blob: 8678b8028ca19361f3714a9c22a1a75ec4efd823 [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 Tarreau9dc6b972019-06-16 21:49:47 +02005 version 2.1
Willy Tarreau6a06a402007-07-15 20:15:28 +02006 willy tarreau
Willy Tarreaudb514072019-07-16 19:15:28 +02007 2019/07/16
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
442.4. Time format
452.5. Examples
Willy Tarreauc57f0e22009-05-10 13:12:33 +020046
473. Global parameters
483.1. Process management and security
493.2. Performance tuning
503.3. Debugging
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +0100513.4. Userlists
Cyril Bontédc4d9032012-04-08 21:57:39 +0200523.5. Peers
Cyril Bonté307ee1e2015-09-28 23:16:06 +0200533.6. Mailers
William Lallemandc9515522019-06-12 16:32:11 +0200543.7. Programs
Willy Tarreauc57f0e22009-05-10 13:12:33 +020055
564. Proxies
574.1. Proxy keywords matrix
584.2. Alphabetically sorted keywords reference
59
Davor Ocelice9ed2812017-12-25 17:49:28 +0100605. Bind and server options
Willy Tarreau086fbf52012-09-24 20:34:51 +0200615.1. Bind options
625.2. Server and default-server options
Baptiste Assmann1fa66662015-04-14 00:28:47 +0200635.3. Server DNS resolution
645.3.1. Global overview
655.3.2. The resolvers section
Willy Tarreauc57f0e22009-05-10 13:12:33 +020066
676. HTTP header manipulation
68
Willy Tarreau74ca5042013-06-11 23:12:07 +0200697. Using ACLs and fetching samples
707.1. ACL basics
717.1.1. Matching booleans
727.1.2. Matching integers
737.1.3. Matching strings
747.1.4. Matching regular expressions (regexes)
757.1.5. Matching arbitrary data blocks
767.1.6. Matching IPv4 and IPv6 addresses
777.2. Using ACLs to form conditions
787.3. Fetching samples
Thierry FOURNIER060762e2014-04-23 13:29:15 +0200797.3.1. Converters
807.3.2. Fetching samples from internal states
817.3.3. Fetching samples at Layer 4
827.3.4. Fetching samples at Layer 5
837.3.5. Fetching samples from buffer contents (Layer 6)
847.3.6. Fetching HTTP samples (Layer 7)
Willy Tarreau74ca5042013-06-11 23:12:07 +0200857.4. Pre-defined ACLs
Willy Tarreauc57f0e22009-05-10 13:12:33 +020086
878. Logging
888.1. Log levels
898.2. Log formats
908.2.1. Default log format
918.2.2. TCP log format
928.2.3. HTTP log format
William Lallemand48940402012-01-30 16:47:22 +0100938.2.4. Custom log format
Willy Tarreau5f51e1a2012-12-03 18:40:10 +0100948.2.5. Error log format
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200958.3. Advanced logging options
968.3.1. Disabling logging of external tests
978.3.2. Logging before waiting for the session to terminate
988.3.3. Raising log level upon errors
998.3.4. Disabling logging of successful connections
1008.4. Timing events
1018.5. Session state at disconnection
1028.6. Non-printable characters
1038.7. Capturing HTTP cookies
1048.8. Capturing HTTP headers
1058.9. Examples of logs
106
Christopher Fauletc3fe5332016-04-07 15:30:10 +02001079. Supported filters
1089.1. Trace
1099.2. HTTP compression
Christopher Fauletf7e4e7e2016-10-27 22:29:49 +02001109.3. Stream Processing Offload Engine (SPOE)
Christopher Faulet99a17a22018-12-11 09:18:27 +01001119.4. Cache
Christopher Fauletc3fe5332016-04-07 15:30:10 +0200112
William Lallemand86d0df02017-11-24 21:36:45 +010011310. Cache
Cyril Bonté7b888f12017-11-26 22:24:31 +010011410.1. Limitation
11510.2. Setup
11610.2.1. Cache section
11710.2.2. Proxy section
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200118
1191. Quick reminder about HTTP
120----------------------------
121
Davor Ocelice9ed2812017-12-25 17:49:28 +0100122When HAProxy is running in HTTP mode, both the request and the response are
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200123fully analyzed and indexed, thus it becomes possible to build matching criteria
124on almost anything found in the contents.
125
126However, it is important to understand how HTTP requests and responses are
127formed, and how HAProxy decomposes them. It will then become easier to write
128correct rules and to debug existing configurations.
129
130
1311.1. The HTTP transaction model
132-------------------------------
133
134The HTTP protocol is transaction-driven. This means that each request will lead
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +0100135to one and only one response. Traditionally, a TCP connection is established
Davor Ocelice9ed2812017-12-25 17:49:28 +0100136from the client to the server, a request is sent by the client through the
137connection, the server responds, and the connection is closed. A new request
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200138will involve a new connection :
139
140 [CON1] [REQ1] ... [RESP1] [CLO1] [CON2] [REQ2] ... [RESP2] [CLO2] ...
141
142In this mode, called the "HTTP close" mode, there are as many connection
143establishments as there are HTTP transactions. Since the connection is closed
144by the server after the response, the client does not need to know the content
145length.
146
147Due to the transactional nature of the protocol, it was possible to improve it
148to avoid closing a connection between two subsequent transactions. In this mode
149however, it is mandatory that the server indicates the content length for each
150response so that the client does not wait indefinitely. For this, a special
151header is used: "Content-length". This mode is called the "keep-alive" mode :
152
153 [CON] [REQ1] ... [RESP1] [REQ2] ... [RESP2] [CLO] ...
154
155Its advantages are a reduced latency between transactions, and less processing
156power required on the server side. It is generally better than the close mode,
157but not always because the clients often limit their concurrent connections to
Patrick Mezard9ec2ec42010-06-12 17:02:45 +0200158a smaller value.
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200159
Willy Tarreau95c4e142017-11-26 12:18:55 +0100160Another improvement in the communications is the pipelining mode. It still uses
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200161keep-alive, but the client does not wait for the first response to send the
162second request. This is useful for fetching large number of images composing a
163page :
164
165 [CON] [REQ1] [REQ2] ... [RESP1] [RESP2] [CLO] ...
166
167This can obviously have a tremendous benefit on performance because the network
168latency is eliminated between subsequent requests. Many HTTP agents do not
169correctly support pipelining since there is no way to associate a response with
170the corresponding request in HTTP. For this reason, it is mandatory for the
Cyril Bonté78caf842010-03-10 22:41:43 +0100171server to reply in the exact same order as the requests were received.
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200172
Willy Tarreau95c4e142017-11-26 12:18:55 +0100173The next improvement is the multiplexed mode, as implemented in HTTP/2. This
174time, each transaction is assigned a single stream identifier, and all streams
175are multiplexed over an existing connection. Many requests can be sent in
176parallel by the client, and responses can arrive in any order since they also
177carry the stream identifier.
178
Willy Tarreau70dffda2014-01-30 03:07:23 +0100179By default HAProxy operates in keep-alive mode with regards to persistent
180connections: for each connection it processes each request and response, and
181leaves the connection idle on both sides between the end of a response and the
Willy Tarreau95c4e142017-11-26 12:18:55 +0100182start of a new request. When it receives HTTP/2 connections from a client, it
183processes all the requests in parallel and leaves the connection idling,
184waiting for new requests, just as if it was a keep-alive HTTP connection.
Patrick Mezard9ec2ec42010-06-12 17:02:45 +0200185
Christopher Faulet315b39c2018-09-21 16:26:19 +0200186HAProxy supports 4 connection modes :
Willy Tarreau70dffda2014-01-30 03:07:23 +0100187 - keep alive : all requests and responses are processed (default)
188 - tunnel : only the first request and response are processed,
Christopher Faulet6c9bbb22019-03-26 21:37:23 +0100189 everything else is forwarded with no analysis (deprecated).
Willy Tarreau70dffda2014-01-30 03:07:23 +0100190 - server close : the server-facing connection is closed after the response.
Christopher Faulet315b39c2018-09-21 16:26:19 +0200191 - close : the connection is actively closed after end of response.
Willy Tarreau70dffda2014-01-30 03:07:23 +0100192
Davor Ocelice9ed2812017-12-25 17:49:28 +0100193For HTTP/2, the connection mode resembles more the "server close" mode : given
194the independence of all streams, there is currently no place to hook the idle
Willy Tarreau95c4e142017-11-26 12:18:55 +0100195server connection after a response, so it is closed after the response. HTTP/2
196is only supported for incoming connections, not on connections going to
197servers.
198
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200199
2001.2. HTTP request
201-----------------
202
203First, let's consider this HTTP request :
204
205 Line Contents
Willy Tarreaud72758d2010-01-12 10:42:19 +0100206 number
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200207 1 GET /serv/login.php?lang=en&profile=2 HTTP/1.1
208 2 Host: www.mydomain.com
209 3 User-agent: my small browser
210 4 Accept: image/jpeg, image/gif
211 5 Accept: image/png
212
213
2141.2.1. The Request line
215-----------------------
216
217Line 1 is the "request line". It is always composed of 3 fields :
218
219 - a METHOD : GET
220 - a URI : /serv/login.php?lang=en&profile=2
221 - a version tag : HTTP/1.1
222
223All of them are delimited by what the standard calls LWS (linear white spaces),
224which are commonly spaces, but can also be tabs or line feeds/carriage returns
225followed by spaces/tabs. The method itself cannot contain any colon (':') and
226is limited to alphabetic letters. All those various combinations make it
227desirable that HAProxy performs the splitting itself rather than leaving it to
228the user to write a complex or inaccurate regular expression.
229
230The URI itself can have several forms :
231
232 - A "relative URI" :
233
234 /serv/login.php?lang=en&profile=2
235
236 It is a complete URL without the host part. This is generally what is
237 received by servers, reverse proxies and transparent proxies.
238
239 - An "absolute URI", also called a "URL" :
240
241 http://192.168.0.12:8080/serv/login.php?lang=en&profile=2
242
243 It is composed of a "scheme" (the protocol name followed by '://'), a host
244 name or address, optionally a colon (':') followed by a port number, then
245 a relative URI beginning at the first slash ('/') after the address part.
246 This is generally what proxies receive, but a server supporting HTTP/1.1
247 must accept this form too.
248
249 - a star ('*') : this form is only accepted in association with the OPTIONS
250 method and is not relayable. It is used to inquiry a next hop's
251 capabilities.
Willy Tarreaud72758d2010-01-12 10:42:19 +0100252
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200253 - an address:port combination : 192.168.0.12:80
254 This is used with the CONNECT method, which is used to establish TCP
255 tunnels through HTTP proxies, generally for HTTPS, but sometimes for
256 other protocols too.
257
258In a relative URI, two sub-parts are identified. The part before the question
259mark is called the "path". It is typically the relative path to static objects
260on the server. The part after the question mark is called the "query string".
261It is mostly used with GET requests sent to dynamic scripts and is very
262specific to the language, framework or application in use.
263
Willy Tarreau95c4e142017-11-26 12:18:55 +0100264HTTP/2 doesn't convey a version information with the request, so the version is
Davor Ocelice9ed2812017-12-25 17:49:28 +0100265assumed to be the same as the one of the underlying protocol (i.e. "HTTP/2").
Willy Tarreau95c4e142017-11-26 12:18:55 +0100266However, haproxy natively processes HTTP/1.x requests and headers, so requests
267received over an HTTP/2 connection are transcoded to HTTP/1.1 before being
268processed. This explains why they still appear as "HTTP/1.1" in haproxy's logs
269as well as in server logs.
270
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200271
2721.2.2. The request headers
273--------------------------
274
275The headers start at the second line. They are composed of a name at the
276beginning of the line, immediately followed by a colon (':'). Traditionally,
277an LWS is added after the colon but that's not required. Then come the values.
278Multiple identical headers may be folded into one single line, delimiting the
279values with commas, provided that their order is respected. This is commonly
280encountered in the "Cookie:" field. A header may span over multiple lines if
281the subsequent lines begin with an LWS. In the example in 1.2, lines 4 and 5
282define a total of 3 values for the "Accept:" header.
283
Davor Ocelice9ed2812017-12-25 17:49:28 +0100284Contrary to a common misconception, header names are not case-sensitive, and
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200285their values are not either if they refer to other header names (such as the
Willy Tarreau95c4e142017-11-26 12:18:55 +0100286"Connection:" header). In HTTP/2, header names are always sent in lower case,
287as can be seen when running in debug mode.
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200288
289The end of the headers is indicated by the first empty line. People often say
290that it's a double line feed, which is not exact, even if a double line feed
291is one valid form of empty line.
292
293Fortunately, HAProxy takes care of all these complex combinations when indexing
294headers, checking values and counting them, so there is no reason to worry
295about the way they could be written, but it is important not to accuse an
296application of being buggy if it does unusual, valid things.
297
298Important note:
Lukas Tribus23953682017-04-28 13:24:30 +0000299 As suggested by RFC7231, HAProxy normalizes headers by replacing line breaks
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200300 in the middle of headers by LWS in order to join multi-line headers. This
301 is necessary for proper analysis and helps less capable HTTP parsers to work
302 correctly and not to be fooled by such complex constructs.
303
304
3051.3. HTTP response
306------------------
307
308An HTTP response looks very much like an HTTP request. Both are called HTTP
309messages. Let's consider this HTTP response :
310
311 Line Contents
Willy Tarreaud72758d2010-01-12 10:42:19 +0100312 number
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200313 1 HTTP/1.1 200 OK
314 2 Content-length: 350
315 3 Content-Type: text/html
316
Willy Tarreau816b9792009-09-15 21:25:21 +0200317As a special case, HTTP supports so called "Informational responses" as status
318codes 1xx. These messages are special in that they don't convey any part of the
319response, they're just used as sort of a signaling message to ask a client to
Willy Tarreau5843d1a2010-02-01 15:13:32 +0100320continue to post its request for instance. In the case of a status 100 response
321the requested information will be carried by the next non-100 response message
322following the informational one. This implies that multiple responses may be
323sent to a single request, and that this only works when keep-alive is enabled
324(1xx messages are HTTP/1.1 only). HAProxy handles these messages and is able to
325correctly forward and skip them, and only process the next non-100 response. As
326such, these messages are neither logged nor transformed, unless explicitly
327state otherwise. Status 101 messages indicate that the protocol is changing
328over the same connection and that haproxy must switch to tunnel mode, just as
329if a CONNECT had occurred. Then the Upgrade header would contain additional
330information about the type of protocol the connection is switching to.
Willy Tarreau816b9792009-09-15 21:25:21 +0200331
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200332
Davor Ocelice9ed2812017-12-25 17:49:28 +01003331.3.1. The response line
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200334------------------------
335
336Line 1 is the "response line". It is always composed of 3 fields :
337
338 - a version tag : HTTP/1.1
339 - a status code : 200
340 - a reason : OK
341
342The status code is always 3-digit. The first digit indicates a general status :
Davor Ocelice9ed2812017-12-25 17:49:28 +0100343 - 1xx = informational message to be skipped (e.g. 100, 101)
344 - 2xx = OK, content is following (e.g. 200, 206)
345 - 3xx = OK, no content following (e.g. 302, 304)
346 - 4xx = error caused by the client (e.g. 401, 403, 404)
347 - 5xx = error caused by the server (e.g. 500, 502, 503)
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200348
Lukas Tribus23953682017-04-28 13:24:30 +0000349Please refer to RFC7231 for the detailed meaning of all such codes. The
Willy Tarreaud72758d2010-01-12 10:42:19 +0100350"reason" field is just a hint, but is not parsed by clients. Anything can be
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200351found there, but it's a common practice to respect the well-established
352messages. It can be composed of one or multiple words, such as "OK", "Found",
353or "Authentication Required".
354
Davor Ocelice9ed2812017-12-25 17:49:28 +0100355HAProxy may emit the following status codes by itself :
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200356
357 Code When / reason
358 200 access to stats page, and when replying to monitoring requests
359 301 when performing a redirection, depending on the configured code
360 302 when performing a redirection, depending on the configured code
361 303 when performing a redirection, depending on the configured code
Willy Tarreaub67fdc42013-03-29 19:28:11 +0100362 307 when performing a redirection, depending on the configured code
363 308 when performing a redirection, depending on the configured code
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200364 400 for an invalid or too large request
365 401 when an authentication is required to perform the action (when
366 accessing the stats page)
367 403 when a request is forbidden by a "block" ACL or "reqdeny" filter
368 408 when the request timeout strikes before the request is complete
369 500 when haproxy encounters an unrecoverable internal error, such as a
370 memory allocation failure, which should never happen
371 502 when the server returns an empty, invalid or incomplete response, or
372 when an "rspdeny" filter blocks the response.
373 503 when no server was available to handle the request, or in response to
374 monitoring requests which match the "monitor fail" condition
375 504 when the response timeout strikes before the server responds
376
377The error 4xx and 5xx codes above may be customized (see "errorloc" in section
3784.2).
379
380
3811.3.2. The response headers
382---------------------------
383
384Response headers work exactly like request headers, and as such, HAProxy uses
385the same parsing function for both. Please refer to paragraph 1.2.2 for more
386details.
387
388
3892. Configuring HAProxy
390----------------------
391
3922.1. Configuration file format
393------------------------------
Willy Tarreau6a06a402007-07-15 20:15:28 +0200394
395HAProxy's configuration process involves 3 major sources of parameters :
396
397 - the arguments from the command-line, which always take precedence
398 - the "global" section, which sets process-wide parameters
399 - the proxies sections which can take form of "defaults", "listen",
400 "frontend" and "backend".
401
Willy Tarreau0ba27502007-12-24 16:55:16 +0100402The configuration file syntax consists in lines beginning with a keyword
403referenced in this manual, optionally followed by one or several parameters
William Lallemandf9873ba2015-05-05 17:37:14 +0200404delimited by spaces.
Willy Tarreau0ba27502007-12-24 16:55:16 +0100405
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200406
William Lallemandf9873ba2015-05-05 17:37:14 +02004072.2. Quoting and escaping
408-------------------------
409
410HAProxy's configuration introduces a quoting and escaping system similar to
411many programming languages. The configuration file supports 3 types: escaping
412with a backslash, weak quoting with double quotes, and strong quoting with
413single quotes.
414
415If spaces have to be entered in strings, then they must be escaped by preceding
416them by a backslash ('\') or by quoting them. Backslashes also have to be
417escaped by doubling or strong quoting them.
418
419Escaping is achieved by preceding a special character by a backslash ('\'):
420
421 \ to mark a space and differentiate it from a delimiter
422 \# to mark a hash and differentiate it from a comment
423 \\ to use a backslash
424 \' to use a single quote and differentiate it from strong quoting
425 \" to use a double quote and differentiate it from weak quoting
426
427Weak quoting is achieved by using double quotes (""). Weak quoting prevents
428the interpretation of:
429
430 space as a parameter separator
431 ' single quote as a strong quoting delimiter
432 # hash as a comment start
433
William Lallemandb2f07452015-05-12 14:27:13 +0200434Weak quoting permits the interpretation of variables, if you want to use a non
435-interpreted dollar within a double quoted string, you should escape it with a
436backslash ("\$"), it does not work outside weak quoting.
437
438Interpretation of escaping and special characters are not prevented by weak
William Lallemandf9873ba2015-05-05 17:37:14 +0200439quoting.
440
441Strong quoting is achieved by using single quotes (''). Inside single quotes,
442nothing is interpreted, it's the efficient way to quote regexes.
443
444Quoted and escaped strings are replaced in memory by their interpreted
445equivalent, it allows you to perform concatenation.
446
447 Example:
448 # those are equivalents:
449 log-format %{+Q}o\ %t\ %s\ %{-Q}r
450 log-format "%{+Q}o %t %s %{-Q}r"
451 log-format '%{+Q}o %t %s %{-Q}r'
452 log-format "%{+Q}o %t"' %s %{-Q}r'
453 log-format "%{+Q}o %t"' %s'\ %{-Q}r
454
455 # those are equivalents:
456 reqrep "^([^\ :]*)\ /static/(.*)" \1\ /\2
457 reqrep "^([^ :]*)\ /static/(.*)" '\1 /\2'
458 reqrep "^([^ :]*)\ /static/(.*)" "\1 /\2"
459 reqrep "^([^ :]*)\ /static/(.*)" "\1\ /\2"
460
461
William Lallemandb2f07452015-05-12 14:27:13 +02004622.3. Environment variables
463--------------------------
464
465HAProxy's configuration supports environment variables. Those variables are
466interpreted only within double quotes. Variables are expanded during the
467configuration parsing. Variable names must be preceded by a dollar ("$") and
468optionally enclosed with braces ("{}") similarly to what is done in Bourne
469shell. Variable names can contain alphanumerical characters or the character
470underscore ("_") but should not start with a digit.
471
472 Example:
473
474 bind "fd@${FD_APP1}"
475
476 log "${LOCAL_SYSLOG}:514" local0 notice # send to local server
477
478 user "$HAPROXY_USER"
479
William Lallemand4d03e432019-06-14 15:35:37 +0200480Some variables are defined by HAProxy, they can be used in the configuration
481file, or could be inherited by a program (See 3.7. Programs):
William Lallemanddaf4cd22018-04-17 16:46:13 +0200482
William Lallemand4d03e432019-06-14 15:35:37 +0200483* HAPROXY_LOCALPEER: defined at the startup of the process which contains the
484 name of the local peer. (See "-L" in the management guide.)
485
486* HAPROXY_CFGFILES: list of the configuration files loaded by HAProxy,
487 separated by semicolons. Can be useful in the case you specified a
488 directory.
489
490* HAPROXY_MWORKER: In master-worker mode, this variable is set to 1.
491
John Roeslerfb2fce12019-07-10 15:45:51 -0500492* HAPROXY_CLI: configured listeners addresses of the stats socket for every
William Lallemand4d03e432019-06-14 15:35:37 +0200493 processes, separated by semicolons.
494
John Roeslerfb2fce12019-07-10 15:45:51 -0500495* HAPROXY_MASTER_CLI: In master-worker mode, listeners addresses of the master
William Lallemand4d03e432019-06-14 15:35:37 +0200496 CLI, separated by semicolons.
497
498See also "external-check command" for other variables.
William Lallemandb2f07452015-05-12 14:27:13 +0200499
5002.4. Time format
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200501----------------
502
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +0100503Some parameters involve values representing time, such as timeouts. These
Willy Tarreau0ba27502007-12-24 16:55:16 +0100504values are generally expressed in milliseconds (unless explicitly stated
505otherwise) but may be expressed in any other unit by suffixing the unit to the
506numeric value. It is important to consider this because it will not be repeated
507for every keyword. Supported units are :
508
509 - us : microseconds. 1 microsecond = 1/1000000 second
510 - ms : milliseconds. 1 millisecond = 1/1000 second. This is the default.
511 - s : seconds. 1s = 1000ms
512 - m : minutes. 1m = 60s = 60000ms
513 - h : hours. 1h = 60m = 3600s = 3600000ms
514 - d : days. 1d = 24h = 1440m = 86400s = 86400000ms
515
516
Lukas Tribusaa83a312017-03-21 09:25:09 +00005172.5. Examples
Patrick Mezard35da19c2010-06-12 17:02:47 +0200518-------------
519
520 # Simple configuration for an HTTP proxy listening on port 80 on all
521 # interfaces and forwarding requests to a single backend "servers" with a
522 # single server "server1" listening on 127.0.0.1:8000
523 global
524 daemon
525 maxconn 256
526
527 defaults
528 mode http
529 timeout connect 5000ms
530 timeout client 50000ms
531 timeout server 50000ms
532
533 frontend http-in
534 bind *:80
535 default_backend servers
536
537 backend servers
538 server server1 127.0.0.1:8000 maxconn 32
539
540
541 # The same configuration defined with a single listen block. Shorter but
542 # less expressive, especially in HTTP mode.
543 global
544 daemon
545 maxconn 256
546
547 defaults
548 mode http
549 timeout connect 5000ms
550 timeout client 50000ms
551 timeout server 50000ms
552
553 listen http-in
554 bind *:80
555 server server1 127.0.0.1:8000 maxconn 32
556
557
558Assuming haproxy is in $PATH, test these configurations in a shell with:
559
Willy Tarreauccb289d2010-12-11 20:19:38 +0100560 $ sudo haproxy -f configuration.conf -c
Patrick Mezard35da19c2010-06-12 17:02:47 +0200561
562
Willy Tarreauc57f0e22009-05-10 13:12:33 +02005633. Global parameters
Willy Tarreau6a06a402007-07-15 20:15:28 +0200564--------------------
565
566Parameters in the "global" section are process-wide and often OS-specific. They
567are generally set once for all and do not need being changed once correct. Some
568of them have command-line equivalents.
569
570The following keywords are supported in the "global" section :
571
572 * Process management and security
Emeric Brunc8e8d122012-10-02 18:42:10 +0200573 - ca-base
Willy Tarreau6a06a402007-07-15 20:15:28 +0200574 - chroot
Emeric Brunc8e8d122012-10-02 18:42:10 +0200575 - crt-base
Baptiste Assmann3493d0f2015-10-12 20:21:23 +0200576 - cpu-map
Willy Tarreau6a06a402007-07-15 20:15:28 +0200577 - daemon
Baptiste Assmann3493d0f2015-10-12 20:21:23 +0200578 - description
579 - deviceatlas-json-file
580 - deviceatlas-log-level
581 - deviceatlas-separator
582 - deviceatlas-properties-cookie
Simon Horman98637e52014-06-20 12:30:16 +0900583 - external-check
Willy Tarreau6a06a402007-07-15 20:15:28 +0200584 - gid
585 - group
Cyril Bonté203ec5a2017-03-23 22:44:13 +0100586 - hard-stop-after
Willy Tarreau6a06a402007-07-15 20:15:28 +0200587 - log
Baptiste Assmann3493d0f2015-10-12 20:21:23 +0200588 - log-tag
Joe Williamsdf5b38f2010-12-29 17:05:48 +0100589 - log-send-hostname
Baptiste Assmann3493d0f2015-10-12 20:21:23 +0200590 - lua-load
William Lallemand27edc4b2019-05-07 17:49:33 +0200591 - mworker-max-reloads
Willy Tarreau6a06a402007-07-15 20:15:28 +0200592 - nbproc
Christopher Fauletbe0faa22017-08-29 15:37:10 +0200593 - nbthread
Baptiste Assmann3493d0f2015-10-12 20:21:23 +0200594 - node
Willy Tarreau6a06a402007-07-15 20:15:28 +0200595 - pidfile
Willy Tarreau1d549722016-02-16 12:41:57 +0100596 - presetenv
597 - resetenv
Willy Tarreau6a06a402007-07-15 20:15:28 +0200598 - uid
599 - ulimit-n
600 - user
Willy Tarreau636848a2019-04-15 19:38:50 +0200601 - set-dumpable
Willy Tarreau1d549722016-02-16 12:41:57 +0100602 - setenv
Willy Tarreaufbee7132007-10-18 13:53:22 +0200603 - stats
Baptiste Assmann3493d0f2015-10-12 20:21:23 +0200604 - ssl-default-bind-ciphers
Dirkjan Bussink415150f2018-09-14 11:14:21 +0200605 - ssl-default-bind-ciphersuites
Baptiste Assmann3493d0f2015-10-12 20:21:23 +0200606 - ssl-default-bind-options
607 - ssl-default-server-ciphers
Dirkjan Bussink415150f2018-09-14 11:14:21 +0200608 - ssl-default-server-ciphersuites
Baptiste Assmann3493d0f2015-10-12 20:21:23 +0200609 - ssl-default-server-options
610 - ssl-dh-param-file
Emeric Brun850efd52014-01-29 12:24:34 +0100611 - ssl-server-verify
Willy Tarreauceb24bc2010-11-09 12:46:41 +0100612 - unix-bind
Willy Tarreau1d549722016-02-16 12:41:57 +0100613 - unsetenv
Thomas Holmesdb04f192015-05-18 13:21:39 +0100614 - 51degrees-data-file
615 - 51degrees-property-name-list
Dragan Dosen93b38d92015-06-29 16:43:25 +0200616 - 51degrees-property-separator
Dragan Dosenae6d39a2015-06-29 16:43:27 +0200617 - 51degrees-cache-size
Willy Tarreaub3cc9f22019-04-19 16:03:32 +0200618 - wurfl-data-file
619 - wurfl-information-list
620 - wurfl-information-list-separator
Willy Tarreaub3cc9f22019-04-19 16:03:32 +0200621 - wurfl-cache-size
Willy Tarreaud72758d2010-01-12 10:42:19 +0100622
Willy Tarreau6a06a402007-07-15 20:15:28 +0200623 * Performance tuning
Willy Tarreau1746eec2014-04-25 10:46:47 +0200624 - max-spread-checks
Willy Tarreau6a06a402007-07-15 20:15:28 +0200625 - maxconn
Willy Tarreau81c25d02011-09-07 15:17:21 +0200626 - maxconnrate
William Lallemandd85f9172012-11-09 17:05:39 +0100627 - maxcomprate
William Lallemand072a2bf2012-11-20 17:01:01 +0100628 - maxcompcpuusage
Willy Tarreauff4f82d2009-02-06 11:28:13 +0100629 - maxpipes
Willy Tarreau93e7c002013-10-07 18:51:07 +0200630 - maxsessrate
Willy Tarreau403edff2012-09-06 11:58:37 +0200631 - maxsslconn
Willy Tarreaue43d5322013-10-07 20:01:52 +0200632 - maxsslrate
Baptiste Assmann3493d0f2015-10-12 20:21:23 +0200633 - maxzlibmem
Willy Tarreau6a06a402007-07-15 20:15:28 +0200634 - noepoll
635 - nokqueue
Emmanuel Hocdet0ba4f482019-04-08 16:53:32 +0000636 - noevports
Willy Tarreau6a06a402007-07-15 20:15:28 +0200637 - nopoll
Willy Tarreauff4f82d2009-02-06 11:28:13 +0100638 - nosplice
Jarno Huuskonen0e82b922014-04-12 18:22:19 +0300639 - nogetaddrinfo
Lukas Tribusa0bcbdc2016-09-12 21:42:20 +0000640 - noreuseport
Willy Tarreau75c62c22018-11-22 11:02:09 +0100641 - profiling.tasks
Willy Tarreaufe255b72007-10-14 23:09:26 +0200642 - spread-checks
Baptiste Assmann5626f482015-08-23 10:00:10 +0200643 - server-state-base
Baptiste Assmannef1f0fc2015-08-23 10:06:39 +0200644 - server-state-file
Grant Zhang872f9c22017-01-21 01:10:18 +0000645 - ssl-engine
Grant Zhangfa6c7ee2017-01-14 01:42:15 +0000646 - ssl-mode-async
Baptiste Assmann3493d0f2015-10-12 20:21:23 +0200647 - tune.buffers.limit
648 - tune.buffers.reserve
Willy Tarreau27a674e2009-08-17 07:23:33 +0200649 - tune.bufsize
Willy Tarreau43961d52010-10-04 20:39:20 +0200650 - tune.chksize
William Lallemandf3747832012-11-09 12:33:10 +0100651 - tune.comp.maxlevel
Willy Tarreaufe20e5b2017-07-27 11:42:14 +0200652 - tune.h2.header-table-size
Willy Tarreaue6baec02017-07-27 11:45:11 +0200653 - tune.h2.initial-window-size
Willy Tarreau5242ef82017-07-27 11:47:28 +0200654 - tune.h2.max-concurrent-streams
Willy Tarreau193b8c62012-11-22 00:17:38 +0100655 - tune.http.cookielen
Stéphane Cottin23e9e932017-05-18 08:58:41 +0200656 - tune.http.logurilen
Willy Tarreauac1932d2011-10-24 19:14:41 +0200657 - tune.http.maxhdr
Willy Tarreau7e312732014-02-12 16:35:14 +0100658 - tune.idletimer
Thierry FOURNIER90da1912015-03-05 11:17:06 +0100659 - tune.lua.forced-yield
Willy Tarreau32f61e22015-03-18 17:54:59 +0100660 - tune.lua.maxmem
Thierry FOURNIER90da1912015-03-05 11:17:06 +0100661 - tune.lua.session-timeout
662 - tune.lua.task-timeout
Thierry FOURNIER7dd784b2015-10-01 14:49:33 +0200663 - tune.lua.service-timeout
Willy Tarreaua0250ba2008-01-06 11:22:57 +0100664 - tune.maxaccept
665 - tune.maxpollevents
Willy Tarreau27a674e2009-08-17 07:23:33 +0200666 - tune.maxrewrite
Willy Tarreauf3045d22015-04-29 16:24:50 +0200667 - tune.pattern.cache-size
Willy Tarreaubd9a0a72011-10-23 21:14:29 +0200668 - tune.pipesize
Willy Tarreaue803de22010-01-21 17:43:04 +0100669 - tune.rcvbuf.client
670 - tune.rcvbuf.server
Willy Tarreaub22fc302015-12-14 12:04:35 +0100671 - tune.recv_enough
Olivier Houchard1599b802018-05-24 18:59:04 +0200672 - tune.runqueue-depth
Willy Tarreaue803de22010-01-21 17:43:04 +0100673 - tune.sndbuf.client
674 - tune.sndbuf.server
Willy Tarreau6ec58db2012-11-16 16:32:15 +0100675 - tune.ssl.cachesize
Willy Tarreaubfd59462013-02-21 07:46:09 +0100676 - tune.ssl.lifetime
Emeric Brun8dc60392014-05-09 13:52:00 +0200677 - tune.ssl.force-private-cache
Willy Tarreaubfd59462013-02-21 07:46:09 +0100678 - tune.ssl.maxrecord
Remi Gacognef46cd6e2014-06-12 14:58:40 +0200679 - tune.ssl.default-dh-param
Christopher Faulet31af49d2015-06-09 17:29:50 +0200680 - tune.ssl.ssl-ctx-cache-size
Thierry FOURNIER5bf77322017-02-25 12:45:22 +0100681 - tune.ssl.capture-cipherlist-size
Thierry FOURNIER4834bc72015-06-06 19:29:07 +0200682 - tune.vars.global-max-size
Christopher Fauletff2613e2016-11-09 11:36:17 +0100683 - tune.vars.proc-max-size
Thierry FOURNIER4834bc72015-06-06 19:29:07 +0200684 - tune.vars.reqres-max-size
685 - tune.vars.sess-max-size
686 - tune.vars.txn-max-size
William Lallemanda509e4c2012-11-07 16:54:34 +0100687 - tune.zlib.memlevel
688 - tune.zlib.windowsize
Willy Tarreaud72758d2010-01-12 10:42:19 +0100689
Willy Tarreau6a06a402007-07-15 20:15:28 +0200690 * Debugging
691 - debug
692 - quiet
Willy Tarreau6a06a402007-07-15 20:15:28 +0200693
694
Willy Tarreauc57f0e22009-05-10 13:12:33 +02006953.1. Process management and security
Willy Tarreau6a06a402007-07-15 20:15:28 +0200696------------------------------------
697
Emeric Brunc8e8d122012-10-02 18:42:10 +0200698ca-base <dir>
699 Assigns a default directory to fetch SSL CA certificates and CRLs from when a
Emeric Brunfd33a262012-10-11 16:28:27 +0200700 relative path is used with "ca-file" or "crl-file" directives. Absolute
701 locations specified in "ca-file" and "crl-file" prevail and ignore "ca-base".
Emeric Brunc8e8d122012-10-02 18:42:10 +0200702
Willy Tarreau6a06a402007-07-15 20:15:28 +0200703chroot <jail dir>
704 Changes current directory to <jail dir> and performs a chroot() there before
705 dropping privileges. This increases the security level in case an unknown
706 vulnerability would be exploited, since it would make it very hard for the
707 attacker to exploit the system. This only works when the process is started
708 with superuser privileges. It is important to ensure that <jail_dir> is both
Davor Ocelice9ed2812017-12-25 17:49:28 +0100709 empty and non-writable to anyone.
Willy Tarreaud72758d2010-01-12 10:42:19 +0100710
Christopher Fauletcb6a9452017-11-22 16:50:41 +0100711cpu-map [auto:]<process-set>[/<thread-set>] <cpu-set>...
712 On Linux 2.6 and above, it is possible to bind a process or a thread to a
713 specific CPU set. This means that the process or the thread will never run on
714 other CPUs. The "cpu-map" directive specifies CPU sets for process or thread
715 sets. The first argument is a process set, eventually followed by a thread
716 set. These sets have the format
717
718 all | odd | even | number[-[number]]
719
720 <number>> must be a number between 1 and 32 or 64, depending on the machine's
Davor Ocelice9ed2812017-12-25 17:49:28 +0100721 word size. Any process IDs above nbproc and any thread IDs above nbthread are
Christopher Fauletcb6a9452017-11-22 16:50:41 +0100722 ignored. It is possible to specify a range with two such number delimited by
723 a dash ('-'). It also is possible to specify all processes at once using
Christopher Faulet1dcb9cb2017-11-22 10:24:40 +0100724 "all", only odd numbers using "odd" or even numbers using "even", just like
725 with the "bind-process" directive. The second and forthcoming arguments are
Davor Ocelice9ed2812017-12-25 17:49:28 +0100726 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 +0100727 range with two such numbers delimited by a dash ('-'). Multiple CPU numbers
Christopher Fauletcb6a9452017-11-22 16:50:41 +0100728 or ranges may be specified, and the processes or threads will be allowed to
Davor Ocelice9ed2812017-12-25 17:49:28 +0100729 bind to all of them. Obviously, multiple "cpu-map" directives may be
Christopher Fauletcb6a9452017-11-22 16:50:41 +0100730 specified. Each "cpu-map" directive will replace the previous ones when they
731 overlap. A thread will be bound on the intersection of its mapping and the
732 one of the process on which it is attached. If the intersection is null, no
733 specific binding will be set for the thread.
