50 +010014051 often makes sense to disable it when communicating with local servers. This
14052 option is also available on global statement "ssl-default-server-options".
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020014053 Use "ssl-min-ver" and "ssl-max-ver" instead.
Willy Tarreau763a95b2012-10-04 23:15:39 +020014054
Emmanuel Hocdet6cb2d1e2017-03-30 14:43:31 +020014055 Supported in default-server: No
14056
Emeric Brun9b3009b2012-10-05 11:55:06 +020014057no-tlsv11
Emeric Brun8694b9a2012-10-05 14:39:07 +020014058 This option disables support for TLSv1.1 when SSL is used to communicate with
Emeric Brunf5da4932012-09-28 19:42:54 +020014059 the server. Note that SSLv2 is disabled in the code and cannot be enabled
14060 using any configuration option. TLSv1 is more expensive than SSLv3 so it
Emeric Brun2c86cbf2014-10-30 15:56:50 +010014061 often makes sense to disable it when communicating with local servers. This
14062 option is also available on global statement "ssl-default-server-options".
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020014063 Use "ssl-min-ver" and "ssl-max-ver" instead.
Willy Tarreau763a95b2012-10-04 23:15:39 +020014064
Emmanuel Hocdet6cb2d1e2017-03-30 14:43:31 +020014065 Supported in default-server: No
14066
Emeric Brun9b3009b2012-10-05 11:55:06 +020014067no-tlsv12
Emeric Brun8694b9a2012-10-05 14:39:07 +020014068 This option disables support for TLSv1.2 when SSL is used to communicate with
Willy Tarreaua0ee1d02012-09-10 09:01:23 +020014069 the server. Note that SSLv2 is disabled in the code and cannot be enabled
14070 using any configuration option. TLSv1 is more expensive than SSLv3 so it
Emeric Brun2c86cbf2014-10-30 15:56:50 +010014071 often makes sense to disable it when communicating with local servers. This
14072 option is also available on global statement "ssl-default-server-options".
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020014073 Use "ssl-min-ver" and "ssl-max-ver" instead.
Emmanuel Hocdet42fb9802017-03-30 19:29:39 +020014074
14075 Supported in default-server: No
14076
14077no-tlsv13
14078 This option disables support for TLSv1.3 when SSL is used to communicate with
14079 the server. Note that SSLv2 is disabled in the code and cannot be enabled
14080 using any configuration option. TLSv1 is more expensive than SSLv3 so it
14081 often makes sense to disable it when communicating with local servers. This
14082 option is also available on global statement "ssl-default-server-options".
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020014083 Use "ssl-min-ver" and "ssl-max-ver" instead.
Willy Tarreaua0ee1d02012-09-10 09:01:23 +020014084
Emmanuel Hocdet6cb2d1e2017-03-30 14:43:31 +020014085 Supported in default-server: No
14086
Frédéric Lécailled2376272017-03-21 18:52:12 +010014087no-verifyhost
14088 This option may be used as "server" setting to reset any "verifyhost"
14089 setting which would have been inherited from "default-server" directive as
14090 default value.
14091 It may also be used as "default-server" setting to reset any previous
14092 "default-server" "verifyhost" setting.
Willy Tarreau763a95b2012-10-04 23:15:39 +020014093
Frédéric Lécaille1b9423d2019-07-04 14:19:06 +020014094no-tfo
14095 This option may be used as "server" setting to reset any "tfo"
14096 setting which would have been inherited from "default-server" directive as
14097 default value.
14098 It may also be used as "default-server" setting to reset any previous
14099 "default-server" "tfo" setting.
14100
Simon Hormanfa461682011-06-25 09:39:49 +090014101non-stick
14102 Never add connections allocated to this sever to a stick-table.
14103 This may be used in conjunction with backup to ensure that
14104 stick-table persistence is disabled for backup servers.
14105
Olivier Houchardc7566002018-11-20 23:33:50 +010014106npn <protocols>
14107 This enables the NPN TLS extension and advertises the specified protocol list
14108 as supported on top of NPN. The protocol list consists in a comma-delimited
14109 list of protocol names, for instance: "http/1.1,http/1.0" (without quotes).
John Roeslerfb2fce12019-07-10 15:45:51 -050014110 This requires that the SSL library is built with support for TLS extensions
Olivier Houchardc7566002018-11-20 23:33:50 +010014111 enabled (check with haproxy -vv). Note that the NPN extension has been
14112 replaced with the ALPN extension (see the "alpn" keyword), though this one is
14113 only available starting with OpenSSL 1.0.2.
14114
Krzysztof Piotr Oledzki97f07b82009-12-15 22:31:24 +010014115observe <mode>
14116 This option enables health adjusting based on observing communication with
14117 the server. By default this functionality is disabled and enabling it also
14118 requires to enable health checks. There are two supported modes: "layer4" and
14119 "layer7". In layer4 mode, only successful/unsuccessful tcp connections are
14120 significant. In layer7, which is only allowed for http proxies, responses
14121 received from server are verified, like valid/wrong http code, unparsable
Willy Tarreau150d1462012-03-10 08:19:02 +010014122 headers, a timeout, etc. Valid status codes include 100 to 499, 501 and 505.
Krzysztof Piotr Oledzki97f07b82009-12-15 22:31:24 +010014123
14124 See also the "check", "on-error" and "error-limit".
14125
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010014126on-error <mode>
Krzysztof Piotr Oledzki97f07b82009-12-15 22:31:24 +010014127 Select what should happen when enough consecutive errors are detected.
14128 Currently, four modes are available:
14129 - fastinter: force fastinter
14130 - fail-check: simulate a failed check, also forces fastinter (default)
14131 - sudden-death: simulate a pre-fatal failed health check, one more failed
14132 check will mark a server down, forces fastinter
14133 - mark-down: mark the server immediately down and force fastinter
14134
14135 See also the "check", "observe" and "error-limit".
14136
Simon Hormane0d1bfb2011-06-21 14:34:58 +090014137on-marked-down <action>
14138 Modify what occurs when a server is marked down.
14139 Currently one action is available:
Justin Karnegeseb2c24a2012-05-24 15:28:52 -070014140 - shutdown-sessions: Shutdown peer sessions. When this setting is enabled,
14141 all connections to the server are immediately terminated when the server
14142 goes down. It might be used if the health check detects more complex cases
14143 than a simple connection status, and long timeouts would cause the service
14144 to remain unresponsive for too long a time. For instance, a health check
14145 might detect that a database is stuck and that there's no chance to reuse
14146 existing connections anymore. Connections killed this way are logged with
14147 a 'D' termination code (for "Down").
Simon Hormane0d1bfb2011-06-21 14:34:58 +090014148
14149 Actions are disabled by default
14150
Justin Karnegeseb2c24a2012-05-24 15:28:52 -070014151on-marked-up <action>
14152 Modify what occurs when a server is marked up.
14153 Currently one action is available:
14154 - shutdown-backup-sessions: Shutdown sessions on all backup servers. This is
14155 done only if the server is not in backup state and if it is not disabled
14156 (it must have an effective weight > 0). This can be used sometimes to force
14157 an active server to take all the traffic back after recovery when dealing
Davor Ocelice9ed2812017-12-25 17:49:28 +010014158 with long sessions (e.g. LDAP, SQL, ...). Doing this can cause more trouble
14159 than it tries to solve (e.g. incomplete transactions), so use this feature
Justin Karnegeseb2c24a2012-05-24 15:28:52 -070014160 with extreme care. Sessions killed because a server comes up are logged
14161 with an 'U' termination code (for "Up").
14162
14163 Actions are disabled by default
14164
Willy Tarreau2f3f4d32020-07-01 07:43:51 +020014165pool-low-conn <max>
14166 Set a low threshold on the number of idling connections for a server, below
14167 which a thread will not try to steal a connection from another thread. This
14168 can be useful to improve CPU usage patterns in scenarios involving many very
14169 fast servers, in order to ensure all threads will keep a few idle connections
14170 all the time instead of letting them accumulate over one thread and migrating
14171 them from thread to thread. Typical values of twice the number of threads
14172 seem to show very good performance already with sub-millisecond response
14173 times. The default is zero, indicating that any idle connection can be used
14174 at any time. It is the recommended setting for normal use. This only applies
14175 to connections that can be shared according to the same principles as those
Willy Tarreau0784db82021-02-19 11:45:22 +010014176 applying to "http-reuse". In case connection sharing between threads would
14177 be disabled via "tune.idle-pool.shared", it can become very important to use
14178 this setting to make sure each thread always has a few connections, or the
14179 connection reuse rate will decrease as thread count increases.
Willy Tarreau2f3f4d32020-07-01 07:43:51 +020014180
Olivier Houchard006e3102018-12-10 18:30:32 +010014181pool-max-conn <max>
14182 Set the maximum number of idling connections for a server. -1 means unlimited
14183 connections, 0 means no idle connections. The default is -1. When idle
14184 connections are enabled, orphaned idle connections which do not belong to any
14185 client session anymore are moved to a dedicated pool so that they remain
14186 usable by future clients. This only applies to connections that can be shared
14187 according to the same principles as those applying to "http-reuse".
14188
Olivier Houchardb7b3faa2018-12-14 18:15:36 +010014189pool-purge-delay <delay>
14190 Sets the delay to start purging idle connections. Each <delay> interval, half
Olivier Houcharda56eebf2019-03-19 16:44:02 +010014191 of the idle connections are closed. 0 means we don't keep any idle connection.
Willy Tarreaufb553652019-06-04 14:06:31 +020014192 The default is 5s.
Olivier Houchardb7b3faa2018-12-14 18:15:36 +010014193
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010014194port <port>
Willy Tarreauc57f0e22009-05-10 13:12:33 +020014195 Using the "port" parameter, it becomes possible to use a different port to
William Dauchy4858fb22021-02-03 22:30:09 +010014196 send health-checks or to probe the agent-check. On some servers, it may be
14197 desirable to dedicate a port to a specific component able to perform complex
14198 tests which are more suitable to health-checks than the application. It is
14199 common to run a simple script in inetd for instance. This parameter is
14200 ignored if the "check" parameter is not set. See also the "addr" parameter.
Willy Tarreauc57f0e22009-05-10 13:12:33 +020014201
Christopher Faulet8ed0a3e2018-04-10 14:45:45 +020014202proto <name>
Christopher Faulet8ed0a3e2018-04-10 14:45:45 +020014203 Forces the multiplexer's protocol to use for the outgoing connections to this
14204 server. It must be compatible with the mode of the backend (TCP or HTTP). It
14205 must also be usable on the backend side. The list of available protocols is
14206 reported in haproxy -vv.
Daniel Corbett67a82712020-07-06 23:01:19 -040014207 Idea behind this option is to bypass the selection of the best multiplexer's
Christopher Faulet8ed0a3e2018-04-10 14:45:45 +020014208 protocol for all connections established to this server.
14209
Willy Tarreauc57f0e22009-05-10 13:12:33 +020014210redir <prefix>
14211 The "redir" parameter enables the redirection mode for all GET and HEAD
14212 requests addressing this server. This means that instead of having HAProxy
14213 forward the request to the server, it will send an "HTTP 302" response with
14214 the "Location" header composed of this prefix immediately followed by the
14215 requested URI beginning at the leading '/' of the path component. That means
14216 that no trailing slash should be used after <prefix>. All invalid requests
14217 will be rejected, and all non-GET or HEAD requests will be normally served by
14218 the server. Note that since the response is completely forged, no header
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +010014219 mangling nor cookie insertion is possible in the response. However, cookies in
Davor Ocelice9ed2812017-12-25 17:49:28 +010014220 requests are still analyzed, making this solution completely usable to direct
Willy Tarreauc57f0e22009-05-10 13:12:33 +020014221 users to a remote location in case of local disaster. Main use consists in
14222 increasing bandwidth for static servers by having the clients directly
14223 connect to them. Note: never use a relative location here, it would cause a
14224 loop between the client and HAProxy!
14225
14226 Example : server srv1 192.168.1.1:80 redir http://image1.mydomain.com check
14227
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010014228rise <count>
Willy Tarreauc57f0e22009-05-10 13:12:33 +020014229 The "rise" parameter states that a server will be considered as operational
14230 after <count> consecutive successful health checks. This value defaults to 2
14231 if unspecified. See also the "check", "inter" and "fall" parameters.
14232
Baptiste Assmann8e2d9432018-06-22 15:04:43 +020014233resolve-opts <option>,<option>,...
14234 Comma separated list of options to apply to DNS resolution linked to this
14235 server.
14236
14237 Available options:
14238
14239 * allow-dup-ip
14240 By default, HAProxy prevents IP address duplication in a backend when DNS
14241 resolution at runtime is in operation.
14242 That said, for some cases, it makes sense that two servers (in the same
14243 backend, being resolved by the same FQDN) have the same IP address.
14244 For such case, simply enable this option.
14245 This is the opposite of prevent-dup-ip.
14246
Daniel Corbettf8716912019-11-17 09:48:56 -050014247 * ignore-weight
14248 Ignore any weight that is set within an SRV record. This is useful when
14249 you would like to control the weights using an alternate method, such as
14250 using an "agent-check" or through the runtime api.
14251
Baptiste Assmann8e2d9432018-06-22 15:04:43 +020014252 * prevent-dup-ip
14253 Ensure HAProxy's default behavior is enforced on a server: prevent re-using
14254 an IP address already set to a server in the same backend and sharing the
14255 same fqdn.
14256 This is the opposite of allow-dup-ip.
14257
14258 Example:
14259 backend b_myapp
14260 default-server init-addr none resolvers dns
14261 server s1 myapp.example.com:80 check resolve-opts allow-dup-ip
14262 server s2 myapp.example.com:81 check resolve-opts allow-dup-ip
14263
14264 With the option allow-dup-ip set:
14265 * if the nameserver returns a single IP address, then both servers will use
14266 it
14267 * If the nameserver returns 2 IP addresses, then each server will pick up a
14268 different address
14269
14270 Default value: not set
14271
Baptiste Assmann1fa66662015-04-14 00:28:47 +020014272resolve-prefer <family>
14273 When DNS resolution is enabled for a server and multiple IP addresses from
14274 different families are returned, HAProxy will prefer using an IP address
14275 from the family mentioned in the "resolve-prefer" parameter.
14276 Available families: "ipv4" and "ipv6"
14277
Baptiste Assmannc4aabae2015-08-04 22:43:06 +020014278 Default value: ipv6
14279
Olivier Doucetaa1ea8a2016-08-05 17:15:20 +020014280 Example:
14281
14282 server s1 app1.domain.com:80 resolvers mydns resolve-prefer ipv6
Baptiste Assmann1fa66662015-04-14 00:28:47 +020014283
Thierry Fournierac88cfe2016-02-17 22:05:30 +010014284resolve-net <network>[,<network[,...]]
John Roeslerfb2fce12019-07-10 15:45:51 -050014285 This option prioritizes the choice of an ip address matching a network. This is
Thierry Fournierac88cfe2016-02-17 22:05:30 +010014286 useful with clouds to prefer a local ip. In some cases, a cloud high
Tim Düsterhus4896c442016-11-29 02:15:19 +010014287 availability service can be announced with many ip addresses on many
Davor Ocelice9ed2812017-12-25 17:49:28 +010014288 different datacenters. The latency between datacenter is not negligible, so
14289 this patch permits to prefer a local datacenter. If no address matches the
Thierry Fournierac88cfe2016-02-17 22:05:30 +010014290 configured network, another address is selected.
14291
Olivier Doucetaa1ea8a2016-08-05 17:15:20 +020014292 Example:
14293
14294 server s1 app1.domain.com:80 resolvers mydns resolve-net 10.0.0.0/8
Thierry Fournierac88cfe2016-02-17 22:05:30 +010014295
Baptiste Assmann1fa66662015-04-14 00:28:47 +020014296resolvers <id>
14297 Points to an existing "resolvers" section to resolve current server's
14298 hostname.
14299
Olivier Doucetaa1ea8a2016-08-05 17:15:20 +020014300 Example:
14301
14302 server s1 app1.domain.com:80 check resolvers mydns
Baptiste Assmann1fa66662015-04-14 00:28:47 +020014303
Olivier Doucetaa1ea8a2016-08-05 17:15:20 +020014304 See also section 5.3
Baptiste Assmann1fa66662015-04-14 00:28:47 +020014305
Willy Tarreau5ab04ec2011-03-20 10:32:26 +010014306send-proxy
14307 The "send-proxy" parameter enforces use of the PROXY protocol over any
14308 connection established to this server. The PROXY protocol informs the other
14309 end about the layer 3/4 addresses of the incoming connection, so that it can
14310 know the client's address or the public address it accessed to, whatever the
Bertrand Jacquin93b227d2016-06-04 15:11:10 +010014311 upper layer protocol. For connections accepted by an "accept-proxy" or
14312 "accept-netscaler-cip" listener, the advertised address will be used. Only
14313 TCPv4 and TCPv6 address families are supported. Other families such as
14314 Unix sockets, will report an UNKNOWN family. Servers using this option can
14315 fully be chained to another instance of haproxy listening with an
14316 "accept-proxy" setting. This setting must not be used if the server isn't
14317 aware of the protocol. When health checks are sent to the server, the PROXY
14318 protocol is automatically used when this option is set, unless there is an
14319 explicit "port" or "addr" directive, in which case an explicit
14320 "check-send-proxy" directive would also be needed to use the PROXY protocol.
Frédéric Lécailled2376272017-03-21 18:52:12 +010014321 See also the "no-send-proxy" option of this section and "accept-proxy" and
14322 "accept-netscaler-cip" option of the "bind" keyword.
Willy Tarreau5ab04ec2011-03-20 10:32:26 +010014323
David Safb76832014-05-08 23:42:08 -040014324send-proxy-v2
14325 The "send-proxy-v2" parameter enforces use of the PROXY protocol version 2
14326 over any connection established to this server. The PROXY protocol informs
14327 the other end about the layer 3/4 addresses of the incoming connection, so
14328 that it can know the client's address or the public address it accessed to,
Emmanuel Hocdet404d9782017-10-24 10:55:14 +020014329 whatever the upper layer protocol. It also send ALPN information if an alpn
14330 have been negotiated. This setting must not be used if the server isn't aware
14331 of this version of the protocol. See also the "no-send-proxy-v2" option of
14332 this section and send-proxy" option of the "bind" keyword.
David Safb76832014-05-08 23:42:08 -040014333
Emmanuel Hocdetf643b802018-02-01 15:20:32 +010014334proxy-v2-options <option>[,<option>]*
Tim Duesterhuscf6e0c82020-03-13 12:34:24 +010014335 The "proxy-v2-options" parameter add options to send in PROXY protocol
14336 version 2 when "send-proxy-v2" is used. Options available are:
14337
14338 - ssl : See also "send-proxy-v2-ssl".
14339 - cert-cn : See also "send-proxy-v2-ssl-cn".
14340 - ssl-cipher: Name of the used cipher.
14341 - cert-sig : Signature algorithm of the used certificate.
14342 - cert-key : Key algorithm of the used certificate
14343 - authority : Host name value passed by the client (only SNI from a TLS
14344 connection is supported).
14345 - crc32c : Checksum of the PROXYv2 header.
14346 - unique-id : Send a unique ID generated using the frontend's
14347 "unique-id-format" within the PROXYv2 header.
14348 This unique-id is primarily meant for "mode tcp". It can
14349 lead to unexpected results in "mode http", because the
14350 generated unique ID is also used for the first HTTP request
14351 within a Keep-Alive connection.
Emmanuel Hocdetf643b802018-02-01 15:20:32 +010014352
David Safb76832014-05-08 23:42:08 -040014353send-proxy-v2-ssl
14354 The "send-proxy-v2-ssl" parameter enforces use of the PROXY protocol version
14355 2 over any connection established to this server. The PROXY protocol informs
14356 the other end about the layer 3/4 addresses of the incoming connection, so
14357 that it can know the client's address or the public address it accessed to,
14358 whatever the upper layer protocol. In addition, the SSL information extension
14359 of the PROXY protocol is added to the PROXY protocol header. This setting
14360 must not be used if the server isn't aware of this version of the protocol.
Frédéric Lécailled2376272017-03-21 18:52:12 +010014361 See also the "no-send-proxy-v2-ssl" option of this section and the
14362 "send-proxy-v2" option of the "bind" keyword.
David Safb76832014-05-08 23:42:08 -040014363
14364send-proxy-v2-ssl-cn
14365 The "send-proxy-v2-ssl" parameter enforces use of the PROXY protocol version
14366 2 over any connection established to this server. The PROXY protocol informs
14367 the other end about the layer 3/4 addresses of the incoming connection, so
14368 that it can know the client's address or the public address it accessed to,
14369 whatever the upper layer protocol. In addition, the SSL information extension
14370 of the PROXY protocol, along along with the Common Name from the subject of
14371 the client certificate (if any), is added to the PROXY protocol header. This
14372 setting must not be used if the server isn't aware of this version of the
Davor Ocelice9ed2812017-12-25 17:49:28 +010014373 protocol. See also the "no-send-proxy-v2-ssl-cn" option of this section and
14374 the "send-proxy-v2" option of the "bind" keyword.
David Safb76832014-05-08 23:42:08 -040014375
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010014376slowstart <start_time_in_ms>
Willy Tarreauc57f0e22009-05-10 13:12:33 +020014377 The "slowstart" parameter for a server accepts a value in milliseconds which
14378 indicates after how long a server which has just come back up will run at
14379 full speed. Just as with every other time-based parameter, it can be entered
14380 in any other explicit unit among { us, ms, s, m, h, d }. The speed grows
14381 linearly from 0 to 100% during this time. The limitation applies to two
14382 parameters :
14383
14384 - maxconn: the number of connections accepted by the server will grow from 1
14385 to 100% of the usual dynamic limit defined by (minconn,maxconn,fullconn).
14386
14387 - weight: when the backend uses a dynamic weighted algorithm, the weight
14388 grows linearly from 1 to 100%. In this case, the weight is updated at every
14389 health-check. For this reason, it is important that the "inter" parameter
14390 is smaller than the "slowstart", in order to maximize the number of steps.
14391
14392 The slowstart never applies when haproxy starts, otherwise it would cause
14393 trouble to running servers. It only applies when a server has been previously
14394 seen as failed.
14395
Willy Tarreau732eac42015-07-09 11:40:25 +020014396sni <expression>
14397 The "sni" parameter evaluates the sample fetch expression, converts it to a
14398 string and uses the result as the host name sent in the SNI TLS extension to
14399 the server. A typical use case is to send the SNI received from the client in
14400 a bridged HTTPS scenario, using the "ssl_fc_sni" sample fetch for the
Willy Tarreau2ab88672017-07-05 18:23:03 +020014401 expression, though alternatives such as req.hdr(host) can also make sense. If
14402 "verify required" is set (which is the recommended setting), the resulting
Willy Tarreauad92a9a2017-07-28 11:38:41 +020014403 name will also be matched against the server certificate's names. See the
Jérôme Magninb36a6d22018-12-09 16:03:40 +010014404 "verify" directive for more details. If you want to set a SNI for health
14405 checks, see the "check-sni" directive for more details.
Willy Tarreau732eac42015-07-09 11:40:25 +020014406
Willy Tarreauc6f4ce82009-06-10 11:09:37 +020014407source <addr>[:<pl>[-<ph>]] [usesrc { <addr2>[:<port2>] | client | clientip } ]
Willy Tarreaubce70882009-09-07 11:51:47 +020014408source <addr>[:<port>] [usesrc { <addr2>[:<port2>] | hdr_ip(<hdr>[,<occ>]) } ]
Willy Tarreauc6f4ce82009-06-10 11:09:37 +020014409source <addr>[:<pl>[-<ph>]] [interface <name>] ...
Willy Tarreauc57f0e22009-05-10 13:12:33 +020014410 The "source" parameter sets the source address which will be used when
14411 connecting to the server. It follows the exact same parameters and principle
14412 as the backend "source" keyword, except that it only applies to the server
14413 referencing it. Please consult the "source" keyword for details.
14414
Willy Tarreauc6f4ce82009-06-10 11:09:37 +020014415 Additionally, the "source" statement on a server line allows one to specify a
14416 source port range by indicating the lower and higher bounds delimited by a
14417 dash ('-'). Some operating systems might require a valid IP address when a
14418 source port range is specified. It is permitted to have the same IP/range for
14419 several servers. Doing so makes it possible to bypass the maximum of 64k
14420 total concurrent connections. The limit will then reach 64k connections per
14421 server.
14422
Lukas Tribus7d56c6d2016-09-13 09:51:15 +000014423 Since Linux 4.2/libc 2.23 IP_BIND_ADDRESS_NO_PORT is set for connections
14424 specifying the source address without port(s).
14425
Willy Tarreaua0ee1d02012-09-10 09:01:23 +020014426ssl
Willy Tarreau44f65392013-06-25 07:56:20 +020014427 This option enables SSL ciphering on outgoing connections to the server. It
14428 is critical to verify server certificates using "verify" when using SSL to
14429 connect to servers, otherwise the communication is prone to trivial man in
14430 the-middle attacks rendering SSL useless. When this option is used, health
14431 checks are automatically sent in SSL too unless there is a "port" or an
14432 "addr" directive indicating the check should be sent to a different location.
Frédéric Lécailled2376272017-03-21 18:52:12 +010014433 See the "no-ssl" to disable "ssl" option and "check-ssl" option to force
14434 SSL health checks.
Willy Tarreau763a95b2012-10-04 23:15:39 +020014435
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020014436ssl-max-ver [ SSLv3 | TLSv1.0 | TLSv1.1 | TLSv1.2 | TLSv1.3 ]
14437 This option enforces use of <version> or lower when SSL is used to communicate
14438 with the server. This option is also available on global statement
14439 "ssl-default-server-options". See also "ssl-min-ver".
14440
14441ssl-min-ver [ SSLv3 | TLSv1.0 | TLSv1.1 | TLSv1.2 | TLSv1.3 ]
14442 This option enforces use of <version> or upper when SSL is used to communicate
14443 with the server. This option is also available on global statement
14444 "ssl-default-server-options". See also "ssl-max-ver".
14445
Frédéric Lécailled2376272017-03-21 18:52:12 +010014446ssl-reuse
14447 This option may be used as "server" setting to reset any "no-ssl-reuse"
14448 setting which would have been inherited from "default-server" directive as
14449 default value.
14450 It may also be used as "default-server" setting to reset any previous
14451 "default-server" "no-ssl-reuse" setting.
14452
14453stick
14454 This option may be used as "server" setting to reset any "non-stick"
14455 setting which would have been inherited from "default-server" directive as
14456 default value.
14457 It may also be used as "default-server" setting to reset any previous
14458 "default-server" "non-stick" setting.
Willy Tarreaua0ee1d02012-09-10 09:01:23 +020014459
Alexander Liu2a54bb72019-05-22 19:44:48 +080014460socks4 <addr>:<port>
John Roeslerfb2fce12019-07-10 15:45:51 -050014461 This option enables upstream socks4 tunnel for outgoing connections to the
Alexander Liu2a54bb72019-05-22 19:44:48 +080014462 server. Using this option won't force the health check to go via socks4 by
14463 default. You will have to use the keyword "check-via-socks4" to enable it.
14464
Willy Tarreau163d4622015-10-13 16:16:41 +020014465tcp-ut <delay>
14466 Sets the TCP User Timeout for all outgoing connections to this server. This
14467 option is available on Linux since version 2.6.37. It allows haproxy to
14468 configure a timeout for sockets which contain data not receiving an
Davor Ocelice9ed2812017-12-25 17:49:28 +010014469 acknowledgment for the configured delay. This is especially useful on
Willy Tarreau163d4622015-10-13 16:16:41 +020014470 long-lived connections experiencing long idle periods such as remote
14471 terminals or database connection pools, where the client and server timeouts
14472 must remain high to allow a long period of idle, but where it is important to
14473 detect that the server has disappeared in order to release all resources
14474 associated with its connection (and the client's session). One typical use
14475 case is also to force dead server connections to die when health checks are
14476 too slow or during a soft reload since health checks are then disabled. The
14477 argument is a delay expressed in milliseconds by default. This only works for
14478 regular TCP connections, and is ignored for other protocols.
14479
Willy Tarreau034c88c2017-01-23 23:36:45 +010014480tfo
14481 This option enables using TCP fast open when connecting to servers, on
14482 systems that support it (currently only the Linux kernel >= 4.11).
14483 See the "tfo" bind option for more information about TCP fast open.
14484 Please note that when using tfo, you should also use the "conn-failure",
14485 "empty-response" and "response-timeout" keywords for "retry-on", or haproxy
Frédéric Lécaille1b9423d2019-07-04 14:19:06 +020014486 won't be able to retry the connection on failure. See also "no-tfo".
Willy Tarreau034c88c2017-01-23 23:36:45 +010014487
Willy Tarreauc57f0e22009-05-10 13:12:33 +020014488track [<proxy>/]<server>
Willy Tarreau32091232014-05-16 13:52:00 +020014489 This option enables ability to set the current state of the server by tracking
14490 another one. It is possible to track a server which itself tracks another
14491 server, provided that at the end of the chain, a server has health checks
14492 enabled. If <proxy> is omitted the current one is used. If disable-on-404 is
Willy Tarreauc57f0e22009-05-10 13:12:33 +020014493 used, it has to be enabled on both proxies.
14494
Frédéric Lécailled2376272017-03-21 18:52:12 +010014495tls-tickets
14496 This option may be used as "server" setting to reset any "no-tls-tickets"
14497 setting which would have been inherited from "default-server" directive as
14498 default value.
Lukas Tribusbdb386d2020-03-10 00:56:09 +010014499 The TLS ticket mechanism is only used up to TLS 1.2.
14500 Forward Secrecy is compromised with TLS tickets, unless ticket keys
14501 are periodically rotated (via reload or by using "tls-ticket-keys").
Frédéric Lécailled2376272017-03-21 18:52:12 +010014502 It may also be used as "default-server" setting to reset any previous
Bjoern Jacke5ab7eb62020-02-13 14:16:16 +010014503 "default-server" "no-tls-tickets" setting.
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010014504
Emeric Brunef42d922012-10-11 16:11:36 +020014505verify [none|required]
14506 This setting is only available when support for OpenSSL was built in. If set
Emeric Brun850efd52014-01-29 12:24:34 +010014507 to 'none', server certificate is not verified. In the other case, The
Willy Tarreauad92a9a2017-07-28 11:38:41 +020014508 certificate provided by the server is verified using CAs from 'ca-file' and
14509 optional CRLs from 'crl-file' after having checked that the names provided in
Davor Ocelice9ed2812017-12-25 17:49:28 +010014510 the certificate's subject and subjectAlternateNames attributes match either
Willy Tarreauad92a9a2017-07-28 11:38:41 +020014511 the name passed using the "sni" directive, or if not provided, the static
14512 host name passed using the "verifyhost" directive. When no name is found, the
14513 certificate's names are ignored. For this reason, without SNI it's important
14514 to use "verifyhost". On verification failure the handshake is aborted. It is
14515 critically important to verify server certificates when using SSL to connect
14516 to servers, otherwise the communication is prone to trivial man-in-the-middle
14517 attacks rendering SSL totally useless. Unless "ssl_server_verify" appears in
14518 the global section, "verify" is set to "required" by default.
Emeric Brunef42d922012-10-11 16:11:36 +020014519
Evan Broderbe554312013-06-27 00:05:25 -070014520verifyhost <hostname>
14521 This setting is only available when support for OpenSSL was built in, and
Willy Tarreauad92a9a2017-07-28 11:38:41 +020014522 only takes effect if 'verify required' is also specified. This directive sets
14523 a default static hostname to check the server's certificate against when no
14524 SNI was used to connect to the server. If SNI is not used, this is the only
14525 way to enable hostname verification. This static hostname, when set, will
14526 also be used for health checks (which cannot provide an SNI value). If none
14527 of the hostnames in the certificate match the specified hostname, the
14528 handshake is aborted. The hostnames in the server-provided certificate may
14529 include wildcards. See also "verify", "sni" and "no-verifyhost" options.
Evan Broderbe554312013-06-27 00:05:25 -070014530
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010014531weight <weight>
Willy Tarreauc57f0e22009-05-10 13:12:33 +020014532 The "weight" parameter is used to adjust the server's weight relative to
14533 other servers. All servers will receive a load proportional to their weight
14534 relative to the sum of all weights, so the higher the weight, the higher the
Willy Tarreau6704d672009-06-15 10:56:05 +020014535 load. The default weight is 1, and the maximal value is 256. A value of 0
14536 means the server will not participate in load-balancing but will still accept
14537 persistent connections. If this parameter is used to distribute the load
14538 according to server's capacity, it is recommended to start with values which
14539 can both grow and shrink, for instance between 10 and 100 to leave enough
14540 room above and below for later adjustments.
Willy Tarreauc57f0e22009-05-10 13:12:33 +020014541
14542
Cyril Bonté46175dd2015-07-02 22:45:32 +0200145435.3. Server IP address resolution using DNS
14544-------------------------------------------
Baptiste Assmann1fa66662015-04-14 00:28:47 +020014545
Baptiste Assmann62b75b42015-09-09 01:11:36 +020014546HAProxy allows using a host name on the server line to retrieve its IP address
14547using name servers. By default, HAProxy resolves the name when parsing the
Thayne McCombscdbcca92021-01-07 21:24:41 -070014548configuration file, at startup and cache the result for the process's life.
Baptiste Assmann1fa66662015-04-14 00:28:47 +020014549This is not sufficient in some cases, such as in Amazon where a server's IP
14550can change after a reboot or an ELB Virtual IP can change based on current
14551workload.
14552This chapter describes how HAProxy can be configured to process server's name
14553resolution at run time.
14554Whether run time server name resolution has been enable or not, HAProxy will
14555carry on doing the first resolution when parsing the configuration.
14556
14557
Cyril Bonté46175dd2015-07-02 22:45:32 +0200145585.3.1. Global overview
14559----------------------
Baptiste Assmann1fa66662015-04-14 00:28:47 +020014560
14561As we've seen in introduction, name resolution in HAProxy occurs at two
14562different steps of the process life:
14563
14564 1. when starting up, HAProxy parses the server line definition and matches a
14565 host name. It uses libc functions to get the host name resolved. This
14566 resolution relies on /etc/resolv.conf file.
14567
Christopher Faulet67957bd2017-09-27 11:00:59 +020014568 2. at run time, HAProxy performs periodically name resolutions for servers
14569 requiring DNS resolutions.
Baptiste Assmann1fa66662015-04-14 00:28:47 +020014570
14571A few other events can trigger a name resolution at run time:
14572 - when a server's health check ends up in a connection timeout: this may be
14573 because the server has a new IP address. So we need to trigger a name
14574 resolution to know this new IP.
14575
Christopher Faulet67957bd2017-09-27 11:00:59 +020014576When using resolvers, the server name can either be a hostname, or a SRV label.
Davor Ocelice9ed2812017-12-25 17:49:28 +010014577HAProxy considers anything that starts with an underscore as a SRV label. If a
Christopher Faulet67957bd2017-09-27 11:00:59 +020014578SRV label is specified, then the corresponding SRV records will be retrieved
14579from the DNS server, and the provided hostnames will be used. The SRV label
14580will be checked periodically, and if any server are added or removed, haproxy
14581will automatically do the same.
Olivier Houchardecfa18d2017-08-07 17:30:03 +020014582
Baptiste Assmann1fa66662015-04-14 00:28:47 +020014583A few things important to notice:
John Roeslerfb2fce12019-07-10 15:45:51 -050014584 - all the name servers are queried in the meantime. HAProxy will process the
Baptiste Assmann1fa66662015-04-14 00:28:47 +020014585 first valid response.
14586
14587 - a resolution is considered as invalid (NX, timeout, refused), when all the
14588 servers return an error.
14589
14590
Cyril Bonté46175dd2015-07-02 22:45:32 +0200145915.3.2. The resolvers section
14592----------------------------
Baptiste Assmann1fa66662015-04-14 00:28:47 +020014593
14594This section is dedicated to host information related to name resolution in
Christopher Faulet67957bd2017-09-27 11:00:59 +020014595HAProxy. There can be as many as resolvers section as needed. Each section can
14596contain many name servers.
Baptiste Assmann1fa66662015-04-14 00:28:47 +020014597
Baptiste Assmann62b75b42015-09-09 01:11:36 +020014598When multiple name servers are configured in a resolvers section, then HAProxy
14599uses the first valid response. In case of invalid responses, only the last one
14600is treated. Purpose is to give the chance to a slow server to deliver a valid
14601answer after a fast faulty or outdated server.
14602
14603When each server returns a different error type, then only the last error is
Christopher Faulet67957bd2017-09-27 11:00:59 +020014604used by HAProxy. The following processing is applied on this error:
Baptiste Assmann62b75b42015-09-09 01:11:36 +020014605
Christopher Faulet67957bd2017-09-27 11:00:59 +020014606 1. HAProxy retries the same DNS query with a new query type. The A queries are
14607 switch to AAAA or the opposite. SRV queries are not concerned here. Timeout
14608 errors are also excluded.
Baptiste Assmann62b75b42015-09-09 01:11:36 +020014609
Christopher Faulet67957bd2017-09-27 11:00:59 +020014610 2. When the fallback on the query type was done (or not applicable), HAProxy
14611 retries the original DNS query, with the preferred query type.
Baptiste Assmann62b75b42015-09-09 01:11:36 +020014612
Christopher Faulet67957bd2017-09-27 11:00:59 +020014613 3. HAProxy retries previous steps <resolve_retires> times. If no valid
14614 response is received after that, it stops the DNS resolution and reports
14615 the error.
Baptiste Assmann62b75b42015-09-09 01:11:36 +020014616
Christopher Faulet67957bd2017-09-27 11:00:59 +020014617For example, with 2 name servers configured in a resolvers section, the
14618following scenarios are possible:
14619
14620 - First response is valid and is applied directly, second response is
14621 ignored
14622
14623 - First response is invalid and second one is valid, then second response is
14624 applied
14625
14626 - First response is a NX domain and second one a truncated response, then
14627 HAProxy retries the query with a new type
14628
14629 - First response is a NX domain and second one is a timeout, then HAProxy
14630 retries the query with a new type
14631
14632 - Query timed out for both name servers, then HAProxy retries it with the
14633 same query type
Baptiste Assmann62b75b42015-09-09 01:11:36 +020014634
Olivier Houcharda8c6db82017-07-06 18:46:47 +020014635As a DNS server may not answer all the IPs in one DNS request, haproxy keeps
14636a cache of previous answers, an answer will be considered obsolete after
Christopher Faulet67957bd2017-09-27 11:00:59 +020014637<hold obsolete> seconds without the IP returned.
Olivier Houcharda8c6db82017-07-06 18:46:47 +020014638
Baptiste Assmann62b75b42015-09-09 01:11:36 +020014639
Baptiste Assmann1fa66662015-04-14 00:28:47 +020014640resolvers <resolvers id>
Davor Ocelice9ed2812017-12-25 17:49:28 +010014641 Creates a new name server list labeled <resolvers id>
Baptiste Assmann1fa66662015-04-14 00:28:47 +020014642
14643A resolvers section accept the following parameters:
14644
Baptiste Assmann2af08fe2017-08-14 00:13:01 +020014645accepted_payload_size <nb>
Davor Ocelice9ed2812017-12-25 17:49:28 +010014646 Defines the maximum payload size accepted by HAProxy and announced to all the
Christopher Faulet67957bd2017-09-27 11:00:59 +020014647 name servers configured in this resolvers section.
Baptiste Assmann2af08fe2017-08-14 00:13:01 +020014648 <nb> is in bytes. If not set, HAProxy announces 512. (minimal value defined
14649 by RFC 6891)
14650
Emeric Brun4c751952021-03-08 16:41:29 +010014651 Note: the maximum allowed value is 65535. Recommended value for UDP is
14652 4096 and it is not recommended to exceed 8192 except if you are sure
14653 that your system and network can handle this (over 65507 makes no sense
14654 since is the maximum UDP payload size). If you are using only TCP
14655 nameservers to handle huge DNS responses, you should put this value
14656 to the max: 65535.
Baptiste Assmann9d8dbbc2017-08-18 23:35:08 +020014657
Baptiste Assmann1fa66662015-04-14 00:28:47 +020014658nameserver <id> <ip>:<port>
Emeric Brun56fc5d92021-02-12 20:05:45 +010014659 UDP DNS server description:
Baptiste Assmann1fa66662015-04-14 00:28:47 +020014660 <id> : label of the server, should be unique
14661 <ip> : IP address of the server
14662 <port> : port where the DNS service actually runs
14663
Emeric Brun56fc5d92021-02-12 20:05:45 +010014664server <name> <address> [param*]
14665 Used to configure a DNS TCP or stream server. This supports for all
14666 "server" parameters found in 5.2 paragraph. Some of these parameters
14667 are irrelevant for DNS resolving. Note: currently 4 queries are pipelined
14668 on the same connections. A batch of idle connections are removed every
14669 5 seconds. "maxconn" can be configured to limit the amount of those
14670 concurrent connections and TLS should also usable if the server supports.
14671
Ben Draut44e609b2018-05-29 15:40:08 -060014672parse-resolv-conf
14673 Adds all nameservers found in /etc/resolv.conf to this resolvers nameservers
14674 list. Ordered as if each nameserver in /etc/resolv.conf was individually
14675 placed in the resolvers section in place of this directive.
14676
Baptiste Assmann1fa66662015-04-14 00:28:47 +020014677hold <status> <period>
14678 Defines <period> during which the last name resolution should be kept based
14679 on last resolution <status>
Baptiste Assmann987e16d2016-11-02 22:23:31 +010014680 <status> : last name resolution status. Acceptable values are "nx",
Olivier Houcharda8c6db82017-07-06 18:46:47 +020014681 "other", "refused", "timeout", "valid", "obsolete".
Baptiste Assmann1fa66662015-04-14 00:28:47 +020014682 <period> : interval between two successive name resolution when the last
14683 answer was in <status>. It follows the HAProxy time format.
14684 <period> is in milliseconds by default.
14685
Baptiste Assmann686408b2017-08-18 10:15:42 +020014686 Default value is 10s for "valid", 0s for "obsolete" and 30s for others.
Baptiste Assmann1fa66662015-04-14 00:28:47 +020014687
Baptiste Assmann1fa66662015-04-14 00:28:47 +020014688resolve_retries <nb>
14689 Defines the number <nb> of queries to send to resolve a server name before
14690 giving up.
14691 Default value: 3
14692
Baptiste Assmann62b75b42015-09-09 01:11:36 +020014693 A retry occurs on name server timeout or when the full sequence of DNS query
14694 type failover is over and we need to start up from the default ANY query
14695 type.
14696
Baptiste Assmann1fa66662015-04-14 00:28:47 +020014697timeout <event> <time>
14698 Defines timeouts related to name resolution
14699 <event> : the event on which the <time> timeout period applies to.
14700 events available are:
Frédéric Lécaille93d33162019-03-06 09:35:59 +010014701 - resolve : default time to trigger name resolutions when no
14702 other time applied.
Christopher Faulet67957bd2017-09-27 11:00:59 +020014703 Default value: 1s
14704 - retry : time between two DNS queries, when no valid response
Frédéric Lécaille93d33162019-03-06 09:35:59 +010014705 have been received.
Christopher Faulet67957bd2017-09-27 11:00:59 +020014706 Default value: 1s
Baptiste Assmann1fa66662015-04-14 00:28:47 +020014707 <time> : time related to the event. It follows the HAProxy time format.
14708 <time> is expressed in milliseconds.
14709
Olivier Doucetaa1ea8a2016-08-05 17:15:20 +020014710 Example:
Baptiste Assmann1fa66662015-04-14 00:28:47 +020014711
14712 resolvers mydns
14713 nameserver dns1 10.0.0.1:53
14714 nameserver dns2 10.0.0.2:53
Ben Draut44e609b2018-05-29 15:40:08 -060014715 parse-resolv-conf
Baptiste Assmann1fa66662015-04-14 00:28:47 +020014716 resolve_retries 3
Christopher Faulet67957bd2017-09-27 11:00:59 +020014717 timeout resolve 1s
Baptiste Assmann1fa66662015-04-14 00:28:47 +020014718 timeout retry 1s
Baptiste Assmann987e16d2016-11-02 22:23:31 +010014719 hold other 30s
14720 hold refused 30s
14721 hold nx 30s
14722 hold timeout 30s
Baptiste Assmann1fa66662015-04-14 00:28:47 +020014723 hold valid 10s
Olivier Houcharda8c6db82017-07-06 18:46:47 +020014724 hold obsolete 30s
Baptiste Assmann1fa66662015-04-14 00:28:47 +020014725
14726
Christopher Faulet87f1f3d2019-07-18 14:51:20 +0200147276. Cache
14728---------
14729
14730HAProxy provides a cache, which was designed to perform cache on small objects
14731(favicon, css...). This is a minimalist low-maintenance cache which runs in
14732RAM.
14733
14734The cache is based on a memory which is shared between processes and threads,
14735this memory is split in blocks of 1k.
14736
14737If an object is not used anymore, it can be deleted to store a new object
14738independently of its expiration date. The oldest objects are deleted first
14739when we try to allocate a new one.
14740
14741The cache uses a hash of the host header and the URI as the key.
14742
14743It's possible to view the status of a cache using the Unix socket command
14744"show cache" consult section 9.3 "Unix Socket commands" of Management Guide
14745for more details.
14746
14747When an object is delivered from the cache, the server name in the log is
14748replaced by "<CACHE>".
14749
14750
147516.1. Limitation
14752----------------
14753
14754The cache won't store and won't deliver objects in these cases:
14755
14756- If the response is not a 200
Remi Tricot-Le Breton4f730832020-11-26 15:51:50 +010014757- If the response contains a Vary header and either the process-vary option is
14758 disabled, or a currently unmanaged header is specified in the Vary value (only
14759 accept-encoding and referer are managed for now)
Christopher Faulet87f1f3d2019-07-18 14:51:20 +020014760- If the Content-Length + the headers size is greater than "max-object-size"
14761- If the response is not cacheable
Remi Tricot-Le Bretond493bc82020-11-26 15:51:29 +010014762- If the response does not have an explicit expiration time (s-maxage or max-age
14763 Cache-Control directives or Expires header) or a validator (ETag or Last-Modified
14764 headers)
Remi Tricot-Le Breton5853c0c2020-12-10 17:58:43 +010014765- If the process-vary option is enabled and there are already max-secondary-entries
14766 entries with the same primary key as the current response
Remi Tricot-Le Breton6ca89162021-01-07 14:50:51 +010014767- If the process-vary option is enabled and the response has an unknown encoding (not
14768 mentioned in https://www.iana.org/assignments/http-parameters/http-parameters.xhtml)
14769 while varying on the accept-encoding client header
Christopher Faulet87f1f3d2019-07-18 14:51:20 +020014770
14771- If the request is not a GET
14772- If the HTTP version of the request is smaller than 1.1
14773- If the request contains an Authorization header
14774
14775
147766.2. Setup
14777-----------
14778
14779To setup a cache, you must define a cache section and use it in a proxy with
14780the corresponding http-request and response actions.
14781
14782
147836.2.1. Cache section
14784---------------------
14785
14786cache <name>
14787 Declare a cache section, allocate a shared cache memory named <name>, the
14788 size of cache is mandatory.
14789
14790total-max-size <megabytes>
14791 Define the size in RAM of the cache in megabytes. This size is split in
14792 blocks of 1kB which are used by the cache entries. Its maximum value is 4095.
14793
14794max-object-size <bytes>
14795 Define the maximum size of the objects to be cached. Must not be greater than
14796 an half of "total-max-size". If not set, it equals to a 256th of the cache size.
14797 All objects with sizes larger than "max-object-size" will not be cached.
14798
14799max-age <seconds>
Remi Tricot-Le Breton5853c0c2020-12-10 17:58:43 +010014800 Define the maximum expiration duration. The expiration is set as the lowest
Christopher Faulet87f1f3d2019-07-18 14:51:20 +020014801 value between the s-maxage or max-age (in this order) directive in the
14802 Cache-Control response header and this value. The default value is 60
14803 seconds, which means that you can't cache an object more than 60 seconds by
14804 default.
14805
Remi Tricot-Le Bretone6cc5b52020-12-23 18:13:53 +010014806process-vary <on/off>
14807 Enable or disable the processing of the Vary header. When disabled, a response
Remi Tricot-Le Breton754b2422020-11-16 15:56:10 +010014808 containing such a header will never be cached. When enabled, we need to calculate
14809 a preliminary hash for a subset of request headers on all the incoming requests
14810 (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 +010014811 for a given request (see RFC 7234#4.1). The default value is off (disabled).
Remi Tricot-Le Breton754b2422020-11-16 15:56:10 +010014812
Remi Tricot-Le Breton5853c0c2020-12-10 17:58:43 +010014813max-secondary-entries <number>
14814 Define the maximum number of simultaneous secondary entries with the same primary
14815 key in the cache. This needs the vary support to be enabled. Its default value is 10
14816 and should be passed a strictly positive integer.
14817
Christopher Faulet87f1f3d2019-07-18 14:51:20 +020014818
148196.2.2. Proxy section
14820---------------------
14821
14822http-request cache-use <name> [ { if | unless } <condition> ]
14823 Try to deliver a cached object from the cache <name>. This directive is also
14824 mandatory to store the cache as it calculates the cache hash. If you want to
14825 use a condition for both storage and delivering that's a good idea to put it
14826 after this one.
14827
14828http-response cache-store <name> [ { if | unless } <condition> ]
14829 Store an http-response within the cache. The storage of the response headers
14830 is done at this step, which means you can use others http-response actions
14831 to modify headers before or after the storage of the response. This action
14832 is responsible for the setup of the cache storage filter.
14833
14834
14835Example:
14836
14837 backend bck1
14838 mode http
14839
14840 http-request cache-use foobar
14841 http-response cache-store foobar
14842 server srv1 127.0.0.1:80
14843
14844 cache foobar
14845 total-max-size 4
14846 max-age 240
14847
14848
Willy Tarreau74ca5042013-06-11 23:12:07 +0200148497. Using ACLs and fetching samples
14850----------------------------------
14851
Davor Ocelice9ed2812017-12-25 17:49:28 +010014852HAProxy is capable of extracting data from request or response streams, from
Willy Tarreau74ca5042013-06-11 23:12:07 +020014853client or server information, from tables, environmental information etc...
14854The action of extracting such data is called fetching a sample. Once retrieved,
14855these samples may be used for various purposes such as a key to a stick-table,
14856but most common usages consist in matching them against predefined constant
14857data called patterns.
14858
14859
148607.1. ACL basics
14861---------------
Willy Tarreauc57f0e22009-05-10 13:12:33 +020014862
14863The use of Access Control Lists (ACL) provides a flexible solution to perform
14864content switching and generally to take decisions based on content extracted
14865from the request, the response or any environmental status. The principle is
14866simple :
14867
Willy Tarreau74ca5042013-06-11 23:12:07 +020014868 - extract a data sample from a stream, table or the environment
Willy Tarreaue6b11e42013-11-26 19:02:32 +010014869 - optionally apply some format conversion to the extracted sample
Willy Tarreau74ca5042013-06-11 23:12:07 +020014870 - apply one or multiple pattern matching methods on this sample
14871 - perform actions only when a pattern matches the sample
Willy Tarreauc57f0e22009-05-10 13:12:33 +020014872
Willy Tarreau74ca5042013-06-11 23:12:07 +020014873The actions generally consist in blocking a request, selecting a backend, or
14874adding a header.
Willy Tarreauc57f0e22009-05-10 13:12:33 +020014875
14876In order to define a test, the "acl" keyword is used. The syntax is :
14877
Willy Tarreau74ca5042013-06-11 23:12:07 +020014878 acl <aclname> <criterion> [flags] [operator] [<value>] ...
Willy Tarreauc57f0e22009-05-10 13:12:33 +020014879
14880This creates a new ACL <aclname> or completes an existing one with new tests.
14881Those tests apply to the portion of request/response specified in <criterion>
14882and may be adjusted with optional flags [flags]. Some criteria also support
Willy Tarreaue6b11e42013-11-26 19:02:32 +010014883an operator which may be specified before the set of values. Optionally some
14884conversion operators may be applied to the sample, and they will be specified
14885as a comma-delimited list of keywords just after the first keyword. The values
14886are of the type supported by the criterion, and are separated by spaces.
Willy Tarreauc57f0e22009-05-10 13:12:33 +020014887
14888ACL names must be formed from upper and lower case letters, digits, '-' (dash),
14889'_' (underscore) , '.' (dot) and ':' (colon). ACL names are case-sensitive,
14890which means that "my_acl" and "My_Acl" are two different ACLs.
14891
14892There is no enforced limit to the number of ACLs. The unused ones do not affect
14893performance, they just consume a small amount of memory.
14894
Willy Tarreau74ca5042013-06-11 23:12:07 +020014895The criterion generally is the name of a sample fetch method, or one of its ACL
14896specific declinations. The default test method is implied by the output type of
14897this sample fetch method. The ACL declinations can describe alternate matching
Willy Tarreaue6b11e42013-11-26 19:02:32 +010014898methods of a same sample fetch method. The sample fetch methods are the only
14899ones supporting a conversion.
Willy Tarreau74ca5042013-06-11 23:12:07 +020014900
14901Sample fetch methods return data which can be of the following types :
14902 - boolean
14903 - integer (signed or unsigned)
14904 - IPv4 or IPv6 address
14905 - string
14906 - data block
14907
Willy Tarreaue6b11e42013-11-26 19:02:32 +010014908Converters transform any of these data into any of these. For example, some
14909converters might convert a string to a lower-case string while other ones
14910would turn a string to an IPv4 address, or apply a netmask to an IP address.
14911The resulting sample is of the type of the last converter applied to the list,
14912which defaults to the type of the sample fetch method.
14913
Thierry FOURNIER2a06e392014-05-11 15:49:55 +020014914Each sample or converter returns data of a specific type, specified with its
14915keyword in this documentation. When an ACL is declared using a standard sample
14916fetch method, certain types automatically involved a default matching method
14917which are summarized in the table below :
14918
14919 +---------------------+-----------------+
14920 | Sample or converter | Default |
14921 | output type | matching method |
14922 +---------------------+-----------------+
14923 | boolean | bool |
14924 +---------------------+-----------------+
14925 | integer | int |
14926 +---------------------+-----------------+
14927 | ip | ip |
14928 +---------------------+-----------------+
14929 | string | str |
14930 +---------------------+-----------------+
14931 | binary | none, use "-m" |
14932 +---------------------+-----------------+
14933
14934Note that in order to match a binary samples, it is mandatory to specify a
14935matching method, see below.
14936
Willy Tarreau74ca5042013-06-11 23:12:07 +020014937The ACL engine can match these types against patterns of the following types :
14938 - boolean
14939 - integer or integer range
14940 - IP address / network
14941 - string (exact, substring, suffix, prefix, subdir, domain)
14942 - regular expression
14943 - hex block
14944
Willy Tarreauc57f0e22009-05-10 13:12:33 +020014945The following ACL flags are currently supported :
14946
Willy Tarreau2b5285d2010-05-09 23:45:24 +020014947 -i : ignore case during matching of all subsequent patterns.
14948 -f : load patterns from a file.
Willy Tarreau74ca5042013-06-11 23:12:07 +020014949 -m : use a specific pattern matching method
Thierry FOURNIERb7729c92014-02-11 16:24:41 +010014950 -n : forbid the DNS resolutions
Thierry FOURNIER9860c412014-01-29 14:23:29 +010014951 -M : load the file pointed by -f like a map file.
Thierry FOURNIER3534d882014-01-20 17:01:44 +010014952 -u : force the unique id of the ACL
Willy Tarreau6a06a402007-07-15 20:15:28 +020014953 -- : force end of flags. Useful when a string looks like one of the flags.
14954
Willy Tarreau74ca5042013-06-11 23:12:07 +020014955The "-f" flag is followed by the name of a file from which all lines will be
14956read as individual values. It is even possible to pass multiple "-f" arguments
14957if the patterns are to be loaded from multiple files. Empty lines as well as
14958lines beginning with a sharp ('#') will be ignored. All leading spaces and tabs
14959will be stripped. If it is absolutely necessary to insert a valid pattern
14960beginning with a sharp, just prefix it with a space so that it is not taken for
14961a comment. Depending on the data type and match method, haproxy may load the
14962lines into a binary tree, allowing very fast lookups. This is true for IPv4 and
14963exact string matching. In this case, duplicates will automatically be removed.
14964
Thierry FOURNIER9860c412014-01-29 14:23:29 +010014965The "-M" flag allows an ACL to use a map file. If this flag is set, the file is
14966parsed as two column file. The first column contains the patterns used by the
14967ACL, and the second column contain the samples. The sample can be used later by
14968a map. This can be useful in some rare cases where an ACL would just be used to
14969check for the existence of a pattern in a map before a mapping is applied.