Willy Tarreaufc6c0322012-11-16 16:12:27 +0100734
Christopher Fauletff4121f2017-11-22 16:38:49 +0100735 Ranges can be partially defined. The higher bound can be omitted. In such
736 case, it is replaced by the corresponding maximum value, 32 or 64 depending
737 on the machine's word size.
738
Christopher Faulet26028f62017-11-22 15:01:51 +0100739 The prefix "auto:" can be added before the process set to let HAProxy
Christopher Fauletcb6a9452017-11-22 16:50:41 +0100740 automatically bind a process or a thread to a CPU by incrementing
741 process/thread and CPU sets. To be valid, both sets must have the same
742 size. No matter the declaration order of the CPU sets, it will be bound from
743 the lowest to the highest bound. Having a process and a thread range with the
744 "auto:" prefix is not supported. Only one range is supported, the other one
745 must be a fixed number.
Christopher Faulet26028f62017-11-22 15:01:51 +0100746
747 Examples:
Christopher Fauletcb6a9452017-11-22 16:50:41 +0100748 cpu-map 1-4 0-3 # bind processes 1 to 4 on the first 4 CPUs
749
750 cpu-map 1/all 0-3 # bind all threads of the first process on the
751 # first 4 CPUs
752
753 cpu-map 1- 0- # will be replaced by "cpu-map 1-64 0-63"
754 # or "cpu-map 1-32 0-31" depending on the machine's
755 # word size.
756
Christopher Faulet26028f62017-11-22 15:01:51 +0100757 # all these lines bind the process 1 to the cpu 0, the process 2 to cpu 1
Christopher Fauletcb6a9452017-11-22 16:50:41 +0100758 # and so on.
Christopher Faulet26028f62017-11-22 15:01:51 +0100759 cpu-map auto:1-4 0-3
760 cpu-map auto:1-4 0-1 2-3
761 cpu-map auto:1-4 3 2 1 0
762
Christopher Fauletcb6a9452017-11-22 16:50:41 +0100763 # all these lines bind the thread 1 to the cpu 0, the thread 2 to cpu 1
764 # and so on.
765 cpu-map auto:1/1-4 0-3
766 cpu-map auto:1/1-4 0-1 2-3
767 cpu-map auto:1/1-4 3 2 1 0
768
Davor Ocelice9ed2812017-12-25 17:49:28 +0100769 # bind each process to exactly one CPU using all/odd/even keyword
Christopher Faulet26028f62017-11-22 15:01:51 +0100770 cpu-map auto:all 0-63
771 cpu-map auto:even 0-31
772 cpu-map auto:odd 32-63
773
774 # invalid cpu-map because process and CPU sets have different sizes.
775 cpu-map auto:1-4 0 # invalid
776 cpu-map auto:1 0-3 # invalid
777
Christopher Fauletcb6a9452017-11-22 16:50:41 +0100778 # invalid cpu-map because automatic binding is used with a process range
779 # and a thread range.
780 cpu-map auto:all/all 0 # invalid
781 cpu-map auto:all/1-4 0 # invalid
782 cpu-map auto:1-4/all 0 # invalid
783
Emeric Brunc8e8d122012-10-02 18:42:10 +0200784crt-base <dir>
785 Assigns a default directory to fetch SSL certificates from when a relative
786 path is used with "crtfile" directives. Absolute locations specified after
787 "crtfile" prevail and ignore "crt-base".
788
Willy Tarreau6a06a402007-07-15 20:15:28 +0200789daemon
790 Makes the process fork into background. This is the recommended mode of
791 operation. It is equivalent to the command line "-D" argument. It can be
Lukas Tribusf46bf952017-11-21 12:39:34 +0100792 disabled by the command line "-db" argument. This option is ignored in
793 systemd mode.
Willy Tarreau6a06a402007-07-15 20:15:28 +0200794
David Carlier8167f302015-06-01 13:50:06 +0200795deviceatlas-json-file <path>
796 Sets the path of the DeviceAtlas JSON data file to be loaded by the API.
Davor Ocelice9ed2812017-12-25 17:49:28 +0100797 The path must be a valid JSON data file and accessible by HAProxy process.
David Carlier8167f302015-06-01 13:50:06 +0200798
799deviceatlas-log-level <value>
Davor Ocelice9ed2812017-12-25 17:49:28 +0100800 Sets the level of information returned by the API. This directive is
David Carlier8167f302015-06-01 13:50:06 +0200801 optional and set to 0 by default if not set.
802
803deviceatlas-separator <char>
804 Sets the character separator for the API properties results. This directive
805 is optional and set to | by default if not set.
806
Cyril Bonté0306c4a2015-10-26 22:37:38 +0100807deviceatlas-properties-cookie <name>
Cyril Bonté307ee1e2015-09-28 23:16:06 +0200808 Sets the client cookie's name used for the detection if the DeviceAtlas
809 Client-side component was used during the request. This directive is optional
810 and set to DAPROPS by default if not set.
David Carlier29b3ca32015-09-25 14:09:21 +0100811
Simon Horman98637e52014-06-20 12:30:16 +0900812external-check
813 Allows the use of an external agent to perform health checks.
814 This is disabled by default as a security precaution.
815 See "option external-check".
816
Willy Tarreau6a06a402007-07-15 20:15:28 +0200817gid <number>
818 Changes the process' group ID to <number>. It is recommended that the group
819 ID is dedicated to HAProxy or to a small set of similar daemons. HAProxy must
820 be started with a user belonging to this group, or with superuser privileges.
Michael Schererab012dd2013-01-12 18:35:19 +0100821 Note that if haproxy is started from a user having supplementary groups, it
822 will only be able to drop these groups if started with superuser privileges.
Willy Tarreau6a06a402007-07-15 20:15:28 +0200823 See also "group" and "uid".
Willy Tarreaud72758d2010-01-12 10:42:19 +0100824
Cyril Bonté203ec5a2017-03-23 22:44:13 +0100825hard-stop-after <time>
826 Defines the maximum time allowed to perform a clean soft-stop.
827
828 Arguments :
829 <time> is the maximum time (by default in milliseconds) for which the
830 instance will remain alive when a soft-stop is received via the
831 SIGUSR1 signal.
832
833 This may be used to ensure that the instance will quit even if connections
834 remain opened during a soft-stop (for example with long timeouts for a proxy
835 in tcp mode). It applies both in TCP and HTTP mode.
836
837 Example:
838 global
839 hard-stop-after 30s
840
Willy Tarreau6a06a402007-07-15 20:15:28 +0200841group <group name>
842 Similar to "gid" but uses the GID of group name <group name> from /etc/group.
843 See also "gid" and "user".
Willy Tarreaud72758d2010-01-12 10:42:19 +0100844
Frédéric Lécailled690dfa2019-04-25 10:52:17 +0200845log <address> [len <length>] [format <format>] [sample <ranges>:<smp_size>]
846 <facility> [max level [min level]]
Cyril Bonté3e954872018-03-20 23:30:27 +0100847 Adds a global syslog server. Several global servers can be defined. They
Davor Ocelice9ed2812017-12-25 17:49:28 +0100848 will receive logs for starts and exits, as well as all logs from proxies
Robert Tsai81ae1952007-12-05 10:47:29 +0100849 configured with "log global".
850
851 <address> can be one of:
852
Willy Tarreau2769aa02007-12-27 18:26:09 +0100853 - An IPv4 address optionally followed by a colon and a UDP port. If
Robert Tsai81ae1952007-12-05 10:47:29 +0100854 no port is specified, 514 is used by default (the standard syslog
855 port).
856
David du Colombier24bb5f52011-03-17 10:40:23 +0100857 - An IPv6 address followed by a colon and optionally a UDP port. If
858 no port is specified, 514 is used by default (the standard syslog
859 port).
860
Willy Tarreau5a32ecc2018-11-12 07:34:59 +0100861 - A filesystem path to a datagram UNIX domain socket, keeping in mind
Robert Tsai81ae1952007-12-05 10:47:29 +0100862 considerations for chroot (be sure the path is accessible inside
863 the chroot) and uid/gid (be sure the path is appropriately
Davor Ocelice9ed2812017-12-25 17:49:28 +0100864 writable).
Robert Tsai81ae1952007-12-05 10:47:29 +0100865
Willy Tarreau5a32ecc2018-11-12 07:34:59 +0100866 - A file descriptor number in the form "fd@<number>", which may point
867 to a pipe, terminal, or socket. In this case unbuffered logs are used
868 and one writev() call per log is performed. This is a bit expensive
869 but acceptable for most workloads. Messages sent this way will not be
870 truncated but may be dropped, in which case the DroppedLogs counter
871 will be incremented. The writev() call is atomic even on pipes for
872 messages up to PIPE_BUF size, which POSIX recommends to be at least
873 512 and which is 4096 bytes on most modern operating systems. Any
874 larger message may be interleaved with messages from other processes.
875 Exceptionally for debugging purposes the file descriptor may also be
876 directed to a file, but doing so will significantly slow haproxy down
877 as non-blocking calls will be ignored. Also there will be no way to
878 purge nor rotate this file without restarting the process. Note that
879 the configured syslog format is preserved, so the output is suitable
Willy Tarreauc1b06452018-11-12 11:57:56 +0100880 for use with a TCP syslog server. See also the "short" and "raw"
881 format below.
Willy Tarreau5a32ecc2018-11-12 07:34:59 +0100882
883 - "stdout" / "stderr", which are respectively aliases for "fd@1" and
884 "fd@2", see above.
885
William Lallemandb2f07452015-05-12 14:27:13 +0200886 You may want to reference some environment variables in the address
887 parameter, see section 2.3 about environment variables.
Willy Tarreaudad36a32013-03-11 01:20:04 +0100888
Willy Tarreau18324f52014-06-27 18:10:07 +0200889 <length> is an optional maximum line length. Log lines larger than this value
890 will be truncated before being sent. The reason is that syslog
891 servers act differently on log line length. All servers support the
892 default value of 1024, but some servers simply drop larger lines
893 while others do log them. If a server supports long lines, it may
894 make sense to set this value here in order to avoid truncating long
895 lines. Similarly, if a server drops long lines, it is preferable to
896 truncate them before sending them. Accepted values are 80 to 65535
897 inclusive. The default value of 1024 is generally fine for all
898 standard usages. Some specific cases of long captures or
Davor Ocelice9ed2812017-12-25 17:49:28 +0100899 JSON-formatted logs may require larger values. You may also need to
900 increase "tune.http.logurilen" if your request URIs are truncated.
Willy Tarreau18324f52014-06-27 18:10:07 +0200901
Dragan Dosen7ad31542015-09-28 17:16:47 +0200902 <format> is the log format used when generating syslog messages. It may be
903 one of the following :
904
905 rfc3164 The RFC3164 syslog message format. This is the default.
906 (https://tools.ietf.org/html/rfc3164)
907
908 rfc5424 The RFC5424 syslog message format.
909 (https://tools.ietf.org/html/rfc5424)
910
Willy Tarreaue8746a02018-11-12 08:45:00 +0100911 short A message containing only a level between angle brackets such as
912 '<3>', followed by the text. The PID, date, time, process name
913 and system name are omitted. This is designed to be used with a
914 local log server. This format is compatible with what the systemd
915 logger consumes.
916
Willy Tarreauc1b06452018-11-12 11:57:56 +0100917 raw A message containing only the text. The level, PID, date, time,
918 process name and system name are omitted. This is designed to be
919 used in containers or during development, where the severity only
920 depends on the file descriptor used (stdout/stderr).
921
Frédéric Lécailled690dfa2019-04-25 10:52:17 +0200922 <ranges> A list of comma-separated ranges to identify the logs to sample.
923 This is used to balance the load of the logs to send to the log
924 server. The limits of the ranges cannot be null. They are numbered
925 from 1. The size or period (in number of logs) of the sample must be
926 set with <sample_size> parameter.
927
928 <sample_size>
929 The size of the sample in number of logs to consider when balancing
930 their logging loads. It is used to balance the load of the logs to
931 send to the syslog server. This size must be greater or equal to the
932 maximum of the high limits of the ranges.
933 (see also <ranges> parameter).
934
Robert Tsai81ae1952007-12-05 10:47:29 +0100935 <facility> must be one of the 24 standard syslog facilities :
Willy Tarreau6a06a402007-07-15 20:15:28 +0200936
Willy Tarreaue8746a02018-11-12 08:45:00 +0100937 kern user mail daemon auth syslog lpr news
938 uucp cron auth2 ftp ntp audit alert cron2
939 local0 local1 local2 local3 local4 local5 local6 local7
940
Willy Tarreauc1b06452018-11-12 11:57:56 +0100941 Note that the facility is ignored for the "short" and "raw"
942 formats, but still required as a positional field. It is
943 recommended to use "daemon" in this case to make it clear that
944 it's only supposed to be used locally.
Willy Tarreau6a06a402007-07-15 20:15:28 +0200945
946 An optional level can be specified to filter outgoing messages. By default,
Willy Tarreauf7edefa2009-05-10 17:20:05 +0200947 all messages are sent. If a maximum level is specified, only messages with a
948 severity at least as important as this level will be sent. An optional minimum
949 level can be specified. If it is set, logs emitted with a more severe level
950 than this one will be capped to this level. This is used to avoid sending
951 "emerg" messages on all terminals on some default syslog configurations.
952 Eight levels are known :
Willy Tarreau6a06a402007-07-15 20:15:28 +0200953
Cyril Bontédc4d9032012-04-08 21:57:39 +0200954 emerg alert crit err warning notice info debug
Willy Tarreau6a06a402007-07-15 20:15:28 +0200955
Joe Williamsdf5b38f2010-12-29 17:05:48 +0100956log-send-hostname [<string>]
957 Sets the hostname field in the syslog header. If optional "string" parameter
958 is set the header is set to the string contents, otherwise uses the hostname
959 of the system. Generally used if one is not relaying logs through an
960 intermediate syslog server or for simply customizing the hostname printed in
961 the logs.
962
Kevinm48936af2010-12-22 16:08:21 +0000963log-tag <string>
964 Sets the tag field in the syslog header to this string. It defaults to the
965 program name as launched from the command line, which usually is "haproxy".
966 Sometimes it can be useful to differentiate between multiple processes
Willy Tarreau094af4e2015-01-07 15:03:42 +0100967 running on the same host. See also the per-proxy "log-tag" directive.
Kevinm48936af2010-12-22 16:08:21 +0000968
Thierry FOURNIER90da1912015-03-05 11:17:06 +0100969lua-load <file>
970 This global directive loads and executes a Lua file. This directive can be
971 used multiple times.
972
William Lallemand4cfede82017-11-24 22:02:34 +0100973master-worker [no-exit-on-failure]
William Lallemande202b1e2017-06-01 17:38:56 +0200974 Master-worker mode. It is equivalent to the command line "-W" argument.
975 This mode will launch a "master" which will monitor the "workers". Using
976 this mode, you can reload HAProxy directly by sending a SIGUSR2 signal to
Davor Ocelice9ed2812017-12-25 17:49:28 +0100977 the master. The master-worker mode is compatible either with the foreground
William Lallemande202b1e2017-06-01 17:38:56 +0200978 or daemon mode. It is recommended to use this mode with multiprocess and
979 systemd.
William Lallemand4cfede82017-11-24 22:02:34 +0100980 By default, if a worker exits with a bad return code, in the case of a
981 segfault for example, all workers will be killed, and the master will leave.
982 It is convenient to combine this behavior with Restart=on-failure in a
983 systemd unit file in order to relaunch the whole process. If you don't want
984 this behavior, you must use the keyword "no-exit-on-failure".
William Lallemande202b1e2017-06-01 17:38:56 +0200985
William Lallemand4cfede82017-11-24 22:02:34 +0100986 See also "-W" in the management guide.
William Lallemande202b1e2017-06-01 17:38:56 +0200987
William Lallemand27edc4b2019-05-07 17:49:33 +0200988mworker-max-reloads <number>
989 In master-worker mode, this option limits the number of time a worker can
John Roeslerfb2fce12019-07-10 15:45:51 -0500990 survive to a reload. If the worker did not leave after a reload, once its
William Lallemand27edc4b2019-05-07 17:49:33 +0200991 number of reloads is greater than this number, the worker will receive a
992 SIGTERM. This option helps to keep under control the number of workers.
993 See also "show proc" in the Management Guide.
994
Willy Tarreau6a06a402007-07-15 20:15:28 +0200995nbproc <number>
996 Creates <number> processes when going daemon. This requires the "daemon"
997 mode. By default, only one process is created, which is the recommended mode
998 of operation. For systems limited to small sets of file descriptors per
Willy Tarreau149ab772019-01-26 14:27:06 +0100999 process, it may be needed to fork multiple daemons. When set to a value
1000 larger than 1, threads are automatically disabled. USING MULTIPLE PROCESSES
Willy Tarreau1f672a82019-01-26 14:20:55 +01001001 IS HARDER TO DEBUG AND IS REALLY DISCOURAGED. See also "daemon" and
1002 "nbthread".
Willy Tarreau6a06a402007-07-15 20:15:28 +02001003
Christopher Fauletbe0faa22017-08-29 15:37:10 +02001004nbthread <number>
1005 This setting is only available when support for threads was built in. It
Willy Tarreau26f6ae12019-02-02 12:56:15 +01001006 makes haproxy run on <number> threads. This is exclusive with "nbproc". While
1007 "nbproc" historically used to be the only way to use multiple processors, it
1008 also involved a number of shortcomings related to the lack of synchronization
1009 between processes (health-checks, peers, stick-tables, stats, ...) which do
1010 not affect threads. As such, any modern configuration is strongly encouraged
Willy Tarreau149ab772019-01-26 14:27:06 +01001011 to migrate away from "nbproc" to "nbthread". "nbthread" also works when
1012 HAProxy is started in foreground. On some platforms supporting CPU affinity,
1013 when nbproc is not used, the default "nbthread" value is automatically set to
1014 the number of CPUs the process is bound to upon startup. This means that the
1015 thread count can easily be adjusted from the calling process using commands
1016 like "taskset" or "cpuset". Otherwise, this value defaults to 1. The default
1017 value is reported in the output of "haproxy -vv". See also "nbproc".
Christopher Fauletbe0faa22017-08-29 15:37:10 +02001018
Willy Tarreau6a06a402007-07-15 20:15:28 +02001019pidfile <pidfile>
Davor Ocelice9ed2812017-12-25 17:49:28 +01001020 Writes PIDs of all daemons into file <pidfile>. This option is equivalent to
Willy Tarreau6a06a402007-07-15 20:15:28 +02001021 the "-p" command line argument. The file must be accessible to the user
1022 starting the process. See also "daemon".
1023
Willy Tarreau1d549722016-02-16 12:41:57 +01001024presetenv <name> <value>
1025 Sets environment variable <name> to value <value>. If the variable exists, it
1026 is NOT overwritten. The changes immediately take effect so that the next line
1027 in the configuration file sees the new value. See also "setenv", "resetenv",
1028 and "unsetenv".
1029
1030resetenv [<name> ...]
1031 Removes all environment variables except the ones specified in argument. It
1032 allows to use a clean controlled environment before setting new values with
1033 setenv or unsetenv. Please note that some internal functions may make use of
1034 some environment variables, such as time manipulation functions, but also
1035 OpenSSL or even external checks. This must be used with extreme care and only
1036 after complete validation. The changes immediately take effect so that the
1037 next line in the configuration file sees the new environment. See also
1038 "setenv", "presetenv", and "unsetenv".
1039
Christopher Fauletff4121f2017-11-22 16:38:49 +01001040stats bind-process [ all | odd | even | <process_num>[-[process_num>]] ] ...
Willy Tarreau35b7b162012-10-22 23:17:18 +02001041 Limits the stats socket to a certain set of processes numbers. By default the
1042 stats socket is bound to all processes, causing a warning to be emitted when
1043 nbproc is greater than 1 because there is no way to select the target process
1044 when connecting. However, by using this setting, it becomes possible to pin
1045 the stats socket to a specific set of processes, typically the first one. The
1046 warning will automatically be disabled when this setting is used, whatever
Willy Tarreaua9db57e2013-01-18 11:29:29 +01001047 the number of processes used. The maximum process ID depends on the machine's
Christopher Fauletff4121f2017-11-22 16:38:49 +01001048 word size (32 or 64). Ranges can be partially defined. The higher bound can
1049 be omitted. In such case, it is replaced by the corresponding maximum
1050 value. A better option consists in using the "process" setting of the "stats
1051 socket" line to force the process on each line.
Willy Tarreau35b7b162012-10-22 23:17:18 +02001052
Baptiste Assmann5626f482015-08-23 10:00:10 +02001053server-state-base <directory>
1054 Specifies the directory prefix to be prepended in front of all servers state
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02001055 file names which do not start with a '/'. See also "server-state-file",
1056 "load-server-state-from-file" and "server-state-file-name".
Baptiste Assmannef1f0fc2015-08-23 10:06:39 +02001057
1058server-state-file <file>
1059 Specifies the path to the file containing state of servers. If the path starts
1060 with a slash ('/'), it is considered absolute, otherwise it is considered
1061 relative to the directory specified using "server-state-base" (if set) or to
1062 the current directory. Before reloading HAProxy, it is possible to save the
1063 servers' current state using the stats command "show servers state". The
1064 output of this command must be written in the file pointed by <file>. When
1065 starting up, before handling traffic, HAProxy will read, load and apply state
1066 for each server found in the file and available in its current running
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02001067 configuration. See also "server-state-base" and "show servers state",
1068 "load-server-state-from-file" and "server-state-file-name"
Baptiste Assmann5626f482015-08-23 10:00:10 +02001069
Willy Tarreau1d549722016-02-16 12:41:57 +01001070setenv <name> <value>
1071 Sets environment variable <name> to value <value>. If the variable exists, it
1072 is overwritten. The changes immediately take effect so that the next line in
1073 the configuration file sees the new value. See also "presetenv", "resetenv",
1074 and "unsetenv".
1075
Willy Tarreau636848a2019-04-15 19:38:50 +02001076set-dumpable
1077 This option is better left disabled by default and enabled only upon a
1078 developer's request. It has no impact on performance nor stability but will
1079 try hard to re-enable core dumps that were possibly disabled by file size
1080 limitations (ulimit -f), core size limitations (ulimit -c), or "dumpability"
1081 of a process after changing its UID/GID (such as /proc/sys/fs/suid_dumpable
1082 on Linux). Core dumps might still be limited by the current directory's
1083 permissions (check what directory the file is started from), the chroot
1084 directory's permission (it may be needed to temporarily disable the chroot
1085 directive or to move it to a dedicated writable location), or any other
1086 system-specific constraint. For example, some Linux flavours are notorious
1087 for replacing the default core file with a path to an executable not even
1088 installed on the system (check /proc/sys/kernel/core_pattern). Often, simply
1089 writing "core", "core.%p" or "/var/log/core/core.%p" addresses the issue.
1090 When trying to enable this option waiting for a rare issue to re-appear, it's
1091 often a good idea to first try to obtain such a dump by issuing, for example,
1092 "kill -11" to the haproxy process and verify that it leaves a core where
1093 expected when dying.
1094
Willy Tarreau610f04b2014-02-13 11:36:41 +01001095ssl-default-bind-ciphers <ciphers>
1096 This setting is only available when support for OpenSSL was built in. It sets
1097 the default string describing the list of cipher algorithms ("cipher suite")
Bertrand Jacquin8cf7c1e2019-02-03 18:35:25 +00001098 that are negotiated during the SSL/TLS handshake up to TLSv1.2 for all
Dirkjan Bussink415150f2018-09-14 11:14:21 +02001099 "bind" lines which do not explicitly define theirs. The format of the string
Bertrand Jacquin4f03ab02019-02-03 18:48:49 +00001100 is defined in "man 1 ciphers" from OpenSSL man pages. For background
1101 information and recommendations see e.g.
1102 (https://wiki.mozilla.org/Security/Server_Side_TLS) and
1103 (https://mozilla.github.io/server-side-tls/ssl-config-generator/). For TLSv1.3
1104 cipher configuration, please check the "ssl-default-bind-ciphersuites" keyword.
1105 Please check the "bind" keyword for more information.
Dirkjan Bussink415150f2018-09-14 11:14:21 +02001106
1107ssl-default-bind-ciphersuites <ciphersuites>
1108 This setting is only available when support for OpenSSL was built in and
1109 OpenSSL 1.1.1 or later was used to build HAProxy. It sets the default string
1110 describing the list of cipher algorithms ("cipher suite") that are negotiated
1111 during the TLSv1.3 handshake for all "bind" lines which do not explicitly define
1112 theirs. The format of the string is defined in
Bertrand Jacquin4f03ab02019-02-03 18:48:49 +00001113 "man 1 ciphers" from OpenSSL man pages under the section "ciphersuites". For
1114 cipher configuration for TLSv1.2 and earlier, please check the
1115 "ssl-default-bind-ciphers" keyword. Please check the "bind" keyword for more
Dirkjan Bussink415150f2018-09-14 11:14:21 +02001116 information.
Willy Tarreau610f04b2014-02-13 11:36:41 +01001117
Emeric Brun2c86cbf2014-10-30 15:56:50 +01001118ssl-default-bind-options [<option>]...
1119 This setting is only available when support for OpenSSL was built in. It sets
1120 default ssl-options to force on all "bind" lines. Please check the "bind"
1121 keyword to see available options.
1122
1123 Example:
1124 global
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +02001125 ssl-default-bind-options ssl-min-ver TLSv1.0 no-tls-tickets
Emeric Brun2c86cbf2014-10-30 15:56:50 +01001126
Willy Tarreau610f04b2014-02-13 11:36:41 +01001127ssl-default-server-ciphers <ciphers>
1128 This setting is only available when support for OpenSSL was built in. It
1129 sets the default string describing the list of cipher algorithms that are
Bertrand Jacquin8cf7c1e2019-02-03 18:35:25 +00001130 negotiated during the SSL/TLS handshake up to TLSv1.2 with the server,
Dirkjan Bussink415150f2018-09-14 11:14:21 +02001131 for all "server" lines which do not explicitly define theirs. The format of
Bertrand Jacquin4f03ab02019-02-03 18:48:49 +00001132 the string is defined in "man 1 ciphers" from OpenSSL man pages. For background
1133 information and recommendations see e.g.
1134 (https://wiki.mozilla.org/Security/Server_Side_TLS) and
1135 (https://mozilla.github.io/server-side-tls/ssl-config-generator/).
1136 For TLSv1.3 cipher configuration, please check the
1137 "ssl-default-server-ciphersuites" keyword. Please check the "server" keyword
1138 for more information.
Dirkjan Bussink415150f2018-09-14 11:14:21 +02001139
1140ssl-default-server-ciphersuites <ciphersuites>
1141 This setting is only available when support for OpenSSL was built in and
1142 OpenSSL 1.1.1 or later was used to build HAProxy. It sets the default
1143 string describing the list of cipher algorithms that are negotiated during
1144 the TLSv1.3 handshake with the server, for all "server" lines which do not
1145 explicitly define theirs. The format of the string is defined in
Bertrand Jacquin4f03ab02019-02-03 18:48:49 +00001146 "man 1 ciphers" from OpenSSL man pages under the section "ciphersuites". For
1147 cipher configuration for TLSv1.2 and earlier, please check the
1148 "ssl-default-server-ciphers" keyword. Please check the "server" keyword for
1149 more information.
Willy Tarreau610f04b2014-02-13 11:36:41 +01001150
Emeric Brun2c86cbf2014-10-30 15:56:50 +01001151ssl-default-server-options [<option>]...
1152 This setting is only available when support for OpenSSL was built in. It sets
1153 default ssl-options to force on all "server" lines. Please check the "server"
1154 keyword to see available options.
1155
Remi Gacogne47783ef2015-05-29 15:53:22 +02001156ssl-dh-param-file <file>
1157 This setting is only available when support for OpenSSL was built in. It sets
1158 the default DH parameters that are used during the SSL/TLS handshake when
1159 ephemeral Diffie-Hellman (DHE) key exchange is used, for all "bind" lines
Davor Ocelice9ed2812017-12-25 17:49:28 +01001160 which do not explicitly define theirs. It will be overridden by custom DH
Remi Gacogne47783ef2015-05-29 15:53:22 +02001161 parameters found in a bind certificate file if any. If custom DH parameters
Cyril Bonté307ee1e2015-09-28 23:16:06 +02001162 are not specified either by using ssl-dh-param-file or by setting them
1163 directly in the certificate file, pre-generated DH parameters of the size
1164 specified by tune.ssl.default-dh-param will be used. Custom parameters are
1165 known to be more secure and therefore their use is recommended.
Remi Gacogne47783ef2015-05-29 15:53:22 +02001166 Custom DH parameters may be generated by using the OpenSSL command
1167 "openssl dhparam <size>", where size should be at least 2048, as 1024-bit DH
1168 parameters should not be considered secure anymore.
1169
Emeric Brun850efd52014-01-29 12:24:34 +01001170ssl-server-verify [none|required]
1171 The default behavior for SSL verify on servers side. If specified to 'none',
1172 servers certificates are not verified. The default is 'required' except if
1173 forced using cmdline option '-dV'.
1174
Willy Tarreauabb175f2012-09-24 12:43:26 +02001175stats socket [<address:port>|<path>] [param*]
1176 Binds a UNIX socket to <path> or a TCPv4/v6 address to <address:port>.
1177 Connections to this socket will return various statistics outputs and even
1178 allow some commands to be issued to change some runtime settings. Please
Willy Tarreau1af20c72017-06-23 16:01:14 +02001179 consult section 9.3 "Unix Socket commands" of Management Guide for more
Kevin Decherf949c7202015-10-13 23:26:44 +02001180 details.
Willy Tarreau6162db22009-10-10 17:13:00 +02001181
Willy Tarreauabb175f2012-09-24 12:43:26 +02001182 All parameters supported by "bind" lines are supported, for instance to
1183 restrict access to some users or their access rights. Please consult
1184 section 5.1 for more information.
Willy Tarreaufbee7132007-10-18 13:53:22 +02001185
1186stats timeout <timeout, in milliseconds>
1187 The default timeout on the stats socket is set to 10 seconds. It is possible
1188 to change this value with "stats timeout". The value must be passed in
Willy Tarreaubefdff12007-12-02 22:27:38 +01001189 milliseconds, or be suffixed by a time unit among { us, ms, s, m, h, d }.
Willy Tarreaufbee7132007-10-18 13:53:22 +02001190
1191stats maxconn <connections>
1192 By default, the stats socket is limited to 10 concurrent connections. It is
1193 possible to change this value with "stats maxconn".
1194
Willy Tarreau6a06a402007-07-15 20:15:28 +02001195uid <number>
1196 Changes the process' user ID to <number>. It is recommended that the user ID
1197 is dedicated to HAProxy or to a small set of similar daemons. HAProxy must
1198 be started with superuser privileges in order to be able to switch to another
1199 one. See also "gid" and "user".
1200
1201ulimit-n <number>
1202 Sets the maximum number of per-process file-descriptors to <number>. By
1203 default, it is automatically computed, so it is recommended not to use this
1204 option.
1205
Willy Tarreauceb24bc2010-11-09 12:46:41 +01001206unix-bind [ prefix <prefix> ] [ mode <mode> ] [ user <user> ] [ uid <uid> ]
1207 [ group <group> ] [ gid <gid> ]
1208
1209 Fixes common settings to UNIX listening sockets declared in "bind" statements.
1210 This is mainly used to simplify declaration of those UNIX sockets and reduce
1211 the risk of errors, since those settings are most commonly required but are
1212 also process-specific. The <prefix> setting can be used to force all socket
1213 path to be relative to that directory. This might be needed to access another
1214 component's chroot. Note that those paths are resolved before haproxy chroots
1215 itself, so they are absolute. The <mode>, <user>, <uid>, <group> and <gid>
1216 all have the same meaning as their homonyms used by the "bind" statement. If
1217 both are specified, the "bind" statement has priority, meaning that the
1218 "unix-bind" settings may be seen as process-wide default settings.
1219
Willy Tarreau1d549722016-02-16 12:41:57 +01001220unsetenv [<name> ...]
1221 Removes environment variables specified in arguments. This can be useful to
1222 hide some sensitive information that are occasionally inherited from the
1223 user's environment during some operations. Variables which did not exist are
1224 silently ignored so that after the operation, it is certain that none of
1225 these variables remain. The changes immediately take effect so that the next
1226 line in the configuration file will not see these variables. See also
1227 "setenv", "presetenv", and "resetenv".
1228
Willy Tarreau6a06a402007-07-15 20:15:28 +02001229user <user name>
1230 Similar to "uid" but uses the UID of user name <user name> from /etc/passwd.
1231 See also "uid" and "group".
1232
Krzysztof Piotr Oledzki48cb2ae2009-10-02 22:51:14 +02001233node <name>
1234 Only letters, digits, hyphen and underscore are allowed, like in DNS names.
1235
1236 This statement is useful in HA configurations where two or more processes or
1237 servers share the same IP address. By setting a different node-name on all
1238 nodes, it becomes easy to immediately spot what server is handling the
1239 traffic.
1240
1241description <text>
1242 Add a text that describes the instance.
1243
1244 Please note that it is required to escape certain characters (# for example)
1245 and this text is inserted into a html page so you should avoid using
1246 "<" and ">" characters.
1247
Thomas Holmesdb04f192015-05-18 13:21:39 +0100124851degrees-data-file <file path>
1249 The path of the 51Degrees data file to provide device detection services. The
Davor Ocelice9ed2812017-12-25 17:49:28 +01001250 file should be unzipped and accessible by HAProxy with relevant permissions.
Thomas Holmesdb04f192015-05-18 13:21:39 +01001251
Dragan Dosenae6d39a2015-06-29 16:43:27 +02001252 Please note that this option is only available when haproxy has been
Thomas Holmesdb04f192015-05-18 13:21:39 +01001253 compiled with USE_51DEGREES.
1254
Ben Shillitof25e8e52016-12-02 14:25:37 +0000125551degrees-property-name-list [<string> ...]
Thomas Holmesdb04f192015-05-18 13:21:39 +01001256 A list of 51Degrees property names to be load from the dataset. A full list
1257 of names is available on the 51Degrees website:
1258 https://51degrees.com/resources/property-dictionary
1259
Dragan Dosenae6d39a2015-06-29 16:43:27 +02001260 Please note that this option is only available when haproxy has been
Thomas Holmesdb04f192015-05-18 13:21:39 +01001261 compiled with USE_51DEGREES.
1262
Dragan Dosen93b38d92015-06-29 16:43:25 +0200126351degrees-property-separator <char>
Thomas Holmesdb04f192015-05-18 13:21:39 +01001264 A char that will be appended to every property value in a response header
1265 containing 51Degrees results. If not set that will be set as ','.
1266
Dragan Dosenae6d39a2015-06-29 16:43:27 +02001267 Please note that this option is only available when haproxy has been
1268 compiled with USE_51DEGREES.
1269
127051degrees-cache-size <number>
1271 Sets the size of the 51Degrees converter cache to <number> entries. This
1272 is an LRU cache which reminds previous device detections and their results.
1273 By default, this cache is disabled.
1274
1275 Please note that this option is only available when haproxy has been
Thomas Holmesdb04f192015-05-18 13:21:39 +01001276 compiled with USE_51DEGREES.
1277
Willy Tarreaub3cc9f22019-04-19 16:03:32 +02001278wurfl-data-file <file path>
1279 The path of the WURFL data file to provide device detection services. The
1280 file should be accessible by HAProxy with relevant permissions.
1281
1282 Please note that this option is only available when haproxy has been compiled
1283 with USE_WURFL=1.
1284
1285wurfl-information-list [<capability>]*
1286 A space-delimited list of WURFL capabilities, virtual capabilities, property
1287 names we plan to use in injected headers. A full list of capability and
1288 virtual capability names is available on the Scientiamobile website :
1289
1290 https://www.scientiamobile.com/wurflCapability
1291
1292 Valid WURFL properties are:
1293 - wurfl_id Contains the device ID of the matched device.
1294
1295 - wurfl_root_id Contains the device root ID of the matched
1296 device.
1297
1298 - wurfl_isdevroot Tells if the matched device is a root device.
1299 Possible values are "TRUE" or "FALSE".
1300
1301 - wurfl_useragent The original useragent coming with this
1302 particular web request.
1303
1304 - wurfl_api_version Contains a string representing the currently
1305 used Libwurfl API version.
1306
Willy Tarreaub3cc9f22019-04-19 16:03:32 +02001307 - wurfl_info A string containing information on the parsed
1308 wurfl.xml and its full path.
1309
1310 - wurfl_last_load_time Contains the UNIX timestamp of the last time
1311 WURFL has been loaded successfully.
1312
1313 - wurfl_normalized_useragent The normalized useragent.
1314
Willy Tarreaub3cc9f22019-04-19 16:03:32 +02001315 Please note that this option is only available when haproxy has been compiled
1316 with USE_WURFL=1.
1317
1318wurfl-information-list-separator <char>
1319 A char that will be used to separate values in a response header containing
1320 WURFL results. If not set that a comma (',') will be used by default.
1321
1322 Please note that this option is only available when haproxy has been compiled
1323 with USE_WURFL=1.
1324
1325wurfl-patch-file [<file path>]
1326 A list of WURFL patch file paths. Note that patches are loaded during startup
1327 thus before the chroot.
1328
1329 Please note that this option is only available when haproxy has been compiled
1330 with USE_WURFL=1.
1331
paulborilebad132c2019-04-18 11:57:04 +02001332wurfl-cache-size <size>
1333 Sets the WURFL Useragent cache size. For faster lookups, already processed user
1334 agents are kept in a LRU cache :
Willy Tarreaub3cc9f22019-04-19 16:03:32 +02001335 - "0" : no cache is used.
paulborilebad132c2019-04-18 11:57:04 +02001336 - <size> : size of lru cache in elements.
Willy Tarreaub3cc9f22019-04-19 16:03:32 +02001337
1338 Please note that this option is only available when haproxy has been compiled
1339 with USE_WURFL=1.
1340
Willy Tarreauc57f0e22009-05-10 13:12:33 +020013413.2. Performance tuning
Willy Tarreau6a06a402007-07-15 20:15:28 +02001342-----------------------
1343
Willy Tarreaubeb859a2018-11-22 18:07:59 +01001344busy-polling
1345 In some situations, especially when dealing with low latency on processors
1346 supporting a variable frequency or when running inside virtual machines, each
1347 time the process waits for an I/O using the poller, the processor goes back
1348 to sleep or is offered to another VM for a long time, and it causes
1349 excessively high latencies. This option provides a solution preventing the
1350 processor from sleeping by always using a null timeout on the pollers. This
1351 results in a significant latency reduction (30 to 100 microseconds observed)
1352 at the expense of a risk to overheat the processor. It may even be used with
1353 threads, in which case improperly bound threads may heavily conflict,
1354 resulting in a worse performance and high values for the CPU stolen fields
1355 in "show info" output, indicating which threads are misconfigured. It is
1356 important not to let the process run on the same processor as the network
1357 interrupts when this option is used. It is also better to avoid using it on
1358 multiple CPU threads sharing the same core. This option is disabled by
1359 default. If it has been enabled, it may still be forcibly disabled by
1360 prefixing it with the "no" keyword. It is ignored by the "select" and
1361 "poll" pollers.
1362
Willy Tarreau1746eec2014-04-25 10:46:47 +02001363max-spread-checks <delay in milliseconds>
1364 By default, haproxy tries to spread the start of health checks across the
1365 smallest health check interval of all the servers in a farm. The principle is
1366 to avoid hammering services running on the same server. But when using large
1367 check intervals (10 seconds or more), the last servers in the farm take some
1368 time before starting to be tested, which can be a problem. This parameter is
1369 used to enforce an upper bound on delay between the first and the last check,
1370 even if the servers' check intervals are larger. When servers run with
1371 shorter intervals, their intervals will be respected though.
1372
Willy Tarreau6a06a402007-07-15 20:15:28 +02001373maxconn <number>
1374 Sets the maximum per-process number of concurrent connections to <number>. It
1375 is equivalent to the command-line argument "-n". Proxies will stop accepting
1376 connections when this limit is reached. The "ulimit-n" parameter is
Willy Tarreau8274e102014-06-19 15:31:25 +02001377 automatically adjusted according to this value. See also "ulimit-n". Note:
1378 the "select" poller cannot reliably use more than 1024 file descriptors on
1379 some platforms. If your platform only supports select and reports "select
1380 FAILED" on startup, you need to reduce maxconn until it works (slightly
Willy Tarreaub28f3442019-03-04 08:13:43 +01001381 below 500 in general). If this value is not set, it will automatically be
1382 calculated based on the current file descriptors limit reported by the
1383 "ulimit -n" command, possibly reduced to a lower value if a memory limit
1384 is enforced, based on the buffer size, memory allocated to compression, SSL
1385 cache size, and use or not of SSL and the associated maxsslconn (which can
1386 also be automatic).
Willy Tarreau6a06a402007-07-15 20:15:28 +02001387
Willy Tarreau81c25d02011-09-07 15:17:21 +02001388maxconnrate <number>
1389 Sets the maximum per-process number of connections per second to <number>.
1390 Proxies will stop accepting connections when this limit is reached. It can be
1391 used to limit the global capacity regardless of each frontend capacity. It is
1392 important to note that this can only be used as a service protection measure,
1393 as there will not necessarily be a fair share between frontends when the
1394 limit is reached, so it's a good idea to also limit each frontend to some
1395 value close to its expected share. Also, lowering tune.maxaccept can improve
1396 fairness.
1397
William Lallemandd85f9172012-11-09 17:05:39 +01001398maxcomprate <number>
1399 Sets the maximum per-process input compression rate to <number> kilobytes
Davor Ocelice9ed2812017-12-25 17:49:28 +01001400 per second. For each session, if the maximum is reached, the compression
William Lallemandd85f9172012-11-09 17:05:39 +01001401 level will be decreased during the session. If the maximum is reached at the
1402 beginning of a session, the session will not compress at all. If the maximum
1403 is not reached, the compression level will be increased up to
Davor Ocelice9ed2812017-12-25 17:49:28 +01001404 tune.comp.maxlevel. A value of zero means there is no limit, this is the
William Lallemandd85f9172012-11-09 17:05:39 +01001405 default value.
1406
William Lallemand072a2bf2012-11-20 17:01:01 +01001407maxcompcpuusage <number>
1408 Sets the maximum CPU usage HAProxy can reach before stopping the compression
1409 for new requests or decreasing the compression level of current requests.
1410 It works like 'maxcomprate' but measures CPU usage instead of incoming data
1411 bandwidth. The value is expressed in percent of the CPU used by haproxy. In
1412 case of multiple processes (nbproc > 1), each process manages its individual
1413 usage. A value of 100 disable the limit. The default value is 100. Setting
1414 a lower value will prevent the compression work from slowing the whole
1415 process down and from introducing high latencies.
1416
Willy Tarreauff4f82d2009-02-06 11:28:13 +01001417maxpipes <number>
1418 Sets the maximum per-process number of pipes to <number>. Currently, pipes
1419 are only used by kernel-based tcp splicing. Since a pipe contains two file
1420 descriptors, the "ulimit-n" value will be increased accordingly. The default
1421 value is maxconn/4, which seems to be more than enough for most heavy usages.
1422 The splice code dynamically allocates and releases pipes, and can fall back
1423 to standard copy, so setting this value too low may only impact performance.
1424
Willy Tarreau93e7c002013-10-07 18:51:07 +02001425maxsessrate <number>
1426 Sets the maximum per-process number of sessions per second to <number>.
1427 Proxies will stop accepting connections when this limit is reached. It can be
1428 used to limit the global capacity regardless of each frontend capacity. It is
1429 important to note that this can only be used as a service protection measure,
1430 as there will not necessarily be a fair share between frontends when the
1431 limit is reached, so it's a good idea to also limit each frontend to some
1432 value close to its expected share. Also, lowering tune.maxaccept can improve
1433 fairness.
1434
Willy Tarreau403edff2012-09-06 11:58:37 +02001435maxsslconn <number>
1436 Sets the maximum per-process number of concurrent SSL connections to
1437 <number>. By default there is no SSL-specific limit, which means that the
1438 global maxconn setting will apply to all connections. Setting this limit
1439 avoids having openssl use too much memory and crash when malloc returns NULL
1440 (since it unfortunately does not reliably check for such conditions). Note
1441 that the limit applies both to incoming and outgoing connections, so one
1442 connection which is deciphered then ciphered accounts for 2 SSL connections.
Willy Tarreaud0256482015-01-15 21:45:22 +01001443 If this value is not set, but a memory limit is enforced, this value will be
1444 automatically computed based on the memory limit, maxconn, the buffer size,
1445 memory allocated to compression, SSL cache size, and use of SSL in either
1446 frontends, backends or both. If neither maxconn nor maxsslconn are specified
1447 when there is a memory limit, haproxy will automatically adjust these values
1448 so that 100% of the connections can be made over SSL with no risk, and will
1449 consider the sides where it is enabled (frontend, backend, both).
Willy Tarreau403edff2012-09-06 11:58:37 +02001450
Willy Tarreaue43d5322013-10-07 20:01:52 +02001451maxsslrate <number>
1452 Sets the maximum per-process number of SSL sessions per second to <number>.
1453 SSL listeners will stop accepting connections when this limit is reached. It
1454 can be used to limit the global SSL CPU usage regardless of each frontend
1455 capacity. It is important to note that this can only be used as a service
1456 protection measure, as there will not necessarily be a fair share between
1457 frontends when the limit is reached, so it's a good idea to also limit each
1458 frontend to some value close to its expected share. It is also important to
1459 note that the sessions are accounted before they enter the SSL stack and not
1460 after, which also protects the stack against bad handshakes. Also, lowering
1461 tune.maxaccept can improve fairness.
1462
William Lallemand9d5f5482012-11-07 16:12:57 +01001463maxzlibmem <number>
1464 Sets the maximum amount of RAM in megabytes per process usable by the zlib.
1465 When the maximum amount is reached, future sessions will not compress as long
1466 as RAM is unavailable. When sets to 0, there is no limit.
William Lallemande3a7d992012-11-20 11:25:20 +01001467 The default value is 0. The value is available in bytes on the UNIX socket
1468 with "show info" on the line "MaxZlibMemUsage", the memory used by zlib is
1469 "ZlibMemUsage" in bytes.
1470
Willy Tarreau6a06a402007-07-15 20:15:28 +02001471noepoll
1472 Disables the use of the "epoll" event polling system on Linux. It is
1473 equivalent to the command-line argument "-de". The next polling system
Willy Tarreaue9f49e72012-11-11 17:42:00 +01001474 used will generally be "poll". See also "nopoll".
Willy Tarreau6a06a402007-07-15 20:15:28 +02001475
1476nokqueue
1477 Disables the use of the "kqueue" event polling system on BSD. It is
1478 equivalent to the command-line argument "-dk". The next polling system
1479 used will generally be "poll". See also "nopoll".
1480
Emmanuel Hocdet0ba4f482019-04-08 16:53:32 +00001481noevports
1482 Disables the use of the event ports event polling system on SunOS systems
1483 derived from Solaris 10 and later. It is equivalent to the command-line
1484 argument "-dv". The next polling system used will generally be "poll". See
1485 also "nopoll".
1486
Willy Tarreau6a06a402007-07-15 20:15:28 +02001487nopoll
1488 Disables the use of the "poll" event polling system. It is equivalent to the
1489 command-line argument "-dp". The next polling system used will be "select".
Willy Tarreau0ba27502007-12-24 16:55:16 +01001490 It should never be needed to disable "poll" since it's available on all
Emmanuel Hocdet0ba4f482019-04-08 16:53:32 +00001491 platforms supported by HAProxy. See also "nokqueue", "noepoll" and
1492 "noevports".
Willy Tarreau6a06a402007-07-15 20:15:28 +02001493
Willy Tarreauff4f82d2009-02-06 11:28:13 +01001494nosplice
1495 Disables the use of kernel tcp splicing between sockets on Linux. It is
Davor Ocelice9ed2812017-12-25 17:49:28 +01001496 equivalent to the command line argument "-dS". Data will then be copied
Willy Tarreauff4f82d2009-02-06 11:28:13 +01001497 using conventional and more portable recv/send calls. Kernel tcp splicing is
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01001498 limited to some very recent instances of kernel 2.6. Most versions between
Willy Tarreauff4f82d2009-02-06 11:28:13 +01001499 2.6.25 and 2.6.28 are buggy and will forward corrupted data, so they must not
1500 be used. This option makes it easier to globally disable kernel splicing in
1501 case of doubt. See also "option splice-auto", "option splice-request" and
1502 "option splice-response".
1503
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03001504nogetaddrinfo
1505 Disables the use of getaddrinfo(3) for name resolving. It is equivalent to
1506 the command line argument "-dG". Deprecated gethostbyname(3) will be used.
1507
Lukas Tribusa0bcbdc2016-09-12 21:42:20 +00001508noreuseport
1509 Disables the use of SO_REUSEPORT - see socket(7). It is equivalent to the
1510 command line argument "-dR".
1511
Willy Tarreaud2d33482019-04-25 17:09:07 +02001512profiling.tasks { auto | on | off }
1513 Enables ('on') or disables ('off') per-task CPU profiling. When set to 'auto'
1514 the profiling automatically turns on a thread when it starts to suffer from
1515 an average latency of 1000 microseconds or higher as reported in the
1516 "avg_loop_us" activity field, and automatically turns off when the latency
John Roeslerfb2fce12019-07-10 15:45:51 -05001517 returns below 990 microseconds (this value is an average over the last 1024
Willy Tarreaud2d33482019-04-25 17:09:07 +02001518 loops so it does not vary quickly and tends to significantly smooth short
1519 spikes). It may also spontaneously trigger from time to time on overloaded
1520 systems, containers, or virtual machines, or when the system swaps (which
1521 must absolutely never happen on a load balancer).
1522
1523 CPU profiling per task can be very convenient to report where the time is
1524 spent and which requests have what effect on which other request. Enabling
1525 it will typically affect the overall's performance by less than 1%, thus it
1526 is recommended to leave it to the default 'auto' value so that it only
1527 operates when a problem is identified. This feature requires a system
Willy Tarreau75c62c22018-11-22 11:02:09 +01001528 supporting the clock_gettime(2) syscall with clock identifiers
1529 CLOCK_MONOTONIC and CLOCK_THREAD_CPUTIME_ID, otherwise the reported time will
1530 be zero. This option may be changed at run time using "set profiling" on the
1531 CLI.
1532
Willy Tarreaufe255b72007-10-14 23:09:26 +02001533spread-checks <0..50, in percent>
Simon Hormand60d6912013-11-25 10:46:36 +09001534 Sometimes it is desirable to avoid sending agent and health checks to
1535 servers at exact intervals, for instance when many logical servers are
1536 located on the same physical server. With the help of this parameter, it
1537 becomes possible to add some randomness in the check interval between 0
1538 and +/- 50%. A value between 2 and 5 seems to show good results. The
1539 default value remains at 0.
Willy Tarreaufe255b72007-10-14 23:09:26 +02001540
Davor Ocelice9ed2812017-12-25 17:49:28 +01001541ssl-engine <name> [algo <comma-separated list of algorithms>]
Grant Zhang872f9c22017-01-21 01:10:18 +00001542 Sets the OpenSSL engine to <name>. List of valid values for <name> may be
Davor Ocelice9ed2812017-12-25 17:49:28 +01001543 obtained using the command "openssl engine". This statement may be used
Grant Zhang872f9c22017-01-21 01:10:18 +00001544 multiple times, it will simply enable multiple crypto engines. Referencing an
1545 unsupported engine will prevent haproxy from starting. Note that many engines
1546 will lead to lower HTTPS performance than pure software with recent
1547 processors. The optional command "algo" sets the default algorithms an ENGINE
1548 will supply using the OPENSSL function ENGINE_set_default_string(). A value
Davor Ocelice9ed2812017-12-25 17:49:28 +01001549 of "ALL" uses the engine for all cryptographic operations. If no list of
1550 algo is specified then the value of "ALL" is used. A comma-separated list
Grant Zhang872f9c22017-01-21 01:10:18 +00001551 of different algorithms may be specified, including: RSA, DSA, DH, EC, RAND,
1552 CIPHERS, DIGESTS, PKEY, PKEY_CRYPTO, PKEY_ASN1. This is the same format that
1553 openssl configuration file uses:
1554 https://www.openssl.org/docs/man1.0.2/apps/config.html
1555
Grant Zhangfa6c7ee2017-01-14 01:42:15 +00001556ssl-mode-async
1557 Adds SSL_MODE_ASYNC mode to the SSL context. This enables asynchronous TLS
Emeric Brun3854e012017-05-17 20:42:48 +02001558 I/O operations if asynchronous capable SSL engines are used. The current
Emeric Brunb5e42a82017-06-06 12:35:14 +00001559 implementation supports a maximum of 32 engines. The Openssl ASYNC API
1560 doesn't support moving read/write buffers and is not compliant with
1561 haproxy's buffer management. So the asynchronous mode is disabled on
John Roeslerfb2fce12019-07-10 15:45:51 -05001562 read/write operations (it is only enabled during initial and renegotiation
Emeric Brunb5e42a82017-06-06 12:35:14 +00001563 handshakes).
Grant Zhangfa6c7ee2017-01-14 01:42:15 +00001564
Willy Tarreau33cb0652014-12-23 22:52:37 +01001565tune.buffers.limit <number>
1566 Sets a hard limit on the number of buffers which may be allocated per process.
1567 The default value is zero which means unlimited. The minimum non-zero value
1568 will always be greater than "tune.buffers.reserve" and should ideally always
1569 be about twice as large. Forcing this value can be particularly useful to
1570 limit the amount of memory a process may take, while retaining a sane
Davor Ocelice9ed2812017-12-25 17:49:28 +01001571 behavior. When this limit is reached, sessions which need a buffer wait for
Willy Tarreau33cb0652014-12-23 22:52:37 +01001572 another one to be released by another session. Since buffers are dynamically
1573 allocated and released, the waiting time is very short and not perceptible
1574 provided that limits remain reasonable. In fact sometimes reducing the limit
1575 may even increase performance by increasing the CPU cache's efficiency. Tests
1576 have shown good results on average HTTP traffic with a limit to 1/10 of the
1577 expected global maxconn setting, which also significantly reduces memory
1578 usage. The memory savings come from the fact that a number of connections
1579 will not allocate 2*tune.bufsize. It is best not to touch this value unless
1580 advised to do so by an haproxy core developer.
1581
Willy Tarreau1058ae72014-12-23 22:40:40 +01001582tune.buffers.reserve <number>
1583 Sets the number of buffers which are pre-allocated and reserved for use only
1584 during memory shortage conditions resulting in failed memory allocations. The
1585 minimum value is 2 and is also the default. There is no reason a user would
1586 want to change this value, it's mostly aimed at haproxy core developers.
1587
Willy Tarreau27a674e2009-08-17 07:23:33 +02001588tune.bufsize <number>
1589 Sets the buffer size to this size (in bytes). Lower values allow more
1590 sessions to coexist in the same amount of RAM, and higher values allow some
1591 applications with very large cookies to work. The default value is 16384 and
1592 can be changed at build time. It is strongly recommended not to change this
1593 from the default value, as very low values will break some services such as
1594 statistics, and values larger than default size will increase memory usage,
1595 possibly causing the system to run out of memory. At least the global maxconn
Willy Tarreau45a66cc2017-11-24 11:28:00 +01001596 parameter should be decreased by the same factor as this one is increased. In
1597 addition, use of HTTP/2 mandates that this value must be 16384 or more. If an
1598 HTTP request is larger than (tune.bufsize - tune.maxrewrite), haproxy will
Dmitry Sivachenkof6f4f7b2012-10-21 18:10:25 +04001599 return HTTP 400 (Bad Request) error. Similarly if an HTTP response is larger
Willy Tarreauc77d3642018-12-12 06:19:42 +01001600 than this size, haproxy will return HTTP 502 (Bad Gateway). Note that the
1601 value set using this parameter will automatically be rounded up to the next
1602 multiple of 8 on 32-bit machines and 16 on 64-bit machines.
Willy Tarreau27a674e2009-08-17 07:23:33 +02001603
Willy Tarreau43961d52010-10-04 20:39:20 +02001604tune.chksize <number>
1605 Sets the check buffer size to this size (in bytes). Higher values may help
1606 find string or regex patterns in very large pages, though doing so may imply
1607 more memory and CPU usage. The default value is 16384 and can be changed at
1608 build time. It is not recommended to change this value, but to use better
1609 checks whenever possible.
1610
William Lallemandf3747832012-11-09 12:33:10 +01001611tune.comp.maxlevel <number>
1612 Sets the maximum compression level. The compression level affects CPU
1613 usage during compression. This value affects CPU usage during compression.
1614 Each session using compression initializes the compression algorithm with
1615 this value. The default value is 1.
1616
Willy Tarreauc299e1e2019-02-27 11:35:12 +01001617tune.fail-alloc
1618 If compiled with DEBUG_FAIL_ALLOC, gives the percentage of chances an
1619 allocation attempt fails. Must be between 0 (no failure) and 100 (no
1620 success). This is useful to debug and make sure memory failures are handled
1621 gracefully.
1622
Willy Tarreaufe20e5b2017-07-27 11:42:14 +02001623tune.h2.header-table-size <number>
1624 Sets the HTTP/2 dynamic header table size. It defaults to 4096 bytes and
1625 cannot be larger than 65536 bytes. A larger value may help certain clients
1626 send more compact requests, depending on their capabilities. This amount of
1627 memory is consumed for each HTTP/2 connection. It is recommended not to
1628 change it.
1629
Willy Tarreaue6baec02017-07-27 11:45:11 +02001630tune.h2.initial-window-size <number>
1631 Sets the HTTP/2 initial window size, which is the number of bytes the client
Davor Ocelice9ed2812017-12-25 17:49:28 +01001632 can upload before waiting for an acknowledgment from haproxy. This setting
1633 only affects payload contents (i.e. the body of POST requests), not headers.
Willy Tarreaue6baec02017-07-27 11:45:11 +02001634 The default value is 65535, which roughly allows up to 5 Mbps of upload
1635 bandwidth per client over a network showing a 100 ms ping time, or 500 Mbps
1636 over a 1-ms local network. It can make sense to increase this value to allow
1637 faster uploads, or to reduce it to increase fairness when dealing with many
1638 clients. It doesn't affect resource usage.
1639
Willy Tarreau5242ef82017-07-27 11:47:28 +02001640tune.h2.max-concurrent-streams <number>
1641 Sets the HTTP/2 maximum number of concurrent streams per connection (ie the
1642 number of outstanding requests on a single connection). The default value is
1643 100. A larger one may slightly improve page load time for complex sites when
1644 visited over high latency networks, but increases the amount of resources a
1645 single client may allocate. A value of zero disables the limit so a single
1646 client may create as many streams as allocatable by haproxy. It is highly
1647 recommended not to change this value.
1648
Willy Tarreaua24b35c2019-02-21 13:24:36 +01001649tune.h2.max-frame-size <number>
1650 Sets the HTTP/2 maximum frame size that haproxy announces it is willing to
1651 receive to its peers. The default value is the largest between 16384 and the
1652 buffer size (tune.bufsize). In any case, haproxy will not announce support
1653 for frame sizes larger than buffers. The main purpose of this setting is to
1654 allow to limit the maximum frame size setting when using large buffers. Too
1655 large frame sizes might have performance impact or cause some peers to
1656 misbehave. It is highly recommended not to change this value.
1657
Willy Tarreau193b8c62012-11-22 00:17:38 +01001658tune.http.cookielen <number>
1659 Sets the maximum length of captured cookies. This is the maximum value that
1660 the "capture cookie xxx len yyy" will be allowed to take, and any upper value
1661 will automatically be truncated to this one. It is important not to set too
1662 high a value because all cookie captures still allocate this size whatever
1663 their configured value (they share a same pool). This value is per request
1664 per response, so the memory allocated is twice this value per connection.
1665 When not specified, the limit is set to 63 characters. It is recommended not
1666 to change this value.
1667
Stéphane Cottin23e9e932017-05-18 08:58:41 +02001668tune.http.logurilen <number>
Davor Ocelice9ed2812017-12-25 17:49:28 +01001669 Sets the maximum length of request URI in logs. This prevents truncating long
1670 request URIs with valuable query strings in log lines. This is not related
Stéphane Cottin23e9e932017-05-18 08:58:41 +02001671 to syslog limits. If you increase this limit, you may also increase the
Davor Ocelice9ed2812017-12-25 17:49:28 +01001672 'log ... len yyy' parameter. Your syslog daemon may also need specific
Stéphane Cottin23e9e932017-05-18 08:58:41 +02001673 configuration directives too.
1674 The default value is 1024.
1675
Willy Tarreauac1932d2011-10-24 19:14:41 +02001676tune.http.maxhdr <number>
1677 Sets the maximum number of headers in a request. When a request comes with a
1678 number of headers greater than this value (including the first line), it is
1679 rejected with a "400 Bad Request" status code. Similarly, too large responses
1680 are blocked with "502 Bad Gateway". The default value is 101, which is enough
1681 for all usages, considering that the widely deployed Apache server uses the
1682 same limit. It can be useful to push this limit further to temporarily allow
Christopher Faulet50174f32017-06-21 16:31:35 +02001683 a buggy application to work by the time it gets fixed. The accepted range is
1684 1..32767. Keep in mind that each new header consumes 32bits of memory for
1685 each session, so don't push this limit too high.
Willy Tarreauac1932d2011-10-24 19:14:41 +02001686
Willy Tarreau7e312732014-02-12 16:35:14 +01001687tune.idletimer <timeout>
1688 Sets the duration after which haproxy will consider that an empty buffer is
1689 probably associated with an idle stream. This is used to optimally adjust
1690 some packet sizes while forwarding large and small data alternatively. The
1691 decision to use splice() or to send large buffers in SSL is modulated by this
1692 parameter. The value is in milliseconds between 0 and 65535. A value of zero
1693 means that haproxy will not try to detect idle streams. The default is 1000,
Davor Ocelice9ed2812017-12-25 17:49:28 +01001694 which seems to correctly detect end user pauses (e.g. read a page before
John Roeslerfb2fce12019-07-10 15:45:51 -05001695 clicking). There should be no reason for changing this value. Please check
Willy Tarreau7e312732014-02-12 16:35:14 +01001696 tune.ssl.maxrecord below.
1697
Willy Tarreau7ac908b2019-02-27 12:02:18 +01001698tune.listener.multi-queue { on | off }
1699 Enables ('on') or disables ('off') the listener's multi-queue accept which
1700 spreads the incoming traffic to all threads a "bind" line is allowed to run
1701 on instead of taking them for itself. This provides a smoother traffic
1702 distribution and scales much better, especially in environments where threads
1703 may be unevenly loaded due to external activity (network interrupts colliding
1704 with one thread for example). This option is enabled by default, but it may
1705 be forcefully disabled for troubleshooting or for situations where it is
1706 estimated that the operating system already provides a good enough
1707 distribution and connections are extremely short-lived.
1708
Thierry FOURNIER90da1912015-03-05 11:17:06 +01001709tune.lua.forced-yield <number>
1710 This directive forces the Lua engine to execute a yield each <number> of
Tim Düsterhus4896c442016-11-29 02:15:19 +01001711 instructions executed. This permits interrupting a long script and allows the
Thierry FOURNIER90da1912015-03-05 11:17:06 +01001712 HAProxy scheduler to process other tasks like accepting connections or
1713 forwarding traffic. The default value is 10000 instructions. If HAProxy often
Davor Ocelice9ed2812017-12-25 17:49:28 +01001714 executes some Lua code but more responsiveness is required, this value can be
Thierry FOURNIER90da1912015-03-05 11:17:06 +01001715 lowered. If the Lua code is quite long and its result is absolutely required
1716 to process the data, the <number> can be increased.
1717
Willy Tarreau32f61e22015-03-18 17:54:59 +01001718tune.lua.maxmem
1719 Sets the maximum amount of RAM in megabytes per process usable by Lua. By
1720 default it is zero which means unlimited. It is important to set a limit to
1721 ensure that a bug in a script will not result in the system running out of
1722 memory.
1723
Thierry FOURNIER90da1912015-03-05 11:17:06 +01001724tune.lua.session-timeout <timeout>
1725 This is the execution timeout for the Lua sessions. This is useful for
Thierry FOURNIER7dd784b2015-10-01 14:49:33 +02001726 preventing infinite loops or spending too much time in Lua. This timeout
1727 counts only the pure Lua runtime. If the Lua does a sleep, the sleep is
Davor Ocelice9ed2812017-12-25 17:49:28 +01001728 not taken in account. The default timeout is 4s.
Thierry FOURNIER90da1912015-03-05 11:17:06 +01001729
1730tune.lua.task-timeout <timeout>
1731 Purpose is the same as "tune.lua.session-timeout", but this timeout is
1732 dedicated to the tasks. By default, this timeout isn't set because a task may
1733 remain alive during of the lifetime of HAProxy. For example, a task used to
1734 check servers.
1735
Thierry FOURNIER7dd784b2015-10-01 14:49:33 +02001736tune.lua.service-timeout <timeout>
1737 This is the execution timeout for the Lua services. This is useful for
1738 preventing infinite loops or spending too much time in Lua. This timeout
1739 counts only the pure Lua runtime. If the Lua does a sleep, the sleep is
Davor Ocelice9ed2812017-12-25 17:49:28 +01001740 not taken in account. The default timeout is 4s.
Thierry FOURNIER7dd784b2015-10-01 14:49:33 +02001741
Willy Tarreaua0250ba2008-01-06 11:22:57 +01001742tune.maxaccept <number>
Willy Tarreau16a21472012-11-19 12:39:59 +01001743 Sets the maximum number of consecutive connections a process may accept in a
1744 row before switching to other work. In single process mode, higher numbers
1745 give better performance at high connection rates. However in multi-process
1746 modes, keeping a bit of fairness between processes generally is better to
1747 increase performance. This value applies individually to each listener, so
1748 that the number of processes a listener is bound to is taken into account.
1749 This value defaults to 64. In multi-process mode, it is divided by twice
1750 the number of processes the listener is bound to. Setting this value to -1
1751 completely disables the limitation. It should normally not be needed to tweak
1752 this value.
Willy Tarreaua0250ba2008-01-06 11:22:57 +01001753
1754tune.maxpollevents <number>
1755 Sets the maximum amount of events that can be processed at once in a call to
1756 the polling system. The default value is adapted to the operating system. It
1757 has been noticed that reducing it below 200 tends to slightly decrease
1758 latency at the expense of network bandwidth, and increasing it above 200
1759 tends to trade latency for slightly increased bandwidth.
1760
Willy Tarreau27a674e2009-08-17 07:23:33 +02001761tune.maxrewrite <number>
1762 Sets the reserved buffer space to this size in bytes. The reserved space is
1763 used for header rewriting or appending. The first reads on sockets will never
1764 fill more than bufsize-maxrewrite. Historically it has defaulted to half of
1765 bufsize, though that does not make much sense since there are rarely large
1766 numbers of headers to add. Setting it too high prevents processing of large
1767 requests or responses. Setting it too low prevents addition of new headers
1768 to already large requests or to POST requests. It is generally wise to set it
1769 to about 1024. It is automatically readjusted to half of bufsize if it is
1770 larger than that. This means you don't have to worry about it when changing
1771 bufsize.
1772
Willy Tarreauf3045d22015-04-29 16:24:50 +02001773tune.pattern.cache-size <number>
1774 Sets the size of the pattern lookup cache to <number> entries. This is an LRU
1775 cache which reminds previous lookups and their results. It is used by ACLs
1776 and maps on slow pattern lookups, namely the ones using the "sub", "reg",
1777 "dir", "dom", "end", "bin" match methods as well as the case-insensitive
1778 strings. It applies to pattern expressions which means that it will be able
1779 to memorize the result of a lookup among all the patterns specified on a
1780 configuration line (including all those loaded from files). It automatically
1781 invalidates entries which are updated using HTTP actions or on the CLI. The
1782 default cache size is set to 10000 entries, which limits its footprint to
1783 about 5 MB on 32-bit systems and 8 MB on 64-bit systems. There is a very low
1784 risk of collision in this cache, which is in the order of the size of the
1785 cache divided by 2^64. Typically, at 10000 requests per second with the
1786 default cache size of 10000 entries, there's 1% chance that a brute force
1787 attack could cause a single collision after 60 years, or 0.1% after 6 years.
1788 This is considered much lower than the risk of a memory corruption caused by
1789 aging components. If this is not acceptable, the cache can be disabled by
1790 setting this parameter to 0.
1791
Willy Tarreaubd9a0a72011-10-23 21:14:29 +02001792tune.pipesize <number>
1793 Sets the kernel pipe buffer size to this size (in bytes). By default, pipes
1794 are the default size for the system. But sometimes when using TCP splicing,
1795 it can improve performance to increase pipe sizes, especially if it is
1796 suspected that pipes are not filled and that many calls to splice() are
1797 performed. This has an impact on the kernel's memory footprint, so this must
1798 not be changed if impacts are not understood.
1799
Olivier Houchard88698d92019-04-16 19:07:22 +02001800tune.pool-low-fd-ratio <number>
1801 This setting sets the max number of file descriptors (in percentage) used by
1802 haproxy globally against the maximum number of file descriptors haproxy can
1803 use before we stop putting connection into the idle pool for reuse. The
1804 default is 20.
1805
1806tune.pool-high-fd-ratio <number>
1807 This setting sets the max number of file descriptors (in percentage) used by
1808 haproxy globally against the maximum number of file descriptors haproxy can
1809 use before we start killing idle connections when we can't reuse a connection
1810 and we have to create a new one. The default is 25 (one quarter of the file
1811 descriptor will mean that roughly half of the maximum front connections can
1812 keep an idle connection behind, anything beyond this probably doesn't make
John Roeslerfb2fce12019-07-10 15:45:51 -05001813 much sense in the general case when targeting connection reuse).
Olivier Houchard88698d92019-04-16 19:07:22 +02001814
Willy Tarreaue803de22010-01-21 17:43:04 +01001815tune.rcvbuf.client <number>
1816tune.rcvbuf.server <number>
1817 Forces the kernel socket receive buffer size on the client or the server side
1818 to the specified value in bytes. This value applies to all TCP/HTTP frontends
1819 and backends. It should normally never be set, and the default size (0) lets
John Roeslerfb2fce12019-07-10 15:45:51 -05001820 the kernel auto-tune this value depending on the amount of available memory.
Davor Ocelice9ed2812017-12-25 17:49:28 +01001821 However it can sometimes help to set it to very low values (e.g. 4096) in
Willy Tarreaue803de22010-01-21 17:43:04 +01001822 order to save kernel memory by preventing it from buffering too large amounts
1823 of received data. Lower values will significantly increase CPU usage though.
1824
Willy Tarreaub22fc302015-12-14 12:04:35 +01001825tune.recv_enough <number>
Davor Ocelice9ed2812017-12-25 17:49:28 +01001826 HAProxy uses some hints to detect that a short read indicates the end of the
Willy Tarreaub22fc302015-12-14 12:04:35 +01001827 socket buffers. One of them is that a read returns more than <recv_enough>
1828 bytes, which defaults to 10136 (7 segments of 1448 each). This default value
1829 may be changed by this setting to better deal with workloads involving lots
1830 of short messages such as telnet or SSH sessions.
1831
Olivier Houchard1599b802018-05-24 18:59:04 +02001832tune.runqueue-depth <number>
John Roeslerfb2fce12019-07-10 15:45:51 -05001833 Sets the maximum amount of task that can be processed at once when running
Olivier Houchard1599b802018-05-24 18:59:04 +02001834 tasks. The default value is 200. Increasing it may incur latency when
1835 dealing with I/Os, making it too small can incur extra overhead.
1836
Willy Tarreaue803de22010-01-21 17:43:04 +01001837tune.sndbuf.client <number>
1838tune.sndbuf.server <number>
1839 Forces the kernel socket send buffer size on the client or the server side to
1840 the specified value in bytes. This value applies to all TCP/HTTP frontends
1841 and backends. It should normally never be set, and the default size (0) lets
John Roeslerfb2fce12019-07-10 15:45:51 -05001842 the kernel auto-tune this value depending on the amount of available memory.
Davor Ocelice9ed2812017-12-25 17:49:28 +01001843 However it can sometimes help to set it to very low values (e.g. 4096) in
Willy Tarreaue803de22010-01-21 17:43:04 +01001844 order to save kernel memory by preventing it from buffering too large amounts
1845 of received data. Lower values will significantly increase CPU usage though.
1846 Another use case is to prevent write timeouts with extremely slow clients due
1847 to the kernel waiting for a large part of the buffer to be read before
1848 notifying haproxy again.
1849
Willy Tarreau6ec58db2012-11-16 16:32:15 +01001850tune.ssl.cachesize <number>
Emeric Brunaf9619d2012-11-28 18:47:52 +01001851 Sets the size of the global SSL session cache, in a number of blocks. A block
1852 is large enough to contain an encoded session without peer certificate.
1853 An encoded session with peer certificate is stored in multiple blocks
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03001854 depending on the size of the peer certificate. A block uses approximately
Emeric Brunaf9619d2012-11-28 18:47:52 +01001855 200 bytes of memory. The default value may be forced at build time, otherwise
Davor Ocelice9ed2812017-12-25 17:49:28 +01001856 defaults to 20000. When the cache is full, the most idle entries are purged
Emeric Brunaf9619d2012-11-28 18:47:52 +01001857 and reassigned. Higher values reduce the occurrence of such a purge, hence
1858 the number of CPU-intensive SSL handshakes by ensuring that all users keep
1859 their session as long as possible. All entries are pre-allocated upon startup
Emeric Brun22890a12012-12-28 14:41:32 +01001860 and are shared between all processes if "nbproc" is greater than 1. Setting
1861 this value to 0 disables the SSL session cache.
Willy Tarreau6ec58db2012-11-16 16:32:15 +01001862
Emeric Brun8dc60392014-05-09 13:52:00 +02001863tune.ssl.force-private-cache
Lukas Tribus27935782018-10-01 02:00:16 +02001864 This option disables SSL session cache sharing between all processes. It
Emeric Brun8dc60392014-05-09 13:52:00 +02001865 should normally not be used since it will force many renegotiations due to
1866 clients hitting a random process. But it may be required on some operating
1867 systems where none of the SSL cache synchronization method may be used. In
1868 this case, adding a first layer of hash-based load balancing before the SSL
1869 layer might limit the impact of the lack of session sharing.
1870
Emeric Brun4f65bff2012-11-16 15:11:00 +01001871tune.ssl.lifetime <timeout>
1872 Sets how long a cached SSL session may remain valid. This time is expressed
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03001873 in seconds and defaults to 300 (5 min). It is important to understand that it
Emeric Brun4f65bff2012-11-16 15:11:00 +01001874 does not guarantee that sessions will last that long, because if the cache is
1875 full, the longest idle sessions will be purged despite their configured
1876 lifetime. The real usefulness of this setting is to prevent sessions from
1877 being used for too long.
1878
Willy Tarreaubfd59462013-02-21 07:46:09 +01001879tune.ssl.maxrecord <number>
1880 Sets the maximum amount of bytes passed to SSL_write() at a time. Default
1881 value 0 means there is no limit. Over SSL/TLS, the client can decipher the
1882 data only once it has received a full record. With large records, it means
1883 that clients might have to download up to 16kB of data before starting to
1884 process them. Limiting the value can improve page load times on browsers
1885 located over high latency or low bandwidth networks. It is suggested to find
1886 optimal values which fit into 1 or 2 TCP segments (generally 1448 bytes over
1887 Ethernet with TCP timestamps enabled, or 1460 when timestamps are disabled),
1888 keeping in mind that SSL/TLS add some overhead. Typical values of 1419 and
1889 2859 gave good results during tests. Use "strace -e trace=write" to find the
Davor Ocelice9ed2812017-12-25 17:49:28 +01001890 best value. HAProxy will automatically switch to this setting after an idle
Willy Tarreau7e312732014-02-12 16:35:14 +01001891 stream has been detected (see tune.idletimer above).