14970
Thierry FOURNIER3534d882014-01-20 17:01:44 +010014971The "-u" flag forces the unique id of the ACL. This unique id is used with the
14972socket interface to identify ACL and dynamically change its values. Note that a
14973file is always identified by its name even if an id is set.
14974
Willy Tarreau74ca5042013-06-11 23:12:07 +020014975Also, note that the "-i" flag applies to subsequent entries and not to entries
14976loaded from files preceding it. For instance :
14977
14978 acl valid-ua hdr(user-agent) -f exact-ua.lst -i -f generic-ua.lst test
14979
14980In this example, each line of "exact-ua.lst" will be exactly matched against
14981the "user-agent" header of the request. Then each line of "generic-ua" will be
14982case-insensitively matched. Then the word "test" will be insensitively matched
14983as well.
14984
14985The "-m" flag is used to select a specific pattern matching method on the input
14986sample. All ACL-specific criteria imply a pattern matching method and generally
14987do not need this flag. However, this flag is useful with generic sample fetch
14988methods to describe how they're going to be matched against the patterns. This
14989is required for sample fetches which return data type for which there is no
Davor Ocelice9ed2812017-12-25 17:49:28 +010014990obvious matching method (e.g. string or binary). When "-m" is specified and
Willy Tarreau74ca5042013-06-11 23:12:07 +020014991followed by a pattern matching method name, this method is used instead of the
14992default one for the criterion. This makes it possible to match contents in ways
14993that were not initially planned, or with sample fetch methods which return a
14994string. The matching method also affects the way the patterns are parsed.
14995
Thierry FOURNIERb7729c92014-02-11 16:24:41 +010014996The "-n" flag forbids the dns resolutions. It is used with the load of ip files.
14997By default, if the parser cannot parse ip address it considers that the parsed
14998string is maybe a domain name and try dns resolution. The flag "-n" disable this
14999resolution. It is useful for detecting malformed ip lists. Note that if the DNS
15000server is not reachable, the haproxy configuration parsing may last many minutes
John Roeslerfb2fce12019-07-10 15:45:51 -050015001waiting for the timeout. During this time no error messages are displayed. The
Thierry FOURNIERb7729c92014-02-11 16:24:41 +010015002flag "-n" disable this behavior. Note also that during the runtime, this
15003function is disabled for the dynamic acl modifications.
15004
Willy Tarreau74ca5042013-06-11 23:12:07 +020015005There are some restrictions however. Not all methods can be used with all
15006sample fetch methods. Also, if "-m" is used in conjunction with "-f", it must
15007be placed first. The pattern matching method must be one of the following :
Willy Tarreau5adeda12013-03-31 22:13:34 +020015008
15009 - "found" : only check if the requested sample could be found in the stream,
15010 but do not compare it against any pattern. It is recommended not
Willy Tarreau74ca5042013-06-11 23:12:07 +020015011 to pass any pattern to avoid confusion. This matching method is
15012 particularly useful to detect presence of certain contents such
15013 as headers, cookies, etc... even if they are empty and without
15014 comparing them to anything nor counting them.
Willy Tarreau5adeda12013-03-31 22:13:34 +020015015
15016 - "bool" : check the value as a boolean. It can only be applied to fetches
15017 which return a boolean or integer value, and takes no pattern.
Willy Tarreau74ca5042013-06-11 23:12:07 +020015018 Value zero or false does not match, all other values do match.
Willy Tarreau5adeda12013-03-31 22:13:34 +020015019
15020 - "int" : match the value as an integer. It can be used with integer and
Willy Tarreau74ca5042013-06-11 23:12:07 +020015021 boolean samples. Boolean false is integer 0, true is integer 1.
Willy Tarreau5adeda12013-03-31 22:13:34 +020015022
15023 - "ip" : match the value as an IPv4 or IPv6 address. It is compatible
Willy Tarreau74ca5042013-06-11 23:12:07 +020015024 with IP address samples only, so it is implied and never needed.
Willy Tarreau5adeda12013-03-31 22:13:34 +020015025
Davor Ocelice9ed2812017-12-25 17:49:28 +010015026 - "bin" : match the contents against a hexadecimal string representing a
Willy Tarreau5adeda12013-03-31 22:13:34 +020015027 binary sequence. This may be used with binary or string samples.
15028
15029 - "len" : match the sample's length as an integer. This may be used with
15030 binary or string samples.
15031
Willy Tarreau74ca5042013-06-11 23:12:07 +020015032 - "str" : exact match : match the contents against a string. This may be
15033 used with binary or string samples.
Willy Tarreau5adeda12013-03-31 22:13:34 +020015034
Willy Tarreau74ca5042013-06-11 23:12:07 +020015035 - "sub" : substring match : check that the contents contain at least one of
15036 the provided string patterns. This may be used with binary or
15037 string samples.
Willy Tarreau5adeda12013-03-31 22:13:34 +020015038
Willy Tarreau74ca5042013-06-11 23:12:07 +020015039 - "reg" : regex match : match the contents against a list of regular
15040 expressions. This may be used with binary or string samples.
Willy Tarreau5adeda12013-03-31 22:13:34 +020015041
Willy Tarreau74ca5042013-06-11 23:12:07 +020015042 - "beg" : prefix match : check that the contents begin like the provided
15043 string patterns. This may be used with binary or string samples.
Willy Tarreau5adeda12013-03-31 22:13:34 +020015044
Willy Tarreau74ca5042013-06-11 23:12:07 +020015045 - "end" : suffix match : check that the contents end like the provided
15046 string patterns. This may be used with binary or string samples.
Willy Tarreau5adeda12013-03-31 22:13:34 +020015047
Willy Tarreau74ca5042013-06-11 23:12:07 +020015048 - "dir" : subdir match : check that a slash-delimited portion of the
15049 contents exactly matches one of the provided string patterns.
Willy Tarreau5adeda12013-03-31 22:13:34 +020015050 This may be used with binary or string samples.
15051
Willy Tarreau74ca5042013-06-11 23:12:07 +020015052 - "dom" : domain match : check that a dot-delimited portion of the contents
15053 exactly match one of the provided string patterns. This may be
15054 used with binary or string samples.
Willy Tarreau5adeda12013-03-31 22:13:34 +020015055
15056For example, to quickly detect the presence of cookie "JSESSIONID" in an HTTP
15057request, it is possible to do :
15058
15059 acl jsess_present cook(JSESSIONID) -m found
15060
15061In order to apply a regular expression on the 500 first bytes of data in the
15062buffer, one would use the following acl :
15063
15064 acl script_tag payload(0,500) -m reg -i <script>
15065
Willy Tarreaue6b11e42013-11-26 19:02:32 +010015066On systems where the regex library is much slower when using "-i", it is
15067possible to convert the sample to lowercase before matching, like this :
15068
15069 acl script_tag payload(0,500),lower -m reg <script>
15070
Willy Tarreau74ca5042013-06-11 23:12:07 +020015071All ACL-specific criteria imply a default matching method. Most often, these
15072criteria are composed by concatenating the name of the original sample fetch
15073method and the matching method. For example, "hdr_beg" applies the "beg" match
15074to samples retrieved using the "hdr" fetch method. Since all ACL-specific
15075criteria rely on a sample fetch method, it is always possible instead to use
15076the original sample fetch method and the explicit matching method using "-m".
Willy Tarreau2b5285d2010-05-09 23:45:24 +020015077
Willy Tarreau74ca5042013-06-11 23:12:07 +020015078If an alternate match is specified using "-m" on an ACL-specific criterion,
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030015079the matching method is simply applied to the underlying sample fetch method.
15080For example, all ACLs below are exact equivalent :
Willy Tarreau2b5285d2010-05-09 23:45:24 +020015081
Willy Tarreau74ca5042013-06-11 23:12:07 +020015082 acl short_form hdr_beg(host) www.
15083 acl alternate1 hdr_beg(host) -m beg www.
15084 acl alternate2 hdr_dom(host) -m beg www.
15085 acl alternate3 hdr(host) -m beg www.
Willy Tarreau2b5285d2010-05-09 23:45:24 +020015086
Willy Tarreau2b5285d2010-05-09 23:45:24 +020015087
Thierry FOURNIER2a06e392014-05-11 15:49:55 +020015088The table below summarizes the compatibility matrix between sample or converter
15089types and the pattern types to fetch against. It indicates for each compatible
15090combination the name of the matching method to be used, surrounded with angle
15091brackets ">" and "<" when the method is the default one and will work by
15092default without "-m".
Willy Tarreau0ba27502007-12-24 16:55:16 +010015093
Willy Tarreau74ca5042013-06-11 23:12:07 +020015094 +-------------------------------------------------+
15095 | Input sample type |
15096 +----------------------+---------+---------+---------+---------+---------+
Thierry FOURNIER2a06e392014-05-11 15:49:55 +020015097 | pattern type | boolean | integer | ip | string | binary |
Willy Tarreau74ca5042013-06-11 23:12:07 +020015098 +----------------------+---------+---------+---------+---------+---------+
15099 | none (presence only) | found | found | found | found | found |
15100 +----------------------+---------+---------+---------+---------+---------+
Thierry FOURNIER2a06e392014-05-11 15:49:55 +020015101 | none (boolean value) |> bool <| bool | | bool | |
Willy Tarreau74ca5042013-06-11 23:12:07 +020015102 +----------------------+---------+---------+---------+---------+---------+
Thierry FOURNIER2a06e392014-05-11 15:49:55 +020015103 | integer (value) | int |> int <| int | int | |
Willy Tarreau74ca5042013-06-11 23:12:07 +020015104 +----------------------+---------+---------+---------+---------+---------+
Thierry FOURNIERe3ded592013-12-06 15:36:54 +010015105 | integer (length) | len | len | len | len | len |
Willy Tarreau74ca5042013-06-11 23:12:07 +020015106 +----------------------+---------+---------+---------+---------+---------+
Thierry FOURNIER2a06e392014-05-11 15:49:55 +020015107 | IP address | | |> ip <| ip | ip |
Willy Tarreau74ca5042013-06-11 23:12:07 +020015108 +----------------------+---------+---------+---------+---------+---------+
Thierry FOURNIER2a06e392014-05-11 15:49:55 +020015109 | exact string | str | str | str |> str <| str |
Willy Tarreau74ca5042013-06-11 23:12:07 +020015110 +----------------------+---------+---------+---------+---------+---------+
Thierry FOURNIERe3ded592013-12-06 15:36:54 +010015111 | prefix | beg | beg | beg | beg | beg |
Willy Tarreau74ca5042013-06-11 23:12:07 +020015112 +----------------------+---------+---------+---------+---------+---------+
Thierry FOURNIERe3ded592013-12-06 15:36:54 +010015113 | suffix | end | end | end | end | end |
Willy Tarreau74ca5042013-06-11 23:12:07 +020015114 +----------------------+---------+---------+---------+---------+---------+
Thierry FOURNIERe3ded592013-12-06 15:36:54 +010015115 | substring | sub | sub | sub | sub | sub |
Willy Tarreau74ca5042013-06-11 23:12:07 +020015116 +----------------------+---------+---------+---------+---------+---------+
Thierry FOURNIERe3ded592013-12-06 15:36:54 +010015117 | subdir | dir | dir | dir | dir | dir |
Willy Tarreau74ca5042013-06-11 23:12:07 +020015118 +----------------------+---------+---------+---------+---------+---------+
Thierry FOURNIERe3ded592013-12-06 15:36:54 +010015119 | domain | dom | dom | dom | dom | dom |
Willy Tarreau74ca5042013-06-11 23:12:07 +020015120 +----------------------+---------+---------+---------+---------+---------+
Thierry FOURNIERe3ded592013-12-06 15:36:54 +010015121 | regex | reg | reg | reg | reg | reg |
Willy Tarreau74ca5042013-06-11 23:12:07 +020015122 +----------------------+---------+---------+---------+---------+---------+
15123 | hex block | | | | bin | bin |
15124 +----------------------+---------+---------+---------+---------+---------+
Willy Tarreau6a06a402007-07-15 20:15:28 +020015125
15126
Willy Tarreau74ca5042013-06-11 23:12:07 +0200151277.1.1. Matching booleans
15128------------------------
15129
15130In order to match a boolean, no value is needed and all values are ignored.
15131Boolean matching is used by default for all fetch methods of type "boolean".
15132When boolean matching is used, the fetched value is returned as-is, which means
15133that a boolean "true" will always match and a boolean "false" will never match.
15134
15135Boolean matching may also be enforced using "-m bool" on fetch methods which
15136return an integer value. Then, integer value 0 is converted to the boolean
15137"false" and all other values are converted to "true".
15138
Willy Tarreau6a06a402007-07-15 20:15:28 +020015139
Willy Tarreau74ca5042013-06-11 23:12:07 +0200151407.1.2. Matching integers
15141------------------------
15142
15143Integer matching applies by default to integer fetch methods. It can also be
15144enforced on boolean fetches using "-m int". In this case, "false" is converted
15145to the integer 0, and "true" is converted to the integer 1.
15146
15147Integer matching also supports integer ranges and operators. Note that integer
15148matching only applies to positive values. A range is a value expressed with a
15149lower and an upper bound separated with a colon, both of which may be omitted.
Willy Tarreau6a06a402007-07-15 20:15:28 +020015150
15151For instance, "1024:65535" is a valid range to represent a range of
15152unprivileged ports, and "1024:" would also work. "0:1023" is a valid
15153representation of privileged ports, and ":1023" would also work.
15154
Willy Tarreau62644772008-07-16 18:36:06 +020015155As a special case, some ACL functions support decimal numbers which are in fact
15156two integers separated by a dot. This is used with some version checks for
15157instance. All integer properties apply to those decimal numbers, including
15158ranges and operators.
15159
Willy Tarreau6a06a402007-07-15 20:15:28 +020015160For an easier usage, comparison operators are also supported. Note that using
Willy Tarreau0ba27502007-12-24 16:55:16 +010015161operators with ranges does not make much sense and is strongly discouraged.
15162Similarly, it does not make much sense to perform order comparisons with a set
15163of values.
Willy Tarreau6a06a402007-07-15 20:15:28 +020015164
Willy Tarreau0ba27502007-12-24 16:55:16 +010015165Available operators for integer matching are :
Willy Tarreau6a06a402007-07-15 20:15:28 +020015166
15167 eq : true if the tested value equals at least one value
15168 ge : true if the tested value is greater than or equal to at least one value
15169 gt : true if the tested value is greater than at least one value
15170 le : true if the tested value is less than or equal to at least one value
15171 lt : true if the tested value is less than at least one value
15172
Willy Tarreau0ba27502007-12-24 16:55:16 +010015173For instance, the following ACL matches any negative Content-Length header :
Willy Tarreau6a06a402007-07-15 20:15:28 +020015174
15175 acl negative-length hdr_val(content-length) lt 0
15176
Willy Tarreau62644772008-07-16 18:36:06 +020015177This one matches SSL versions between 3.0 and 3.1 (inclusive) :
15178
15179 acl sslv3 req_ssl_ver 3:3.1
15180
Willy Tarreau6a06a402007-07-15 20:15:28 +020015181
Willy Tarreau74ca5042013-06-11 23:12:07 +0200151827.1.3. Matching strings
15183-----------------------
15184
15185String matching applies to string or binary fetch methods, and exists in 6
15186different forms :
15187
15188 - exact match (-m str) : the extracted string must exactly match the
Davor Ocelice9ed2812017-12-25 17:49:28 +010015189 patterns;
Willy Tarreau74ca5042013-06-11 23:12:07 +020015190
15191 - substring match (-m sub) : the patterns are looked up inside the
Davor Ocelice9ed2812017-12-25 17:49:28 +010015192 extracted string, and the ACL matches if any of them is found inside;
Willy Tarreau74ca5042013-06-11 23:12:07 +020015193
15194 - prefix match (-m beg) : the patterns are compared with the beginning of
15195 the extracted string, and the ACL matches if any of them matches.
15196
15197 - suffix match (-m end) : the patterns are compared with the end of the
15198 extracted string, and the ACL matches if any of them matches.
15199
Baptiste Assmann33db6002016-03-06 23:32:10 +010015200 - subdir match (-m dir) : the patterns are looked up inside the extracted
Willy Tarreau74ca5042013-06-11 23:12:07 +020015201 string, delimited with slashes ("/"), and the ACL matches if any of them
15202 matches.
15203
15204 - domain match (-m dom) : the patterns are looked up inside the extracted
15205 string, delimited with dots ("."), and the ACL matches if any of them
15206 matches.
Willy Tarreau6a06a402007-07-15 20:15:28 +020015207
15208String matching applies to verbatim strings as they are passed, with the
15209exception of the backslash ("\") which makes it possible to escape some
15210characters such as the space. If the "-i" flag is passed before the first
15211string, then the matching will be performed ignoring the case. In order
15212to match the string "-i", either set it second, or pass the "--" flag
Willy Tarreau0ba27502007-12-24 16:55:16 +010015213before the first string. Same applies of course to match the string "--".
Willy Tarreau6a06a402007-07-15 20:15:28 +020015214
Mathias Weiersmuellercb250fc2019-12-02 09:43:40 +010015215Do not use string matches for binary fetches which might contain null bytes
15216(0x00), as the comparison stops at the occurrence of the first null byte.
15217Instead, convert the binary fetch to a hex string with the hex converter first.
15218
15219Example:
15220 # matches if the string <tag> is present in the binary sample
15221 acl tag_found req.payload(0,0),hex -m sub 3C7461673E
15222
Willy Tarreau6a06a402007-07-15 20:15:28 +020015223
Willy Tarreau74ca5042013-06-11 23:12:07 +0200152247.1.4. Matching regular expressions (regexes)
15225---------------------------------------------
Willy Tarreau6a06a402007-07-15 20:15:28 +020015226
15227Just like with string matching, regex matching applies to verbatim strings as
15228they are passed, with the exception of the backslash ("\") which makes it
15229possible to escape some characters such as the space. If the "-i" flag is
15230passed before the first regex, then the matching will be performed ignoring
15231the case. In order to match the string "-i", either set it second, or pass
Willy Tarreau0ba27502007-12-24 16:55:16 +010015232the "--" flag before the first string. Same principle applies of course to
15233match the string "--".
Willy Tarreau6a06a402007-07-15 20:15:28 +020015234
15235
Willy Tarreau74ca5042013-06-11 23:12:07 +0200152367.1.5. Matching arbitrary data blocks
15237-------------------------------------
15238
15239It is possible to match some extracted samples against a binary block which may
15240not safely be represented as a string. For this, the patterns must be passed as
15241a series of hexadecimal digits in an even number, when the match method is set
15242to binary. Each sequence of two digits will represent a byte. The hexadecimal
15243digits may be used upper or lower case.
15244
15245Example :
15246 # match "Hello\n" in the input stream (\x48 \x65 \x6c \x6c \x6f \x0a)
15247 acl hello payload(0,6) -m bin 48656c6c6f0a
15248
15249
152507.1.6. Matching IPv4 and IPv6 addresses
15251---------------------------------------
Willy Tarreau6a06a402007-07-15 20:15:28 +020015252
15253IPv4 addresses values can be specified either as plain addresses or with a
15254netmask appended, in which case the IPv4 address matches whenever it is
15255within the network. Plain addresses may also be replaced with a resolvable
Willy Tarreaud2a4aa22008-01-31 15:28:22 +010015256host name, but this practice is generally discouraged as it makes it more
Willy Tarreau0ba27502007-12-24 16:55:16 +010015257difficult to read and debug configurations. If hostnames are used, you should
15258at least ensure that they are present in /etc/hosts so that the configuration
15259does not depend on any random DNS match at the moment the configuration is
15260parsed.
Willy Tarreau6a06a402007-07-15 20:15:28 +020015261
Daniel Schnellereba56342016-04-13 00:26:52 +020015262The dotted IPv4 address notation is supported in both regular as well as the
15263abbreviated form with all-0-octets omitted:
15264
15265 +------------------+------------------+------------------+
15266 | Example 1 | Example 2 | Example 3 |
15267 +------------------+------------------+------------------+
15268 | 192.168.0.1 | 10.0.0.12 | 127.0.0.1 |
15269 | 192.168.1 | 10.12 | 127.1 |
15270 | 192.168.0.1/22 | 10.0.0.12/8 | 127.0.0.1/8 |
15271 | 192.168.1/22 | 10.12/8 | 127.1/8 |
15272 +------------------+------------------+------------------+
15273
15274Notice that this is different from RFC 4632 CIDR address notation in which
15275192.168.42/24 would be equivalent to 192.168.42.0/24.
15276
Willy Tarreauceb4ac92012-04-28 00:41:46 +020015277IPv6 may be entered in their usual form, with or without a netmask appended.
15278Only bit counts are accepted for IPv6 netmasks. In order to avoid any risk of
15279trouble with randomly resolved IP addresses, host names are never allowed in
15280IPv6 patterns.
15281
15282HAProxy is also able to match IPv4 addresses with IPv6 addresses in the
15283following situations :
15284 - tested address is IPv4, pattern address is IPv4, the match applies
15285 in IPv4 using the supplied mask if any.
15286 - tested address is IPv6, pattern address is IPv6, the match applies
15287 in IPv6 using the supplied mask if any.
15288 - tested address is IPv6, pattern address is IPv4, the match applies in IPv4
15289 using the pattern's mask if the IPv6 address matches with 2002:IPV4::,
15290 ::IPV4 or ::ffff:IPV4, otherwise it fails.
15291 - tested address is IPv4, pattern address is IPv6, the IPv4 address is first
15292 converted to IPv6 by prefixing ::ffff: in front of it, then the match is
15293 applied in IPv6 using the supplied IPv6 mask.
15294
Willy Tarreau74ca5042013-06-11 23:12:07 +020015295
152967.2. Using ACLs to form conditions
15297----------------------------------
15298
15299Some actions are only performed upon a valid condition. A condition is a
15300combination of ACLs with operators. 3 operators are supported :
15301
15302 - AND (implicit)
15303 - OR (explicit with the "or" keyword or the "||" operator)
15304 - Negation with the exclamation mark ("!")
Willy Tarreau6a06a402007-07-15 20:15:28 +020015305
Willy Tarreau74ca5042013-06-11 23:12:07 +020015306A condition is formed as a disjunctive form:
Willy Tarreau6a06a402007-07-15 20:15:28 +020015307
Willy Tarreau74ca5042013-06-11 23:12:07 +020015308 [!]acl1 [!]acl2 ... [!]acln { or [!]acl1 [!]acl2 ... [!]acln } ...
Willy Tarreaubef91e72013-03-31 23:14:46 +020015309
Willy Tarreau74ca5042013-06-11 23:12:07 +020015310Such conditions are generally used after an "if" or "unless" statement,
15311indicating when the condition will trigger the action.
Willy Tarreaubef91e72013-03-31 23:14:46 +020015312
Willy Tarreau74ca5042013-06-11 23:12:07 +020015313For instance, to block HTTP requests to the "*" URL with methods other than
15314"OPTIONS", as well as POST requests without content-length, and GET or HEAD
15315requests with a content-length greater than 0, and finally every request which
15316is not either GET/HEAD/POST/OPTIONS !
15317
15318 acl missing_cl hdr_cnt(Content-length) eq 0
Jarno Huuskonen84c51ec2017-04-03 14:20:34 +030015319 http-request deny if HTTP_URL_STAR !METH_OPTIONS || METH_POST missing_cl
15320 http-request deny if METH_GET HTTP_CONTENT
15321 http-request deny unless METH_GET or METH_POST or METH_OPTIONS
Willy Tarreau74ca5042013-06-11 23:12:07 +020015322
15323To select a different backend for requests to static contents on the "www" site
15324and to every request on the "img", "video", "download" and "ftp" hosts :
15325
15326 acl url_static path_beg /static /images /img /css
15327 acl url_static path_end .gif .png .jpg .css .js
15328 acl host_www hdr_beg(host) -i www
15329 acl host_static hdr_beg(host) -i img. video. download. ftp.
15330
Davor Ocelice9ed2812017-12-25 17:49:28 +010015331 # now use backend "static" for all static-only hosts, and for static URLs
Willy Tarreau74ca5042013-06-11 23:12:07 +020015332 # of host "www". Use backend "www" for the rest.
15333 use_backend static if host_static or host_www url_static
15334 use_backend www if host_www
15335
15336It is also possible to form rules using "anonymous ACLs". Those are unnamed ACL
15337expressions that are built on the fly without needing to be declared. They must
15338be enclosed between braces, with a space before and after each brace (because
15339the braces must be seen as independent words). Example :
15340
15341 The following rule :
15342
15343 acl missing_cl hdr_cnt(Content-length) eq 0
Jarno Huuskonen84c51ec2017-04-03 14:20:34 +030015344 http-request deny if METH_POST missing_cl
Willy Tarreau74ca5042013-06-11 23:12:07 +020015345
15346 Can also be written that way :
15347
Jarno Huuskonen84c51ec2017-04-03 14:20:34 +030015348 http-request deny if METH_POST { hdr_cnt(Content-length) eq 0 }
Willy Tarreau74ca5042013-06-11 23:12:07 +020015349
15350It is generally not recommended to use this construct because it's a lot easier
15351to leave errors in the configuration when written that way. However, for very
15352simple rules matching only one source IP address for instance, it can make more
15353sense to use them than to declare ACLs with random names. Another example of
15354good use is the following :
15355
15356 With named ACLs :
15357
15358 acl site_dead nbsrv(dynamic) lt 2
15359 acl site_dead nbsrv(static) lt 2
15360 monitor fail if site_dead
15361
15362 With anonymous ACLs :
15363
15364 monitor fail if { nbsrv(dynamic) lt 2 } || { nbsrv(static) lt 2 }
15365
Jarno Huuskonen84c51ec2017-04-03 14:20:34 +030015366See section 4.2 for detailed help on the "http-request deny" and "use_backend"
15367keywords.
Willy Tarreau74ca5042013-06-11 23:12:07 +020015368
15369
153707.3. Fetching samples
15371---------------------
15372
15373Historically, sample fetch methods were only used to retrieve data to match
15374against patterns using ACLs. With the arrival of stick-tables, a new class of
15375sample fetch methods was created, most often sharing the same syntax as their
15376ACL counterpart. These sample fetch methods are also known as "fetches". As
15377of now, ACLs and fetches have converged. All ACL fetch methods have been made
15378available as fetch methods, and ACLs may use any sample fetch method as well.
15379
15380This section details all available sample fetch methods and their output type.
15381Some sample fetch methods have deprecated aliases that are used to maintain
15382compatibility with existing configurations. They are then explicitly marked as
15383deprecated and should not be used in new setups.
15384
15385The ACL derivatives are also indicated when available, with their respective
15386matching methods. These ones all have a well defined default pattern matching
15387method, so it is never necessary (though allowed) to pass the "-m" option to
15388indicate how the sample will be matched using ACLs.
15389
15390As indicated in the sample type versus matching compatibility matrix above,
15391when using a generic sample fetch method in an ACL, the "-m" option is
15392mandatory unless the sample type is one of boolean, integer, IPv4 or IPv6. When
15393the same keyword exists as an ACL keyword and as a standard fetch method, the
15394ACL engine will automatically pick the ACL-only one by default.
15395
15396Some of these keywords support one or multiple mandatory arguments, and one or
15397multiple optional arguments. These arguments are strongly typed and are checked
15398when the configuration is parsed so that there is no risk of running with an
Davor Ocelice9ed2812017-12-25 17:49:28 +010015399incorrect argument (e.g. an unresolved backend name). Fetch function arguments
15400are passed between parenthesis and are delimited by commas. When an argument
Willy Tarreau74ca5042013-06-11 23:12:07 +020015401is optional, it will be indicated below between square brackets ('[ ]'). When
15402all arguments are optional, the parenthesis may be omitted.
15403
15404Thus, the syntax of a standard sample fetch method is one of the following :
15405 - name
15406 - name(arg1)
15407 - name(arg1,arg2)
15408
Thierry FOURNIER060762e2014-04-23 13:29:15 +020015409
154107.3.1. Converters
15411-----------------
15412
Willy Tarreaue6b11e42013-11-26 19:02:32 +010015413Sample fetch methods may be combined with transformations to be applied on top
15414of the fetched sample (also called "converters"). These combinations form what
15415is called "sample expressions" and the result is a "sample". Initially this
15416was only supported by "stick on" and "stick store-request" directives but this
Davor Ocelice9ed2812017-12-25 17:49:28 +010015417has now be extended to all places where samples may be used (ACLs, log-format,
Willy Tarreaue6b11e42013-11-26 19:02:32 +010015418unique-id-format, add-header, ...).
15419
15420These transformations are enumerated as a series of specific keywords after the
15421sample fetch method. These keywords may equally be appended immediately after
15422the fetch keyword's argument, delimited by a comma. These keywords can also
Davor Ocelice9ed2812017-12-25 17:49:28 +010015423support some arguments (e.g. a netmask) which must be passed in parenthesis.
Willy Tarreau0ba27502007-12-24 16:55:16 +010015424
Willy Tarreau97707872015-01-27 15:12:13 +010015425A certain category of converters are bitwise and arithmetic operators which
15426support performing basic operations on integers. Some bitwise operations are
15427supported (and, or, xor, cpl) and some arithmetic operations are supported
15428(add, sub, mul, div, mod, neg). Some comparators are provided (odd, even, not,
15429bool) which make it possible to report a match without having to write an ACL.
15430
Willy Tarreau74ca5042013-06-11 23:12:07 +020015431The currently available list of transformation keywords include :
Willy Tarreau0ba27502007-12-24 16:55:16 +010015432
Ben Shillitof25e8e52016-12-02 14:25:37 +00001543351d.single(<prop>[,<prop>*])
15434 Returns values for the properties requested as a string, where values are
15435 separated by the delimiter specified with "51degrees-property-separator".
15436 The device is identified using the User-Agent header passed to the
15437 converter. The function can be passed up to five property names, and if a
15438 property name can't be found, the value "NoData" is returned.
15439
15440 Example :
Davor Ocelice9ed2812017-12-25 17:49:28 +010015441 # Here the header "X-51D-DeviceTypeMobileTablet" is added to the request,
15442 # containing values for the three properties requested by using the
Ben Shillitof25e8e52016-12-02 14:25:37 +000015443 # User-Agent passed to the converter.
15444 frontend http-in
15445 bind *:8081
15446 default_backend servers
15447 http-request set-header X-51D-DeviceTypeMobileTablet \
15448 %[req.fhdr(User-Agent),51d.single(DeviceType,IsMobile,IsTablet)]
15449
Willy Tarreau97707872015-01-27 15:12:13 +010015450add(<value>)
Thierry FOURNIER07ee64e2015-07-06 23:43:03 +020015451 Adds <value> to the input value of type signed integer, and returns the
Thierry FOURNIER5d86fae2015-07-07 21:10:16 +020015452 result as a signed integer. <value> can be a numeric value or a variable
Daniel Schneller0b547052016-03-21 20:46:57 +010015453 name. The name of the variable starts with an indication about its scope. The
15454 scopes allowed are:
Christopher Fauletff2613e2016-11-09 11:36:17 +010015455 "proc" : the variable is shared with the whole process
Daniel Schneller0b547052016-03-21 20:46:57 +010015456 "sess" : the variable is shared with the whole session
15457 "txn" : the variable is shared with the transaction (request and response)
15458 "req" : the variable is shared only during request processing
15459 "res" : the variable is shared only during response processing
Davor Ocelice9ed2812017-12-25 17:49:28 +010015460 This prefix is followed by a name. The separator is a '.'. The name may only
Christopher Fauletb71557a2016-10-31 10:49:03 +010015461 contain characters 'a-z', 'A-Z', '0-9', '.' and '_'.
Willy Tarreau97707872015-01-27 15:12:13 +010015462
Nenad Merdanovicc31499d2019-03-23 11:00:32 +010015463aes_gcm_dec(<bits>,<nonce>,<key>,<aead_tag>)
15464 Decrypts the raw byte input using the AES128-GCM, AES192-GCM or
15465 AES256-GCM algorithm, depending on the <bits> parameter. All other parameters
15466 need to be base64 encoded and the returned result is in raw byte format.
15467 If the <aead_tag> validation fails, the converter doesn't return any data.
15468 The <nonce>, <key> and <aead_tag> can either be strings or variables. This
15469 converter requires at least OpenSSL 1.0.1.
15470
15471 Example:
15472 http-response set-header X-Decrypted-Text %[var(txn.enc),\
15473 aes_gcm_dec(128,txn.nonce,Zm9vb2Zvb29mb29wZm9vbw==,txn.aead_tag)]
15474
Willy Tarreau97707872015-01-27 15:12:13 +010015475and(<value>)
Thierry FOURNIER07ee64e2015-07-06 23:43:03 +020015476 Performs a bitwise "AND" between <value> and the input value of type signed
Thierry FOURNIER5d86fae2015-07-07 21:10:16 +020015477 integer, and returns the result as an signed integer. <value> can be a
Daniel Schneller0b547052016-03-21 20:46:57 +010015478 numeric value or a variable name. The name of the variable starts with an
15479 indication about its scope. The scopes allowed are:
Christopher Fauletff2613e2016-11-09 11:36:17 +010015480 "proc" : the variable is shared with the whole process
Daniel Schneller0b547052016-03-21 20:46:57 +010015481 "sess" : the variable is shared with the whole session
15482 "txn" : the variable is shared with the transaction (request and response)
15483 "req" : the variable is shared only during request processing
15484 "res" : the variable is shared only during response processing
Davor Ocelice9ed2812017-12-25 17:49:28 +010015485 This prefix is followed by a name. The separator is a '.'. The name may only
Christopher Fauletb71557a2016-10-31 10:49:03 +010015486 contain characters 'a-z', 'A-Z', '0-9', '.' and '_'.
Willy Tarreau97707872015-01-27 15:12:13 +010015487
Holger Just1bfc24b2017-05-06 00:56:53 +020015488b64dec
15489 Converts (decodes) a base64 encoded input string to its binary
15490 representation. It performs the inverse operation of base64().
15491
Emeric Brun53d1a982014-04-30 18:21:37 +020015492base64
15493 Converts a binary input sample to a base64 string. It is used to log or
Davor Ocelice9ed2812017-12-25 17:49:28 +010015494 transfer binary content in a way that can be reliably transferred (e.g.
Emeric Brun53d1a982014-04-30 18:21:37 +020015495 an SSL ID can be copied in a header).
15496
Willy Tarreau97707872015-01-27 15:12:13 +010015497bool
Thierry FOURNIER07ee64e2015-07-06 23:43:03 +020015498 Returns a boolean TRUE if the input value of type signed integer is
Willy Tarreau97707872015-01-27 15:12:13 +010015499 non-null, otherwise returns FALSE. Used in conjunction with and(), it can be
Davor Ocelice9ed2812017-12-25 17:49:28 +010015500 used to report true/false for bit testing on input values (e.g. verify the
Willy Tarreau97707872015-01-27 15:12:13 +010015501 presence of a flag).
15502
Emeric Brun54c4ac82014-11-03 15:32:43 +010015503bytes(<offset>[,<length>])
15504 Extracts some bytes from an input binary sample. The result is a binary
15505 sample starting at an offset (in bytes) of the original sample and
Tim Düsterhus4896c442016-11-29 02:15:19 +010015506 optionally truncated at the given length.
Emeric Brun54c4ac82014-11-03 15:32:43 +010015507
Willy Tarreau280f42b2018-02-19 15:34:12 +010015508concat([<start>],[<var>],[<end>])
15509 Concatenates up to 3 fields after the current sample which is then turned to
15510 a string. The first one, <start>, is a constant string, that will be appended
15511 immediately after the existing sample. It may be omitted if not used. The
15512 second one, <var>, is a variable name. The variable will be looked up, its
15513 contents converted to a string, and it will be appended immediately after the
15514 <first> part. If the variable is not found, nothing is appended. It may be
15515 omitted as well. The third field, <end> is a constant string that will be
15516 appended after the variable. It may also be omitted. Together, these elements
15517 allow to concatenate variables with delimiters to an existing set of
15518 variables. This can be used to build new variables made of a succession of
Willy Tarreauef21fac2020-02-14 13:37:20 +010015519 other variables, such as colon-delimited values. If commas or closing
Daniel Corbett67a82712020-07-06 23:01:19 -040015520 parenthesis are needed as delimiters, they must be protected by quotes or
Willy Tarreauef21fac2020-02-14 13:37:20 +010015521 backslashes, themselves protected so that they are not stripped by the first
15522 level parser. See examples below.
Willy Tarreau280f42b2018-02-19 15:34:12 +010015523
15524 Example:
15525 tcp-request session set-var(sess.src) src
15526 tcp-request session set-var(sess.dn) ssl_c_s_dn
15527 tcp-request session set-var(txn.sig) str(),concat(<ip=,sess.ip,>),concat(<dn=,sess.dn,>)
Willy Tarreauef21fac2020-02-14 13:37:20 +010015528 tcp-request session set-var(txn.ipport) "str(),concat('addr=(',sess.ip),concat(',',sess.port,')')"
Willy Tarreau280f42b2018-02-19 15:34:12 +010015529 http-request set-header x-hap-sig %[var(txn.sig)]
15530
Willy Tarreau97707872015-01-27 15:12:13 +010015531cpl
Thierry FOURNIER07ee64e2015-07-06 23:43:03 +020015532 Takes the input value of type signed integer, applies a ones-complement
15533 (flips all bits) and returns the result as an signed integer.
Willy Tarreau97707872015-01-27 15:12:13 +010015534
Willy Tarreau80599772015-01-20 19:35:24 +010015535crc32([<avalanche>])
15536 Hashes a binary input sample into an unsigned 32-bit quantity using the CRC32
15537 hash function. Optionally, it is possible to apply a full avalanche hash
15538 function to the output if the optional <avalanche> argument equals 1. This
15539 converter uses the same functions as used by the various hash-based load
15540 balancing algorithms, so it will provide exactly the same results. It is
15541 provided for compatibility with other software which want a CRC32 to be
15542 computed on some input keys, so it follows the most common implementation as
15543 found in Ethernet, Gzip, PNG, etc... It is slower than the other algorithms
15544 but may provide a better or at least less predictable distribution. It must
15545 not be used for security purposes as a 32-bit hash is trivial to break. See
Emmanuel Hocdet50791a72018-03-21 11:19:01 +010015546 also "djb2", "sdbm", "wt6", "crc32c" and the "hash-type" directive.
15547
15548crc32c([<avalanche>])
15549 Hashes a binary input sample into an unsigned 32-bit quantity using the CRC32C
15550 hash function. Optionally, it is possible to apply a full avalanche hash
15551 function to the output if the optional <avalanche> argument equals 1. This
15552 converter uses the same functions as described in RFC4960, Appendix B [8].
15553 It is provided for compatibility with other software which want a CRC32C to be
15554 computed on some input keys. It is slower than the other algorithms and it must
15555 not be used for security purposes as a 32-bit hash is trivial to break. See
15556 also "djb2", "sdbm", "wt6", "crc32" and the "hash-type" directive.
Willy Tarreau80599772015-01-20 19:35:24 +010015557
Christopher Fauletea159d62020-04-01 16:21:44 +020015558cut_crlf
15559 Cuts the string representation of the input sample on the first carriage
15560 return ('\r') or newline ('\n') character found. Only the string length is
15561 updated.
15562
David Carlier29b3ca32015-09-25 14:09:21 +010015563da-csv-conv(<prop>[,<prop>*])
David Carlier4542b102015-06-01 13:54:29 +020015564 Asks the DeviceAtlas converter to identify the User Agent string passed on
15565 input, and to emit a string made of the concatenation of the properties
15566 enumerated in argument, delimited by the separator defined by the global
15567 keyword "deviceatlas-property-separator", or by default the pipe character
David Carlier840b0242016-03-16 10:09:55 +000015568 ('|'). There's a limit of 12 different properties imposed by the haproxy
David Carlier4542b102015-06-01 13:54:29 +020015569 configuration language.
15570
15571 Example:
15572 frontend www
Cyril Bonté307ee1e2015-09-28 23:16:06 +020015573 bind *:8881
15574 default_backend servers
David Carlier840b0242016-03-16 10:09:55 +000015575 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 +020015576
Willy Tarreau0851fd52019-12-17 10:07:25 +010015577debug([<prefix][,<destination>])
15578 This converter is used as debug tool. It takes a capture of the input sample
15579 and sends it to event sink <destination>, which may designate a ring buffer
15580 such as "buf0", as well as "stdout", or "stderr". Available sinks may be
15581 checked at run time by issuing "show events" on the CLI. When not specified,
15582 the output will be "buf0", which may be consulted via the CLI's "show events"
15583 command. An optional prefix <prefix> may be passed to help distinguish
15584 outputs from multiple expressions. It will then appear before the colon in
15585 the output message. The input sample is passed as-is on the output, so that
15586 it is safe to insert the debug converter anywhere in a chain, even with non-
15587 printable sample types.
15588
15589 Example:
15590 tcp-request connection track-sc0 src,debug(track-sc)
Thierry FOURNIER9687c772015-05-07 15:46:29 +020015591
Patrick Gansterer8e366512020-04-22 16:47:57 +020015592digest(<algorithm>)
15593 Converts a binary input sample to a message digest. The result is a binary
15594 sample. The <algorithm> must be an OpenSSL message digest name (e.g. sha256).
15595
15596 Please note that this converter is only available when haproxy has been
15597 compiled with USE_OPENSSL.
15598
Willy Tarreau97707872015-01-27 15:12:13 +010015599div(<value>)
Thierry FOURNIER07ee64e2015-07-06 23:43:03 +020015600 Divides the input value of type signed integer by <value>, and returns the
15601 result as an signed integer. If <value> is null, the largest unsigned
Thierry FOURNIER5d86fae2015-07-07 21:10:16 +020015602 integer is returned (typically 2^63-1). <value> can be a numeric value or a
Daniel Schneller0b547052016-03-21 20:46:57 +010015603 variable name. The name of the variable starts with an indication about its
15604 scope. The scopes allowed are:
Christopher Fauletff2613e2016-11-09 11:36:17 +010015605 "proc" : the variable is shared with the whole process
Daniel Schneller0b547052016-03-21 20:46:57 +010015606 "sess" : the variable is shared with the whole session
15607 "txn" : the variable is shared with the transaction (request and response)
15608 "req" : the variable is shared only during request processing
15609 "res" : the variable is shared only during response processing
Davor Ocelice9ed2812017-12-25 17:49:28 +010015610 This prefix is followed by a name. The separator is a '.'. The name may only
Christopher Fauletb71557a2016-10-31 10:49:03 +010015611 contain characters 'a-z', 'A-Z', '0-9', '.' and '_'.
Willy Tarreau97707872015-01-27 15:12:13 +010015612
Willy Tarreau23ec4ca2014-07-15 20:15:37 +020015613djb2([<avalanche>])
15614 Hashes a binary input sample into an unsigned 32-bit quantity using the DJB2
15615 hash function. Optionally, it is possible to apply a full avalanche hash
15616 function to the output if the optional <avalanche> argument equals 1. This
15617 converter uses the same functions as used by the various hash-based load
15618 balancing algorithms, so it will provide exactly the same results. It is
15619 mostly intended for debugging, but can be used as a stick-table entry to
15620 collect rough statistics. It must not be used for security purposes as a
Emmanuel Hocdet50791a72018-03-21 11:19:01 +010015621 32-bit hash is trivial to break. See also "crc32", "sdbm", "wt6", "crc32c",
15622 and the "hash-type" directive.
Willy Tarreau23ec4ca2014-07-15 20:15:37 +020015623
Willy Tarreau97707872015-01-27 15:12:13 +010015624even
Thierry FOURNIER07ee64e2015-07-06 23:43:03 +020015625 Returns a boolean TRUE if the input value of type signed integer is even
Willy Tarreau97707872015-01-27 15:12:13 +010015626 otherwise returns FALSE. It is functionally equivalent to "not,and(1),bool".
15627
Marcin Deranek9631a282018-04-16 14:30:46 +020015628field(<index>,<delimiters>[,<count>])
15629 Extracts the substring at the given index counting from the beginning
15630 (positive index) or from the end (negative index) considering given delimiters
15631 from an input string. Indexes start at 1 or -1 and delimiters are a string
15632 formatted list of chars. Optionally you can specify <count> of fields to
15633 extract (default: 1). Value of 0 indicates extraction of all remaining
15634 fields.
15635
15636 Example :
15637 str(f1_f2_f3__f5),field(5,_) # f5
15638 str(f1_f2_f3__f5),field(2,_,0) # f2_f3__f5
15639 str(f1_f2_f3__f5),field(2,_,2) # f2_f3
15640 str(f1_f2_f3__f5),field(-2,_,3) # f2_f3_
15641 str(f1_f2_f3__f5),field(-3,_,0) # f1_f2_f3
Emeric Brunf399b0d2014-11-03 17:07:03 +010015642
Baptiste Assmanne138dda2020-10-22 15:39:03 +020015643fix_is_valid
15644 Parses a binary payload and performs sanity checks regarding FIX (Financial
15645 Information eXchange):
15646
15647 - checks that all tag IDs and values are not empty and the tags IDs are well
15648 numeric
Ilya Shipitsin2272d8a2020-12-21 01:22:40 +050015649 - checks the BeginString tag is the first tag with a valid FIX version
Baptiste Assmanne138dda2020-10-22 15:39:03 +020015650 - checks the BodyLength tag is the second one with the right body length
Christopher Fauleted4bef72021-03-18 17:40:56 +010015651 - checks the MsgType tag is the third tag.
Baptiste Assmanne138dda2020-10-22 15:39:03 +020015652 - checks that last tag in the message is the CheckSum tag with a valid
15653 checksum
15654
15655 Due to current HAProxy design, only the first message sent by the client and
15656 the server can be parsed.
15657
15658 This converter returns a boolean, true if the payload contains a valid FIX
15659 message, false if not.
15660
15661 See also the fix_tag_value converter.
15662
15663 Example:
15664 tcp-request inspect-delay 10s
15665 tcp-request content reject unless { req.payload(0,0),fix_is_valid }
15666
15667fix_tag_value(<tag>)
15668 Parses a FIX (Financial Information eXchange) message and extracts the value
15669 from the tag <tag>. <tag> can be a string or an integer pointing to the
15670 desired tag. Any integer value is accepted, but only the following strings
15671 are translated into their integer equivalent: BeginString, BodyLength,
Daniel Corbettbefef702021-03-09 23:00:34 -050015672 MsgType, SenderCompID, TargetCompID, CheckSum. More tag names can be easily
Baptiste Assmanne138dda2020-10-22 15:39:03 +020015673 added.
15674
15675 Due to current HAProxy design, only the first message sent by the client and
15676 the server can be parsed. No message validation is performed by this
15677 converter. It is highly recommended to validate the message first using
15678 fix_is_valid converter.
15679
15680 See also the fix_is_valid converter.
15681
15682 Example:
15683 tcp-request inspect-delay 10s
15684 tcp-request content reject unless { req.payload(0,0),fix_is_valid }
15685 # MsgType tag ID is 35, so both lines below will return the same content
15686 tcp-request content set-var(txn.foo) req.payload(0,0),fix_tag_value(35)
15687 tcp-request content set-var(txn.bar) req.payload(0,0),fix_tag_value(MsgType)
15688
Thierry FOURNIER060762e2014-04-23 13:29:15 +020015689hex
Davor Ocelice9ed2812017-12-25 17:49:28 +010015690 Converts a binary input sample to a hex string containing two hex digits per
Thierry FOURNIER060762e2014-04-23 13:29:15 +020015691 input byte. It is used to log or transfer hex dumps of some binary input data
Davor Ocelice9ed2812017-12-25 17:49:28 +010015692 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 +020015693 header).
Thierry FOURNIER2f49d6d2014-03-12 15:01:52 +010015694
Dragan Dosen3f957b22017-10-24 09:27:34 +020015695hex2i
15696 Converts a hex string containing two hex digits per input byte to an
John Roeslerfb2fce12019-07-10 15:45:51 -050015697 integer. If the input value cannot be converted, then zero is returned.
Dragan Dosen3f957b22017-10-24 09:27:34 +020015698
Christopher Faulet4ccc12f2020-04-01 09:08:32 +020015699htonl
15700 Converts the input integer value to its 32-bit binary representation in the
15701 network byte order. Because sample fetches own signed 64-bit integer, when
15702 this converter is used, the input integer value is first casted to an
15703 unsigned 32-bit integer.
15704
Tim Duesterhusa3082092021-01-21 17:40:49 +010015705hmac(<algorithm>,<key>)
Patrick Gansterer8e366512020-04-22 16:47:57 +020015706 Converts a binary input sample to a message authentication code with the given
15707 key. The result is a binary sample. The <algorithm> must be one of the
15708 registered OpenSSL message digest names (e.g. sha256). The <key> parameter must
15709 be base64 encoded and can either be a string or a variable.
15710
15711 Please note that this converter is only available when haproxy has been
15712 compiled with USE_OPENSSL.
15713
Cyril Bonté6bcd1822019-11-05 23:13:59 +010015714http_date([<offset],[<unit>])
Thierry FOURNIER060762e2014-04-23 13:29:15 +020015715 Converts an integer supposed to contain a date since epoch to a string
15716 representing this date in a format suitable for use in HTTP header fields. If
Damien Claisseae6f1252019-10-30 15:57:28 +000015717 an offset value is specified, then it is added to the date before the
15718 conversion is operated. This is particularly useful to emit Date header fields,
15719 Expires values in responses when combined with a positive offset, or
15720 Last-Modified values when the offset is negative.
15721 If a unit value is specified, then consider the timestamp as either
15722 "s" for seconds (default behavior), "ms" for milliseconds, or "us" for
15723 microseconds since epoch. Offset is assumed to have the same unit as
15724 input timestamp.
Willy Tarreau74ca5042013-06-11 23:12:07 +020015725
Tim Duesterhus3943e4f2020-09-11 14:25:23 +020015726iif(<true>,<false>)
15727 Returns the <true> string if the input value is true. Returns the <false>
15728 string otherwise.
15729
15730 Example:
Tim Duesterhus870713b2020-09-11 17:13:12 +020015731 http-request set-header x-forwarded-proto %[ssl_fc,iif(https,http)]
Tim Duesterhus3943e4f2020-09-11 14:25:23 +020015732
Willy Tarreaud9f316a2014-07-10 14:03:38 +020015733in_table(<table>)
15734 Uses the string representation of the input sample to perform a look up in
15735 the specified table. If the key is not found in the table, a boolean false
15736 is returned. Otherwise a boolean true is returned. This can be used to verify
Davor Ocelice9ed2812017-12-25 17:49:28 +010015737 the presence of a certain key in a table tracking some elements (e.g. whether
Willy Tarreaud9f316a2014-07-10 14:03:38 +020015738 or not a source IP address or an Authorization header was already seen).
15739
Tim Duesterhusa3082092021-01-21 17:40:49 +010015740ipmask(<mask4>,[<mask6>])
Tim Duesterhus1478aa72018-01-25 16:24:51 +010015741 Apply a mask to an IP address, and use the result for lookups and storage.
Willy Tarreauffcb2e42014-07-10 16:29:08 +020015742 This can be used to make all hosts within a certain mask to share the same
Tim Duesterhus1478aa72018-01-25 16:24:51 +010015743 table entries and as such use the same server. The mask4 can be passed in
15744 dotted form (e.g. 255.255.255.0) or in CIDR form (e.g. 24). The mask6 can
15745 be passed in quadruplet form (e.g. ffff:ffff::) or in CIDR form (e.g. 64).
15746 If no mask6 is given IPv6 addresses will fail to convert for backwards
15747 compatibility reasons.
Willy Tarreauffcb2e42014-07-10 16:29:08 +020015748
Thierry FOURNIER317e1c42014-08-12 10:20:47 +020015749json([<input-code>])
Davor Ocelice9ed2812017-12-25 17:49:28 +010015750 Escapes the input string and produces an ASCII output string ready to use as a
Thierry FOURNIER317e1c42014-08-12 10:20:47 +020015751 JSON string. The converter tries to decode the input string according to the
Herve COMMOWICK8dfe8632016-08-05 12:01:20 +020015752 <input-code> parameter. It can be "ascii", "utf8", "utf8s", "utf8p" or
Thierry FOURNIER317e1c42014-08-12 10:20:47 +020015753 "utf8ps". The "ascii" decoder never fails. The "utf8" decoder detects 3 types
15754 of errors:
15755 - bad UTF-8 sequence (lone continuation byte, bad number of continuation
15756 bytes, ...)
15757 - invalid range (the decoded value is within a UTF-8 prohibited range),
15758 - code overlong (the value is encoded with more bytes than necessary).
15759
15760 The UTF-8 JSON encoding can produce a "too long value" error when the UTF-8
15761 character is greater than 0xffff because the JSON string escape specification
15762 only authorizes 4 hex digits for the value encoding. The UTF-8 decoder exists
15763 in 4 variants designated by a combination of two suffix letters : "p" for
15764 "permissive" and "s" for "silently ignore". The behaviors of the decoders
15765 are :
Davor Ocelice9ed2812017-12-25 17:49:28 +010015766 - "ascii" : never fails;
15767 - "utf8" : fails on any detected errors;
15768 - "utf8s" : never fails, but removes characters corresponding to errors;
Thierry FOURNIER317e1c42014-08-12 10:20:47 +020015769 - "utf8p" : accepts and fixes the overlong errors, but fails on any other
Davor Ocelice9ed2812017-12-25 17:49:28 +010015770 error;
Thierry FOURNIER317e1c42014-08-12 10:20:47 +020015771 - "utf8ps" : never fails, accepts and fixes the overlong errors, but removes
15772 characters corresponding to the other errors.
15773
15774 This converter is particularly useful for building properly escaped JSON for
Davor Ocelice9ed2812017-12-25 17:49:28 +010015775 logging to servers which consume JSON-formatted traffic logs.
Thierry FOURNIER317e1c42014-08-12 10:20:47 +020015776
15777 Example:
Thierry FOURNIER317e1c42014-08-12 10:20:47 +020015778 capture request header Host len 15
Herve COMMOWICK8dfe8632016-08-05 12:01:20 +020015779 capture request header user-agent len 150
15780 log-format '{"ip":"%[src]","user-agent":"%[capture.req.hdr(1),json(utf8s)]"}'
Thierry FOURNIER317e1c42014-08-12 10:20:47 +020015781
15782 Input request from client 127.0.0.1:
15783 GET / HTTP/1.0
15784 User-Agent: Very "Ugly" UA 1/2
15785
15786 Output log:
15787 {"ip":"127.0.0.1","user-agent":"Very \"Ugly\" UA 1\/2"}
15788
Thierry FOURNIER060762e2014-04-23 13:29:15 +020015789language(<value>[,<default>])
15790 Returns the value with the highest q-factor from a list as extracted from the
15791 "accept-language" header using "req.fhdr". Values with no q-factor have a
15792 q-factor of 1. Values with a q-factor of 0 are dropped. Only values which
15793 belong to the list of semi-colon delimited <values> will be considered. The
15794 argument <value> syntax is "lang[;lang[;lang[;...]]]". If no value matches the
15795 given list and a default value is provided, it is returned. Note that language
15796 names may have a variant after a dash ('-'). If this variant is present in the
15797 list, it will be matched, but if it is not, only the base language is checked.
15798 The match is case-sensitive, and the output string is always one of those
Davor Ocelice9ed2812017-12-25 17:49:28 +010015799 provided in arguments. The ordering of arguments is meaningless, only the
Thierry FOURNIER060762e2014-04-23 13:29:15 +020015800 ordering of the values in the request counts, as the first value among
15801 multiple sharing the same q-factor is used.
Thierry FOURNIERad903512014-04-11 17:51:01 +020015802
Thierry FOURNIER060762e2014-04-23 13:29:15 +020015803 Example :
Thierry FOURNIERad903512014-04-11 17:51:01 +020015804
Thierry FOURNIER060762e2014-04-23 13:29:15 +020015805 # this configuration switches to the backend matching a
15806 # given language based on the request :
Thierry FOURNIERad903512014-04-11 17:51:01 +020015807
Thierry FOURNIER060762e2014-04-23 13:29:15 +020015808 acl es req.fhdr(accept-language),language(es;fr;en) -m str es
15809 acl fr req.fhdr(accept-language),language(es;fr;en) -m str fr
15810 acl en req.fhdr(accept-language),language(es;fr;en) -m str en
15811 use_backend spanish if es
15812 use_backend french if fr
15813 use_backend english if en
15814 default_backend choose_your_language
Thierry FOURNIERad903512014-04-11 17:51:01 +020015815
Willy Tarreau60a2ee72017-12-15 07:13:48 +010015816length
Etienne Carriereed0d24e2017-12-13 13:41:34 +010015817 Get the length of the string. This can only be placed after a string
15818 sample fetch function or after a transformation keyword returning a string
15819 type. The result is of type integer.
15820
Willy Tarreauffcb2e42014-07-10 16:29:08 +020015821lower
15822 Convert a string sample to lower case. This can only be placed after a string
15823 sample fetch function or after a transformation keyword returning a string
15824 type. The result is of type string.