Willy Tarreaubfd59462013-02-21 07:46:09 +01001892
Remi Gacognef46cd6e2014-06-12 14:58:40 +02001893tune.ssl.default-dh-param <number>
1894 Sets the maximum size of the Diffie-Hellman parameters used for generating
1895 the ephemeral/temporary Diffie-Hellman key in case of DHE key exchange. The
1896 final size will try to match the size of the server's RSA (or DSA) key (e.g,
1897 a 2048 bits temporary DH key for a 2048 bits RSA key), but will not exceed
1898 this maximum value. Default value if 1024. Only 1024 or higher values are
1899 allowed. Higher values will increase the CPU load, and values greater than
1900 1024 bits are not supported by Java 7 and earlier clients. This value is not
Remi Gacogne47783ef2015-05-29 15:53:22 +02001901 used if static Diffie-Hellman parameters are supplied either directly
1902 in the certificate file or by using the ssl-dh-param-file parameter.
Remi Gacognef46cd6e2014-06-12 14:58:40 +02001903
Christopher Faulet31af49d2015-06-09 17:29:50 +02001904tune.ssl.ssl-ctx-cache-size <number>
1905 Sets the size of the cache used to store generated certificates to <number>
1906 entries. This is a LRU cache. Because generating a SSL certificate
1907 dynamically is expensive, they are cached. The default cache size is set to
1908 1000 entries.
1909
Thierry FOURNIER5bf77322017-02-25 12:45:22 +01001910tune.ssl.capture-cipherlist-size <number>
1911 Sets the maximum size of the buffer used for capturing client-hello cipher
1912 list. If the value is 0 (default value) the capture is disabled, otherwise
1913 a buffer is allocated for each SSL/TLS connection.
1914
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02001915tune.vars.global-max-size <size>
Christopher Fauletff2613e2016-11-09 11:36:17 +01001916tune.vars.proc-max-size <size>
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02001917tune.vars.reqres-max-size <size>
1918tune.vars.sess-max-size <size>
1919tune.vars.txn-max-size <size>
Christopher Fauletff2613e2016-11-09 11:36:17 +01001920 These five tunes help to manage the maximum amount of memory used by the
1921 variables system. "global" limits the overall amount of memory available for
1922 all scopes. "proc" limits the memory for the process scope, "sess" limits the
1923 memory for the session scope, "txn" for the transaction scope, and "reqres"
1924 limits the memory for each request or response processing.
1925 Memory accounting is hierarchical, meaning more coarse grained limits include
1926 the finer grained ones: "proc" includes "sess", "sess" includes "txn", and
1927 "txn" includes "reqres".
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02001928
Daniel Schneller0b547052016-03-21 20:46:57 +01001929 For example, when "tune.vars.sess-max-size" is limited to 100,
1930 "tune.vars.txn-max-size" and "tune.vars.reqres-max-size" cannot exceed
1931 100 either. If we create a variable "txn.var" that contains 100 bytes,
1932 all available space is consumed.
1933 Notice that exceeding the limits at runtime will not result in an error
1934 message, but values might be cut off or corrupted. So make sure to accurately
1935 plan for the amount of space needed to store all your variables.
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02001936
William Lallemanda509e4c2012-11-07 16:54:34 +01001937tune.zlib.memlevel <number>
1938 Sets the memLevel parameter in zlib initialization for each session. It
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03001939 defines how much memory should be allocated for the internal compression
William Lallemanda509e4c2012-11-07 16:54:34 +01001940 state. A value of 1 uses minimum memory but is slow and reduces compression
Davor Ocelice9ed2812017-12-25 17:49:28 +01001941 ratio, a value of 9 uses maximum memory for optimal speed. Can be a value
William Lallemanda509e4c2012-11-07 16:54:34 +01001942 between 1 and 9. The default value is 8.
1943
1944tune.zlib.windowsize <number>
1945 Sets the window size (the size of the history buffer) as a parameter of the
1946 zlib initialization for each session. Larger values of this parameter result
Davor Ocelice9ed2812017-12-25 17:49:28 +01001947 in better compression at the expense of memory usage. Can be a value between
1948 8 and 15. The default value is 15.
Willy Tarreau6a06a402007-07-15 20:15:28 +02001949
Willy Tarreauc57f0e22009-05-10 13:12:33 +020019503.3. Debugging
1951--------------
Willy Tarreau6a06a402007-07-15 20:15:28 +02001952
1953debug
1954 Enables debug mode which dumps to stdout all exchanges, and disables forking
1955 into background. It is the equivalent of the command-line argument "-d". It
1956 should never be used in a production configuration since it may prevent full
1957 system startup.
1958
1959quiet
1960 Do not display any message during startup. It is equivalent to the command-
1961 line argument "-q".
1962
Emeric Brunf099e792010-09-27 12:05:28 +02001963
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +010019643.4. Userlists
1965--------------
1966It is possible to control access to frontend/backend/listen sections or to
1967http stats by allowing only authenticated and authorized users. To do this,
1968it is required to create at least one userlist and to define users.
1969
1970userlist <listname>
Cyril Bonté78caf842010-03-10 22:41:43 +01001971 Creates new userlist with name <listname>. Many independent userlists can be
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01001972 used to store authentication & authorization data for independent customers.
1973
1974group <groupname> [users <user>,<user>,(...)]
Cyril Bonté78caf842010-03-10 22:41:43 +01001975 Adds group <groupname> to the current userlist. It is also possible to
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01001976 attach users to this group by using a comma separated list of names
1977 proceeded by "users" keyword.
1978
Cyril Bontéf0c60612010-02-06 14:44:47 +01001979user <username> [password|insecure-password <password>]
1980 [groups <group>,<group>,(...)]
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01001981 Adds user <username> to the current userlist. Both secure (encrypted) and
1982 insecure (unencrypted) passwords can be used. Encrypted passwords are
Daniel Schnellerd06f31c2017-11-06 16:51:04 +01001983 evaluated using the crypt(3) function, so depending on the system's
1984 capabilities, different algorithms are supported. For example, modern Glibc
1985 based Linux systems support MD5, SHA-256, SHA-512, and, of course, the
1986 classic DES-based method of encrypting passwords.
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01001987
Daniel Schnellerd06f31c2017-11-06 16:51:04 +01001988 Attention: Be aware that using encrypted passwords might cause significantly
1989 increased CPU usage, depending on the number of requests, and the algorithm
1990 used. For any of the hashed variants, the password for each request must
1991 be processed through the chosen algorithm, before it can be compared to the
1992 value specified in the config file. Most current algorithms are deliberately
1993 designed to be expensive to compute to achieve resistance against brute
1994 force attacks. They do not simply salt/hash the clear text password once,
1995 but thousands of times. This can quickly become a major factor in haproxy's
1996 overall CPU consumption!
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01001997
1998 Example:
Cyril Bontéf0c60612010-02-06 14:44:47 +01001999 userlist L1
2000 group G1 users tiger,scott
2001 group G2 users xdb,scott
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01002002
Cyril Bontéf0c60612010-02-06 14:44:47 +01002003 user tiger password $6$k6y3o.eP$JlKBx9za9667qe4(...)xHSwRv6J.C0/D7cV91
2004 user scott insecure-password elgato
2005 user xdb insecure-password hello
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01002006
Cyril Bontéf0c60612010-02-06 14:44:47 +01002007 userlist L2
2008 group G1
2009 group G2
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01002010
Cyril Bontéf0c60612010-02-06 14:44:47 +01002011 user tiger password $6$k6y3o.eP$JlKBx(...)xHSwRv6J.C0/D7cV91 groups G1
2012 user scott insecure-password elgato groups G1,G2
2013 user xdb insecure-password hello groups G2
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01002014
2015 Please note that both lists are functionally identical.
Willy Tarreau6a06a402007-07-15 20:15:28 +02002016
Emeric Brunf099e792010-09-27 12:05:28 +02002017
20183.5. Peers
Cyril Bontédc4d9032012-04-08 21:57:39 +02002019----------
Emeric Brun94900952015-06-11 18:25:54 +02002020It is possible to propagate entries of any data-types in stick-tables between
2021several haproxy instances over TCP connections in a multi-master fashion. Each
2022instance pushes its local updates and insertions to remote peers. The pushed
2023values overwrite remote ones without aggregation. Interrupted exchanges are
2024automatically detected and recovered from the last known point.
2025In addition, during a soft restart, the old process connects to the new one
2026using such a TCP connection to push all its entries before the new process
2027tries to connect to other peers. That ensures very fast replication during a
2028reload, it typically takes a fraction of a second even for large tables.
2029Note that Server IDs are used to identify servers remotely, so it is important
2030that configurations look similar or at least that the same IDs are forced on
2031each server on all participants.
Emeric Brunf099e792010-09-27 12:05:28 +02002032
2033peers <peersect>
Jamie Gloudon801a0a32012-08-25 00:18:33 -04002034 Creates a new peer list with name <peersect>. It is an independent section,
Emeric Brunf099e792010-09-27 12:05:28 +02002035 which is referenced by one or more stick-tables.
2036
Frédéric Lécaille2f167b32019-01-11 14:13:54 +01002037bind [<address>]:<port_range> [, ...] [param*]
2038 Defines the binding parameters of the local peer of this "peers" section.
2039 Such lines are not supported with "peer" line in the same "peers" section.
2040
Willy Tarreau77e4bd12015-05-01 20:02:17 +02002041disabled
2042 Disables a peers section. It disables both listening and any synchronization
2043 related to this section. This is provided to disable synchronization of stick
2044 tables without having to comment out all "peers" references.
2045
Frédéric Lécaille2f167b32019-01-11 14:13:54 +01002046default-bind [param*]
2047 Defines the binding parameters for the local peer, excepted its address.
2048
2049default-server [param*]
2050 Change default options for a server in a "peers" section.
2051
2052 Arguments:
2053 <param*> is a list of parameters for this server. The "default-server"
2054 keyword accepts an important number of options and has a complete
2055 section dedicated to it. Please refer to section 5 for more
2056 details.
2057
2058
2059 See also: "server" and section 5 about server options
2060
Willy Tarreau77e4bd12015-05-01 20:02:17 +02002061enable
2062 This re-enables a disabled peers section which was previously disabled.
2063
Frédéric Lécaille2f167b32019-01-11 14:13:54 +01002064peer <peername> <ip>:<port> [param*]
Emeric Brunf099e792010-09-27 12:05:28 +02002065 Defines a peer inside a peers section.
2066 If <peername> is set to the local peer name (by default hostname, or forced
2067 using "-L" command line option), haproxy will listen for incoming remote peer
2068 connection on <ip>:<port>. Otherwise, <ip>:<port> defines where to connect to
2069 to join the remote peer, and <peername> is used at the protocol level to
2070 identify and validate the remote peer on the server side.
2071
2072 During a soft restart, local peer <ip>:<port> is used by the old instance to
2073 connect the new one and initiate a complete replication (teaching process).
2074
2075 It is strongly recommended to have the exact same peers declaration on all
2076 peers and to only rely on the "-L" command line argument to change the local
2077 peer name. This makes it easier to maintain coherent configuration files
2078 across all peers.
2079
William Lallemandb2f07452015-05-12 14:27:13 +02002080 You may want to reference some environment variables in the address
2081 parameter, see section 2.3 about environment variables.
Willy Tarreaudad36a32013-03-11 01:20:04 +01002082
Frédéric Lécaille2f167b32019-01-11 14:13:54 +01002083 Note: "peer" keyword may transparently be replaced by "server" keyword (see
2084 "server" keyword explanation below).
2085
2086server <peername> [<ip>:<port>] [param*]
Michael Prokop4438c602019-05-24 10:25:45 +02002087 As previously mentioned, "peer" keyword may be replaced by "server" keyword
Frédéric Lécaille2f167b32019-01-11 14:13:54 +01002088 with a support for all "server" parameters found in 5.2 paragraph.
2089 If the underlying peer is local, <ip>:<port> parameters must not be present.
2090 These parameters must be provided on a "bind" line (see "bind" keyword
2091 of this "peers" section).
2092 Some of these parameters are irrelevant for "peers" sections.
2093
2094
Cyril Bontédc4d9032012-04-08 21:57:39 +02002095 Example:
Frédéric Lécaille2f167b32019-01-11 14:13:54 +01002096 # The old way.
Emeric Brunf099e792010-09-27 12:05:28 +02002097 peers mypeers
Willy Tarreauf7b30a92010-12-06 22:59:17 +01002098 peer haproxy1 192.168.0.1:1024
2099 peer haproxy2 192.168.0.2:1024
2100 peer haproxy3 10.2.0.1:1024
Emeric Brunf099e792010-09-27 12:05:28 +02002101
2102 backend mybackend
2103 mode tcp
2104 balance roundrobin
2105 stick-table type ip size 20k peers mypeers
2106 stick on src
2107
Willy Tarreauf7b30a92010-12-06 22:59:17 +01002108 server srv1 192.168.0.30:80
2109 server srv2 192.168.0.31:80
Emeric Brunf099e792010-09-27 12:05:28 +02002110
Frédéric Lécaille2f167b32019-01-11 14:13:54 +01002111 Example:
2112 peers mypeers
2113 bind 127.0.0.11:10001 ssl crt mycerts/pem
2114 default-server ssl verify none
2115 server hostA 127.0.0.10:10000
2116 server hostB #local peer
Emeric Brunf099e792010-09-27 12:05:28 +02002117
Frédéric Lécaille4f5b77c2019-03-18 14:05:58 +01002118
2119table <tablename> type {ip | integer | string [len <length>] | binary [len <length>]}
2120 size <size> [expire <expire>] [nopurge] [store <data_type>]*
2121
2122 Configure a stickiness table for the current section. This line is parsed
2123 exactly the same way as the "stick-table" keyword in others section, except
John Roeslerfb2fce12019-07-10 15:45:51 -05002124 for the "peers" argument which is not required here and with an additional
Frédéric Lécaille4f5b77c2019-03-18 14:05:58 +01002125 mandatory first parameter to designate the stick-table. Contrary to others
2126 sections, there may be several "table" lines in "peers" sections (see also
2127 "stick-table" keyword).
2128
2129 Also be aware of the fact that "peers" sections have their own stick-table
2130 namespaces to avoid collisions between stick-table names identical in
2131 different "peers" section. This is internally handled prepending the "peers"
2132 sections names to the name of the stick-tables followed by a '/' character.
2133 If somewhere else in the configuration file you have to refer to such
2134 stick-tables declared in "peers" sections you must use the prefixed version
2135 of the stick-table name as follows:
2136
2137 peers mypeers
2138 peer A ...
2139 peer B ...
2140 table t1 ...
2141
2142 frontend fe1
2143 tcp-request content track-sc0 src table mypeers/t1
2144
2145 This is also this prefixed version of the stick-table names which must be
2146 used to refer to stick-tables through the CLI.
2147
2148 About "peers" protocol, as only "peers" belonging to the same section may
2149 communicate with each others, there is no need to do such a distinction.
2150 Several "peers" sections may declare stick-tables with the same name.
2151 This is shorter version of the stick-table name which is sent over the network.
2152 There is only a '/' character as prefix to avoid stick-table name collisions between
2153 stick-tables declared as backends and stick-table declared in "peers" sections
2154 as follows in this weird but supported configuration:
2155
2156 peers mypeers
2157 peer A ...
2158 peer B ...
2159 table t1 type string size 10m store gpc0
2160
2161 backend t1
2162 stick-table type string size 10m store gpc0 peers mypeers
2163
2164 Here "t1" table declared in "mypeeers" section has "mypeers/t1" as global name.
2165 "t1" table declared as a backend as "t1" as global name. But at peer protocol
2166 level the former table is named "/t1", the latter is again named "t1".
2167
Simon Horman51a1cf62015-02-03 13:00:44 +090021683.6. Mailers
2169------------
2170It is possible to send email alerts when the state of servers changes.
2171If configured email alerts are sent to each mailer that is configured
2172in a mailers section. Email is sent to mailers using SMTP.
2173
Pieter Baauw386a1272015-08-16 15:26:24 +02002174mailers <mailersect>
Simon Horman51a1cf62015-02-03 13:00:44 +09002175 Creates a new mailer list with the name <mailersect>. It is an
2176 independent section which is referenced by one or more proxies.
2177
2178mailer <mailername> <ip>:<port>
2179 Defines a mailer inside a mailers section.
2180
2181 Example:
2182 mailers mymailers
2183 mailer smtp1 192.168.0.1:587
2184 mailer smtp2 192.168.0.2:587
2185
2186 backend mybackend
2187 mode tcp
2188 balance roundrobin
2189
2190 email-alert mailers mymailers
2191 email-alert from test1@horms.org
2192 email-alert to test2@horms.org
2193
2194 server srv1 192.168.0.30:80
2195 server srv2 192.168.0.31:80
2196
Pieter Baauw235fcfc2016-02-13 15:33:40 +01002197timeout mail <time>
2198 Defines the time available for a mail/connection to be made and send to
2199 the mail-server. If not defined the default value is 10 seconds. To allow
2200 for at least two SYN-ACK packets to be send during initial TCP handshake it
2201 is advised to keep this value above 4 seconds.
2202
2203 Example:
2204 mailers mymailers
2205 timeout mail 20s
2206 mailer smtp1 192.168.0.1:587
Simon Horman51a1cf62015-02-03 13:00:44 +09002207
William Lallemandc9515522019-06-12 16:32:11 +020022083.7. Programs
2209-------------
2210In master-worker mode, it is possible to launch external binaries with the
2211master, these processes are called programs. These programs are launched and
2212managed the same way as the workers.
2213
2214During a reload of HAProxy, those processes are dealing with the same
2215sequence as a worker:
2216
2217 - the master is re-executed
2218 - the master sends a SIGUSR1 signal to the program
2219 - if "option start-on-reload" is not disabled, the master launches a new
2220 instance of the program
2221
2222During a stop, or restart, a SIGTERM is sent to the programs.
2223
2224program <name>
2225 This is a new program section, this section will create an instance <name>
2226 which is visible in "show proc" on the master CLI. (See "9.4. Master CLI" in
2227 the management guide).
2228
2229command <command> [arguments*]
2230 Define the command to start with optional arguments. The command is looked
2231 up in the current PATH if it does not include an absolute path. This is a
2232 mandatory option of the program section. Arguments containing spaces must
2233 be enclosed in quotes or double quotes or be prefixed by a backslash.
2234
Andrew Heberle97236962019-07-12 11:50:26 +08002235user <user name>
2236 Changes the executed command user ID to the <user name> from /etc/passwd.
2237 See also "group".
2238
2239group <group name>
2240 Changes the executed command group ID to the <group name> from /etc/group.
2241 See also "user".
2242
William Lallemandc9515522019-06-12 16:32:11 +02002243option start-on-reload
2244no option start-on-reload
2245 Start (or not) a new instance of the program upon a reload of the master.
2246 The default is to start a new instance. This option may only be used in a
2247 program section.
2248
2249
Willy Tarreauc57f0e22009-05-10 13:12:33 +020022504. Proxies
Willy Tarreau6a06a402007-07-15 20:15:28 +02002251----------
Willy Tarreau0ba27502007-12-24 16:55:16 +01002252
Willy Tarreau6a06a402007-07-15 20:15:28 +02002253Proxy configuration can be located in a set of sections :
William Lallemand6e62fb62015-04-28 16:55:23 +02002254 - defaults [<name>]
Willy Tarreau6a06a402007-07-15 20:15:28 +02002255 - frontend <name>
2256 - backend <name>
2257 - listen <name>
2258
2259A "defaults" section sets default parameters for all other sections following
2260its declaration. Those default parameters are reset by the next "defaults"
2261section. See below for the list of parameters which can be set in a "defaults"
Willy Tarreau0ba27502007-12-24 16:55:16 +01002262section. The name is optional but its use is encouraged for better readability.
Willy Tarreau6a06a402007-07-15 20:15:28 +02002263
2264A "frontend" section describes a set of listening sockets accepting client
2265connections.
2266
2267A "backend" section describes a set of servers to which the proxy will connect
2268to forward incoming connections.
2269
2270A "listen" section defines a complete proxy with its frontend and backend
2271parts combined in one section. It is generally useful for TCP-only traffic.
2272
Willy Tarreau0ba27502007-12-24 16:55:16 +01002273All proxy names must be formed from upper and lower case letters, digits,
2274'-' (dash), '_' (underscore) , '.' (dot) and ':' (colon). ACL names are
2275case-sensitive, which means that "www" and "WWW" are two different proxies.
2276
2277Historically, all proxy names could overlap, it just caused troubles in the
2278logs. Since the introduction of content switching, it is mandatory that two
2279proxies with overlapping capabilities (frontend/backend) have different names.
2280However, it is still permitted that a frontend and a backend share the same
2281name, as this configuration seems to be commonly encountered.
2282
2283Right now, two major proxy modes are supported : "tcp", also known as layer 4,
2284and "http", also known as layer 7. In layer 4 mode, HAProxy simply forwards
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01002285bidirectional traffic between two sides. In layer 7 mode, HAProxy analyzes the
Willy Tarreau0ba27502007-12-24 16:55:16 +01002286protocol, and can interact with it by allowing, blocking, switching, adding,
2287modifying, or removing arbitrary contents in requests or responses, based on
2288arbitrary criteria.
2289
Willy Tarreau70dffda2014-01-30 03:07:23 +01002290In HTTP mode, the processing applied to requests and responses flowing over
2291a connection depends in the combination of the frontend's HTTP options and
Christopher Faulet315b39c2018-09-21 16:26:19 +02002292the backend's. HAProxy supports 4 connection modes :
Willy Tarreau70dffda2014-01-30 03:07:23 +01002293
2294 - KAL : keep alive ("option http-keep-alive") which is the default mode : all
2295 requests and responses are processed, and connections remain open but idle
2296 between responses and new requests.
2297
2298 - TUN: tunnel ("option http-tunnel") : this was the default mode for versions
2299 1.0 to 1.5-dev21 : only the first request and response are processed, and
2300 everything else is forwarded with no analysis at all. This mode should not
Christopher Faulet6c9bbb22019-03-26 21:37:23 +01002301 be used as it creates lots of trouble with logging and HTTP processing.
2302 And because it cannot work in HTTP/2, this option is deprecated and it is
2303 only supported on legacy HTTP frontends. In HTX, it is ignored and a
2304 warning is emitted during HAProxy startup.
Willy Tarreau70dffda2014-01-30 03:07:23 +01002305
Willy Tarreau70dffda2014-01-30 03:07:23 +01002306 - SCL: server close ("option http-server-close") : the server-facing
2307 connection is closed after the end of the response is received, but the
2308 client-facing connection remains open.
2309
Christopher Faulet315b39c2018-09-21 16:26:19 +02002310 - CLO: close ("option httpclose"): the connection is closed after the end of
2311 the response and "Connection: close" appended in both directions.
Willy Tarreau70dffda2014-01-30 03:07:23 +01002312
2313The effective mode that will be applied to a connection passing through a
2314frontend and a backend can be determined by both proxy modes according to the
2315following matrix, but in short, the modes are symmetric, keep-alive is the
Christopher Faulet315b39c2018-09-21 16:26:19 +02002316weakest option and close is the strongest.
Willy Tarreau70dffda2014-01-30 03:07:23 +01002317
Christopher Faulet315b39c2018-09-21 16:26:19 +02002318 Backend mode
Willy Tarreau70dffda2014-01-30 03:07:23 +01002319
Christopher Faulet315b39c2018-09-21 16:26:19 +02002320 | KAL | SCL | CLO
2321 ----+-----+-----+----
2322 KAL | KAL | SCL | CLO
2323 ----+-----+-----+----
2324 TUN | TUN | SCL | CLO
2325 Frontend ----+-----+-----+----
2326 mode SCL | SCL | SCL | CLO
2327 ----+-----+-----+----
2328 CLO | CLO | CLO | CLO
Willy Tarreau70dffda2014-01-30 03:07:23 +01002329
Willy Tarreau0ba27502007-12-24 16:55:16 +01002330
Willy Tarreau70dffda2014-01-30 03:07:23 +01002331
Willy Tarreauc57f0e22009-05-10 13:12:33 +020023324.1. Proxy keywords matrix
2333--------------------------
Willy Tarreau0ba27502007-12-24 16:55:16 +01002334
Willy Tarreauc57f0e22009-05-10 13:12:33 +02002335The following list of keywords is supported. Most of them may only be used in a
2336limited set of section types. Some of them are marked as "deprecated" because
2337they are inherited from an old syntax which may be confusing or functionally
2338limited, and there are new recommended keywords to replace them. Keywords
Davor Ocelice9ed2812017-12-25 17:49:28 +01002339marked with "(*)" can be optionally inverted using the "no" prefix, e.g. "no
Willy Tarreauc57f0e22009-05-10 13:12:33 +02002340option contstats". This makes sense when the option has been enabled by default
Willy Tarreau3842f002009-06-14 11:39:52 +02002341and must be disabled for a specific instance. Such options may also be prefixed
2342with "default" in order to restore default settings regardless of what has been
2343specified in a previous "defaults" section.
Willy Tarreau0ba27502007-12-24 16:55:16 +01002344
Willy Tarreau6a06a402007-07-15 20:15:28 +02002345
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01002346 keyword defaults frontend listen backend
2347------------------------------------+----------+----------+---------+---------
2348acl - X X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01002349backlog X X X -
2350balance X - X X
2351bind - X X -
2352bind-process X X X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01002353capture cookie - X X -
2354capture request header - X X -
2355capture response header - X X -
William Lallemand82fe75c2012-10-23 10:25:10 +02002356compression X X X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01002357cookie X - X X
Thierry FOURNIERa0a1b752015-05-26 17:44:32 +02002358declare capture - X X -
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01002359default-server X - X X
2360default_backend X X X -
2361description - X X X
2362disabled X X X X
2363dispatch - - X X
Simon Horman51a1cf62015-02-03 13:00:44 +09002364email-alert from X X X X
Simon Horman64e34162015-02-06 11:11:57 +09002365email-alert level X X X X
Simon Horman51a1cf62015-02-03 13:00:44 +09002366email-alert mailers X X X X
2367email-alert myhostname X X X X
2368email-alert to X X X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01002369enabled X X X X
2370errorfile X X X X
2371errorloc X X X X
2372errorloc302 X X X X
2373-- keyword -------------------------- defaults - frontend - listen -- backend -
2374errorloc303 X X X X
Cyril Bonté4288c5a2018-03-12 22:02:59 +01002375force-persist - - X X
Christopher Fauletc3fe5332016-04-07 15:30:10 +02002376filter - X X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01002377fullconn X - X X
2378grace X X X X
2379hash-type X - X X
2380http-check disable-on-404 X - X X
Willy Tarreaubd741542010-03-16 18:46:54 +01002381http-check expect - - X X
Willy Tarreau7ab6aff2010-10-12 06:30:16 +02002382http-check send-state X - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01002383http-request - X X X
Willy Tarreaue365c0b2013-06-11 16:06:12 +02002384http-response - X X X
Willy Tarreau30631952015-08-06 15:05:24 +02002385http-reuse X - X X
Baptiste Assmann2c42ef52013-10-09 21:57:02 +02002386http-send-name-header - - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01002387id - X X X
Cyril Bonté4288c5a2018-03-12 22:02:59 +01002388ignore-persist - - X X
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02002389load-server-state-from-file X - X X
William Lallemand0f99e342011-10-12 17:50:54 +02002390log (*) X X X X
Willy Tarreaufb4e7ea2015-01-07 14:55:17 +01002391log-format X X X -
Dragan Dosen7ad31542015-09-28 17:16:47 +02002392log-format-sd X X X -
Willy Tarreau094af4e2015-01-07 15:03:42 +01002393log-tag X X X X
Willy Tarreauc35362a2014-04-25 13:58:37 +02002394max-keep-alive-queue X - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01002395maxconn X X X -
2396mode X X X X
2397monitor fail - X X -
2398monitor-net X X X -
2399monitor-uri X X X -
2400option abortonclose (*) X - X X
2401option accept-invalid-http-request (*) X X X -
2402option accept-invalid-http-response (*) X - X X
2403option allbackups (*) X - X X
2404option checkcache (*) X - X X
2405option clitcpka (*) X X X -
2406option contstats (*) X X X -
2407option dontlog-normal (*) X X X -
2408option dontlognull (*) X X X -
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01002409-- keyword -------------------------- defaults - frontend - listen -- backend -
2410option forwardfor X X X X
Willy Tarreau9fbe18e2015-05-01 22:42:08 +02002411option http-buffer-request (*) X X X X
Willy Tarreau82649f92015-05-01 22:40:51 +02002412option http-ignore-probes (*) X X X -
Willy Tarreau16bfb022010-01-16 19:48:41 +01002413option http-keep-alive (*) X X X X
Willy Tarreau96e31212011-05-30 18:10:30 +02002414option http-no-delay (*) X X X X
Christopher Faulet98db9762018-09-21 10:25:19 +02002415option http-pretend-keepalive (*) X - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01002416option http-server-close (*) X X X X
Christopher Faulet6c9bbb22019-03-26 21:37:23 +01002417option http-tunnel (deprecated) (*) X X X -
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01002418option http-use-proxy-header (*) X X X -
Willy Tarreau68ad3a42018-10-22 11:49:15 +02002419option http-use-htx (*) X X X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01002420option httpchk X - X X
2421option httpclose (*) X X X X
Freddy Spierenburge88b7732019-03-25 14:35:17 +01002422option httplog X X X -
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01002423option http_proxy (*) X X X X
Jamie Gloudon801a0a32012-08-25 00:18:33 -04002424option independent-streams (*) X X X X
Gabor Lekenyb4c81e42010-09-29 18:17:05 +02002425option ldap-check X - X X
Simon Horman98637e52014-06-20 12:30:16 +09002426option external-check X - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01002427option log-health-checks (*) X - X X
2428option log-separate-errors (*) X X X -
2429option logasap (*) X X X -
2430option mysql-check X - X X
2431option nolinger (*) X X X X
2432option originalto X X X X
2433option persist (*) X - X X
Baptiste Assmann809e22a2015-10-12 20:22:55 +02002434option pgsql-check X - X X
2435option prefer-last-server (*) X - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01002436option redispatch (*) X - X X
Hervé COMMOWICKec032d62011-08-05 16:23:48 +02002437option redis-check X - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01002438option smtpchk X - X X
2439option socket-stats (*) X X X -
2440option splice-auto (*) X X X X
2441option splice-request (*) X X X X
2442option splice-response (*) X X X X
Christopher Fauletba7bc162016-11-07 21:07:38 +01002443option spop-check - - - X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01002444option srvtcpka (*) X - X X
2445option ssl-hello-chk X - X X
2446-- keyword -------------------------- defaults - frontend - listen -- backend -
Willy Tarreaued179852013-12-16 01:07:00 +01002447option tcp-check X - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01002448option tcp-smart-accept (*) X X X -
2449option tcp-smart-connect (*) X - X X
2450option tcpka X X X X
2451option tcplog X X X X
2452option transparent (*) X - X X
Simon Horman98637e52014-06-20 12:30:16 +09002453external-check command X - X X
2454external-check path X - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01002455persist rdp-cookie X - X X
2456rate-limit sessions X X X -
2457redirect - X X X
Willy Tarreau96d51952019-05-22 20:34:35 +02002458reqadd (deprecated) - X X X
2459reqallow (deprecated) - X X X
2460reqdel (deprecated) - X X X
2461reqdeny (deprecated) - X X X
2462reqiallow (deprecated) - X X X
2463reqidel (deprecated) - X X X
2464reqideny (deprecated) - X X X
2465reqipass (deprecated) - X X X
2466reqirep (deprecated) - X X X
2467reqitarpit (deprecated) - X X X
2468reqpass (deprecated) - X X X
2469reqrep (deprecated) - X X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01002470-- keyword -------------------------- defaults - frontend - listen -- backend -
Willy Tarreau96d51952019-05-22 20:34:35 +02002471reqtarpit (deprecated) - X X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01002472retries X - X X
Olivier Houcharda254a372019-04-05 15:30:12 +02002473retry-on X - X X
Willy Tarreau96d51952019-05-22 20:34:35 +02002474rspadd (deprecated) - X X X
2475rspdel (deprecated) - X X X
2476rspdeny (deprecated) - X X X
2477rspidel (deprecated) - X X X
2478rspideny (deprecated) - X X X
2479rspirep (deprecated) - X X X
2480rsprep (deprecated) - X X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01002481server - - X X
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02002482server-state-file-name X - X X
Frédéric Lécaillecb4502e2017-04-20 13:36:25 +02002483server-template - - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01002484source X - X X
Baptiste Assmann5a549212015-10-12 20:30:24 +02002485stats admin - X X X
2486stats auth X X X X
2487stats enable X X X X
2488stats hide-version X X X X
2489stats http-request - X X X
2490stats realm X X X X
2491stats refresh X X X X
2492stats scope X X X X
2493stats show-desc X X X X
2494stats show-legends X X X X
2495stats show-node X X X X
2496stats uri X X X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01002497-- keyword -------------------------- defaults - frontend - listen -- backend -
2498stick match - - X X
2499stick on - - X X
2500stick store-request - - X X
Willy Tarreaud8dc99f2011-07-01 11:33:25 +02002501stick store-response - - X X
Adam Spiers68af3c12017-04-06 16:31:39 +01002502stick-table - X X X
Willy Tarreau938c7fe2014-04-25 14:21:39 +02002503tcp-check connect - - X X
2504tcp-check expect - - X X
2505tcp-check send - - X X
2506tcp-check send-binary - - X X
Willy Tarreaue9656522010-08-17 15:40:09 +02002507tcp-request connection - X X -
2508tcp-request content - X X X
Willy Tarreaua56235c2010-09-14 11:31:36 +02002509tcp-request inspect-delay - X X X
Willy Tarreau4f614292016-10-21 17:49:36 +02002510tcp-request session - X X -
Emeric Brun0a3b67f2010-09-24 15:34:53 +02002511tcp-response content - - X X
2512tcp-response inspect-delay - - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01002513timeout check X - X X
2514timeout client X X X -
Willy Tarreau05cdd962014-05-10 14:30:07 +02002515timeout client-fin X X X -
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01002516timeout connect X - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01002517timeout http-keep-alive X X X X
2518timeout http-request X X X X
2519timeout queue X - X X
2520timeout server X - X X
Willy Tarreau05cdd962014-05-10 14:30:07 +02002521timeout server-fin X - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01002522timeout tarpit X X X X
Willy Tarreauce887fd2012-05-12 12:50:00 +02002523timeout tunnel X - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01002524transparent (deprecated) X - X X
William Lallemanda73203e2012-03-12 12:48:57 +01002525unique-id-format X X X -
2526unique-id-header X X X -
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01002527use_backend - X X -
Willy Tarreau4a5cade2012-04-05 21:09:48 +02002528use-server - - X X
Willy Tarreau5c6f7b32010-02-26 13:34:29 +01002529------------------------------------+----------+----------+---------+---------
2530 keyword defaults frontend listen backend
Willy Tarreau6a06a402007-07-15 20:15:28 +02002531
Willy Tarreau0ba27502007-12-24 16:55:16 +01002532
Willy Tarreauc57f0e22009-05-10 13:12:33 +020025334.2. Alphabetically sorted keywords reference
2534---------------------------------------------
Willy Tarreau0ba27502007-12-24 16:55:16 +01002535
2536This section provides a description of each keyword and its usage.
2537
2538
2539acl <aclname> <criterion> [flags] [operator] <value> ...
2540 Declare or complete an access list.
2541 May be used in sections : defaults | frontend | listen | backend
2542 no | yes | yes | yes
2543 Example:
2544 acl invalid_src src 0.0.0.0/7 224.0.0.0/3
2545 acl invalid_src src_port 0:1023
2546 acl local_dst hdr(host) -i localhost
2547
Willy Tarreauc57f0e22009-05-10 13:12:33 +02002548 See section 7 about ACL usage.
Willy Tarreau0ba27502007-12-24 16:55:16 +01002549
2550
Willy Tarreauc73ce2b2008-01-06 10:55:10 +01002551backlog <conns>
2552 Give hints to the system about the approximate listen backlog desired size
2553 May be used in sections : defaults | frontend | listen | backend
2554 yes | yes | yes | no
2555 Arguments :
2556 <conns> is the number of pending connections. Depending on the operating
2557 system, it may represent the number of already acknowledged
Cyril Bontédc4d9032012-04-08 21:57:39 +02002558 connections, of non-acknowledged ones, or both.
Willy Tarreauc73ce2b2008-01-06 10:55:10 +01002559
2560 In order to protect against SYN flood attacks, one solution is to increase
2561 the system's SYN backlog size. Depending on the system, sometimes it is just
2562 tunable via a system parameter, sometimes it is not adjustable at all, and
2563 sometimes the system relies on hints given by the application at the time of
2564 the listen() syscall. By default, HAProxy passes the frontend's maxconn value
2565 to the listen() syscall. On systems which can make use of this value, it can
2566 sometimes be useful to be able to specify a different value, hence this
2567 backlog parameter.
2568
2569 On Linux 2.4, the parameter is ignored by the system. On Linux 2.6, it is
2570 used as a hint and the system accepts up to the smallest greater power of
2571 two, and never more than some limits (usually 32768).
2572
2573 See also : "maxconn" and the target operating system's tuning guide.
2574
2575
Willy Tarreau0ba27502007-12-24 16:55:16 +01002576balance <algorithm> [ <arguments> ]
Willy Tarreau226071e2014-04-10 11:55:45 +02002577balance url_param <param> [check_post]
Willy Tarreau0ba27502007-12-24 16:55:16 +01002578 Define the load balancing algorithm to be used in a backend.
2579 May be used in sections : defaults | frontend | listen | backend
2580 yes | no | yes | yes
2581 Arguments :
2582 <algorithm> is the algorithm used to select a server when doing load
2583 balancing. This only applies when no persistence information
2584 is available, or when a connection is redispatched to another
2585 server. <algorithm> may be one of the following :
2586
2587 roundrobin Each server is used in turns, according to their weights.
2588 This is the smoothest and fairest algorithm when the server's
2589 processing time remains equally distributed. This algorithm
2590 is dynamic, which means that server weights may be adjusted
Willy Tarreau9757a382009-10-03 12:56:50 +02002591 on the fly for slow starts for instance. It is limited by
Godbacha34bdc02013-07-22 07:44:53 +08002592 design to 4095 active servers per backend. Note that in some
Willy Tarreau9757a382009-10-03 12:56:50 +02002593 large farms, when a server becomes up after having been down
2594 for a very short time, it may sometimes take a few hundreds
2595 requests for it to be re-integrated into the farm and start
2596 receiving traffic. This is normal, though very rare. It is
2597 indicated here in case you would have the chance to observe
2598 it, so that you don't worry.
2599
2600 static-rr Each server is used in turns, according to their weights.
2601 This algorithm is as similar to roundrobin except that it is
2602 static, which means that changing a server's weight on the
2603 fly will have no effect. On the other hand, it has no design
2604 limitation on the number of servers, and when a server goes
2605 up, it is always immediately reintroduced into the farm, once
2606 the full map is recomputed. It also uses slightly less CPU to
2607 run (around -1%).
Willy Tarreau0ba27502007-12-24 16:55:16 +01002608
Willy Tarreau2d2a7f82008-03-17 12:07:56 +01002609 leastconn The server with the lowest number of connections receives the
2610 connection. Round-robin is performed within groups of servers
2611 of the same load to ensure that all servers will be used. Use
2612 of this algorithm is recommended where very long sessions are
2613 expected, such as LDAP, SQL, TSE, etc... but is not very well
2614 suited for protocols using short sessions such as HTTP. This
2615 algorithm is dynamic, which means that server weights may be
2616 adjusted on the fly for slow starts for instance.
2617
Willy Tarreauf09c6602012-02-13 17:12:08 +01002618 first The first server with available connection slots receives the
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03002619 connection. The servers are chosen from the lowest numeric
Willy Tarreauf09c6602012-02-13 17:12:08 +01002620 identifier to the highest (see server parameter "id"), which
2621 defaults to the server's position in the farm. Once a server
Willy Tarreau64559c52012-04-07 09:08:45 +02002622 reaches its maxconn value, the next server is used. It does
Willy Tarreauf09c6602012-02-13 17:12:08 +01002623 not make sense to use this algorithm without setting maxconn.
2624 The purpose of this algorithm is to always use the smallest
2625 number of servers so that extra servers can be powered off
2626 during non-intensive hours. This algorithm ignores the server
2627 weight, and brings more benefit to long session such as RDP
Willy Tarreau64559c52012-04-07 09:08:45 +02002628 or IMAP than HTTP, though it can be useful there too. In
2629 order to use this algorithm efficiently, it is recommended
2630 that a cloud controller regularly checks server usage to turn
2631 them off when unused, and regularly checks backend queue to
2632 turn new servers on when the queue inflates. Alternatively,
2633 using "http-check send-state" may inform servers on the load.
Willy Tarreauf09c6602012-02-13 17:12:08 +01002634
Willy Tarreau0ba27502007-12-24 16:55:16 +01002635 source The source IP address is hashed and divided by the total
2636 weight of the running servers to designate which server will
2637 receive the request. This ensures that the same client IP
2638 address will always reach the same server as long as no
2639 server goes down or up. If the hash result changes due to the
2640 number of running servers changing, many clients will be
2641 directed to a different server. This algorithm is generally
2642 used in TCP mode where no cookie may be inserted. It may also
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01002643 be used on the Internet to provide a best-effort stickiness
Willy Tarreau0ba27502007-12-24 16:55:16 +01002644 to clients which refuse session cookies. This algorithm is
Willy Tarreau6b2e11b2009-10-01 07:52:15 +02002645 static by default, which means that changing a server's
2646 weight on the fly will have no effect, but this can be
2647 changed using "hash-type".
Willy Tarreau0ba27502007-12-24 16:55:16 +01002648
Oskar Stolc8dc41842012-05-19 10:19:54 +01002649 uri This algorithm hashes either the left part of the URI (before
2650 the question mark) or the whole URI (if the "whole" parameter
2651 is present) and divides the hash value by the total weight of
2652 the running servers. The result designates which server will
2653 receive the request. This ensures that the same URI will
2654 always be directed to the same server as long as no server
2655 goes up or down. This is used with proxy caches and
2656 anti-virus proxies in order to maximize the cache hit rate.
2657 Note that this algorithm may only be used in an HTTP backend.
2658 This algorithm is static by default, which means that
2659 changing a server's weight on the fly will have no effect,
2660 but this can be changed using "hash-type".
Willy Tarreau0ba27502007-12-24 16:55:16 +01002661
Oskar Stolc8dc41842012-05-19 10:19:54 +01002662 This algorithm supports two optional parameters "len" and
Marek Majkowski9c30fc12008-04-27 23:25:55 +02002663 "depth", both followed by a positive integer number. These
2664 options may be helpful when it is needed to balance servers
2665 based on the beginning of the URI only. The "len" parameter
2666 indicates that the algorithm should only consider that many
2667 characters at the beginning of the URI to compute the hash.
2668 Note that having "len" set to 1 rarely makes sense since most
2669 URIs start with a leading "/".
2670
2671 The "depth" parameter indicates the maximum directory depth
2672 to be used to compute the hash. One level is counted for each
2673 slash in the request. If both parameters are specified, the
2674 evaluation stops when either is reached.
2675
Willy Tarreau0ba27502007-12-24 16:55:16 +01002676 url_param The URL parameter specified in argument will be looked up in
matt.farnsworth@nokia.com1c2ab962008-04-14 20:47:37 +02002677 the query string of each HTTP GET request.
2678
2679 If the modifier "check_post" is used, then an HTTP POST
Cyril Bontédc4d9032012-04-08 21:57:39 +02002680 request entity will be searched for the parameter argument,
2681 when it is not found in a query string after a question mark
Willy Tarreau226071e2014-04-10 11:55:45 +02002682 ('?') in the URL. The message body will only start to be
2683 analyzed once either the advertised amount of data has been
2684 received or the request buffer is full. In the unlikely event
2685 that chunked encoding is used, only the first chunk is
Cyril Bontédc4d9032012-04-08 21:57:39 +02002686 scanned. Parameter values separated by a chunk boundary, may
Willy Tarreau226071e2014-04-10 11:55:45 +02002687 be randomly balanced if at all. This keyword used to support
2688 an optional <max_wait> parameter which is now ignored.
matt.farnsworth@nokia.com1c2ab962008-04-14 20:47:37 +02002689
2690 If the parameter is found followed by an equal sign ('=') and
2691 a value, then the value is hashed and divided by the total
2692 weight of the running servers. The result designates which
2693 server will receive the request.
2694
2695 This is used to track user identifiers in requests and ensure
2696 that a same user ID will always be sent to the same server as
2697 long as no server goes up or down. If no value is found or if
2698 the parameter is not found, then a round robin algorithm is
2699 applied. Note that this algorithm may only be used in an HTTP
Willy Tarreau6b2e11b2009-10-01 07:52:15 +02002700 backend. This algorithm is static by default, which means
2701 that changing a server's weight on the fly will have no
2702 effect, but this can be changed using "hash-type".
Willy Tarreau0ba27502007-12-24 16:55:16 +01002703
Cyril Bontédc4d9032012-04-08 21:57:39 +02002704 hdr(<name>) The HTTP header <name> will be looked up in each HTTP
2705 request. Just as with the equivalent ACL 'hdr()' function,
2706 the header name in parenthesis is not case sensitive. If the
2707 header is absent or if it does not contain any value, the
2708 roundrobin algorithm is applied instead.
Benoitaffb4812009-03-25 13:02:10 +01002709
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01002710 An optional 'use_domain_only' parameter is available, for
Benoitaffb4812009-03-25 13:02:10 +01002711 reducing the hash algorithm to the main domain part with some
2712 specific headers such as 'Host'. For instance, in the Host
2713 value "haproxy.1wt.eu", only "1wt" will be considered.
2714
Willy Tarreau6b2e11b2009-10-01 07:52:15 +02002715 This algorithm is static by default, which means that
2716 changing a server's weight on the fly will have no effect,
2717 but this can be changed using "hash-type".
2718
Willy Tarreau21c741a2019-01-14 18:14:27 +01002719 random
2720 random(<draws>)
2721 A random number will be used as the key for the consistent
Willy Tarreau760e81d2018-05-03 07:20:40 +02002722 hashing function. This means that the servers' weights are
2723 respected, dynamic weight changes immediately take effect, as
2724 well as new server additions. Random load balancing can be
2725 useful with large farms or when servers are frequently added
Willy Tarreau21c741a2019-01-14 18:14:27 +01002726 or removed as it may avoid the hammering effect that could
2727 result from roundrobin or leastconn in this situation. The
2728 hash-balance-factor directive can be used to further improve
2729 fairness of the load balancing, especially in situations
2730 where servers show highly variable response times. When an
2731 argument <draws> is present, it must be an integer value one
2732 or greater, indicating the number of draws before selecting
2733 the least loaded of these servers. It was indeed demonstrated
2734 that picking the least loaded of two servers is enough to
2735 significantly improve the fairness of the algorithm, by
2736 always avoiding to pick the most loaded server within a farm
2737 and getting rid of any bias that could be induced by the
2738 unfair distribution of the consistent list. Higher values N
2739 will take away N-1 of the highest loaded servers at the
2740 expense of performance. With very high values, the algorithm
2741 will converge towards the leastconn's result but much slower.
2742 The default value is 2, which generally shows very good
2743 distribution and performance. This algorithm is also known as
2744 the Power of Two Random Choices and is described here :
2745 http://www.eecs.harvard.edu/~michaelm/postscripts/handbook2001.pdf
Willy Tarreau760e81d2018-05-03 07:20:40 +02002746
Emeric Brun736aa232009-06-30 17:56:00 +02002747 rdp-cookie
Hervé COMMOWICKa3eb39c2011-08-05 18:48:51 +02002748 rdp-cookie(<name>)
Emeric Brun736aa232009-06-30 17:56:00 +02002749 The RDP cookie <name> (or "mstshash" if omitted) will be
2750 looked up and hashed for each incoming TCP request. Just as
2751 with the equivalent ACL 'req_rdp_cookie()' function, the name
2752 is not case-sensitive. This mechanism is useful as a degraded
2753 persistence mode, as it makes it possible to always send the
2754 same user (or the same session ID) to the same server. If the
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01002755 cookie is not found, the normal roundrobin algorithm is
Emeric Brun736aa232009-06-30 17:56:00 +02002756 used instead.
2757
2758 Note that for this to work, the frontend must ensure that an
2759 RDP cookie is already present in the request buffer. For this
2760 you must use 'tcp-request content accept' rule combined with
2761 a 'req_rdp_cookie_cnt' ACL.
2762
Willy Tarreau6b2e11b2009-10-01 07:52:15 +02002763 This algorithm is static by default, which means that
2764 changing a server's weight on the fly will have no effect,
2765 but this can be changed using "hash-type".
2766
Cyril Bontédc4d9032012-04-08 21:57:39 +02002767 See also the rdp_cookie pattern fetch function.
Simon Hormanab814e02011-06-24 14:50:20 +09002768
Willy Tarreau0ba27502007-12-24 16:55:16 +01002769 <arguments> is an optional list of arguments which may be needed by some
Marek Majkowski9c30fc12008-04-27 23:25:55 +02002770 algorithms. Right now, only "url_param" and "uri" support an
2771 optional argument.
matt.farnsworth@nokia.com1c2ab962008-04-14 20:47:37 +02002772
Willy Tarreau3cd9af22009-03-15 14:06:41 +01002773 The load balancing algorithm of a backend is set to roundrobin when no other
2774 algorithm, mode nor option have been set. The algorithm may only be set once
2775 for each backend.
Willy Tarreau0ba27502007-12-24 16:55:16 +01002776
Lukas Tribus80512b12018-10-27 20:07:40 +02002777 With authentication schemes that require the same connection like NTLM, URI
John Roeslerfb2fce12019-07-10 15:45:51 -05002778 based algorithms must not be used, as they would cause subsequent requests
Lukas Tribus80512b12018-10-27 20:07:40 +02002779 to be routed to different backend servers, breaking the invalid assumptions
2780 NTLM relies on.
2781
Willy Tarreau0ba27502007-12-24 16:55:16 +01002782 Examples :
2783 balance roundrobin
2784 balance url_param userid
matt.farnsworth@nokia.com1c2ab962008-04-14 20:47:37 +02002785 balance url_param session_id check_post 64
Benoitaffb4812009-03-25 13:02:10 +01002786 balance hdr(User-Agent)
2787 balance hdr(host)
2788 balance hdr(Host) use_domain_only
matt.farnsworth@nokia.com1c2ab962008-04-14 20:47:37 +02002789
2790 Note: the following caveats and limitations on using the "check_post"
2791 extension with "url_param" must be considered :
2792
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01002793 - all POST requests are eligible for consideration, because there is no way
matt.farnsworth@nokia.com1c2ab962008-04-14 20:47:37 +02002794 to determine if the parameters will be found in the body or entity which
2795 may contain binary data. Therefore another method may be required to
2796 restrict consideration of POST requests that have no URL parameters in
2797 the body. (see acl reqideny http_end)
2798
2799 - using a <max_wait> value larger than the request buffer size does not
2800 make sense and is useless. The buffer size is set at build time, and
2801 defaults to 16 kB.
2802
2803 - Content-Encoding is not supported, the parameter search will probably
2804 fail; and load balancing will fall back to Round Robin.
2805
2806 - Expect: 100-continue is not supported, load balancing will fall back to
2807 Round Robin.
2808
Lukas Tribus23953682017-04-28 13:24:30 +00002809 - Transfer-Encoding (RFC7230 3.3.1) is only supported in the first chunk.
matt.farnsworth@nokia.com1c2ab962008-04-14 20:47:37 +02002810 If the entire parameter value is not present in the first chunk, the
2811 selection of server is undefined (actually, defined by how little
2812 actually appeared in the first chunk).
2813
2814 - This feature does not support generation of a 100, 411 or 501 response.
2815
2816 - In some cases, requesting "check_post" MAY attempt to scan the entire
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01002817 contents of a message body. Scanning normally terminates when linear
matt.farnsworth@nokia.com1c2ab962008-04-14 20:47:37 +02002818 white space or control characters are found, indicating the end of what
2819 might be a URL parameter list. This is probably not a concern with SGML
2820 type message bodies.
Willy Tarreau0ba27502007-12-24 16:55:16 +01002821
Willy Tarreau294d0f02015-08-10 19:40:12 +02002822 See also : "dispatch", "cookie", "transparent", "hash-type" and "http_proxy".
Willy Tarreau0ba27502007-12-24 16:55:16 +01002823
2824
Willy Tarreaub6205fd2012-09-24 12:27:33 +02002825bind [<address>]:<port_range> [, ...] [param*]
2826bind /<path> [, ...] [param*]
Willy Tarreau0ba27502007-12-24 16:55:16 +01002827 Define one or several listening addresses and/or ports in a frontend.
2828 May be used in sections : defaults | frontend | listen | backend
2829 no | yes | yes | no
2830 Arguments :
Willy Tarreaub1e52e82008-01-13 14:49:51 +01002831 <address> is optional and can be a host name, an IPv4 address, an IPv6
2832 address, or '*'. It designates the address the frontend will
2833 listen on. If unset, all IPv4 addresses of the system will be
2834 listened on. The same will apply for '*' or the system's
David du Colombier9c938da2011-03-17 10:40:27 +01002835 special address "0.0.0.0". The IPv6 equivalent is '::'.
Willy Tarreau24709282013-03-10 21:32:12 +01002836 Optionally, an address family prefix may be used before the
2837 address to force the family regardless of the address format,
2838 which can be useful to specify a path to a unix socket with
2839 no slash ('/'). Currently supported prefixes are :
2840 - 'ipv4@' -> address is always IPv4
2841 - 'ipv6@' -> address is always IPv6
2842 - 'unix@' -> address is a path to a local unix socket
Willy Tarreau70f72e02014-07-08 00:37:50 +02002843 - 'abns@' -> address is in abstract namespace (Linux only).
2844 Note: since abstract sockets are not "rebindable", they
2845 do not cope well with multi-process mode during
2846 soft-restart, so it is better to avoid them if
2847 nbproc is greater than 1. The effect is that if the
2848 new process fails to start, only one of the old ones
2849 will be able to rebind to the socket.
Willy Tarreau40aa0702013-03-10 23:51:38 +01002850 - 'fd@<n>' -> use file descriptor <n> inherited from the
2851 parent. The fd must be bound and may or may not already
2852 be listening.
William Lallemand2fe7dd02018-09-11 16:51:29 +02002853 - 'sockpair@<n>'-> like fd@ but you must use the fd of a
2854 connected unix socket or of a socketpair. The bind waits
2855 to receive a FD over the unix socket and uses it as if it
2856 was the FD of an accept(). Should be used carefully.
William Lallemandb2f07452015-05-12 14:27:13 +02002857 You may want to reference some environment variables in the
2858 address parameter, see section 2.3 about environment
2859 variables.
Willy Tarreaub1e52e82008-01-13 14:49:51 +01002860
Willy Tarreauc5011ca2010-03-22 11:53:56 +01002861 <port_range> is either a unique TCP port, or a port range for which the
2862 proxy will accept connections for the IP address specified
Willy Tarreauceb24bc2010-11-09 12:46:41 +01002863 above. The port is mandatory for TCP listeners. Note that in
2864 the case of an IPv6 address, the port is always the number
2865 after the last colon (':'). A range can either be :
Willy Tarreauc5011ca2010-03-22 11:53:56 +01002866 - a numerical port (ex: '80')
2867 - a dash-delimited ports range explicitly stating the lower
2868 and upper bounds (ex: '2000-2100') which are included in
2869 the range.
2870
2871 Particular care must be taken against port ranges, because
2872 every <address:port> couple consumes one socket (= a file
2873 descriptor), so it's easy to consume lots of descriptors
2874 with a simple range, and to run out of sockets. Also, each
2875 <address:port> couple must be used only once among all
2876 instances running on a same system. Please note that binding
2877 to ports lower than 1024 generally require particular
Jamie Gloudon801a0a32012-08-25 00:18:33 -04002878 privileges to start the program, which are independent of
Willy Tarreauc5011ca2010-03-22 11:53:56 +01002879 the 'uid' parameter.
Willy Tarreau0ba27502007-12-24 16:55:16 +01002880
Willy Tarreauceb24bc2010-11-09 12:46:41 +01002881 <path> is a UNIX socket path beginning with a slash ('/'). This is
Davor Ocelice9ed2812017-12-25 17:49:28 +01002882 alternative to the TCP listening port. HAProxy will then
Willy Tarreauceb24bc2010-11-09 12:46:41 +01002883 receive UNIX connections on the socket located at this place.
2884 The path must begin with a slash and by default is absolute.
2885 It can be relative to the prefix defined by "unix-bind" in
2886 the global section. Note that the total length of the prefix
2887 followed by the socket path cannot exceed some system limits
2888 for UNIX sockets, which commonly are set to 107 characters.
2889
Willy Tarreaub6205fd2012-09-24 12:27:33 +02002890 <param*> is a list of parameters common to all sockets declared on the
2891 same line. These numerous parameters depend on OS and build
2892 options and have a complete section dedicated to them. Please
2893 refer to section 5 to for more details.
Willy Tarreaua0ee1d02012-09-10 09:01:23 +02002894
Willy Tarreau0ba27502007-12-24 16:55:16 +01002895 It is possible to specify a list of address:port combinations delimited by
2896 commas. The frontend will then listen on all of these addresses. There is no
2897 fixed limit to the number of addresses and ports which can be listened on in
2898 a frontend, as well as there is no limit to the number of "bind" statements
2899 in a frontend.
2900
2901 Example :
2902 listen http_proxy
2903 bind :80,:443
2904 bind 10.0.0.1:10080,10.0.0.1:10443
Willy Tarreauceb24bc2010-11-09 12:46:41 +01002905 bind /var/run/ssl-frontend.sock user root mode 600 accept-proxy
Willy Tarreau0ba27502007-12-24 16:55:16 +01002906
Willy Tarreaua0ee1d02012-09-10 09:01:23 +02002907 listen http_https_proxy
2908 bind :80
Cyril Bonté0d44fc62012-10-09 22:45:33 +02002909 bind :443 ssl crt /etc/haproxy/site.pem
Willy Tarreaua0ee1d02012-09-10 09:01:23 +02002910
Willy Tarreau24709282013-03-10 21:32:12 +01002911 listen http_https_proxy_explicit
2912 bind ipv6@:80
2913 bind ipv4@public_ssl:443 ssl crt /etc/haproxy/site.pem
2914 bind unix@ssl-frontend.sock user root mode 600 accept-proxy
2915
Willy Tarreaudad36a32013-03-11 01:20:04 +01002916 listen external_bind_app1
William Lallemandb2f07452015-05-12 14:27:13 +02002917 bind "fd@${FD_APP1}"
Willy Tarreaudad36a32013-03-11 01:20:04 +01002918
Willy Tarreau55dcaf62015-09-27 15:03:15 +02002919 Note: regarding Linux's abstract namespace sockets, HAProxy uses the whole
2920 sun_path length is used for the address length. Some other programs
2921 such as socat use the string length only by default. Pass the option
2922 ",unix-tightsocklen=0" to any abstract socket definition in socat to
2923 make it compatible with HAProxy's.
2924
Willy Tarreauceb24bc2010-11-09 12:46:41 +01002925 See also : "source", "option forwardfor", "unix-bind" and the PROXY protocol
Willy Tarreaub6205fd2012-09-24 12:27:33 +02002926 documentation, and section 5 about bind options.
Willy Tarreau0ba27502007-12-24 16:55:16 +01002927
2928
Christopher Fauletff4121f2017-11-22 16:38:49 +01002929bind-process [ all | odd | even | <process_num>[-[<process_num>]] ] ...
Willy Tarreau0b9c02c2009-02-04 22:05:05 +01002930 Limit visibility of an instance to a certain set of processes numbers.
2931 May be used in sections : defaults | frontend | listen | backend
2932 yes | yes | yes | yes
2933 Arguments :
2934 all All process will see this instance. This is the default. It
2935 may be used to override a default value.
2936
Willy Tarreaua9db57e2013-01-18 11:29:29 +01002937 odd This instance will be enabled on processes 1,3,5,...63. This
Willy Tarreau0b9c02c2009-02-04 22:05:05 +01002938 option may be combined with other numbers.
2939
Willy Tarreaua9db57e2013-01-18 11:29:29 +01002940 even This instance will be enabled on processes 2,4,6,...64. This
Willy Tarreau0b9c02c2009-02-04 22:05:05 +01002941 option may be combined with other numbers. Do not use it
2942 with less than 2 processes otherwise some instances might be
2943 missing from all processes.
2944
Christopher Fauletff4121f2017-11-22 16:38:49 +01002945 process_num The instance will be enabled on this process number or range,
Willy Tarreaua9db57e2013-01-18 11:29:29 +01002946 whose values must all be between 1 and 32 or 64 depending on
Christopher Fauletff4121f2017-11-22 16:38:49 +01002947 the machine's word size. Ranges can be partially defined. The
2948 higher bound can be omitted. In such case, it is replaced by
2949 the corresponding maximum value. If a proxy is bound to
2950 process numbers greater than the configured global.nbproc, it
2951 will either be forced to process #1 if a single process was
Willy Tarreau102df612014-05-07 23:56:38 +02002952 specified, or to all processes otherwise.
Willy Tarreau0b9c02c2009-02-04 22:05:05 +01002953
2954 This keyword limits binding of certain instances to certain processes. This
2955 is useful in order not to have too many processes listening to the same
2956 ports. For instance, on a dual-core machine, it might make sense to set
2957 'nbproc 2' in the global section, then distributes the listeners among 'odd'
2958 and 'even' instances.
2959
Willy Tarreaua9db57e2013-01-18 11:29:29 +01002960 At the moment, it is not possible to reference more than 32 or 64 processes
2961 using this keyword, but this should be more than enough for most setups.
2962 Please note that 'all' really means all processes regardless of the machine's
2963 word size, and is not limited to the first 32 or 64.
Willy Tarreau0b9c02c2009-02-04 22:05:05 +01002964
Willy Tarreau6ae1ba62014-05-07 19:01:58 +02002965 Each "bind" line may further be limited to a subset of the proxy's processes,
2966 please consult the "process" bind keyword in section 5.1.
2967
Willy Tarreaub369a042014-09-16 13:21:03 +02002968 When a frontend has no explicit "bind-process" line, it tries to bind to all
2969 the processes referenced by its "bind" lines. That means that frontends can
2970 easily adapt to their listeners' processes.
2971
Willy Tarreau0b9c02c2009-02-04 22:05:05 +01002972 If some backends are referenced by frontends bound to other processes, the
2973 backend automatically inherits the frontend's processes.
2974
2975 Example :
2976 listen app_ip1
2977 bind 10.0.0.1:80
Willy Tarreaubfcd3112010-10-23 11:22:08 +02002978 bind-process odd
Willy Tarreau0b9c02c2009-02-04 22:05:05 +01002979
2980 listen app_ip2
2981 bind 10.0.0.2:80
Willy Tarreaubfcd3112010-10-23 11:22:08 +02002982 bind-process even
Willy Tarreau0b9c02c2009-02-04 22:05:05 +01002983
2984 listen management
2985 bind 10.0.0.3:80
Willy Tarreaubfcd3112010-10-23 11:22:08 +02002986 bind-process 1 2 3 4
Willy Tarreau0b9c02c2009-02-04 22:05:05 +01002987
Willy Tarreau110ecc12012-11-15 17:50:01 +01002988 listen management
2989 bind 10.0.0.4:80
2990 bind-process 1-4
2991
Willy Tarreau6ae1ba62014-05-07 19:01:58 +02002992 See also : "nbproc" in global section, and "process" in section 5.1.
Willy Tarreau0b9c02c2009-02-04 22:05:05 +01002993
2994
Willy Tarreau0ba27502007-12-24 16:55:16 +01002995capture cookie <name> len <length>
2996 Capture and log a cookie in the request and in the response.
2997 May be used in sections : defaults | frontend | listen | backend
2998 no | yes | yes | no
2999 Arguments :
3000 <name> is the beginning of the name of the cookie to capture. In order
3001 to match the exact name, simply suffix the name with an equal
3002 sign ('='). The full name will appear in the logs, which is
3003 useful with application servers which adjust both the cookie name
Davor Ocelice9ed2812017-12-25 17:49:28 +01003004 and value (e.g. ASPSESSIONXXX).
Willy Tarreau0ba27502007-12-24 16:55:16 +01003005
3006 <length> is the maximum number of characters to report in the logs, which
3007 include the cookie name, the equal sign and the value, all in the
3008 standard "name=value" form. The string will be truncated on the
3009 right if it exceeds <length>.
3010
3011 Only the first cookie is captured. Both the "cookie" request headers and the
3012 "set-cookie" response headers are monitored. This is particularly useful to
3013 check for application bugs causing session crossing or stealing between
3014 users, because generally the user's cookies can only change on a login page.
3015
3016 When the cookie was not presented by the client, the associated log column
3017 will report "-". When a request does not cause a cookie to be assigned by the
3018 server, a "-" is reported in the response column.
3019
3020 The capture is performed in the frontend only because it is necessary that
3021 the log format does not change for a given frontend depending on the
3022 backends. This may change in the future. Note that there can be only one
Willy Tarreau193b8c62012-11-22 00:17:38 +01003023 "capture cookie" statement in a frontend. The maximum capture length is set
3024 by the global "tune.http.cookielen" setting and defaults to 63 characters. It
3025 is not possible to specify a capture in a "defaults" section.
Willy Tarreau0ba27502007-12-24 16:55:16 +01003026
3027 Example:
3028 capture cookie ASPSESSION len 32
3029
3030 See also : "capture request header", "capture response header" as well as
Willy Tarreauc57f0e22009-05-10 13:12:33 +02003031 section 8 about logging.
Willy Tarreau0ba27502007-12-24 16:55:16 +01003032
3033
3034capture request header <name> len <length>
Willy Tarreau4460d032012-11-21 23:37:37 +01003035 Capture and log the last occurrence of the specified request header.
Willy Tarreau0ba27502007-12-24 16:55:16 +01003036 May be used in sections : defaults | frontend | listen | backend
3037 no | yes | yes | no
3038 Arguments :
3039 <name> is the name of the header to capture. The header names are not
Willy Tarreaud2a4aa22008-01-31 15:28:22 +01003040 case-sensitive, but it is a common practice to write them as they
Willy Tarreau0ba27502007-12-24 16:55:16 +01003041 appear in the requests, with the first letter of each word in
3042 upper case. The header name will not appear in the logs, only the
3043 value is reported, but the position in the logs is respected.
3044
3045 <length> is the maximum number of characters to extract from the value and
3046 report in the logs. The string will be truncated on the right if
3047 it exceeds <length>.
3048
Willy Tarreau4460d032012-11-21 23:37:37 +01003049 The complete value of the last occurrence of the header is captured. The
Willy Tarreau0ba27502007-12-24 16:55:16 +01003050 value will be added to the logs between braces ('{}'). If multiple headers
3051 are captured, they will be delimited by a vertical bar ('|') and will appear
Willy Tarreaucc6c8912009-02-22 10:53:55 +01003052 in the same order they were declared in the configuration. Non-existent
3053 headers will be logged just as an empty string. Common uses for request
3054 header captures include the "Host" field in virtual hosting environments, the
3055 "Content-length" when uploads are supported, "User-agent" to quickly
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01003056 differentiate between real users and robots, and "X-Forwarded-For" in proxied
Willy Tarreaucc6c8912009-02-22 10:53:55 +01003057 environments to find where the request came from.
3058
3059 Note that when capturing headers such as "User-agent", some spaces may be
3060 logged, making the log analysis more difficult. Thus be careful about what
3061 you log if you know your log parser is not smart enough to rely on the
3062 braces.
Willy Tarreau0ba27502007-12-24 16:55:16 +01003063
Willy Tarreau0900abb2012-11-22 00:21:46 +01003064 There is no limit to the number of captured request headers nor to their
3065 length, though it is wise to keep them low to limit memory usage per session.
3066 In order to keep log format consistent for a same frontend, header captures
3067 can only be declared in a frontend. It is not possible to specify a capture
3068 in a "defaults" section.
Willy Tarreau0ba27502007-12-24 16:55:16 +01003069
3070 Example:
3071 capture request header Host len 15
3072 capture request header X-Forwarded-For len 15
Cyril Bontéd1b0f7c2015-10-26 22:37:39 +01003073 capture request header Referer len 15
Willy Tarreau0ba27502007-12-24 16:55:16 +01003074
Willy Tarreauc57f0e22009-05-10 13:12:33 +02003075 See also : "capture cookie", "capture response header" as well as section 8
Willy Tarreau0ba27502007-12-24 16:55:16 +01003076 about logging.
3077
3078
3079capture response header <name> len <length>
Willy Tarreau4460d032012-11-21 23:37:37 +01003080 Capture and log the last occurrence of the specified response header.
Willy Tarreau0ba27502007-12-24 16:55:16 +01003081 May be used in sections : defaults | frontend | listen | backend
3082 no | yes | yes | no
3083 Arguments :
3084 <name> is the name of the header to capture. The header names are not
Willy Tarreaud2a4aa22008-01-31 15:28:22 +01003085 case-sensitive, but it is a common practice to write them as they
Willy Tarreau0ba27502007-12-24 16:55:16 +01003086 appear in the response, with the first letter of each word in
3087 upper case. The header name will not appear in the logs, only the
3088 value is reported, but the position in the logs is respected.
3089
3090 <length> is the maximum number of characters to extract from the value and
3091 report in the logs. The string will be truncated on the right if
3092 it exceeds <length>.
3093
Willy Tarreau4460d032012-11-21 23:37:37 +01003094 The complete value of the last occurrence of the header is captured. The
Willy Tarreau0ba27502007-12-24 16:55:16 +01003095 result will be added to the logs between braces ('{}') after the captured
3096 request headers. If multiple headers are captured, they will be delimited by
3097 a vertical bar ('|') and will appear in the same order they were declared in
Willy Tarreaucc6c8912009-02-22 10:53:55 +01003098 the configuration. Non-existent headers will be logged just as an empty
3099 string. Common uses for response header captures include the "Content-length"
3100 header which indicates how many bytes are expected to be returned, the
3101 "Location" header to track redirections.
Willy Tarreau0ba27502007-12-24 16:55:16 +01003102
Willy Tarreau0900abb2012-11-22 00:21:46 +01003103 There is no limit to the number of captured response headers nor to their
3104 length, though it is wise to keep them low to limit memory usage per session.
3105 In order to keep log format consistent for a same frontend, header captures
3106 can only be declared in a frontend. It is not possible to specify a capture
3107 in a "defaults" section.
Willy Tarreau0ba27502007-12-24 16:55:16 +01003108
3109 Example:
3110 capture response header Content-length len 9
3111 capture response header Location len 15
3112
Willy Tarreauc57f0e22009-05-10 13:12:33 +02003113 See also : "capture cookie", "capture request header" as well as section 8
Willy Tarreau0ba27502007-12-24 16:55:16 +01003114 about logging.
3115
3116
Cyril Bonté316a8cf2012-11-11 13:38:27 +01003117compression algo <algorithm> ...
3118compression type <mime type> ...
Willy Tarreau70737d12012-10-27 00:34:28 +02003119compression offload
William Lallemand82fe75c2012-10-23 10:25:10 +02003120 Enable HTTP compression.
3121 May be used in sections : defaults | frontend | listen | backend
3122 yes | yes | yes | yes
3123 Arguments :
Cyril Bonté316a8cf2012-11-11 13:38:27 +01003124 algo is followed by the list of supported compression algorithms.
3125 type is followed by the list of MIME types that will be compressed.
3126 offload makes haproxy work as a compression offloader only (see notes).
3127
3128 The currently supported algorithms are :
Willy Tarreauc91840a2015-03-28 17:00:39 +01003129 identity this is mostly for debugging, and it was useful for developing
3130 the compression feature. Identity does not apply any change on
3131 data.
Cyril Bonté316a8cf2012-11-11 13:38:27 +01003132
Willy Tarreauc91840a2015-03-28 17:00:39 +01003133 gzip applies gzip compression. This setting is only available when
Baptiste Assmannf085d632015-12-21 17:57:32 +01003134 support for zlib or libslz was built in.
Willy Tarreauc91840a2015-03-28 17:00:39 +01003135
3136 deflate same as "gzip", but with deflate algorithm and zlib format.
3137 Note that this algorithm has ambiguous support on many
3138 browsers and no support at all from recent ones. It is
3139 strongly recommended not to use it for anything else than
3140 experimentation. This setting is only available when support
Baptiste Assmannf085d632015-12-21 17:57:32 +01003141 for zlib or libslz was built in.
Cyril Bonté316a8cf2012-11-11 13:38:27 +01003142
Willy Tarreauc91840a2015-03-28 17:00:39 +01003143 raw-deflate same as "deflate" without the zlib wrapper, and used as an
3144 alternative when the browser wants "deflate". All major
3145 browsers understand it and despite violating the standards,
3146 it is known to work better than "deflate", at least on MSIE
3147 and some versions of Safari. Do not use it in conjunction
3148 with "deflate", use either one or the other since both react
3149 to the same Accept-Encoding token. This setting is only
Baptiste Assmannf085d632015-12-21 17:57:32 +01003150 available when support for zlib or libslz was built in.
Cyril Bonté316a8cf2012-11-11 13:38:27 +01003151
Dmitry Sivachenko87c208b2012-11-22 20:03:26 +04003152 Compression will be activated depending on the Accept-Encoding request
Cyril Bonté316a8cf2012-11-11 13:38:27 +01003153 header. With identity, it does not take care of that header.
Dmitry Sivachenkoc9f3b452012-11-28 17:47:11 +04003154 If backend servers support HTTP compression, these directives
3155 will be no-op: haproxy will see the compressed response and will not
3156 compress again. If backend servers do not support HTTP compression and
3157 there is Accept-Encoding header in request, haproxy will compress the
3158 matching response.
Willy Tarreau70737d12012-10-27 00:34:28 +02003159
3160 The "offload" setting makes haproxy remove the Accept-Encoding header to
3161 prevent backend servers from compressing responses. It is strongly
3162 recommended not to do this because this means that all the compression work
3163 will be done on the single point where haproxy is located. However in some
3164 deployment scenarios, haproxy may be installed in front of a buggy gateway
Dmitry Sivachenkoc9f3b452012-11-28 17:47:11 +04003165 with broken HTTP compression implementation which can't be turned off.
3166 In that case haproxy can be used to prevent that gateway from emitting
3167 invalid payloads. In this case, simply removing the header in the
3168 configuration does not work because it applies before the header is parsed,
3169 so that prevents haproxy from compressing. The "offload" setting should
Willy Tarreauffea9fd2014-07-12 16:37:02 +02003170 then be used for such scenarios. Note: for now, the "offload" setting is
3171 ignored when set in a defaults section.