15825
Willy Tarreau0dbfdba2014-07-10 16:37:47 +020015826ltime(<format>[,<offset>])
15827 Converts an integer supposed to contain a date since epoch to a string
15828 representing this date in local time using a format defined by the <format>
15829 string using strftime(3). The purpose is to allow any date format to be used
15830 in logs. An optional <offset> in seconds may be applied to the input date
15831 (positive or negative). See the strftime() man page for the format supported
15832 by your operating system. See also the utime converter.
15833
15834 Example :
15835
15836 # Emit two colons, one with the local time and another with ip:port
Davor Ocelice9ed2812017-12-25 17:49:28 +010015837 # e.g. 20140710162350 127.0.0.1:57325
Willy Tarreau0dbfdba2014-07-10 16:37:47 +020015838 log-format %[date,ltime(%Y%m%d%H%M%S)]\ %ci:%cp
15839
Christopher Faulet51fc9d12020-04-01 17:24:41 +020015840ltrim(<chars>)
15841 Skips any characters from <chars> from the beginning of the string
15842 representation of the input sample.
15843
Thierry FOURNIER060762e2014-04-23 13:29:15 +020015844map(<map_file>[,<default_value>])
15845map_<match_type>(<map_file>[,<default_value>])
15846map_<match_type>_<output_type>(<map_file>[,<default_value>])
15847 Search the input value from <map_file> using the <match_type> matching method,
15848 and return the associated value converted to the type <output_type>. If the
15849 input value cannot be found in the <map_file>, the converter returns the
15850 <default_value>. If the <default_value> is not set, the converter fails and
15851 acts as if no input value could be fetched. If the <match_type> is not set, it
15852 defaults to "str". Likewise, if the <output_type> is not set, it defaults to
15853 "str". For convenience, the "map" keyword is an alias for "map_str" and maps a
15854 string to another string.
Thierry FOURNIERd5f624d2013-11-26 11:52:33 +010015855
Thierry FOURNIER060762e2014-04-23 13:29:15 +020015856 It is important to avoid overlapping between the keys : IP addresses and
15857 strings are stored in trees, so the first of the finest match will be used.
15858 Other keys are stored in lists, so the first matching occurrence will be used.
Thierry FOURNIERd5f624d2013-11-26 11:52:33 +010015859
Tim Düsterhus4896c442016-11-29 02:15:19 +010015860 The following array contains the list of all map functions available sorted by
Thierry FOURNIER060762e2014-04-23 13:29:15 +020015861 input type, match type and output type.
Thierry FOURNIERd5f624d2013-11-26 11:52:33 +010015862
Thierry FOURNIER060762e2014-04-23 13:29:15 +020015863 input type | match method | output type str | output type int | output type ip
15864 -----------+--------------+-----------------+-----------------+---------------
15865 str | str | map_str | map_str_int | map_str_ip
15866 -----------+--------------+-----------------+-----------------+---------------
Willy Tarreau787a4c02014-05-10 07:55:30 +020015867 str | beg | map_beg | map_beg_int | map_end_ip
15868 -----------+--------------+-----------------+-----------------+---------------
Thierry FOURNIER060762e2014-04-23 13:29:15 +020015869 str | sub | map_sub | map_sub_int | map_sub_ip
15870 -----------+--------------+-----------------+-----------------+---------------
15871 str | dir | map_dir | map_dir_int | map_dir_ip
15872 -----------+--------------+-----------------+-----------------+---------------
15873 str | dom | map_dom | map_dom_int | map_dom_ip
15874 -----------+--------------+-----------------+-----------------+---------------
15875 str | end | map_end | map_end_int | map_end_ip
15876 -----------+--------------+-----------------+-----------------+---------------
Ruoshan Huang3c5e3742016-12-02 16:25:31 +080015877 str | reg | map_reg | map_reg_int | map_reg_ip
15878 -----------+--------------+-----------------+-----------------+---------------
15879 str | reg | map_regm | map_reg_int | map_reg_ip
Thierry FOURNIER060762e2014-04-23 13:29:15 +020015880 -----------+--------------+-----------------+-----------------+---------------
15881 int | int | map_int | map_int_int | map_int_ip
15882 -----------+--------------+-----------------+-----------------+---------------
15883 ip | ip | map_ip | map_ip_int | map_ip_ip
15884 -----------+--------------+-----------------+-----------------+---------------
Thierry FOURNIERd5f624d2013-11-26 11:52:33 +010015885
Thierry Fournier8feaa662016-02-10 22:55:20 +010015886 The special map called "map_regm" expect matching zone in the regular
15887 expression and modify the output replacing back reference (like "\1") by
15888 the corresponding match text.
15889
Thierry FOURNIER060762e2014-04-23 13:29:15 +020015890 The file contains one key + value per line. Lines which start with '#' are
15891 ignored, just like empty lines. Leading tabs and spaces are stripped. The key
15892 is then the first "word" (series of non-space/tabs characters), and the value
15893 is what follows this series of space/tab till the end of the line excluding
15894 trailing spaces/tabs.
Thierry FOURNIERd5f624d2013-11-26 11:52:33 +010015895
Thierry FOURNIER060762e2014-04-23 13:29:15 +020015896 Example :
15897
15898 # this is a comment and is ignored
15899 2.22.246.0/23 United Kingdom \n
15900 <-><-----------><--><------------><---->
15901 | | | | `- trailing spaces ignored
15902 | | | `---------- value
15903 | | `-------------------- middle spaces ignored
15904 | `---------------------------- key
15905 `------------------------------------ leading spaces ignored
15906
Willy Tarreau97707872015-01-27 15:12:13 +010015907mod(<value>)
Thierry FOURNIER07ee64e2015-07-06 23:43:03 +020015908 Divides the input value of type signed integer by <value>, and returns the
15909 remainder as an signed integer. If <value> is null, then zero is returned.
Thierry FOURNIER5d86fae2015-07-07 21:10:16 +020015910 <value> can be a numeric value or a variable name. The name of the variable
Daniel Schneller0b547052016-03-21 20:46:57 +010015911 starts with an indication about its scope. The scopes allowed are:
Christopher Fauletff2613e2016-11-09 11:36:17 +010015912 "proc" : the variable is shared with the whole process
Daniel Schneller0b547052016-03-21 20:46:57 +010015913 "sess" : the variable is shared with the whole session
15914 "txn" : the variable is shared with the transaction (request and response)
15915 "req" : the variable is shared only during request processing
15916 "res" : the variable is shared only during response processing
Davor Ocelice9ed2812017-12-25 17:49:28 +010015917 This prefix is followed by a name. The separator is a '.'. The name may only
Christopher Fauletb71557a2016-10-31 10:49:03 +010015918 contain characters 'a-z', 'A-Z', '0-9', '.' and '_'.
Willy Tarreau97707872015-01-27 15:12:13 +010015919
Baptiste Assmanne279ca62020-10-27 18:10:06 +010015920mqtt_field_value(<packettype>,<fieldname or property ID>)
15921 Returns value of <fieldname> found in input MQTT payload of type
15922 <packettype>.
15923 <packettype> can be either a string (case insensitive matching) or a numeric
15924 value corresponding to the type of packet we're supposed to extract data
15925 from.
15926 Supported string and integers can be found here:
15927 https://docs.oasis-open.org/mqtt/mqtt/v3.1.1/os/mqtt-v3.1.1-os.html#_Toc398718021
15928 https://docs.oasis-open.org/mqtt/mqtt/v5.0/os/mqtt-v5.0-os.html#_Toc3901022
15929
15930 <fieldname> depends on <packettype> and can be any of the following below.
15931 (note that <fieldname> matching is case insensitive).
15932 <property id> can only be found in MQTT v5.0 streams. check this table:
15933 https://docs.oasis-open.org/mqtt/mqtt/v5.0/os/mqtt-v5.0-os.html#_Toc3901029
15934
15935 - CONNECT (or 1): flags, protocol_name, protocol_version, client_identifier,
15936 will_topic, will_payload, username, password, keepalive
15937 OR any property ID as a numeric value (for MQTT v5.0
15938 packets only):
15939 17: Session Expiry Interval
15940 33: Receive Maximum
15941 39: Maximum Packet Size
15942 34: Topic Alias Maximum
15943 25: Request Response Information
15944 23: Request Problem Information
15945 21: Authentication Method
15946 22: Authentication Data
15947 18: Will Delay Interval
15948 1: Payload Format Indicator
15949 2: Message Expiry Interval
15950 3: Content Type
15951 8: Response Topic
15952 9: Correlation Data
15953 Not supported yet:
15954 38: User Property
15955
15956 - CONNACK (or 2): flags, protocol_version, reason_code
15957 OR any property ID as a numeric value (for MQTT v5.0
15958 packets only):
15959 17: Session Expiry Interval
15960 33: Receive Maximum
15961 36: Maximum QoS
15962 37: Retain Available
15963 39: Maximum Packet Size
15964 18: Assigned Client Identifier
15965 34: Topic Alias Maximum
15966 31: Reason String
15967 40; Wildcard Subscription Available
15968 41: Subscription Identifiers Available
15969 42: Shared Subscription Available
15970 19: Server Keep Alive
15971 26: Response Information
15972 28: Server Reference
15973 21: Authentication Method
15974 22: Authentication Data
15975 Not supported yet:
15976 38: User Property
15977
15978 Due to current HAProxy design, only the first message sent by the client and
15979 the server can be parsed. Thus this converter can extract data only from
15980 CONNECT and CONNACK packet types. CONNECT is the first message sent by the
15981 client and CONNACK is the first response sent by the server.
15982
15983 Example:
15984
15985 acl data_in_buffer req.len ge 4
15986 tcp-request content set-var(txn.username) \
15987 req.payload(0,0),mqtt_field_value(connect,protocol_name) \
15988 if data_in_buffer
15989 # do the same as above
15990 tcp-request content set-var(txn.username) \
15991 req.payload(0,0),mqtt_field_value(1,protocol_name) \
15992 if data_in_buffer
15993
15994mqtt_is_valid
15995 Checks that the binary input is a valid MQTT packet. It returns a boolean.
15996
15997 Due to current HAProxy design, only the first message sent by the client and
15998 the server can be parsed. Thus this converter can extract data only from
15999 CONNECT and CONNACK packet types. CONNECT is the first message sent by the
16000 client and CONNACK is the first response sent by the server.
16001
16002 Example:
16003
16004 acl data_in_buffer req.len ge 4
16005 tcp-request content reject unless req.payload(0,0),mqtt_is_valid
16006
Willy Tarreau97707872015-01-27 15:12:13 +010016007mul(<value>)
Thierry FOURNIER07ee64e2015-07-06 23:43:03 +020016008 Multiplies the input value of type signed integer by <value>, and returns
Thierry FOURNIER00c005c2015-07-08 01:10:21 +020016009 the product as an signed integer. In case of overflow, the largest possible
16010 value for the sign is returned so that the operation doesn't wrap around.
Thierry FOURNIER5d86fae2015-07-07 21:10:16 +020016011 <value> can be a numeric value or a variable name. The name of the variable
Daniel Schneller0b547052016-03-21 20:46:57 +010016012 starts with an indication about its scope. The scopes allowed are:
Christopher Fauletff2613e2016-11-09 11:36:17 +010016013 "proc" : the variable is shared with the whole process
Daniel Schneller0b547052016-03-21 20:46:57 +010016014 "sess" : the variable is shared with the whole session
16015 "txn" : the variable is shared with the transaction (request and response)
16016 "req" : the variable is shared only during request processing
16017 "res" : the variable is shared only during response processing
Davor Ocelice9ed2812017-12-25 17:49:28 +010016018 This prefix is followed by a name. The separator is a '.'. The name may only
Christopher Fauletb71557a2016-10-31 10:49:03 +010016019 contain characters 'a-z', 'A-Z', '0-9', '.' and '_'.
Willy Tarreau97707872015-01-27 15:12:13 +010016020
Nenad Merdanovicb7e7c472017-03-12 21:56:55 +010016021nbsrv
16022 Takes an input value of type string, interprets it as a backend name and
16023 returns the number of usable servers in that backend. Can be used in places
16024 where we want to look up a backend from a dynamic name, like a result of a
16025 map lookup.
16026
Willy Tarreau97707872015-01-27 15:12:13 +010016027neg
Thierry FOURNIER07ee64e2015-07-06 23:43:03 +020016028 Takes the input value of type signed integer, computes the opposite value,
16029 and returns the remainder as an signed integer. 0 is identity. This operator
16030 is provided for reversed subtracts : in order to subtract the input from a
16031 constant, simply perform a "neg,add(value)".
Willy Tarreau97707872015-01-27 15:12:13 +010016032
16033not
Thierry FOURNIER07ee64e2015-07-06 23:43:03 +020016034 Returns a boolean FALSE if the input value of type signed integer is
Willy Tarreau97707872015-01-27 15:12:13 +010016035 non-null, otherwise returns TRUE. Used in conjunction with and(), it can be
Davor Ocelice9ed2812017-12-25 17:49:28 +010016036 used to report true/false for bit testing on input values (e.g. verify the
Willy Tarreau97707872015-01-27 15:12:13 +010016037 absence of a flag).
16038
16039odd
Thierry FOURNIER07ee64e2015-07-06 23:43:03 +020016040 Returns a boolean TRUE if the input value of type signed integer is odd
Willy Tarreau97707872015-01-27 15:12:13 +010016041 otherwise returns FALSE. It is functionally equivalent to "and(1),bool".
16042
16043or(<value>)
Thierry FOURNIER07ee64e2015-07-06 23:43:03 +020016044 Performs a bitwise "OR" between <value> and the input value of type signed
Thierry FOURNIER5d86fae2015-07-07 21:10:16 +020016045 integer, and returns the result as an signed integer. <value> can be a
Daniel Schneller0b547052016-03-21 20:46:57 +010016046 numeric value or a variable name. The name of the variable starts with an
16047 indication about its scope. The scopes allowed are:
Christopher Fauletff2613e2016-11-09 11:36:17 +010016048 "proc" : the variable is shared with the whole process
Daniel Schneller0b547052016-03-21 20:46:57 +010016049 "sess" : the variable is shared with the whole session
16050 "txn" : the variable is shared with the transaction (request and response)
16051 "req" : the variable is shared only during request processing
16052 "res" : the variable is shared only during response processing
Davor Ocelice9ed2812017-12-25 17:49:28 +010016053 This prefix is followed by a name. The separator is a '.'. The name may only
Christopher Fauletb71557a2016-10-31 10:49:03 +010016054 contain characters 'a-z', 'A-Z', '0-9', '.' and '_'.
Willy Tarreau97707872015-01-27 15:12:13 +010016055
Frédéric Lécaillebfe61382019-03-06 14:34:36 +010016056protobuf(<field_number>,[<field_type>])
16057 This extracts the protocol buffers message field in raw mode of an input binary
16058 sample representation of a protocol buffer message with <field_number> as field
16059 number (dotted notation) if <field_type> is not present, or as an integer sample
16060 if this field is present (see also "ungrpc" below).
16061 The list of the authorized types is the following one: "int32", "int64", "uint32",
16062 "uint64", "sint32", "sint64", "bool", "enum" for the "varint" wire type 0
16063 "fixed64", "sfixed64", "double" for the 64bit wire type 1, "fixed32", "sfixed32",
16064 "float" for the wire type 5. Note that "string" is considered as a length-delimited
16065 type, so it does not require any <field_type> argument to be extracted.
16066 More information may be found here about the protocol buffers message field types:
16067 https://developers.google.com/protocol-buffers/docs/encoding
16068
Willy Tarreauc4dc3502015-01-23 20:39:28 +010016069regsub(<regex>,<subst>[,<flags>])
Willy Tarreau7eda8492015-01-20 19:47:06 +010016070 Applies a regex-based substitution to the input string. It does the same
16071 operation as the well-known "sed" utility with "s/<regex>/<subst>/". By
16072 default it will replace in the input string the first occurrence of the
16073 largest part matching the regular expression <regex> with the substitution
16074 string <subst>. It is possible to replace all occurrences instead by adding
16075 the flag "g" in the third argument <flags>. It is also possible to make the
16076 regex case insensitive by adding the flag "i" in <flags>. Since <flags> is a
16077 string, it is made up from the concatenation of all desired flags. Thus if
16078 both "i" and "g" are desired, using "gi" or "ig" will have the same effect.
Willy Tarreauef21fac2020-02-14 13:37:20 +010016079 The first use of this converter is to replace certain characters or sequence
16080 of characters with other ones.
16081
16082 It is highly recommended to enclose the regex part using protected quotes to
16083 improve clarity and never have a closing parenthesis from the regex mixed up
16084 with the parenthesis from the function. Just like in Bourne shell, the first
16085 level of quotes is processed when delimiting word groups on the line, a
16086 second level is usable for argument. It is recommended to use single quotes
16087 outside since these ones do not try to resolve backslashes nor dollar signs.
Willy Tarreau7eda8492015-01-20 19:47:06 +010016088
Willy Tarreaucd0d2ed2020-02-14 17:33:06 +010016089 Examples:
Willy Tarreau7eda8492015-01-20 19:47:06 +010016090
16091 # de-duplicate "/" in header "x-path".
16092 # input: x-path: /////a///b/c/xzxyz/
16093 # output: x-path: /a/b/c/xzxyz/
Willy Tarreauef21fac2020-02-14 13:37:20 +010016094 http-request set-header x-path "%[hdr(x-path),regsub('/+','/','g')]"
Willy Tarreau7eda8492015-01-20 19:47:06 +010016095
Willy Tarreaucd0d2ed2020-02-14 17:33:06 +010016096 # copy query string to x-query and drop all leading '?', ';' and '&'
16097 http-request set-header x-query "%[query,regsub([?;&]*,'')]"
16098
Jerome Magnin07e1e3c2020-02-16 19:20:19 +010016099 # capture groups and backreferences
16100 # both lines do the same.
Willy Tarreau465dc7d2020-10-08 18:05:56 +020016101 http-request redirect location %[url,'regsub("(foo|bar)([0-9]+)?","\2\1",i)']
Jerome Magnin07e1e3c2020-02-16 19:20:19 +010016102 http-request redirect location %[url,regsub(\"(foo|bar)([0-9]+)?\",\"\2\1\",i)]
16103
Thierry FOURNIER35ab2752015-05-28 13:22:03 +020016104capture-req(<id>)
16105 Capture the string entry in the request slot <id> and returns the entry as
16106 is. If the slot doesn't exist, the capture fails silently.
16107
16108 See also: "declare capture", "http-request capture",
Baptiste Assmann5ac425c2015-10-21 23:13:46 +020016109 "http-response capture", "capture.req.hdr" and
16110 "capture.res.hdr" (sample fetches).
Thierry FOURNIER35ab2752015-05-28 13:22:03 +020016111
16112capture-res(<id>)
16113 Capture the string entry in the response slot <id> and returns the entry as
16114 is. If the slot doesn't exist, the capture fails silently.
16115
16116 See also: "declare capture", "http-request capture",
Baptiste Assmann5ac425c2015-10-21 23:13:46 +020016117 "http-response capture", "capture.req.hdr" and
16118 "capture.res.hdr" (sample fetches).
Thierry FOURNIER35ab2752015-05-28 13:22:03 +020016119
Christopher Faulet568415a2020-04-01 17:24:47 +020016120rtrim(<chars>)
16121 Skips any characters from <chars> from the end of the string representation
16122 of the input sample.
16123
Willy Tarreau23ec4ca2014-07-15 20:15:37 +020016124sdbm([<avalanche>])
16125 Hashes a binary input sample into an unsigned 32-bit quantity using the SDBM
16126 hash function. Optionally, it is possible to apply a full avalanche hash
16127 function to the output if the optional <avalanche> argument equals 1. This
16128 converter uses the same functions as used by the various hash-based load
16129 balancing algorithms, so it will provide exactly the same results. It is
16130 mostly intended for debugging, but can be used as a stick-table entry to
16131 collect rough statistics. It must not be used for security purposes as a
Emmanuel Hocdet50791a72018-03-21 11:19:01 +010016132 32-bit hash is trivial to break. See also "crc32", "djb2", "wt6", "crc32c",
16133 and the "hash-type" directive.
Willy Tarreau23ec4ca2014-07-15 20:15:37 +020016134
Tim Duesterhusf38175c2020-06-09 11:48:42 +020016135secure_memcmp(<var>)
16136 Compares the contents of <var> with the input value. Both values are treated
16137 as a binary string. Returns a boolean indicating whether both binary strings
16138 match.
16139
16140 If both binary strings have the same length then the comparison will be
16141 performed in constant time.
16142
16143 Please note that this converter is only available when haproxy has been
16144 compiled with USE_OPENSSL.
16145
16146 Example :
16147
16148 http-request set-var(txn.token) hdr(token)
16149 # Check whether the token sent by the client matches the secret token
16150 # value, without leaking the contents using a timing attack.
16151 acl token_given str(my_secret_token),secure_memcmp(txn.token)
16152
Tim Duesterhusef4e45c2021-01-21 17:40:50 +010016153set-var(<var>)
Davor Ocelice9ed2812017-12-25 17:49:28 +010016154 Sets a variable with the input content and returns the content on the output
16155 as-is. The variable keeps the value and the associated input type. The name of
16156 the variable starts with an indication about its scope. The scopes allowed are:
Christopher Fauletff2613e2016-11-09 11:36:17 +010016157 "proc" : the variable is shared with the whole process
Daniel Schneller0b547052016-03-21 20:46:57 +010016158 "sess" : the variable is shared with the whole session
16159 "txn" : the variable is shared with the transaction (request and
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020016160 response),
Daniel Schneller0b547052016-03-21 20:46:57 +010016161 "req" : the variable is shared only during request processing,
16162 "res" : the variable is shared only during response processing.
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020016163 This prefix is followed by a name. The separator is a '.'. The name may only
Christopher Fauletb71557a2016-10-31 10:49:03 +010016164 contain characters 'a-z', 'A-Z', '0-9', '.' and '_'.
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020016165
Dragan Dosen6e5a9ca2017-10-24 09:18:23 +020016166sha1
Tim Duesterhusd4376302019-06-17 12:41:44 +020016167 Converts a binary input sample to a SHA-1 digest. The result is a binary
Dragan Dosen6e5a9ca2017-10-24 09:18:23 +020016168 sample with length of 20 bytes.
16169
Tim Duesterhusd4376302019-06-17 12:41:44 +020016170sha2([<bits>])
16171 Converts a binary input sample to a digest in the SHA-2 family. The result
16172 is a binary sample with length of <bits>/8 bytes.
16173
16174 Valid values for <bits> are 224, 256, 384, 512, each corresponding to
16175 SHA-<bits>. The default value is 256.
16176
16177 Please note that this converter is only available when haproxy has been
16178 compiled with USE_OPENSSL.
16179
Nenad Merdanovic177adc92019-08-27 01:58:13 +020016180srv_queue
16181 Takes an input value of type string, either a server name or <backend>/<server>
16182 format and returns the number of queued sessions on that server. Can be used
16183 in places where we want to look up queued sessions from a dynamic name, like a
16184 cookie value (e.g. req.cook(SRVID),srv_queue) and then make a decision to break
16185 persistence or direct a request elsewhere.
16186
Tim Duesterhusca097c12018-04-27 21:18:45 +020016187strcmp(<var>)
16188 Compares the contents of <var> with the input value of type string. Returns
16189 the result as a signed integer compatible with strcmp(3): 0 if both strings
16190 are identical. A value less than 0 if the left string is lexicographically
16191 smaller than the right string or if the left string is shorter. A value greater
16192 than 0 otherwise (right string greater than left string or the right string is
16193 shorter).
16194
Tim Duesterhusf38175c2020-06-09 11:48:42 +020016195 See also the secure_memcmp converter if you need to compare two binary
16196 strings in constant time.
16197
Tim Duesterhusca097c12018-04-27 21:18:45 +020016198 Example :
16199
16200 http-request set-var(txn.host) hdr(host)
16201 # Check whether the client is attempting domain fronting.
16202 acl ssl_sni_http_host_match ssl_fc_sni,strcmp(txn.host) eq 0
16203
16204
Willy Tarreau97707872015-01-27 15:12:13 +010016205sub(<value>)
Thierry FOURNIER07ee64e2015-07-06 23:43:03 +020016206 Subtracts <value> from the input value of type signed integer, and returns
16207 the result as an signed integer. Note: in order to subtract the input from
Thierry FOURNIER5d86fae2015-07-07 21:10:16 +020016208 a constant, simply perform a "neg,add(value)". <value> can be a numeric value
Daniel Schneller0b547052016-03-21 20:46:57 +010016209 or a variable name. The name of the variable starts with an indication about
16210 its scope. The scopes allowed are:
Christopher Fauletff2613e2016-11-09 11:36:17 +010016211 "proc" : the variable is shared with the whole process
Daniel Schneller0b547052016-03-21 20:46:57 +010016212 "sess" : the variable is shared with the whole session
16213 "txn" : the variable is shared with the transaction (request and
Thierry FOURNIER5d86fae2015-07-07 21:10:16 +020016214 response),
Daniel Schneller0b547052016-03-21 20:46:57 +010016215 "req" : the variable is shared only during request processing,
16216 "res" : the variable is shared only during response processing.
Thierry FOURNIER5d86fae2015-07-07 21:10:16 +020016217 This prefix is followed by a name. The separator is a '.'. The name may only
Christopher Fauletb71557a2016-10-31 10:49:03 +010016218 contain characters 'a-z', 'A-Z', '0-9', '.' and '_'.
Willy Tarreau97707872015-01-27 15:12:13 +010016219
Willy Tarreaud9f316a2014-07-10 14:03:38 +020016220table_bytes_in_rate(<table>)
16221 Uses the string representation of the input sample to perform a look up in
16222 the specified table. If the key is not found in the table, integer value zero
16223 is returned. Otherwise the converter returns the average client-to-server
16224 bytes rate associated with the input sample in the designated table, measured
16225 in amount of bytes over the period configured in the table. See also the
16226 sc_bytes_in_rate sample fetch keyword.
16227
16228
16229table_bytes_out_rate(<table>)
16230 Uses the string representation of the input sample to perform a look up in
16231 the specified table. If the key is not found in the table, integer value zero
16232 is returned. Otherwise the converter returns the average server-to-client
16233 bytes rate associated with the input sample in the designated table, measured
16234 in amount of bytes over the period configured in the table. See also the
16235 sc_bytes_out_rate sample fetch keyword.
16236
16237table_conn_cnt(<table>)
16238 Uses the string representation of the input sample to perform a look up in
16239 the specified table. If the key is not found in the table, integer value zero
Davor Ocelice9ed2812017-12-25 17:49:28 +010016240 is returned. Otherwise the converter returns the cumulative number of incoming
Willy Tarreaud9f316a2014-07-10 14:03:38 +020016241 connections associated with the input sample in the designated table. See
16242 also the sc_conn_cnt sample fetch keyword.
16243
16244table_conn_cur(<table>)
16245 Uses the string representation of the input sample to perform a look up in
16246 the specified table. If the key is not found in the table, integer value zero
16247 is returned. Otherwise the converter returns the current amount of concurrent
16248 tracked connections associated with the input sample in the designated table.
16249 See also the sc_conn_cur sample fetch keyword.
16250
16251table_conn_rate(<table>)
16252 Uses the string representation of the input sample to perform a look up in
16253 the specified table. If the key is not found in the table, integer value zero
16254 is returned. Otherwise the converter returns the average incoming connection
16255 rate associated with the input sample in the designated table. See also the
16256 sc_conn_rate sample fetch keyword.
16257
Thierry FOURNIER236657b2015-08-19 08:25:14 +020016258table_gpt0(<table>)
16259 Uses the string representation of the input sample to perform a look up in
16260 the specified table. If the key is not found in the table, boolean value zero
16261 is returned. Otherwise the converter returns the current value of the first
16262 general purpose tag associated with the input sample in the designated table.
16263 See also the sc_get_gpt0 sample fetch keyword.
16264
Willy Tarreaud9f316a2014-07-10 14:03:38 +020016265table_gpc0(<table>)
16266 Uses the string representation of the input sample to perform a look up in
16267 the specified table. If the key is not found in the table, integer value zero
16268 is returned. Otherwise the converter returns the current value of the first
16269 general purpose counter associated with the input sample in the designated
16270 table. See also the sc_get_gpc0 sample fetch keyword.
16271
16272table_gpc0_rate(<table>)
16273 Uses the string representation of the input sample to perform a look up in
16274 the specified table. If the key is not found in the table, integer value zero
16275 is returned. Otherwise the converter returns the frequency which the gpc0
16276 counter was incremented over the configured period in the table, associated
16277 with the input sample in the designated table. See also the sc_get_gpc0_rate
16278 sample fetch keyword.
16279
Frédéric Lécaille6778b272018-01-29 15:22:53 +010016280table_gpc1(<table>)
16281 Uses the string representation of the input sample to perform a look up in
16282 the specified table. If the key is not found in the table, integer value zero
16283 is returned. Otherwise the converter returns the current value of the second
16284 general purpose counter associated with the input sample in the designated
16285 table. See also the sc_get_gpc1 sample fetch keyword.
16286
16287table_gpc1_rate(<table>)
16288 Uses the string representation of the input sample to perform a look up in
16289 the specified table. If the key is not found in the table, integer value zero
16290 is returned. Otherwise the converter returns the frequency which the gpc1
16291 counter was incremented over the configured period in the table, associated
16292 with the input sample in the designated table. See also the sc_get_gpc1_rate
16293 sample fetch keyword.
16294
Willy Tarreaud9f316a2014-07-10 14:03:38 +020016295table_http_err_cnt(<table>)
16296 Uses the string representation of the input sample to perform a look up in
16297 the specified table. If the key is not found in the table, integer value zero
Davor Ocelice9ed2812017-12-25 17:49:28 +010016298 is returned. Otherwise the converter returns the cumulative number of HTTP
Willy Tarreaud9f316a2014-07-10 14:03:38 +020016299 errors associated with the input sample in the designated table. See also the
16300 sc_http_err_cnt sample fetch keyword.
16301
16302table_http_err_rate(<table>)
16303 Uses the string representation of the input sample to perform a look up in
16304 the specified table. If the key is not found in the table, integer value zero
16305 is returned. Otherwise the average rate of HTTP errors associated with the
16306 input sample in the designated table, measured in amount of errors over the
16307 period configured in the table. See also the sc_http_err_rate sample fetch
16308 keyword.
16309
Willy Tarreau826f3ab2021-02-10 12:07:15 +010016310table_http_fail_cnt(<table>)
16311 Uses the string representation of the input sample to perform a look up in
16312 the specified table. If the key is not found in the table, integer value zero
16313 is returned. Otherwise the converter returns the cumulative number of HTTP
16314 failures associated with the input sample in the designated table. See also
16315 the sc_http_fail_cnt sample fetch keyword.
16316
16317table_http_fail_rate(<table>)
16318 Uses the string representation of the input sample to perform a look up in
16319 the specified table. If the key is not found in the table, integer value zero
16320 is returned. Otherwise the average rate of HTTP failures associated with the
16321 input sample in the designated table, measured in amount of failures over the
16322 period configured in the table. See also the sc_http_fail_rate sample fetch
16323 keyword.
16324
Willy Tarreaud9f316a2014-07-10 14:03:38 +020016325table_http_req_cnt(<table>)
16326 Uses the string representation of the input sample to perform a look up in
16327 the specified table. If the key is not found in the table, integer value zero
Davor Ocelice9ed2812017-12-25 17:49:28 +010016328 is returned. Otherwise the converter returns the cumulative number of HTTP
Willy Tarreaud9f316a2014-07-10 14:03:38 +020016329 requests associated with the input sample in the designated table. See also
16330 the sc_http_req_cnt sample fetch keyword.
16331
16332table_http_req_rate(<table>)
16333 Uses the string representation of the input sample to perform a look up in
16334 the specified table. If the key is not found in the table, integer value zero
16335 is returned. Otherwise the average rate of HTTP requests associated with the
16336 input sample in the designated table, measured in amount of requests over the
16337 period configured in the table. See also the sc_http_req_rate sample fetch
16338 keyword.
16339
16340table_kbytes_in(<table>)
16341 Uses the string representation of the input sample to perform a look up in
16342 the specified table. If the key is not found in the table, integer value zero
Davor Ocelice9ed2812017-12-25 17:49:28 +010016343 is returned. Otherwise the converter returns the cumulative number of client-
Willy Tarreaud9f316a2014-07-10 14:03:38 +020016344 to-server data associated with the input sample in the designated table,
16345 measured in kilobytes. The test is currently performed on 32-bit integers,
16346 which limits values to 4 terabytes. See also the sc_kbytes_in sample fetch
16347 keyword.
16348
16349table_kbytes_out(<table>)
16350 Uses the string representation of the input sample to perform a look up in
16351 the specified table. If the key is not found in the table, integer value zero
Davor Ocelice9ed2812017-12-25 17:49:28 +010016352 is returned. Otherwise the converter returns the cumulative number of server-
Willy Tarreaud9f316a2014-07-10 14:03:38 +020016353 to-client data associated with the input sample in the designated table,
16354 measured in kilobytes. The test is currently performed on 32-bit integers,
16355 which limits values to 4 terabytes. See also the sc_kbytes_out sample fetch
16356 keyword.
16357
16358table_server_id(<table>)
16359 Uses the string representation of the input sample to perform a look up in
16360 the specified table. If the key is not found in the table, integer value zero
16361 is returned. Otherwise the converter returns the server ID associated with
16362 the input sample in the designated table. A server ID is associated to a
16363 sample by a "stick" rule when a connection to a server succeeds. A server ID
16364 zero means that no server is associated with this key.
16365
16366table_sess_cnt(<table>)
16367 Uses the string representation of the input sample to perform a look up in
16368 the specified table. If the key is not found in the table, integer value zero
Davor Ocelice9ed2812017-12-25 17:49:28 +010016369 is returned. Otherwise the converter returns the cumulative number of incoming
Willy Tarreaud9f316a2014-07-10 14:03:38 +020016370 sessions associated with the input sample in the designated table. Note that
16371 a session here refers to an incoming connection being accepted by the
16372 "tcp-request connection" rulesets. See also the sc_sess_cnt sample fetch
16373 keyword.
16374
16375table_sess_rate(<table>)
16376 Uses the string representation of the input sample to perform a look up in
16377 the specified table. If the key is not found in the table, integer value zero
16378 is returned. Otherwise the converter returns the average incoming session
16379 rate associated with the input sample in the designated table. Note that a
16380 session here refers to an incoming connection being accepted by the
16381 "tcp-request connection" rulesets. See also the sc_sess_rate sample fetch
16382 keyword.
16383
16384table_trackers(<table>)
16385 Uses the string representation of the input sample to perform a look up in
16386 the specified table. If the key is not found in the table, integer value zero
16387 is returned. Otherwise the converter returns the current amount of concurrent
16388 connections tracking the same key as the input sample in the designated
16389 table. It differs from table_conn_cur in that it does not rely on any stored
16390 information but on the table's reference count (the "use" value which is
16391 returned by "show table" on the CLI). This may sometimes be more suited for
16392 layer7 tracking. It can be used to tell a server how many concurrent
16393 connections there are from a given address for example. See also the
16394 sc_trackers sample fetch keyword.
16395
Willy Tarreauffcb2e42014-07-10 16:29:08 +020016396upper
16397 Convert a string sample to upper case. This can only be placed after a string
16398 sample fetch function or after a transformation keyword returning a string
16399 type. The result is of type string.
16400
Willy Tarreau62ba9ba2020-04-23 17:54:47 +020016401url_dec([<in_form>])
16402 Takes an url-encoded string provided as input and returns the decoded version
16403 as output. The input and the output are of type string. If the <in_form>
16404 argument is set to a non-zero integer value, the input string is assumed to
16405 be part of a form or query string and the '+' character will be turned into a
16406 space (' '). Otherwise this will only happen after a question mark indicating
16407 a query string ('?').
Thierry FOURNIER82ff3c92015-05-07 15:46:20 +020016408
William Dauchy888b0ae2021-01-06 23:39:50 +010016409url_enc([<enc_type>])
16410 Takes a string provided as input and returns the encoded version as output.
16411 The input and the output are of type string. By default the type of encoding
16412 is meant for `query` type. There is no other type supported for now but the
16413 optional argument is here for future changes.
16414
Frédéric Lécaille756d97f2019-03-04 19:03:48 +010016415ungrpc(<field_number>,[<field_type>])
Frédéric Lécaille50290fb2019-02-27 14:34:51 +010016416 This extracts the protocol buffers message field in raw mode of an input binary
Frédéric Lécaillebfe61382019-03-06 14:34:36 +010016417 sample representation of a gRPC message with <field_number> as field number
16418 (dotted notation) if <field_type> is not present, or as an integer sample if this
16419 field is present.
Frédéric Lécaille756d97f2019-03-04 19:03:48 +010016420 The list of the authorized types is the following one: "int32", "int64", "uint32",
16421 "uint64", "sint32", "sint64", "bool", "enum" for the "varint" wire type 0
16422 "fixed64", "sfixed64", "double" for the 64bit wire type 1, "fixed32", "sfixed32",
16423 "float" for the wire type 5. Note that "string" is considered as a length-delimited
Frédéric Lécaille93d33162019-03-06 09:35:59 +010016424 type, so it does not require any <field_type> argument to be extracted.
Frédéric Lécaille756d97f2019-03-04 19:03:48 +010016425 More information may be found here about the protocol buffers message field types:
16426 https://developers.google.com/protocol-buffers/docs/encoding
Frédéric Lécaille50290fb2019-02-27 14:34:51 +010016427
16428 Example:
16429 // with such a protocol buffer .proto file content adapted from
16430 // https://github.com/grpc/grpc/blob/master/examples/protos/route_guide.proto
16431
16432 message Point {
16433 int32 latitude = 1;
16434 int32 longitude = 2;
16435 }
16436
16437 message PPoint {
16438 Point point = 59;
16439 }
16440
16441 message Rectangle {
16442 // One corner of the rectangle.
16443 PPoint lo = 48;
16444 // The other corner of the rectangle.
16445 PPoint hi = 49;
16446 }
16447
Peter Gervaidf4c9d22020-06-11 18:05:11 +020016448 let's say a body request is made of a "Rectangle" object value (two PPoint
16449 protocol buffers messages), the four protocol buffers fields could be
16450 extracted with these "ungrpc" directives:
Frédéric Lécaille50290fb2019-02-27 14:34:51 +010016451
Frédéric Lécaille756d97f2019-03-04 19:03:48 +010016452 req.body,ungrpc(48.59.1,int32) # "latitude" of "lo" first PPoint
16453 req.body,ungrpc(48.59.2,int32) # "longitude" of "lo" first PPoint
John Roeslerfb2fce12019-07-10 15:45:51 -050016454 req.body,ungrpc(49.59.1,int32) # "latitude" of "hi" second PPoint
Frédéric Lécaille756d97f2019-03-04 19:03:48 +010016455 req.body,ungrpc(49.59.2,int32) # "longitude" of "hi" second PPoint
16456
Peter Gervaidf4c9d22020-06-11 18:05:11 +020016457 We could also extract the intermediary 48.59 field as a binary sample as follows:
Frédéric Lécaille756d97f2019-03-04 19:03:48 +010016458
Frédéric Lécaille93d33162019-03-06 09:35:59 +010016459 req.body,ungrpc(48.59)
Frédéric Lécaille756d97f2019-03-04 19:03:48 +010016460
Peter Gervaidf4c9d22020-06-11 18:05:11 +020016461 As a gRPC message is always made of a gRPC header followed by protocol buffers
16462 messages, in the previous example the "latitude" of "lo" first PPoint
16463 could be extracted with these equivalent directives:
Frédéric Lécaillebfe61382019-03-06 14:34:36 +010016464
16465 req.body,ungrpc(48.59),protobuf(1,int32)
16466 req.body,ungrpc(48),protobuf(59.1,int32)
16467 req.body,ungrpc(48),protobuf(59),protobuf(1,int32)
16468
Peter Gervaidf4c9d22020-06-11 18:05:11 +020016469 Note that the first convert must be "ungrpc", the remaining ones must be
16470 "protobuf" and only the last one may have or not a second argument to
16471 interpret the previous binary sample.
Frédéric Lécaillebfe61382019-03-06 14:34:36 +010016472
Frédéric Lécaille50290fb2019-02-27 14:34:51 +010016473
Tim Duesterhusef4e45c2021-01-21 17:40:50 +010016474unset-var(<var>)
Christopher Faulet85d79c92016-11-09 16:54:56 +010016475 Unsets a variable if the input content is defined. The name of the variable
16476 starts with an indication about its scope. The scopes allowed are:
16477 "proc" : the variable is shared with the whole process
16478 "sess" : the variable is shared with the whole session
16479 "txn" : the variable is shared with the transaction (request and
16480 response),
16481 "req" : the variable is shared only during request processing,
16482 "res" : the variable is shared only during response processing.
16483 This prefix is followed by a name. The separator is a '.'. The name may only
16484 contain characters 'a-z', 'A-Z', '0-9', '.' and '_'.
16485
Willy Tarreau0dbfdba2014-07-10 16:37:47 +020016486utime(<format>[,<offset>])
16487 Converts an integer supposed to contain a date since epoch to a string
16488 representing this date in UTC time using a format defined by the <format>
16489 string using strftime(3). The purpose is to allow any date format to be used
16490 in logs. An optional <offset> in seconds may be applied to the input date
16491 (positive or negative). See the strftime() man page for the format supported
16492 by your operating system. See also the ltime converter.
16493
16494 Example :
16495
16496 # Emit two colons, one with the UTC time and another with ip:port
Davor Ocelice9ed2812017-12-25 17:49:28 +010016497 # e.g. 20140710162350 127.0.0.1:57325
Willy Tarreau0dbfdba2014-07-10 16:37:47 +020016498 log-format %[date,utime(%Y%m%d%H%M%S)]\ %ci:%cp
16499
Marcin Deranek9631a282018-04-16 14:30:46 +020016500word(<index>,<delimiters>[,<count>])
16501 Extracts the nth word counting from the beginning (positive index) or from
16502 the end (negative index) considering given delimiters from an input string.
16503 Indexes start at 1 or -1 and delimiters are a string formatted list of chars.
Jerome Magnin88209322020-01-28 13:33:44 +010016504 Delimiters at the beginning or end of the input string are ignored.
Marcin Deranek9631a282018-04-16 14:30:46 +020016505 Optionally you can specify <count> of words to extract (default: 1).
16506 Value of 0 indicates extraction of all remaining words.
16507
16508 Example :
16509 str(f1_f2_f3__f5),word(4,_) # f5
16510 str(f1_f2_f3__f5),word(2,_,0) # f2_f3__f5
16511 str(f1_f2_f3__f5),word(3,_,2) # f3__f5
16512 str(f1_f2_f3__f5),word(-2,_,3) # f1_f2_f3
16513 str(f1_f2_f3__f5),word(-3,_,0) # f1_f2
Jerome Magnin88209322020-01-28 13:33:44 +010016514 str(/f1/f2/f3/f4),word(1,/) # f1
Emeric Brunc9a0f6d2014-11-25 14:09:01 +010016515
Willy Tarreau23ec4ca2014-07-15 20:15:37 +020016516wt6([<avalanche>])
16517 Hashes a binary input sample into an unsigned 32-bit quantity using the WT6
16518 hash function. Optionally, it is possible to apply a full avalanche hash
16519 function to the output if the optional <avalanche> argument equals 1. This
16520 converter uses the same functions as used by the various hash-based load
16521 balancing algorithms, so it will provide exactly the same results. It is
16522 mostly intended for debugging, but can be used as a stick-table entry to
16523 collect rough statistics. It must not be used for security purposes as a
Emmanuel Hocdet50791a72018-03-21 11:19:01 +010016524 32-bit hash is trivial to break. See also "crc32", "djb2", "sdbm", "crc32c",
16525 and the "hash-type" directive.
Willy Tarreau23ec4ca2014-07-15 20:15:37 +020016526
Willy Tarreau97707872015-01-27 15:12:13 +010016527xor(<value>)
16528 Performs a bitwise "XOR" (exclusive OR) between <value> and the input value
Thierry FOURNIER07ee64e2015-07-06 23:43:03 +020016529 of type signed integer, and returns the result as an signed integer.
Thierry FOURNIER5d86fae2015-07-07 21:10:16 +020016530 <value> can be a numeric value or a variable name. The name of the variable
Daniel Schneller0b547052016-03-21 20:46:57 +010016531 starts with an indication about its scope. The scopes allowed are:
Christopher Fauletff2613e2016-11-09 11:36:17 +010016532 "proc" : the variable is shared with the whole process
Daniel Schneller0b547052016-03-21 20:46:57 +010016533 "sess" : the variable is shared with the whole session
16534 "txn" : the variable is shared with the transaction (request and
Thierry FOURNIER5d86fae2015-07-07 21:10:16 +020016535 response),
Daniel Schneller0b547052016-03-21 20:46:57 +010016536 "req" : the variable is shared only during request processing,
16537 "res" : the variable is shared only during response processing.
Thierry FOURNIER5d86fae2015-07-07 21:10:16 +020016538 This prefix is followed by a name. The separator is a '.'. The name may only
Christopher Fauletb71557a2016-10-31 10:49:03 +010016539 contain characters 'a-z', 'A-Z', '0-9', '.' and '_'.
Willy Tarreau97707872015-01-27 15:12:13 +010016540
Dragan Dosen04bf0cc2020-12-22 21:44:33 +010016541xxh3([<seed>])
16542 Hashes a binary input sample into a signed 64-bit quantity using the XXH3
16543 64-bit variant of the XXhash hash function. This hash supports a seed which
16544 defaults to zero but a different value maybe passed as the <seed> argument.
16545 This hash is known to be very good and very fast so it can be used to hash
16546 URLs and/or URL parameters for use as stick-table keys to collect statistics
16547 with a low collision rate, though care must be taken as the algorithm is not
16548 considered as cryptographically secure.
16549
Thierry FOURNIER01e09742016-12-26 11:46:11 +010016550xxh32([<seed>])
16551 Hashes a binary input sample into an unsigned 32-bit quantity using the 32-bit
16552 variant of the XXHash hash function. This hash supports a seed which defaults
16553 to zero but a different value maybe passed as the <seed> argument. This hash
16554 is known to be very good and very fast so it can be used to hash URLs and/or
16555 URL parameters for use as stick-table keys to collect statistics with a low
16556 collision rate, though care must be taken as the algorithm is not considered
16557 as cryptographically secure.
16558
16559xxh64([<seed>])
16560 Hashes a binary input sample into a signed 64-bit quantity using the 64-bit
16561 variant of the XXHash hash function. This hash supports a seed which defaults
16562 to zero but a different value maybe passed as the <seed> argument. This hash
16563 is known to be very good and very fast so it can be used to hash URLs and/or
16564 URL parameters for use as stick-table keys to collect statistics with a low
16565 collision rate, though care must be taken as the algorithm is not considered
16566 as cryptographically secure.
16567
Thierry FOURNIERd5f624d2013-11-26 11:52:33 +010016568
Thierry FOURNIER060762e2014-04-23 13:29:15 +0200165697.3.2. Fetching samples from internal states
Willy Tarreau74ca5042013-06-11 23:12:07 +020016570--------------------------------------------
16571
16572A first set of sample fetch methods applies to internal information which does
16573not even relate to any client information. These ones are sometimes used with
16574"monitor-fail" directives to report an internal status to external watchers.
16575The sample fetch methods described in this section are usable anywhere.
16576
16577always_false : boolean
16578 Always returns the boolean "false" value. It may be used with ACLs as a
16579 temporary replacement for another one when adjusting configurations.
16580
16581always_true : boolean
16582 Always returns the boolean "true" value. It may be used with ACLs as a
16583 temporary replacement for another one when adjusting configurations.
16584
16585avg_queue([<backend>]) : integer
Willy Tarreaud63335a2010-02-26 12:56:52 +010016586 Returns the total number of queued connections of the designated backend
Willy Tarreau74ca5042013-06-11 23:12:07 +020016587 divided by the number of active servers. The current backend is used if no
16588 backend is specified. This is very similar to "queue" except that the size of
16589 the farm is considered, in order to give a more accurate measurement of the
16590 time it may take for a new connection to be processed. The main usage is with
16591 ACL to return a sorry page to new users when it becomes certain they will get
16592 a degraded service, or to pass to the backend servers in a header so that
16593 they decide to work in degraded mode or to disable some functions to speed up
16594 the processing a bit. Note that in the event there would not be any active
16595 server anymore, twice the number of queued connections would be considered as
16596 the measured value. This is a fair estimate, as we expect one server to get
16597 back soon anyway, but we still prefer to send new traffic to another backend
16598 if in better shape. See also the "queue", "be_conn", and "be_sess_rate"
16599 sample fetches.
Krzysztof Piotr Oledzki346f76d2010-01-12 21:59:30 +010016600
Willy Tarreau74ca5042013-06-11 23:12:07 +020016601be_conn([<backend>]) : integer
Willy Tarreaua36af912009-10-10 12:02:45 +020016602 Applies to the number of currently established connections on the backend,
16603 possibly including the connection being evaluated. If no backend name is
16604 specified, the current one is used. But it is also possible to check another
16605 backend. It can be used to use a specific farm when the nominal one is full.
Patrick Hemmer4cdf3ab2018-06-14 17:10:27 -040016606 See also the "fe_conn", "queue", "be_conn_free", and "be_sess_rate" criteria.
16607
16608be_conn_free([<backend>]) : integer
16609 Returns an integer value corresponding to the number of available connections
16610 across available servers in the backend. Queue slots are not included. Backup
16611 servers are also not included, unless all other servers are down. If no
16612 backend name is specified, the current one is used. But it is also possible
16613 to check another backend. It can be used to use a specific farm when the
Patrick Hemmer155e93e2018-06-14 18:01:35 -040016614 nominal one is full. See also the "be_conn", "connslots", and "srv_conn_free"
16615 criteria.
Patrick Hemmer4cdf3ab2018-06-14 17:10:27 -040016616
16617 OTHER CAVEATS AND NOTES: if any of the server maxconn, or maxqueue is 0
16618 (meaning unlimited), then this fetch clearly does not make sense, in which
16619 case the value returned will be -1.
Willy Tarreau6a06a402007-07-15 20:15:28 +020016620
Willy Tarreau74ca5042013-06-11 23:12:07 +020016621be_sess_rate([<backend>]) : integer
16622 Returns an integer value corresponding to the sessions creation rate on the
16623 backend, in number of new sessions per second. This is used with ACLs to
16624 switch to an alternate backend when an expensive or fragile one reaches too
Davor Ocelice9ed2812017-12-25 17:49:28 +010016625 high a session rate, or to limit abuse of service (e.g. prevent sucking of an
Willy Tarreau74ca5042013-06-11 23:12:07 +020016626 online dictionary). It can also be useful to add this element to logs using a
16627 log-format directive.
Willy Tarreaud63335a2010-02-26 12:56:52 +010016628
16629 Example :
16630 # Redirect to an error page if the dictionary is requested too often
16631 backend dynamic
16632 mode http
16633 acl being_scanned be_sess_rate gt 100
16634 redirect location /denied.html if being_scanned
Willy Tarreau0ba27502007-12-24 16:55:16 +010016635
Davor Ocelice9ed2812017-12-25 17:49:28 +010016636bin(<hex>) : bin
Thierry FOURNIERcc103292015-06-06 19:30:17 +020016637 Returns a binary chain. The input is the hexadecimal representation
16638 of the string.
16639
16640bool(<bool>) : bool
16641 Returns a boolean value. <bool> can be 'true', 'false', '1' or '0'.
16642 'false' and '0' are the same. 'true' and '1' are the same.
16643
Willy Tarreau74ca5042013-06-11 23:12:07 +020016644connslots([<backend>]) : integer
16645 Returns an integer value corresponding to the number of connection slots
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030016646 still available in the backend, by totaling the maximum amount of
Willy Tarreau74ca5042013-06-11 23:12:07 +020016647 connections on all servers and the maximum queue size. This is probably only
16648 used with ACLs.
Tait Clarridge7896d522012-12-05 21:39:31 -050016649
Jeffrey 'jf' Lim5051d7b2008-09-04 01:03:03 +080016650 The basic idea here is to be able to measure the number of connection "slots"
Willy Tarreau55165fe2009-05-10 12:02:55 +020016651 still available (connection + queue), so that anything beyond that (intended
Jeffrey 'jf' Lim5051d7b2008-09-04 01:03:03 +080016652 usage; see "use_backend" keyword) can be redirected to a different backend.
16653
Willy Tarreau55165fe2009-05-10 12:02:55 +020016654 'connslots' = number of available server connection slots, + number of
16655 available server queue slots.
Jeffrey 'jf' Lim5051d7b2008-09-04 01:03:03 +080016656
Willy Tarreaua36af912009-10-10 12:02:45 +020016657 Note that while "fe_conn" may be used, "connslots" comes in especially
Willy Tarreau55165fe2009-05-10 12:02:55 +020016658 useful when you have a case of traffic going to one single ip, splitting into
Willy Tarreau74ca5042013-06-11 23:12:07 +020016659 multiple backends (perhaps using ACLs to do name-based load balancing) and
Willy Tarreau55165fe2009-05-10 12:02:55 +020016660 you want to be able to differentiate between different backends, and their
Davor Ocelice9ed2812017-12-25 17:49:28 +010016661 available "connslots". Also, whereas "nbsrv" only measures servers that are
Willy Tarreau74ca5042013-06-11 23:12:07 +020016662 actually *down*, this fetch is more fine-grained and looks into the number of
Willy Tarreaua36af912009-10-10 12:02:45 +020016663 available connection slots as well. See also "queue" and "avg_queue".
Jeffrey 'jf' Lim5051d7b2008-09-04 01:03:03 +080016664
Willy Tarreau55165fe2009-05-10 12:02:55 +020016665 OTHER CAVEATS AND NOTES: at this point in time, the code does not take care
16666 of dynamic connections. Also, if any of the server maxconn, or maxqueue is 0,
Willy Tarreau74ca5042013-06-11 23:12:07 +020016667 then this fetch clearly does not make sense, in which case the value returned
Willy Tarreau55165fe2009-05-10 12:02:55 +020016668 will be -1.
Jeffrey 'jf' Lim5051d7b2008-09-04 01:03:03 +080016669
Willy Tarreau70fe9442018-11-22 16:07:39 +010016670cpu_calls : integer
16671 Returns the number of calls to the task processing the stream or current
16672 request since it was allocated. This number is reset for each new request on
16673 the same connections in case of HTTP keep-alive. This value should usually be
16674 low and stable (around 2 calls for a typically simple request) but may become
16675 high if some processing (compression, caching or analysis) is performed. This
16676 is purely for performance monitoring purposes.
16677
16678cpu_ns_avg : integer
16679 Returns the average number of nanoseconds spent in each call to the task
16680 processing the stream or current request. This number is reset for each new
16681 request on the same connections in case of HTTP keep-alive. This value
16682 indicates the overall cost of processing the request or the connection for
16683 each call. There is no good nor bad value but the time spent in a call
16684 automatically causes latency for other processing (see lat_ns_avg below),
16685 and may affect other connection's apparent response time. Certain operations
16686 like compression, complex regex matching or heavy Lua operations may directly
16687 affect this value, and having it in the logs will make it easier to spot the
16688 faulty processing that needs to be fixed to recover decent performance.
16689 Note: this value is exactly cpu_ns_tot divided by cpu_calls.
16690
16691cpu_ns_tot : integer
16692 Returns the total number of nanoseconds spent in each call to the task
16693 processing the stream or current request. This number is reset for each new
16694 request on the same connections in case of HTTP keep-alive. This value
16695 indicates the overall cost of processing the request or the connection for
16696 each call. There is no good nor bad value but the time spent in a call
16697 automatically causes latency for other processing (see lat_ns_avg below),
16698 induces CPU costs on the machine, and may affect other connection's apparent
16699 response time. Certain operations like compression, complex regex matching or
16700 heavy Lua operations may directly affect this value, and having it in the
16701 logs will make it easier to spot the faulty processing that needs to be fixed
16702 to recover decent performance. The value may be artificially high due to a
16703 high cpu_calls count, for example when processing many HTTP chunks, and for
16704 this reason it is often preferred to log cpu_ns_avg instead.
16705
Cyril Bonté6bcd1822019-11-05 23:13:59 +010016706date([<offset>],[<unit>]) : integer
Willy Tarreau6236d3a2013-07-25 14:28:25 +020016707 Returns the current date as the epoch (number of seconds since 01/01/1970).
Damien Claisseae6f1252019-10-30 15:57:28 +000016708
16709 If an offset value is specified, then it is added to the current date before
16710 returning the value. This is particularly useful to compute relative dates,
16711 as both positive and negative offsets are allowed.
Willy Tarreau276fae92013-07-25 14:36:01 +020016712 It is useful combined with the http_date converter.