William Lallemand82fe75c2012-10-23 10:25:10 +02003172
William Lallemand05097442012-11-20 12:14:28 +01003173 Compression is disabled when:
Baptiste Assmann650d53d2013-01-05 15:44:44 +01003174 * the request does not advertise a supported compression algorithm in the
3175 "Accept-Encoding" header
3176 * the response message is not HTTP/1.1
Tim Duesterhusbb48c9a2019-01-30 23:46:04 +01003177 * HTTP status code is not one of 200, 201, 202, or 203
Baptiste Assmann650d53d2013-01-05 15:44:44 +01003178 * response contain neither a "Content-Length" header nor a
3179 "Transfer-Encoding" whose last value is "chunked"
3180 * response contains a "Content-Type" header whose first value starts with
3181 "multipart"
3182 * the response contains the "no-transform" value in the "Cache-control"
3183 header
3184 * User-Agent matches "Mozilla/4" unless it is MSIE 6 with XP SP2, or MSIE 7
3185 and later
3186 * The response contains a "Content-Encoding" header, indicating that the
3187 response is already compressed (see compression offload)
Tim Duesterhusbb48c9a2019-01-30 23:46:04 +01003188 * The response contains an invalid "ETag" header or multiple ETag headers
William Lallemand05097442012-11-20 12:14:28 +01003189
Tim Duesterhusb229f012019-01-29 16:38:56 +01003190 Note: The compression does not emit the Warning header.
William Lallemand05097442012-11-20 12:14:28 +01003191
William Lallemand82fe75c2012-10-23 10:25:10 +02003192 Examples :
3193 compression algo gzip
3194 compression type text/html text/plain
Willy Tarreau0ba27502007-12-24 16:55:16 +01003195
Christopher Fauletc3fe5332016-04-07 15:30:10 +02003196
Willy Tarreau55165fe2009-05-10 12:02:55 +02003197cookie <name> [ rewrite | insert | prefix ] [ indirect ] [ nocache ]
Willy Tarreau4992dd22012-05-31 21:02:17 +02003198 [ postonly ] [ preserve ] [ httponly ] [ secure ]
3199 [ domain <domain> ]* [ maxidle <idle> ] [ maxlife <life> ]
Olivier Houchard4e694042017-03-14 20:01:29 +01003200 [ dynamic ]
Willy Tarreau0ba27502007-12-24 16:55:16 +01003201 Enable cookie-based persistence in a backend.
3202 May be used in sections : defaults | frontend | listen | backend
3203 yes | no | yes | yes
3204 Arguments :
3205 <name> is the name of the cookie which will be monitored, modified or
3206 inserted in order to bring persistence. This cookie is sent to
3207 the client via a "Set-Cookie" header in the response, and is
3208 brought back by the client in a "Cookie" header in all requests.
3209 Special care should be taken to choose a name which does not
3210 conflict with any likely application cookie. Also, if the same
Davor Ocelice9ed2812017-12-25 17:49:28 +01003211 backends are subject to be used by the same clients (e.g.
Willy Tarreau0ba27502007-12-24 16:55:16 +01003212 HTTP/HTTPS), care should be taken to use different cookie names
3213 between all backends if persistence between them is not desired.
3214
3215 rewrite This keyword indicates that the cookie will be provided by the
3216 server and that haproxy will have to modify its value to set the
3217 server's identifier in it. This mode is handy when the management
3218 of complex combinations of "Set-cookie" and "Cache-control"
3219 headers is left to the application. The application can then
3220 decide whether or not it is appropriate to emit a persistence
Lukas Tribusf01a9cd2016-02-03 18:09:37 +01003221 cookie. Since all responses should be monitored, this mode
3222 doesn't work in HTTP tunnel mode. Unless the application
Davor Ocelice9ed2812017-12-25 17:49:28 +01003223 behavior is very complex and/or broken, it is advised not to
Lukas Tribusf01a9cd2016-02-03 18:09:37 +01003224 start with this mode for new deployments. This keyword is
3225 incompatible with "insert" and "prefix".
Willy Tarreau0ba27502007-12-24 16:55:16 +01003226
3227 insert This keyword indicates that the persistence cookie will have to
Willy Tarreaua79094d2010-08-31 22:54:15 +02003228 be inserted by haproxy in server responses if the client did not
Willy Tarreauba4c5be2010-10-23 12:46:42 +02003229
Willy Tarreaua79094d2010-08-31 22:54:15 +02003230 already have a cookie that would have permitted it to access this
Willy Tarreauba4c5be2010-10-23 12:46:42 +02003231 server. When used without the "preserve" option, if the server
Michael Prokop4438c602019-05-24 10:25:45 +02003232 emits a cookie with the same name, it will be removed before
Davor Ocelice9ed2812017-12-25 17:49:28 +01003233 processing. For this reason, this mode can be used to upgrade
Willy Tarreauba4c5be2010-10-23 12:46:42 +02003234 existing configurations running in the "rewrite" mode. The cookie
3235 will only be a session cookie and will not be stored on the
3236 client's disk. By default, unless the "indirect" option is added,
3237 the server will see the cookies emitted by the client. Due to
3238 caching effects, it is generally wise to add the "nocache" or
3239 "postonly" keywords (see below). The "insert" keyword is not
3240 compatible with "rewrite" and "prefix".
Willy Tarreau0ba27502007-12-24 16:55:16 +01003241
3242 prefix This keyword indicates that instead of relying on a dedicated
3243 cookie for the persistence, an existing one will be completed.
3244 This may be needed in some specific environments where the client
3245 does not support more than one single cookie and the application
3246 already needs it. In this case, whenever the server sets a cookie
3247 named <name>, it will be prefixed with the server's identifier
3248 and a delimiter. The prefix will be removed from all client
3249 requests so that the server still finds the cookie it emitted.
3250 Since all requests and responses are subject to being modified,
Lukas Tribusf01a9cd2016-02-03 18:09:37 +01003251 this mode doesn't work with tunnel mode. The "prefix" keyword is
Willy Tarreau37229df2011-10-17 12:24:55 +02003252 not compatible with "rewrite" and "insert". Note: it is highly
3253 recommended not to use "indirect" with "prefix", otherwise server
3254 cookie updates would not be sent to clients.
Willy Tarreau0ba27502007-12-24 16:55:16 +01003255
Willy Tarreaua79094d2010-08-31 22:54:15 +02003256 indirect When this option is specified, no cookie will be emitted to a
3257 client which already has a valid one for the server which has
3258 processed the request. If the server sets such a cookie itself,
Willy Tarreauba4c5be2010-10-23 12:46:42 +02003259 it will be removed, unless the "preserve" option is also set. In
3260 "insert" mode, this will additionally remove cookies from the
3261 requests transmitted to the server, making the persistence
3262 mechanism totally transparent from an application point of view.
Willy Tarreau37229df2011-10-17 12:24:55 +02003263 Note: it is highly recommended not to use "indirect" with
3264 "prefix", otherwise server cookie updates would not be sent to
3265 clients.
Willy Tarreau0ba27502007-12-24 16:55:16 +01003266
3267 nocache This option is recommended in conjunction with the insert mode
3268 when there is a cache between the client and HAProxy, as it
3269 ensures that a cacheable response will be tagged non-cacheable if
3270 a cookie needs to be inserted. This is important because if all
3271 persistence cookies are added on a cacheable home page for
3272 instance, then all customers will then fetch the page from an
3273 outer cache and will all share the same persistence cookie,
3274 leading to one server receiving much more traffic than others.
3275 See also the "insert" and "postonly" options.
3276
3277 postonly This option ensures that cookie insertion will only be performed
3278 on responses to POST requests. It is an alternative to the
3279 "nocache" option, because POST responses are not cacheable, so
3280 this ensures that the persistence cookie will never get cached.
3281 Since most sites do not need any sort of persistence before the
3282 first POST which generally is a login request, this is a very
3283 efficient method to optimize caching without risking to find a
3284 persistence cookie in the cache.
3285 See also the "insert" and "nocache" options.
3286
Willy Tarreauba4c5be2010-10-23 12:46:42 +02003287 preserve This option may only be used with "insert" and/or "indirect". It
3288 allows the server to emit the persistence cookie itself. In this
3289 case, if a cookie is found in the response, haproxy will leave it
3290 untouched. This is useful in order to end persistence after a
3291 logout request for instance. For this, the server just has to
Davor Ocelice9ed2812017-12-25 17:49:28 +01003292 emit a cookie with an invalid value (e.g. empty) or with a date in
Willy Tarreauba4c5be2010-10-23 12:46:42 +02003293 the past. By combining this mechanism with the "disable-on-404"
3294 check option, it is possible to perform a completely graceful
3295 shutdown because users will definitely leave the server after
3296 they logout.
3297
Willy Tarreau4992dd22012-05-31 21:02:17 +02003298 httponly This option tells haproxy to add an "HttpOnly" cookie attribute
3299 when a cookie is inserted. This attribute is used so that a
3300 user agent doesn't share the cookie with non-HTTP components.
3301 Please check RFC6265 for more information on this attribute.
3302
3303 secure This option tells haproxy to add a "Secure" cookie attribute when
3304 a cookie is inserted. This attribute is used so that a user agent
3305 never emits this cookie over non-secure channels, which means
3306 that a cookie learned with this flag will be presented only over
3307 SSL/TLS connections. Please check RFC6265 for more information on
3308 this attribute.
3309
Krzysztof Piotr Oledzkiefe3b6f2008-05-23 23:49:32 +02003310 domain This option allows to specify the domain at which a cookie is
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01003311 inserted. It requires exactly one parameter: a valid domain
Willy Tarreau68a897b2009-12-03 23:28:34 +01003312 name. If the domain begins with a dot, the browser is allowed to
3313 use it for any host ending with that name. It is also possible to
3314 specify several domain names by invoking this option multiple
3315 times. Some browsers might have small limits on the number of
3316 domains, so be careful when doing that. For the record, sending
3317 10 domains to MSIE 6 or Firefox 2 works as expected.
Krzysztof Piotr Oledzkiefe3b6f2008-05-23 23:49:32 +02003318
Willy Tarreau996a92c2010-10-13 19:30:47 +02003319 maxidle This option allows inserted cookies to be ignored after some idle
3320 time. It only works with insert-mode cookies. When a cookie is
3321 sent to the client, the date this cookie was emitted is sent too.
3322 Upon further presentations of this cookie, if the date is older
3323 than the delay indicated by the parameter (in seconds), it will
3324 be ignored. Otherwise, it will be refreshed if needed when the
3325 response is sent to the client. This is particularly useful to
3326 prevent users who never close their browsers from remaining for
Davor Ocelice9ed2812017-12-25 17:49:28 +01003327 too long on the same server (e.g. after a farm size change). When
Willy Tarreau996a92c2010-10-13 19:30:47 +02003328 this option is set and a cookie has no date, it is always
3329 accepted, but gets refreshed in the response. This maintains the
3330 ability for admins to access their sites. Cookies that have a
3331 date in the future further than 24 hours are ignored. Doing so
3332 lets admins fix timezone issues without risking kicking users off
3333 the site.
3334
3335 maxlife This option allows inserted cookies to be ignored after some life
3336 time, whether they're in use or not. It only works with insert
3337 mode cookies. When a cookie is first sent to the client, the date
3338 this cookie was emitted is sent too. Upon further presentations
3339 of this cookie, if the date is older than the delay indicated by
3340 the parameter (in seconds), it will be ignored. If the cookie in
3341 the request has no date, it is accepted and a date will be set.
3342 Cookies that have a date in the future further than 24 hours are
3343 ignored. Doing so lets admins fix timezone issues without risking
3344 kicking users off the site. Contrary to maxidle, this value is
3345 not refreshed, only the first visit date counts. Both maxidle and
3346 maxlife may be used at the time. This is particularly useful to
3347 prevent users who never close their browsers from remaining for
Davor Ocelice9ed2812017-12-25 17:49:28 +01003348 too long on the same server (e.g. after a farm size change). This
Willy Tarreau996a92c2010-10-13 19:30:47 +02003349 is stronger than the maxidle method in that it forces a
3350 redispatch after some absolute delay.
3351
Olivier Houchard4e694042017-03-14 20:01:29 +01003352 dynamic Activate dynamic cookies. When used, a session cookie is
3353 dynamically created for each server, based on the IP and port
3354 of the server, and a secret key, specified in the
3355 "dynamic-cookie-key" backend directive.
3356 The cookie will be regenerated each time the IP address change,
3357 and is only generated for IPv4/IPv6.
3358
Willy Tarreau0ba27502007-12-24 16:55:16 +01003359 There can be only one persistence cookie per HTTP backend, and it can be
3360 declared in a defaults section. The value of the cookie will be the value
3361 indicated after the "cookie" keyword in a "server" statement. If no cookie
3362 is declared for a given server, the cookie is not set.
Willy Tarreau6a06a402007-07-15 20:15:28 +02003363
Willy Tarreau0ba27502007-12-24 16:55:16 +01003364 Examples :
3365 cookie JSESSIONID prefix
3366 cookie SRV insert indirect nocache
3367 cookie SRV insert postonly indirect
Willy Tarreau996a92c2010-10-13 19:30:47 +02003368 cookie SRV insert indirect nocache maxidle 30m maxlife 8h
Willy Tarreau0ba27502007-12-24 16:55:16 +01003369
Willy Tarreau294d0f02015-08-10 19:40:12 +02003370 See also : "balance source", "capture cookie", "server" and "ignore-persist".
Willy Tarreau0ba27502007-12-24 16:55:16 +01003371
Willy Tarreau983e01e2010-01-11 18:42:06 +01003372
Thierry FOURNIERa0a1b752015-05-26 17:44:32 +02003373declare capture [ request | response ] len <length>
3374 Declares a capture slot.
3375 May be used in sections : defaults | frontend | listen | backend
3376 no | yes | yes | no
3377 Arguments:
3378 <length> is the length allowed for the capture.
3379
3380 This declaration is only available in the frontend or listen section, but the
3381 reserved slot can be used in the backends. The "request" keyword allocates a
3382 capture slot for use in the request, and "response" allocates a capture slot
3383 for use in the response.
3384
3385 See also: "capture-req", "capture-res" (sample converters),
Baptiste Assmann5ac425c2015-10-21 23:13:46 +02003386 "capture.req.hdr", "capture.res.hdr" (sample fetches),
Thierry FOURNIERa0a1b752015-05-26 17:44:32 +02003387 "http-request capture" and "http-response capture".
3388
3389
Krzysztof Piotr Oledzkic6df0662010-01-05 16:38:49 +01003390default-server [param*]
3391 Change default options for a server in a backend
3392 May be used in sections : defaults | frontend | listen | backend
3393 yes | no | yes | yes
3394 Arguments:
Willy Tarreau983e01e2010-01-11 18:42:06 +01003395 <param*> is a list of parameters for this server. The "default-server"
3396 keyword accepts an important number of options and has a complete
3397 section dedicated to it. Please refer to section 5 for more
3398 details.
Krzysztof Piotr Oledzkic6df0662010-01-05 16:38:49 +01003399
Willy Tarreau983e01e2010-01-11 18:42:06 +01003400 Example :
Krzysztof Piotr Oledzkic6df0662010-01-05 16:38:49 +01003401 default-server inter 1000 weight 13
3402
3403 See also: "server" and section 5 about server options
Willy Tarreau0ba27502007-12-24 16:55:16 +01003404
Willy Tarreau983e01e2010-01-11 18:42:06 +01003405
Willy Tarreau0ba27502007-12-24 16:55:16 +01003406default_backend <backend>
3407 Specify the backend to use when no "use_backend" rule has been matched.
3408 May be used in sections : defaults | frontend | listen | backend
3409 yes | yes | yes | no
3410 Arguments :
3411 <backend> is the name of the backend to use.
3412
3413 When doing content-switching between frontend and backends using the
3414 "use_backend" keyword, it is often useful to indicate which backend will be
3415 used when no rule has matched. It generally is the dynamic backend which
3416 will catch all undetermined requests.
3417
Willy Tarreau0ba27502007-12-24 16:55:16 +01003418 Example :
3419
3420 use_backend dynamic if url_dyn
3421 use_backend static if url_css url_img extension_img
3422 default_backend dynamic
3423
Willy Tarreau98d04852015-05-26 12:18:29 +02003424 See also : "use_backend"
Willy Tarreau2769aa02007-12-27 18:26:09 +01003425
Willy Tarreau0ba27502007-12-24 16:55:16 +01003426
Baptiste Assmann27f51342013-10-09 06:51:49 +02003427description <string>
3428 Describe a listen, frontend or backend.
3429 May be used in sections : defaults | frontend | listen | backend
3430 no | yes | yes | yes
3431 Arguments : string
3432
3433 Allows to add a sentence to describe the related object in the HAProxy HTML
3434 stats page. The description will be printed on the right of the object name
3435 it describes.
3436 No need to backslash spaces in the <string> arguments.
3437
3438
Willy Tarreau0ba27502007-12-24 16:55:16 +01003439disabled
3440 Disable a proxy, frontend or backend.
3441 May be used in sections : defaults | frontend | listen | backend
3442 yes | yes | yes | yes
3443 Arguments : none
3444
3445 The "disabled" keyword is used to disable an instance, mainly in order to
3446 liberate a listening port or to temporarily disable a service. The instance
3447 will still be created and its configuration will be checked, but it will be
3448 created in the "stopped" state and will appear as such in the statistics. It
3449 will not receive any traffic nor will it send any health-checks or logs. It
3450 is possible to disable many instances at once by adding the "disabled"
3451 keyword in a "defaults" section.
3452
3453 See also : "enabled"
3454
3455
Willy Tarreau5ce94572010-06-07 14:35:41 +02003456dispatch <address>:<port>
3457 Set a default server address
3458 May be used in sections : defaults | frontend | listen | backend
3459 no | no | yes | yes
Cyril Bonté108cf6e2012-04-21 23:30:29 +02003460 Arguments :
Willy Tarreau5ce94572010-06-07 14:35:41 +02003461
3462 <address> is the IPv4 address of the default server. Alternatively, a
3463 resolvable hostname is supported, but this name will be resolved
3464 during start-up.
3465
3466 <ports> is a mandatory port specification. All connections will be sent
3467 to this port, and it is not permitted to use port offsets as is
3468 possible with normal servers.
3469
Willy Tarreau787aed52011-04-15 06:45:37 +02003470 The "dispatch" keyword designates a default server for use when no other
Willy Tarreau5ce94572010-06-07 14:35:41 +02003471 server can take the connection. In the past it was used to forward non
3472 persistent connections to an auxiliary load balancer. Due to its simple
3473 syntax, it has also been used for simple TCP relays. It is recommended not to
3474 use it for more clarity, and to use the "server" directive instead.
3475
3476 See also : "server"
3477
Olivier Houchard4e694042017-03-14 20:01:29 +01003478
3479dynamic-cookie-key <string>
3480 Set the dynamic cookie secret key for a backend.
3481 May be used in sections : defaults | frontend | listen | backend
3482 yes | no | yes | yes
3483 Arguments : The secret key to be used.
3484
3485 When dynamic cookies are enabled (see the "dynamic" directive for cookie),
Davor Ocelice9ed2812017-12-25 17:49:28 +01003486 a dynamic cookie is created for each server (unless one is explicitly
Olivier Houchard4e694042017-03-14 20:01:29 +01003487 specified on the "server" line), using a hash of the IP address of the
3488 server, the TCP port, and the secret key.
Davor Ocelice9ed2812017-12-25 17:49:28 +01003489 That way, we can ensure session persistence across multiple load-balancers,
Olivier Houchard4e694042017-03-14 20:01:29 +01003490 even if servers are dynamically added or removed.
Willy Tarreau5ce94572010-06-07 14:35:41 +02003491
Willy Tarreau0ba27502007-12-24 16:55:16 +01003492enabled
3493 Enable a proxy, frontend or backend.
3494 May be used in sections : defaults | frontend | listen | backend
3495 yes | yes | yes | yes
3496 Arguments : none
3497
3498 The "enabled" keyword is used to explicitly enable an instance, when the
3499 defaults has been set to "disabled". This is very rarely used.
3500
3501 See also : "disabled"
3502
3503
3504errorfile <code> <file>
3505 Return a file contents instead of errors generated by HAProxy
3506 May be used in sections : defaults | frontend | listen | backend
3507 yes | yes | yes | yes
3508 Arguments :
3509 <code> is the HTTP status code. Currently, HAProxy is capable of
Olivier Houchard51a76d82017-10-02 16:12:07 +02003510 generating codes 200, 400, 403, 405, 408, 425, 429, 500, 502,
3511 503, and 504.
Willy Tarreau0ba27502007-12-24 16:55:16 +01003512
3513 <file> designates a file containing the full HTTP response. It is
Willy Tarreaud2a4aa22008-01-31 15:28:22 +01003514 recommended to follow the common practice of appending ".http" to
Willy Tarreau0ba27502007-12-24 16:55:16 +01003515 the filename so that people do not confuse the response with HTML
Willy Tarreau59140a22009-02-22 12:02:30 +01003516 error pages, and to use absolute paths, since files are read
3517 before any chroot is performed.
Willy Tarreau0ba27502007-12-24 16:55:16 +01003518
3519 It is important to understand that this keyword is not meant to rewrite
3520 errors returned by the server, but errors detected and returned by HAProxy.
3521 This is why the list of supported errors is limited to a small set.
3522
Willy Tarreauae94d4d2011-05-11 16:28:49 +02003523 Code 200 is emitted in response to requests matching a "monitor-uri" rule.
3524
Willy Tarreau0ba27502007-12-24 16:55:16 +01003525 The files are returned verbatim on the TCP socket. This allows any trick such
3526 as redirections to another URL or site, as well as tricks to clean cookies,
3527 force enable or disable caching, etc... The package provides default error
3528 files returning the same contents as default errors.
3529
Willy Tarreau59140a22009-02-22 12:02:30 +01003530 The files should not exceed the configured buffer size (BUFSIZE), which
3531 generally is 8 or 16 kB, otherwise they will be truncated. It is also wise
Davor Ocelice9ed2812017-12-25 17:49:28 +01003532 not to put any reference to local contents (e.g. images) in order to avoid
Willy Tarreau59140a22009-02-22 12:02:30 +01003533 loops between the client and HAProxy when all servers are down, causing an
3534 error to be returned instead of an image. For better HTTP compliance, it is
3535 recommended that all header lines end with CR-LF and not LF alone.
3536
Willy Tarreau0ba27502007-12-24 16:55:16 +01003537 The files are read at the same time as the configuration and kept in memory.
3538 For this reason, the errors continue to be returned even when the process is
3539 chrooted, and no file change is considered while the process is running. A
Willy Tarreauc27debf2008-01-06 08:57:02 +01003540 simple method for developing those files consists in associating them to the
Willy Tarreau0ba27502007-12-24 16:55:16 +01003541 403 status code and interrogating a blocked URL.
3542
3543 See also : "errorloc", "errorloc302", "errorloc303"
3544
Willy Tarreau59140a22009-02-22 12:02:30 +01003545 Example :
3546 errorfile 400 /etc/haproxy/errorfiles/400badreq.http
Willy Tarreau989222a2016-01-15 10:26:26 +01003547 errorfile 408 /dev/null # work around Chrome pre-connect bug
Willy Tarreau59140a22009-02-22 12:02:30 +01003548 errorfile 403 /etc/haproxy/errorfiles/403forbid.http
3549 errorfile 503 /etc/haproxy/errorfiles/503sorry.http
3550
Willy Tarreau2769aa02007-12-27 18:26:09 +01003551
3552errorloc <code> <url>
3553errorloc302 <code> <url>
3554 Return an HTTP redirection to a URL instead of errors generated by HAProxy
3555 May be used in sections : defaults | frontend | listen | backend
3556 yes | yes | yes | yes
3557 Arguments :
3558 <code> is the HTTP status code. Currently, HAProxy is capable of
Olivier Houchard51a76d82017-10-02 16:12:07 +02003559 generating codes 200, 400, 403, 405, 408, 425, 429, 500, 502,
3560 503, and 504.
Willy Tarreau2769aa02007-12-27 18:26:09 +01003561
3562 <url> it is the exact contents of the "Location" header. It may contain
3563 either a relative URI to an error page hosted on the same site,
3564 or an absolute URI designating an error page on another site.
3565 Special care should be given to relative URIs to avoid redirect
Davor Ocelice9ed2812017-12-25 17:49:28 +01003566 loops if the URI itself may generate the same error (e.g. 500).
Willy Tarreau2769aa02007-12-27 18:26:09 +01003567
3568 It is important to understand that this keyword is not meant to rewrite
3569 errors returned by the server, but errors detected and returned by HAProxy.
3570 This is why the list of supported errors is limited to a small set.
3571
Willy Tarreauae94d4d2011-05-11 16:28:49 +02003572 Code 200 is emitted in response to requests matching a "monitor-uri" rule.
3573
Willy Tarreau2769aa02007-12-27 18:26:09 +01003574 Note that both keyword return the HTTP 302 status code, which tells the
3575 client to fetch the designated URL using the same HTTP method. This can be
3576 quite problematic in case of non-GET methods such as POST, because the URL
3577 sent to the client might not be allowed for something other than GET. To
Willy Tarreau989222a2016-01-15 10:26:26 +01003578 work around this problem, please use "errorloc303" which send the HTTP 303
Willy Tarreau2769aa02007-12-27 18:26:09 +01003579 status code, indicating to the client that the URL must be fetched with a GET
3580 request.
3581
3582 See also : "errorfile", "errorloc303"
3583
3584
3585errorloc303 <code> <url>
3586 Return an HTTP redirection to a URL instead of errors generated by HAProxy
3587 May be used in sections : defaults | frontend | listen | backend
3588 yes | yes | yes | yes
3589 Arguments :
3590 <code> is the HTTP status code. Currently, HAProxy is capable of
Olivier Houchard51a76d82017-10-02 16:12:07 +02003591 generating codes 200, 400, 403, 405, 408, 425, 429, 500, 502,
3592 503, and 504.
Willy Tarreau2769aa02007-12-27 18:26:09 +01003593
3594 <url> it is the exact contents of the "Location" header. It may contain
3595 either a relative URI to an error page hosted on the same site,
3596 or an absolute URI designating an error page on another site.
3597 Special care should be given to relative URIs to avoid redirect
Davor Ocelice9ed2812017-12-25 17:49:28 +01003598 loops if the URI itself may generate the same error (e.g. 500).
Willy Tarreau2769aa02007-12-27 18:26:09 +01003599
3600 It is important to understand that this keyword is not meant to rewrite
3601 errors returned by the server, but errors detected and returned by HAProxy.
3602 This is why the list of supported errors is limited to a small set.
3603
Willy Tarreauae94d4d2011-05-11 16:28:49 +02003604 Code 200 is emitted in response to requests matching a "monitor-uri" rule.
3605
Willy Tarreau2769aa02007-12-27 18:26:09 +01003606 Note that both keyword return the HTTP 303 status code, which tells the
3607 client to fetch the designated URL using the same HTTP GET method. This
3608 solves the usual problems associated with "errorloc" and the 302 code. It is
3609 possible that some very old browsers designed before HTTP/1.1 do not support
Willy Tarreaud2a4aa22008-01-31 15:28:22 +01003610 it, but no such problem has been reported till now.
Willy Tarreau2769aa02007-12-27 18:26:09 +01003611
3612 See also : "errorfile", "errorloc", "errorloc302"
3613
3614
Simon Horman51a1cf62015-02-03 13:00:44 +09003615email-alert from <emailaddr>
3616 Declare the from email address to be used in both the envelope and header
Davor Ocelice9ed2812017-12-25 17:49:28 +01003617 of email alerts. This is the address that email alerts are sent from.
Simon Horman51a1cf62015-02-03 13:00:44 +09003618 May be used in sections: defaults | frontend | listen | backend
3619 yes | yes | yes | yes
3620
3621 Arguments :
3622
3623 <emailaddr> is the from email address to use when sending email alerts
3624
3625 Also requires "email-alert mailers" and "email-alert to" to be set
3626 and if so sending email alerts is enabled for the proxy.
3627
Simon Horman64e34162015-02-06 11:11:57 +09003628 See also : "email-alert level", "email-alert mailers",
Cyril Bonté307ee1e2015-09-28 23:16:06 +02003629 "email-alert myhostname", "email-alert to", section 3.6 about
3630 mailers.
Simon Horman64e34162015-02-06 11:11:57 +09003631
3632
3633email-alert level <level>
3634 Declare the maximum log level of messages for which email alerts will be
3635 sent. This acts as a filter on the sending of email alerts.
3636 May be used in sections: defaults | frontend | listen | backend
3637 yes | yes | yes | yes
3638
3639 Arguments :
3640
3641 <level> One of the 8 syslog levels:
3642 emerg alert crit err warning notice info debug
3643 The above syslog levels are ordered from lowest to highest.
3644
3645 By default level is alert
3646
3647 Also requires "email-alert from", "email-alert mailers" and
3648 "email-alert to" to be set and if so sending email alerts is enabled
3649 for the proxy.
3650
Simon Horman1421e212015-04-30 13:10:35 +09003651 Alerts are sent when :
3652
3653 * An un-paused server is marked as down and <level> is alert or lower
3654 * A paused server is marked as down and <level> is notice or lower
3655 * A server is marked as up or enters the drain state and <level>
3656 is notice or lower
3657 * "option log-health-checks" is enabled, <level> is info or lower,
3658 and a health check status update occurs
3659
Simon Horman64e34162015-02-06 11:11:57 +09003660 See also : "email-alert from", "email-alert mailers",
3661 "email-alert myhostname", "email-alert to",
Simon Horman51a1cf62015-02-03 13:00:44 +09003662 section 3.6 about mailers.
3663
3664
3665email-alert mailers <mailersect>
3666 Declare the mailers to be used when sending email alerts
3667 May be used in sections: defaults | frontend | listen | backend
3668 yes | yes | yes | yes
3669
3670 Arguments :
3671
3672 <mailersect> is the name of the mailers section to send email alerts.
3673
3674 Also requires "email-alert from" and "email-alert to" to be set
3675 and if so sending email alerts is enabled for the proxy.
3676
Simon Horman64e34162015-02-06 11:11:57 +09003677 See also : "email-alert from", "email-alert level", "email-alert myhostname",
3678 "email-alert to", section 3.6 about mailers.
Simon Horman51a1cf62015-02-03 13:00:44 +09003679
3680
3681email-alert myhostname <hostname>
3682 Declare the to hostname address to be used when communicating with
3683 mailers.
3684 May be used in sections: defaults | frontend | listen | backend
3685 yes | yes | yes | yes
3686
3687 Arguments :
3688
Baptiste Assmann738bad92015-12-21 15:27:53 +01003689 <hostname> is the hostname to use when communicating with mailers
Simon Horman51a1cf62015-02-03 13:00:44 +09003690
3691 By default the systems hostname is used.
3692
3693 Also requires "email-alert from", "email-alert mailers" and
3694 "email-alert to" to be set and if so sending email alerts is enabled
3695 for the proxy.
3696
Simon Horman64e34162015-02-06 11:11:57 +09003697 See also : "email-alert from", "email-alert level", "email-alert mailers",
3698 "email-alert to", section 3.6 about mailers.
Simon Horman51a1cf62015-02-03 13:00:44 +09003699
3700
3701email-alert to <emailaddr>
Davor Ocelice9ed2812017-12-25 17:49:28 +01003702 Declare both the recipient address in the envelope and to address in the
Simon Horman51a1cf62015-02-03 13:00:44 +09003703 header of email alerts. This is the address that email alerts are sent to.
3704 May be used in sections: defaults | frontend | listen | backend
3705 yes | yes | yes | yes
3706
3707 Arguments :
3708
3709 <emailaddr> is the to email address to use when sending email alerts
3710
3711 Also requires "email-alert mailers" and "email-alert to" to be set
3712 and if so sending email alerts is enabled for the proxy.
3713
Simon Horman64e34162015-02-06 11:11:57 +09003714 See also : "email-alert from", "email-alert level", "email-alert mailers",
Simon Horman51a1cf62015-02-03 13:00:44 +09003715 "email-alert myhostname", section 3.6 about mailers.
3716
3717
Willy Tarreau4de91492010-01-22 19:10:05 +01003718force-persist { if | unless } <condition>
3719 Declare a condition to force persistence on down servers
3720 May be used in sections: defaults | frontend | listen | backend
Cyril Bonté4288c5a2018-03-12 22:02:59 +01003721 no | no | yes | yes
Willy Tarreau4de91492010-01-22 19:10:05 +01003722
3723 By default, requests are not dispatched to down servers. It is possible to
3724 force this using "option persist", but it is unconditional and redispatches
3725 to a valid server if "option redispatch" is set. That leaves with very little
3726 possibilities to force some requests to reach a server which is artificially
3727 marked down for maintenance operations.
3728
3729 The "force-persist" statement allows one to declare various ACL-based
3730 conditions which, when met, will cause a request to ignore the down status of
3731 a server and still try to connect to it. That makes it possible to start a
3732 server, still replying an error to the health checks, and run a specially
3733 configured browser to test the service. Among the handy methods, one could
3734 use a specific source IP address, or a specific cookie. The cookie also has
3735 the advantage that it can easily be added/removed on the browser from a test
3736 page. Once the service is validated, it is then possible to open the service
3737 to the world by returning a valid response to health checks.
3738
3739 The forced persistence is enabled when an "if" condition is met, or unless an
3740 "unless" condition is met. The final redispatch is always disabled when this
3741 is used.
3742
Cyril Bonté0d4bf012010-04-25 23:21:46 +02003743 See also : "option redispatch", "ignore-persist", "persist",
Cyril Bontéa8e7bbc2010-04-25 22:29:29 +02003744 and section 7 about ACL usage.
Willy Tarreau4de91492010-01-22 19:10:05 +01003745
Christopher Fauletc3fe5332016-04-07 15:30:10 +02003746
3747filter <name> [param*]
3748 Add the filter <name> in the filter list attached to the proxy.
3749 May be used in sections : defaults | frontend | listen | backend
3750 no | yes | yes | yes
3751 Arguments :
3752 <name> is the name of the filter. Officially supported filters are
3753 referenced in section 9.
3754
Tim Düsterhus4896c442016-11-29 02:15:19 +01003755 <param*> is a list of parameters accepted by the filter <name>. The
Christopher Fauletc3fe5332016-04-07 15:30:10 +02003756 parsing of these parameters are the responsibility of the
Tim Düsterhus4896c442016-11-29 02:15:19 +01003757 filter. Please refer to the documentation of the corresponding
3758 filter (section 9) for all details on the supported parameters.
Christopher Fauletc3fe5332016-04-07 15:30:10 +02003759
3760 Multiple occurrences of the filter line can be used for the same proxy. The
3761 same filter can be referenced many times if needed.
3762
3763 Example:
3764 listen
3765 bind *:80
3766
3767 filter trace name BEFORE-HTTP-COMP
3768 filter compression
3769 filter trace name AFTER-HTTP-COMP
3770
3771 compression algo gzip
3772 compression offload
3773
3774 server srv1 192.168.0.1:80
3775
3776 See also : section 9.
3777
Willy Tarreau4de91492010-01-22 19:10:05 +01003778
Willy Tarreau2769aa02007-12-27 18:26:09 +01003779fullconn <conns>
3780 Specify at what backend load the servers will reach their maxconn
3781 May be used in sections : defaults | frontend | listen | backend
3782 yes | no | yes | yes
3783 Arguments :
3784 <conns> is the number of connections on the backend which will make the
3785 servers use the maximal number of connections.
3786
Willy Tarreau198a7442008-01-17 12:05:32 +01003787 When a server has a "maxconn" parameter specified, it means that its number
Willy Tarreau2769aa02007-12-27 18:26:09 +01003788 of concurrent connections will never go higher. Additionally, if it has a
Willy Tarreau198a7442008-01-17 12:05:32 +01003789 "minconn" parameter, it indicates a dynamic limit following the backend's
Willy Tarreau2769aa02007-12-27 18:26:09 +01003790 load. The server will then always accept at least <minconn> connections,
3791 never more than <maxconn>, and the limit will be on the ramp between both
3792 values when the backend has less than <conns> concurrent connections. This
3793 makes it possible to limit the load on the servers during normal loads, but
3794 push it further for important loads without overloading the servers during
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01003795 exceptional loads.
Willy Tarreau2769aa02007-12-27 18:26:09 +01003796
Willy Tarreaufbb78422011-06-05 15:38:35 +02003797 Since it's hard to get this value right, haproxy automatically sets it to
3798 10% of the sum of the maxconns of all frontends that may branch to this
Bertrand Jacquin702d44f2013-11-19 11:43:06 +01003799 backend (based on "use_backend" and "default_backend" rules). That way it's
3800 safe to leave it unset. However, "use_backend" involving dynamic names are
3801 not counted since there is no way to know if they could match or not.
Willy Tarreaufbb78422011-06-05 15:38:35 +02003802
Willy Tarreau2769aa02007-12-27 18:26:09 +01003803 Example :
3804 # The servers will accept between 100 and 1000 concurrent connections each
3805 # and the maximum of 1000 will be reached when the backend reaches 10000
3806 # connections.
3807 backend dynamic
3808 fullconn 10000
3809 server srv1 dyn1:80 minconn 100 maxconn 1000
3810 server srv2 dyn2:80 minconn 100 maxconn 1000
3811
3812 See also : "maxconn", "server"
3813
3814
3815grace <time>
3816 Maintain a proxy operational for some time after a soft stop
3817 May be used in sections : defaults | frontend | listen | backend
Cyril Bonté99ed3272010-01-24 23:29:44 +01003818 yes | yes | yes | yes
Willy Tarreau2769aa02007-12-27 18:26:09 +01003819 Arguments :
3820 <time> is the time (by default in milliseconds) for which the instance
3821 will remain operational with the frontend sockets still listening
3822 when a soft-stop is received via the SIGUSR1 signal.