16713
Damien Claisseae6f1252019-10-30 15:57:28 +000016714 <unit> is facultative, and can be set to "s" for seconds (default behavior),
16715 "ms" for milliseconds or "us" for microseconds.
16716 If unit is set, return value is an integer reflecting either seconds,
16717 milliseconds or microseconds since epoch, plus offset.
16718 It is useful when a time resolution of less than a second is needed.
16719
Willy Tarreau276fae92013-07-25 14:36:01 +020016720 Example :
16721
16722 # set an expires header to now+1 hour in every response
16723 http-response set-header Expires %[date(3600),http_date]
Willy Tarreau6236d3a2013-07-25 14:28:25 +020016724
Damien Claisseae6f1252019-10-30 15:57:28 +000016725 # set an expires header to now+1 hour in every response, with
16726 # millisecond granularity
16727 http-response set-header Expires %[date(3600000,ms),http_date(0,ms)]
16728
Etienne Carrierea792a0a2018-01-17 13:43:24 +010016729date_us : integer
16730 Return the microseconds part of the date (the "second" part is returned by
16731 date sample). This sample is coherent with the date sample as it is comes
16732 from the same timeval structure.
16733
Willy Tarreaud716f9b2017-10-13 11:03:15 +020016734distcc_body(<token>[,<occ>]) : binary
16735 Parses a distcc message and returns the body associated to occurrence #<occ>
16736 of the token <token>. Occurrences start at 1, and when unspecified, any may
16737 match though in practice only the first one is checked for now. This can be
16738 used to extract file names or arguments in files built using distcc through
16739 haproxy. Please refer to distcc's protocol documentation for the complete
16740 list of supported tokens.
16741
16742distcc_param(<token>[,<occ>]) : integer
16743 Parses a distcc message and returns the parameter associated to occurrence
16744 #<occ> of the token <token>. Occurrences start at 1, and when unspecified,
16745 any may match though in practice only the first one is checked for now. This
16746 can be used to extract certain information such as the protocol version, the
16747 file size or the argument in files built using distcc through haproxy.
16748 Another use case consists in waiting for the start of the preprocessed file
16749 contents before connecting to the server to avoid keeping idle connections.
16750 Please refer to distcc's protocol documentation for the complete list of
16751 supported tokens.
16752
16753 Example :
16754 # wait up to 20s for the pre-processed file to be uploaded
16755 tcp-request inspect-delay 20s
16756 tcp-request content accept if { distcc_param(DOTI) -m found }
16757 # send large files to the big farm
16758 use_backend big_farm if { distcc_param(DOTI) gt 1000000 }
16759
Willy Tarreau595ec542013-06-12 21:34:28 +020016760env(<name>) : string
16761 Returns a string containing the value of environment variable <name>. As a
16762 reminder, environment variables are per-process and are sampled when the
16763 process starts. This can be useful to pass some information to a next hop
16764 server, or with ACLs to take specific action when the process is started a
16765 certain way.
16766
16767 Examples :
16768 # Pass the Via header to next hop with the local hostname in it
16769 http-request add-header Via 1.1\ %[env(HOSTNAME)]
16770
16771 # reject cookie-less requests when the STOP environment variable is set
16772 http-request deny if !{ cook(SESSIONID) -m found } { env(STOP) -m found }
16773
Willy Tarreau74ca5042013-06-11 23:12:07 +020016774fe_conn([<frontend>]) : integer
16775 Returns the number of currently established connections on the frontend,
Willy Tarreaud63335a2010-02-26 12:56:52 +010016776 possibly including the connection being evaluated. If no frontend name is
16777 specified, the current one is used. But it is also possible to check another
Willy Tarreau74ca5042013-06-11 23:12:07 +020016778 frontend. It can be used to return a sorry page before hard-blocking, or to
16779 use a specific backend to drain new requests when the farm is considered
Davor Ocelice9ed2812017-12-25 17:49:28 +010016780 full. This is mostly used with ACLs but can also be used to pass some
Willy Tarreau74ca5042013-06-11 23:12:07 +020016781 statistics to servers in HTTP headers. See also the "dst_conn", "be_conn",
16782 "fe_sess_rate" fetches.
Willy Tarreaua36af912009-10-10 12:02:45 +020016783
Nenad Merdanovicad9a7e92016-10-03 04:57:37 +020016784fe_req_rate([<frontend>]) : integer
16785 Returns an integer value corresponding to the number of HTTP requests per
16786 second sent to a frontend. This number can differ from "fe_sess_rate" in
16787 situations where client-side keep-alive is enabled.
16788
Willy Tarreau74ca5042013-06-11 23:12:07 +020016789fe_sess_rate([<frontend>]) : integer
16790 Returns an integer value corresponding to the sessions creation rate on the
16791 frontend, in number of new sessions per second. This is used with ACLs to
16792 limit the incoming session rate to an acceptable range in order to prevent
16793 abuse of service at the earliest moment, for example when combined with other
16794 layer 4 ACLs in order to force the clients to wait a bit for the rate to go
16795 down below the limit. It can also be useful to add this element to logs using
16796 a log-format directive. See also the "rate-limit sessions" directive for use
16797 in frontends.
Willy Tarreau079ff0a2009-03-05 21:34:28 +010016798
16799 Example :
16800 # This frontend limits incoming mails to 10/s with a max of 100
16801 # concurrent connections. We accept any connection below 10/s, and
16802 # force excess clients to wait for 100 ms. Since clients are limited to
16803 # 100 max, there cannot be more than 10 incoming mails per second.
16804 frontend mail
16805 bind :25
16806 mode tcp
16807 maxconn 100
16808 acl too_fast fe_sess_rate ge 10
16809 tcp-request inspect-delay 100ms
16810 tcp-request content accept if ! too_fast
16811 tcp-request content accept if WAIT_END
Willy Tarreaud72758d2010-01-12 10:42:19 +010016812
Nenad Merdanovic807a6e72017-03-12 22:00:00 +010016813hostname : string
16814 Returns the system hostname.
16815
Thierry FOURNIER07ee64e2015-07-06 23:43:03 +020016816int(<integer>) : signed integer
16817 Returns a signed integer.
16818
Thierry FOURNIERcc103292015-06-06 19:30:17 +020016819ipv4(<ipv4>) : ipv4
16820 Returns an ipv4.
16821
16822ipv6(<ipv6>) : ipv6
16823 Returns an ipv6.
16824
Willy Tarreau70fe9442018-11-22 16:07:39 +010016825lat_ns_avg : integer
16826 Returns the average number of nanoseconds spent between the moment the task
16827 handling the stream is woken up and the moment it is effectively called. This
16828 number is reset for each new request on the same connections in case of HTTP
16829 keep-alive. This value indicates the overall latency inflicted to the current
16830 request by all other requests being processed in parallel, and is a direct
16831 indicator of perceived performance due to noisy neighbours. In order to keep
16832 the value low, it is possible to reduce the scheduler's run queue depth using
16833 "tune.runqueue-depth", to reduce the number of concurrent events processed at
16834 once using "tune.maxpollevents", to decrease the stream's nice value using
Willy Tarreaue7723bd2020-06-24 11:11:02 +020016835 the "nice" option on the "bind" lines or in the frontend, to enable low
16836 latency scheduling using "tune.sched.low-latency", or to look for other heavy
16837 requests in logs (those exhibiting large values of "cpu_ns_avg"), whose
16838 processing needs to be adjusted or fixed. Compression of large buffers could
16839 be a culprit, like heavy regex or long lists of regex. Note: this value is
16840 exactly lat_ns_tot divided by cpu_calls.
Willy Tarreau70fe9442018-11-22 16:07:39 +010016841
16842lat_ns_tot : integer
16843 Returns the total number of nanoseconds spent between the moment the task
16844 handling the stream is woken up and the moment it is effectively called. This
16845 number is reset for each new request on the same connections in case of HTTP
16846 keep-alive. This value indicates the overall latency inflicted to the current
16847 request by all other requests being processed in parallel, and is a direct
16848 indicator of perceived performance due to noisy neighbours. In order to keep
16849 the value low, it is possible to reduce the scheduler's run queue depth using
16850 "tune.runqueue-depth", to reduce the number of concurrent events processed at
16851 once using "tune.maxpollevents", to decrease the stream's nice value using
Willy Tarreaue7723bd2020-06-24 11:11:02 +020016852 the "nice" option on the "bind" lines or in the frontend, to enable low
16853 latency scheduling using "tune.sched.low-latency", or to look for other heavy
16854 requests in logs (those exhibiting large values of "cpu_ns_avg"), whose
16855 processing needs to be adjusted or fixed. Compression of large buffers could
16856 be a culprit, like heavy regex or long lists of regex. Note: while it
Willy Tarreau70fe9442018-11-22 16:07:39 +010016857 may intuitively seem that the total latency adds to a transfer time, it is
16858 almost never true because while a task waits for the CPU, network buffers
16859 continue to fill up and the next call will process more at once. The value
16860 may be artificially high due to a high cpu_calls count, for example when
16861 processing many HTTP chunks, and for this reason it is often preferred to log
16862 lat_ns_avg instead, which is a more relevant performance indicator.
16863
Thierry FOURNIERcc103292015-06-06 19:30:17 +020016864meth(<method>) : method
16865 Returns a method.
16866
Willy Tarreau0f30d262014-11-24 16:02:05 +010016867nbproc : integer
16868 Returns an integer value corresponding to the number of processes that were
16869 started (it equals the global "nbproc" setting). This is useful for logging
16870 and debugging purposes.
16871
Willy Tarreau74ca5042013-06-11 23:12:07 +020016872nbsrv([<backend>]) : integer
16873 Returns an integer value corresponding to the number of usable servers of
16874 either the current backend or the named backend. This is mostly used with
16875 ACLs but can also be useful when added to logs. This is normally used to
Willy Tarreaud63335a2010-02-26 12:56:52 +010016876 switch to an alternate backend when the number of servers is too low to
16877 to handle some load. It is useful to report a failure when combined with
16878 "monitor fail".
Willy Tarreau079ff0a2009-03-05 21:34:28 +010016879
Patrick Hemmerfabb24f2018-08-13 14:07:57 -040016880prio_class : integer
16881 Returns the priority class of the current session for http mode or connection
16882 for tcp mode. The value will be that set by the last call to "http-request
16883 set-priority-class" or "tcp-request content set-priority-class".
16884
16885prio_offset : integer
16886 Returns the priority offset of the current session for http mode or
16887 connection for tcp mode. The value will be that set by the last call to
16888 "http-request set-priority-offset" or "tcp-request content
16889 set-priority-offset".
16890
Willy Tarreau0f30d262014-11-24 16:02:05 +010016891proc : integer
16892 Returns an integer value corresponding to the position of the process calling
16893 the function, between 1 and global.nbproc. This is useful for logging and
16894 debugging purposes.
16895
Willy Tarreau74ca5042013-06-11 23:12:07 +020016896queue([<backend>]) : integer
Willy Tarreaud63335a2010-02-26 12:56:52 +010016897 Returns the total number of queued connections of the designated backend,
16898 including all the connections in server queues. If no backend name is
16899 specified, the current one is used, but it is also possible to check another
Willy Tarreau74ca5042013-06-11 23:12:07 +020016900 one. This is useful with ACLs or to pass statistics to backend servers. This
16901 can be used to take actions when queuing goes above a known level, generally
16902 indicating a surge of traffic or a massive slowdown on the servers. One
16903 possible action could be to reject new users but still accept old ones. See
16904 also the "avg_queue", "be_conn", and "be_sess_rate" fetches.
16905
Willy Tarreau84310e22014-02-14 11:59:04 +010016906rand([<range>]) : integer
16907 Returns a random integer value within a range of <range> possible values,
16908 starting at zero. If the range is not specified, it defaults to 2^32, which
16909 gives numbers between 0 and 4294967295. It can be useful to pass some values
16910 needed to take some routing decisions for example, or just for debugging
16911 purposes. This random must not be used for security purposes.
16912
Luca Schimweg8a694b82019-09-10 15:42:52 +020016913uuid([<version>]) : string
16914 Returns a UUID following the RFC4122 standard. If the version is not
16915 specified, a UUID version 4 (fully random) is returned.
16916 Currently, only version 4 is supported.
16917
Willy Tarreau74ca5042013-06-11 23:12:07 +020016918srv_conn([<backend>/]<server>) : integer
16919 Returns an integer value corresponding to the number of currently established
16920 connections on the designated server, possibly including the connection being
16921 evaluated. If <backend> is omitted, then the server is looked up in the
16922 current backend. It can be used to use a specific farm when one server is
16923 full, or to inform the server about our view of the number of active
Patrick Hemmer155e93e2018-06-14 18:01:35 -040016924 connections with it. See also the "fe_conn", "be_conn", "queue", and
16925 "srv_conn_free" fetch methods.
16926
16927srv_conn_free([<backend>/]<server>) : integer
16928 Returns an integer value corresponding to the number of available connections
16929 on the designated server, possibly including the connection being evaluated.
16930 The value does not include queue slots. If <backend> is omitted, then the
16931 server is looked up in the current backend. It can be used to use a specific
16932 farm when one server is full, or to inform the server about our view of the
16933 number of active connections with it. See also the "be_conn_free" and
16934 "srv_conn" fetch methods.
16935
16936 OTHER CAVEATS AND NOTES: If the server maxconn is 0, then this fetch clearly
16937 does not make sense, in which case the value returned will be -1.
Willy Tarreau74ca5042013-06-11 23:12:07 +020016938
16939srv_is_up([<backend>/]<server>) : boolean
16940 Returns true when the designated server is UP, and false when it is either
16941 DOWN or in maintenance mode. If <backend> is omitted, then the server is
16942 looked up in the current backend. It is mainly used to take action based on
Davor Ocelice9ed2812017-12-25 17:49:28 +010016943 an external status reported via a health check (e.g. a geographical site's
Willy Tarreau74ca5042013-06-11 23:12:07 +020016944 availability). Another possible use which is more of a hack consists in
16945 using dummy servers as boolean variables that can be enabled or disabled from
16946 the CLI, so that rules depending on those ACLs can be tweaked in realtime.
16947
Willy Tarreauff2b7af2017-10-13 11:46:26 +020016948srv_queue([<backend>/]<server>) : integer
16949 Returns an integer value corresponding to the number of connections currently
16950 pending in the designated server's queue. If <backend> is omitted, then the
16951 server is looked up in the current backend. It can sometimes be used together
16952 with the "use-server" directive to force to use a known faster server when it
16953 is not much loaded. See also the "srv_conn", "avg_queue" and "queue" sample
16954 fetch methods.
16955
Willy Tarreau74ca5042013-06-11 23:12:07 +020016956srv_sess_rate([<backend>/]<server>) : integer
16957 Returns an integer corresponding to the sessions creation rate on the
16958 designated server, in number of new sessions per second. If <backend> is
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030016959 omitted, then the server is looked up in the current backend. This is mostly
Willy Tarreau74ca5042013-06-11 23:12:07 +020016960 used with ACLs but can make sense with logs too. This is used to switch to an
16961 alternate backend when an expensive or fragile one reaches too high a session
Davor Ocelice9ed2812017-12-25 17:49:28 +010016962 rate, or to limit abuse of service (e.g. prevent latent requests from
Willy Tarreau74ca5042013-06-11 23:12:07 +020016963 overloading servers).
16964
16965 Example :
16966 # Redirect to a separate back
16967 acl srv1_full srv_sess_rate(be1/srv1) gt 50
16968 acl srv2_full srv_sess_rate(be1/srv2) gt 50
16969 use_backend be2 if srv1_full or srv2_full
16970
Christopher Faulet1bea8652020-07-10 16:03:45 +020016971srv_iweight([<backend>/]<server>): integer
16972 Returns an integer corresponding to the server's initial weight. If <backend>
16973 is omitted, then the server is looked up in the current backend. See also
16974 "srv_weight" and "srv_uweight".
16975
16976srv_uweight([<backend>/]<server>): integer
16977 Returns an integer corresponding to the user visible server's weight. If
16978 <backend> is omitted, then the server is looked up in the current
16979 backend. See also "srv_weight" and "srv_iweight".
16980
16981srv_weight([<backend>/]<server>): integer
16982 Returns an integer corresponding to the current (or effective) server's
16983 weight. If <backend> is omitted, then the server is looked up in the current
16984 backend. See also "srv_iweight" and "srv_uweight".
16985
Willy Tarreau0f30d262014-11-24 16:02:05 +010016986stopping : boolean
16987 Returns TRUE if the process calling the function is currently stopping. This
16988 can be useful for logging, or for relaxing certain checks or helping close
16989 certain connections upon graceful shutdown.
16990
Thierry FOURNIERcc103292015-06-06 19:30:17 +020016991str(<string>) : string
16992 Returns a string.
16993
Willy Tarreau74ca5042013-06-11 23:12:07 +020016994table_avl([<table>]) : integer
16995 Returns the total number of available entries in the current proxy's
16996 stick-table or in the designated stick-table. See also table_cnt.
16997
16998table_cnt([<table>]) : integer
16999 Returns the total number of entries currently in use in the current proxy's
17000 stick-table or in the designated stick-table. See also src_conn_cnt and
17001 table_avl for other entry counting methods.
17002
Christopher Faulet34adb2a2017-11-21 21:45:38 +010017003thread : integer
17004 Returns an integer value corresponding to the position of the thread calling
17005 the function, between 0 and (global.nbthread-1). This is useful for logging
17006 and debugging purposes.
17007
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020017008var(<var-name>) : undefined
17009 Returns a variable with the stored type. If the variable is not set, the
Daniel Schneller0b547052016-03-21 20:46:57 +010017010 sample fetch fails. The name of the variable starts with an indication
17011 about its scope. The scopes allowed are:
Christopher Fauletff2613e2016-11-09 11:36:17 +010017012 "proc" : the variable is shared with the whole process
Daniel Schneller0b547052016-03-21 20:46:57 +010017013 "sess" : the variable is shared with the whole session
17014 "txn" : the variable is shared with the transaction (request and
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020017015 response),
Daniel Schneller0b547052016-03-21 20:46:57 +010017016 "req" : the variable is shared only during request processing,
17017 "res" : the variable is shared only during response processing.
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020017018 This prefix is followed by a name. The separator is a '.'. The name may only
Christopher Fauletb71557a2016-10-31 10:49:03 +010017019 contain characters 'a-z', 'A-Z', '0-9', '.' and '_'.
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020017020
Thierry FOURNIER060762e2014-04-23 13:29:15 +0200170217.3.3. Fetching samples at Layer 4
Willy Tarreau74ca5042013-06-11 23:12:07 +020017022----------------------------------
17023
17024The layer 4 usually describes just the transport layer which in haproxy is
17025closest to the connection, where no content is yet made available. The fetch
17026methods described here are usable as low as the "tcp-request connection" rule
17027sets unless they require some future information. Those generally include
17028TCP/IP addresses and ports, as well as elements from stick-tables related to
Willy Tarreau4d4149c2013-07-23 19:33:46 +020017029the incoming connection. For retrieving a value from a sticky counters, the
17030counter number can be explicitly set as 0, 1, or 2 using the pre-defined
Moemen MHEDHBI9cf46342018-09-25 17:50:53 +020017031"sc0_", "sc1_", or "sc2_" prefix. These three pre-defined prefixes can only be
17032used if MAX_SESS_STKCTR value does not exceed 3, otherwise the counter number
17033can be specified as the first integer argument when using the "sc_" prefix.
17034Starting from "sc_0" to "sc_N" where N is (MAX_SESS_STKCTR-1). An optional
17035table may be specified with the "sc*" form, in which case the currently
17036tracked key will be looked up into this alternate table instead of the table
17037currently being tracked.
Willy Tarreau74ca5042013-06-11 23:12:07 +020017038
Jérôme Magnin35e53a62019-01-16 14:38:37 +010017039bc_http_major : integer
Jérôme Magnin86577422018-12-07 09:03:11 +010017040 Returns the backend connection's HTTP major version encoding, which may be 1
17041 for HTTP/0.9 to HTTP/1.1 or 2 for HTTP/2. Note, this is based on the on-wire
17042 encoding and not the version present in the request header.
17043
Willy Tarreau74ca5042013-06-11 23:12:07 +020017044be_id : integer
17045 Returns an integer containing the current backend's id. It can be used in
Christopher Fauletd1b44642020-04-30 09:51:15 +020017046 frontends with responses to check which backend processed the request. It can
17047 also be used in a tcp-check or an http-check ruleset.
Willy Tarreau74ca5042013-06-11 23:12:07 +020017048
Marcin Deranekd2471c22016-12-12 14:08:05 +010017049be_name : string
17050 Returns a string containing the current backend's name. It can be used in
Christopher Fauletd1b44642020-04-30 09:51:15 +020017051 frontends with responses to check which backend processed the request. It can
17052 also be used in a tcp-check or an http-check ruleset.
Marcin Deranekd2471c22016-12-12 14:08:05 +010017053
Amaury Denoyelled91d7792020-12-10 13:43:56 +010017054be_server_timeout : integer
17055 Returns the configuration value in millisecond for the server timeout of the
17056 current backend. This timeout can be overwritten by a "set-timeout" rule. See
17057 also the "cur_server_timeout".
17058
17059be_tunnel_timeout : integer
17060 Returns the configuration value in millisecond for the tunnel timeout of the
17061 current backend. This timeout can be overwritten by a "set-timeout" rule. See
17062 also the "cur_tunnel_timeout".
17063
Amaury Denoyellef7719a22020-12-10 13:43:58 +010017064cur_server_timeout : integer
17065 Returns the currently applied server timeout in millisecond for the stream.
17066 In the default case, this will be equal to be_server_timeout unless a
17067 "set-timeout" rule has been applied. See also "be_server_timeout".
17068
17069cur_tunnel_timeout : integer
17070 Returns the currently applied tunnel timeout in millisecond for the stream.
17071 In the default case, this will be equal to be_tunnel_timeout unless a
17072 "set-timeout" rule has been applied. See also "be_tunnel_timeout".
17073
Willy Tarreau74ca5042013-06-11 23:12:07 +020017074dst : ip
17075 This is the destination IPv4 address of the connection on the client side,
17076 which is the address the client connected to. It can be useful when running
17077 in transparent mode. It is of type IP and works on both IPv4 and IPv6 tables.
17078 On IPv6 tables, IPv4 address is mapped to its IPv6 equivalent, according to
Willy Tarreau64ded3d2019-01-23 10:02:15 +010017079 RFC 4291. When the incoming connection passed through address translation or
17080 redirection involving connection tracking, the original destination address
17081 before the redirection will be reported. On Linux systems, the source and
17082 destination may seldom appear reversed if the nf_conntrack_tcp_loose sysctl
17083 is set, because a late response may reopen a timed out connection and switch
17084 what is believed to be the source and the destination.
Willy Tarreau74ca5042013-06-11 23:12:07 +020017085
17086dst_conn : integer
17087 Returns an integer value corresponding to the number of currently established
17088 connections on the same socket including the one being evaluated. It is
17089 normally used with ACLs but can as well be used to pass the information to
17090 servers in an HTTP header or in logs. It can be used to either return a sorry
17091 page before hard-blocking, or to use a specific backend to drain new requests
17092 when the socket is considered saturated. This offers the ability to assign
17093 different limits to different listening ports or addresses. See also the
17094 "fe_conn" and "be_conn" fetches.
Willy Tarreaud63335a2010-02-26 12:56:52 +010017095
Willy Tarreau16e01562016-08-09 16:46:18 +020017096dst_is_local : boolean
17097 Returns true if the destination address of the incoming connection is local
17098 to the system, or false if the address doesn't exist on the system, meaning
17099 that it was intercepted in transparent mode. It can be useful to apply
17100 certain rules by default to forwarded traffic and other rules to the traffic
Davor Ocelice9ed2812017-12-25 17:49:28 +010017101 targeting the real address of the machine. For example the stats page could
Willy Tarreau16e01562016-08-09 16:46:18 +020017102 be delivered only on this address, or SSH access could be locally redirected.
17103 Please note that the check involves a few system calls, so it's better to do
17104 it only once per connection.
17105
Willy Tarreau74ca5042013-06-11 23:12:07 +020017106dst_port : integer
17107 Returns an integer value corresponding to the destination TCP port of the
17108 connection on the client side, which is the port the client connected to.
17109 This might be used when running in transparent mode, when assigning dynamic
17110 ports to some clients for a whole application session, to stick all users to
17111 a same server, or to pass the destination port information to a server using
17112 an HTTP header.
17113
Willy Tarreau60ca10a2017-08-18 15:26:54 +020017114fc_http_major : integer
17115 Reports the front connection's HTTP major version encoding, which may be 1
17116 for HTTP/0.9 to HTTP/1.1 or 2 for HTTP/2. Note, this is based on the on-wire
17117 encoding and not on the version present in the request header.
17118
Geoff Simmons7185b782019-08-27 18:31:16 +020017119fc_pp_authority : string
17120 Returns the authority TLV sent by the client in the PROXY protocol header,
17121 if any.
17122
Tim Duesterhusd1b15b62020-03-13 12:34:23 +010017123fc_pp_unique_id : string
17124 Returns the unique ID TLV sent by the client in the PROXY protocol header,
17125 if any.
17126
Emeric Brun4f603012017-01-05 15:11:44 +010017127fc_rcvd_proxy : boolean
17128 Returns true if the client initiated the connection with a PROXY protocol
17129 header.
17130
Thierry Fournier / OZON.IO6310bef2016-07-24 20:16:50 +020017131fc_rtt(<unit>) : integer
17132 Returns the Round Trip Time (RTT) measured by the kernel for the client
17133 connection. <unit> is facultative, by default the unit is milliseconds. <unit>
17134 can be set to "ms" for milliseconds or "us" for microseconds. If the server
17135 connection is not established, if the connection is not TCP or if the
17136 operating system does not support TCP_INFO, for example Linux kernels before
17137 2.4, the sample fetch fails.
17138
17139fc_rttvar(<unit>) : integer
17140 Returns the Round Trip Time (RTT) variance measured by the kernel for the
17141 client connection. <unit> is facultative, by default the unit is milliseconds.
17142 <unit> can be set to "ms" for milliseconds or "us" for microseconds. If the
17143 server connection is not established, if the connection is not TCP or if the
17144 operating system does not support TCP_INFO, for example Linux kernels before
17145 2.4, the sample fetch fails.
17146
Christopher Fauletba0c53e2019-10-17 14:40:48 +020017147fc_unacked : integer
Joe Williams30fcd392016-08-10 07:06:44 -070017148 Returns the unacked counter measured by the kernel for the client connection.
17149 If the server connection is not established, if the connection is not TCP or
17150 if the operating system does not support TCP_INFO, for example Linux kernels
17151 before 2.4, the sample fetch fails.
17152
Christopher Fauletba0c53e2019-10-17 14:40:48 +020017153fc_sacked : integer
Joe Williams30fcd392016-08-10 07:06:44 -070017154 Returns the sacked counter measured by the kernel for the client connection.
17155 If the server connection is not established, if the connection is not TCP or
17156 if the operating system does not support TCP_INFO, for example Linux kernels
17157 before 2.4, the sample fetch fails.
17158
Christopher Fauletba0c53e2019-10-17 14:40:48 +020017159fc_retrans : integer
Joe Williams30fcd392016-08-10 07:06:44 -070017160 Returns the retransmits counter measured by the kernel for the client
17161 connection. If the server connection is not established, if the connection is
17162 not TCP or if the operating system does not support TCP_INFO, for example
17163 Linux kernels before 2.4, the sample fetch fails.
17164
Christopher Fauletba0c53e2019-10-17 14:40:48 +020017165fc_fackets : integer
Joe Williams30fcd392016-08-10 07:06:44 -070017166 Returns the fack counter measured by the kernel for the client
17167 connection. If the server connection is not established, if the connection is
17168 not TCP or if the operating system does not support TCP_INFO, for example
17169 Linux kernels before 2.4, the sample fetch fails.
17170
Christopher Fauletba0c53e2019-10-17 14:40:48 +020017171fc_lost : integer
Joe Williams30fcd392016-08-10 07:06:44 -070017172 Returns the lost counter measured by the kernel for the client
17173 connection. If the server connection is not established, if the connection is
17174 not TCP or if the operating system does not support TCP_INFO, for example
17175 Linux kernels before 2.4, the sample fetch fails.
17176
Christopher Fauletba0c53e2019-10-17 14:40:48 +020017177fc_reordering : integer
Joe Williams30fcd392016-08-10 07:06:44 -070017178 Returns the reordering counter measured by the kernel for the client
17179 connection. If the server connection is not established, if the connection is
17180 not TCP or if the operating system does not support TCP_INFO, for example
17181 Linux kernels before 2.4, the sample fetch fails.
17182
Marcin Deranek9a66dfb2018-04-13 14:37:50 +020017183fe_defbe : string
17184 Returns a string containing the frontend's default backend name. It can be
17185 used in frontends to check which backend will handle requests by default.
17186
Willy Tarreau74ca5042013-06-11 23:12:07 +020017187fe_id : integer
17188 Returns an integer containing the current frontend's id. It can be used in
Marcin Deranek6e413ed2016-12-13 12:40:01 +010017189 backends to check from which frontend it was called, or to stick all users
Willy Tarreau74ca5042013-06-11 23:12:07 +020017190 coming via a same frontend to the same server.
17191
Marcin Deranekd2471c22016-12-12 14:08:05 +010017192fe_name : string
17193 Returns a string containing the current frontend's name. It can be used in
17194 backends to check from which frontend it was called, or to stick all users
17195 coming via a same frontend to the same server.
17196
Amaury Denoyelleda184d52020-12-10 13:43:55 +010017197fe_client_timeout : integer
17198 Returns the configuration value in millisecond for the client timeout of the
17199 current frontend.
17200
Cyril Bonté62ba8702014-04-22 23:52:25 +020017201sc_bytes_in_rate(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020017202sc0_bytes_in_rate([<table>]) : integer
17203sc1_bytes_in_rate([<table>]) : integer
17204sc2_bytes_in_rate([<table>]) : integer
Willy Tarreaue9656522010-08-17 15:40:09 +020017205 Returns the average client-to-server bytes rate from the currently tracked
17206 counters, measured in amount of bytes over the period configured in the
17207 table. See also src_bytes_in_rate.
17208
Cyril Bonté62ba8702014-04-22 23:52:25 +020017209sc_bytes_out_rate(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020017210sc0_bytes_out_rate([<table>]) : integer
17211sc1_bytes_out_rate([<table>]) : integer
17212sc2_bytes_out_rate([<table>]) : integer
Willy Tarreaue9656522010-08-17 15:40:09 +020017213 Returns the average server-to-client bytes rate from the currently tracked
17214 counters, measured in amount of bytes over the period configured in the
17215 table. See also src_bytes_out_rate.
17216
Cyril Bonté62ba8702014-04-22 23:52:25 +020017217sc_clr_gpc0(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020017218sc0_clr_gpc0([<table>]) : integer
17219sc1_clr_gpc0([<table>]) : integer
17220sc2_clr_gpc0([<table>]) : integer
Willy Tarreauf73cd112011-08-13 01:45:16 +020017221 Clears the first General Purpose Counter associated to the currently tracked
17222 counters, and returns its previous value. Before the first invocation, the
Willy Tarreau869948b2013-01-04 14:14:57 +010017223 stored value is zero, so first invocation will always return zero. This is
17224 typically used as a second ACL in an expression in order to mark a connection
17225 when a first ACL was verified :
Willy Tarreauf73cd112011-08-13 01:45:16 +020017226
Jarno Huuskonen676f6222017-03-30 09:19:45 +030017227 Example:
Willy Tarreauf73cd112011-08-13 01:45:16 +020017228 # block if 5 consecutive requests continue to come faster than 10 sess
17229 # per second, and reset the counter as soon as the traffic slows down.
Willy Tarreaube4a3ef2013-06-17 15:04:07 +020017230 acl abuse sc0_http_req_rate gt 10
17231 acl kill sc0_inc_gpc0 gt 5
17232 acl save sc0_clr_gpc0 ge 0
Willy Tarreauf73cd112011-08-13 01:45:16 +020017233 tcp-request connection accept if !abuse save
17234 tcp-request connection reject if abuse kill
17235
Frédéric Lécaille6778b272018-01-29 15:22:53 +010017236sc_clr_gpc1(<ctr>[,<table>]) : integer
17237sc0_clr_gpc1([<table>]) : integer
17238sc1_clr_gpc1([<table>]) : integer
17239sc2_clr_gpc1([<table>]) : integer
17240 Clears the second General Purpose Counter associated to the currently tracked
17241 counters, and returns its previous value. Before the first invocation, the
17242 stored value is zero, so first invocation will always return zero. This is
17243 typically used as a second ACL in an expression in order to mark a connection
17244 when a first ACL was verified.
17245
Cyril Bonté62ba8702014-04-22 23:52:25 +020017246sc_conn_cnt(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020017247sc0_conn_cnt([<table>]) : integer
17248sc1_conn_cnt([<table>]) : integer
17249sc2_conn_cnt([<table>]) : integer
Davor Ocelice9ed2812017-12-25 17:49:28 +010017250 Returns the cumulative number of incoming connections from currently tracked
Willy Tarreaue9656522010-08-17 15:40:09 +020017251 counters. See also src_conn_cnt.
17252
Cyril Bonté62ba8702014-04-22 23:52:25 +020017253sc_conn_cur(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020017254sc0_conn_cur([<table>]) : integer
17255sc1_conn_cur([<table>]) : integer
17256sc2_conn_cur([<table>]) : integer
Willy Tarreaue9656522010-08-17 15:40:09 +020017257 Returns the current amount of concurrent connections tracking the same
17258 tracked counters. This number is automatically incremented when tracking
17259 begins and decremented when tracking stops. See also src_conn_cur.
17260
Cyril Bonté62ba8702014-04-22 23:52:25 +020017261sc_conn_rate(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020017262sc0_conn_rate([<table>]) : integer
17263sc1_conn_rate([<table>]) : integer
17264sc2_conn_rate([<table>]) : integer
Willy Tarreaue9656522010-08-17 15:40:09 +020017265 Returns the average connection rate from the currently tracked counters,
17266 measured in amount of connections over the period configured in the table.
17267 See also src_conn_rate.
17268
Cyril Bonté62ba8702014-04-22 23:52:25 +020017269sc_get_gpc0(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020017270sc0_get_gpc0([<table>]) : integer
17271sc1_get_gpc0([<table>]) : integer
17272sc2_get_gpc0([<table>]) : integer
Willy Tarreaue9656522010-08-17 15:40:09 +020017273 Returns the value of the first General Purpose Counter associated to the
Willy Tarreau4d4149c2013-07-23 19:33:46 +020017274 currently tracked counters. See also src_get_gpc0 and sc/sc0/sc1/sc2_inc_gpc0.
Willy Tarreauba2ffd12013-05-29 15:54:14 +020017275
Frédéric Lécaille6778b272018-01-29 15:22:53 +010017276sc_get_gpc1(<ctr>[,<table>]) : integer
17277sc0_get_gpc1([<table>]) : integer
17278sc1_get_gpc1([<table>]) : integer
17279sc2_get_gpc1([<table>]) : integer
17280 Returns the value of the second General Purpose Counter associated to the
17281 currently tracked counters. See also src_get_gpc1 and sc/sc0/sc1/sc2_inc_gpc1.
17282
Thierry FOURNIER236657b2015-08-19 08:25:14 +020017283sc_get_gpt0(<ctr>[,<table>]) : integer
17284sc0_get_gpt0([<table>]) : integer
17285sc1_get_gpt0([<table>]) : integer
17286sc2_get_gpt0([<table>]) : integer
17287 Returns the value of the first General Purpose Tag associated to the
17288 currently tracked counters. See also src_get_gpt0.
17289
Cyril Bonté62ba8702014-04-22 23:52:25 +020017290sc_gpc0_rate(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020017291sc0_gpc0_rate([<table>]) : integer
17292sc1_gpc0_rate([<table>]) : integer
17293sc2_gpc0_rate([<table>]) : integer
Willy Tarreauba2ffd12013-05-29 15:54:14 +020017294 Returns the average increment rate of the first General Purpose Counter
17295 associated to the currently tracked counters. It reports the frequency
17296 which the gpc0 counter was incremented over the configured period. See also
Willy Tarreau4d4149c2013-07-23 19:33:46 +020017297 src_gpc0_rate, sc/sc0/sc1/sc2_get_gpc0, and sc/sc0/sc1/sc2_inc_gpc0. Note
17298 that the "gpc0_rate" counter must be stored in the stick-table for a value to
17299 be returned, as "gpc0" only holds the event count.
Willy Tarreaue9656522010-08-17 15:40:09 +020017300
Frédéric Lécaille6778b272018-01-29 15:22:53 +010017301sc_gpc1_rate(<ctr>[,<table>]) : integer
17302sc0_gpc1_rate([<table>]) : integer
17303sc1_gpc1_rate([<table>]) : integer
17304sc2_gpc1_rate([<table>]) : integer
17305 Returns the average increment rate of the second General Purpose Counter
17306 associated to the currently tracked counters. It reports the frequency
17307 which the gpc1 counter was incremented over the configured period. See also
17308 src_gpcA_rate, sc/sc0/sc1/sc2_get_gpc1, and sc/sc0/sc1/sc2_inc_gpc1. Note
17309 that the "gpc1_rate" counter must be stored in the stick-table for a value to
17310 be returned, as "gpc1" only holds the event count.
17311
Cyril Bonté62ba8702014-04-22 23:52:25 +020017312sc_http_err_cnt(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020017313sc0_http_err_cnt([<table>]) : integer
17314sc1_http_err_cnt([<table>]) : integer
17315sc2_http_err_cnt([<table>]) : integer
Davor Ocelice9ed2812017-12-25 17:49:28 +010017316 Returns the cumulative number of HTTP errors from the currently tracked
Willy Tarreaue9656522010-08-17 15:40:09 +020017317 counters. This includes the both request errors and 4xx error responses.
17318 See also src_http_err_cnt.
17319
Cyril Bonté62ba8702014-04-22 23:52:25 +020017320sc_http_err_rate(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020017321sc0_http_err_rate([<table>]) : integer
17322sc1_http_err_rate([<table>]) : integer
17323sc2_http_err_rate([<table>]) : integer
Willy Tarreaue9656522010-08-17 15:40:09 +020017324 Returns the average rate of HTTP errors from the currently tracked counters,
17325 measured in amount of errors over the period configured in the table. This
17326 includes the both request errors and 4xx error responses. See also
17327 src_http_err_rate.
17328
Willy Tarreau826f3ab2021-02-10 12:07:15 +010017329sc_http_fail_cnt(<ctr>[,<table>]) : integer
17330sc0_http_fail_cnt([<table>]) : integer
17331sc1_http_fail_cnt([<table>]) : integer
17332sc2_http_fail_cnt([<table>]) : integer
17333 Returns the cumulative number of HTTP response failures from the currently
17334 tracked counters. This includes the both response errors and 5xx status codes
17335 other than 501 and 505. See also src_http_fail_cnt.
17336
17337sc_http_fail_rate(<ctr>[,<table>]) : integer
17338sc0_http_fail_rate([<table>]) : integer
17339sc1_http_fail_rate([<table>]) : integer
17340sc2_http_fail_rate([<table>]) : integer
17341 Returns the average rate of HTTP response failures from the currently tracked
17342 counters, measured in amount of failures over the period configured in the
17343 table. This includes the both response errors and 5xx status codes other than
17344 501 and 505. See also src_http_fail_rate.
17345
Cyril Bonté62ba8702014-04-22 23:52:25 +020017346sc_http_req_cnt(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020017347sc0_http_req_cnt([<table>]) : integer
17348sc1_http_req_cnt([<table>]) : integer
17349sc2_http_req_cnt([<table>]) : integer
Davor Ocelice9ed2812017-12-25 17:49:28 +010017350 Returns the cumulative number of HTTP requests from the currently tracked
Willy Tarreaue9656522010-08-17 15:40:09 +020017351 counters. This includes every started request, valid or not. See also
17352 src_http_req_cnt.
17353
Cyril Bonté62ba8702014-04-22 23:52:25 +020017354sc_http_req_rate(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020017355sc0_http_req_rate([<table>]) : integer
17356sc1_http_req_rate([<table>]) : integer
17357sc2_http_req_rate([<table>]) : integer
Willy Tarreaue9656522010-08-17 15:40:09 +020017358 Returns the average rate of HTTP requests from the currently tracked
17359 counters, measured in amount of requests over the period configured in
17360 the table. This includes every started request, valid or not. See also
17361 src_http_req_rate.
17362
Cyril Bonté62ba8702014-04-22 23:52:25 +020017363sc_inc_gpc0(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020017364sc0_inc_gpc0([<table>]) : integer
17365sc1_inc_gpc0([<table>]) : integer
17366sc2_inc_gpc0([<table>]) : integer
Willy Tarreaue9656522010-08-17 15:40:09 +020017367 Increments the first General Purpose Counter associated to the currently
Willy Tarreau869948b2013-01-04 14:14:57 +010017368 tracked counters, and returns its new value. Before the first invocation,
17369 the stored value is zero, so first invocation will increase it to 1 and will
17370 return 1. This is typically used as a second ACL in an expression in order
17371 to mark a connection when a first ACL was verified :
Willy Tarreaue9656522010-08-17 15:40:09 +020017372
Jarno Huuskonen676f6222017-03-30 09:19:45 +030017373 Example:
Willy Tarreaube4a3ef2013-06-17 15:04:07 +020017374 acl abuse sc0_http_req_rate gt 10
17375 acl kill sc0_inc_gpc0 gt 0
Willy Tarreaue9656522010-08-17 15:40:09 +020017376 tcp-request connection reject if abuse kill
17377
Frédéric Lécaille6778b272018-01-29 15:22:53 +010017378sc_inc_gpc1(<ctr>[,<table>]) : integer
17379sc0_inc_gpc1([<table>]) : integer
17380sc1_inc_gpc1([<table>]) : integer
17381sc2_inc_gpc1([<table>]) : integer
17382 Increments the second General Purpose Counter associated to the currently
17383 tracked counters, and returns its new value. Before the first invocation,
17384 the stored value is zero, so first invocation will increase it to 1 and will
17385 return 1. This is typically used as a second ACL in an expression in order
17386 to mark a connection when a first ACL was verified.
17387
Cyril Bonté62ba8702014-04-22 23:52:25 +020017388sc_kbytes_in(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020017389sc0_kbytes_in([<table>]) : integer
17390sc1_kbytes_in([<table>]) : integer
17391sc2_kbytes_in([<table>]) : integer
Willy Tarreaua01b9742014-07-10 15:29:24 +020017392 Returns the total amount of client-to-server data from the currently tracked
17393 counters, measured in kilobytes. The test is currently performed on 32-bit
17394 integers, which limits values to 4 terabytes. See also src_kbytes_in.
Willy Tarreaue9656522010-08-17 15:40:09 +020017395
Cyril Bonté62ba8702014-04-22 23:52:25 +020017396sc_kbytes_out(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020017397sc0_kbytes_out([<table>]) : integer
17398sc1_kbytes_out([<table>]) : integer
17399sc2_kbytes_out([<table>]) : integer
Willy Tarreaua01b9742014-07-10 15:29:24 +020017400 Returns the total amount of server-to-client data from the currently tracked
17401 counters, measured in kilobytes. The test is currently performed on 32-bit
17402 integers, which limits values to 4 terabytes. See also src_kbytes_out.
Willy Tarreaue9656522010-08-17 15:40:09 +020017403
Cyril Bonté62ba8702014-04-22 23:52:25 +020017404sc_sess_cnt(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020017405sc0_sess_cnt([<table>]) : integer
17406sc1_sess_cnt([<table>]) : integer
17407sc2_sess_cnt([<table>]) : integer
Davor Ocelice9ed2812017-12-25 17:49:28 +010017408 Returns the cumulative number of incoming connections that were transformed
Willy Tarreaue9656522010-08-17 15:40:09 +020017409 into sessions, which means that they were accepted by a "tcp-request
17410 connection" rule, from the currently tracked counters. A backend may count
17411 more sessions than connections because each connection could result in many
Jamie Gloudonaaa21002012-08-25 00:18:33 -040017412 backend sessions if some HTTP keep-alive is performed over the connection
Willy Tarreaue9656522010-08-17 15:40:09 +020017413 with the client. See also src_sess_cnt.
17414
Cyril Bonté62ba8702014-04-22 23:52:25 +020017415sc_sess_rate(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020017416sc0_sess_rate([<table>]) : integer
17417sc1_sess_rate([<table>]) : integer
17418sc2_sess_rate([<table>]) : integer
Willy Tarreaue9656522010-08-17 15:40:09 +020017419 Returns the average session rate from the currently tracked counters,
17420 measured in amount of sessions over the period configured in the table. A
17421 session is a connection that got past the early "tcp-request connection"
17422 rules. A backend may count more sessions than connections because each
17423 connection could result in many backend sessions if some HTTP keep-alive is
Jamie Gloudonaaa21002012-08-25 00:18:33 -040017424 performed over the connection with the client. See also src_sess_rate.
Willy Tarreaue9656522010-08-17 15:40:09 +020017425
Cyril Bonté62ba8702014-04-22 23:52:25 +020017426sc_tracked(<ctr>[,<table>]) : boolean
Willy Tarreau0f791d42013-07-23 19:56:43 +020017427sc0_tracked([<table>]) : boolean
17428sc1_tracked([<table>]) : boolean
17429sc2_tracked([<table>]) : boolean
Willy Tarreau6f1615f2013-06-03 15:15:22 +020017430 Returns true if the designated session counter is currently being tracked by
17431 the current session. This can be useful when deciding whether or not we want
17432 to set some values in a header passed to the server.
17433
Cyril Bonté62ba8702014-04-22 23:52:25 +020017434sc_trackers(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020017435sc0_trackers([<table>]) : integer
17436sc1_trackers([<table>]) : integer
17437sc2_trackers([<table>]) : integer
Willy Tarreau2406db42012-12-09 12:16:43 +010017438 Returns the current amount of concurrent connections tracking the same
17439 tracked counters. This number is automatically incremented when tracking
Willy Tarreaube4a3ef2013-06-17 15:04:07 +020017440 begins and decremented when tracking stops. It differs from sc0_conn_cur in
Willy Tarreau2406db42012-12-09 12:16:43 +010017441 that it does not rely on any stored information but on the table's reference
17442 count (the "use" value which is returned by "show table" on the CLI). This
Willy Tarreau74ca5042013-06-11 23:12:07 +020017443 may sometimes be more suited for layer7 tracking. It can be used to tell a
17444 server how many concurrent connections there are from a given address for
17445 example.
Willy Tarreau2406db42012-12-09 12:16:43 +010017446
Willy Tarreau74ca5042013-06-11 23:12:07 +020017447so_id : integer
17448 Returns an integer containing the current listening socket's id. It is useful
17449 in frontends involving many "bind" lines, or to stick all users coming via a
17450 same socket to the same server.
Willy Tarreaud63335a2010-02-26 12:56:52 +010017451
Jerome Magnineb421b22020-03-27 22:08:40 +010017452so_name : string
17453 Returns a string containing the current listening socket's name, as defined
17454 with name on a "bind" line. It can serve the same purposes as so_id but with
17455 strings instead of integers.
17456
Willy Tarreau74ca5042013-06-11 23:12:07 +020017457src : ip
Davor Ocelice9ed2812017-12-25 17:49:28 +010017458 This is the source IPv4 address of the client of the session. It is of type
Willy Tarreau74ca5042013-06-11 23:12:07 +020017459 IP and works on both IPv4 and IPv6 tables. On IPv6 tables, IPv4 addresses are
17460 mapped to their IPv6 equivalent, according to RFC 4291. Note that it is the
17461 TCP-level source address which is used, and not the address of a client
Bertrand Jacquin93b227d2016-06-04 15:11:10 +010017462 behind a proxy. However if the "accept-proxy" or "accept-netscaler-cip" bind
17463 directive is used, it can be the address of a client behind another
17464 PROXY-protocol compatible component for all rule sets except
Willy Tarreau64ded3d2019-01-23 10:02:15 +010017465 "tcp-request connection" which sees the real address. When the incoming
17466 connection passed through address translation or redirection involving
17467 connection tracking, the original destination address before the redirection
17468 will be reported. On Linux systems, the source and destination may seldom
17469 appear reversed if the nf_conntrack_tcp_loose sysctl is set, because a late
17470 response may reopen a timed out connection and switch what is believed to be
17471 the source and the destination.
Willy Tarreaud63335a2010-02-26 12:56:52 +010017472
Thierry FOURNIERd5f624d2013-11-26 11:52:33 +010017473 Example:
17474 # add an HTTP header in requests with the originating address' country
17475 http-request set-header X-Country %[src,map_ip(geoip.lst)]
17476
Willy Tarreau74ca5042013-06-11 23:12:07 +020017477src_bytes_in_rate([<table>]) : integer
17478 Returns the average bytes rate from the incoming connection's source address
17479 in the current proxy's stick-table or in the designated stick-table, measured
17480 in amount of bytes over the period configured in the table. If the address is
Willy Tarreau4d4149c2013-07-23 19:33:46 +020017481 not found, zero is returned. See also sc/sc0/sc1/sc2_bytes_in_rate.
Willy Tarreauc9705a12010-07-27 20:05:50 +020017482
Willy Tarreau74ca5042013-06-11 23:12:07 +020017483src_bytes_out_rate([<table>]) : integer
17484 Returns the average bytes rate to the incoming connection's source address in
17485 the current proxy's stick-table or in the designated stick-table, measured in
Willy Tarreauc9705a12010-07-27 20:05:50 +020017486 amount of bytes over the period configured in the table. If the address is
Willy Tarreau4d4149c2013-07-23 19:33:46 +020017487 not found, zero is returned. See also sc/sc0/sc1/sc2_bytes_out_rate.
Willy Tarreauc9705a12010-07-27 20:05:50 +020017488
Willy Tarreau74ca5042013-06-11 23:12:07 +020017489src_clr_gpc0([<table>]) : integer
17490 Clears the first General Purpose Counter associated to the incoming
17491 connection's source address in the current proxy's stick-table or in the
17492 designated stick-table, and returns its previous value. If the address is not
17493 found, an entry is created and 0 is returned. This is typically used as a
17494 second ACL in an expression in order to mark a connection when a first ACL
17495 was verified :
Willy Tarreauf73cd112011-08-13 01:45:16 +020017496
Jarno Huuskonen676f6222017-03-30 09:19:45 +030017497 Example:
Willy Tarreauf73cd112011-08-13 01:45:16 +020017498 # block if 5 consecutive requests continue to come faster than 10 sess
17499 # per second, and reset the counter as soon as the traffic slows down.
17500 acl abuse src_http_req_rate gt 10
17501 acl kill src_inc_gpc0 gt 5
Willy Tarreau869948b2013-01-04 14:14:57 +010017502 acl save src_clr_gpc0 ge 0
Willy Tarreauf73cd112011-08-13 01:45:16 +020017503 tcp-request connection accept if !abuse save
17504 tcp-request connection reject if abuse kill
17505
Frédéric Lécaille6778b272018-01-29 15:22:53 +010017506src_clr_gpc1([<table>]) : integer
17507 Clears the second General Purpose Counter associated to the incoming
17508 connection's source address in the current proxy's stick-table or in the
17509 designated stick-table, and returns its previous value. If the address is not
17510 found, an entry is created and 0 is returned. This is typically used as a
17511 second ACL in an expression in order to mark a connection when a first ACL
17512 was verified.
17513
Willy Tarreau74ca5042013-06-11 23:12:07 +020017514src_conn_cnt([<table>]) : integer
Davor Ocelice9ed2812017-12-25 17:49:28 +010017515 Returns the cumulative number of connections initiated from the current
Willy Tarreau74ca5042013-06-11 23:12:07 +020017516 incoming connection's source address in the current proxy's stick-table or in
Willy Tarreauc9705a12010-07-27 20:05:50 +020017517 the designated stick-table. If the address is not found, zero is returned.
Willy Tarreau4d4149c2013-07-23 19:33:46 +020017518 See also sc/sc0/sc1/sc2_conn_cnt.
Willy Tarreauc9705a12010-07-27 20:05:50 +020017519
Willy Tarreau74ca5042013-06-11 23:12:07 +020017520src_conn_cur([<table>]) : integer
Willy Tarreauc9705a12010-07-27 20:05:50 +020017521 Returns the current amount of concurrent connections initiated from the
Willy Tarreau74ca5042013-06-11 23:12:07 +020017522 current incoming connection's source address in the current proxy's
17523 stick-table or in the designated stick-table. If the address is not found,
Willy Tarreau4d4149c2013-07-23 19:33:46 +020017524 zero is returned. See also sc/sc0/sc1/sc2_conn_cur.
Willy Tarreauc9705a12010-07-27 20:05:50 +020017525
Willy Tarreau74ca5042013-06-11 23:12:07 +020017526src_conn_rate([<table>]) : integer
17527 Returns the average connection rate from the incoming connection's source
17528 address in the current proxy's stick-table or in the designated stick-table,
17529 measured in amount of connections over the period configured in the table. If
Willy Tarreau4d4149c2013-07-23 19:33:46 +020017530 the address is not found, zero is returned. See also sc/sc0/sc1/sc2_conn_rate.
Willy Tarreauc9705a12010-07-27 20:05:50 +020017531
Willy Tarreau74ca5042013-06-11 23:12:07 +020017532src_get_gpc0([<table>]) : integer
Willy Tarreauc9705a12010-07-27 20:05:50 +020017533 Returns the value of the first General Purpose Counter associated to the
Willy Tarreau74ca5042013-06-11 23:12:07 +020017534 incoming connection's source address in the current proxy's stick-table or in
Willy Tarreauc9705a12010-07-27 20:05:50 +020017535 the designated stick-table. If the address is not found, zero is returned.
Willy Tarreau4d4149c2013-07-23 19:33:46 +020017536 See also sc/sc0/sc1/sc2_get_gpc0 and src_inc_gpc0.
Willy Tarreauc9705a12010-07-27 20:05:50 +020017537
Frédéric Lécaille6778b272018-01-29 15:22:53 +010017538src_get_gpc1([<table>]) : integer
17539 Returns the value of the second General Purpose Counter associated to the
17540 incoming connection's source address in the current proxy's stick-table or in
17541 the designated stick-table. If the address is not found, zero is returned.
17542 See also sc/sc0/sc1/sc2_get_gpc1 and src_inc_gpc1.
17543
Thierry FOURNIER236657b2015-08-19 08:25:14 +020017544src_get_gpt0([<table>]) : integer
17545 Returns the value of the first General Purpose Tag associated to the
17546 incoming connection's source address in the current proxy's stick-table or in
17547 the designated stick-table. If the address is not found, zero is returned.
17548 See also sc/sc0/sc1/sc2_get_gpt0.
17549
Willy Tarreau74ca5042013-06-11 23:12:07 +020017550src_gpc0_rate([<table>]) : integer
Willy Tarreauba2ffd12013-05-29 15:54:14 +020017551 Returns the average increment rate of the first General Purpose Counter
Willy Tarreau74ca5042013-06-11 23:12:07 +020017552 associated to the incoming connection's source address in the current proxy's
Willy Tarreauba2ffd12013-05-29 15:54:14 +020017553 stick-table or in the designated stick-table. It reports the frequency
17554 which the gpc0 counter was incremented over the configured period. See also
Willy Tarreau4d4149c2013-07-23 19:33:46 +020017555 sc/sc0/sc1/sc2_gpc0_rate, src_get_gpc0, and sc/sc0/sc1/sc2_inc_gpc0. Note
17556 that the "gpc0_rate" counter must be stored in the stick-table for a value to
17557 be returned, as "gpc0" only holds the event count.
Willy Tarreauba2ffd12013-05-29 15:54:14 +020017558
Frédéric Lécaille6778b272018-01-29 15:22:53 +010017559src_gpc1_rate([<table>]) : integer
17560 Returns the average increment rate of the second General Purpose Counter
17561 associated to the incoming connection's source address in the current proxy's
17562 stick-table or in the designated stick-table. It reports the frequency
17563 which the gpc1 counter was incremented over the configured period. See also
17564 sc/sc0/sc1/sc2_gpc1_rate, src_get_gpc1, and sc/sc0/sc1/sc2_inc_gpc1. Note
17565 that the "gpc1_rate" counter must be stored in the stick-table for a value to
17566 be returned, as "gpc1" only holds the event count.
17567
Willy Tarreau74ca5042013-06-11 23:12:07 +020017568src_http_err_cnt([<table>]) : integer
Davor Ocelice9ed2812017-12-25 17:49:28 +010017569 Returns the cumulative number of HTTP errors from the incoming connection's
Willy Tarreau74ca5042013-06-11 23:12:07 +020017570 source address in the current proxy's stick-table or in the designated
Willy Tarreauc9705a12010-07-27 20:05:50 +020017571 stick-table. This includes the both request errors and 4xx error responses.
Willy Tarreau4d4149c2013-07-23 19:33:46 +020017572 See also sc/sc0/sc1/sc2_http_err_cnt. If the address is not found, zero is
Willy Tarreau74ca5042013-06-11 23:12:07 +020017573 returned.
Willy Tarreauc9705a12010-07-27 20:05:50 +020017574
Willy Tarreau74ca5042013-06-11 23:12:07 +020017575src_http_err_rate([<table>]) : integer
17576 Returns the average rate of HTTP errors from the incoming connection's source
17577 address in the current proxy's stick-table or in the designated stick-table,
17578 measured in amount of errors over the period configured in the table. This
17579 includes the both request errors and 4xx error responses. If the address is
Willy Tarreau4d4149c2013-07-23 19:33:46 +020017580 not found, zero is returned. See also sc/sc0/sc1/sc2_http_err_rate.
Willy Tarreauc9705a12010-07-27 20:05:50 +020017581
Willy Tarreau826f3ab2021-02-10 12:07:15 +010017582src_http_fail_cnt([<table>]) : integer
17583 Returns the cumulative number of HTTP response failures triggered by the
17584 incoming connection's source address in the current proxy's stick-table or in
Ilya Shipitsin0de36ad2021-02-20 00:23:36 +050017585 the designated stick-table. This includes the both response errors and 5xx
Willy Tarreau826f3ab2021-02-10 12:07:15 +010017586 status codes other than 501 and 505. See also sc/sc0/sc1/sc2_http_fail_cnt.
17587 If the address is not found, zero is returned.
17588
17589src_http_fail_rate([<table>]) : integer
17590 Returns the average rate of HTTP response failures triggered by the incoming
17591 connection's source address in the current proxy's stick-table or in the
17592 designated stick-table, measured in amount of failures over the period
17593 configured in the table. This includes the both response errors and 5xx
17594 status codes other than 501 and 505. If the address is not found, zero is
17595 returned. See also sc/sc0/sc1/sc2_http_fail_rate.
17596
Willy Tarreau74ca5042013-06-11 23:12:07 +020017597src_http_req_cnt([<table>]) : integer
Davor Ocelice9ed2812017-12-25 17:49:28 +010017598 Returns the cumulative number of HTTP requests from the incoming connection's
Willy Tarreau74ca5042013-06-11 23:12:07 +020017599 source address in the current proxy's stick-table or in the designated stick-
17600 table. This includes every started request, valid or not. If the address is
Willy Tarreau4d4149c2013-07-23 19:33:46 +020017601 not found, zero is returned. See also sc/sc0/sc1/sc2_http_req_cnt.
Willy Tarreauc9705a12010-07-27 20:05:50 +020017602
Willy Tarreau74ca5042013-06-11 23:12:07 +020017603src_http_req_rate([<table>]) : integer
17604 Returns the average rate of HTTP requests from the incoming connection's
17605 source address in the current proxy's stick-table or in the designated stick-
17606 table, measured in amount of requests over the period configured in the
Willy Tarreauc9705a12010-07-27 20:05:50 +020017607 table. This includes every started request, valid or not. If the address is
Willy Tarreau4d4149c2013-07-23 19:33:46 +020017608 not found, zero is returned. See also sc/sc0/sc1/sc2_http_req_rate.
Willy Tarreauc9705a12010-07-27 20:05:50 +020017609
Willy Tarreau74ca5042013-06-11 23:12:07 +020017610src_inc_gpc0([<table>]) : integer
17611 Increments the first General Purpose Counter associated to the incoming
17612 connection's source address in the current proxy's stick-table or in the
17613 designated stick-table, and returns its new value. If the address is not
Willy Tarreaube4a3ef2013-06-17 15:04:07 +020017614 found, an entry is created and 1 is returned. See also sc0/sc2/sc2_inc_gpc0.
Willy Tarreau74ca5042013-06-11 23:12:07 +020017615 This is typically used as a second ACL in an expression in order to mark a
17616 connection when a first ACL was verified :
Willy Tarreauc9705a12010-07-27 20:05:50 +020017617
Jarno Huuskonen676f6222017-03-30 09:19:45 +030017618 Example:
Willy Tarreauc9705a12010-07-27 20:05:50 +020017619 acl abuse src_http_req_rate gt 10
Willy Tarreau869948b2013-01-04 14:14:57 +010017620 acl kill src_inc_gpc0 gt 0
Willy Tarreaue9656522010-08-17 15:40:09 +020017621 tcp-request connection reject if abuse kill
Willy Tarreauc9705a12010-07-27 20:05:50 +020017622
Frédéric Lécaille6778b272018-01-29 15:22:53 +010017623src_inc_gpc1([<table>]) : integer
17624 Increments the second General Purpose Counter associated to the incoming
17625 connection's source address in the current proxy's stick-table or in the
17626 designated stick-table, and returns its new value. If the address is not
17627 found, an entry is created and 1 is returned. See also sc0/sc2/sc2_inc_gpc1.
17628 This is typically used as a second ACL in an expression in order to mark a
17629 connection when a first ACL was verified.
17630
Willy Tarreau16e01562016-08-09 16:46:18 +020017631src_is_local : boolean
17632 Returns true if the source address of the incoming connection is local to the
17633 system, or false if the address doesn't exist on the system, meaning that it
17634 comes from a remote machine. Note that UNIX addresses are considered local.
17635 It can be useful to apply certain access restrictions based on where the
Davor Ocelice9ed2812017-12-25 17:49:28 +010017636 client comes from (e.g. require auth or https for remote machines). Please
Willy Tarreau16e01562016-08-09 16:46:18 +020017637 note that the check involves a few system calls, so it's better to do it only
17638 once per connection.
17639
Willy Tarreau74ca5042013-06-11 23:12:07 +020017640src_kbytes_in([<table>]) : integer
Willy Tarreaua01b9742014-07-10 15:29:24 +020017641 Returns the total amount of data received from the incoming connection's
17642 source address in the current proxy's stick-table or in the designated
17643 stick-table, measured in kilobytes. If the address is not found, zero is
17644 returned. The test is currently performed on 32-bit integers, which limits
17645 values to 4 terabytes. See also sc/sc0/sc1/sc2_kbytes_in.
Willy Tarreauc9705a12010-07-27 20:05:50 +020017646
Willy Tarreau74ca5042013-06-11 23:12:07 +020017647src_kbytes_out([<table>]) : integer
Willy Tarreaua01b9742014-07-10 15:29:24 +020017648 Returns the total amount of data sent to the incoming connection's source
17649 address in the current proxy's stick-table or in the designated stick-table,
17650 measured in kilobytes. If the address is not found, zero is returned. The
17651 test is currently performed on 32-bit integers, which limits values to 4
17652 terabytes. See also sc/sc0/sc1/sc2_kbytes_out.
Willy Tarreaua975b8f2010-06-05 19:13:27 +020017653
Willy Tarreau74ca5042013-06-11 23:12:07 +020017654src_port : integer
17655 Returns an integer value corresponding to the TCP source port of the
17656 connection on the client side, which is the port the client connected from.
17657 Usage of this function is very limited as modern protocols do not care much
17658 about source ports nowadays.
Willy Tarreau079ff0a2009-03-05 21:34:28 +010017659
Willy Tarreau74ca5042013-06-11 23:12:07 +020017660src_sess_cnt([<table>]) : integer
Davor Ocelice9ed2812017-12-25 17:49:28 +010017661 Returns the cumulative number of connections initiated from the incoming
Willy Tarreauc9705a12010-07-27 20:05:50 +020017662 connection's source IPv4 address in the current proxy's stick-table or in the
17663 designated stick-table, that were transformed into sessions, which means that
17664 they were accepted by "tcp-request" rules. If the address is not found, zero
Willy Tarreau4d4149c2013-07-23 19:33:46 +020017665 is returned. See also sc/sc0/sc1/sc2_sess_cnt.
Willy Tarreauc9705a12010-07-27 20:05:50 +020017666
Willy Tarreau74ca5042013-06-11 23:12:07 +020017667src_sess_rate([<table>]) : integer
17668 Returns the average session rate from the incoming connection's source
17669 address in the current proxy's stick-table or in the designated stick-table,
17670 measured in amount of sessions over the period configured in the table. A
17671 session is a connection that went past the early "tcp-request" rules. If the
Willy Tarreau4d4149c2013-07-23 19:33:46 +020017672 address is not found, zero is returned. See also sc/sc0/sc1/sc2_sess_rate.
Willy Tarreauc9705a12010-07-27 20:05:50 +020017673
Willy Tarreau74ca5042013-06-11 23:12:07 +020017674src_updt_conn_cnt([<table>]) : integer
17675 Creates or updates the entry associated to the incoming connection's source
17676 address in the current proxy's stick-table or in the designated stick-table.
17677 This table must be configured to store the "conn_cnt" data type, otherwise
17678 the match will be ignored. The current count is incremented by one, and the
17679 expiration timer refreshed. The updated count is returned, so this match
17680 can't return zero. This was used to reject service abusers based on their
17681 source address. Note: it is recommended to use the more complete "track-sc*"
17682 actions in "tcp-request" rules instead.
Willy Tarreaua975b8f2010-06-05 19:13:27 +020017683
17684 Example :
17685 # This frontend limits incoming SSH connections to 3 per 10 second for
17686 # each source address, and rejects excess connections until a 10 second
17687 # silence is observed. At most 20 addresses are tracked.
17688 listen ssh
17689 bind :22
17690 mode tcp
17691 maxconn 100
Willy Tarreauc9705a12010-07-27 20:05:50 +020017692 stick-table type ip size 20 expire 10s store conn_cnt
Willy Tarreau74ca5042013-06-11 23:12:07 +020017693 tcp-request content reject if { src_updt_conn_cnt gt 3 }
Willy Tarreaua975b8f2010-06-05 19:13:27 +020017694 server local 127.0.0.1:22
17695
Willy Tarreau74ca5042013-06-11 23:12:07 +020017696srv_id : integer
17697 Returns an integer containing the server's id when processing the response.
17698 While it's almost only used with ACLs, it may be used for logging or
Christopher Fauletd1b44642020-04-30 09:51:15 +020017699 debugging. It can also be used in a tcp-check or an http-check ruleset.
Hervé COMMOWICKdaa824e2011-08-05 12:09:44 +020017700
vkill1dfd1652019-10-30 16:58:14 +080017701srv_name : string
17702 Returns a string containing the server's name when processing the response.
17703 While it's almost only used with ACLs, it may be used for logging or
Christopher Fauletd1b44642020-04-30 09:51:15 +020017704 debugging. It can also be used in a tcp-check or an http-check ruleset.
vkill1dfd1652019-10-30 16:58:14 +080017705
Thierry FOURNIER060762e2014-04-23 13:29:15 +0200177067.3.4. Fetching samples at Layer 5
Willy Tarreau74ca5042013-06-11 23:12:07 +020017707----------------------------------
Willy Tarreau0b1cd942010-05-16 22:18:27 +020017708
Willy Tarreau74ca5042013-06-11 23:12:07 +020017709The layer 5 usually describes just the session layer which in haproxy is
17710closest to the session once all the connection handshakes are finished, but
17711when no content is yet made available. The fetch methods described here are
17712usable as low as the "tcp-request content" rule sets unless they require some
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030017713future information. Those generally include the results of SSL negotiations.
Willy Tarreauc735a072011-03-29 00:57:02 +020017714
Ben Shillitof25e8e52016-12-02 14:25:37 +00001771551d.all(<prop>[,<prop>*]) : string
17716 Returns values for the properties requested as a string, where values are
17717 separated by the delimiter specified with "51degrees-property-separator".
17718 The device is identified using all the important HTTP headers from the
17719 request. The function can be passed up to five property names, and if a
17720 property name can't be found, the value "NoData" is returned.
17721
17722 Example :
17723 # Here the header "X-51D-DeviceTypeMobileTablet" is added to the request
17724 # containing the three properties requested using all relevant headers from
17725 # the request.
17726 frontend http-in
17727 bind *:8081
17728 default_backend servers
17729 http-request set-header X-51D-DeviceTypeMobileTablet \
17730 %[51d.all(DeviceType,IsMobile,IsTablet)]
17731
Emeric Brun645ae792014-04-30 14:21:06 +020017732ssl_bc : boolean
17733 Returns true when the back connection was made via an SSL/TLS transport
17734 layer and is locally deciphered. This means the outgoing connection was made
Christopher Fauletd92ea7f2020-04-30 10:03:55 +020017735 other a server with the "ssl" option. It can be used in a tcp-check or an
17736 http-check ruleset.
Emeric Brun645ae792014-04-30 14:21:06 +020017737
17738ssl_bc_alg_keysize : integer
17739 Returns the symmetric cipher key size supported in bits when the outgoing
Christopher Fauletd92ea7f2020-04-30 10:03:55 +020017740 connection was made over an SSL/TLS transport layer. It can be used in a
17741 tcp-check or an http-check ruleset.
Emeric Brun645ae792014-04-30 14:21:06 +020017742
Olivier Houchard6b77f492018-11-22 18:18:29 +010017743ssl_bc_alpn : string
17744 This extracts the Application Layer Protocol Negotiation field from an
17745 outgoing connection made via a TLS transport layer.
Michael Prokop4438c602019-05-24 10:25:45 +020017746 The result is a string containing the protocol name negotiated with the
Olivier Houchard6b77f492018-11-22 18:18:29 +010017747 server. The SSL library must have been built with support for TLS
17748 extensions enabled (check haproxy -vv). Note that the TLS ALPN extension is
17749 not advertised unless the "alpn" keyword on the "server" line specifies a
17750 protocol list. Also, nothing forces the server to pick a protocol from this
17751 list, any other one may be requested. The TLS ALPN extension is meant to
Christopher Fauletd92ea7f2020-04-30 10:03:55 +020017752 replace the TLS NPN extension. See also "ssl_bc_npn". It can be used in a
17753 tcp-check or an http-check ruleset.
Olivier Houchard6b77f492018-11-22 18:18:29 +010017754
Emeric Brun645ae792014-04-30 14:21:06 +020017755ssl_bc_cipher : string
17756 Returns the name of the used cipher when the outgoing connection was made
Christopher Fauletd92ea7f2020-04-30 10:03:55 +020017757 over an SSL/TLS transport layer. It can be used in a tcp-check or an
17758 http-check ruleset.
Emeric Brun645ae792014-04-30 14:21:06 +020017759
Patrick Hemmer65674662019-06-04 08:13:03 -040017760ssl_bc_client_random : binary
17761 Returns the client random of the back connection when the incoming connection
17762 was made over an SSL/TLS transport layer. It is useful to to decrypt traffic
17763 sent using ephemeral ciphers. This requires OpenSSL >= 1.1.0, or BoringSSL.
Christopher Fauletd92ea7f2020-04-30 10:03:55 +020017764 It can be used in a tcp-check or an http-check ruleset.
Patrick Hemmer65674662019-06-04 08:13:03 -040017765
Emeric Brun74f7ffa2018-02-19 16:14:12 +010017766ssl_bc_is_resumed : boolean
17767 Returns true when the back connection was made over an SSL/TLS transport
17768 layer and the newly created SSL session was resumed using a cached
Christopher Fauletd92ea7f2020-04-30 10:03:55 +020017769 session or a TLS ticket. It can be used in a tcp-check or an http-check
17770 ruleset.
Emeric Brun74f7ffa2018-02-19 16:14:12 +010017771
Olivier Houchard6b77f492018-11-22 18:18:29 +010017772ssl_bc_npn : string
17773 This extracts the Next Protocol Negotiation field from an outgoing connection
17774 made via a TLS transport layer. The result is a string containing the
Michael Prokop4438c602019-05-24 10:25:45 +020017775 protocol name negotiated with the server . The SSL library must have been
Olivier Houchard6b77f492018-11-22 18:18:29 +010017776 built with support for TLS extensions enabled (check haproxy -vv). Note that
17777 the TLS NPN extension is not advertised unless the "npn" keyword on the
17778 "server" line specifies a protocol list. Also, nothing forces the server to
17779 pick a protocol from this list, any other one may be used. Please note that
Christopher Fauletd92ea7f2020-04-30 10:03:55 +020017780 the TLS NPN extension was replaced with ALPN. It can be used in a tcp-check
17781 or an http-check ruleset.
Olivier Houchard6b77f492018-11-22 18:18:29 +010017782
Emeric Brun645ae792014-04-30 14:21:06 +020017783ssl_bc_protocol : string
17784 Returns the name of the used protocol when the outgoing connection was made
Christopher Fauletd92ea7f2020-04-30 10:03:55 +020017785 over an SSL/TLS transport layer. It can be used in a tcp-check or an
17786 http-check ruleset.
Emeric Brun645ae792014-04-30 14:21:06 +020017787
Emeric Brunb73a9b02014-04-30 18:49:19 +020017788ssl_bc_unique_id : binary
Emeric Brun645ae792014-04-30 14:21:06 +020017789 When the outgoing connection was made over an SSL/TLS transport layer,
Emeric Brunb73a9b02014-04-30 18:49:19 +020017790 returns the TLS unique ID as defined in RFC5929 section 3. The unique id
Christopher Fauletd92ea7f2020-04-30 10:03:55 +020017791 can be encoded to base64 using the converter: "ssl_bc_unique_id,base64". It
17792 can be used in a tcp-check or an http-check ruleset.
Emeric Brun645ae792014-04-30 14:21:06 +020017793
Patrick Hemmer65674662019-06-04 08:13:03 -040017794ssl_bc_server_random : binary
17795 Returns the server random of the back connection when the incoming connection
17796 was made over an SSL/TLS transport layer. It is useful to to decrypt traffic
17797 sent using ephemeral ciphers. This requires OpenSSL >= 1.1.0, or BoringSSL.
Christopher Fauletd92ea7f2020-04-30 10:03:55 +020017798 It can be used in a tcp-check or an http-check ruleset.
Patrick Hemmer65674662019-06-04 08:13:03 -040017799
Emeric Brun645ae792014-04-30 14:21:06 +020017800ssl_bc_session_id : binary
17801 Returns the SSL ID of the back connection when the outgoing connection was
17802 made over an SSL/TLS transport layer. It is useful to log if we want to know
Christopher Fauletd92ea7f2020-04-30 10:03:55 +020017803 if session was reused or not. It can be used in a tcp-check or an http-check
17804 ruleset.
Emeric Brun645ae792014-04-30 14:21:06 +020017805
Patrick Hemmere0275472018-04-28 19:15:51 -040017806ssl_bc_session_key : binary
17807 Returns the SSL session master key of the back connection when the outgoing
17808 connection was made over an SSL/TLS transport layer. It is useful to decrypt
17809 traffic sent using ephemeral ciphers. This requires OpenSSL >= 1.1.0, or
Christopher Fauletd92ea7f2020-04-30 10:03:55 +020017810 BoringSSL. It can be used in a tcp-check or an http-check ruleset.
Patrick Hemmere0275472018-04-28 19:15:51 -040017811
Emeric Brun645ae792014-04-30 14:21:06 +020017812ssl_bc_use_keysize : integer
17813 Returns the symmetric cipher key size used in bits when the outgoing
Christopher Fauletd92ea7f2020-04-30 10:03:55 +020017814 connection was made over an SSL/TLS transport layer. It can be used in a
17815 tcp-check or an http-check ruleset.
Emeric Brun645ae792014-04-30 14:21:06 +020017816
Willy Tarreau74ca5042013-06-11 23:12:07 +020017817ssl_c_ca_err : integer
17818 When the incoming connection was made over an SSL/TLS transport layer,
17819 returns the ID of the first error detected during verification of the client
17820 certificate at depth > 0, or 0 if no error was encountered during this
17821 verification process. Please refer to your SSL library's documentation to
17822 find the exhaustive list of error codes.
Willy Tarreauc735a072011-03-29 00:57:02 +020017823
Willy Tarreau74ca5042013-06-11 23:12:07 +020017824ssl_c_ca_err_depth : integer
17825 When the incoming connection was made over an SSL/TLS transport layer,
17826 returns the depth in the CA chain of the first error detected during the
17827 verification of the client certificate. If no error is encountered, 0 is
17828 returned.
Willy Tarreau0ba27502007-12-24 16:55:16 +010017829
Christopher Faulet70d10d12020-11-06 12:10:33 +010017830ssl_c_chain_der : binary
William Dauchya598b502020-08-06 18:11:38 +020017831 Returns the DER formatted chain certificate presented by the client when the
17832 incoming connection was made over an SSL/TLS transport layer. When used for
17833 an ACL, the value(s) to match against can be passed in hexadecimal form. One
Ilya Shipitsin2272d8a2020-12-21 01:22:40 +050017834 can parse the result with any lib accepting ASN.1 DER data. It currently
William Dauchya598b502020-08-06 18:11:38 +020017835 does not support resumed sessions.
17836
Christopher Faulet70d10d12020-11-06 12:10:33 +010017837ssl_c_der : binary
17838 Returns the DER formatted certificate presented by the client when the
17839 incoming connection was made over an SSL/TLS transport layer. When used for
17840 an ACL, the value(s) to match against can be passed in hexadecimal form.
17841
Willy Tarreau74ca5042013-06-11 23:12:07 +020017842ssl_c_err : integer
17843 When the incoming connection was made over an SSL/TLS transport layer,
17844 returns the ID of the first error detected during verification at depth 0, or
17845 0 if no error was encountered during this verification process. Please refer
17846 to your SSL library's documentation to find the exhaustive list of error
17847 codes.
Willy Tarreau62644772008-07-16 18:36:06 +020017848
Elliot Otchet71f82972020-01-15 08:12:14 -050017849ssl_c_i_dn([<entry>[,<occ>[,<format>]]]) : string
Willy Tarreau74ca5042013-06-11 23:12:07 +020017850 When the incoming connection was made over an SSL/TLS transport layer,
17851 returns the full distinguished name of the issuer of the certificate
17852 presented by the client when no <entry> is specified, or the value of the
17853 first given entry found from the beginning of the DN. If a positive/negative
17854 occurrence number is specified as the optional second argument, it returns
17855 the value of the nth given entry value from the beginning/end of the DN.
17856 For instance, "ssl_c_i_dn(OU,2)" the second organization unit, and
17857 "ssl_c_i_dn(CN)" retrieves the common name.
Elliot Otchet71f82972020-01-15 08:12:14 -050017858 The <format> parameter allows you to receive the DN suitable for
17859 consumption by different protocols. Currently supported is rfc2253 for
17860 LDAP v3.
17861 If you'd like to modify the format only you can specify an empty string
17862 and zero for the first two parameters. Example: ssl_c_i_dn(,0,rfc2253)
Willy Tarreau62644772008-07-16 18:36:06 +020017863
Willy Tarreau74ca5042013-06-11 23:12:07 +020017864ssl_c_key_alg : string
17865 Returns the name of the algorithm used to generate the key of the certificate
17866 presented by the client when the incoming connection was made over an SSL/TLS
17867 transport layer.
Willy Tarreau62644772008-07-16 18:36:06 +020017868
Willy Tarreau74ca5042013-06-11 23:12:07 +020017869ssl_c_notafter : string
17870 Returns the end date presented by the client as a formatted string
17871 YYMMDDhhmmss[Z] when the incoming connection was made over an SSL/TLS
17872 transport layer.
Emeric Brunbede3d02009-06-30 17:54:00 +020017873
Willy Tarreau74ca5042013-06-11 23:12:07 +020017874ssl_c_notbefore : string
17875 Returns the start date presented by the client as a formatted string
17876 YYMMDDhhmmss[Z] when the incoming connection was made over an SSL/TLS
17877 transport layer.
Willy Tarreaub6672b52011-12-12 17:23:41 +010017878
Elliot Otchet71f82972020-01-15 08:12:14 -050017879ssl_c_s_dn([<entry>[,<occ>[,<format>]]]) : string
Willy Tarreau74ca5042013-06-11 23:12:07 +020017880 When the incoming connection was made over an SSL/TLS transport layer,
17881 returns the full distinguished name of the subject of the certificate
17882 presented by the client when no <entry> is specified, or the value of the
17883 first given entry found from the beginning of the DN. If a positive/negative
17884 occurrence number is specified as the optional second argument, it returns
17885 the value of the nth given entry value from the beginning/end of the DN.
17886 For instance, "ssl_c_s_dn(OU,2)" the second organization unit, and
17887 "ssl_c_s_dn(CN)" retrieves the common name.
Elliot Otchet71f82972020-01-15 08:12:14 -050017888 The <format> parameter allows you to receive the DN suitable for
17889 consumption by different protocols. Currently supported is rfc2253 for
17890 LDAP v3.
17891 If you'd like to modify the format only you can specify an empty string
17892 and zero for the first two parameters. Example: ssl_c_s_dn(,0,rfc2253)
Willy Tarreaub6672b52011-12-12 17:23:41 +010017893
Willy Tarreau74ca5042013-06-11 23:12:07 +020017894ssl_c_serial : binary
17895 Returns the serial of the certificate presented by the client when the
17896 incoming connection was made over an SSL/TLS transport layer. When used for
17897 an ACL, the value(s) to match against can be passed in hexadecimal form.
Emeric Brun2525b6b2012-10-18 15:59:43 +020017898
Willy Tarreau74ca5042013-06-11 23:12:07 +020017899ssl_c_sha1 : binary
17900 Returns the SHA-1 fingerprint of the certificate presented by the client when
17901 the incoming connection was made over an SSL/TLS transport layer. This can be
17902 used to stick a client to a server, or to pass this information to a server.
Willy Tarreau2d0caa32014-07-02 19:01:22 +020017903 Note that the output is binary, so if you want to pass that signature to the
17904 server, you need to encode it in hex or base64, such as in the example below:
17905
Jarno Huuskonen676f6222017-03-30 09:19:45 +030017906 Example:
Willy Tarreau2d0caa32014-07-02 19:01:22 +020017907 http-request set-header X-SSL-Client-SHA1 %[ssl_c_sha1,hex]
Emeric Brun2525b6b2012-10-18 15:59:43 +020017908
Willy Tarreau74ca5042013-06-11 23:12:07 +020017909ssl_c_sig_alg : string
17910 Returns the name of the algorithm used to sign the certificate presented by
17911 the client when the incoming connection was made over an SSL/TLS transport
17912 layer.
Emeric Brun87855892012-10-17 17:39:35 +020017913
Willy Tarreau74ca5042013-06-11 23:12:07 +020017914ssl_c_used : boolean
17915 Returns true if current SSL session uses a client certificate even if current
17916 connection uses SSL session resumption. See also "ssl_fc_has_crt".
Emeric Brun7f56e742012-10-19 18:15:40 +020017917
Willy Tarreau74ca5042013-06-11 23:12:07 +020017918ssl_c_verify : integer
17919 Returns the verify result error ID when the incoming connection was made over
17920 an SSL/TLS transport layer, otherwise zero if no error is encountered. Please
17921 refer to your SSL library's documentation for an exhaustive list of error
17922 codes.
Emeric Brunce5ad802012-10-22 14:11:22 +020017923
Willy Tarreau74ca5042013-06-11 23:12:07 +020017924ssl_c_version : integer
17925 Returns the version of the certificate presented by the client when the
17926 incoming connection was made over an SSL/TLS transport layer.
Emeric Brunce5ad802012-10-22 14:11:22 +020017927
Emeric Brun43e79582014-10-29 19:03:26 +010017928ssl_f_der : binary
17929 Returns the DER formatted certificate presented by the frontend when the
17930 incoming connection was made over an SSL/TLS transport layer. When used for
17931 an ACL, the value(s) to match against can be passed in hexadecimal form.
17932
Elliot Otchet71f82972020-01-15 08:12:14 -050017933ssl_f_i_dn([<entry>[,<occ>[,<format>]]]) : string
Willy Tarreau74ca5042013-06-11 23:12:07 +020017934 When the incoming connection was made over an SSL/TLS transport layer,
17935 returns the full distinguished name of the issuer of the certificate
17936 presented by the frontend when no <entry> is specified, or the value of the
17937 first given entry found from the beginning of the DN. If a positive/negative
Emeric Brun87855892012-10-17 17:39:35 +020017938 occurrence number is specified as the optional second argument, it returns
Willy Tarreau74ca5042013-06-11 23:12:07 +020017939 the value of the nth given entry value from the beginning/end of the DN.
17940 For instance, "ssl_f_i_dn(OU,2)" the second organization unit, and
17941 "ssl_f_i_dn(CN)" retrieves the common name.
Elliot Otchet71f82972020-01-15 08:12:14 -050017942 The <format> parameter allows you to receive the DN suitable for
17943 consumption by different protocols. Currently supported is rfc2253 for
17944 LDAP v3.
17945 If you'd like to modify the format only you can specify an empty string
17946 and zero for the first two parameters. Example: ssl_f_i_dn(,0,rfc2253)
Emeric Brun87855892012-10-17 17:39:35 +020017947
Willy Tarreau74ca5042013-06-11 23:12:07 +020017948ssl_f_key_alg : string
17949 Returns the name of the algorithm used to generate the key of the certificate
17950 presented by the frontend when the incoming connection was made over an
17951 SSL/TLS transport layer.
Emeric Brun7f56e742012-10-19 18:15:40 +020017952
Willy Tarreau74ca5042013-06-11 23:12:07 +020017953ssl_f_notafter : string
17954 Returns the end date presented by the frontend as a formatted string
17955 YYMMDDhhmmss[Z] when the incoming connection was made over an SSL/TLS
17956 transport layer.
Emeric Brun2525b6b2012-10-18 15:59:43 +020017957
Willy Tarreau74ca5042013-06-11 23:12:07 +020017958ssl_f_notbefore : string
17959 Returns the start date presented by the frontend as a formatted string
17960 YYMMDDhhmmss[Z] when the incoming connection was made over an SSL/TLS
17961 transport layer.
Emeric Brun87855892012-10-17 17:39:35 +020017962
Elliot Otchet71f82972020-01-15 08:12:14 -050017963ssl_f_s_dn([<entry>[,<occ>[,<format>]]]) : string
Willy Tarreau74ca5042013-06-11 23:12:07 +020017964 When the incoming connection was made over an SSL/TLS transport layer,
17965 returns the full distinguished name of the subject of the certificate
17966 presented by the frontend when no <entry> is specified, or the value of the
17967 first given entry found from the beginning of the DN. If a positive/negative
17968 occurrence number is specified as the optional second argument, it returns
17969 the value of the nth given entry value from the beginning/end of the DN.
17970 For instance, "ssl_f_s_dn(OU,2)" the second organization unit, and
17971 "ssl_f_s_dn(CN)" retrieves the common name.
Elliot Otchet71f82972020-01-15 08:12:14 -050017972 The <format> parameter allows you to receive the DN suitable for
17973 consumption by different protocols. Currently supported is rfc2253 for
17974 LDAP v3.
17975 If you'd like to modify the format only you can specify an empty string
17976 and zero for the first two parameters. Example: ssl_f_s_dn(,0,rfc2253)
Emeric Brunce5ad802012-10-22 14:11:22 +020017977
Willy Tarreau74ca5042013-06-11 23:12:07 +020017978ssl_f_serial : binary
17979 Returns the serial of the certificate presented by the frontend when the
17980 incoming connection was made over an SSL/TLS transport layer. When used for
17981 an ACL, the value(s) to match against can be passed in hexadecimal form.
Emeric Brun87855892012-10-17 17:39:35 +020017982
Emeric Brun55f4fa82014-04-30 17:11:25 +020017983ssl_f_sha1 : binary
17984 Returns the SHA-1 fingerprint of the certificate presented by the frontend
17985 when the incoming connection was made over an SSL/TLS transport layer. This
17986 can be used to know which certificate was chosen using SNI.
17987
Willy Tarreau74ca5042013-06-11 23:12:07 +020017988ssl_f_sig_alg : string
17989 Returns the name of the algorithm used to sign the certificate presented by
17990 the frontend when the incoming connection was made over an SSL/TLS transport
17991 layer.
Emeric Brun7f56e742012-10-19 18:15:40 +020017992
Willy Tarreau74ca5042013-06-11 23:12:07 +020017993ssl_f_version : integer
17994 Returns the version of the certificate presented by the frontend when the
17995 incoming connection was made over an SSL/TLS transport layer.
17996
17997ssl_fc : boolean
Emeric Brun2525b6b2012-10-18 15:59:43 +020017998 Returns true when the front connection was made via an SSL/TLS transport
17999 layer and is locally deciphered. This means it has matched a socket declared
18000 with a "bind" line having the "ssl" option.
18001
Willy Tarreau74ca5042013-06-11 23:12:07 +020018002 Example :
18003 # This passes "X-Proto: https" to servers when client connects over SSL
18004 listen http-https
18005 bind :80
18006 bind :443 ssl crt /etc/haproxy.pem
18007 http-request add-header X-Proto https if { ssl_fc }
18008
18009ssl_fc_alg_keysize : integer
18010 Returns the symmetric cipher key size supported in bits when the incoming
18011 connection was made over an SSL/TLS transport layer.
18012
18013ssl_fc_alpn : string
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030018014 This extracts the Application Layer Protocol Negotiation field from an
Willy Tarreau74ca5042013-06-11 23:12:07 +020018015 incoming connection made via a TLS transport layer and locally deciphered by
18016 haproxy. The result is a string containing the protocol name advertised by
18017 the client. The SSL library must have been built with support for TLS
18018 extensions enabled (check haproxy -vv). Note that the TLS ALPN extension is
18019 not advertised unless the "alpn" keyword on the "bind" line specifies a
18020 protocol list. Also, nothing forces the client to pick a protocol from this
18021 list, any other one may be requested. The TLS ALPN extension is meant to
18022 replace the TLS NPN extension. See also "ssl_fc_npn".
18023
Willy Tarreau74ca5042013-06-11 23:12:07 +020018024ssl_fc_cipher : string
18025 Returns the name of the used cipher when the incoming connection was made
18026 over an SSL/TLS transport layer.
Willy Tarreauab861d32013-04-02 02:30:41 +020018027
Thierry FOURNIER5bf77322017-02-25 12:45:22 +010018028ssl_fc_cipherlist_bin : binary
18029 Returns the binary form of the client hello cipher list. The maximum returned
18030 value length is according with the value of
Emmanuel Hocdetaaee7502017-03-07 18:34:58 +010018031 "tune.ssl.capture-cipherlist-size".
Thierry FOURNIER5bf77322017-02-25 12:45:22 +010018032
18033ssl_fc_cipherlist_hex : string
18034 Returns the binary form of the client hello cipher list encoded as
18035 hexadecimal. The maximum returned value length is according with the value of
Emmanuel Hocdetaaee7502017-03-07 18:34:58 +010018036 "tune.ssl.capture-cipherlist-size".
Thierry FOURNIER5bf77322017-02-25 12:45:22 +010018037
18038ssl_fc_cipherlist_str : string
18039 Returns the decoded text form of the client hello cipher list. The maximum
18040 number of ciphers returned is according with the value of
18041 "tune.ssl.capture-cipherlist-size". Note that this sample-fetch is only
Davor Ocelice9ed2812017-12-25 17:49:28 +010018042 available with OpenSSL >= 1.0.2. If the function is not enabled, this
Emmanuel Hocdetddcde192017-09-01 17:32:08 +020018043 sample-fetch returns the hash like "ssl_fc_cipherlist_xxh".
Thierry FOURNIER5bf77322017-02-25 12:45:22 +010018044
18045ssl_fc_cipherlist_xxh : integer
18046 Returns a xxh64 of the cipher list. This hash can be return only is the value
18047 "tune.ssl.capture-cipherlist-size" is set greater than 0, however the hash
Emmanuel Hocdetaaee7502017-03-07 18:34:58 +010018048 take in account all the data of the cipher list.
Thierry FOURNIER5bf77322017-02-25 12:45:22 +010018049
Patrick Hemmer65674662019-06-04 08:13:03 -040018050ssl_fc_client_random : binary
18051 Returns the client random of the front connection when the incoming connection
18052 was made over an SSL/TLS transport layer. It is useful to to decrypt traffic
18053 sent using ephemeral ciphers. This requires OpenSSL >= 1.1.0, or BoringSSL.
18054
William Lallemand7d42ef52020-07-06 11:41:30 +020018055ssl_fc_client_early_traffic_secret : string
18056 Return the CLIENT_EARLY_TRAFFIC_SECRET as an hexadecimal string for the
18057 front connection when the incoming connection was made over a TLS 1.3
18058 transport layer.
18059 Require OpenSSL >= 1.1.1. This is one of the keys dumped by the OpenSSL
18060 keylog callback to generate the SSLKEYLOGFILE. The SSL Key logging must be
18061 activated with "tune.ssl.keylog on" in the global section. See also
18062 "tune.ssl.keylog"
18063
18064ssl_fc_client_handshake_traffic_secret : string
18065 Return the CLIENT_HANDSHAKE_TRAFFIC_SECRET as an hexadecimal string for the
18066 front connection when the incoming connection was made over a TLS 1.3
18067 transport layer.
18068 Require OpenSSL >= 1.1.1. This is one of the keys dumped by the OpenSSL
18069 keylog callback to generate the SSLKEYLOGFILE. The SSL Key logging must be
18070 activated with "tune.ssl.keylog on" in the global section. See also
18071 "tune.ssl.keylog"
18072
18073ssl_fc_client_traffic_secret_0 : string
18074 Return the CLIENT_TRAFFIC_SECRET_0 as an hexadecimal string for the
18075 front connection when the incoming connection was made over a TLS 1.3
18076 transport layer.
18077 Require OpenSSL >= 1.1.1. This is one of the keys dumped by the OpenSSL
18078 keylog callback to generate the SSLKEYLOGFILE. The SSL Key logging must be
18079 activated with "tune.ssl.keylog on" in the global section. See also
18080 "tune.ssl.keylog"
18081
18082ssl_fc_exporter_secret : string
18083 Return the EXPORTER_SECRET as an hexadecimal string for the
18084 front connection when the incoming connection was made over a TLS 1.3
18085 transport layer.
18086 Require OpenSSL >= 1.1.1. This is one of the keys dumped by the OpenSSL
18087 keylog callback to generate the SSLKEYLOGFILE. The SSL Key logging must be
18088 activated with "tune.ssl.keylog on" in the global section. See also
18089 "tune.ssl.keylog"
18090
18091ssl_fc_early_exporter_secret : string
18092 Return the EARLY_EXPORTER_SECRET as an hexadecimal string for the
18093 front connection when the incoming connection was made over an TLS 1.3
18094 transport layer.
18095 Require OpenSSL >= 1.1.1. This is one of the keys dumped by the OpenSSL
18096 keylog callback to generate the SSLKEYLOGFILE. The SSL Key logging must be
18097 activated with "tune.ssl.keylog on" in the global section. See also
18098 "tune.ssl.keylog"
18099
Willy Tarreau74ca5042013-06-11 23:12:07 +020018100ssl_fc_has_crt : boolean
Emeric Brun2525b6b2012-10-18 15:59:43 +020018101 Returns true if a client certificate is present in an incoming connection over
18102 SSL/TLS transport layer. Useful if 'verify' statement is set to 'optional'.
Emeric Brun9143d372012-12-20 15:44:16 +010018103 Note: on SSL session resumption with Session ID or TLS ticket, client
18104 certificate is not present in the current connection but may be retrieved
18105 from the cache or the ticket. So prefer "ssl_c_used" if you want to check if
18106 current SSL session uses a client certificate.
Emeric Brun2525b6b2012-10-18 15:59:43 +020018107
Olivier Houchardccaa7de2017-10-02 11:51:03 +020018108ssl_fc_has_early : boolean
18109 Returns true if early data were sent, and the handshake didn't happen yet. As
18110 it has security implications, it is useful to be able to refuse those, or
18111 wait until the handshake happened.
18112
Willy Tarreau74ca5042013-06-11 23:12:07 +020018113ssl_fc_has_sni : boolean
18114 This checks for the presence of a Server Name Indication TLS extension (SNI)
Willy Tarreauf7bc57c2012-10-03 00:19:48 +020018115 in an incoming connection was made over an SSL/TLS transport layer. Returns
18116 true when the incoming connection presents a TLS SNI field. This requires
John Roeslerfb2fce12019-07-10 15:45:51 -050018117 that the SSL library is built with support for TLS extensions enabled (check
Willy Tarreauf7bc57c2012-10-03 00:19:48 +020018118 haproxy -vv).
Willy Tarreau7875d092012-09-10 08:20:03 +020018119
Nenad Merdanovic1516fe32016-05-17 03:31:21 +020018120ssl_fc_is_resumed : boolean
Nenad Merdanovic26ea8222015-05-18 02:28:57 +020018121 Returns true if the SSL/TLS session has been resumed through the use of
Jérôme Magnin4a326cb2018-01-15 14:01:17 +010018122 SSL session cache or TLS tickets on an incoming connection over an SSL/TLS
18123 transport layer.
Nenad Merdanovic26ea8222015-05-18 02:28:57 +020018124
Willy Tarreau74ca5042013-06-11 23:12:07 +020018125ssl_fc_npn : string
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030018126 This extracts the Next Protocol Negotiation field from an incoming connection
Willy Tarreau74ca5042013-06-11 23:12:07 +020018127 made via a TLS transport layer and locally deciphered by haproxy. The result
18128 is a string containing the protocol name advertised by the client. The SSL
18129 library must have been built with support for TLS extensions enabled (check
18130 haproxy -vv). Note that the TLS NPN extension is not advertised unless the
18131 "npn" keyword on the "bind" line specifies a protocol list. Also, nothing
18132 forces the client to pick a protocol from this list, any other one may be
18133 requested. Please note that the TLS NPN extension was replaced with ALPN.
Willy Tarreaua33c6542012-10-15 13:19:06 +020018134
Willy Tarreau74ca5042013-06-11 23:12:07 +020018135ssl_fc_protocol : string
18136 Returns the name of the used protocol when the incoming connection was made
18137 over an SSL/TLS transport layer.
Willy Tarreau7875d092012-09-10 08:20:03 +020018138
Emeric Brunb73a9b02014-04-30 18:49:19 +020018139ssl_fc_unique_id : binary
David Sc1ad52e2014-04-08 18:48:47 -040018140 When the incoming connection was made over an SSL/TLS transport layer,
Emeric Brunb73a9b02014-04-30 18:49:19 +020018141 returns the TLS unique ID as defined in RFC5929 section 3. The unique id
18142 can be encoded to base64 using the converter: "ssl_bc_unique_id,base64".
David Sc1ad52e2014-04-08 18:48:47 -040018143
William Lallemand7d42ef52020-07-06 11:41:30 +020018144ssl_fc_server_handshake_traffic_secret : string
18145 Return the SERVER_HANDSHAKE_TRAFFIC_SECRET as an hexadecimal string for the
18146 front connection when the incoming connection was made over a TLS 1.3
18147 transport layer.
18148 Require OpenSSL >= 1.1.1. This is one of the keys dumped by the OpenSSL
18149 keylog callback to generate the SSLKEYLOGFILE. The SSL Key logging must be
18150 activated with "tune.ssl.keylog on" in the global section. See also
18151 "tune.ssl.keylog"
18152
18153ssl_fc_server_traffic_secret_0 : string
18154 Return the SERVER_TRAFFIC_SECRET_0 as an hexadecimal string for the
18155 front connection when the incoming connection was made over an TLS 1.3
18156 transport layer.
18157 Require OpenSSL >= 1.1.1. This is one of the keys dumped by the OpenSSL
18158 keylog callback to generate the SSLKEYLOGFILE. The SSL Key logging must be
18159 activated with "tune.ssl.keylog on" in the global section. See also
18160 "tune.ssl.keylog"
18161
Patrick Hemmer65674662019-06-04 08:13:03 -040018162ssl_fc_server_random : binary
18163 Returns the server random of the front connection when the incoming connection
18164 was made over an SSL/TLS transport layer. It is useful to to decrypt traffic
18165 sent using ephemeral ciphers. This requires OpenSSL >= 1.1.0, or BoringSSL.
18166
Willy Tarreau74ca5042013-06-11 23:12:07 +020018167ssl_fc_session_id : binary
18168 Returns the SSL ID of the front connection when the incoming connection was
18169 made over an SSL/TLS transport layer. It is useful to stick a given client to
18170 a server. It is important to note that some browsers refresh their session ID
18171 every few minutes.
Willy Tarreau7875d092012-09-10 08:20:03 +020018172
Patrick Hemmere0275472018-04-28 19:15:51 -040018173ssl_fc_session_key : binary
18174 Returns the SSL session master key of the front connection when the incoming
18175 connection was made over an SSL/TLS transport layer. It is useful to decrypt
18176 traffic sent using ephemeral ciphers. This requires OpenSSL >= 1.1.0, or
18177 BoringSSL.
18178
18179
Willy Tarreau74ca5042013-06-11 23:12:07 +020018180ssl_fc_sni : string
18181 This extracts the Server Name Indication TLS extension (SNI) field from an
18182 incoming connection made via an SSL/TLS transport layer and locally
18183 deciphered by haproxy. The result (when present) typically is a string
18184 matching the HTTPS host name (253 chars or less). The SSL library must have
18185 been built with support for TLS extensions enabled (check haproxy -vv).
18186
18187 This fetch is different from "req_ssl_sni" above in that it applies to the
18188 connection being deciphered by haproxy and not to SSL contents being blindly
18189 forwarded. See also "ssl_fc_sni_end" and "ssl_fc_sni_reg" below. This
John Roeslerfb2fce12019-07-10 15:45:51 -050018190 requires that the SSL library is built with support for TLS extensions
Cyril Bonté9c1eb1e2012-10-09 22:45:34 +020018191 enabled (check haproxy -vv).
Willy Tarreau62644772008-07-16 18:36:06 +020018192
Willy Tarreau74ca5042013-06-11 23:12:07 +020018193 ACL derivatives :
Willy Tarreau74ca5042013-06-11 23:12:07 +020018194 ssl_fc_sni_end : suffix match
18195 ssl_fc_sni_reg : regex match
Emeric Brun589fcad2012-10-16 14:13:26 +020018196
Willy Tarreau74ca5042013-06-11 23:12:07 +020018197ssl_fc_use_keysize : integer
18198 Returns the symmetric cipher key size used in bits when the incoming
18199 connection was made over an SSL/TLS transport layer.
Willy Tarreaub6fb4202008-07-20 11:18:28 +020018200
William Lallemandbfa3e812020-06-25 20:07:18 +020018201ssl_s_der : binary
18202 Returns the DER formatted certificate presented by the server when the
18203 outgoing connection was made over an SSL/TLS transport layer. When used for
18204 an ACL, the value(s) to match against can be passed in hexadecimal form.
18205
William Dauchya598b502020-08-06 18:11:38 +020018206ssl_s_chain_der : binary
18207 Returns the DER formatted chain certificate presented by the server when the
18208 outgoing connection was made over an SSL/TLS transport layer. When used for
18209 an ACL, the value(s) to match against can be passed in hexadecimal form. One
Ilya Shipitsin2272d8a2020-12-21 01:22:40 +050018210 can parse the result with any lib accepting ASN.1 DER data. It currently
William Dauchya598b502020-08-06 18:11:38 +020018211 does not support resumed sessions.
18212
William Lallemandbfa3e812020-06-25 20:07:18 +020018213ssl_s_key_alg : string
18214 Returns the name of the algorithm used to generate the key of the certificate
18215 presented by the server when the outgoing connection was made over an
18216 SSL/TLS transport layer.
18217
18218ssl_s_notafter : string
18219 Returns the end date presented by the server as a formatted string
18220 YYMMDDhhmmss[Z] when the outgoing connection was made over an SSL/TLS
18221 transport layer.
18222
18223ssl_s_notbefore : string
18224 Returns the start date presented by the server as a formatted string
18225 YYMMDDhhmmss[Z] when the outgoing connection was made over an SSL/TLS
18226 transport layer.
18227
18228ssl_s_i_dn([<entry>[,<occ>[,<format>]]]) : string
18229 When the outgoing connection was made over an SSL/TLS transport layer,
18230 returns the full distinguished name of the issuer of the certificate
18231 presented by the server when no <entry> is specified, or the value of the
18232 first given entry found from the beginning of the DN. If a positive/negative
18233 occurrence number is specified as the optional second argument, it returns
18234 the value of the nth given entry value from the beginning/end of the DN.
William Lallemand8f600c82020-06-26 09:55:06 +020018235 For instance, "ssl_s_i_dn(OU,2)" the second organization unit, and
18236 "ssl_s_i_dn(CN)" retrieves the common name.
William Lallemandbfa3e812020-06-25 20:07:18 +020018237 The <format> parameter allows you to receive the DN suitable for
18238 consumption by different protocols. Currently supported is rfc2253 for
18239 LDAP v3.
18240 If you'd like to modify the format only you can specify an empty string
18241 and zero for the first two parameters. Example: ssl_s_i_dn(,0,rfc2253)
18242
18243ssl_s_s_dn([<entry>[,<occ>[,<format>]]]) : string
18244 When the outgoing connection was made over an SSL/TLS transport layer,
18245 returns the full distinguished name of the subject of the certificate
18246 presented by the server when no <entry> is specified, or the value of the
18247 first given entry found from the beginning of the DN. If a positive/negative
18248 occurrence number is specified as the optional second argument, it returns
18249 the value of the nth given entry value from the beginning/end of the DN.
William Lallemand8f600c82020-06-26 09:55:06 +020018250 For instance, "ssl_s_s_dn(OU,2)" the second organization unit, and
18251 "ssl_s_s_dn(CN)" retrieves the common name.
William Lallemandbfa3e812020-06-25 20:07:18 +020018252 The <format> parameter allows you to receive the DN suitable for
18253 consumption by different protocols. Currently supported is rfc2253 for
18254 LDAP v3.
18255 If you'd like to modify the format only you can specify an empty string
18256 and zero for the first two parameters. Example: ssl_s_s_dn(,0,rfc2253)
18257
18258ssl_s_serial : binary
18259 Returns the serial of the certificate presented by the server when the
18260 outgoing connection was made over an SSL/TLS transport layer. When used for
18261 an ACL, the value(s) to match against can be passed in hexadecimal form.
18262
18263ssl_s_sha1 : binary
18264 Returns the SHA-1 fingerprint of the certificate presented by the server
18265 when the outgoing connection was made over an SSL/TLS transport layer. This
18266 can be used to know which certificate was chosen using SNI.
18267
18268ssl_s_sig_alg : string
18269 Returns the name of the algorithm used to sign the certificate presented by
18270 the server when the outgoing connection was made over an SSL/TLS transport
18271 layer.
18272
18273ssl_s_version : integer
18274 Returns the version of the certificate presented by the server when the
18275 outgoing connection was made over an SSL/TLS transport layer.
Willy Tarreaub6fb4202008-07-20 11:18:28 +020018276
Thierry FOURNIER060762e2014-04-23 13:29:15 +0200182777.3.5. Fetching samples from buffer contents (Layer 6)
Willy Tarreau74ca5042013-06-11 23:12:07 +020018278------------------------------------------------------
Willy Tarreaub6fb4202008-07-20 11:18:28 +020018279
Willy Tarreau74ca5042013-06-11 23:12:07 +020018280Fetching samples from buffer contents is a bit different from the previous
18281sample fetches above because the sampled data are ephemeral. These data can
18282only be used when they're available and will be lost when they're forwarded.
18283For this reason, samples fetched from buffer contents during a request cannot
18284be used in a response for example. Even while the data are being fetched, they
18285can change. Sometimes it is necessary to set some delays or combine multiple
18286sample fetch methods to ensure that the expected data are complete and usable,
18287for example through TCP request content inspection. Please see the "tcp-request
18288content" keyword for more detailed information on the subject.
Willy Tarreau62644772008-07-16 18:36:06 +020018289
Willy Tarreau74ca5042013-06-11 23:12:07 +020018290payload(<offset>,<length>) : binary (deprecated)
Davor Ocelice9ed2812017-12-25 17:49:28 +010018291 This is an alias for "req.payload" when used in the context of a request (e.g.
Willy Tarreau74ca5042013-06-11 23:12:07 +020018292 "stick on", "stick match"), and for "res.payload" when used in the context of
18293 a response such as in "stick store response".
Willy Tarreau0ba27502007-12-24 16:55:16 +010018294
Willy Tarreau74ca5042013-06-11 23:12:07 +020018295payload_lv(<offset1>,<length>[,<offset2>]) : binary (deprecated)
18296 This is an alias for "req.payload_lv" when used in the context of a request
Davor Ocelice9ed2812017-12-25 17:49:28 +010018297 (e.g. "stick on", "stick match"), and for "res.payload_lv" when used in the
Willy Tarreau74ca5042013-06-11 23:12:07 +020018298 context of a response such as in "stick store response".
Willy Tarreau0ba27502007-12-24 16:55:16 +010018299
Willy Tarreau74ca5042013-06-11 23:12:07 +020018300req.len : integer
18301req_len : integer (deprecated)
18302 Returns an integer value corresponding to the number of bytes present in the
18303 request buffer. This is mostly used in ACL. It is important to understand
18304 that this test does not return false as long as the buffer is changing. This
18305 means that a check with equality to zero will almost always immediately match
18306 at the beginning of the session, while a test for more data will wait for
18307 that data to come in and return false only when haproxy is certain that no
18308 more data will come in. This test was designed to be used with TCP request
18309 content inspection.
Willy Tarreaua7ad50c2012-04-29 15:39:40 +020018310
Willy Tarreau74ca5042013-06-11 23:12:07 +020018311req.payload(<offset>,<length>) : binary
18312 This extracts a binary block of <length> bytes and starting at byte <offset>
Willy Tarreau00f00842013-08-02 11:07:32 +020018313 in the request buffer. As a special case, if the <length> argument is zero,
18314 the the whole buffer from <offset> to the end is extracted. This can be used
18315 with ACLs in order to check for the presence of some content in a buffer at
18316 any location.
Willy Tarreaua7ad50c2012-04-29 15:39:40 +020018317
Willy Tarreau74ca5042013-06-11 23:12:07 +020018318 ACL alternatives :
18319 payload(<offset>,<length>) : hex binary match
Willy Tarreaua7ad50c2012-04-29 15:39:40 +020018320
Willy Tarreau74ca5042013-06-11 23:12:07 +020018321req.payload_lv(<offset1>,<length>[,<offset2>]) : binary
18322 This extracts a binary block whose size is specified at <offset1> for <length>
18323 bytes, and which starts at <offset2> if specified or just after the length in
18324 the request buffer. The <offset2> parameter also supports relative offsets if
18325 prepended with a '+' or '-' sign.
Willy Tarreaua7ad50c2012-04-29 15:39:40 +020018326
Willy Tarreau74ca5042013-06-11 23:12:07 +020018327 ACL alternatives :
18328 payload_lv(<offset1>,<length>[,<offset2>]) : hex binary match
Willy Tarreaua7ad50c2012-04-29 15:39:40 +020018329
Willy Tarreau74ca5042013-06-11 23:12:07 +020018330 Example : please consult the example from the "stick store-response" keyword.
Willy Tarreaua7ad50c2012-04-29 15:39:40 +020018331
Willy Tarreau74ca5042013-06-11 23:12:07 +020018332req.proto_http : boolean
18333req_proto_http : boolean (deprecated)
18334 Returns true when data in the request buffer look like HTTP and correctly
18335 parses as such. It is the same parser as the common HTTP request parser which
18336 is used so there should be no surprises. The test does not match until the
18337 request is complete, failed or timed out. This test may be used to report the
18338 protocol in TCP logs, but the biggest use is to block TCP request analysis
18339 until a complete HTTP request is present in the buffer, for example to track
18340 a header.
Willy Tarreaua7ad50c2012-04-29 15:39:40 +020018341
Willy Tarreau74ca5042013-06-11 23:12:07 +020018342 Example:
18343 # track request counts per "base" (concatenation of Host+URL)
18344 tcp-request inspect-delay 10s
18345 tcp-request content reject if !HTTP
Willy Tarreaube4a3ef2013-06-17 15:04:07 +020018346 tcp-request content track-sc0 base table req-rate
Willy Tarreaua7ad50c2012-04-29 15:39:40 +020018347
Willy Tarreau74ca5042013-06-11 23:12:07 +020018348req.rdp_cookie([<name>]) : string
18349rdp_cookie([<name>]) : string (deprecated)
18350 When the request buffer looks like the RDP protocol, extracts the RDP cookie
18351 <name>, or any cookie if unspecified. The parser only checks for the first
18352 cookie, as illustrated in the RDP protocol specification. The cookie name is
18353 case insensitive. Generally the "MSTS" cookie name will be used, as it can
18354 contain the user name of the client connecting to the server if properly
18355 configured on the client. The "MSTSHASH" cookie is often used as well for
18356 session stickiness to servers.
Willy Tarreau04aa6a92012-04-06 18:57:55 +020018357
Willy Tarreau74ca5042013-06-11 23:12:07 +020018358 This differs from "balance rdp-cookie" in that any balancing algorithm may be
18359 used and thus the distribution of clients to backend servers is not linked to
18360 a hash of the RDP cookie. It is envisaged that using a balancing algorithm
18361 such as "balance roundrobin" or "balance leastconn" will lead to a more even
18362 distribution of clients to backend servers than the hash used by "balance
18363 rdp-cookie".