3823
3824 This may be used to ensure that the services disappear in a certain order.
3825 This was designed so that frontends which are dedicated to monitoring by an
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01003826 external equipment fail immediately while other ones remain up for the time
Willy Tarreau2769aa02007-12-27 18:26:09 +01003827 needed by the equipment to detect the failure.
3828
3829 Note that currently, there is very little benefit in using this parameter,
3830 and it may in fact complicate the soft-reconfiguration process more than
3831 simplify it.
3832
Willy Tarreau0ba27502007-12-24 16:55:16 +01003833
Andrew Rodland17be45e2016-10-25 17:04:12 -04003834hash-balance-factor <factor>
3835 Specify the balancing factor for bounded-load consistent hashing
3836 May be used in sections : defaults | frontend | listen | backend
3837 yes | no | no | yes
3838 Arguments :
3839 <factor> is the control for the maximum number of concurrent requests to
3840 send to a server, expressed as a percentage of the average number
Frédéric Lécaille93d33162019-03-06 09:35:59 +01003841 of concurrent requests across all of the active servers.
Andrew Rodland17be45e2016-10-25 17:04:12 -04003842
3843 Specifying a "hash-balance-factor" for a server with "hash-type consistent"
3844 enables an algorithm that prevents any one server from getting too many
3845 requests at once, even if some hash buckets receive many more requests than
3846 others. Setting <factor> to 0 (the default) disables the feature. Otherwise,
3847 <factor> is a percentage greater than 100. For example, if <factor> is 150,
3848 then no server will be allowed to have a load more than 1.5 times the average.
3849 If server weights are used, they will be respected.
3850
3851 If the first-choice server is disqualified, the algorithm will choose another
3852 server based on the request hash, until a server with additional capacity is
3853 found. A higher <factor> allows more imbalance between the servers, while a
3854 lower <factor> means that more servers will be checked on average, affecting
3855 performance. Reasonable values are from 125 to 200.
3856
Willy Tarreau760e81d2018-05-03 07:20:40 +02003857 This setting is also used by "balance random" which internally relies on the
3858 consistent hashing mechanism.
3859
Andrew Rodland17be45e2016-10-25 17:04:12 -04003860 See also : "balance" and "hash-type".
3861
3862
Bhaskar Maddalab6c0ac92013-11-05 11:54:02 -05003863hash-type <method> <function> <modifier>
Willy Tarreau6b2e11b2009-10-01 07:52:15 +02003864 Specify a method to use for mapping hashes to servers
3865 May be used in sections : defaults | frontend | listen | backend
3866 yes | no | yes | yes
3867 Arguments :
Bhaskar98634f02013-10-29 23:30:51 -04003868 <method> is the method used to select a server from the hash computed by
3869 the <function> :
Willy Tarreau6b2e11b2009-10-01 07:52:15 +02003870
Bhaskar98634f02013-10-29 23:30:51 -04003871 map-based the hash table is a static array containing all alive servers.
3872 The hashes will be very smooth, will consider weights, but
3873 will be static in that weight changes while a server is up
3874 will be ignored. This means that there will be no slow start.
3875 Also, since a server is selected by its position in the array,
3876 most mappings are changed when the server count changes. This
3877 means that when a server goes up or down, or when a server is
3878 added to a farm, most connections will be redistributed to
3879 different servers. This can be inconvenient with caches for
3880 instance.
Willy Tarreau798a39c2010-11-24 15:04:29 +01003881
Bhaskar98634f02013-10-29 23:30:51 -04003882 consistent the hash table is a tree filled with many occurrences of each
3883 server. The hash key is looked up in the tree and the closest
3884 server is chosen. This hash is dynamic, it supports changing
3885 weights while the servers are up, so it is compatible with the
3886 slow start feature. It has the advantage that when a server
3887 goes up or down, only its associations are moved. When a
3888 server is added to the farm, only a few part of the mappings
3889 are redistributed, making it an ideal method for caches.
3890 However, due to its principle, the distribution will never be
3891 very smooth and it may sometimes be necessary to adjust a
3892 server's weight or its ID to get a more balanced distribution.
3893 In order to get the same distribution on multiple load
3894 balancers, it is important that all servers have the exact
Bhaskar Maddalab6c0ac92013-11-05 11:54:02 -05003895 same IDs. Note: consistent hash uses sdbm and avalanche if no
3896 hash function is specified.
Bhaskar98634f02013-10-29 23:30:51 -04003897
3898 <function> is the hash function to be used :
3899
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03003900 sdbm this function was created initially for sdbm (a public-domain
Bhaskar98634f02013-10-29 23:30:51 -04003901 reimplementation of ndbm) database library. It was found to do
3902 well in scrambling bits, causing better distribution of the keys
3903 and fewer splits. It also happens to be a good general hashing
Bhaskar Maddalab6c0ac92013-11-05 11:54:02 -05003904 function with good distribution, unless the total server weight
3905 is a multiple of 64, in which case applying the avalanche
3906 modifier may help.
Bhaskar98634f02013-10-29 23:30:51 -04003907
3908 djb2 this function was first proposed by Dan Bernstein many years ago
3909 on comp.lang.c. Studies have shown that for certain workload this
Bhaskar Maddalab6c0ac92013-11-05 11:54:02 -05003910 function provides a better distribution than sdbm. It generally
3911 works well with text-based inputs though it can perform extremely
3912 poorly with numeric-only input or when the total server weight is
3913 a multiple of 33, unless the avalanche modifier is also used.
3914
Willy Tarreaua0f42712013-11-14 14:30:35 +01003915 wt6 this function was designed for haproxy while testing other
3916 functions in the past. It is not as smooth as the other ones, but
3917 is much less sensible to the input data set or to the number of
3918 servers. It can make sense as an alternative to sdbm+avalanche or
3919 djb2+avalanche for consistent hashing or when hashing on numeric
3920 data such as a source IP address or a visitor identifier in a URL
3921 parameter.
3922
Willy Tarreau324f07f2015-01-20 19:44:50 +01003923 crc32 this is the most common CRC32 implementation as used in Ethernet,
3924 gzip, PNG, etc. It is slower than the other ones but may provide
3925 a better distribution or less predictable results especially when
3926 used on strings.
3927
Bhaskar Maddalab6c0ac92013-11-05 11:54:02 -05003928 <modifier> indicates an optional method applied after hashing the key :
3929
3930 avalanche This directive indicates that the result from the hash
3931 function above should not be used in its raw form but that
3932 a 4-byte full avalanche hash must be applied first. The
3933 purpose of this step is to mix the resulting bits from the
3934 previous hash in order to avoid any undesired effect when
3935 the input contains some limited values or when the number of
3936 servers is a multiple of one of the hash's components (64
3937 for SDBM, 33 for DJB2). Enabling avalanche tends to make the
3938 result less predictable, but it's also not as smooth as when
3939 using the original function. Some testing might be needed
3940 with some workloads. This hash is one of the many proposed
3941 by Bob Jenkins.
Willy Tarreau6b2e11b2009-10-01 07:52:15 +02003942
Bhaskar98634f02013-10-29 23:30:51 -04003943 The default hash type is "map-based" and is recommended for most usages. The
3944 default function is "sdbm", the selection of a function should be based on
3945 the range of the values being hashed.
Willy Tarreau6b2e11b2009-10-01 07:52:15 +02003946
Andrew Rodland17be45e2016-10-25 17:04:12 -04003947 See also : "balance", "hash-balance-factor", "server"
Willy Tarreau6b2e11b2009-10-01 07:52:15 +02003948
3949
Willy Tarreau0ba27502007-12-24 16:55:16 +01003950http-check disable-on-404
3951 Enable a maintenance mode upon HTTP/404 response to health-checks
3952 May be used in sections : defaults | frontend | listen | backend
Willy Tarreau2769aa02007-12-27 18:26:09 +01003953 yes | no | yes | yes
Willy Tarreau0ba27502007-12-24 16:55:16 +01003954 Arguments : none
3955
3956 When this option is set, a server which returns an HTTP code 404 will be
3957 excluded from further load-balancing, but will still receive persistent
3958 connections. This provides a very convenient method for Web administrators
3959 to perform a graceful shutdown of their servers. It is also important to note
3960 that a server which is detected as failed while it was in this mode will not
3961 generate an alert, just a notice. If the server responds 2xx or 3xx again, it
3962 will immediately be reinserted into the farm. The status on the stats page
3963 reports "NOLB" for a server in this mode. It is important to note that this
Willy Tarreaubd741542010-03-16 18:46:54 +01003964 option only works in conjunction with the "httpchk" option. If this option
3965 is used with "http-check expect", then it has precedence over it so that 404
3966 responses will still be considered as soft-stop.
3967
3968 See also : "option httpchk", "http-check expect"
3969
3970
3971http-check expect [!] <match> <pattern>
Jamie Gloudonaaa21002012-08-25 00:18:33 -04003972 Make HTTP health checks consider response contents or specific status codes
Willy Tarreaubd741542010-03-16 18:46:54 +01003973 May be used in sections : defaults | frontend | listen | backend
Willy Tarreau1ee51a62011-08-19 20:04:17 +02003974 yes | no | yes | yes
Willy Tarreaubd741542010-03-16 18:46:54 +01003975 Arguments :
3976 <match> is a keyword indicating how to look for a specific pattern in the
3977 response. The keyword may be one of "status", "rstatus",
Jamie Gloudonaaa21002012-08-25 00:18:33 -04003978 "string", or "rstring". The keyword may be preceded by an
Willy Tarreaubd741542010-03-16 18:46:54 +01003979 exclamation mark ("!") to negate the match. Spaces are allowed
3980 between the exclamation mark and the keyword. See below for more
3981 details on the supported keywords.
3982
3983 <pattern> is the pattern to look for. It may be a string or a regular
3984 expression. If the pattern contains spaces, they must be escaped
3985 with the usual backslash ('\').
3986
3987 By default, "option httpchk" considers that response statuses 2xx and 3xx
3988 are valid, and that others are invalid. When "http-check expect" is used,
3989 it defines what is considered valid or invalid. Only one "http-check"
3990 statement is supported in a backend. If a server fails to respond or times
3991 out, the check obviously fails. The available matches are :
3992
3993 status <string> : test the exact string match for the HTTP status code.
Jamie Gloudonaaa21002012-08-25 00:18:33 -04003994 A health check response will be considered valid if the
Willy Tarreaubd741542010-03-16 18:46:54 +01003995 response's status code is exactly this string. If the
3996 "status" keyword is prefixed with "!", then the response
3997 will be considered invalid if the status code matches.
3998
3999 rstatus <regex> : test a regular expression for the HTTP status code.
Jamie Gloudonaaa21002012-08-25 00:18:33 -04004000 A health check response will be considered valid if the
Willy Tarreaubd741542010-03-16 18:46:54 +01004001 response's status code matches the expression. If the
4002 "rstatus" keyword is prefixed with "!", then the response
4003 will be considered invalid if the status code matches.
4004 This is mostly used to check for multiple codes.
4005
4006 string <string> : test the exact string match in the HTTP response body.
Jamie Gloudonaaa21002012-08-25 00:18:33 -04004007 A health check response will be considered valid if the
Willy Tarreaubd741542010-03-16 18:46:54 +01004008 response's body contains this exact string. If the
4009 "string" keyword is prefixed with "!", then the response
4010 will be considered invalid if the body contains this
4011 string. This can be used to look for a mandatory word at
4012 the end of a dynamic page, or to detect a failure when a
Davor Ocelice9ed2812017-12-25 17:49:28 +01004013 specific error appears on the check page (e.g. a stack
Willy Tarreaubd741542010-03-16 18:46:54 +01004014 trace).
4015
4016 rstring <regex> : test a regular expression on the HTTP response body.
Jamie Gloudonaaa21002012-08-25 00:18:33 -04004017 A health check response will be considered valid if the
Willy Tarreaubd741542010-03-16 18:46:54 +01004018 response's body matches this expression. If the "rstring"
4019 keyword is prefixed with "!", then the response will be
4020 considered invalid if the body matches the expression.
4021 This can be used to look for a mandatory word at the end
4022 of a dynamic page, or to detect a failure when a specific
Davor Ocelice9ed2812017-12-25 17:49:28 +01004023 error appears on the check page (e.g. a stack trace).
Willy Tarreaubd741542010-03-16 18:46:54 +01004024
4025 It is important to note that the responses will be limited to a certain size
4026 defined by the global "tune.chksize" option, which defaults to 16384 bytes.
4027 Thus, too large responses may not contain the mandatory pattern when using
4028 "string" or "rstring". If a large response is absolutely required, it is
4029 possible to change the default max size by setting the global variable.
4030 However, it is worth keeping in mind that parsing very large responses can
4031 waste some CPU cycles, especially when regular expressions are used, and that
4032 it is always better to focus the checks on smaller resources.
4033
Cyril Bonté32602d22015-01-30 00:07:07 +01004034 Also "http-check expect" doesn't support HTTP keep-alive. Keep in mind that it
4035 will automatically append a "Connection: close" header, meaning that this
4036 header should not be present in the request provided by "option httpchk".
4037
Willy Tarreaubd741542010-03-16 18:46:54 +01004038 Last, if "http-check expect" is combined with "http-check disable-on-404",
4039 then this last one has precedence when the server responds with 404.
4040
4041 Examples :
4042 # only accept status 200 as valid
Willy Tarreau8f2a1e72011-01-06 16:36:10 +01004043 http-check expect status 200
Willy Tarreaubd741542010-03-16 18:46:54 +01004044
4045 # consider SQL errors as errors
Willy Tarreau8f2a1e72011-01-06 16:36:10 +01004046 http-check expect ! string SQL\ Error
Willy Tarreaubd741542010-03-16 18:46:54 +01004047
4048 # consider status 5xx only as errors
Willy Tarreau8f2a1e72011-01-06 16:36:10 +01004049 http-check expect ! rstatus ^5
Willy Tarreaubd741542010-03-16 18:46:54 +01004050
4051 # check that we have a correct hexadecimal tag before /html
Jarno Huuskonene5ae7022017-04-03 14:36:21 +03004052 http-check expect rstring <!--tag:[0-9a-f]*--></html>
Willy Tarreau0ba27502007-12-24 16:55:16 +01004053
Willy Tarreaubd741542010-03-16 18:46:54 +01004054 See also : "option httpchk", "http-check disable-on-404"
Willy Tarreau2769aa02007-12-27 18:26:09 +01004055
4056
Willy Tarreauef781042010-01-27 11:53:01 +01004057http-check send-state
4058 Enable emission of a state header with HTTP health checks
4059 May be used in sections : defaults | frontend | listen | backend
4060 yes | no | yes | yes
4061 Arguments : none
4062
4063 When this option is set, haproxy will systematically send a special header
4064 "X-Haproxy-Server-State" with a list of parameters indicating to each server
4065 how they are seen by haproxy. This can be used for instance when a server is
4066 manipulated without access to haproxy and the operator needs to know whether
4067 haproxy still sees it up or not, or if the server is the last one in a farm.
4068
4069 The header is composed of fields delimited by semi-colons, the first of which
4070 is a word ("UP", "DOWN", "NOLB"), possibly followed by a number of valid
4071 checks on the total number before transition, just as appears in the stats
4072 interface. Next headers are in the form "<variable>=<value>", indicating in
4073 no specific order some values available in the stats interface :
Joseph Lynch514061c2015-01-15 17:52:59 -08004074 - a variable "address", containing the address of the backend server.
4075 This corresponds to the <address> field in the server declaration. For
4076 unix domain sockets, it will read "unix".
4077
4078 - a variable "port", containing the port of the backend server. This
4079 corresponds to the <port> field in the server declaration. For unix
4080 domain sockets, it will read "unix".
4081
Willy Tarreauef781042010-01-27 11:53:01 +01004082 - a variable "name", containing the name of the backend followed by a slash
4083 ("/") then the name of the server. This can be used when a server is
4084 checked in multiple backends.
4085
4086 - a variable "node" containing the name of the haproxy node, as set in the
4087 global "node" variable, otherwise the system's hostname if unspecified.
4088
4089 - a variable "weight" indicating the weight of the server, a slash ("/")
4090 and the total weight of the farm (just counting usable servers). This
4091 helps to know if other servers are available to handle the load when this
4092 one fails.
4093
4094 - a variable "scur" indicating the current number of concurrent connections
4095 on the server, followed by a slash ("/") then the total number of
4096 connections on all servers of the same backend.
4097
4098 - a variable "qcur" indicating the current number of requests in the
4099 server's queue.
4100
4101 Example of a header received by the application server :
4102 >>> X-Haproxy-Server-State: UP 2/3; name=bck/srv2; node=lb1; weight=1/2; \
4103 scur=13/22; qcur=0
4104
4105 See also : "option httpchk", "http-check disable-on-404"
4106
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004107
4108http-request <action> [options...] [ { if | unless } <condition> ]
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01004109 Access control for Layer 7 requests
4110
4111 May be used in sections: defaults | frontend | listen | backend
4112 no | yes | yes | yes
4113
Willy Tarreau20b0de52012-12-24 15:45:22 +01004114 The http-request statement defines a set of rules which apply to layer 7
4115 processing. The rules are evaluated in their declaration order when they are
4116 met in a frontend, listen or backend section. Any rule may optionally be
4117 followed by an ACL-based condition, in which case it will only be evaluated
4118 if the condition is true.
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01004119
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004120 The first keyword is the rule's action. The supported actions are described
4121 below.
Willy Tarreau20b0de52012-12-24 15:45:22 +01004122
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004123 There is no limit to the number of http-request statements per instance.
Willy Tarreau20b0de52012-12-24 15:45:22 +01004124
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004125 It is important to know that http-request rules are processed very early in
4126 the HTTP processing, just after "block" rules and before "reqdel" or "reqrep"
4127 or "reqadd" rules. That way, headers added by "add-header"/"set-header" are
4128 visible by almost all further ACL rules.
Willy Tarreau53275e82017-11-24 07:52:01 +01004129
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004130 Using "reqadd"/"reqdel"/"reqrep" to manipulate request headers is discouraged
4131 in newer versions (>= 1.5). But if you need to use regular expression to
4132 delete headers, you can still use "reqdel". Also please use
4133 "http-request deny/allow/tarpit" instead of "reqdeny"/"reqpass"/"reqtarpit".
Willy Tarreauccbcc372012-12-27 12:37:57 +01004134
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004135 Example:
4136 acl nagios src 192.168.129.3
4137 acl local_net src 192.168.0.0/16
4138 acl auth_ok http_auth(L1)
Willy Tarreau20b0de52012-12-24 15:45:22 +01004139
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004140 http-request allow if nagios
4141 http-request allow if local_net auth_ok
4142 http-request auth realm Gimme if local_net auth_ok
4143 http-request deny
Willy Tarreau81499eb2012-12-27 12:19:02 +01004144
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004145 Example:
4146 acl key req.hdr(X-Add-Acl-Key) -m found
4147 acl add path /addacl
4148 acl del path /delacl
Willy Tarreau20b0de52012-12-24 15:45:22 +01004149
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004150 acl myhost hdr(Host) -f myhost.lst
Willy Tarreau20b0de52012-12-24 15:45:22 +01004151
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004152 http-request add-acl(myhost.lst) %[req.hdr(X-Add-Acl-Key)] if key add
4153 http-request del-acl(myhost.lst) %[req.hdr(X-Add-Acl-Key)] if key del
Thierry FOURNIERdad3d1d2014-04-22 18:07:25 +02004154
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004155 Example:
4156 acl value req.hdr(X-Value) -m found
4157 acl setmap path /setmap
4158 acl delmap path /delmap
Sasha Pachev218f0642014-06-16 12:05:59 -06004159
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004160 use_backend bk_appli if { hdr(Host),map_str(map.lst) -m found }
Sasha Pachev218f0642014-06-16 12:05:59 -06004161
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004162 http-request set-map(map.lst) %[src] %[req.hdr(X-Value)] if setmap value
4163 http-request del-map(map.lst) %[src] if delmap
Sasha Pachev218f0642014-06-16 12:05:59 -06004164
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004165 See also : "stats http-request", section 3.4 about userlists and section 7
4166 about ACL usage.
Sasha Pachev218f0642014-06-16 12:05:59 -06004167
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004168http-request add-acl(<file-name>) <key fmt> [ { if | unless } <condition> ]
Sasha Pachev218f0642014-06-16 12:05:59 -06004169
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004170 This is used to add a new entry into an ACL. The ACL must be loaded from a
4171 file (even a dummy empty file). The file name of the ACL to be updated is
4172 passed between parentheses. It takes one argument: <key fmt>, which follows
4173 log-format rules, to collect content of the new entry. It performs a lookup
4174 in the ACL before insertion, to avoid duplicated (or more) values. This
4175 lookup is done by a linear search and can be expensive with large lists!
4176 It is the equivalent of the "add acl" command from the stats socket, but can
4177 be triggered by an HTTP request.
Sasha Pachev218f0642014-06-16 12:05:59 -06004178
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004179http-request add-header <name> <fmt> [ { if | unless } <condition> ]
Sasha Pachev218f0642014-06-16 12:05:59 -06004180
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004181 This appends an HTTP header field whose name is specified in <name> and
4182 whose value is defined by <fmt> which follows the log-format rules (see
4183 Custom Log Format in section 8.2.4). This is particularly useful to pass
4184 connection-specific information to the server (e.g. the client's SSL
4185 certificate), or to combine several headers into one. This rule is not
4186 final, so it is possible to add other similar rules. Note that header
4187 addition is performed immediately, so one rule might reuse the resulting
4188 header from a previous rule.
Sasha Pachev218f0642014-06-16 12:05:59 -06004189
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004190http-request allow [ { if | unless } <condition> ]
Sasha Pachev218f0642014-06-16 12:05:59 -06004191
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004192 This stops the evaluation of the rules and lets the request pass the check.
4193 No further "http-request" rules are evaluated.
Sasha Pachev218f0642014-06-16 12:05:59 -06004194
Sasha Pachev218f0642014-06-16 12:05:59 -06004195
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004196http-request auth [realm <realm>] [ { if | unless } <condition> ]
Sasha Pachev218f0642014-06-16 12:05:59 -06004197
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004198 This stops the evaluation of the rules and immediately responds with an
4199 HTTP 401 or 407 error code to invite the user to present a valid user name
4200 and password. No further "http-request" rules are evaluated. An optional
4201 "realm" parameter is supported, it sets the authentication realm that is
4202 returned with the response (typically the application's name).
Sasha Pachev218f0642014-06-16 12:05:59 -06004203
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004204 Example:
4205 acl auth_ok http_auth_group(L1) G1
4206 http-request auth unless auth_ok
Sasha Pachev218f0642014-06-16 12:05:59 -06004207
Jarno Huuskonen251a6b72019-01-04 14:05:02 +02004208http-request cache-use <name> [ { if | unless } <condition> ]
Sasha Pachev218f0642014-06-16 12:05:59 -06004209
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004210 See section 10.2 about cache setup.
Willy Tarreaua0dc23f2015-01-22 20:46:11 +01004211
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004212http-request capture <sample> [ len <length> | id <id> ]
4213 [ { if | unless } <condition> ]
Willy Tarreaua0dc23f2015-01-22 20:46:11 +01004214
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004215 This captures sample expression <sample> from the request buffer, and
4216 converts it to a string of at most <len> characters. The resulting string is
4217 stored into the next request "capture" slot, so it will possibly appear next
4218 to some captured HTTP headers. It will then automatically appear in the logs,
4219 and it will be possible to extract it using sample fetch rules to feed it
4220 into headers or anything. The length should be limited given that this size
4221 will be allocated for each capture during the whole session life.
4222 Please check section 7.3 (Fetching samples) and "capture request header" for
4223 more information.
Willy Tarreaua0dc23f2015-01-22 20:46:11 +01004224
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004225 If the keyword "id" is used instead of "len", the action tries to store the
4226 captured string in a previously declared capture slot. This is useful to run
4227 captures in backends. The slot id can be declared by a previous directive
4228 "http-request capture" or with the "declare capture" keyword. If the slot
4229 <id> doesn't exist, then HAProxy fails parsing the configuration to prevent
4230 unexpected behavior at run time.
Willy Tarreaua0dc23f2015-01-22 20:46:11 +01004231
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004232http-request del-acl(<file-name>) <key fmt> [ { if | unless } <condition> ]
Willy Tarreaua0dc23f2015-01-22 20:46:11 +01004233
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004234 This is used to delete an entry from an ACL. The ACL must be loaded from a
4235 file (even a dummy empty file). The file name of the ACL to be updated is
4236 passed between parentheses. It takes one argument: <key fmt>, which follows
4237 log-format rules, to collect content of the entry to delete.
4238 It is the equivalent of the "del acl" command from the stats socket, but can
4239 be triggered by an HTTP request.
Willy Tarreaua0dc23f2015-01-22 20:46:11 +01004240
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004241http-request del-header <name> [ { if | unless } <condition> ]
Willy Tarreauf4c43c12013-06-11 17:01:13 +02004242
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004243 This removes all HTTP header fields whose name is specified in <name>.
Willy Tarreau9a355ec2013-06-11 17:45:46 +02004244
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004245http-request del-map(<file-name>) <key fmt> [ { if | unless } <condition> ]
Willy Tarreau42cf39e2013-06-11 18:51:32 +02004246
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004247 This is used to delete an entry from a MAP. The MAP must be loaded from a
4248 file (even a dummy empty file). The file name of the MAP to be updated is
4249 passed between parentheses. It takes one argument: <key fmt>, which follows
4250 log-format rules, to collect content of the entry to delete.
4251 It takes one argument: "file name" It is the equivalent of the "del map"
4252 command from the stats socket, but can be triggered by an HTTP request.
Willy Tarreau51347ed2013-06-11 19:34:13 +02004253
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004254http-request deny [deny_status <status>] [ { if | unless } <condition> ]
Patrick Hemmer268a7072018-05-11 12:52:31 -04004255
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004256 This stops the evaluation of the rules and immediately rejects the request
4257 and emits an HTTP 403 error, or optionally the status code specified as an
4258 argument to "deny_status". The list of permitted status codes is limited to
4259 those that can be overridden by the "errorfile" directive.
4260 No further "http-request" rules are evaluated.
Patrick Hemmer268a7072018-05-11 12:52:31 -04004261
Olivier Houchard602bf7d2019-05-10 13:59:15 +02004262http-request disable-l7-retry [ { if | unless } <condition> ]
4263 This disables any attempt to retry the request if it fails for any other
4264 reason than a connection failure. This can be useful for example to make
4265 sure POST requests aren't retried on failure.
4266
Baptiste Assmann333939c2019-01-21 08:34:50 +01004267http-request do-resolve(<var>,<resolvers>,[ipv4,ipv6]) <expr> :
4268
4269 This action performs a DNS resolution of the output of <expr> and stores
4270 the result in the variable <var>. It uses the DNS resolvers section
4271 pointed by <resolvers>.
4272 It is possible to choose a resolution preference using the optional
4273 arguments 'ipv4' or 'ipv6'.
4274 When performing the DNS resolution, the client side connection is on
4275 pause waiting till the end of the resolution.
4276 If an IP address can be found, it is stored into <var>. If any kind of
4277 error occurs, then <var> is not set.
4278 One can use this action to discover a server IP address at run time and
4279 based on information found in the request (IE a Host header).
4280 If this action is used to find the server's IP address (using the
4281 "set-dst" action), then the server IP address in the backend must be set
4282 to 0.0.0.0.
4283
4284 Example:
4285 resolvers mydns
4286 nameserver local 127.0.0.53:53
4287 nameserver google 8.8.8.8:53
4288 timeout retry 1s
4289 hold valid 10s
4290 hold nx 3s
4291 hold other 3s
4292 hold obsolete 0s
4293 accepted_payload_size 8192
4294
4295 frontend fe
4296 bind 10.42.0.1:80
4297 http-request do-resolve(txn.myip,mydns,ipv4) hdr(Host),lower
4298 http-request capture var(txn.myip) len 40
4299
4300 # return 503 when the variable is not set,
4301 # which mean DNS resolution error
4302 use_backend b_503 unless { var(txn.myip) -m found }
4303
4304 default_backend be
4305
4306 backend b_503
4307 # dummy backend used to return 503.
4308 # one can use the errorfile directive to send a nice
4309 # 503 error page to end users
4310
4311 backend be
4312 # rule to prevent HAProxy from reconnecting to services
4313 # on the local network (forged DNS name used to scan the network)
4314 http-request deny if { var(txn.myip) -m ip 127.0.0.0/8 10.0.0.0/8 }
4315 http-request set-dst var(txn.myip)
4316 server clear 0.0.0.0:0
4317
4318 NOTE: Don't forget to set the "protection" rules to ensure HAProxy won't
4319 be used to scan the network or worst won't loop over itself...
4320
Frédéric Lécaille06f5b642018-11-12 11:01:10 +01004321http-request early-hint <name> <fmt> [ { if | unless } <condition> ]
4322
4323 This is used to build an HTTP 103 Early Hints response prior to any other one.
4324 This appends an HTTP header field to this response whose name is specified in
4325 <name> and whose value is defined by <fmt> which follows the log-format rules
4326 (see Custom Log Format in section 8.2.4). This is particularly useful to pass
Frédéric Lécaille3aac1062018-11-13 09:42:13 +01004327 to the client some Link headers to preload resources required to render the
4328 HTML documents.
Frédéric Lécaille06f5b642018-11-12 11:01:10 +01004329
4330 See RFC 8297 for more information.
4331
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004332http-request redirect <rule> [ { if | unless } <condition> ]
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02004333
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004334 This performs an HTTP redirection based on a redirect rule. This is exactly
4335 the same as the "redirect" statement except that it inserts a redirect rule
4336 which can be processed in the middle of other "http-request" rules and that
4337 these rules use the "log-format" strings. See the "redirect" keyword for the
4338 rule's syntax.
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02004339
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004340http-request reject [ { if | unless } <condition> ]
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02004341
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004342 This stops the evaluation of the rules and immediately closes the connection
4343 without sending any response. It acts similarly to the
4344 "tcp-request content reject" rules. It can be useful to force an immediate
4345 connection closure on HTTP/2 connections.
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02004346
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004347http-request replace-header <name> <match-regex> <replace-fmt>
4348 [ { if | unless } <condition> ]
Willy Tarreaua9083d02015-05-08 15:27:59 +02004349
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004350 This matches the regular expression in all occurrences of header field
4351 <name> according to <match-regex>, and replaces them with the
4352 <replace-fmt> argument. Format characters are allowed in replace-fmt and
4353 work like in <fmt> arguments in "http-request add-header". The match is
4354 only case-sensitive. It is important to understand that this action only
4355 considers whole header lines, regardless of the number of values they may
4356 contain. This usage is suited to headers naturally containing commas in
4357 their value, such as If-Modified-Since and so on.
Thierry FOURNIER82bf70d2015-05-26 17:58:29 +02004358
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004359 Example:
4360 http-request replace-header Cookie foo=([^;]*);(.*) foo=\1;ip=%bi;\2
William Lallemand86d0df02017-11-24 21:36:45 +01004361
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004362 # applied to:
4363 Cookie: foo=foobar; expires=Tue, 14-Jun-2016 01:40:45 GMT;
Willy Tarreau09448f72014-06-25 18:12:15 +02004364
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004365 # outputs:
4366 Cookie: foo=foobar;ip=192.168.1.20; expires=Tue, 14-Jun-2016 01:40:45 GMT;
Willy Tarreau09448f72014-06-25 18:12:15 +02004367
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004368 # assuming the backend IP is 192.168.1.20
Willy Tarreau09448f72014-06-25 18:12:15 +02004369
Willy Tarreau33810222019-06-12 17:44:02 +02004370http-request replace-uri <match-regex> <replace-fmt>
4371 [ { if | unless } <condition> ]
4372
4373 This matches the regular expression in the URI part of the request
4374 according to <match-regex>, and replaces it with the <replace-fmt>
4375 argument. Standard back-references using the backslash ('\') followed by a
4376 number are supported. The <fmt> field is interpreted as a log-format string
4377 so it may contain special expressions just like the <fmt> argument passed
4378 to "http-request set-uri". The match is exclusively case-sensitive. Any
4379 optional scheme, authority or query string are considered in the matching
4380 part of the URI. It is worth noting that regular expressions may be more
4381 expensive to evaluate than certain ACLs, so rare replacements may benefit
4382 from a condition to avoid performing the evaluation at all if it does not
4383 match.
4384
4385 Example:
4386 # prefix /foo : turn /bar?q=1 into /foo/bar?q=1 :
4387 http-request replace-uri (.*) /foo\1
4388
4389 # suffix /foo : turn /bar?q=1 into /bar/foo?q=1 :
4390 http-request replace-uri ([^?]*)(\?(.*))? \1/foo\2
4391
4392 # strip /foo : turn /foo/bar?q=1 into /bar?q=1
4393 http-request replace-uri /foo/(.*) /\1
4394 # or more efficient if only some requests match :
4395 http-request replace-uri /foo/(.*) /\1 if { url_beg /foo/ }
4396
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004397http-request replace-value <name> <match-regex> <replace-fmt>
4398 [ { if | unless } <condition> ]
Willy Tarreau09448f72014-06-25 18:12:15 +02004399
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004400 This works like "replace-header" except that it matches the regex against
4401 every comma-delimited value of the header field <name> instead of the
4402 entire header. This is suited for all headers which are allowed to carry
4403 more than one value. An example could be the Accept header.
Willy Tarreau09448f72014-06-25 18:12:15 +02004404
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004405 Example:
4406 http-request replace-value X-Forwarded-For ^192\.168\.(.*)$ 172.16.\1
Thierry FOURNIER236657b2015-08-19 08:25:14 +02004407
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004408 # applied to:
4409 X-Forwarded-For: 192.168.10.1, 192.168.13.24, 10.0.0.37
Thierry FOURNIERe0627bd2015-08-04 08:20:33 +02004410
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004411 # outputs:
4412 X-Forwarded-For: 172.16.10.1, 172.16.13.24, 10.0.0.37
Frédéric Lécaille6778b272018-01-29 15:22:53 +01004413
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004414http-request sc-inc-gpc0(<sc-id>) [ { if | unless } <condition> ]
4415http-request sc-inc-gpc1(<sc-id>) [ { if | unless } <condition> ]
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02004416
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004417 This actions increments the GPC0 or GPC1 counter according with the sticky
4418 counter designated by <sc-id>. If an error occurs, this action silently fails
4419 and the actions evaluation continues.
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02004420
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004421http-request sc-set-gpt0(<sc-id>) <int> [ { if | unless } <condition> ]
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02004422
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004423 This action sets the GPT0 tag according to the sticky counter designated by
4424 <sc-id> and the value of <int>. The expected result is a boolean. If an error
4425 occurs, this action silently fails and the actions evaluation continues.
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02004426
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004427http-request set-dst <expr> [ { if | unless } <condition> ]
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02004428
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004429 This is used to set the destination IP address to the value of specified
4430 expression. Useful when a proxy in front of HAProxy rewrites destination IP,
4431 but provides the correct IP in a HTTP header; or you want to mask the IP for
4432 privacy. If you want to connect to the new address/port, use '0.0.0.0:0' as a
4433 server address in the backend.
Christopher Faulet85d79c92016-11-09 16:54:56 +01004434
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004435 Arguments:
4436 <expr> Is a standard HAProxy expression formed by a sample-fetch followed
4437 by some converters.
Christopher Faulet85d79c92016-11-09 16:54:56 +01004438
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004439 Example:
4440 http-request set-dst hdr(x-dst)
4441 http-request set-dst dst,ipmask(24)
Christopher Faulet85d79c92016-11-09 16:54:56 +01004442
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004443 When possible, set-dst preserves the original destination port as long as the
4444 address family allows it, otherwise the destination port is set to 0.
Adis Nezirovic2fbcafc2015-07-06 15:44:30 +02004445
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004446http-request set-dst-port <expr> [ { if | unless } <condition> ]
Adis Nezirovic2fbcafc2015-07-06 15:44:30 +02004447
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004448 This is used to set the destination port address to the value of specified
4449 expression. If you want to connect to the new address/port, use '0.0.0.0:0'
4450 as a server address in the backend.
Adis Nezirovic2fbcafc2015-07-06 15:44:30 +02004451
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004452 Arguments:
4453 <expr> Is a standard HAProxy expression formed by a sample-fetch
4454 followed by some converters.
Adis Nezirovic2fbcafc2015-07-06 15:44:30 +02004455
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004456 Example:
4457 http-request set-dst-port hdr(x-port)
4458 http-request set-dst-port int(4000)
Adis Nezirovic2fbcafc2015-07-06 15:44:30 +02004459
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004460 When possible, set-dst-port preserves the original destination address as
4461 long as the address family supports a port, otherwise it forces the
4462 destination address to IPv4 "0.0.0.0" before rewriting the port.
William Lallemand44be6402016-05-25 01:51:35 +02004463
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004464http-request set-header <name> <fmt> [ { if | unless } <condition> ]
William Lallemand44be6402016-05-25 01:51:35 +02004465
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004466 This does the same as "http-request add-header" except that the header name
4467 is first removed if it existed. This is useful when passing security
4468 information to the server, where the header must not be manipulated by
4469 external users. Note that the new value is computed before the removal so it
4470 is possible to concatenate a value to an existing header.
William Lallemand44be6402016-05-25 01:51:35 +02004471
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004472 Example:
4473 http-request set-header X-Haproxy-Current-Date %T
4474 http-request set-header X-SSL %[ssl_fc]
4475 http-request set-header X-SSL-Session_ID %[ssl_fc_session_id,hex]
4476 http-request set-header X-SSL-Client-Verify %[ssl_c_verify]
4477 http-request set-header X-SSL-Client-DN %{+Q}[ssl_c_s_dn]
4478 http-request set-header X-SSL-Client-CN %{+Q}[ssl_c_s_dn(cn)]
4479 http-request set-header X-SSL-Issuer %{+Q}[ssl_c_i_dn]
4480 http-request set-header X-SSL-Client-NotBefore %{+Q}[ssl_c_notbefore]
4481 http-request set-header X-SSL-Client-NotAfter %{+Q}[ssl_c_notafter]
William Lallemand44be6402016-05-25 01:51:35 +02004482
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004483http-request set-log-level <level> [ { if | unless } <condition> ]
William Lallemand44be6402016-05-25 01:51:35 +02004484
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004485 This is used to change the log level of the current request when a certain
4486 condition is met. Valid levels are the 8 syslog levels (see the "log"
4487 keyword) plus the special level "silent" which disables logging for this
4488 request. This rule is not final so the last matching rule wins. This rule
4489 can be useful to disable health checks coming from another equipment.