Willy Tarreau04aa6a92012-04-06 18:57:55 +020018364
Willy Tarreau74ca5042013-06-11 23:12:07 +020018365 ACL derivatives :
18366 req_rdp_cookie([<name>]) : exact string match
Willy Tarreau04aa6a92012-04-06 18:57:55 +020018367
Willy Tarreau74ca5042013-06-11 23:12:07 +020018368 Example :
18369 listen tse-farm
18370 bind 0.0.0.0:3389
18371 # wait up to 5s for an RDP cookie in the request
18372 tcp-request inspect-delay 5s
18373 tcp-request content accept if RDP_COOKIE
18374 # apply RDP cookie persistence
18375 persist rdp-cookie
18376 # Persist based on the mstshash cookie
18377 # This is only useful makes sense if
18378 # balance rdp-cookie is not used
18379 stick-table type string size 204800
18380 stick on req.rdp_cookie(mstshash)
18381 server srv1 1.1.1.1:3389
18382 server srv1 1.1.1.2:3389
Willy Tarreau04aa6a92012-04-06 18:57:55 +020018383
Willy Tarreau74ca5042013-06-11 23:12:07 +020018384 See also : "balance rdp-cookie", "persist rdp-cookie", "tcp-request" and the
18385 "req_rdp_cookie" ACL.
Willy Tarreau04aa6a92012-04-06 18:57:55 +020018386
Willy Tarreau74ca5042013-06-11 23:12:07 +020018387req.rdp_cookie_cnt([name]) : integer
18388rdp_cookie_cnt([name]) : integer (deprecated)
18389 Tries to parse the request buffer as RDP protocol, then returns an integer
18390 corresponding to the number of RDP cookies found. If an optional cookie name
18391 is passed, only cookies matching this name are considered. This is mostly
18392 used in ACL.
Willy Tarreau04aa6a92012-04-06 18:57:55 +020018393
Willy Tarreau74ca5042013-06-11 23:12:07 +020018394 ACL derivatives :
18395 req_rdp_cookie_cnt([<name>]) : integer match
Willy Tarreau04aa6a92012-04-06 18:57:55 +020018396
Alex Zorin4afdd132018-12-30 13:56:28 +110018397req.ssl_alpn : string
18398 Returns a string containing the values of the Application-Layer Protocol
18399 Negotiation (ALPN) TLS extension (RFC7301), sent by the client within the SSL
18400 ClientHello message. Note that this only applies to raw contents found in the
18401 request buffer and not to the contents deciphered via an SSL data layer, so
18402 this will not work with "bind" lines having the "ssl" option. This is useful
18403 in ACL to make a routing decision based upon the ALPN preferences of a TLS
Jarno Huuskonene504f812019-01-03 07:56:49 +020018404 client, like in the example below. See also "ssl_fc_alpn".
Alex Zorin4afdd132018-12-30 13:56:28 +110018405
18406 Examples :
18407 # Wait for a client hello for at most 5 seconds
18408 tcp-request inspect-delay 5s
18409 tcp-request content accept if { req_ssl_hello_type 1 }
Jarno Huuskonene504f812019-01-03 07:56:49 +020018410 use_backend bk_acme if { req.ssl_alpn acme-tls/1 }
Alex Zorin4afdd132018-12-30 13:56:28 +110018411 default_backend bk_default
18412
Nenad Merdanovic5fc7d7e2015-07-07 22:00:17 +020018413req.ssl_ec_ext : boolean
18414 Returns a boolean identifying if client sent the Supported Elliptic Curves
18415 Extension as defined in RFC4492, section 5.1. within the SSL ClientHello
Cyril Bonté307ee1e2015-09-28 23:16:06 +020018416 message. This can be used to present ECC compatible clients with EC
18417 certificate and to use RSA for all others, on the same IP address. Note that
18418 this only applies to raw contents found in the request buffer and not to
18419 contents deciphered via an SSL data layer, so this will not work with "bind"
18420 lines having the "ssl" option.
Nenad Merdanovic5fc7d7e2015-07-07 22:00:17 +020018421
Willy Tarreau74ca5042013-06-11 23:12:07 +020018422req.ssl_hello_type : integer
18423req_ssl_hello_type : integer (deprecated)
18424 Returns an integer value containing the type of the SSL hello message found
18425 in the request buffer if the buffer contains data that parse as a complete
18426 SSL (v3 or superior) client hello message. Note that this only applies to raw
18427 contents found in the request buffer and not to contents deciphered via an
18428 SSL data layer, so this will not work with "bind" lines having the "ssl"
18429 option. This is mostly used in ACL to detect presence of an SSL hello message
18430 that is supposed to contain an SSL session ID usable for stickiness.
Willy Tarreau04aa6a92012-04-06 18:57:55 +020018431
Willy Tarreau74ca5042013-06-11 23:12:07 +020018432req.ssl_sni : string
18433req_ssl_sni : string (deprecated)
18434 Returns a string containing the value of the Server Name TLS extension sent
18435 by a client in a TLS stream passing through the request buffer if the buffer
18436 contains data that parse as a complete SSL (v3 or superior) client hello
18437 message. Note that this only applies to raw contents found in the request
18438 buffer and not to contents deciphered via an SSL data layer, so this will not
Lukas Tribusa267b5d2020-07-19 00:25:06 +020018439 work with "bind" lines having the "ssl" option. This will only work for actual
18440 implicit TLS based protocols like HTTPS (443), IMAPS (993), SMTPS (465),
18441 however it will not work for explicit TLS based protocols, like SMTP (25/587)
18442 or IMAP (143). SNI normally contains the name of the host the client tries to
18443 connect to (for recent browsers). SNI is useful for allowing or denying access
18444 to certain hosts when SSL/TLS is used by the client. This test was designed to
18445 be used with TCP request content inspection. If content switching is needed,
18446 it is recommended to first wait for a complete client hello (type 1), like in
18447 the example below. See also "ssl_fc_sni".
Willy Tarreau04aa6a92012-04-06 18:57:55 +020018448
Willy Tarreau74ca5042013-06-11 23:12:07 +020018449 ACL derivatives :
18450 req_ssl_sni : exact string match
Willy Tarreau04aa6a92012-04-06 18:57:55 +020018451
Willy Tarreau74ca5042013-06-11 23:12:07 +020018452 Examples :
18453 # Wait for a client hello for at most 5 seconds
18454 tcp-request inspect-delay 5s
18455 tcp-request content accept if { req_ssl_hello_type 1 }
18456 use_backend bk_allow if { req_ssl_sni -f allowed_sites }
18457 default_backend bk_sorry_page
Willy Tarreau04aa6a92012-04-06 18:57:55 +020018458
Pradeep Jindalbb2acf52015-09-29 10:12:57 +053018459req.ssl_st_ext : integer
18460 Returns 0 if the client didn't send a SessionTicket TLS Extension (RFC5077)
18461 Returns 1 if the client sent SessionTicket TLS Extension
18462 Returns 2 if the client also sent non-zero length TLS SessionTicket
18463 Note that this only applies to raw contents found in the request buffer and
18464 not to contents deciphered via an SSL data layer, so this will not work with
18465 "bind" lines having the "ssl" option. This can for example be used to detect
18466 whether the client sent a SessionTicket or not and stick it accordingly, if
18467 no SessionTicket then stick on SessionID or don't stick as there's no server
18468 side state is there when SessionTickets are in use.
18469
Willy Tarreau74ca5042013-06-11 23:12:07 +020018470req.ssl_ver : integer
18471req_ssl_ver : integer (deprecated)
18472 Returns an integer value containing the version of the SSL/TLS protocol of a
18473 stream present in the request buffer. Both SSLv2 hello messages and SSLv3
18474 messages are supported. TLSv1 is announced as SSL version 3.1. The value is
18475 composed of the major version multiplied by 65536, added to the minor
18476 version. Note that this only applies to raw contents found in the request
18477 buffer and not to contents deciphered via an SSL data layer, so this will not
18478 work with "bind" lines having the "ssl" option. The ACL version of the test
Davor Ocelice9ed2812017-12-25 17:49:28 +010018479 matches against a decimal notation in the form MAJOR.MINOR (e.g. 3.1). This
Willy Tarreau74ca5042013-06-11 23:12:07 +020018480 fetch is mostly used in ACL.
Willy Tarreaud63335a2010-02-26 12:56:52 +010018481
Willy Tarreau74ca5042013-06-11 23:12:07 +020018482 ACL derivatives :
18483 req_ssl_ver : decimal match
Willy Tarreaud63335a2010-02-26 12:56:52 +010018484
Willy Tarreau47e8eba2013-09-11 23:28:46 +020018485res.len : integer
18486 Returns an integer value corresponding to the number of bytes present in the
18487 response buffer. This is mostly used in ACL. It is important to understand
18488 that this test does not return false as long as the buffer is changing. This
18489 means that a check with equality to zero will almost always immediately match
18490 at the beginning of the session, while a test for more data will wait for
18491 that data to come in and return false only when haproxy is certain that no
18492 more data will come in. This test was designed to be used with TCP response
Christopher Faulete596d182020-05-05 17:46:34 +020018493 content inspection. But it may also be used in tcp-check based expect rules.
Willy Tarreau47e8eba2013-09-11 23:28:46 +020018494
Willy Tarreau74ca5042013-06-11 23:12:07 +020018495res.payload(<offset>,<length>) : binary
18496 This extracts a binary block of <length> bytes and starting at byte <offset>
Willy Tarreau00f00842013-08-02 11:07:32 +020018497 in the response buffer. As a special case, if the <length> argument is zero,
Christopher Faulete596d182020-05-05 17:46:34 +020018498 the whole buffer from <offset> to the end is extracted. This can be used
Willy Tarreau00f00842013-08-02 11:07:32 +020018499 with ACLs in order to check for the presence of some content in a buffer at
Christopher Faulete596d182020-05-05 17:46:34 +020018500 any location. It may also be used in tcp-check based expect rules.
Willy Tarreaud63335a2010-02-26 12:56:52 +010018501
Willy Tarreau74ca5042013-06-11 23:12:07 +020018502res.payload_lv(<offset1>,<length>[,<offset2>]) : binary
18503 This extracts a binary block whose size is specified at <offset1> for <length>
18504 bytes, and which starts at <offset2> if specified or just after the length in
18505 the response buffer. The <offset2> parameter also supports relative offsets
Christopher Faulete596d182020-05-05 17:46:34 +020018506 if prepended with a '+' or '-' sign. It may also be used in tcp-check based
18507 expect rules.
Willy Tarreaud63335a2010-02-26 12:56:52 +010018508
Willy Tarreau74ca5042013-06-11 23:12:07 +020018509 Example : please consult the example from the "stick store-response" keyword.
Willy Tarreaud63335a2010-02-26 12:56:52 +010018510
Willy Tarreau971f7b62015-09-29 14:06:59 +020018511res.ssl_hello_type : integer
18512rep_ssl_hello_type : integer (deprecated)
18513 Returns an integer value containing the type of the SSL hello message found
18514 in the response buffer if the buffer contains data that parses as a complete
18515 SSL (v3 or superior) hello message. Note that this only applies to raw
18516 contents found in the response buffer and not to contents deciphered via an
18517 SSL data layer, so this will not work with "server" lines having the "ssl"
18518 option. This is mostly used in ACL to detect presence of an SSL hello message
18519 that is supposed to contain an SSL session ID usable for stickiness.
18520
Willy Tarreau74ca5042013-06-11 23:12:07 +020018521wait_end : boolean
18522 This fetch either returns true when the inspection period is over, or does
18523 not fetch. It is only used in ACLs, in conjunction with content analysis to
Davor Ocelice9ed2812017-12-25 17:49:28 +010018524 avoid returning a wrong verdict early. It may also be used to delay some
Willy Tarreau74ca5042013-06-11 23:12:07 +020018525 actions, such as a delayed reject for some special addresses. Since it either
18526 stops the rules evaluation or immediately returns true, it is recommended to
Davor Ocelice9ed2812017-12-25 17:49:28 +010018527 use this acl as the last one in a rule. Please note that the default ACL
Willy Tarreau74ca5042013-06-11 23:12:07 +020018528 "WAIT_END" is always usable without prior declaration. This test was designed
18529 to be used with TCP request content inspection.
Willy Tarreaud63335a2010-02-26 12:56:52 +010018530
Willy Tarreau74ca5042013-06-11 23:12:07 +020018531 Examples :
18532 # delay every incoming request by 2 seconds
18533 tcp-request inspect-delay 2s
18534 tcp-request content accept if WAIT_END
Willy Tarreaud63335a2010-02-26 12:56:52 +010018535
Willy Tarreau74ca5042013-06-11 23:12:07 +020018536 # don't immediately tell bad guys they are rejected
18537 tcp-request inspect-delay 10s
18538 acl goodguys src 10.0.0.0/24
18539 acl badguys src 10.0.1.0/24
18540 tcp-request content accept if goodguys
18541 tcp-request content reject if badguys WAIT_END
18542 tcp-request content reject
18543
18544
Thierry FOURNIER060762e2014-04-23 13:29:15 +0200185457.3.6. Fetching HTTP samples (Layer 7)
Willy Tarreau74ca5042013-06-11 23:12:07 +020018546--------------------------------------
18547
18548It is possible to fetch samples from HTTP contents, requests and responses.
18549This application layer is also called layer 7. It is only possible to fetch the
18550data in this section when a full HTTP request or response has been parsed from
18551its respective request or response buffer. This is always the case with all
18552HTTP specific rules and for sections running with "mode http". When using TCP
18553content inspection, it may be necessary to support an inspection delay in order
18554to let the request or response come in first. These fetches may require a bit
18555more CPU resources than the layer 4 ones, but not much since the request and
18556response are indexed.
18557
18558base : string
18559 This returns the concatenation of the first Host header and the path part of
18560 the request, which starts at the first slash and ends before the question
18561 mark. It can be useful in virtual hosted environments to detect URL abuses as
18562 well as to improve shared caches efficiency. Using this with a limited size
18563 stick table also allows one to collect statistics about most commonly
18564 requested objects by host/path. With ACLs it can allow simple content
18565 switching rules involving the host and the path at the same time, such as
18566 "www.example.com/favicon.ico". See also "path" and "uri".
18567
18568 ACL derivatives :
18569 base : exact string match
18570 base_beg : prefix match
18571 base_dir : subdir match
18572 base_dom : domain match
18573 base_end : suffix match
18574 base_len : length match
18575 base_reg : regex match
18576 base_sub : substring match
18577
18578base32 : integer
18579 This returns a 32-bit hash of the value returned by the "base" fetch method
18580 above. This is useful to track per-URL activity on high traffic sites without
18581 having to store all URLs. Instead a shorter hash is stored, saving a lot of
Willy Tarreau23ec4ca2014-07-15 20:15:37 +020018582 memory. The output type is an unsigned integer. The hash function used is
18583 SDBM with full avalanche on the output. Technically, base32 is exactly equal
18584 to "base,sdbm(1)".
Willy Tarreau74ca5042013-06-11 23:12:07 +020018585
18586base32+src : binary
18587 This returns the concatenation of the base32 fetch above and the src fetch
18588 below. The resulting type is of type binary, with a size of 8 or 20 bytes
18589 depending on the source address family. This can be used to track per-IP,
18590 per-URL counters.
18591
Yves Lafonb4d37082021-02-11 11:01:28 +010018592baseq : string
18593 This returns the concatenation of the first Host header and the path part of
18594 the request with the query-string, which starts at the first slash. Using this
18595 instead of "base" allows one to properly identify the target resource, for
18596 statistics or caching use cases. See also "path", "pathq" and "base".
18597
William Lallemand65ad6e12014-01-31 15:08:02 +010018598capture.req.hdr(<idx>) : string
18599 This extracts the content of the header captured by the "capture request
18600 header", idx is the position of the capture keyword in the configuration.
18601 The first entry is an index of 0. See also: "capture request header".
18602
18603capture.req.method : string
18604 This extracts the METHOD of an HTTP request. It can be used in both request
18605 and response. Unlike "method", it can be used in both request and response
18606 because it's allocated.
18607
18608capture.req.uri : string
18609 This extracts the request's URI, which starts at the first slash and ends
18610 before the first space in the request (without the host part). Unlike "path"
18611 and "url", it can be used in both request and response because it's
18612 allocated.
18613
Willy Tarreau3c1b5ec2014-04-24 23:41:57 +020018614capture.req.ver : string
18615 This extracts the request's HTTP version and returns either "HTTP/1.0" or
18616 "HTTP/1.1". Unlike "req.ver", it can be used in both request, response, and
18617 logs because it relies on a persistent flag.
18618
William Lallemand65ad6e12014-01-31 15:08:02 +010018619capture.res.hdr(<idx>) : string
18620 This extracts the content of the header captured by the "capture response
18621 header", idx is the position of the capture keyword in the configuration.
18622 The first entry is an index of 0.
18623 See also: "capture response header"
18624
Willy Tarreau3c1b5ec2014-04-24 23:41:57 +020018625capture.res.ver : string
18626 This extracts the response's HTTP version and returns either "HTTP/1.0" or
18627 "HTTP/1.1". Unlike "res.ver", it can be used in logs because it relies on a
18628 persistent flag.
18629
Willy Tarreaua5910cc2015-05-02 00:46:08 +020018630req.body : binary
Christopher Fauletaf4dc4c2020-05-05 17:33:25 +020018631 This returns the HTTP request's available body as a block of data. It is
18632 recommended to use "option http-buffer-request" to be sure to wait, as much
18633 as possible, for the request's body.
Willy Tarreaua5910cc2015-05-02 00:46:08 +020018634
Thierry FOURNIER9826c772015-05-20 15:50:54 +020018635req.body_param([<name>) : string
18636 This fetch assumes that the body of the POST request is url-encoded. The user
18637 can check if the "content-type" contains the value
18638 "application/x-www-form-urlencoded". This extracts the first occurrence of the
18639 parameter <name> in the body, which ends before '&'. The parameter name is
18640 case-sensitive. If no name is given, any parameter will match, and the first
18641 one will be returned. The result is a string corresponding to the value of the
18642 parameter <name> as presented in the request body (no URL decoding is
18643 performed). Note that the ACL version of this fetch iterates over multiple
18644 parameters and will iteratively report all parameters values if no name is
18645 given.
18646
Willy Tarreaua5910cc2015-05-02 00:46:08 +020018647req.body_len : integer
18648 This returns the length of the HTTP request's available body in bytes. It may
18649 be lower than the advertised length if the body is larger than the buffer. It
Christopher Fauletaf4dc4c2020-05-05 17:33:25 +020018650 is recommended to use "option http-buffer-request" to be sure to wait, as
18651 much as possible, for the request's body.
Willy Tarreaua5910cc2015-05-02 00:46:08 +020018652
18653req.body_size : integer
18654 This returns the advertised length of the HTTP request's body in bytes. It
Christopher Fauletaf4dc4c2020-05-05 17:33:25 +020018655 will represent the advertised Content-Length header, or the size of the
18656 available data in case of chunked encoding.
Willy Tarreaua5910cc2015-05-02 00:46:08 +020018657
Willy Tarreau74ca5042013-06-11 23:12:07 +020018658req.cook([<name>]) : string
18659cook([<name>]) : string (deprecated)
18660 This extracts the last occurrence of the cookie name <name> on a "Cookie"
18661 header line from the request, and returns its value as string. If no name is
18662 specified, the first cookie value is returned. When used with ACLs, all
18663 matching cookies are evaluated. Spaces around the name and the value are
18664 ignored as requested by the Cookie header specification (RFC6265). The cookie
18665 name is case-sensitive. Empty cookies are valid, so an empty cookie may very
18666 well return an empty value if it is present. Use the "found" match to detect
18667 presence. Use the res.cook() variant for response cookies sent by the server.
18668
18669 ACL derivatives :
18670 cook([<name>]) : exact string match
18671 cook_beg([<name>]) : prefix match
18672 cook_dir([<name>]) : subdir match
18673 cook_dom([<name>]) : domain match
18674 cook_end([<name>]) : suffix match
18675 cook_len([<name>]) : length match
18676 cook_reg([<name>]) : regex match
18677 cook_sub([<name>]) : substring match
Willy Tarreaud63335a2010-02-26 12:56:52 +010018678
Willy Tarreau74ca5042013-06-11 23:12:07 +020018679req.cook_cnt([<name>]) : integer
18680cook_cnt([<name>]) : integer (deprecated)
18681 Returns an integer value representing the number of occurrences of the cookie
18682 <name> in the request, or all cookies if <name> is not specified.
Willy Tarreaud63335a2010-02-26 12:56:52 +010018683
Willy Tarreau74ca5042013-06-11 23:12:07 +020018684req.cook_val([<name>]) : integer
18685cook_val([<name>]) : integer (deprecated)
18686 This extracts the last occurrence of the cookie name <name> on a "Cookie"
18687 header line from the request, and converts its value to an integer which is
18688 returned. If no name is specified, the first cookie value is returned. When
18689 used in ACLs, all matching names are iterated over until a value matches.
Willy Tarreau0e698542011-09-16 08:32:32 +020018690
Willy Tarreau74ca5042013-06-11 23:12:07 +020018691cookie([<name>]) : string (deprecated)
18692 This extracts the last occurrence of the cookie name <name> on a "Cookie"
18693 header line from the request, or a "Set-Cookie" header from the response, and
18694 returns its value as a string. A typical use is to get multiple clients
18695 sharing a same profile use the same server. This can be similar to what
Willy Tarreau294d0f02015-08-10 19:40:12 +020018696 "appsession" did with the "request-learn" statement, but with support for
Willy Tarreau74ca5042013-06-11 23:12:07 +020018697 multi-peer synchronization and state keeping across restarts. If no name is
18698 specified, the first cookie value is returned. This fetch should not be used
18699 anymore and should be replaced by req.cook() or res.cook() instead as it
18700 ambiguously uses the direction based on the context where it is used.
Willy Tarreaud63335a2010-02-26 12:56:52 +010018701
Willy Tarreau74ca5042013-06-11 23:12:07 +020018702hdr([<name>[,<occ>]]) : string
18703 This is equivalent to req.hdr() when used on requests, and to res.hdr() when
18704 used on responses. Please refer to these respective fetches for more details.
18705 In case of doubt about the fetch direction, please use the explicit ones.
18706 Note that contrary to the hdr() sample fetch method, the hdr_* ACL keywords
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030018707 unambiguously apply to the request headers.
Willy Tarreaud63335a2010-02-26 12:56:52 +010018708
Willy Tarreau74ca5042013-06-11 23:12:07 +020018709req.fhdr(<name>[,<occ>]) : string
Tim Duesterhus27c70ae2021-01-23 17:50:21 +010018710 This returns the full value of the last occurrence of header <name> in an
18711 HTTP request. It differs from req.hdr() in that any commas present in the
18712 value are returned and are not used as delimiters. This is sometimes useful
18713 with headers such as User-Agent.
18714
18715 When used from an ACL, all occurrences are iterated over until a match is
18716 found.
18717
Willy Tarreau74ca5042013-06-11 23:12:07 +020018718 Optionally, a specific occurrence might be specified as a position number.
18719 Positive values indicate a position from the first occurrence, with 1 being
18720 the first one. Negative values indicate positions relative to the last one,
Tim Duesterhus27c70ae2021-01-23 17:50:21 +010018721 with -1 being the last one.
Willy Tarreaud63335a2010-02-26 12:56:52 +010018722
Willy Tarreau74ca5042013-06-11 23:12:07 +020018723req.fhdr_cnt([<name>]) : integer
18724 Returns an integer value representing the number of occurrences of request
18725 header field name <name>, or the total number of header fields if <name> is
Tim Duesterhus27c70ae2021-01-23 17:50:21 +010018726 not specified. Like req.fhdr() it differs from res.hdr_cnt() by not splitting
18727 headers at commas.
Willy Tarreaud63335a2010-02-26 12:56:52 +010018728
Willy Tarreau74ca5042013-06-11 23:12:07 +020018729req.hdr([<name>[,<occ>]]) : string
Tim Duesterhus27c70ae2021-01-23 17:50:21 +010018730 This returns the last comma-separated value of the header <name> in an HTTP
18731 request. The fetch considers any comma as a delimiter for distinct values.
18732 This is useful if you need to process headers that are defined to be a list
18733 of values, such as Accept, or X-Forwarded-For. If full-line headers are
18734 desired instead, use req.fhdr(). Please carefully check RFC 7231 to know how
18735 certain headers are supposed to be parsed. Also, some of them are case
18736 insensitive (e.g. Connection).
18737
18738 When used from an ACL, all occurrences are iterated over until a match is
18739 found.
18740
Willy Tarreau74ca5042013-06-11 23:12:07 +020018741 Optionally, a specific occurrence might be specified as a position number.
18742 Positive values indicate a position from the first occurrence, with 1 being
18743 the first one. Negative values indicate positions relative to the last one,
Tim Duesterhus27c70ae2021-01-23 17:50:21 +010018744 with -1 being the last one.
18745
18746 A typical use is with the X-Forwarded-For header once converted to IP,
18747 associated with an IP stick-table.
Willy Tarreaud63335a2010-02-26 12:56:52 +010018748
Willy Tarreau74ca5042013-06-11 23:12:07 +020018749 ACL derivatives :
18750 hdr([<name>[,<occ>]]) : exact string match
18751 hdr_beg([<name>[,<occ>]]) : prefix match
18752 hdr_dir([<name>[,<occ>]]) : subdir match
18753 hdr_dom([<name>[,<occ>]]) : domain match
18754 hdr_end([<name>[,<occ>]]) : suffix match
18755 hdr_len([<name>[,<occ>]]) : length match
18756 hdr_reg([<name>[,<occ>]]) : regex match
18757 hdr_sub([<name>[,<occ>]]) : substring match
18758
18759req.hdr_cnt([<name>]) : integer
18760hdr_cnt([<header>]) : integer (deprecated)
18761 Returns an integer value representing the number of occurrences of request
18762 header field name <name>, or the total number of header field values if
Tim Duesterhus27c70ae2021-01-23 17:50:21 +010018763 <name> is not specified. Like req.hdr() it counts each comma separated
18764 part of the header's value. If counting of full-line headers is desired,
18765 then req.fhdr_cnt() should be used instead.
18766
18767 With ACLs, it can be used to detect presence, absence or abuse of a specific
18768 header, as well as to block request smuggling attacks by rejecting requests
18769 which contain more than one of certain headers.
18770
18771 Refer to req.hdr() for more information on header matching.
Willy Tarreau74ca5042013-06-11 23:12:07 +020018772
18773req.hdr_ip([<name>[,<occ>]]) : ip
18774hdr_ip([<name>[,<occ>]]) : ip (deprecated)
18775 This extracts the last occurrence of header <name> in an HTTP request,
18776 converts it to an IPv4 or IPv6 address and returns this address. When used
18777 with ACLs, all occurrences are checked, and if <name> is omitted, every value
Willy Tarreau7b0e00d2021-03-25 14:12:29 +010018778 of every header is checked. The parser strictly adheres to the format
18779 described in RFC7239, with the extension that IPv4 addresses may optionally
18780 be followed by a colon (':') and a valid decimal port number (0 to 65535),
18781 which will be silently dropped. All other forms will not match and will
18782 cause the address to be ignored.
Tim Duesterhus27c70ae2021-01-23 17:50:21 +010018783
18784 The <occ> parameter is processed as with req.hdr().
18785
18786 A typical use is with the X-Forwarded-For and X-Client-IP headers.
Willy Tarreau74ca5042013-06-11 23:12:07 +020018787
18788req.hdr_val([<name>[,<occ>]]) : integer
18789hdr_val([<name>[,<occ>]]) : integer (deprecated)
18790 This extracts the last occurrence of header <name> in an HTTP request, and
18791 converts it to an integer value. When used with ACLs, all occurrences are
18792 checked, and if <name> is omitted, every value of every header is checked.
Tim Duesterhus27c70ae2021-01-23 17:50:21 +010018793
18794 The <occ> parameter is processed as with req.hdr().
18795
18796 A typical use is with the X-Forwarded-For header.
Willy Tarreau74ca5042013-06-11 23:12:07 +020018797
Christopher Faulet687a68e2020-11-24 17:13:24 +010018798req.hdrs : string
18799 Returns the current request headers as string including the last empty line
18800 separating headers from the request body. The last empty line can be used to
18801 detect a truncated header block. This sample fetch is useful for some SPOE
18802 headers analyzers and for advanced logging.
18803
18804req.hdrs_bin : binary
18805 Returns the current request headers contained in preparsed binary form. This
18806 is useful for offloading some processing with SPOE. Each string is described
18807 by a length followed by the number of bytes indicated in the length. The
18808 length is represented using the variable integer encoding detailed in the
18809 SPOE documentation. The end of the list is marked by a couple of empty header
18810 names and values (length of 0 for both).
18811
18812 *(<str:header-name><str:header-value>)<empty string><empty string>
Frédéric Lécailleec891192019-02-26 15:02:35 +010018813
Christopher Faulet687a68e2020-11-24 17:13:24 +010018814 int: refer to the SPOE documentation for the encoding
18815 str: <int:length><bytes>
Frédéric Lécailleec891192019-02-26 15:02:35 +010018816
Willy Tarreau74ca5042013-06-11 23:12:07 +020018817http_auth(<userlist>) : boolean
18818 Returns a boolean indicating whether the authentication data received from
18819 the client match a username & password stored in the specified userlist. This
18820 fetch function is not really useful outside of ACLs. Currently only http
18821 basic auth is supported.
18822
Thierry FOURNIER9eec0a62014-01-22 18:38:02 +010018823http_auth_group(<userlist>) : string
18824 Returns a string corresponding to the user name found in the authentication
18825 data received from the client if both the user name and password are valid
18826 according to the specified userlist. The main purpose is to use it in ACLs
18827 where it is then checked whether the user belongs to any group within a list.
Willy Tarreau74ca5042013-06-11 23:12:07 +020018828 This fetch function is not really useful outside of ACLs. Currently only http
18829 basic auth is supported.
18830
18831 ACL derivatives :
Thierry FOURNIER9eec0a62014-01-22 18:38:02 +010018832 http_auth_group(<userlist>) : group ...
18833 Returns true when the user extracted from the request and whose password is
18834 valid according to the specified userlist belongs to at least one of the
18835 groups.
Willy Tarreau74ca5042013-06-11 23:12:07 +020018836
Christopher Fauleta4063562019-08-02 11:51:37 +020018837http_auth_pass : string
Willy Tarreauc9c6cdb2020-03-05 16:03:58 +010018838 Returns the user's password found in the authentication data received from
18839 the client, as supplied in the Authorization header. Not checks are
18840 performed by this sample fetch. Only Basic authentication is supported.
Christopher Fauleta4063562019-08-02 11:51:37 +020018841
18842http_auth_type : string
Willy Tarreauc9c6cdb2020-03-05 16:03:58 +010018843 Returns the authentication method found in the authentication data received from
18844 the client, as supplied in the Authorization header. Not checks are
18845 performed by this sample fetch. Only Basic authentication is supported.
Christopher Fauleta4063562019-08-02 11:51:37 +020018846
18847http_auth_user : string
Willy Tarreauc9c6cdb2020-03-05 16:03:58 +010018848 Returns the user name found in the authentication data received from the
18849 client, as supplied in the Authorization header. Not checks are performed by
18850 this sample fetch. Only Basic authentication is supported.
Christopher Fauleta4063562019-08-02 11:51:37 +020018851
Willy Tarreau74ca5042013-06-11 23:12:07 +020018852http_first_req : boolean
Willy Tarreau7f18e522010-10-22 20:04:13 +020018853 Returns true when the request being processed is the first one of the
18854 connection. This can be used to add or remove headers that may be missing
Willy Tarreau74ca5042013-06-11 23:12:07 +020018855 from some requests when a request is not the first one, or to help grouping
18856 requests in the logs.
Willy Tarreau7f18e522010-10-22 20:04:13 +020018857
Willy Tarreau74ca5042013-06-11 23:12:07 +020018858method : integer + string
18859 Returns an integer value corresponding to the method in the HTTP request. For
18860 example, "GET" equals 1 (check sources to establish the matching). Value 9
18861 means "other method" and may be converted to a string extracted from the
18862 stream. This should not be used directly as a sample, this is only meant to
18863 be used from ACLs, which transparently convert methods from patterns to these
18864 integer + string values. Some predefined ACL already check for most common
18865 methods.
Willy Tarreau6a06a402007-07-15 20:15:28 +020018866
Willy Tarreau74ca5042013-06-11 23:12:07 +020018867 ACL derivatives :
18868 method : case insensitive method match
Willy Tarreau6a06a402007-07-15 20:15:28 +020018869
Willy Tarreau74ca5042013-06-11 23:12:07 +020018870 Example :
18871 # only accept GET and HEAD requests
18872 acl valid_method method GET HEAD
18873 http-request deny if ! valid_method
Willy Tarreau6a06a402007-07-15 20:15:28 +020018874
Willy Tarreau74ca5042013-06-11 23:12:07 +020018875path : string
18876 This extracts the request's URL path, which starts at the first slash and
18877 ends before the question mark (without the host part). A typical use is with
18878 prefetch-capable caches, and with portals which need to aggregate multiple
18879 information from databases and keep them in caches. Note that with outgoing
18880 caches, it would be wiser to use "url" instead. With ACLs, it's typically
Davor Ocelice9ed2812017-12-25 17:49:28 +010018881 used to match exact file names (e.g. "/login.php"), or directory parts using
Willy Tarreau74ca5042013-06-11 23:12:07 +020018882 the derivative forms. See also the "url" and "base" fetch methods.
Willy Tarreau6a06a402007-07-15 20:15:28 +020018883
Willy Tarreau74ca5042013-06-11 23:12:07 +020018884 ACL derivatives :
18885 path : exact string match
18886 path_beg : prefix match
18887 path_dir : subdir match
18888 path_dom : domain match
18889 path_end : suffix match
18890 path_len : length match
18891 path_reg : regex match
18892 path_sub : substring match
Willy Tarreau6a06a402007-07-15 20:15:28 +020018893
Christopher Faulete720c322020-09-02 17:25:18 +020018894pathq : string
18895 This extracts the request's URL path with the query-string, which starts at
18896 the first slash. This sample fetch is pretty handy to always retrieve a
18897 relative URI, excluding the scheme and the authority part, if any. Indeed,
18898 while it is the common representation for an HTTP/1.1 request target, in
18899 HTTP/2, an absolute URI is often used. This sample fetch will return the same
18900 result in both cases.
18901
Willy Tarreau49ad95c2015-01-19 15:06:26 +010018902query : string
18903 This extracts the request's query string, which starts after the first
18904 question mark. If no question mark is present, this fetch returns nothing. If
18905 a question mark is present but nothing follows, it returns an empty string.
18906 This means it's possible to easily know whether a query string is present
Tim Düsterhus4896c442016-11-29 02:15:19 +010018907 using the "found" matching method. This fetch is the complement of "path"
Willy Tarreau49ad95c2015-01-19 15:06:26 +010018908 which stops before the question mark.
18909
Willy Tarreaueb27ec72015-02-20 13:55:29 +010018910req.hdr_names([<delim>]) : string
18911 This builds a string made from the concatenation of all header names as they
18912 appear in the request when the rule is evaluated. The default delimiter is
18913 the comma (',') but it may be overridden as an optional argument <delim>. In
18914 this case, only the first character of <delim> is considered.
18915
Willy Tarreau74ca5042013-06-11 23:12:07 +020018916req.ver : string
18917req_ver : string (deprecated)
18918 Returns the version string from the HTTP request, for example "1.1". This can
18919 be useful for logs, but is mostly there for ACL. Some predefined ACL already
18920 check for versions 1.0 and 1.1.
Willy Tarreaud63335a2010-02-26 12:56:52 +010018921
Willy Tarreau74ca5042013-06-11 23:12:07 +020018922 ACL derivatives :
18923 req_ver : exact string match
Willy Tarreau0e698542011-09-16 08:32:32 +020018924
Christopher Faulete596d182020-05-05 17:46:34 +020018925res.body : binary
18926 This returns the HTTP response's available body as a block of data. Unlike
18927 the request side, there is no directive to wait for the response's body. This
Tim Duesterhus27c70ae2021-01-23 17:50:21 +010018928 sample fetch is really useful (and usable) in the health-check context.
18929
18930 It may be used in tcp-check based expect rules.
Christopher Faulete596d182020-05-05 17:46:34 +020018931
18932res.body_len : integer
18933 This returns the length of the HTTP response available body in bytes. Unlike
18934 the request side, there is no directive to wait for the response's body. This
Tim Duesterhus27c70ae2021-01-23 17:50:21 +010018935 sample fetch is really useful (and usable) in the health-check context.
18936
18937 It may be used in tcp-check based expect rules.
Christopher Faulete596d182020-05-05 17:46:34 +020018938
18939res.body_size : integer
18940 This returns the advertised length of the HTTP response body in bytes. It
18941 will represent the advertised Content-Length header, or the size of the
18942 available data in case of chunked encoding. Unlike the request side, there is
18943 no directive to wait for the response body. This sample fetch is really
Tim Duesterhus27c70ae2021-01-23 17:50:21 +010018944 useful (and usable) in the health-check context.
18945
18946 It may be used in tcp-check based expect rules.
Christopher Faulete596d182020-05-05 17:46:34 +020018947
Remi Tricot-Le Bretonbf971212020-10-27 11:55:57 +010018948res.cache_hit : boolean
18949 Returns the boolean "true" value if the response has been built out of an
18950 HTTP cache entry, otherwise returns boolean "false".
18951
18952res.cache_name : string
18953 Returns a string containing the name of the HTTP cache that was used to
18954 build the HTTP response if res.cache_hit is true, otherwise returns an
18955 empty string.
18956
Willy Tarreau74ca5042013-06-11 23:12:07 +020018957res.comp : boolean
18958 Returns the boolean "true" value if the response has been compressed by
18959 HAProxy, otherwise returns boolean "false". This may be used to add
18960 information in the logs.
Willy Tarreau6a06a402007-07-15 20:15:28 +020018961
Willy Tarreau74ca5042013-06-11 23:12:07 +020018962res.comp_algo : string
18963 Returns a string containing the name of the algorithm used if the response
18964 was compressed by HAProxy, for example : "deflate". This may be used to add
18965 some information in the logs.
Willy Tarreaud63335a2010-02-26 12:56:52 +010018966
Willy Tarreau74ca5042013-06-11 23:12:07 +020018967res.cook([<name>]) : string
18968scook([<name>]) : string (deprecated)
18969 This extracts the last occurrence of the cookie name <name> on a "Set-Cookie"
18970 header line from the response, and returns its value as string. If no name is
Tim Duesterhus27c70ae2021-01-23 17:50:21 +010018971 specified, the first cookie value is returned.
18972
18973 It may be used in tcp-check based expect rules.
Willy Tarreau0ce3aa02012-04-25 18:46:33 +020018974
Willy Tarreau74ca5042013-06-11 23:12:07 +020018975 ACL derivatives :
18976 scook([<name>] : exact string match
Willy Tarreau0ce3aa02012-04-25 18:46:33 +020018977
Willy Tarreau74ca5042013-06-11 23:12:07 +020018978res.cook_cnt([<name>]) : integer
18979scook_cnt([<name>]) : integer (deprecated)
18980 Returns an integer value representing the number of occurrences of the cookie
18981 <name> in the response, or all cookies if <name> is not specified. This is
Tim Duesterhus27c70ae2021-01-23 17:50:21 +010018982 mostly useful when combined with ACLs to detect suspicious responses.
18983
18984 It may be used in tcp-check based expect rules.
Willy Tarreaud63335a2010-02-26 12:56:52 +010018985
Willy Tarreau74ca5042013-06-11 23:12:07 +020018986res.cook_val([<name>]) : integer
18987scook_val([<name>]) : integer (deprecated)
18988 This extracts the last occurrence of the cookie name <name> on a "Set-Cookie"
18989 header line from the response, and converts its value to an integer which is
Tim Duesterhus27c70ae2021-01-23 17:50:21 +010018990 returned. If no name is specified, the first cookie value is returned.
18991
18992 It may be used in tcp-check based expect rules.
Willy Tarreaud63335a2010-02-26 12:56:52 +010018993
Willy Tarreau74ca5042013-06-11 23:12:07 +020018994res.fhdr([<name>[,<occ>]]) : string
Tim Duesterhus27c70ae2021-01-23 17:50:21 +010018995 This fetch works like the req.fhdr() fetch with the difference that it acts
18996 on the headers within an HTTP response.
18997
18998 Like req.fhdr() the res.fhdr() fetch returns full values. If the header is
18999 defined to be a list you should use res.hdr().
19000
19001 This fetch is sometimes useful with headers such as Date or Expires.
19002
19003 It may be used in tcp-check based expect rules.
Willy Tarreau6a06a402007-07-15 20:15:28 +020019004
Willy Tarreau74ca5042013-06-11 23:12:07 +020019005res.fhdr_cnt([<name>]) : integer
Tim Duesterhus27c70ae2021-01-23 17:50:21 +010019006 This fetch works like the req.fhdr_cnt() fetch with the difference that it
19007 acts on the headers within an HTTP response.
19008
19009 Like req.fhdr_cnt() the res.fhdr_cnt() fetch acts on full values. If the
19010 header is defined to be a list you should use res.hdr_cnt().
19011
19012 It may be used in tcp-check based expect rules.
Willy Tarreau6a06a402007-07-15 20:15:28 +020019013
Willy Tarreau74ca5042013-06-11 23:12:07 +020019014res.hdr([<name>[,<occ>]]) : string
19015shdr([<name>[,<occ>]]) : string (deprecated)
Tim Duesterhus27c70ae2021-01-23 17:50:21 +010019016 This fetch works like the req.hdr() fetch with the difference that it acts
19017 on the headers within an HTTP response.
19018
Ilya Shipitsinacf84592021-02-06 22:29:08 +050019019 Like req.hdr() the res.hdr() fetch considers the comma to be a delimiter. If
Tim Duesterhus27c70ae2021-01-23 17:50:21 +010019020 this is not desired res.fhdr() should be used.
19021
19022 It may be used in tcp-check based expect rules.
Willy Tarreau6a06a402007-07-15 20:15:28 +020019023
Willy Tarreau74ca5042013-06-11 23:12:07 +020019024 ACL derivatives :
19025 shdr([<name>[,<occ>]]) : exact string match
19026 shdr_beg([<name>[,<occ>]]) : prefix match
19027 shdr_dir([<name>[,<occ>]]) : subdir match
19028 shdr_dom([<name>[,<occ>]]) : domain match
19029 shdr_end([<name>[,<occ>]]) : suffix match
19030 shdr_len([<name>[,<occ>]]) : length match
19031 shdr_reg([<name>[,<occ>]]) : regex match
19032 shdr_sub([<name>[,<occ>]]) : substring match
19033
19034res.hdr_cnt([<name>]) : integer
19035shdr_cnt([<name>]) : integer (deprecated)
Tim Duesterhus27c70ae2021-01-23 17:50:21 +010019036 This fetch works like the req.hdr_cnt() fetch with the difference that it
19037 acts on the headers within an HTTP response.
19038
19039 Like req.hdr_cnt() the res.hdr_cnt() fetch considers the comma to be a
Ilya Shipitsinacf84592021-02-06 22:29:08 +050019040 delimiter. If this is not desired res.fhdr_cnt() should be used.
Tim Duesterhus27c70ae2021-01-23 17:50:21 +010019041
19042 It may be used in tcp-check based expect rules.
Willy Tarreau6a06a402007-07-15 20:15:28 +020019043
Willy Tarreau74ca5042013-06-11 23:12:07 +020019044res.hdr_ip([<name>[,<occ>]]) : ip
19045shdr_ip([<name>[,<occ>]]) : ip (deprecated)
Tim Duesterhus27c70ae2021-01-23 17:50:21 +010019046 This fetch works like the req.hdr_ip() fetch with the difference that it
19047 acts on the headers within an HTTP response.
19048
19049 This can be useful to learn some data into a stick table.
19050
19051 It may be used in tcp-check based expect rules.
Willy Tarreau6a06a402007-07-15 20:15:28 +020019052
Willy Tarreaueb27ec72015-02-20 13:55:29 +010019053res.hdr_names([<delim>]) : string
19054 This builds a string made from the concatenation of all header names as they
19055 appear in the response when the rule is evaluated. The default delimiter is
19056 the comma (',') but it may be overridden as an optional argument <delim>. In
Tim Duesterhus27c70ae2021-01-23 17:50:21 +010019057 this case, only the first character of <delim> is considered.
19058
19059 It may be used in tcp-check based expect rules.
Willy Tarreaueb27ec72015-02-20 13:55:29 +010019060
Willy Tarreau74ca5042013-06-11 23:12:07 +020019061res.hdr_val([<name>[,<occ>]]) : integer
19062shdr_val([<name>[,<occ>]]) : integer (deprecated)
Tim Duesterhus27c70ae2021-01-23 17:50:21 +010019063 This fetch works like the req.hdr_val() fetch with the difference that it
19064 acts on the headers within an HTTP response.
19065
19066 This can be useful to learn some data into a stick table.
19067
19068 It may be used in tcp-check based expect rules.
Christopher Faulete596d182020-05-05 17:46:34 +020019069
19070res.hdrs : string
19071 Returns the current response headers as string including the last empty line
19072 separating headers from the request body. The last empty line can be used to
19073 detect a truncated header block. This sample fetch is useful for some SPOE
Tim Duesterhus27c70ae2021-01-23 17:50:21 +010019074 headers analyzers and for advanced logging.
19075
19076 It may also be used in tcp-check based expect rules.
Christopher Faulete596d182020-05-05 17:46:34 +020019077
19078res.hdrs_bin : binary
19079 Returns the current response headers contained in preparsed binary form. This
19080 is useful for offloading some processing with SPOE. It may be used in
19081 tcp-check based expect rules. Each string is described by a length followed
19082 by the number of bytes indicated in the length. The length is represented
19083 using the variable integer encoding detailed in the SPOE documentation. The
19084 end of the list is marked by a couple of empty header names and values
19085 (length of 0 for both).
19086
19087 *(<str:header-name><str:header-value>)<empty string><empty string>
19088
19089 int: refer to the SPOE documentation for the encoding
19090 str: <int:length><bytes>
Alexandre Cassen5eb1a902007-11-29 15:43:32 +010019091
Willy Tarreau74ca5042013-06-11 23:12:07 +020019092res.ver : string
19093resp_ver : string (deprecated)
19094 Returns the version string from the HTTP response, for example "1.1". This
Tim Duesterhus27c70ae2021-01-23 17:50:21 +010019095 can be useful for logs, but is mostly there for ACL.
19096
19097 It may be used in tcp-check based expect rules.
Willy Tarreau0e698542011-09-16 08:32:32 +020019098
Willy Tarreau74ca5042013-06-11 23:12:07 +020019099 ACL derivatives :
19100 resp_ver : exact string match
Alexandre Cassen5eb1a902007-11-29 15:43:32 +010019101
Willy Tarreau74ca5042013-06-11 23:12:07 +020019102set-cookie([<name>]) : string (deprecated)
19103 This extracts the last occurrence of the cookie name <name> on a "Set-Cookie"
19104 header line from the response and uses the corresponding value to match. This
Willy Tarreau294d0f02015-08-10 19:40:12 +020019105 can be comparable to what "appsession" did with default options, but with
Willy Tarreau74ca5042013-06-11 23:12:07 +020019106 support for multi-peer synchronization and state keeping across restarts.
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +010019107
Willy Tarreau74ca5042013-06-11 23:12:07 +020019108 This fetch function is deprecated and has been superseded by the "res.cook"
19109 fetch. This keyword will disappear soon.
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +010019110
Willy Tarreau74ca5042013-06-11 23:12:07 +020019111status : integer
19112 Returns an integer containing the HTTP status code in the HTTP response, for
19113 example, 302. It is mostly used within ACLs and integer ranges, for example,
Tim Duesterhus27c70ae2021-01-23 17:50:21 +010019114 to remove any Location header if the response is not a 3xx.
19115
19116 It may be used in tcp-check based expect rules.
Willy Tarreau25c1ebc2012-04-25 16:21:44 +020019117
Thierry Fournier0e00dca2016-04-07 15:47:40 +020019118unique-id : string
19119 Returns the unique-id attached to the request. The directive
19120 "unique-id-format" must be set. If it is not set, the unique-id sample fetch
19121 fails. Note that the unique-id is usually used with HTTP requests, however this
19122 sample fetch can be used with other protocols. Obviously, if it is used with
19123 other protocols than HTTP, the unique-id-format directive must not contain
19124 HTTP parts. See: unique-id-format and unique-id-header
19125
Willy Tarreau74ca5042013-06-11 23:12:07 +020019126url : string
19127 This extracts the request's URL as presented in the request. A typical use is
19128 with prefetch-capable caches, and with portals which need to aggregate
19129 multiple information from databases and keep them in caches. With ACLs, using
19130 "path" is preferred over using "url", because clients may send a full URL as
19131 is normally done with proxies. The only real use is to match "*" which does
19132 not match in "path", and for which there is already a predefined ACL. See
19133 also "path" and "base".
Willy Tarreau25c1ebc2012-04-25 16:21:44 +020019134
Willy Tarreau74ca5042013-06-11 23:12:07 +020019135 ACL derivatives :
19136 url : exact string match
19137 url_beg : prefix match
19138 url_dir : subdir match
19139 url_dom : domain match
19140 url_end : suffix match
19141 url_len : length match
19142 url_reg : regex match
19143 url_sub : substring match
Willy Tarreau25c1ebc2012-04-25 16:21:44 +020019144
Willy Tarreau74ca5042013-06-11 23:12:07 +020019145url_ip : ip
19146 This extracts the IP address from the request's URL when the host part is
19147 presented as an IP address. Its use is very limited. For instance, a
19148 monitoring system might use this field as an alternative for the source IP in
19149 order to test what path a given source address would follow, or to force an
19150 entry in a table for a given source address. With ACLs it can be used to
19151 restrict access to certain systems through a proxy, for example when combined
19152 with option "http_proxy".
Willy Tarreau25c1ebc2012-04-25 16:21:44 +020019153
Willy Tarreau74ca5042013-06-11 23:12:07 +020019154url_port : integer
19155 This extracts the port part from the request's URL. Note that if the port is
19156 not specified in the request, port 80 is assumed. With ACLs it can be used to
19157 restrict access to certain systems through a proxy, for example when combined
19158 with option "http_proxy".
Willy Tarreau25c1ebc2012-04-25 16:21:44 +020019159
Willy Tarreau1ede1da2015-05-07 16:06:18 +020019160urlp([<name>[,<delim>]]) : string
19161url_param([<name>[,<delim>]]) : string
Willy Tarreau74ca5042013-06-11 23:12:07 +020019162 This extracts the first occurrence of the parameter <name> in the query
19163 string, which begins after either '?' or <delim>, and which ends before '&',
Willy Tarreau1ede1da2015-05-07 16:06:18 +020019164 ';' or <delim>. The parameter name is case-sensitive. If no name is given,
19165 any parameter will match, and the first one will be returned. The result is
19166 a string corresponding to the value of the parameter <name> as presented in
19167 the request (no URL decoding is performed). This can be used for session
Willy Tarreau74ca5042013-06-11 23:12:07 +020019168 stickiness based on a client ID, to extract an application cookie passed as a
19169 URL parameter, or in ACLs to apply some checks. Note that the ACL version of
Willy Tarreau1ede1da2015-05-07 16:06:18 +020019170 this fetch iterates over multiple parameters and will iteratively report all
19171 parameters values if no name is given
Willy Tarreau25c1ebc2012-04-25 16:21:44 +020019172
Willy Tarreau74ca5042013-06-11 23:12:07 +020019173 ACL derivatives :
19174 urlp(<name>[,<delim>]) : exact string match
19175 urlp_beg(<name>[,<delim>]) : prefix match
19176 urlp_dir(<name>[,<delim>]) : subdir match
19177 urlp_dom(<name>[,<delim>]) : domain match
19178 urlp_end(<name>[,<delim>]) : suffix match
19179 urlp_len(<name>[,<delim>]) : length match
19180 urlp_reg(<name>[,<delim>]) : regex match
19181 urlp_sub(<name>[,<delim>]) : substring match
Willy Tarreau25c1ebc2012-04-25 16:21:44 +020019182
Willy Tarreau25c1ebc2012-04-25 16:21:44 +020019183
Willy Tarreau74ca5042013-06-11 23:12:07 +020019184 Example :
19185 # match http://example.com/foo?PHPSESSIONID=some_id
19186 stick on urlp(PHPSESSIONID)
19187 # match http://example.com/foo;JSESSIONID=some_id
19188 stick on urlp(JSESSIONID,;)
Willy Tarreau25c1ebc2012-04-25 16:21:44 +020019189
Jarno Huuskonen676f6222017-03-30 09:19:45 +030019190urlp_val([<name>[,<delim>]]) : integer
Willy Tarreau74ca5042013-06-11 23:12:07 +020019191 See "urlp" above. This one extracts the URL parameter <name> in the request
19192 and converts it to an integer value. This can be used for session stickiness
19193 based on a user ID for example, or with ACLs to match a page number or price.
Willy Tarreaua9fddca2012-07-31 07:51:48 +020019194
Dragan Dosen0070cd52016-06-16 12:19:49 +020019195url32 : integer
19196 This returns a 32-bit hash of the value obtained by concatenating the first
19197 Host header and the whole URL including parameters (not only the path part of
19198 the request, as in the "base32" fetch above). This is useful to track per-URL
19199 activity. A shorter hash is stored, saving a lot of memory. The output type
19200 is an unsigned integer.
19201
19202url32+src : binary
19203 This returns the concatenation of the "url32" fetch and the "src" fetch. The
19204 resulting type is of type binary, with a size of 8 or 20 bytes depending on
19205 the source address family. This can be used to track per-IP, per-URL counters.
19206
Christopher Faulet16032ab2020-04-30 11:30:00 +020019207
Christopher Faulete596d182020-05-05 17:46:34 +0200192087.3.7. Fetching samples for developers
Christopher Fauletd47941d2020-01-08 14:40:19 +010019209---------------------------------------
19210
19211This set of sample fetch methods is reserved to developers and must never be
19212used on a production environment, except on developer demand, for debugging
19213purposes. Moreover, no special care will be taken on backwards compatibility.
19214There is no warranty the following sample fetches will never change, be renamed
19215or simply removed. So be really careful if you should use one of them. To avoid
19216any ambiguity, these sample fetches are placed in the dedicated scope "internal",
19217for instance "internal.strm.is_htx".
19218
19219internal.htx.data : integer
19220 Returns the size in bytes used by data in the HTX message associated to a
19221 channel. The channel is chosen depending on the sample direction.
19222
19223internal.htx.free : integer
19224 Returns the free space (size - used) in bytes in the HTX message associated
19225 to a channel. The channel is chosen depending on the sample direction.
19226
19227internal.htx.free_data : integer
19228 Returns the free space for the data in bytes in the HTX message associated to
19229 a channel. The channel is chosen depending on the sample direction.
19230
19231internal.htx.has_eom : boolean
Christopher Fauletd1ac2b92020-12-02 19:12:22 +010019232 Returns true if the HTX message associated to a channel contains the
19233 end-of-message flag (EOM). Otherwise, it returns false. The channel is chosen
19234 depending on the sample direction.
Christopher Fauletd47941d2020-01-08 14:40:19 +010019235
19236internal.htx.nbblks : integer
19237 Returns the number of blocks present in the HTX message associated to a
19238 channel. The channel is chosen depending on the sample direction.
19239
19240internal.htx.size : integer
19241 Returns the total size in bytes of the HTX message associated to a
19242 channel. The channel is chosen depending on the sample direction.
19243
19244internal.htx.used : integer
19245 Returns the total size used in bytes (data + metadata) in the HTX message
19246 associated to a channel. The channel is chosen depending on the sample
19247 direction.
19248
19249internal.htx_blk.size(<idx>) : integer
19250 Returns the size of the block at the position <idx> in the HTX message
19251 associated to a channel or 0 if it does not exist. The channel is chosen
19252 depending on the sample direction. <idx> may be any positive integer or one
19253 of the special value :
19254 * head : The oldest inserted block
19255 * tail : The newest inserted block
Ilya Shipitsin8525fd92020-02-29 12:34:59 +050019256 * first : The first block where to (re)start the analysis
Christopher Fauletd47941d2020-01-08 14:40:19 +010019257
19258internal.htx_blk.type(<idx>) : string
19259 Returns the type of the block at the position <idx> in the HTX message
19260 associated to a channel or "HTX_BLK_UNUSED" if it does not exist. The channel
19261 is chosen depending on the sample direction. <idx> may be any positive
19262 integer or one of the special value :
19263 * head : The oldest inserted block
19264 * tail : The newest inserted block
Ilya Shipitsin8525fd92020-02-29 12:34:59 +050019265 * first : The first block where to (re)start the analysis
Christopher Fauletd47941d2020-01-08 14:40:19 +010019266
19267internal.htx_blk.data(<idx>) : binary
19268 Returns the value of the DATA block at the position <idx> in the HTX message
19269 associated to a channel or an empty string if it does not exist or if it is
19270 not a DATA block. The channel is chosen depending on the sample direction.
19271 <idx> may be any positive integer or one of the special value :
19272
19273 * head : The oldest inserted block
19274 * tail : The newest inserted block
Ilya Shipitsin8525fd92020-02-29 12:34:59 +050019275 * first : The first block where to (re)start the analysis
Christopher Fauletd47941d2020-01-08 14:40:19 +010019276
19277internal.htx_blk.hdrname(<idx>) : string
19278 Returns the header name of the HEADER block at the position <idx> in the HTX
19279 message associated to a channel or an empty string if it does not exist or if
19280 it is not an HEADER block. The channel is chosen depending on the sample
19281 direction. <idx> may be any positive integer or one of the special value :
19282
19283 * head : The oldest inserted block
19284 * tail : The newest inserted block
Ilya Shipitsin8525fd92020-02-29 12:34:59 +050019285 * first : The first block where to (re)start the analysis
Christopher Fauletd47941d2020-01-08 14:40:19 +010019286
19287internal.htx_blk.hdrval(<idx>) : string
19288 Returns the header value of the HEADER block at the position <idx> in the HTX
19289 message associated to a channel or an empty string if it does not exist or if
19290 it is not an HEADER block. The channel is chosen depending on the sample
19291 direction. <idx> may be any positive integer or one of the special value :
19292
19293 * head : The oldest inserted block
19294 * tail : The newest inserted block
Ilya Shipitsin8525fd92020-02-29 12:34:59 +050019295 * first : The first block where to (re)start the analysis
Christopher Fauletd47941d2020-01-08 14:40:19 +010019296
19297internal.htx_blk.start_line(<idx>) : string
19298 Returns the value of the REQ_SL or RES_SL block at the position <idx> in the
19299 HTX message associated to a channel or an empty string if it does not exist
19300 or if it is not a SL block. The channel is chosen depending on the sample
19301 direction. <idx> may be any positive integer or one of the special value :
19302
19303 * head : The oldest inserted block
19304 * tail : The newest inserted block
Ilya Shipitsin8525fd92020-02-29 12:34:59 +050019305 * first : The first block where to (re)start the analysis
Christopher Fauletd47941d2020-01-08 14:40:19 +010019306
19307internal.strm.is_htx : boolean
19308 Returns true if the current stream is an HTX stream. It means the data in the
19309 channels buffers are stored using the internal HTX representation. Otherwise,
19310 it returns false.
19311
19312
Willy Tarreau74ca5042013-06-11 23:12:07 +0200193137.4. Pre-defined ACLs
Willy Tarreauc57f0e22009-05-10 13:12:33 +020019314---------------------
Willy Tarreauced27012008-01-17 20:35:34 +010019315
Willy Tarreauc57f0e22009-05-10 13:12:33 +020019316Some predefined ACLs are hard-coded so that they do not have to be declared in
19317every frontend which needs them. They all have their names in upper case in
Patrick Mézard2382ad62010-05-09 10:43:32 +020019318order to avoid confusion. Their equivalence is provided below.
Willy Tarreauced27012008-01-17 20:35:34 +010019319
Willy Tarreauc57f0e22009-05-10 13:12:33 +020019320ACL name Equivalent to Usage
19321---------------+-----------------------------+---------------------------------
Willy Tarreauc57f0e22009-05-10 13:12:33 +020019322FALSE always_false never match
Willy Tarreau2492d5b2009-07-11 00:06:00 +020019323HTTP req_proto_http match if protocol is valid HTTP
Willy Tarreauc57f0e22009-05-10 13:12:33 +020019324HTTP_1.0 req_ver 1.0 match HTTP version 1.0
19325HTTP_1.1 req_ver 1.1 match HTTP version 1.1
Willy Tarreaud63335a2010-02-26 12:56:52 +010019326HTTP_CONTENT hdr_val(content-length) gt 0 match an existing content-length
19327HTTP_URL_ABS url_reg ^[^/:]*:// match absolute URL with scheme
19328HTTP_URL_SLASH url_beg / match URL beginning with "/"
19329HTTP_URL_STAR url * match URL equal to "*"
19330LOCALHOST src 127.0.0.1/8 match connection from local host
Willy Tarreauc57f0e22009-05-10 13:12:33 +020019331METH_CONNECT method CONNECT match HTTP CONNECT method
Daniel Schneller9ff96c72016-04-11 17:45:29 +020019332METH_DELETE method DELETE match HTTP DELETE method
Willy Tarreauc57f0e22009-05-10 13:12:33 +020019333METH_GET method GET HEAD match HTTP GET or HEAD method
19334METH_HEAD method HEAD match HTTP HEAD method
19335METH_OPTIONS method OPTIONS match HTTP OPTIONS method
19336METH_POST method POST match HTTP POST method
Daniel Schneller9ff96c72016-04-11 17:45:29 +020019337METH_PUT method PUT match HTTP PUT method
Willy Tarreauc57f0e22009-05-10 13:12:33 +020019338METH_TRACE method TRACE match HTTP TRACE method
Emeric Brunbede3d02009-06-30 17:54:00 +020019339RDP_COOKIE req_rdp_cookie_cnt gt 0 match presence of an RDP cookie
Willy Tarreauc57f0e22009-05-10 13:12:33 +020019340REQ_CONTENT req_len gt 0 match data in the request buffer
Willy Tarreaud63335a2010-02-26 12:56:52 +010019341TRUE always_true always match
Willy Tarreauc57f0e22009-05-10 13:12:33 +020019342WAIT_END wait_end wait for end of content analysis
19343---------------+-----------------------------+---------------------------------
Willy Tarreauced27012008-01-17 20:35:34 +010019344
Willy Tarreaub937b7e2010-01-12 15:27:54 +010019345
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200193468. Logging
19347----------
Willy Tarreau844e3c52008-01-11 16:28:18 +010019348
Willy Tarreaucc6c8912009-02-22 10:53:55 +010019349One of HAProxy's strong points certainly lies is its precise logs. It probably
19350provides the finest level of information available for such a product, which is
19351very important for troubleshooting complex environments. Standard information
19352provided in logs include client ports, TCP/HTTP state timers, precise session
19353state at termination and precise termination cause, information about decisions
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +010019354to direct traffic to a server, and of course the ability to capture arbitrary
Willy Tarreaucc6c8912009-02-22 10:53:55 +010019355headers.
19356
19357In order to improve administrators reactivity, it offers a great transparency
19358about encountered problems, both internal and external, and it is possible to
19359send logs to different sources at the same time with different level filters :
19360
19361 - global process-level logs (system errors, start/stop, etc..)
19362 - per-instance system and internal errors (lack of resource, bugs, ...)
19363 - per-instance external troubles (servers up/down, max connections)
19364 - per-instance activity (client connections), either at the establishment or
19365 at the termination.
Davor Ocelice9ed2812017-12-25 17:49:28 +010019366 - per-request control of log-level, e.g.
Jim Freeman9e8714b2015-05-26 09:16:34 -060019367 http-request set-log-level silent if sensitive_request
Willy Tarreaucc6c8912009-02-22 10:53:55 +010019368
19369The ability to distribute different levels of logs to different log servers
19370allow several production teams to interact and to fix their problems as soon
19371as possible. For example, the system team might monitor system-wide errors,
19372while the application team might be monitoring the up/down for their servers in
19373real time, and the security team might analyze the activity logs with one hour
19374delay.
19375
19376
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200193778.1. Log levels
19378---------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010019379
Simon Hormandf791f52011-05-29 15:01:10 +090019380TCP and HTTP connections can be logged with information such as the date, time,
Willy Tarreaucc6c8912009-02-22 10:53:55 +010019381source IP address, destination address, connection duration, response times,
Simon Hormandf791f52011-05-29 15:01:10 +090019382HTTP request, HTTP return code, number of bytes transmitted, conditions
19383in which the session ended, and even exchanged cookies values. For example
19384track a particular user's problems. All messages may be sent to up to two
19385syslog servers. Check the "log" keyword in section 4.2 for more information
19386about log facilities.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010019387
19388
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200193898.2. Log formats
19390----------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010019391
William Lallemand48940402012-01-30 16:47:22 +010019392HAProxy supports 5 log formats. Several fields are common between these formats
Simon Hormandf791f52011-05-29 15:01:10 +090019393and will be detailed in the following sections. A few of them may vary
19394slightly with the configuration, due to indicators specific to certain
19395options. The supported formats are as follows :
Willy Tarreaucc6c8912009-02-22 10:53:55 +010019396
19397 - the default format, which is very basic and very rarely used. It only
19398 provides very basic information about the incoming connection at the moment
19399 it is accepted : source IP:port, destination IP:port, and frontend-name.
19400 This mode will eventually disappear so it will not be described to great
19401 extents.
19402
19403 - the TCP format, which is more advanced. This format is enabled when "option
19404 tcplog" is set on the frontend. HAProxy will then usually wait for the
19405 connection to terminate before logging. This format provides much richer
19406 information, such as timers, connection counts, queue size, etc... This
19407 format is recommended for pure TCP proxies.
19408
19409 - the HTTP format, which is the most advanced for HTTP proxying. This format
19410 is enabled when "option httplog" is set on the frontend. It provides the
19411 same information as the TCP format with some HTTP-specific fields such as
19412 the request, the status code, and captures of headers and cookies. This
19413 format is recommended for HTTP proxies.
19414
Emeric Brun3a058f32009-06-30 18:26:00 +020019415 - the CLF HTTP format, which is equivalent to the HTTP format, but with the
19416 fields arranged in the same order as the CLF format. In this mode, all
19417 timers, captures, flags, etc... appear one per field after the end of the
19418 common fields, in the same order they appear in the standard HTTP format.
19419
William Lallemand48940402012-01-30 16:47:22 +010019420 - the custom log format, allows you to make your own log line.
19421
Willy Tarreaucc6c8912009-02-22 10:53:55 +010019422Next sections will go deeper into details for each of these formats. Format
19423specification will be performed on a "field" basis. Unless stated otherwise, a
19424field is a portion of text delimited by any number of spaces. Since syslog
19425servers are susceptible of inserting fields at the beginning of a line, it is
19426always assumed that the first field is the one containing the process name and
19427identifier.
19428
19429Note : Since log lines may be quite long, the log examples in sections below
19430 might be broken into multiple lines. The example log lines will be
19431 prefixed with 3 closing angle brackets ('>>>') and each time a log is
19432 broken into multiple lines, each non-final line will end with a
19433 backslash ('\') and the next line will start indented by two characters.
19434
19435
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200194368.2.1. Default log format
19437-------------------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010019438
19439This format is used when no specific option is set. The log is emitted as soon
19440as the connection is accepted. One should note that this currently is the only
19441format which logs the request's destination IP and ports.
19442
19443 Example :
19444 listen www
19445 mode http
19446 log global
19447 server srv1 127.0.0.1:8000
19448
19449 >>> Feb 6 12:12:09 localhost \
19450 haproxy[14385]: Connect from 10.0.1.2:33312 to 10.0.3.31:8012 \
19451 (www/HTTP)
19452
19453 Field Format Extract from the example above
19454 1 process_name '[' pid ']:' haproxy[14385]:
19455 2 'Connect from' Connect from
19456 3 source_ip ':' source_port 10.0.1.2:33312
19457 4 'to' to
19458 5 destination_ip ':' destination_port 10.0.3.31:8012
19459 6 '(' frontend_name '/' mode ')' (www/HTTP)
19460
19461Detailed fields description :
19462 - "source_ip" is the IP address of the client which initiated the connection.
19463 - "source_port" is the TCP port of the client which initiated the connection.
19464 - "destination_ip" is the IP address the client connected to.
19465 - "destination_port" is the TCP port the client connected to.
19466 - "frontend_name" is the name of the frontend (or listener) which received
19467 and processed the connection.
19468 - "mode is the mode the frontend is operating (TCP or HTTP).
19469
Willy Tarreauceb24bc2010-11-09 12:46:41 +010019470In case of a UNIX socket, the source and destination addresses are marked as
19471"unix:" and the ports reflect the internal ID of the socket which accepted the
19472connection (the same ID as reported in the stats).
19473
Willy Tarreaucc6c8912009-02-22 10:53:55 +010019474It is advised not to use this deprecated format for newer installations as it
19475will eventually disappear.
19476
19477
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200194788.2.2. TCP log format
19479---------------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010019480
19481The TCP format is used when "option tcplog" is specified in the frontend, and
19482is the recommended format for pure TCP proxies. It provides a lot of precious
19483information for troubleshooting. Since this format includes timers and byte
19484counts, the log is normally emitted at the end of the session. It can be
19485emitted earlier if "option logasap" is specified, which makes sense in most
19486environments with long sessions such as remote terminals. Sessions which match
19487the "monitor" rules are never logged. It is also possible not to emit logs for
19488sessions for which no data were exchanged between the client and the server, by
Willy Tarreauc9bd0cc2009-05-10 11:57:02 +020019489specifying "option dontlognull" in the frontend. Successful connections will
19490not be logged if "option dontlog-normal" is specified in the frontend. A few
19491fields may slightly vary depending on some configuration options, those are
19492marked with a star ('*') after the field name below.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010019493
19494 Example :
19495 frontend fnt
19496 mode tcp
19497 option tcplog
19498 log global
19499 default_backend bck
19500
19501 backend bck
19502 server srv1 127.0.0.1:8000
19503
19504 >>> Feb 6 12:12:56 localhost \
19505 haproxy[14387]: 10.0.1.2:33313 [06/Feb/2009:12:12:51.443] fnt \
19506 bck/srv1 0/0/5007 212 -- 0/0/0/0/3 0/0
19507
19508 Field Format Extract from the example above
19509 1 process_name '[' pid ']:' haproxy[14387]:
19510 2 client_ip ':' client_port 10.0.1.2:33313
19511 3 '[' accept_date ']' [06/Feb/2009:12:12:51.443]
19512 4 frontend_name fnt
19513 5 backend_name '/' server_name bck/srv1
19514 6 Tw '/' Tc '/' Tt* 0/0/5007
19515 7 bytes_read* 212
19516 8 termination_state --
19517 9 actconn '/' feconn '/' beconn '/' srv_conn '/' retries* 0/0/0/0/3
19518 10 srv_queue '/' backend_queue 0/0
19519
19520Detailed fields description :
19521 - "client_ip" is the IP address of the client which initiated the TCP
Willy Tarreauceb24bc2010-11-09 12:46:41 +010019522 connection to haproxy. If the connection was accepted on a UNIX socket
19523 instead, the IP address would be replaced with the word "unix". Note that
19524 when the connection is accepted on a socket configured with "accept-proxy"
Bertrand Jacquin93b227d2016-06-04 15:11:10 +010019525 and the PROXY protocol is correctly used, or with a "accept-netscaler-cip"
Davor Ocelice9ed2812017-12-25 17:49:28 +010019526 and the NetScaler Client IP insertion protocol is correctly used, then the
Bertrand Jacquin93b227d2016-06-04 15:11:10 +010019527 logs will reflect the forwarded connection's information.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010019528
19529 - "client_port" is the TCP port of the client which initiated the connection.
Willy Tarreauceb24bc2010-11-09 12:46:41 +010019530 If the connection was accepted on a UNIX socket instead, the port would be
19531 replaced with the ID of the accepting socket, which is also reported in the
19532 stats interface.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010019533
19534 - "accept_date" is the exact date when the connection was received by haproxy
19535 (which might be very slightly different from the date observed on the
19536 network if there was some queuing in the system's backlog). This is usually
Willy Tarreau590a0512018-09-05 11:56:48 +020019537 the same date which may appear in any upstream firewall's log. When used in
19538 HTTP mode, the accept_date field will be reset to the first moment the
19539 connection is ready to receive a new request (end of previous response for
19540 HTTP/1, immediately after previous request for HTTP/2).
Willy Tarreaucc6c8912009-02-22 10:53:55 +010019541
19542 - "frontend_name" is the name of the frontend (or listener) which received
19543 and processed the connection.
19544
19545 - "backend_name" is the name of the backend (or listener) which was selected
19546 to manage the connection to the server. This will be the same as the
19547 frontend if no switching rule has been applied, which is common for TCP
19548 applications.
19549
19550 - "server_name" is the name of the last server to which the connection was
19551 sent, which might differ from the first one if there were connection errors
19552 and a redispatch occurred. Note that this server belongs to the backend
19553 which processed the request. If the connection was aborted before reaching
19554 a server, "<NOSRV>" is indicated instead of a server name.
19555
19556 - "Tw" is the total time in milliseconds spent waiting in the various queues.
19557 It can be "-1" if the connection was aborted before reaching the queue.
19558 See "Timers" below for more details.
19559
19560 - "Tc" is the total time in milliseconds spent waiting for the connection to
19561 establish to the final server, including retries. It can be "-1" if the
19562 connection was aborted before a connection could be established. See
19563 "Timers" below for more details.
19564
19565 - "Tt" is the total time in milliseconds elapsed between the accept and the
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030019566 last close. It covers all possible processing. There is one exception, if
Willy Tarreaucc6c8912009-02-22 10:53:55 +010019567 "option logasap" was specified, then the time counting stops at the moment
19568 the log is emitted. In this case, a '+' sign is prepended before the value,
19569 indicating that the final one will be larger. See "Timers" below for more
19570 details.
19571
19572 - "bytes_read" is the total number of bytes transmitted from the server to
19573 the client when the log is emitted. If "option logasap" is specified, the
19574 this value will be prefixed with a '+' sign indicating that the final one
19575 may be larger. Please note that this value is a 64-bit counter, so log
19576 analysis tools must be able to handle it without overflowing.
19577
19578 - "termination_state" is the condition the session was in when the session
19579 ended. This indicates the session state, which side caused the end of
19580 session to happen, and for what reason (timeout, error, ...). The normal
19581 flags should be "--", indicating the session was closed by either end with
19582 no data remaining in buffers. See below "Session state at disconnection"
19583 for more details.
19584
19585 - "actconn" is the total number of concurrent connections on the process when
Jamie Gloudonaaa21002012-08-25 00:18:33 -040019586 the session was logged. It is useful to detect when some per-process system
Willy Tarreaucc6c8912009-02-22 10:53:55 +010019587 limits have been reached. For instance, if actconn is close to 512 when
19588 multiple connection errors occur, chances are high that the system limits
19589 the process to use a maximum of 1024 file descriptors and that all of them
Willy Tarreauc57f0e22009-05-10 13:12:33 +020019590 are used. See section 3 "Global parameters" to find how to tune the system.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010019591
19592 - "feconn" is the total number of concurrent connections on the frontend when
19593 the session was logged. It is useful to estimate the amount of resource
19594 required to sustain high loads, and to detect when the frontend's "maxconn"
19595 has been reached. Most often when this value increases by huge jumps, it is
19596 because there is congestion on the backend servers, but sometimes it can be
19597 caused by a denial of service attack.
19598
19599 - "beconn" is the total number of concurrent connections handled by the
19600 backend when the session was logged. It includes the total number of
19601 concurrent connections active on servers as well as the number of
19602 connections pending in queues. It is useful to estimate the amount of
19603 additional servers needed to support high loads for a given application.
19604 Most often when this value increases by huge jumps, it is because there is
19605 congestion on the backend servers, but sometimes it can be caused by a
19606 denial of service attack.
19607
19608 - "srv_conn" is the total number of concurrent connections still active on
19609 the server when the session was logged. It can never exceed the server's
19610 configured "maxconn" parameter. If this value is very often close or equal
19611 to the server's "maxconn", it means that traffic regulation is involved a
19612 lot, meaning that either the server's maxconn value is too low, or that
19613 there aren't enough servers to process the load with an optimal response
19614 time. When only one of the server's "srv_conn" is high, it usually means
19615 that this server has some trouble causing the connections to take longer to
19616 be processed than on other servers.
19617
19618 - "retries" is the number of connection retries experienced by this session
19619 when trying to connect to the server. It must normally be zero, unless a
19620 server is being stopped at the same moment the connection was attempted.
19621 Frequent retries generally indicate either a network problem between
19622 haproxy and the server, or a misconfigured system backlog on the server
19623 preventing new connections from being queued. This field may optionally be
19624 prefixed with a '+' sign, indicating that the session has experienced a
19625 redispatch after the maximal retry count has been reached on the initial
19626 server. In this case, the server name appearing in the log is the one the
19627 connection was redispatched to, and not the first one, though both may
19628 sometimes be the same in case of hashing for instance. So as a general rule
19629 of thumb, when a '+' is present in front of the retry count, this count
19630 should not be attributed to the logged server.
19631
19632 - "srv_queue" is the total number of requests which were processed before
19633 this one in the server queue. It is zero when the request has not gone
19634 through the server queue. It makes it possible to estimate the approximate
19635 server's response time by dividing the time spent in queue by the number of
19636 requests in the queue. It is worth noting that if a session experiences a
19637 redispatch and passes through two server queues, their positions will be
Davor Ocelice9ed2812017-12-25 17:49:28 +010019638 cumulative. A request should not pass through both the server queue and the
Willy Tarreaucc6c8912009-02-22 10:53:55 +010019639 backend queue unless a redispatch occurs.
19640
19641 - "backend_queue" is the total number of requests which were processed before
19642 this one in the backend's global queue. It is zero when the request has not
19643 gone through the global queue. It makes it possible to estimate the average
19644 queue length, which easily translates into a number of missing servers when
19645 divided by a server's "maxconn" parameter. It is worth noting that if a
19646 session experiences a redispatch, it may pass twice in the backend's queue,
Davor Ocelice9ed2812017-12-25 17:49:28 +010019647 and then both positions will be cumulative. A request should not pass
Willy Tarreaucc6c8912009-02-22 10:53:55 +010019648 through both the server queue and the backend queue unless a redispatch
19649 occurs.
19650
19651
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200196528.2.3. HTTP log format
19653----------------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010019654
19655The HTTP format is the most complete and the best suited for HTTP proxies. It
19656is enabled by when "option httplog" is specified in the frontend. It provides
19657the same level of information as the TCP format with additional features which
19658are specific to the HTTP protocol. Just like the TCP format, the log is usually
19659emitted at the end of the session, unless "option logasap" is specified, which
19660generally only makes sense for download sites. A session which matches the
19661"monitor" rules will never logged. It is also possible not to log sessions for
19662which no data were sent by the client by specifying "option dontlognull" in the
Willy Tarreauc9bd0cc2009-05-10 11:57:02 +020019663frontend. Successful connections will not be logged if "option dontlog-normal"
19664is specified in the frontend.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010019665
19666Most fields are shared with the TCP log, some being different. A few fields may
19667slightly vary depending on some configuration options. Those ones are marked
19668with a star ('*') after the field name below.
19669
19670 Example :
19671 frontend http-in
19672 mode http
19673 option httplog
19674 log global
19675 default_backend bck
19676
19677 backend static
19678 server srv1 127.0.0.1:8000
19679
19680 >>> Feb 6 12:14:14 localhost \
19681 haproxy[14389]: 10.0.1.2:33317 [06/Feb/2009:12:14:14.655] http-in \
19682 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 +010019683 {} "GET /index.html HTTP/1.1"
Willy Tarreaucc6c8912009-02-22 10:53:55 +010019684
19685 Field Format Extract from the example above
19686 1 process_name '[' pid ']:' haproxy[14389]:
19687 2 client_ip ':' client_port 10.0.1.2:33317
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020019688 3 '[' request_date ']' [06/Feb/2009:12:14:14.655]
Willy Tarreaucc6c8912009-02-22 10:53:55 +010019689 4 frontend_name http-in
19690 5 backend_name '/' server_name static/srv1
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020019691 6 TR '/' Tw '/' Tc '/' Tr '/' Ta* 10/0/30/69/109
Willy Tarreaucc6c8912009-02-22 10:53:55 +010019692 7 status_code 200
19693 8 bytes_read* 2750
19694 9 captured_request_cookie -
19695 10 captured_response_cookie -
19696 11 termination_state ----
19697 12 actconn '/' feconn '/' beconn '/' srv_conn '/' retries* 1/1/1/1/0
19698 13 srv_queue '/' backend_queue 0/0
19699 14 '{' captured_request_headers* '}' {haproxy.1wt.eu}
19700 15 '{' captured_response_headers* '}' {}
19701 16 '"' http_request '"' "GET /index.html HTTP/1.1"
Willy Tarreaud72758d2010-01-12 10:42:19 +010019702
Willy Tarreaucc6c8912009-02-22 10:53:55 +010019703Detailed fields description :
19704 - "client_ip" is the IP address of the client which initiated the TCP
Willy Tarreauceb24bc2010-11-09 12:46:41 +010019705 connection to haproxy. If the connection was accepted on a UNIX socket
19706 instead, the IP address would be replaced with the word "unix". Note that
19707 when the connection is accepted on a socket configured with "accept-proxy"
Bertrand Jacquin93b227d2016-06-04 15:11:10 +010019708 and the PROXY protocol is correctly used, or with a "accept-netscaler-cip"
Davor Ocelice9ed2812017-12-25 17:49:28 +010019709 and the NetScaler Client IP insertion protocol is correctly used, then the
Bertrand Jacquin93b227d2016-06-04 15:11:10 +010019710 logs will reflect the forwarded connection's information.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010019711
19712 - "client_port" is the TCP port of the client which initiated the connection.
Willy Tarreauceb24bc2010-11-09 12:46:41 +010019713 If the connection was accepted on a UNIX socket instead, the port would be
19714 replaced with the ID of the accepting socket, which is also reported in the
19715 stats interface.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010019716
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020019717 - "request_date" is the exact date when the first byte of the HTTP request
19718 was received by haproxy (log field %tr).
Willy Tarreaucc6c8912009-02-22 10:53:55 +010019719
19720 - "frontend_name" is the name of the frontend (or listener) which received
19721 and processed the connection.
19722
19723 - "backend_name" is the name of the backend (or listener) which was selected
19724 to manage the connection to the server. This will be the same as the
19725 frontend if no switching rule has been applied.
19726
19727 - "server_name" is the name of the last server to which the connection was
19728 sent, which might differ from the first one if there were connection errors
19729 and a redispatch occurred. Note that this server belongs to the backend
19730 which processed the request. If the request was aborted before reaching a
19731 server, "<NOSRV>" is indicated instead of a server name. If the request was
19732 intercepted by the stats subsystem, "<STATS>" is indicated instead.
19733
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020019734 - "TR" is the total time in milliseconds spent waiting for a full HTTP
19735 request from the client (not counting body) after the first byte was
19736 received. It can be "-1" if the connection was aborted before a complete
John Roeslerfb2fce12019-07-10 15:45:51 -050019737 request could be received or a bad request was received. It should
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020019738 always be very small because a request generally fits in one single packet.
19739 Large times here generally indicate network issues between the client and
Willy Tarreau590a0512018-09-05 11:56:48 +020019740 haproxy or requests being typed by hand. See section 8.4 "Timing Events"
19741 for more details.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010019742
19743 - "Tw" is the total time in milliseconds spent waiting in the various queues.
19744 It can be "-1" if the connection was aborted before reaching the queue.
Willy Tarreau590a0512018-09-05 11:56:48 +020019745 See section 8.4 "Timing Events" for more details.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010019746
19747 - "Tc" is the total time in milliseconds spent waiting for the connection to
19748 establish to the final server, including retries. It can be "-1" if the
Willy Tarreau590a0512018-09-05 11:56:48 +020019749 request was aborted before a connection could be established. See section
19750 8.4 "Timing Events" for more details.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010019751
19752 - "Tr" is the total time in milliseconds spent waiting for the server to send
19753 a full HTTP response, not counting data. It can be "-1" if the request was
19754 aborted before a complete response could be received. It generally matches
19755 the server's processing time for the request, though it may be altered by
19756 the amount of data sent by the client to the server. Large times here on
Willy Tarreau590a0512018-09-05 11:56:48 +020019757 "GET" requests generally indicate an overloaded server. See section 8.4
19758 "Timing Events" for more details.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010019759
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020019760 - "Ta" is the time the request remained active in haproxy, which is the total
19761 time in milliseconds elapsed between the first byte of the request was
19762 received and the last byte of response was sent. It covers all possible
19763 processing except the handshake (see Th) and idle time (see Ti). There is
19764 one exception, if "option logasap" was specified, then the time counting
19765 stops at the moment the log is emitted. In this case, a '+' sign is
19766 prepended before the value, indicating that the final one will be larger.
Willy Tarreau590a0512018-09-05 11:56:48 +020019767 See section 8.4 "Timing Events" for more details.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010019768
19769 - "status_code" is the HTTP status code returned to the client. This status
19770 is generally set by the server, but it might also be set by haproxy when
19771 the server cannot be reached or when its response is blocked by haproxy.
19772
19773 - "bytes_read" is the total number of bytes transmitted to the client when
19774 the log is emitted. This does include HTTP headers. If "option logasap" is
John Roeslerfb2fce12019-07-10 15:45:51 -050019775 specified, this value will be prefixed with a '+' sign indicating that
Willy Tarreaucc6c8912009-02-22 10:53:55 +010019776 the final one may be larger. Please note that this value is a 64-bit
19777 counter, so log analysis tools must be able to handle it without
19778 overflowing.
19779
19780 - "captured_request_cookie" is an optional "name=value" entry indicating that
19781 the client had this cookie in the request. The cookie name and its maximum
19782 length are defined by the "capture cookie" statement in the frontend
19783 configuration. The field is a single dash ('-') when the option is not
19784 set. Only one cookie may be captured, it is generally used to track session
19785 ID exchanges between a client and a server to detect session crossing
19786 between clients due to application bugs. For more details, please consult
19787 the section "Capturing HTTP headers and cookies" below.
19788
19789 - "captured_response_cookie" is an optional "name=value" entry indicating
19790 that the server has returned a cookie with its response. The cookie name
19791 and its maximum length are defined by the "capture cookie" statement in the
19792 frontend configuration. The field is a single dash ('-') when the option is
19793 not set. Only one cookie may be captured, it is generally used to track
19794 session ID exchanges between a client and a server to detect session
19795 crossing between clients due to application bugs. For more details, please
19796 consult the section "Capturing HTTP headers and cookies" below.
19797
19798 - "termination_state" is the condition the session was in when the session
19799 ended. This indicates the session state, which side caused the end of
19800 session to happen, for what reason (timeout, error, ...), just like in TCP
19801 logs, and information about persistence operations on cookies in the last
19802 two characters. The normal flags should begin with "--", indicating the
19803 session was closed by either end with no data remaining in buffers. See
19804 below "Session state at disconnection" for more details.
19805
19806 - "actconn" is the total number of concurrent connections on the process when
Jamie Gloudonaaa21002012-08-25 00:18:33 -040019807 the session was logged. It is useful to detect when some per-process system
Willy Tarreaucc6c8912009-02-22 10:53:55 +010019808 limits have been reached. For instance, if actconn is close to 512 or 1024
19809 when multiple connection errors occur, chances are high that the system
19810 limits the process to use a maximum of 1024 file descriptors and that all
Willy Tarreauc57f0e22009-05-10 13:12:33 +020019811 of them are used. See section 3 "Global parameters" to find how to tune the
Willy Tarreaucc6c8912009-02-22 10:53:55 +010019812 system.
19813
19814 - "feconn" is the total number of concurrent connections on the frontend when
19815 the session was logged. It is useful to estimate the amount of resource
19816 required to sustain high loads, and to detect when the frontend's "maxconn"
19817 has been reached. Most often when this value increases by huge jumps, it is
19818 because there is congestion on the backend servers, but sometimes it can be
19819 caused by a denial of service attack.
19820
19821 - "beconn" is the total number of concurrent connections handled by the
19822 backend when the session was logged. It includes the total number of
19823 concurrent connections active on servers as well as the number of
19824 connections pending in queues. It is useful to estimate the amount of
19825 additional servers needed to support high loads for a given application.
19826 Most often when this value increases by huge jumps, it is because there is
19827 congestion on the backend servers, but sometimes it can be caused by a
19828 denial of service attack.
19829
19830 - "srv_conn" is the total number of concurrent connections still active on
19831 the server when the session was logged. It can never exceed the server's
19832 configured "maxconn" parameter. If this value is very often close or equal
19833 to the server's "maxconn", it means that traffic regulation is involved a
19834 lot, meaning that either the server's maxconn value is too low, or that
19835 there aren't enough servers to process the load with an optimal response
19836 time. When only one of the server's "srv_conn" is high, it usually means
19837 that this server has some trouble causing the requests to take longer to be
19838 processed than on other servers.
19839
19840 - "retries" is the number of connection retries experienced by this session
19841 when trying to connect to the server. It must normally be zero, unless a
19842 server is being stopped at the same moment the connection was attempted.
19843 Frequent retries generally indicate either a network problem between
19844 haproxy and the server, or a misconfigured system backlog on the server
19845 preventing new connections from being queued. This field may optionally be
19846 prefixed with a '+' sign, indicating that the session has experienced a
19847 redispatch after the maximal retry count has been reached on the initial
19848 server. In this case, the server name appearing in the log is the one the
19849 connection was redispatched to, and not the first one, though both may
19850 sometimes be the same in case of hashing for instance. So as a general rule
19851 of thumb, when a '+' is present in front of the retry count, this count
19852 should not be attributed to the logged server.
19853
19854 - "srv_queue" is the total number of requests which were processed before
19855 this one in the server queue. It is zero when the request has not gone
19856 through the server queue. It makes it possible to estimate the approximate
19857 server's response time by dividing the time spent in queue by the number of
19858 requests in the queue. It is worth noting that if a session experiences a
19859 redispatch and passes through two server queues, their positions will be
Davor Ocelice9ed2812017-12-25 17:49:28 +010019860 cumulative. A request should not pass through both the server queue and the
Willy Tarreaucc6c8912009-02-22 10:53:55 +010019861 backend queue unless a redispatch occurs.
19862
19863 - "backend_queue" is the total number of requests which were processed before
19864 this one in the backend's global queue. It is zero when the request has not
19865 gone through the global queue. It makes it possible to estimate the average
19866 queue length, which easily translates into a number of missing servers when
19867 divided by a server's "maxconn" parameter. It is worth noting that if a
19868 session experiences a redispatch, it may pass twice in the backend's queue,
Davor Ocelice9ed2812017-12-25 17:49:28 +010019869 and then both positions will be cumulative. A request should not pass
Willy Tarreaucc6c8912009-02-22 10:53:55 +010019870 through both the server queue and the backend queue unless a redispatch
19871 occurs.
19872
19873 - "captured_request_headers" is a list of headers captured in the request due
19874 to the presence of the "capture request header" statement in the frontend.
19875 Multiple headers can be captured, they will be delimited by a vertical bar
19876 ('|'). When no capture is enabled, the braces do not appear, causing a
19877 shift of remaining fields. It is important to note that this field may
19878 contain spaces, and that using it requires a smarter log parser than when
19879 it's not used. Please consult the section "Capturing HTTP headers and
19880 cookies" below for more details.
19881
19882 - "captured_response_headers" is a list of headers captured in the response
19883 due to the presence of the "capture response header" statement in the
19884 frontend. Multiple headers can be captured, they will be delimited by a
19885 vertical bar ('|'). When no capture is enabled, the braces do not appear,
19886 causing a shift of remaining fields. It is important to note that this
19887 field may contain spaces, and that using it requires a smarter log parser
19888 than when it's not used. Please consult the section "Capturing HTTP headers
19889 and cookies" below for more details.
19890
19891 - "http_request" is the complete HTTP request line, including the method,
19892 request and HTTP version string. Non-printable characters are encoded (see
19893 below the section "Non-printable characters"). This is always the last
19894 field, and it is always delimited by quotes and is the only one which can
19895 contain quotes. If new fields are added to the log format, they will be
19896 added before this field. This field might be truncated if the request is
19897 huge and does not fit in the standard syslog buffer (1024 characters). This
19898 is the reason why this field must always remain the last one.
19899
19900
Cyril Bontédc4d9032012-04-08 21:57:39 +0200199018.2.4. Custom log format
19902------------------------
William Lallemand48940402012-01-30 16:47:22 +010019903
Willy Tarreau2beef582012-12-20 17:22:52 +010019904The directive log-format allows you to customize the logs in http mode and tcp
William Lallemandbddd4fd2012-02-27 11:23:10 +010019905mode. It takes a string as argument.
William Lallemand48940402012-01-30 16:47:22 +010019906
Davor Ocelice9ed2812017-12-25 17:49:28 +010019907HAProxy understands some log format variables. % precedes log format variables.
William Lallemand48940402012-01-30 16:47:22 +010019908Variables can take arguments using braces ('{}'), and multiple arguments are
19909separated by commas within the braces. Flags may be added or removed by
19910prefixing them with a '+' or '-' sign.
19911
19912Special variable "%o" may be used to propagate its flags to all other
19913variables on the same format string. This is particularly handy with quoted
Dragan Dosen835b9212016-02-12 13:23:03 +010019914("Q") and escaped ("E") string formats.
William Lallemand48940402012-01-30 16:47:22 +010019915
Willy Tarreauc8368452012-12-21 00:09:23 +010019916If a variable is named between square brackets ('[' .. ']') then it is used
Willy Tarreaube722a22014-06-13 16:31:59 +020019917as a sample expression rule (see section 7.3). This it useful to add some
Willy Tarreauc8368452012-12-21 00:09:23 +010019918less common information such as the client's SSL certificate's DN, or to log
19919the key that would be used to store an entry into a stick table.
19920
Dragan Dosen1e3b16f2020-06-23 18:16:44 +020019921Note: spaces must be escaped. In configuration directives "log-format",
19922"log-format-sd" and "unique-id-format", spaces are considered as
19923delimiters and are merged. In order to emit a verbatim '%', it must be
19924preceded by another '%' resulting in '%%'.
William Lallemand48940402012-01-30 16:47:22 +010019925
Dragan Dosen835b9212016-02-12 13:23:03 +010019926Note: when using the RFC5424 syslog message format, the characters '"',
19927'\' and ']' inside PARAM-VALUE should be escaped with '\' as prefix (see
19928https://tools.ietf.org/html/rfc5424#section-6.3.3 for more details). In
19929such cases, the use of the flag "E" should be considered.
19930
William Lallemand48940402012-01-30 16:47:22 +010019931Flags are :
19932 * Q: quote a string
Jamie Gloudonaaa21002012-08-25 00:18:33 -040019933 * X: hexadecimal representation (IPs, Ports, %Ts, %rt, %pid)
Dragan Dosen835b9212016-02-12 13:23:03 +010019934 * E: escape characters '"', '\' and ']' in a string with '\' as prefix
19935 (intended purpose is for the RFC5424 structured-data log formats)
William Lallemand48940402012-01-30 16:47:22 +010019936
19937 Example:
19938
19939 log-format %T\ %t\ Some\ Text
19940 log-format %{+Q}o\ %t\ %s\ %{-Q}r
19941
Dragan Dosen835b9212016-02-12 13:23:03 +010019942 log-format-sd %{+Q,+E}o\ [exampleSDID@1234\ header=%[capture.req.hdr(0)]]
19943
William Lallemand48940402012-01-30 16:47:22 +010019944At the moment, the default HTTP format is defined this way :
19945
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020019946 log-format "%ci:%cp [%tr] %ft %b/%s %TR/%Tw/%Tc/%Tr/%Ta %ST %B %CC \
19947 %CS %tsc %ac/%fc/%bc/%sc/%rc %sq/%bq %hr %hs %{+Q}r"
William Lallemand48940402012-01-30 16:47:22 +010019948
William Lallemandbddd4fd2012-02-27 11:23:10 +010019949the default CLF format is defined this way :
William Lallemand48940402012-01-30 16:47:22 +010019950
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020019951 log-format "%{+Q}o %{-Q}ci - - [%trg] %r %ST %B \"\" \"\" %cp \
19952 %ms %ft %b %s %TR %Tw %Tc %Tr %Ta %tsc %ac %fc \
19953 %bc %sc %rc %sq %bq %CC %CS %hrl %hsl"
William Lallemand48940402012-01-30 16:47:22 +010019954
William Lallemandbddd4fd2012-02-27 11:23:10 +010019955and the default TCP format is defined this way :
19956
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020019957 log-format "%ci:%cp [%t] %ft %b/%s %Tw/%Tc/%Tt %B %ts \
19958 %ac/%fc/%bc/%sc/%rc %sq/%bq"
William Lallemandbddd4fd2012-02-27 11:23:10 +010019959
William Lallemand48940402012-01-30 16:47:22 +010019960Please refer to the table below for currently defined variables :
19961
William Lallemandbddd4fd2012-02-27 11:23:10 +010019962 +---+------+-----------------------------------------------+-------------+
Willy Tarreauffc3fcd2012-10-12 20:17:54 +020019963 | R | var | field name (8.2.2 and 8.2.3 for description) | type |
William Lallemandbddd4fd2012-02-27 11:23:10 +010019964 +---+------+-----------------------------------------------+-------------+
19965 | | %o | special variable, apply flags on all next var | |
19966 +---+------+-----------------------------------------------+-------------+
Willy Tarreau2beef582012-12-20 17:22:52 +010019967 | | %B | bytes_read (from server to client) | numeric |
19968 | H | %CC | captured_request_cookie | string |
19969 | H | %CS | captured_response_cookie | string |
William Lallemand5f232402012-04-05 18:02:55 +020019970 | | %H | hostname | string |
Andrew Hayworth0ebc55f2015-04-27 21:37:03 +000019971 | H | %HM | HTTP method (ex: POST) | string |
Maciej Zdeb21acc332020-11-26 10:45:52 +000019972 | H | %HP | HTTP request URI without query string | string |
Maciej Zdebfcdfd852020-11-30 18:27:47 +000019973 | H | %HPO | HTTP path only (without host nor query string)| string |
Andrew Hayworthe63ac872015-07-31 16:14:16 +000019974 | H | %HQ | HTTP request URI query string (ex: ?bar=baz) | string |
Andrew Hayworth0ebc55f2015-04-27 21:37:03 +000019975 | H | %HU | HTTP request URI (ex: /foo?bar=baz) | string |
19976 | H | %HV | HTTP version (ex: HTTP/1.0) | string |
William Lallemanda73203e2012-03-12 12:48:57 +010019977 | | %ID | unique-id | string |
Willy Tarreau4bf99632014-06-13 12:21:40 +020019978 | | %ST | status_code | numeric |
William Lallemand5f232402012-04-05 18:02:55 +020019979 | | %T | gmt_date_time | date |
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020019980 | | %Ta | Active time of the request (from TR to end) | numeric |
William Lallemandbddd4fd2012-02-27 11:23:10 +010019981 | | %Tc | Tc | numeric |
Willy Tarreau27b639d2016-05-17 17:55:27 +020019982 | | %Td | Td = Tt - (Tq + Tw + Tc + Tr) | numeric |
Yuxans Yao4e25b012012-10-19 10:36:09 +080019983 | | %Tl | local_date_time | date |
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020019984 | | %Th | connection handshake time (SSL, PROXY proto) | numeric |
19985 | H | %Ti | idle time before the HTTP request | numeric |
19986 | H | %Tq | Th + Ti + TR | numeric |
19987 | H | %TR | time to receive the full request from 1st byte| numeric |
19988 | H | %Tr | Tr (response time) | numeric |
William Lallemand5f232402012-04-05 18:02:55 +020019989 | | %Ts | timestamp | numeric |
William Lallemandbddd4fd2012-02-27 11:23:10 +010019990 | | %Tt | Tt | numeric |
Damien Claisse57c8eb92020-04-28 12:09:19 +000019991 | | %Tu | Tu | numeric |
William Lallemandbddd4fd2012-02-27 11:23:10 +010019992 | | %Tw | Tw | numeric |
Willy Tarreau2beef582012-12-20 17:22:52 +010019993 | | %U | bytes_uploaded (from client to server) | numeric |
William Lallemandbddd4fd2012-02-27 11:23:10 +010019994 | | %ac | actconn | numeric |
19995 | | %b | backend_name | string |
Willy Tarreau2beef582012-12-20 17:22:52 +010019996 | | %bc | beconn (backend concurrent connections) | numeric |
19997 | | %bi | backend_source_ip (connecting address) | IP |
19998 | | %bp | backend_source_port (connecting address) | numeric |
William Lallemandbddd4fd2012-02-27 11:23:10 +010019999 | | %bq | backend_queue | numeric |
Willy Tarreau2beef582012-12-20 17:22:52 +010020000 | | %ci | client_ip (accepted address) | IP |
20001 | | %cp | client_port (accepted address) | numeric |
William Lallemandbddd4fd2012-02-27 11:23:10 +010020002 | | %f | frontend_name | string |
Willy Tarreau2beef582012-12-20 17:22:52 +010020003 | | %fc | feconn (frontend concurrent connections) | numeric |
20004 | | %fi | frontend_ip (accepting address) | IP |
20005 | | %fp | frontend_port (accepting address) | numeric |
Willy Tarreau773d65f2012-10-12 14:56:11 +020020006 | | %ft | frontend_name_transport ('~' suffix for SSL) | string |
Willy Tarreau7346acb2014-08-28 15:03:15 +020020007 | | %lc | frontend_log_counter | numeric |
Willy Tarreaud9ed3d22014-06-13 12:23:06 +020020008 | | %hr | captured_request_headers default style | string |
20009 | | %hrl | captured_request_headers CLF style | string list |
20010 | | %hs | captured_response_headers default style | string |
20011 | | %hsl | captured_response_headers CLF style | string list |
Willy Tarreau812c88e2015-08-09 10:56:35 +020020012 | | %ms | accept date milliseconds (left-padded with 0) | numeric |
William Lallemand5f232402012-04-05 18:02:55 +020020013 | | %pid | PID | numeric |
Willy Tarreauffc3fcd2012-10-12 20:17:54 +020020014 | H | %r | http_request | string |
William Lallemandbddd4fd2012-02-27 11:23:10 +010020015 | | %rc | retries | numeric |
Willy Tarreau1f0da242014-01-25 11:01:50 +010020016 | | %rt | request_counter (HTTP req or TCP session) | numeric |
William Lallemandbddd4fd2012-02-27 11:23:10 +010020017 | | %s | server_name | string |
Willy Tarreau2beef582012-12-20 17:22:52 +010020018 | | %sc | srv_conn (server concurrent connections) | numeric |
20019 | | %si | server_IP (target address) | IP |
20020 | | %sp | server_port (target address) | numeric |
William Lallemandbddd4fd2012-02-27 11:23:10 +010020021 | | %sq | srv_queue | numeric |
Willy Tarreauffc3fcd2012-10-12 20:17:54 +020020022 | S | %sslc| ssl_ciphers (ex: AES-SHA) | string |
20023 | S | %sslv| ssl_version (ex: TLSv1) | string |
Willy Tarreau2beef582012-12-20 17:22:52 +010020024 | | %t | date_time (with millisecond resolution) | date |
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020020025 | H | %tr | date_time of HTTP request | date |
20026 | H | %trg | gmt_date_time of start of HTTP request | date |
Jens Bissinger15c64ff2018-08-23 14:11:27 +020020027 | H | %trl | local_date_time of start of HTTP request | date |
William Lallemandbddd4fd2012-02-27 11:23:10 +010020028 | | %ts | termination_state | string |
Willy Tarreauffc3fcd2012-10-12 20:17:54 +020020029 | H | %tsc | termination_state with cookie status | string |
William Lallemandbddd4fd2012-02-27 11:23:10 +010020030 +---+------+-----------------------------------------------+-------------+
William Lallemand48940402012-01-30 16:47:22 +010020031
Willy Tarreauffc3fcd2012-10-12 20:17:54 +020020032 R = Restrictions : H = mode http only ; S = SSL only
William Lallemand48940402012-01-30 16:47:22 +010020033
Willy Tarreau5f51e1a2012-12-03 18:40:10 +010020034
200358.2.5. Error log format
20036-----------------------
20037
20038When an incoming connection fails due to an SSL handshake or an invalid PROXY
20039protocol header, haproxy will log the event using a shorter, fixed line format.
20040By default, logs are emitted at the LOG_INFO level, unless the option
20041"log-separate-errors" is set in the backend, in which case the LOG_ERR level
Davor Ocelice9ed2812017-12-25 17:49:28 +010020042will be used. Connections on which no data are exchanged (e.g. probes) are not
Willy Tarreau5f51e1a2012-12-03 18:40:10 +010020043logged if the "dontlognull" option is set.
20044
20045The format looks like this :
20046
20047 >>> Dec 3 18:27:14 localhost \
20048 haproxy[6103]: 127.0.0.1:56059 [03/Dec/2012:17:35:10.380] frt/f1: \
20049 Connection error during SSL handshake
20050
20051 Field Format Extract from the example above
20052 1 process_name '[' pid ']:' haproxy[6103]:
20053 2 client_ip ':' client_port 127.0.0.1:56059
20054 3 '[' accept_date ']' [03/Dec/2012:17:35:10.380]
20055 4 frontend_name "/" bind_name ":" frt/f1:
20056 5 message Connection error during SSL handshake
20057
20058These fields just provide minimal information to help debugging connection
20059failures.
20060
20061
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200200628.3. Advanced logging options
20063-----------------------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020064
20065Some advanced logging options are often looked for but are not easy to find out
20066just by looking at the various options. Here is an entry point for the few
20067options which can enable better logging. Please refer to the keywords reference
20068for more information about their usage.
20069
20070
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200200718.3.1. Disabling logging of external tests
20072------------------------------------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020073
20074It is quite common to have some monitoring tools perform health checks on
20075haproxy. Sometimes it will be a layer 3 load-balancer such as LVS or any
20076commercial load-balancer, and sometimes it will simply be a more complete
20077monitoring system such as Nagios. When the tests are very frequent, users often
20078ask how to disable logging for those checks. There are three possibilities :
20079
20080 - if connections come from everywhere and are just TCP probes, it is often
20081 desired to simply disable logging of connections without data exchange, by
20082 setting "option dontlognull" in the frontend. It also disables logging of
20083 port scans, which may or may not be desired.
20084
Willy Tarreau9e9919d2020-10-14 15:55:23 +020020085 - it is possible to use the "http-request set-log-level silent" action using
20086 a variety of conditions (source networks, paths, user-agents, etc).
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020087
20088 - if the tests are performed on a known URI, use "monitor-uri" to declare
20089 this URI as dedicated to monitoring. Any host sending this request will
20090 only get the result of a health-check, and the request will not be logged.
20091
20092
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200200938.3.2. Logging before waiting for the session to terminate
20094----------------------------------------------------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020095
20096The problem with logging at end of connection is that you have no clue about
20097what is happening during very long sessions, such as remote terminal sessions
20098or large file downloads. This problem can be worked around by specifying
Davor Ocelice9ed2812017-12-25 17:49:28 +010020099"option logasap" in the frontend. HAProxy will then log as soon as possible,
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020100just before data transfer begins. This means that in case of TCP, it will still
20101log the connection status to the server, and in case of HTTP, it will log just
20102after processing the server headers. In this case, the number of bytes reported
20103is the number of header bytes sent to the client. In order to avoid confusion
20104with normal logs, the total time field and the number of bytes are prefixed
20105with a '+' sign which means that real numbers are certainly larger.
20106
20107
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200201088.3.3. Raising log level upon errors
20109------------------------------------
Willy Tarreauc9bd0cc2009-05-10 11:57:02 +020020110
20111Sometimes it is more convenient to separate normal traffic from errors logs,
20112for instance in order to ease error monitoring from log files. When the option
20113"log-separate-errors" is used, connections which experience errors, timeouts,
20114retries, redispatches or HTTP status codes 5xx will see their syslog level
20115raised from "info" to "err". This will help a syslog daemon store the log in
20116a separate file. It is very important to keep the errors in the normal traffic
20117file too, so that log ordering is not altered. You should also be careful if
20118you already have configured your syslog daemon to store all logs higher than
20119"notice" in an "admin" file, because the "err" level is higher than "notice".
20120
20121
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200201228.3.4. Disabling logging of successful connections
20123--------------------------------------------------
Willy Tarreauc9bd0cc2009-05-10 11:57:02 +020020124
20125Although this may sound strange at first, some large sites have to deal with
20126multiple thousands of logs per second and are experiencing difficulties keeping
20127them intact for a long time or detecting errors within them. If the option
20128"dontlog-normal" is set on the frontend, all normal connections will not be
20129logged. In this regard, a normal connection is defined as one without any
20130error, timeout, retry nor redispatch. In HTTP, the status code is checked too,
20131and a response with a status 5xx is not considered normal and will be logged
20132too. Of course, doing is is really discouraged as it will remove most of the
20133useful information from the logs. Do this only if you have no other
20134alternative.
20135
20136
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200201378.4. Timing events
20138------------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020139
20140Timers provide a great help in troubleshooting network problems. All values are
20141reported in milliseconds (ms). These timers should be used in conjunction with
20142the session termination flags. In TCP mode with "option tcplog" set on the
20143frontend, 3 control points are reported under the form "Tw/Tc/Tt", and in HTTP
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020020144mode, 5 control points are reported under the form "TR/Tw/Tc/Tr/Ta". In
20145addition, three other measures are provided, "Th", "Ti", and "Tq".
20146
Guillaume de Lafondf27cddc2016-12-23 17:32:43 +010020147Timings events in HTTP mode:
20148
20149 first request 2nd request
20150 |<-------------------------------->|<-------------- ...
20151 t tr t tr ...
20152 ---|----|----|----|----|----|----|----|----|--
20153 : Th Ti TR Tw Tc Tr Td : Ti ...
20154 :<---- Tq ---->: :
20155 :<-------------- Tt -------------->:
Damien Claisse57c8eb92020-04-28 12:09:19 +000020156 :<-- -----Tu--------------->:
Guillaume de Lafondf27cddc2016-12-23 17:32:43 +010020157 :<--------- Ta --------->:
20158
20159Timings events in TCP mode:
20160
20161 TCP session
20162 |<----------------->|
20163 t t
20164 ---|----|----|----|----|---
20165 | Th Tw Tc Td |
20166 |<------ Tt ------->|
20167
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020020168 - Th: total time to accept tcp connection and execute handshakes for low level
Davor Ocelice9ed2812017-12-25 17:49:28 +010020169 protocols. Currently, these protocols are proxy-protocol and SSL. This may
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020020170 only happen once during the whole connection's lifetime. A large time here
20171 may indicate that the client only pre-established the connection without
20172 speaking, that it is experiencing network issues preventing it from
Davor Ocelice9ed2812017-12-25 17:49:28 +010020173 completing a handshake in a reasonable time (e.g. MTU issues), or that an
Willy Tarreau590a0512018-09-05 11:56:48 +020020174 SSL handshake was very expensive to compute. Please note that this time is
20175 reported only before the first request, so it is safe to average it over
20176 all request to calculate the amortized value. The second and subsequent
20177 request will always report zero here.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020178
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020020179 - Ti: is the idle time before the HTTP request (HTTP mode only). This timer
20180 counts between the end of the handshakes and the first byte of the HTTP
20181 request. When dealing with a second request in keep-alive mode, it starts
Willy Tarreau590a0512018-09-05 11:56:48 +020020182 to count after the end of the transmission the previous response. When a
20183 multiplexed protocol such as HTTP/2 is used, it starts to count immediately
20184 after the previous request. Some browsers pre-establish connections to a
20185 server in order to reduce the latency of a future request, and keep them
20186 pending until they need it. This delay will be reported as the idle time. A
20187 value of -1 indicates that nothing was received on the connection.
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020020188
20189 - TR: total time to get the client request (HTTP mode only). It's the time
20190 elapsed between the first bytes received and the moment the proxy received
20191 the empty line marking the end of the HTTP headers. The value "-1"
20192 indicates that the end of headers has never been seen. This happens when
20193 the client closes prematurely or times out. This time is usually very short
20194 since most requests fit in a single packet. A large time may indicate a
20195 request typed by hand during a test.
20196
20197 - Tq: total time to get the client request from the accept date or since the
20198 emission of the last byte of the previous response (HTTP mode only). It's
Davor Ocelice9ed2812017-12-25 17:49:28 +010020199 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 +020020200 returns -1 as well. This timer used to be very useful before the arrival of
20201 HTTP keep-alive and browsers' pre-connect feature. It's recommended to drop
20202 it in favor of TR nowadays, as the idle time adds a lot of noise to the
20203 reports.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020204
20205 - Tw: total time spent in the queues waiting for a connection slot. It
20206 accounts for backend queue as well as the server queues, and depends on the
20207 queue size, and the time needed for the server to complete previous
20208 requests. The value "-1" means that the request was killed before reaching
20209 the queue, which is generally what happens with invalid or denied requests.
20210
20211 - Tc: total time to establish the TCP connection to the server. It's the time
20212 elapsed between the moment the proxy sent the connection request, and the
20213 moment it was acknowledged by the server, or between the TCP SYN packet and
20214 the matching SYN/ACK packet in return. The value "-1" means that the
20215 connection never established.
20216
20217 - Tr: server response time (HTTP mode only). It's the time elapsed between
20218 the moment the TCP connection was established to the server and the moment
20219 the server sent its complete response headers. It purely shows its request
20220 processing time, without the network overhead due to the data transmission.
20221 It is worth noting that when the client has data to send to the server, for
20222 instance during a POST request, the time already runs, and this can distort
20223 apparent response time. For this reason, it's generally wise not to trust
20224 too much this field for POST requests initiated from clients behind an
20225 untrusted network. A value of "-1" here means that the last the response
20226 header (empty line) was never seen, most likely because the server timeout
20227 stroke before the server managed to process the request.
20228
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020020229 - Ta: total active time for the HTTP request, between the moment the proxy
20230 received the first byte of the request header and the emission of the last
20231 byte of the response body. The exception is when the "logasap" option is
20232 specified. In this case, it only equals (TR+Tw+Tc+Tr), and is prefixed with
20233 a '+' sign. From this field, we can deduce "Td", the data transmission time,
20234 by subtracting other timers when valid :
20235
20236 Td = Ta - (TR + Tw + Tc + Tr)
20237
20238 Timers with "-1" values have to be excluded from this equation. Note that
20239 "Ta" can never be negative.
20240
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020241 - Tt: total session duration time, between the moment the proxy accepted it
20242 and the moment both ends were closed. The exception is when the "logasap"
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020020243 option is specified. In this case, it only equals (Th+Ti+TR+Tw+Tc+Tr), and
20244 is prefixed with a '+' sign. From this field, we can deduce "Td", the data
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030020245 transmission time, by subtracting other timers when valid :
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020246
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020020247 Td = Tt - (Th + Ti + TR + Tw + Tc + Tr)
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020248
20249 Timers with "-1" values have to be excluded from this equation. In TCP
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020020250 mode, "Ti", "Tq" and "Tr" have to be excluded too. Note that "Tt" can never
20251 be negative and that for HTTP, Tt is simply equal to (Th+Ti+Ta).
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020252
Damien Claisse57c8eb92020-04-28 12:09:19 +000020253 - Tu: total estimated time as seen from client, between the moment the proxy
20254 accepted it and the moment both ends were closed, without idle time.
20255 This is useful to roughly measure end-to-end time as a user would see it,
20256 without idle time pollution from keep-alive time between requests. This
20257 timer in only an estimation of time seen by user as it assumes network
20258 latency is the same in both directions. The exception is when the "logasap"
20259 option is specified. In this case, it only equals (Th+TR+Tw+Tc+Tr), and is
20260 prefixed with a '+' sign.
20261
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020262These timers provide precious indications on trouble causes. Since the TCP
20263protocol defines retransmit delays of 3, 6, 12... seconds, we know for sure
20264that timers close to multiples of 3s are nearly always related to lost packets
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020020265due to network problems (wires, negotiation, congestion). Moreover, if "Ta" or
20266"Tt" is close to a timeout value specified in the configuration, it often means
20267that a session has been aborted on timeout.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020268
20269Most common cases :
20270
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020020271 - If "Th" or "Ti" are close to 3000, a packet has probably been lost between
20272 the client and the proxy. This is very rare on local networks but might
20273 happen when clients are on far remote networks and send large requests. It
20274 may happen that values larger than usual appear here without any network
20275 cause. Sometimes, during an attack or just after a resource starvation has
20276 ended, haproxy may accept thousands of connections in a few milliseconds.
20277 The time spent accepting these connections will inevitably slightly delay
20278 processing of other connections, and it can happen that request times in the
20279 order of a few tens of milliseconds are measured after a few thousands of
20280 new connections have been accepted at once. Using one of the keep-alive
20281 modes may display larger idle times since "Ti" measures the time spent
Patrick Mezard105faca2010-06-12 17:02:46 +020020282 waiting for additional requests.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020283
20284 - If "Tc" is close to 3000, a packet has probably been lost between the
20285 server and the proxy during the server connection phase. This value should
20286 always be very low, such as 1 ms on local networks and less than a few tens
20287 of ms on remote networks.
20288
Willy Tarreau55165fe2009-05-10 12:02:55 +020020289 - If "Tr" is nearly always lower than 3000 except some rare values which seem
20290 to be the average majored by 3000, there are probably some packets lost
20291 between the proxy and the server.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020292
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020020293 - If "Ta" is large even for small byte counts, it generally is because
20294 neither the client nor the server decides to close the connection while
20295 haproxy is running in tunnel mode and both have agreed on a keep-alive
20296 connection mode. In order to solve this issue, it will be needed to specify
20297 one of the HTTP options to manipulate keep-alive or close options on either
20298 the frontend or the backend. Having the smallest possible 'Ta' or 'Tt' is
20299 important when connection regulation is used with the "maxconn" option on
20300 the servers, since no new connection will be sent to the server until
20301 another one is released.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020302
20303Other noticeable HTTP log cases ('xx' means any value to be ignored) :
20304
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020020305 TR/Tw/Tc/Tr/+Ta The "option logasap" is present on the frontend and the log
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020306 was emitted before the data phase. All the timers are valid
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020020307 except "Ta" which is shorter than reality.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020308
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020020309 -1/xx/xx/xx/Ta The client was not able to send a complete request in time
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020310 or it aborted too early. Check the session termination flags
20311 then "timeout http-request" and "timeout client" settings.
20312
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020020313 TR/-1/xx/xx/Ta It was not possible to process the request, maybe because
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020314 servers were out of order, because the request was invalid
20315 or forbidden by ACL rules. Check the session termination
20316 flags.
20317
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020020318 TR/Tw/-1/xx/Ta The connection could not establish on the server. Either it
20319 actively refused it or it timed out after Ta-(TR+Tw) ms.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020320 Check the session termination flags, then check the
20321 "timeout connect" setting. Note that the tarpit action might
20322 return similar-looking patterns, with "Tw" equal to the time
20323 the client connection was maintained open.
20324
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020020325 TR/Tw/Tc/-1/Ta The server has accepted the connection but did not return
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030020326 a complete response in time, or it closed its connection
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020020327 unexpectedly after Ta-(TR+Tw+Tc) ms. Check the session
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020328 termination flags, then check the "timeout server" setting.
20329
20330
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200203318.5. Session state at disconnection
20332-----------------------------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020333
20334TCP and HTTP logs provide a session termination indicator in the
20335"termination_state" field, just before the number of active connections. It is
203362-characters long in TCP mode, and is extended to 4 characters in HTTP mode,
20337each of which has a special meaning :
20338
20339 - On the first character, a code reporting the first event which caused the
20340 session to terminate :
20341
20342 C : the TCP session was unexpectedly aborted by the client.
20343
20344 S : the TCP session was unexpectedly aborted by the server, or the
20345 server explicitly refused it.
20346
20347 P : the session was prematurely aborted by the proxy, because of a
20348 connection limit enforcement, because a DENY filter was matched,
20349 because of a security check which detected and blocked a dangerous
20350 error in server response which might have caused information leak
Davor Ocelice9ed2812017-12-25 17:49:28 +010020351 (e.g. cacheable cookie).
Willy Tarreau570f2212013-06-10 16:42:09 +020020352
20353 L : the session was locally processed by haproxy and was not passed to
20354 a server. This is what happens for stats and redirects.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020355
20356 R : a resource on the proxy has been exhausted (memory, sockets, source
20357 ports, ...). Usually, this appears during the connection phase, and
20358 system logs should contain a copy of the precise error. If this
20359 happens, it must be considered as a very serious anomaly which
20360 should be fixed as soon as possible by any means.
20361
20362 I : an internal error was identified by the proxy during a self-check.
20363 This should NEVER happen, and you are encouraged to report any log
20364 containing this, because this would almost certainly be a bug. It
20365 would be wise to preventively restart the process after such an
20366 event too, in case it would be caused by memory corruption.
20367
Simon Horman752dc4a2011-06-21 14:34:59 +090020368 D : the session was killed by haproxy because the server was detected
20369 as down and was configured to kill all connections when going down.
20370
Justin Karnegeseb2c24a2012-05-24 15:28:52 -070020371 U : the session was killed by haproxy on this backup server because an
20372 active server was detected as up and was configured to kill all
20373 backup connections when going up.
20374
Willy Tarreaua2a64e92011-09-07 23:01:56 +020020375 K : the session was actively killed by an admin operating on haproxy.
20376
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020377 c : the client-side timeout expired while waiting for the client to
20378 send or receive data.
20379
20380 s : the server-side timeout expired while waiting for the server to
20381 send or receive data.
20382
20383 - : normal session completion, both the client and the server closed
20384 with nothing left in the buffers.
20385
20386 - on the second character, the TCP or HTTP session state when it was closed :
20387
Willy Tarreauf7b30a92010-12-06 22:59:17 +010020388 R : the proxy was waiting for a complete, valid REQUEST from the client
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020389 (HTTP mode only). Nothing was sent to any server.
20390
20391 Q : the proxy was waiting in the QUEUE for a connection slot. This can
20392 only happen when servers have a 'maxconn' parameter set. It can
20393 also happen in the global queue after a redispatch consecutive to
20394 a failed attempt to connect to a dying server. If no redispatch is
20395 reported, then no connection attempt was made to any server.
20396
20397 C : the proxy was waiting for the CONNECTION to establish on the
20398 server. The server might at most have noticed a connection attempt.
20399
20400 H : the proxy was waiting for complete, valid response HEADERS from the
20401 server (HTTP only).
20402
20403 D : the session was in the DATA phase.
20404
20405 L : the proxy was still transmitting LAST data to the client while the
20406 server had already finished. This one is very rare as it can only
20407 happen when the client dies while receiving the last packets.
20408
20409 T : the request was tarpitted. It has been held open with the client
20410 during the whole "timeout tarpit" duration or until the client
20411 closed, both of which will be reported in the "Tw" timer.
20412
20413 - : normal session completion after end of data transfer.
20414
20415 - the third character tells whether the persistence cookie was provided by
20416 the client (only in HTTP mode) :
20417
20418 N : the client provided NO cookie. This is usually the case for new
20419 visitors, so counting the number of occurrences of this flag in the
20420 logs generally indicate a valid trend for the site frequentation.
20421
20422 I : the client provided an INVALID cookie matching no known server.
20423 This might be caused by a recent configuration change, mixed
Cyril Bontéa8e7bbc2010-04-25 22:29:29 +020020424 cookies between HTTP/HTTPS sites, persistence conditionally
20425 ignored, or an attack.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020426
20427 D : the client provided a cookie designating a server which was DOWN,
20428 so either "option persist" was used and the client was sent to
20429 this server, or it was not set and the client was redispatched to
20430 another server.
20431
Willy Tarreau996a92c2010-10-13 19:30:47 +020020432 V : the client provided a VALID cookie, and was sent to the associated
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020433 server.
20434
Willy Tarreau996a92c2010-10-13 19:30:47 +020020435 E : the client provided a valid cookie, but with a last date which was
20436 older than what is allowed by the "maxidle" cookie parameter, so
20437 the cookie is consider EXPIRED and is ignored. The request will be
20438 redispatched just as if there was no cookie.
20439
20440 O : the client provided a valid cookie, but with a first date which was
20441 older than what is allowed by the "maxlife" cookie parameter, so
20442 the cookie is consider too OLD and is ignored. The request will be
20443 redispatched just as if there was no cookie.
20444
Willy Tarreauc89ccb62012-04-05 21:18:22 +020020445 U : a cookie was present but was not used to select the server because
20446 some other server selection mechanism was used instead (typically a
20447 "use-server" rule).
20448
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020449 - : does not apply (no cookie set in configuration).
20450
20451 - the last character reports what operations were performed on the persistence
20452 cookie returned by the server (only in HTTP mode) :
20453
20454 N : NO cookie was provided by the server, and none was inserted either.
20455
20456 I : no cookie was provided by the server, and the proxy INSERTED one.
20457 Note that in "cookie insert" mode, if the server provides a cookie,
20458 it will still be overwritten and reported as "I" here.
20459
Willy Tarreau996a92c2010-10-13 19:30:47 +020020460 U : the proxy UPDATED the last date in the cookie that was presented by
20461 the client. This can only happen in insert mode with "maxidle". It
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030020462 happens every time there is activity at a different date than the
Willy Tarreau996a92c2010-10-13 19:30:47 +020020463 date indicated in the cookie. If any other change happens, such as
20464 a redispatch, then the cookie will be marked as inserted instead.
20465
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020466 P : a cookie was PROVIDED by the server and transmitted as-is.
20467
20468 R : the cookie provided by the server was REWRITTEN by the proxy, which
20469 happens in "cookie rewrite" or "cookie prefix" modes.
20470
20471 D : the cookie provided by the server was DELETED by the proxy.
20472
20473 - : does not apply (no cookie set in configuration).
20474
Willy Tarreau996a92c2010-10-13 19:30:47 +020020475The combination of the two first flags gives a lot of information about what
20476was happening when the session terminated, and why it did terminate. It can be
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020477helpful to detect server saturation, network troubles, local system resource
20478starvation, attacks, etc...
20479
20480The most common termination flags combinations are indicated below. They are
20481alphabetically sorted, with the lowercase set just after the upper case for
20482easier finding and understanding.
20483
20484 Flags Reason
20485
20486 -- Normal termination.
20487
20488 CC The client aborted before the connection could be established to the
20489 server. This can happen when haproxy tries to connect to a recently
20490 dead (or unchecked) server, and the client aborts while haproxy is
20491 waiting for the server to respond or for "timeout connect" to expire.
20492
20493 CD The client unexpectedly aborted during data transfer. This can be
20494 caused by a browser crash, by an intermediate equipment between the
20495 client and haproxy which decided to actively break the connection,
20496 by network routing issues between the client and haproxy, or by a
20497 keep-alive session between the server and the client terminated first
20498 by the client.
Willy Tarreaud72758d2010-01-12 10:42:19 +010020499
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020500 cD The client did not send nor acknowledge any data for as long as the
20501 "timeout client" delay. This is often caused by network failures on
Cyril Bontédc4d9032012-04-08 21:57:39 +020020502 the client side, or the client simply leaving the net uncleanly.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020503
20504 CH The client aborted while waiting for the server to start responding.
20505 It might be the server taking too long to respond or the client
20506 clicking the 'Stop' button too fast.
20507
20508 cH The "timeout client" stroke while waiting for client data during a
20509 POST request. This is sometimes caused by too large TCP MSS values
20510 for PPPoE networks which cannot transport full-sized packets. It can
20511 also happen when client timeout is smaller than server timeout and
20512 the server takes too long to respond.
20513
20514 CQ The client aborted while its session was queued, waiting for a server
20515 with enough empty slots to accept it. It might be that either all the
20516 servers were saturated or that the assigned server was taking too
20517 long a time to respond.
20518
20519 CR The client aborted before sending a full HTTP request. Most likely
20520 the request was typed by hand using a telnet client, and aborted
20521 too early. The HTTP status code is likely a 400 here. Sometimes this
20522 might also be caused by an IDS killing the connection between haproxy
Willy Tarreau0f228a02015-05-01 15:37:53 +020020523 and the client. "option http-ignore-probes" can be used to ignore
20524 connections without any data transfer.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020525
20526 cR The "timeout http-request" stroke before the client sent a full HTTP
20527 request. This is sometimes caused by too large TCP MSS values on the
20528 client side for PPPoE networks which cannot transport full-sized
20529 packets, or by clients sending requests by hand and not typing fast
20530 enough, or forgetting to enter the empty line at the end of the
Willy Tarreau2705a612014-05-23 17:38:34 +020020531 request. The HTTP status code is likely a 408 here. Note: recently,
Willy Tarreau0f228a02015-05-01 15:37:53 +020020532 some browsers started to implement a "pre-connect" feature consisting
20533 in speculatively connecting to some recently visited web sites just
20534 in case the user would like to visit them. This results in many
20535 connections being established to web sites, which end up in 408
20536 Request Timeout if the timeout strikes first, or 400 Bad Request when
20537 the browser decides to close them first. These ones pollute the log
20538 and feed the error counters. Some versions of some browsers have even
20539 been reported to display the error code. It is possible to work
Davor Ocelice9ed2812017-12-25 17:49:28 +010020540 around the undesirable effects of this behavior by adding "option
Willy Tarreau0f228a02015-05-01 15:37:53 +020020541 http-ignore-probes" in the frontend, resulting in connections with
20542 zero data transfer to be totally ignored. This will definitely hide
20543 the errors of people experiencing connectivity issues though.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020544
20545 CT The client aborted while its session was tarpitted. It is important to
20546 check if this happens on valid requests, in order to be sure that no
Willy Tarreau55165fe2009-05-10 12:02:55 +020020547 wrong tarpit rules have been written. If a lot of them happen, it
20548 might make sense to lower the "timeout tarpit" value to something
20549 closer to the average reported "Tw" timer, in order not to consume
20550 resources for just a few attackers.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020551
Willy Tarreau570f2212013-06-10 16:42:09 +020020552 LR The request was intercepted and locally handled by haproxy. Generally
20553 it means that this was a redirect or a stats request.
20554
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +010020555 SC The server or an equipment between it and haproxy explicitly refused
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020556 the TCP connection (the proxy received a TCP RST or an ICMP message
20557 in return). Under some circumstances, it can also be the network
Davor Ocelice9ed2812017-12-25 17:49:28 +010020558 stack telling the proxy that the server is unreachable (e.g. no route,
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020559 or no ARP response on local network). When this happens in HTTP mode,
20560 the status code is likely a 502 or 503 here.
20561
20562 sC The "timeout connect" stroke before a connection to the server could
20563 complete. When this happens in HTTP mode, the status code is likely a
20564 503 or 504 here.
20565
20566 SD The connection to the server died with an error during the data
20567 transfer. This usually means that haproxy has received an RST from
20568 the server or an ICMP message from an intermediate equipment while
20569 exchanging data with the server. This can be caused by a server crash
20570 or by a network issue on an intermediate equipment.
20571
20572 sD The server did not send nor acknowledge any data for as long as the
20573 "timeout server" setting during the data phase. This is often caused
Davor Ocelice9ed2812017-12-25 17:49:28 +010020574 by too short timeouts on L4 equipment before the server (firewalls,
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020575 load-balancers, ...), as well as keep-alive sessions maintained
20576 between the client and the server expiring first on haproxy.
20577
20578 SH The server aborted before sending its full HTTP response headers, or
20579 it crashed while processing the request. Since a server aborting at
20580 this moment is very rare, it would be wise to inspect its logs to
20581 control whether it crashed and why. The logged request may indicate a
20582 small set of faulty requests, demonstrating bugs in the application.
20583 Sometimes this might also be caused by an IDS killing the connection
20584 between haproxy and the server.
20585
20586 sH The "timeout server" stroke before the server could return its
20587 response headers. This is the most common anomaly, indicating too
20588 long transactions, probably caused by server or database saturation.
20589 The immediate workaround consists in increasing the "timeout server"
20590 setting, but it is important to keep in mind that the user experience
20591 will suffer from these long response times. The only long term
20592 solution is to fix the application.
20593
20594 sQ The session spent too much time in queue and has been expired. See
20595 the "timeout queue" and "timeout connect" settings to find out how to
20596 fix this if it happens too often. If it often happens massively in
20597 short periods, it may indicate general problems on the affected
20598 servers due to I/O or database congestion, or saturation caused by
20599 external attacks.
20600
20601 PC The proxy refused to establish a connection to the server because the
Thayne McCombscdbcca92021-01-07 21:24:41 -070020602 process's socket limit has been reached while attempting to connect.
Cyril Bontédc4d9032012-04-08 21:57:39 +020020603 The global "maxconn" parameter may be increased in the configuration
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020604 so that it does not happen anymore. This status is very rare and
20605 might happen when the global "ulimit-n" parameter is forced by hand.
20606
Willy Tarreaued2fd2d2010-12-29 11:23:27 +010020607 PD The proxy blocked an incorrectly formatted chunked encoded message in
20608 a request or a response, after the server has emitted its headers. In
20609 most cases, this will indicate an invalid message from the server to
Davor Ocelice9ed2812017-12-25 17:49:28 +010020610 the client. HAProxy supports chunk sizes of up to 2GB - 1 (2147483647
Willy Tarreauf3a3e132013-08-31 08:16:26 +020020611 bytes). Any larger size will be considered as an error.
Willy Tarreaued2fd2d2010-12-29 11:23:27 +010020612
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020613 PH The proxy blocked the server's response, because it was invalid,
20614 incomplete, dangerous (cache control), or matched a security filter.
20615 In any case, an HTTP 502 error is sent to the client. One possible
20616 cause for this error is an invalid syntax in an HTTP header name
Willy Tarreaued2fd2d2010-12-29 11:23:27 +010020617 containing unauthorized characters. It is also possible but quite
20618 rare, that the proxy blocked a chunked-encoding request from the
20619 client due to an invalid syntax, before the server responded. In this
20620 case, an HTTP 400 error is sent to the client and reported in the
20621 logs.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020622
20623 PR The proxy blocked the client's HTTP request, either because of an
20624 invalid HTTP syntax, in which case it returned an HTTP 400 error to
20625 the client, or because a deny filter matched, in which case it
20626 returned an HTTP 403 error.
20627
20628 PT The proxy blocked the client's request and has tarpitted its
20629 connection before returning it a 500 server error. Nothing was sent
20630 to the server. The connection was maintained open for as long as
20631 reported by the "Tw" timer field.
20632
20633 RC A local resource has been exhausted (memory, sockets, source ports)
20634 preventing the connection to the server from establishing. The error
20635 logs will tell precisely what was missing. This is very rare and can
20636 only be solved by proper system tuning.
20637
Willy Tarreau996a92c2010-10-13 19:30:47 +020020638The combination of the two last flags gives a lot of information about how
20639persistence was handled by the client, the server and by haproxy. This is very
20640important to troubleshoot disconnections, when users complain they have to
20641re-authenticate. The commonly encountered flags are :
20642
20643 -- Persistence cookie is not enabled.
20644
20645 NN No cookie was provided by the client, none was inserted in the
20646 response. For instance, this can be in insert mode with "postonly"
20647 set on a GET request.
20648
20649 II A cookie designating an invalid server was provided by the client,
20650 a valid one was inserted in the response. This typically happens when
Jamie Gloudonaaa21002012-08-25 00:18:33 -040020651 a "server" entry is removed from the configuration, since its cookie
Willy Tarreau996a92c2010-10-13 19:30:47 +020020652 value can be presented by a client when no other server knows it.
20653
20654 NI No cookie was provided by the client, one was inserted in the
20655 response. This typically happens for first requests from every user
20656 in "insert" mode, which makes it an easy way to count real users.
20657
20658 VN A cookie was provided by the client, none was inserted in the
20659 response. This happens for most responses for which the client has
20660 already got a cookie.
20661
20662 VU A cookie was provided by the client, with a last visit date which is
20663 not completely up-to-date, so an updated cookie was provided in
20664 response. This can also happen if there was no date at all, or if
20665 there was a date but the "maxidle" parameter was not set, so that the
20666 cookie can be switched to unlimited time.
20667
20668 EI A cookie was provided by the client, with a last visit date which is
20669 too old for the "maxidle" parameter, so the cookie was ignored and a
20670 new cookie was inserted in the response.
20671
20672 OI A cookie was provided by the client, with a first visit date which is
20673 too old for the "maxlife" parameter, so the cookie was ignored and a
20674 new cookie was inserted in the response.
20675
20676 DI The server designated by the cookie was down, a new server was
20677 selected and a new cookie was emitted in the response.
20678
20679 VI The server designated by the cookie was not marked dead but could not
20680 be reached. A redispatch happened and selected another one, which was
20681 then advertised in the response.
20682
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020683
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200206848.6. Non-printable characters
20685-----------------------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020686
20687In order not to cause trouble to log analysis tools or terminals during log
20688consulting, non-printable characters are not sent as-is into log files, but are
20689converted to the two-digits hexadecimal representation of their ASCII code,
20690prefixed by the character '#'. The only characters that can be logged without
20691being escaped are comprised between 32 and 126 (inclusive). Obviously, the
20692escape character '#' itself is also encoded to avoid any ambiguity ("#23"). It
20693is the same for the character '"' which becomes "#22", as well as '{', '|' and
20694'}' when logging headers.
20695
20696Note that the space character (' ') is not encoded in headers, which can cause
20697issues for tools relying on space count to locate fields. A typical header
20698containing spaces is "User-Agent".
20699
20700Last, it has been observed that some syslog daemons such as syslog-ng escape
20701the quote ('"') with a backslash ('\'). The reverse operation can safely be
20702performed since no quote may appear anywhere else in the logs.
20703
20704
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200207058.7. Capturing HTTP cookies
20706---------------------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020707
20708Cookie capture simplifies the tracking a complete user session. This can be
20709achieved using the "capture cookie" statement in the frontend. Please refer to
Willy Tarreauc57f0e22009-05-10 13:12:33 +020020710section 4.2 for more details. Only one cookie can be captured, and the same
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020711cookie will simultaneously be checked in the request ("Cookie:" header) and in
20712the response ("Set-Cookie:" header). The respective values will be reported in
20713the HTTP logs at the "captured_request_cookie" and "captured_response_cookie"
Willy Tarreauc57f0e22009-05-10 13:12:33 +020020714locations (see section 8.2.3 about HTTP log format). When either cookie is
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020715not seen, a dash ('-') replaces the value. This way, it's easy to detect when a
20716user switches to a new session for example, because the server will reassign it
20717a new cookie. It is also possible to detect if a server unexpectedly sets a
20718wrong cookie to a client, leading to session crossing.
20719
20720 Examples :
20721 # capture the first cookie whose name starts with "ASPSESSION"
20722 capture cookie ASPSESSION len 32
20723
20724 # capture the first cookie whose name is exactly "vgnvisitor"
20725 capture cookie vgnvisitor= len 32
20726
20727
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200207288.8. Capturing HTTP headers
20729---------------------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020730
20731Header captures are useful to track unique request identifiers set by an upper
20732proxy, virtual host names, user-agents, POST content-length, referrers, etc. In
20733the response, one can search for information about the response length, how the
20734server asked the cache to behave, or an object location during a redirection.
20735
20736Header captures are performed using the "capture request header" and "capture
20737response header" statements in the frontend. Please consult their definition in
Willy Tarreauc57f0e22009-05-10 13:12:33 +020020738section 4.2 for more details.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020739
20740It is possible to include both request headers and response headers at the same
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +010020741time. Non-existent headers are logged as empty strings, and if one header
20742appears more than once, only its last occurrence will be logged. Request headers
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020743are grouped within braces '{' and '}' in the same order as they were declared,
20744and delimited with a vertical bar '|' without any space. Response headers
20745follow the same representation, but are displayed after a space following the
20746request headers block. These blocks are displayed just before the HTTP request
20747in the logs.
20748
Willy Tarreaud9ed3d22014-06-13 12:23:06 +020020749As a special case, it is possible to specify an HTTP header capture in a TCP
20750frontend. The purpose is to enable logging of headers which will be parsed in
20751an HTTP backend if the request is then switched to this HTTP backend.
20752
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020753 Example :
20754 # This instance chains to the outgoing proxy
20755 listen proxy-out
20756 mode http
20757 option httplog
20758 option logasap
20759 log global
20760 server cache1 192.168.1.1:3128
20761
20762 # log the name of the virtual server
20763 capture request header Host len 20
20764
20765 # log the amount of data uploaded during a POST
20766 capture request header Content-Length len 10
20767
20768 # log the beginning of the referrer
20769 capture request header Referer len 20
20770
20771 # server name (useful for outgoing proxies only)
20772 capture response header Server len 20
20773
20774 # logging the content-length is useful with "option logasap"
20775 capture response header Content-Length len 10
20776
Davor Ocelice9ed2812017-12-25 17:49:28 +010020777 # log the expected cache behavior on the response
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020778 capture response header Cache-Control len 8
20779
20780 # the Via header will report the next proxy's name
20781 capture response header Via len 20
20782
20783 # log the URL location during a redirection
20784 capture response header Location len 20
20785
20786 >>> Aug 9 20:26:09 localhost \
20787 haproxy[2022]: 127.0.0.1:34014 [09/Aug/2004:20:26:09] proxy-out \
20788 proxy-out/cache1 0/0/0/162/+162 200 +350 - - ---- 0/0/0/0/0 0/0 \
20789 {fr.adserver.yahoo.co||http://fr.f416.mail.} {|864|private||} \
20790 "GET http://fr.adserver.yahoo.com/"
20791
20792 >>> Aug 9 20:30:46 localhost \
20793 haproxy[2022]: 127.0.0.1:34020 [09/Aug/2004:20:30:46] proxy-out \
20794 proxy-out/cache1 0/0/0/182/+182 200 +279 - - ---- 0/0/0/0/0 0/0 \
20795 {w.ods.org||} {Formilux/0.1.8|3495|||} \
Willy Tarreaud72758d2010-01-12 10:42:19 +010020796 "GET http://trafic.1wt.eu/ HTTP/1.1"
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020797
20798 >>> Aug 9 20:30:46 localhost \
20799 haproxy[2022]: 127.0.0.1:34028 [09/Aug/2004:20:30:46] proxy-out \
20800 proxy-out/cache1 0/0/2/126/+128 301 +223 - - ---- 0/0/0/0/0 0/0 \
20801 {www.sytadin.equipement.gouv.fr||http://trafic.1wt.eu/} \
20802 {Apache|230|||http://www.sytadin.} \
Willy Tarreaud72758d2010-01-12 10:42:19 +010020803 "GET http://www.sytadin.equipement.gouv.fr/ HTTP/1.1"
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020804
20805
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200208068.9. Examples of logs
20807---------------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020808
20809These are real-world examples of logs accompanied with an explanation. Some of
20810them have been made up by hand. The syslog part has been removed for better
20811reading. Their sole purpose is to explain how to decipher them.
20812
20813 >>> haproxy[674]: 127.0.0.1:33318 [15/Oct/2003:08:31:57.130] px-http \
20814 px-http/srv1 6559/0/7/147/6723 200 243 - - ---- 5/3/3/1/0 0/0 \
20815 "HEAD / HTTP/1.0"
20816
20817 => long request (6.5s) entered by hand through 'telnet'. The server replied
20818 in 147 ms, and the session ended normally ('----')
20819
20820 >>> haproxy[674]: 127.0.0.1:33319 [15/Oct/2003:08:31:57.149] px-http \
20821 px-http/srv1 6559/1230/7/147/6870 200 243 - - ---- 324/239/239/99/0 \
20822 0/9 "HEAD / HTTP/1.0"
20823
20824 => Idem, but the request was queued in the global queue behind 9 other
20825 requests, and waited there for 1230 ms.
20826
20827 >>> haproxy[674]: 127.0.0.1:33320 [15/Oct/2003:08:32:17.654] px-http \
20828 px-http/srv1 9/0/7/14/+30 200 +243 - - ---- 3/3/3/1/0 0/0 \
20829 "GET /image.iso HTTP/1.0"
20830
20831 => request for a long data transfer. The "logasap" option was specified, so
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +010020832 the log was produced just before transferring data. The server replied in
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020833 14 ms, 243 bytes of headers were sent to the client, and total time from
20834 accept to first data byte is 30 ms.
20835
20836 >>> haproxy[674]: 127.0.0.1:33320 [15/Oct/2003:08:32:17.925] px-http \
20837 px-http/srv1 9/0/7/14/30 502 243 - - PH-- 3/2/2/0/0 0/0 \
20838 "GET /cgi-bin/bug.cgi? HTTP/1.0"
20839
Christopher Faulet87f1f3d2019-07-18 14:51:20 +020020840 => the proxy blocked a server response either because of an "http-response
20841 deny" rule, or because the response was improperly formatted and not
20842 HTTP-compliant, or because it blocked sensitive information which risked
20843 being cached. In this case, the response is replaced with a "502 bad
20844 gateway". The flags ("PH--") tell us that it was haproxy who decided to
20845 return the 502 and not the server.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020846
20847 >>> haproxy[18113]: 127.0.0.1:34548 [15/Oct/2003:15:18:55.798] px-http \
Willy Tarreaud72758d2010-01-12 10:42:19 +010020848 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 +010020849
20850 => the client never completed its request and aborted itself ("C---") after
20851 8.5s, while the proxy was waiting for the request headers ("-R--").
20852 Nothing was sent to any server.
20853
20854 >>> haproxy[18113]: 127.0.0.1:34549 [15/Oct/2003:15:19:06.103] px-http \
20855 px-http/<NOSRV> -1/-1/-1/-1/50001 408 0 - - cR-- 2/2/2/0/0 0/0 ""
20856
20857 => The client never completed its request, which was aborted by the
20858 time-out ("c---") after 50s, while the proxy was waiting for the request
Davor Ocelice9ed2812017-12-25 17:49:28 +010020859 headers ("-R--"). Nothing was sent to any server, but the proxy could
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020860 send a 408 return code to the client.
20861
20862 >>> haproxy[18989]: 127.0.0.1:34550 [15/Oct/2003:15:24:28.312] px-tcp \
20863 px-tcp/srv1 0/0/5007 0 cD 0/0/0/0/0 0/0
20864
20865 => This log was produced with "option tcplog". The client timed out after
20866 5 seconds ("c----").
20867
20868 >>> haproxy[18989]: 10.0.0.1:34552 [15/Oct/2003:15:26:31.462] px-http \
20869 px-http/srv1 3183/-1/-1/-1/11215 503 0 - - SC-- 205/202/202/115/3 \
Willy Tarreaud72758d2010-01-12 10:42:19 +010020870 0/0 "HEAD / HTTP/1.0"
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020871
20872 => The request took 3s to complete (probably a network problem), and the
Willy Tarreauc57f0e22009-05-10 13:12:33 +020020873 connection to the server failed ('SC--') after 4 attempts of 2 seconds
Willy Tarreaucc6c8912009-02-22 10:53:55 +010020874 (config says 'retries 3'), and no redispatch (otherwise we would have
20875 seen "/+3"). Status code 503 was returned to the client. There were 115
20876 connections on this server, 202 connections on this proxy, and 205 on
20877 the global process. It is possible that the server refused the
20878 connection because of too many already established.
Willy Tarreau844e3c52008-01-11 16:28:18 +010020879
Willy Tarreau52b2d222011-09-07 23:48:48 +020020880
Christopher Fauletc3fe5332016-04-07 15:30:10 +0200208819. Supported filters
20882--------------------
20883
20884Here are listed officially supported filters with the list of parameters they
20885accept. Depending on compile options, some of these filters might be
20886unavailable. The list of available filters is reported in haproxy -vv.
20887
20888See also : "filter"
20889
208909.1. Trace
20891----------
20892
Christopher Fauletc41d8bd2020-11-17 10:43:26 +010020893filter trace [name <name>] [random-forwarding] [hexdump]
Christopher Fauletc3fe5332016-04-07 15:30:10 +020020894
20895 Arguments:
20896 <name> is an arbitrary name that will be reported in
20897 messages. If no name is provided, "TRACE" is used.
20898
Christopher Faulet96a577a2020-11-17 10:45:05 +010020899 <quiet> inhibits trace messages.
Christopher Fauletc3fe5332016-04-07 15:30:10 +020020900
Davor Ocelice9ed2812017-12-25 17:49:28 +010020901 <random-forwarding> enables the random forwarding of parsed data. By
Christopher Fauletc3fe5332016-04-07 15:30:10 +020020902 default, this filter forwards all previously parsed
20903 data. With this parameter, it only forwards a random
20904 amount of the parsed data.
20905
Davor Ocelice9ed2812017-12-25 17:49:28 +010020906 <hexdump> dumps all forwarded data to the server and the client.
Christopher Faulet31bfe1f2016-12-09 17:42:38 +010020907
Christopher Fauletc3fe5332016-04-07 15:30:10 +020020908This filter can be used as a base to develop new filters. It defines all
20909callbacks and print a message on the standard error stream (stderr) with useful
20910information for all of them. It may be useful to debug the activity of other
20911filters or, quite simply, HAProxy's activity.
20912
20913Using <random-parsing> and/or <random-forwarding> parameters is a good way to
20914tests the behavior of a filter that parses data exchanged between a client and
20915a server by adding some latencies in the processing.
20916
20917
209189.2. HTTP compression
20919---------------------
20920
20921filter compression
20922
20923The HTTP compression has been moved in a filter in HAProxy 1.7. "compression"
20924keyword must still be used to enable and configure the HTTP compression. And
Christopher Fauletb30b3102019-09-12 23:03:09 +020020925when no other filter is used, it is enough. When used with the cache or the
20926fcgi-app enabled, it is also enough. In this case, the compression is always
20927done after the response is stored in the cache. But it is mandatory to
20928explicitly use a filter line to enable the HTTP compression when at least one
20929filter other than the cache or the fcgi-app is used for the same
20930listener/frontend/backend. This is important to know the filters evaluation
20931order.
Christopher Fauletc3fe5332016-04-07 15:30:10 +020020932
Christopher Fauletb30b3102019-09-12 23:03:09 +020020933See also : "compression", section 9.4 about the cache filter and section 9.5
20934 about the fcgi-app filter.
Christopher Fauletc3fe5332016-04-07 15:30:10 +020020935
20936
Christopher Fauletf7e4e7e2016-10-27 22:29:49 +0200209379.3. Stream Processing Offload Engine (SPOE)
20938--------------------------------------------
20939
20940filter spoe [engine <name>] config <file>
20941
20942 Arguments :
20943
20944 <name> is the engine name that will be used to find the right scope in
20945 the configuration file. If not provided, all the file will be
20946 parsed.
20947
20948 <file> is the path of the engine configuration file. This file can
20949 contain configuration of several engines. In this case, each
20950 part must be placed in its own scope.
20951
20952The Stream Processing Offload Engine (SPOE) is a filter communicating with
20953external components. It allows the offload of some specifics processing on the
Davor Ocelice9ed2812017-12-25 17:49:28 +010020954streams in tiered applications. These external components and information
Christopher Fauletf7e4e7e2016-10-27 22:29:49 +020020955exchanged with them are configured in dedicated files, for the main part. It
20956also requires dedicated backends, defined in HAProxy configuration.
20957
20958SPOE communicates with external components using an in-house binary protocol,
20959the Stream Processing Offload Protocol (SPOP).
20960
Tim Düsterhus4896c442016-11-29 02:15:19 +010020961For all information about the SPOE configuration and the SPOP specification, see
Christopher Fauletf7e4e7e2016-10-27 22:29:49 +020020962"doc/SPOE.txt".
20963
Christopher Faulet99a17a22018-12-11 09:18:27 +0100209649.4. Cache
20965----------
20966
20967filter cache <name>
20968
20969 Arguments :
20970
20971 <name> is name of the cache section this filter will use.
20972
20973The cache uses a filter to store cacheable responses. The HTTP rules
20974"cache-store" and "cache-use" must be used to define how and when to use a
John Roeslerfb2fce12019-07-10 15:45:51 -050020975cache. By default the corresponding filter is implicitly defined. And when no
Christopher Fauletb30b3102019-09-12 23:03:09 +020020976other filters than fcgi-app or compression are used, it is enough. In such
20977case, the compression filter is always evaluated after the cache filter. But it
20978is mandatory to explicitly use a filter line to use a cache when at least one
20979filter other than the compression or the fcgi-app is used for the same
Christopher Faulet27d93c32018-12-15 22:32:02 +010020980listener/frontend/backend. This is important to know the filters evaluation
20981order.
Christopher Faulet99a17a22018-12-11 09:18:27 +010020982
Christopher Fauletb30b3102019-09-12 23:03:09 +020020983See also : section 9.2 about the compression filter, section 9.5 about the
20984 fcgi-app filter and section 6 about cache.
20985
20986
209879.5. Fcgi-app
20988-------------
20989
Daniel Corbett67a82712020-07-06 23:01:19 -040020990filter fcgi-app <name>
Christopher Fauletb30b3102019-09-12 23:03:09 +020020991
20992 Arguments :
20993
20994 <name> is name of the fcgi-app section this filter will use.
20995
20996The FastCGI application uses a filter to evaluate all custom parameters on the
20997request path, and to process the headers on the response path. the <name> must
20998reference an existing fcgi-app section. The directive "use-fcgi-app" should be
20999used to define the application to use. By default the corresponding filter is
21000implicitly defined. And when no other filters than cache or compression are
21001used, it is enough. But it is mandatory to explicitly use a filter line to a
21002fcgi-app when at least one filter other than the compression or the cache is
21003used for the same backend. This is important to know the filters evaluation
21004order.
21005
21006See also: "use-fcgi-app", section 9.2 about the compression filter, section 9.4
21007 about the cache filter and section 10 about FastCGI application.
21008
21009
Miroslav Zagoracdc32cd92020-12-13 18:32:57 +0100210109.6. OpenTracing
21011----------------
21012
21013The OpenTracing filter adds native support for using distributed tracing in
21014HAProxy. This is enabled by sending an OpenTracing compliant request to one
21015of the supported tracers such as Datadog, Jaeger, Lightstep and Zipkin tracers.
21016Please note: tracers are not listed by any preference, but alphabetically.
21017
21018This feature is only enabled when haproxy was built with USE_OT=1.
21019
21020The OpenTracing filter activation is done explicitly by specifying it in the
21021HAProxy configuration. If this is not done, the OpenTracing filter in no way
21022participates in the work of HAProxy.
21023
21024filter opentracing [id <id>] config <file>
21025
21026 Arguments :
21027
21028 <id> is the OpenTracing filter id that will be used to find the
21029 right scope in the configuration file. If no filter id is
21030 specified, 'ot-filter' is used as default. If scope is not
21031 specified in the configuration file, it applies to all defined
21032 OpenTracing filters.
21033
21034 <file> is the path of the OpenTracing configuration file. The same
21035 file can contain configurations for multiple OpenTracing
21036 filters simultaneously. In that case we do not need to define
21037 scope so the same configuration applies to all filters or each
21038 filter must have its own scope defined.
21039
21040More detailed documentation related to the operation, configuration and use
21041of the filter can be found in the contrib/opentracing directory.
21042
21043
Christopher Fauletb30b3102019-09-12 23:03:09 +02002104410. FastCGI applications
21045-------------------------
21046
21047HAProxy is able to send HTTP requests to Responder FastCGI applications. This
21048feature was added in HAProxy 2.1. To do so, servers must be configured to use
21049the FastCGI protocol (using the keyword "proto fcgi" on the server line) and a
21050FastCGI application must be configured and used by the backend managing these
21051servers (using the keyword "use-fcgi-app" into the proxy section). Several
21052FastCGI applications may be defined, but only one can be used at a time by a
21053backend.
21054
21055HAProxy implements all features of the FastCGI specification for Responder
21056application. Especially it is able to multiplex several requests on a simple
21057connection.
21058
2105910.1. Setup
21060-----------
21061
2106210.1.1. Fcgi-app section
21063--------------------------
21064
21065fcgi-app <name>
21066 Declare a FastCGI application named <name>. To be valid, at least the
21067 document root must be defined.
21068
21069acl <aclname> <criterion> [flags] [operator] <value> ...
21070 Declare or complete an access list.
21071
21072 See "acl" keyword in section 4.2 and section 7 about ACL usage for
21073 details. ACLs defined for a FastCGI application are private. They cannot be
21074 used by any other application or by any proxy. In the same way, ACLs defined
21075 in any other section are not usable by a FastCGI application. However,
21076 Pre-defined ACLs are available.
21077
21078docroot <path>
21079 Define the document root on the remote host. <path> will be used to build
21080 the default value of FastCGI parameters SCRIPT_FILENAME and
21081 PATH_TRANSLATED. It is a mandatory setting.
21082
21083index <script-name>
21084 Define the script name that will be appended after an URI that ends with a
21085 slash ("/") to set the default value of the FastCGI parameter SCRIPT_NAME. It
21086 is an optional setting.
21087
21088 Example :
21089 index index.php
21090
21091log-stderr global
21092log-stderr <address> [len <length>] [format <format>]
Jan Wagner3e678602020-12-17 22:22:32 +010021093 [sample <ranges>:<sample_size>] <facility> [<level> [<minlevel>]]
Christopher Fauletb30b3102019-09-12 23:03:09 +020021094 Enable logging of STDERR messages reported by the FastCGI application.
21095
21096 See "log" keyword in section 4.2 for details. It is an optional setting. By
21097 default STDERR messages are ignored.
21098
21099pass-header <name> [ { if | unless } <condition> ]
21100 Specify the name of a request header which will be passed to the FastCGI
21101 application. It may optionally be followed by an ACL-based condition, in
21102 which case it will only be evaluated if the condition is true.
21103
21104 Most request headers are already available to the FastCGI application,
21105 prefixed with "HTTP_". Thus, this directive is only required to pass headers
21106 that are purposefully omitted. Currently, the headers "Authorization",
21107 "Proxy-Authorization" and hop-by-hop headers are omitted.
21108
21109 Note that the headers "Content-type" and "Content-length" are never passed to
21110 the FastCGI application because they are already converted into parameters.
21111
21112path-info <regex>
Christopher Faulet28cb3662020-02-14 14:47:37 +010021113 Define a regular expression to extract the script-name and the path-info from
Christopher Faulet6c57f2d2020-02-14 16:55:52 +010021114 the URL-decoded path. Thus, <regex> may have two captures: the first one to
21115 capture the script name and the second one to capture the path-info. The
21116 first one is mandatory, the second one is optional. This way, it is possible
21117 to extract the script-name from the path ignoring the path-info. It is an
21118 optional setting. If it is not defined, no matching is performed on the
21119 path. and the FastCGI parameters PATH_INFO and PATH_TRANSLATED are not
21120 filled.
Christopher Faulet28cb3662020-02-14 14:47:37 +010021121
21122 For security reason, when this regular expression is defined, the newline and
Ilya Shipitsin8525fd92020-02-29 12:34:59 +050021123 the null characters are forbidden from the path, once URL-decoded. The reason
Christopher Faulet28cb3662020-02-14 14:47:37 +010021124 to such limitation is because otherwise the matching always fails (due to a
21125 limitation one the way regular expression are executed in HAProxy). So if one
21126 of these two characters is found in the URL-decoded path, an error is
21127 returned to the client. The principle of least astonishment is applied here.
Christopher Fauletb30b3102019-09-12 23:03:09 +020021128
21129 Example :
Christopher Faulet6c57f2d2020-02-14 16:55:52 +010021130 path-info ^(/.+\.php)(/.*)?$ # both script-name and path-info may be set
21131 path-info ^(/.+\.php) # the path-info is ignored
Christopher Fauletb30b3102019-09-12 23:03:09 +020021132
21133option get-values
21134no option get-values
21135 Enable or disable the retrieve of variables about connection management.
21136
Daniel Corbett67a82712020-07-06 23:01:19 -040021137 HAProxy is able to send the record FCGI_GET_VALUES on connection
Christopher Fauletb30b3102019-09-12 23:03:09 +020021138 establishment to retrieve the value for following variables:
21139
21140 * FCGI_MAX_REQS The maximum number of concurrent requests this
21141 application will accept.
21142
William Lallemand93e548e2019-09-30 13:54:02 +020021143 * FCGI_MPXS_CONNS "0" if this application does not multiplex connections,
21144 "1" otherwise.
Christopher Fauletb30b3102019-09-12 23:03:09 +020021145
21146 Some FastCGI applications does not support this feature. Some others close
Ilya Shipitsin11057a32020-06-21 21:18:27 +050021147 the connection immediately after sending their response. So, by default, this
Christopher Fauletb30b3102019-09-12 23:03:09 +020021148 option is disabled.
21149
21150 Note that the maximum number of concurrent requests accepted by a FastCGI
21151 application is a connection variable. It only limits the number of streams
21152 per connection. If the global load must be limited on the application, the
21153 server parameters "maxconn" and "pool-max-conn" must be set. In addition, if
21154 an application does not support connection multiplexing, the maximum number
21155 of concurrent requests is automatically set to 1.
21156
21157option keep-conn
21158no option keep-conn
21159 Instruct the FastCGI application to keep the connection open or not after
21160 sending a response.
21161
21162 If disabled, the FastCGI application closes the connection after responding
21163 to this request. By default, this option is enabled.
21164
21165option max-reqs <reqs>
21166 Define the maximum number of concurrent requests this application will
21167 accept.
21168
21169 This option may be overwritten if the variable FCGI_MAX_REQS is retrieved
21170 during connection establishment. Furthermore, if the application does not
21171 support connection multiplexing, this option will be ignored. By default set
21172 to 1.
21173
21174option mpxs-conns
21175no option mpxs-conns
21176 Enable or disable the support of connection multiplexing.
21177
21178 This option may be overwritten if the variable FCGI_MPXS_CONNS is retrieved
21179 during connection establishment. It is disabled by default.
21180
21181set-param <name> <fmt> [ { if | unless } <condition> ]
21182 Set a FastCGI parameter that should be passed to this application. Its
21183 value, defined by <fmt> must follows the log-format rules (see section 8.2.4
21184 "Custom Log format"). It may optionally be followed by an ACL-based
21185 condition, in which case it will only be evaluated if the condition is true.
21186
21187 With this directive, it is possible to overwrite the value of default FastCGI
21188 parameters. If the value is evaluated to an empty string, the rule is
21189 ignored. These directives are evaluated in their declaration order.
21190
21191 Example :
21192 # PHP only, required if PHP was built with --enable-force-cgi-redirect
21193 set-param REDIRECT_STATUS 200
21194
21195 set-param PHP_AUTH_DIGEST %[req.hdr(Authorization)]
21196
21197
2119810.1.2. Proxy section
21199---------------------
21200
21201use-fcgi-app <name>
21202 Define the FastCGI application to use for the backend.
21203
21204 Arguments :
21205 <name> is the name of the FastCGI application to use.
21206
21207 This keyword is only available for HTTP proxies with the backend capability
21208 and with at least one FastCGI server. However, FastCGI servers can be mixed
21209 with HTTP servers. But except there is a good reason to do so, it is not
21210 recommended (see section 10.3 about the limitations for details). Only one
21211 application may be defined at a time per backend.
21212
21213 Note that, once a FastCGI application is referenced for a backend, depending
21214 on the configuration some processing may be done even if the request is not
21215 sent to a FastCGI server. Rules to set parameters or pass headers to an
21216 application are evaluated.
21217
21218
2121910.1.3. Example
21220---------------
21221
21222 frontend front-http
21223 mode http
21224 bind *:80
21225 bind *:
21226
21227 use_backend back-dynamic if { path_reg ^/.+\.php(/.*)?$ }
21228 default_backend back-static
21229
21230 backend back-static
21231 mode http
21232 server www A.B.C.D:80
21233
21234 backend back-dynamic
21235 mode http
21236 use-fcgi-app php-fpm
21237 server php-fpm A.B.C.D:9000 proto fcgi
21238
21239 fcgi-app php-fpm
21240 log-stderr global
21241 option keep-conn
21242
21243 docroot /var/www/my-app
21244 index index.php
21245 path-info ^(/.+\.php)(/.*)?$
21246
21247
2124810.2. Default parameters
21249------------------------
21250
21251A Responder FastCGI application has the same purpose as a CGI/1.1 program. In
21252the CGI/1.1 specification (RFC3875), several variables must be passed to the
Ilya Shipitsin8525fd92020-02-29 12:34:59 +050021253script. So HAProxy set them and some others commonly used by FastCGI
Christopher Fauletb30b3102019-09-12 23:03:09 +020021254applications. All these variables may be overwritten, with caution though.
21255
21256 +-------------------+-----------------------------------------------------+
21257 | AUTH_TYPE | Identifies the mechanism, if any, used by HAProxy |
21258 | | to authenticate the user. Concretely, only the |
21259 | | BASIC authentication mechanism is supported. |
21260 | | |
21261 +-------------------+-----------------------------------------------------+
21262 | CONTENT_LENGTH | Contains the size of the message-body attached to |
21263 | | the request. It means only requests with a known |
21264 | | size are considered as valid and sent to the |
21265 | | application. |
21266 | | |
21267 +-------------------+-----------------------------------------------------+
21268 | CONTENT_TYPE | Contains the type of the message-body attached to |
21269 | | the request. It may not be set. |
21270 | | |
21271 +-------------------+-----------------------------------------------------+
21272 | DOCUMENT_ROOT | Contains the document root on the remote host under |
21273 | | which the script should be executed, as defined in |
21274 | | the application's configuration. |
21275 | | |
21276 +-------------------+-----------------------------------------------------+
21277 | GATEWAY_INTERFACE | Contains the dialect of CGI being used by HAProxy |
21278 | | to communicate with the FastCGI application. |
21279 | | Concretely, it is set to "CGI/1.1". |
21280 | | |
21281 +-------------------+-----------------------------------------------------+
21282 | PATH_INFO | Contains the portion of the URI path hierarchy |
21283 | | following the part that identifies the script |
21284 | | itself. To be set, the directive "path-info" must |
21285 | | be defined. |
21286 | | |
21287 +-------------------+-----------------------------------------------------+
21288 | PATH_TRANSLATED | If PATH_INFO is set, it is its translated version. |
21289 | | It is the concatenation of DOCUMENT_ROOT and |
21290 | | PATH_INFO. If PATH_INFO is not set, this parameters |
21291 | | is not set too. |
21292 | | |
21293 +-------------------+-----------------------------------------------------+
21294 | QUERY_STRING | Contains the request's query string. It may not be |
21295 | | set. |
21296 | | |
21297 +-------------------+-----------------------------------------------------+
21298 | REMOTE_ADDR | Contains the network address of the client sending |
21299 | | the request. |
21300 | | |
21301 +-------------------+-----------------------------------------------------+
21302 | REMOTE_USER | Contains the user identification string supplied by |
21303 | | client as part of user authentication. |
21304 | | |
21305 +-------------------+-----------------------------------------------------+
21306 | REQUEST_METHOD | Contains the method which should be used by the |
21307 | | script to process the request. |
21308 | | |
21309 +-------------------+-----------------------------------------------------+
21310 | REQUEST_URI | Contains the request's URI. |
21311 | | |
21312 +-------------------+-----------------------------------------------------+
21313 | SCRIPT_FILENAME | Contains the absolute pathname of the script. it is |
21314 | | the concatenation of DOCUMENT_ROOT and SCRIPT_NAME. |
21315 | | |
21316 +-------------------+-----------------------------------------------------+
21317 | SCRIPT_NAME | Contains the name of the script. If the directive |
21318 | | "path-info" is defined, it is the first part of the |
21319 | | URI path hierarchy, ending with the script name. |
21320 | | Otherwise, it is the entire URI path. |
21321 | | |
21322 +-------------------+-----------------------------------------------------+
21323 | SERVER_NAME | Contains the name of the server host to which the |
21324 | | client request is directed. It is the value of the |
21325 | | header "Host", if defined. Otherwise, the |
21326 | | destination address of the connection on the client |
21327 | | side. |
21328 | | |
21329 +-------------------+-----------------------------------------------------+
21330 | SERVER_PORT | Contains the destination TCP port of the connection |
21331 | | on the client side, which is the port the client |
21332 | | connected to. |
21333 | | |
21334 +-------------------+-----------------------------------------------------+
21335 | SERVER_PROTOCOL | Contains the request's protocol. |
21336 | | |
21337 +-------------------+-----------------------------------------------------+
21338 | HTTPS | Set to a non-empty value ("on") if the script was |
21339 | | queried through the HTTPS protocol. |
21340 | | |
21341 +-------------------+-----------------------------------------------------+
21342
21343
2134410.3. Limitations
21345------------------
21346
21347The current implementation have some limitations. The first one is about the
21348way some request headers are hidden to the FastCGI applications. This happens
21349during the headers analysis, on the backend side, before the connection
21350establishment. At this stage, HAProxy know the backend is using a FastCGI
21351application but it don't know if the request will be routed to a FastCGI server
21352or not. But to hide request headers, it simply removes them from the HTX
21353message. So, if the request is finally routed to an HTTP server, it never see
21354these headers. For this reason, it is not recommended to mix FastCGI servers
21355and HTTP servers under the same backend.
21356
21357Similarly, the rules "set-param" and "pass-header" are evaluated during the
21358request headers analysis. So the evaluation is always performed, even if the
21359requests is finally forwarded to an HTTP server.
21360
21361About the rules "set-param", when a rule is applied, a pseudo header is added
21362into the HTX message. So, the same way than for HTTP header rewrites, it may
21363fail if the buffer is full. The rules "set-param" will compete with
21364"http-request" ones.
21365
21366Finally, all FastCGI params and HTTP headers are sent into a unique record
21367FCGI_PARAM. Encoding of this record must be done in one pass, otherwise a
21368processing error is returned. It means the record FCGI_PARAM, once encoded,
21369must not exceeds the size of a buffer. However, there is no reserve to respect
21370here.
William Lallemand86d0df02017-11-24 21:36:45 +010021371
Willy Tarreau0ba27502007-12-24 16:55:16 +010021372/*
21373 * Local variables:
21374 * fill-column: 79
21375 * End:
21376 */