William Lallemand13e9b0c2016-05-25 02:34:07 +02004490
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004491http-request set-map(<file-name>) <key fmt> <value fmt>
4492 [ { if | unless } <condition> ]
William Lallemand13e9b0c2016-05-25 02:34:07 +02004493
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004494 This is used to add a new entry into a MAP. The MAP must be loaded from a
4495 file (even a dummy empty file). The file name of the MAP to be updated is
4496 passed between parentheses. It takes 2 arguments: <key fmt>, which follows
4497 log-format rules, used to collect MAP key, and <value fmt>, which follows
4498 log-format rules, used to collect content for the new entry.
4499 It performs a lookup in the MAP before insertion, to avoid duplicated (or
4500 more) values. This lookup is done by a linear search and can be expensive
4501 with large lists! It is the equivalent of the "set map" command from the
4502 stats socket, but can be triggered by an HTTP request.
William Lallemand13e9b0c2016-05-25 02:34:07 +02004503
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004504http-request set-mark <mark> [ { if | unless } <condition> ]
William Lallemand13e9b0c2016-05-25 02:34:07 +02004505
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004506 This is used to set the Netfilter MARK on all packets sent to the client to
4507 the value passed in <mark> on platforms which support it. This value is an
4508 unsigned 32 bit value which can be matched by netfilter and by the routing
4509 table. It can be expressed both in decimal or hexadecimal format (prefixed by
4510 "0x"). This can be useful to force certain packets to take a different route
4511 (for example a cheaper network path for bulk downloads). This works on Linux
4512 kernels 2.6.32 and above and requires admin privileges.
Willy Tarreau00005ce2016-10-21 15:07:45 +02004513
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004514http-request set-method <fmt> [ { if | unless } <condition> ]
William Lallemand13e9b0c2016-05-25 02:34:07 +02004515
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004516 This rewrites the request method with the result of the evaluation of format
4517 string <fmt>. There should be very few valid reasons for having to do so as
4518 this is more likely to break something than to fix it.
William Lallemand13e9b0c2016-05-25 02:34:07 +02004519
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004520http-request set-nice <nice> [ { if | unless } <condition> ]
William Lallemand13e9b0c2016-05-25 02:34:07 +02004521
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004522 This sets the "nice" factor of the current request being processed. It only
4523 has effect against the other requests being processed at the same time.
4524 The default value is 0, unless altered by the "nice" setting on the "bind"
4525 line. The accepted range is -1024..1024. The higher the value, the nicest
4526 the request will be. Lower values will make the request more important than
4527 other ones. This can be useful to improve the speed of some requests, or
4528 lower the priority of non-important requests. Using this setting without
4529 prior experimentation can cause some major slowdown.
William Lallemand13e9b0c2016-05-25 02:34:07 +02004530
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004531http-request set-path <fmt> [ { if | unless } <condition> ]
Willy Tarreau00005ce2016-10-21 15:07:45 +02004532
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004533 This rewrites the request path with the result of the evaluation of format
4534 string <fmt>. The query string, if any, is left intact. If a scheme and
4535 authority is found before the path, they are left intact as well. If the
4536 request doesn't have a path ("*"), this one is replaced with the format.
4537 This can be used to prepend a directory component in front of a path for
4538 example. See also "http-request set-query" and "http-request set-uri".
Willy Tarreau2d392c22015-08-24 01:43:45 +02004539
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004540 Example :
4541 # prepend the host name before the path
4542 http-request set-path /%[hdr(host)]%[path]
Christopher Faulet76c09ef2017-09-21 11:03:52 +02004543
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004544http-request set-priority-class <expr> [ { if | unless } <condition> ]
Olivier Houchardccaa7de2017-10-02 11:51:03 +02004545
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004546 This is used to set the queue priority class of the current request.
4547 The value must be a sample expression which converts to an integer in the
4548 range -2047..2047. Results outside this range will be truncated.
4549 The priority class determines the order in which queued requests are
4550 processed. Lower values have higher priority.
Christopher Faulet76c09ef2017-09-21 11:03:52 +02004551
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004552http-request set-priority-offset <expr> [ { if | unless } <condition> ]
Christopher Faulet76c09ef2017-09-21 11:03:52 +02004553
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004554 This is used to set the queue priority timestamp offset of the current
4555 request. The value must be a sample expression which converts to an integer
4556 in the range -524287..524287. Results outside this range will be truncated.
4557 When a request is queued, it is ordered first by the priority class, then by
4558 the current timestamp adjusted by the given offset in milliseconds. Lower
4559 values have higher priority.
4560 Note that the resulting timestamp is is only tracked with enough precision
4561 for 524,287ms (8m44s287ms). If the request is queued long enough to where the
4562 adjusted timestamp exceeds this value, it will be misidentified as highest
4563 priority. Thus it is important to set "timeout queue" to a value, where when
4564 combined with the offset, does not exceed this limit.
Christopher Faulet76c09ef2017-09-21 11:03:52 +02004565
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004566http-request set-query <fmt> [ { if | unless } <condition> ]
Willy Tarreau20b0de52012-12-24 15:45:22 +01004567
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004568 This rewrites the request's query string which appears after the first
4569 question mark ("?") with the result of the evaluation of format string <fmt>.
4570 The part prior to the question mark is left intact. If the request doesn't
4571 contain a question mark and the new value is not empty, then one is added at
4572 the end of the URI, followed by the new value. If a question mark was
4573 present, it will never be removed even if the value is empty. This can be
4574 used to add or remove parameters from the query string.
Ruoshan Huangeb5a3632015-12-08 21:00:23 +08004575
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004576 See also "http-request set-query" and "http-request set-uri".
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01004577
4578 Example:
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004579 # replace "%3D" with "=" in the query string
4580 http-request set-query %[query,regsub(%3D,=,g)]
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01004581
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004582http-request set-src <expr> [ { if | unless } <condition> ]
4583 This is used to set the source IP address to the value of specified
4584 expression. Useful when a proxy in front of HAProxy rewrites source IP, but
4585 provides the correct IP in a HTTP header; or you want to mask source IP for
4586 privacy.
4587
4588 Arguments :
4589 <expr> Is a standard HAProxy expression formed by a sample-fetch followed
4590 by some converters.
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01004591
Cyril Bonté78caf842010-03-10 22:41:43 +01004592 Example:
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004593 http-request set-src hdr(x-forwarded-for)
4594 http-request set-src src,ipmask(24)
4595
4596 When possible, set-src preserves the original source port as long as the
4597 address family allows it, otherwise the source port is set to 0.
4598
4599http-request set-src-port <expr> [ { if | unless } <condition> ]
4600
4601 This is used to set the source port address to the value of specified
4602 expression.
4603
4604 Arguments:
4605 <expr> Is a standard HAProxy expression formed by a sample-fetch followed
4606 by some converters.
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +01004607
Willy Tarreau20b0de52012-12-24 15:45:22 +01004608 Example:
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004609 http-request set-src-port hdr(x-port)
4610 http-request set-src-port int(4000)
4611
4612 When possible, set-src-port preserves the original source address as long as
4613 the address family supports a port, otherwise it forces the source address to
4614 IPv4 "0.0.0.0" before rewriting the port.
4615
4616http-request set-tos <tos> [ { if | unless } <condition> ]
4617
4618 This is used to set the TOS or DSCP field value of packets sent to the client
4619 to the value passed in <tos> on platforms which support this. This value
4620 represents the whole 8 bits of the IP TOS field, and can be expressed both in
4621 decimal or hexadecimal format (prefixed by "0x"). Note that only the 6 higher
4622 bits are used in DSCP or TOS, and the two lower bits are always 0. This can
4623 be used to adjust some routing behavior on border routers based on some
4624 information from the request.
4625
4626 See RFC 2474, 2597, 3260 and 4594 for more information.
4627
4628http-request set-uri <fmt> [ { if | unless } <condition> ]
4629
4630 This rewrites the request URI with the result of the evaluation of format
4631 string <fmt>. The scheme, authority, path and query string are all replaced
4632 at once. This can be used to rewrite hosts in front of proxies, or to
4633 perform complex modifications to the URI such as moving parts between the
4634 path and the query string.
4635 See also "http-request set-path" and "http-request set-query".
4636
4637http-request set-var(<var-name>) <expr> [ { if | unless } <condition> ]
4638
4639 This is used to set the contents of a variable. The variable is declared
4640 inline.
4641
4642 Arguments:
4643 <var-name> The name of the variable starts with an indication about its
4644 scope. The scopes allowed are:
4645 "proc" : the variable is shared with the whole process
4646 "sess" : the variable is shared with the whole session
4647 "txn" : the variable is shared with the transaction
4648 (request and response)
4649 "req" : the variable is shared only during request
4650 processing
4651 "res" : the variable is shared only during response
4652 processing
4653 This prefix is followed by a name. The separator is a '.'.
4654 The name may only contain characters 'a-z', 'A-Z', '0-9'
4655 and '_'.
4656
4657 <expr> Is a standard HAProxy expression formed by a sample-fetch
4658 followed by some converters.
Willy Tarreau20b0de52012-12-24 15:45:22 +01004659
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02004660 Example:
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004661 http-request set-var(req.my_var) req.fhdr(user-agent),lower
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02004662
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004663http-request send-spoe-group <engine-name> <group-name>
4664 [ { if | unless } <condition> ]
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02004665
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004666 This action is used to trigger sending of a group of SPOE messages. To do so,
4667 the SPOE engine used to send messages must be defined, as well as the SPOE
4668 group to send. Of course, the SPOE engine must refer to an existing SPOE
4669 filter. If not engine name is provided on the SPOE filter line, the SPOE
4670 agent name must be used.
4671
4672 Arguments:
4673 <engine-name> The SPOE engine name.
4674
4675 <group-name> The SPOE group name as specified in the engine
4676 configuration.
4677
4678http-request silent-drop [ { if | unless } <condition> ]
4679
4680 This stops the evaluation of the rules and makes the client-facing connection
4681 suddenly disappear using a system-dependent way that tries to prevent the
4682 client from being notified. The effect it then that the client still sees an
4683 established connection while there's none on HAProxy. The purpose is to
4684 achieve a comparable effect to "tarpit" except that it doesn't use any local
4685 resource at all on the machine running HAProxy. It can resist much higher
4686 loads than "tarpit", and slow down stronger attackers. It is important to
4687 understand the impact of using this mechanism. All stateful equipment placed
4688 between the client and HAProxy (firewalls, proxies, load balancers) will also
4689 keep the established connection for a long time and may suffer from this
4690 action.
4691 On modern Linux systems running with enough privileges, the TCP_REPAIR socket
4692 option is used to block the emission of a TCP reset. On other systems, the
4693 socket's TTL is reduced to 1 so that the TCP reset doesn't pass the first
4694 router, though it's still delivered to local networks. Do not use it unless
4695 you fully understand how it works.
4696
4697http-request tarpit [deny_status <status>] [ { if | unless } <condition> ]
4698
4699 This stops the evaluation of the rules and immediately blocks the request
4700 without responding for a delay specified by "timeout tarpit" or
4701 "timeout connect" if the former is not set. After that delay, if the client
4702 is still connected, an HTTP error 500 (or optionally the status code
4703 specified as an argument to "deny_status") is returned so that the client
4704 does not suspect it has been tarpitted. Logs will report the flags "PT".
4705 The goal of the tarpit rule is to slow down robots during an attack when
4706 they're limited on the number of concurrent requests. It can be very
4707 efficient against very dumb robots, and will significantly reduce the load
4708 on firewalls compared to a "deny" rule. But when facing "correctly"
4709 developed robots, it can make things worse by forcing haproxy and the front
4710 firewall to support insane number of concurrent connections.
4711 See also the "silent-drop" action.
4712
4713http-request track-sc0 <key> [table <table>] [ { if | unless } <condition> ]
4714http-request track-sc1 <key> [table <table>] [ { if | unless } <condition> ]
4715http-request track-sc2 <key> [table <table>] [ { if | unless } <condition> ]
4716
4717 This enables tracking of sticky counters from current request. These rules do
4718 not stop evaluation and do not change default action. The number of counters
4719 that may be simultaneously tracked by the same connection is set in
4720 MAX_SESS_STKCTR at build time (reported in haproxy -vv) which defaults to 3,
4721 so the track-sc number is between 0 and (MAX_SESS_STCKTR-1). The first
4722 "track-sc0" rule executed enables tracking of the counters of the specified
4723 table as the first set. The first "track-sc1" rule executed enables tracking
4724 of the counters of the specified table as the second set. The first
4725 "track-sc2" rule executed enables tracking of the counters of the specified
4726 table as the third set. It is a recommended practice to use the first set of
4727 counters for the per-frontend counters and the second set for the per-backend
4728 ones. But this is just a guideline, all may be used everywhere.
4729
4730 Arguments :
4731 <key> is mandatory, and is a sample expression rule as described in
4732 section 7.3. It describes what elements of the incoming request or
4733 connection will be analyzed, extracted, combined, and used to
4734 select which table entry to update the counters.
4735
4736 <table> is an optional table to be used instead of the default one, which
4737 is the stick-table declared in the current proxy. All the counters
4738 for the matches and updates for the key will then be performed in
4739 that table until the session ends.
4740
4741 Once a "track-sc*" rule is executed, the key is looked up in the table and if
4742 it is not found, an entry is allocated for it. Then a pointer to that entry
4743 is kept during all the session's life, and this entry's counters are updated
4744 as often as possible, every time the session's counters are updated, and also
4745 systematically when the session ends. Counters are only updated for events
4746 that happen after the tracking has been started. As an exception, connection
4747 counters and request counters are systematically updated so that they reflect
4748 useful information.
4749
4750 If the entry tracks concurrent connection counters, one connection is counted
4751 for as long as the entry is tracked, and the entry will not expire during
4752 that time. Tracking counters also provides a performance advantage over just
4753 checking the keys, because only one table lookup is performed for all ACL
4754 checks that make use of it.
4755
4756http-request unset-var(<var-name>) [ { if | unless } <condition> ]
4757
4758 This is used to unset a variable. See above for details about <var-name>.
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02004759
4760 Example:
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004761 http-request unset-var(req.my_var)
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02004762
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004763http-request wait-for-handshake [ { if | unless } <condition> ]
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02004764
Cyril Bontéc6ad23b2018-10-17 00:14:50 +02004765 This will delay the processing of the request until the SSL handshake
4766 happened. This is mostly useful to delay processing early data until we're
4767 sure they are valid.
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02004768
Willy Tarreauef781042010-01-27 11:53:01 +01004769
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02004770http-response <action> <options...> [ { if | unless } <condition> ]
Willy Tarreaue365c0b2013-06-11 16:06:12 +02004771 Access control for Layer 7 responses
4772
4773 May be used in sections: defaults | frontend | listen | backend
4774 no | yes | yes | yes
4775
4776 The http-response statement defines a set of rules which apply to layer 7
4777 processing. The rules are evaluated in their declaration order when they are
4778 met in a frontend, listen or backend section. Any rule may optionally be
4779 followed by an ACL-based condition, in which case it will only be evaluated
4780 if the condition is true. Since these rules apply on responses, the backend
4781 rules are applied first, followed by the frontend's rules.
4782
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02004783 The first keyword is the rule's action. The supported actions are described
4784 below.
Willy Tarreaue365c0b2013-06-11 16:06:12 +02004785
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02004786 There is no limit to the number of http-response statements per instance.
Willy Tarreaue365c0b2013-06-11 16:06:12 +02004787
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02004788 It is important to know that http-response rules are processed very early in
4789 the HTTP processing, before "rspdel" or "rsprep" or "rspadd" rules. That way,
4790 headers added by "add-header"/"set-header" are visible by almost all further
4791 ACL rules.
Willy Tarreaue365c0b2013-06-11 16:06:12 +02004792
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02004793 Using "rspadd"/"rspdel"/"rsprep" to manipulate request headers is discouraged
4794 in newer versions (>= 1.5). But if you need to use regular expression to
4795 delete headers, you can still use "rspdel". Also please use
4796 "http-response deny" instead of "rspdeny".
Willy Tarreaue365c0b2013-06-11 16:06:12 +02004797
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02004798 Example:
4799 acl key_acl res.hdr(X-Acl-Key) -m found
Thierry FOURNIERdad3d1d2014-04-22 18:07:25 +02004800
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02004801 acl myhost hdr(Host) -f myhost.lst
Sasha Pachev218f0642014-06-16 12:05:59 -06004802
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02004803 http-response add-acl(myhost.lst) %[res.hdr(X-Acl-Key)] if key_acl
4804 http-response del-acl(myhost.lst) %[res.hdr(X-Acl-Key)] if key_acl
Sasha Pachev218f0642014-06-16 12:05:59 -06004805
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02004806 Example:
4807 acl value res.hdr(X-Value) -m found
Sasha Pachev218f0642014-06-16 12:05:59 -06004808
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02004809 use_backend bk_appli if { hdr(Host),map_str(map.lst) -m found }
Sasha Pachev218f0642014-06-16 12:05:59 -06004810
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02004811 http-response set-map(map.lst) %[src] %[res.hdr(X-Value)] if value
4812 http-response del-map(map.lst) %[src] if ! value
Sasha Pachev218f0642014-06-16 12:05:59 -06004813
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02004814 See also : "http-request", section 3.4 about userlists and section 7 about
4815 ACL usage.
Sasha Pachev218f0642014-06-16 12:05:59 -06004816
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02004817http-response add-acl(<file-name>) <key fmt> [ { if | unless } <condition> ]
Sasha Pachev218f0642014-06-16 12:05:59 -06004818
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02004819 This is used to add a new entry into an ACL. The ACL must be loaded from a
4820 file (even a dummy empty file). The file name of the ACL to be updated is
4821 passed between parentheses. It takes one argument: <key fmt>, which follows
4822 log-format rules, to collect content of the new entry. It performs a lookup
4823 in the ACL before insertion, to avoid duplicated (or more) values.
4824 This lookup is done by a linear search and can be expensive with large lists!
4825 It is the equivalent of the "add acl" command from the stats socket, but can
4826 be triggered by an HTTP response.
Sasha Pachev218f0642014-06-16 12:05:59 -06004827
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02004828http-response add-header <name> <fmt> [ { if | unless } <condition> ]
Sasha Pachev218f0642014-06-16 12:05:59 -06004829
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02004830 This appends an HTTP header field whose name is specified in <name> and whose
4831 value is defined by <fmt> which follows the log-format rules (see Custom Log
4832 Format in section 8.2.4). This may be used to send a cookie to a client for
4833 example, or to pass some internal information.
4834 This rule is not final, so it is possible to add other similar rules.
4835 Note that header addition is performed immediately, so one rule might reuse
4836 the resulting header from a previous rule.
Sasha Pachev218f0642014-06-16 12:05:59 -06004837
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02004838http-response allow [ { if | unless } <condition> ]
Sasha Pachev218f0642014-06-16 12:05:59 -06004839
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02004840 This stops the evaluation of the rules and lets the response pass the check.
4841 No further "http-response" rules are evaluated for the current section.
Sasha Pachev218f0642014-06-16 12:05:59 -06004842
Jarno Huuskonen251a6b72019-01-04 14:05:02 +02004843http-response cache-store <name> [ { if | unless } <condition> ]
Sasha Pachev218f0642014-06-16 12:05:59 -06004844
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02004845 See section 10.2 about cache setup.
Sasha Pachev218f0642014-06-16 12:05:59 -06004846
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02004847http-response capture <sample> id <id> [ { if | unless } <condition> ]
Sasha Pachev218f0642014-06-16 12:05:59 -06004848
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02004849 This captures sample expression <sample> from the response buffer, and
4850 converts it to a string. The resulting string is stored into the next request
4851 "capture" slot, so it will possibly appear next to some captured HTTP
4852 headers. It will then automatically appear in the logs, and it will be
4853 possible to extract it using sample fetch rules to feed it into headers or
4854 anything. Please check section 7.3 (Fetching samples) and
4855 "capture response header" for more information.
Thierry FOURNIER35d70ef2015-08-26 16:21:56 +02004856
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02004857 The keyword "id" is the id of the capture slot which is used for storing the
4858 string. The capture slot must be defined in an associated frontend.
4859 This is useful to run captures in backends. The slot id can be declared by a
4860 previous directive "http-response capture" or with the "declare capture"
4861 keyword.
4862 If the slot <id> doesn't exist, then HAProxy fails parsing the configuration
4863 to prevent unexpected behavior at run time.
Thierry FOURNIER35d70ef2015-08-26 16:21:56 +02004864
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02004865http-response del-acl(<file-name>) <key fmt> [ { if | unless } <condition> ]
Thierry FOURNIER35d70ef2015-08-26 16:21:56 +02004866
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02004867 This is used to delete an entry from an ACL. The ACL must be loaded from a
4868 file (even a dummy empty file). The file name of the ACL to be updated is
4869 passed between parentheses. It takes one argument: <key fmt>, which follows
4870 log-format rules, to collect content of the entry to delete.
4871 It is the equivalent of the "del acl" command from the stats socket, but can
4872 be triggered by an HTTP response.
Willy Tarreauf4c43c12013-06-11 17:01:13 +02004873
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02004874http-response del-header <name> [ { if | unless } <condition> ]
Willy Tarreau9a355ec2013-06-11 17:45:46 +02004875
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02004876 This removes all HTTP header fields whose name is specified in <name>.
Willy Tarreau42cf39e2013-06-11 18:51:32 +02004877
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02004878http-response del-map(<file-name>) <key fmt> [ { if | unless } <condition> ]
Willy Tarreau51347ed2013-06-11 19:34:13 +02004879
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02004880 This is used to delete an entry from a MAP. The MAP must be loaded from a
4881 file (even a dummy empty file). The file name of the MAP to be updated is
4882 passed between parentheses. It takes one argument: <key fmt>, which follows
4883 log-format rules, to collect content of the entry to delete.
4884 It takes one argument: "file name" It is the equivalent of the "del map"
4885 command from the stats socket, but can be triggered by an HTTP response.
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02004886
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02004887http-response deny [ { if | unless } <condition> ]
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02004888
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02004889 This stops the evaluation of the rules and immediately rejects the response
4890 and emits an HTTP 502 error. No further "http-response" rules are evaluated.
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02004891
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02004892http-response redirect <rule> [ { if | unless } <condition> ]
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02004893
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02004894 This performs an HTTP redirection based on a redirect rule.
4895 This supports a format string similarly to "http-request redirect" rules,
4896 with the exception that only the "location" type of redirect is possible on
4897 the response. See the "redirect" keyword for the rule's syntax. When a
4898 redirect rule is applied during a response, connections to the server are
4899 closed so that no data can be forwarded from the server to the client.
Thierry FOURNIERe80fada2015-05-26 18:06:31 +02004900
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02004901http-response replace-header <name> <regex-match> <replace-fmt>
4902 [ { if | unless } <condition> ]
Thierry FOURNIERe80fada2015-05-26 18:06:31 +02004903
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02004904 This matches the regular expression in all occurrences of header field <name>
4905 according to <match-regex>, and replaces them with the <replace-fmt> argument.
4906 Format characters are allowed in replace-fmt and work like in <fmt> arguments
4907 in "add-header". The match is only case-sensitive. It is important to
4908 understand that this action only considers whole header lines, regardless of
4909 the number of values they may contain. This usage is suited to headers
4910 naturally containing commas in their value, such as Set-Cookie, Expires and
4911 so on.
William Lallemand86d0df02017-11-24 21:36:45 +01004912
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02004913 Example:
4914 http-response replace-header Set-Cookie (C=[^;]*);(.*) \1;ip=%bi;\2
Willy Tarreau51d861a2015-05-22 17:30:48 +02004915
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02004916 # applied to:
4917 Set-Cookie: C=1; expires=Tue, 14-Jun-2016 01:40:45 GMT
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02004918
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02004919 # outputs:
4920 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 +02004921
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02004922 # assuming the backend IP is 192.168.1.20.
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02004923
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02004924http-response replace-value <name> <regex-match> <replace-fmt>
4925 [ { if | unless } <condition> ]
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02004926
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02004927 This works like "replace-header" except that it matches the regex against
4928 every comma-delimited value of the header field <name> instead of the entire
4929 header. This is suited for all headers which are allowed to carry more than
4930 one value. An example could be the Accept header.
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02004931
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02004932 Example:
4933 http-response replace-value Cache-control ^public$ private
Christopher Faulet85d79c92016-11-09 16:54:56 +01004934
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02004935 # applied to:
4936 Cache-Control: max-age=3600, public
Christopher Faulet85d79c92016-11-09 16:54:56 +01004937
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02004938 # outputs:
4939 Cache-Control: max-age=3600, private
Christopher Faulet85d79c92016-11-09 16:54:56 +01004940
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02004941http-response sc-inc-gpc0(<sc-id>) [ { if | unless } <condition> ]
4942http-response sc-inc-gpc1(<sc-id>) [ { if | unless } <condition> ]
Ruoshan Huange4edc6b2016-07-14 15:07:45 +08004943
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02004944 This action increments the GPC0 or GPC1 counter according with the sticky
4945 counter designated by <sc-id>. If an error occurs, this action silently fails
4946 and the actions evaluation continues.
Thierry FOURNIER236657b2015-08-19 08:25:14 +02004947
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02004948http-response sc-set-gpt0(<sc-id>) <int> [ { if | unless } <condition> ]
Thierry FOURNIERe0627bd2015-08-04 08:20:33 +02004949
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02004950 This action sets the GPT0 tag according to the sticky counter designated by
4951 <sc-id> and the value of <int>. The expected result is a boolean. If an error
4952 occurs, this action silently fails and the actions evaluation continues.
Frédéric Lécaille6778b272018-01-29 15:22:53 +01004953
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02004954http-response send-spoe-group [ { if | unless } <condition> ]
Willy Tarreau2d392c22015-08-24 01:43:45 +02004955
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02004956 This action is used to trigger sending of a group of SPOE messages. To do so,
4957 the SPOE engine used to send messages must be defined, as well as the SPOE
4958 group to send. Of course, the SPOE engine must refer to an existing SPOE
4959 filter. If not engine name is provided on the SPOE filter line, the SPOE
4960 agent name must be used.
Christopher Faulet76c09ef2017-09-21 11:03:52 +02004961
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02004962 Arguments:
4963 <engine-name> The SPOE engine name.
Christopher Faulet76c09ef2017-09-21 11:03:52 +02004964
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02004965 <group-name> The SPOE group name as specified in the engine
4966 configuration.
Christopher Faulet76c09ef2017-09-21 11:03:52 +02004967
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02004968http-response set-header <name> <fmt> [ { if | unless } <condition> ]
Willy Tarreaue365c0b2013-06-11 16:06:12 +02004969
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02004970 This does the same as "add-header" except that the header name is first
4971 removed if it existed. This is useful when passing security information to
4972 the server, where the header must not be manipulated by external users.
Willy Tarreaue365c0b2013-06-11 16:06:12 +02004973
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02004974http-response set-log-level <level> [ { if | unless } <condition> ]
4975
4976 This is used to change the log level of the current request when a certain
4977 condition is met. Valid levels are the 8 syslog levels (see the "log"
4978 keyword) plus the special level "silent" which disables logging for this
4979 request. This rule is not final so the last matching rule wins. This rule can
4980 be useful to disable health checks coming from another equipment.
4981
4982http-response set-map(<file-name>) <key fmt> <value fmt>
4983
4984 This is used to add a new entry into a MAP. The MAP must be loaded from a
4985 file (even a dummy empty file). The file name of the MAP to be updated is
4986 passed between parentheses. It takes 2 arguments: <key fmt>, which follows
4987 log-format rules, used to collect MAP key, and <value fmt>, which follows
4988 log-format rules, used to collect content for the new entry. It performs a
4989 lookup in the MAP before insertion, to avoid duplicated (or more) values.
4990 This lookup is done by a linear search and can be expensive with large lists!
4991 It is the equivalent of the "set map" command from the stats socket, but can
4992 be triggered by an HTTP response.
4993
4994http-response set-mark <mark> [ { if | unless } <condition> ]
4995
4996 This is used to set the Netfilter MARK on all packets sent to the client to
4997 the value passed in <mark> on platforms which support it. This value is an
4998 unsigned 32 bit value which can be matched by netfilter and by the routing
4999 table. It can be expressed both in decimal or hexadecimal format (prefixed
5000 by "0x"). This can be useful to force certain packets to take a different
5001 route (for example a cheaper network path for bulk downloads). This works on
5002 Linux kernels 2.6.32 and above and requires admin privileges.
5003
5004http-response set-nice <nice> [ { if | unless } <condition> ]
5005
5006 This sets the "nice" factor of the current request being processed.
5007 It only has effect against the other requests being processed at the same
5008 time. The default value is 0, unless altered by the "nice" setting on the
5009 "bind" line. The accepted range is -1024..1024. The higher the value, the
5010 nicest the request will be. Lower values will make the request more important
5011 than other ones. This can be useful to improve the speed of some requests, or
5012 lower the priority of non-important requests. Using this setting without
5013 prior experimentation can cause some major slowdown.
5014
5015http-response set-status <status> [reason <str>]
5016 [ { if | unless } <condition> ]
5017
5018 This replaces the response status code with <status> which must be an integer
5019 between 100 and 999. Optionally, a custom reason text can be provided defined
5020 by <str>, or the default reason for the specified code will be used as a
5021 fallback.
Ruoshan Huangeb5a3632015-12-08 21:00:23 +08005022
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02005023 Example:
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02005024 # return "431 Request Header Fields Too Large"
5025 http-response set-status 431
5026 # return "503 Slow Down", custom reason
5027 http-response set-status 503 reason "Slow Down".
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02005028
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02005029http-response set-tos <tos> [ { if | unless } <condition> ]
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02005030
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02005031 This is used to set the TOS or DSCP field value of packets sent to the client
5032 to the value passed in <tos> on platforms which support this.
5033 This value represents the whole 8 bits of the IP TOS field, and can be
5034 expressed both in decimal or hexadecimal format (prefixed by "0x"). Note that
5035 only the 6 higher bits are used in DSCP or TOS, and the two lower bits are
5036 always 0. This can be used to adjust some routing behavior on border routers
5037 based on some information from the request.
5038
5039 See RFC 2474, 2597, 3260 and 4594 for more information.
5040
5041http-response set-var(<var-name>) <expr> [ { if | unless } <condition> ]
5042
5043 This is used to set the contents of a variable. The variable is declared
5044 inline.
5045
5046 Arguments:
5047 <var-name> The name of the variable starts with an indication about its
5048 scope. The scopes allowed are:
5049 "proc" : the variable is shared with the whole process
5050 "sess" : the variable is shared with the whole session
5051 "txn" : the variable is shared with the transaction
5052 (request and response)
5053 "req" : the variable is shared only during request
5054 processing
5055 "res" : the variable is shared only during response
5056 processing
5057 This prefix is followed by a name. The separator is a '.'.
5058 The name may only contain characters 'a-z', 'A-Z', '0-9', '.'
5059 and '_'.
5060
5061 <expr> Is a standard HAProxy expression formed by a sample-fetch
5062 followed by some converters.
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02005063
5064 Example:
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02005065 http-response set-var(sess.last_redir) res.hdr(location)
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02005066
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02005067http-response silent-drop [ { if | unless } <condition> ]
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02005068
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02005069 This stops the evaluation of the rules and makes the client-facing connection
5070 suddenly disappear using a system-dependent way that tries to prevent the
5071 client from being notified. The effect it then that the client still sees an
5072 established connection while there's none on HAProxy. The purpose is to
5073 achieve a comparable effect to "tarpit" except that it doesn't use any local
5074 resource at all on the machine running HAProxy. It can resist much higher
5075 loads than "tarpit", and slow down stronger attackers. It is important to
5076 understand the impact of using this mechanism. All stateful equipment placed
5077 between the client and HAProxy (firewalls, proxies, load balancers) will also
5078 keep the established connection for a long time and may suffer from this
5079 action.
5080 On modern Linux systems running with enough privileges, the TCP_REPAIR socket
5081 option is used to block the emission of a TCP reset. On other systems, the
5082 socket's TTL is reduced to 1 so that the TCP reset doesn't pass the first
5083 router, though it's still delivered to local networks. Do not use it unless
5084 you fully understand how it works.
Baptiste Assmannfabcbe02014-04-24 22:16:59 +02005085
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02005086http-response track-sc0 <key> [table <table>] [ { if | unless } <condition> ]
5087http-response track-sc1 <key> [table <table>] [ { if | unless } <condition> ]
5088http-response track-sc2 <key> [table <table>] [ { if | unless } <condition> ]
Willy Tarreaue365c0b2013-06-11 16:06:12 +02005089
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02005090 This enables tracking of sticky counters from current response. Please refer
5091 to "http-request track-sc" for a complete description. The only difference
5092 from "http-request track-sc" is the <key> sample expression can only make use
5093 of samples in response (e.g. res.*, status etc.) and samples below Layer 6
5094 (e.g. SSL-related samples, see section 7.3.4). If the sample is not
5095 supported, haproxy will fail and warn while parsing the config.
5096
5097http-response unset-var(<var-name>) [ { if | unless } <condition> ]
5098
5099 This is used to unset a variable. See "http-response set-var" for details
5100 about <var-name>.
5101
5102 Example:
5103 http-response unset-var(sess.last_redir)
5104
Baptiste Assmann5ecb77f2013-10-06 23:24:13 +02005105
Willy Tarreau30631952015-08-06 15:05:24 +02005106http-reuse { never | safe | aggressive | always }
5107 Declare how idle HTTP connections may be shared between requests
5108
5109 May be used in sections: defaults | frontend | listen | backend
5110 yes | no | yes | yes
5111
5112 By default, a connection established between haproxy and the backend server
Olivier Houchard86006a52018-12-14 19:37:49 +01005113 which is considered safe for reuse is moved back to the server's idle
5114 connections pool so that any other request can make use of it. This is the
5115 "safe" strategy below.
Willy Tarreau30631952015-08-06 15:05:24 +02005116
5117 The argument indicates the desired connection reuse strategy :
5118
Olivier Houchard86006a52018-12-14 19:37:49 +01005119 - "never" : idle connections are never shared between sessions. This mode
5120 may be enforced to cancel a different strategy inherited from
5121 a defaults section or for troubleshooting. For example, if an
5122 old bogus application considers that multiple requests over
5123 the same connection come from the same client and it is not
5124 possible to fix the application, it may be desirable to
5125 disable connection sharing in a single backend. An example of
5126 such an application could be an old haproxy using cookie
5127 insertion in tunnel mode and not checking any request past the
5128 first one.
Willy Tarreau30631952015-08-06 15:05:24 +02005129
Olivier Houchard86006a52018-12-14 19:37:49 +01005130 - "safe" : this is the default and the recommended strategy. The first
5131 request of a session is always sent over its own connection,
5132 and only subsequent requests may be dispatched over other
5133 existing connections. This ensures that in case the server
5134 closes the connection when the request is being sent, the
5135 browser can decide to silently retry it. Since it is exactly
5136 equivalent to regular keep-alive, there should be no side
5137 effects.
Willy Tarreau30631952015-08-06 15:05:24 +02005138
5139 - "aggressive" : this mode may be useful in webservices environments where
5140 all servers are not necessarily known and where it would be
5141 appreciable to deliver most first requests over existing
5142 connections. In this case, first requests are only delivered
5143 over existing connections that have been reused at least once,
5144 proving that the server correctly supports connection reuse.
5145 It should only be used when it's sure that the client can
5146 retry a failed request once in a while and where the benefit
Michael Prokop4438c602019-05-24 10:25:45 +02005147 of aggressive connection reuse significantly outweighs the
Willy Tarreau30631952015-08-06 15:05:24 +02005148 downsides of rare connection failures.
5149
5150 - "always" : this mode is only recommended when the path to the server is
5151 known for never breaking existing connections quickly after
5152 releasing them. It allows the first request of a session to be
5153 sent to an existing connection. This can provide a significant
5154 performance increase over the "safe" strategy when the backend
5155 is a cache farm, since such components tend to show a
Davor Ocelice9ed2812017-12-25 17:49:28 +01005156 consistent behavior and will benefit from the connection
Willy Tarreau30631952015-08-06 15:05:24 +02005157 sharing. It is recommended that the "http-keep-alive" timeout
5158 remains low in this mode so that no dead connections remain
5159 usable. In most cases, this will lead to the same performance
5160 gains as "aggressive" but with more risks. It should only be
5161 used when it improves the situation over "aggressive".
5162
5163 When http connection sharing is enabled, a great care is taken to respect the
Davor Ocelice9ed2812017-12-25 17:49:28 +01005164 connection properties and compatibility. Specifically :
5165 - connections made with "usesrc" followed by a client-dependent value
5166 ("client", "clientip", "hdr_ip") are marked private and never shared;
Willy Tarreau30631952015-08-06 15:05:24 +02005167
5168 - connections sent to a server with a TLS SNI extension are marked private
Davor Ocelice9ed2812017-12-25 17:49:28 +01005169 and are never shared;
Willy Tarreau30631952015-08-06 15:05:24 +02005170
Lukas Tribusfd9b68c2018-10-27 20:06:59 +02005171 - connections with certain bogus authentication schemes (relying on the
5172 connection) like NTLM are detected, marked private and are never shared;
Willy Tarreau30631952015-08-06 15:05:24 +02005173
5174 No connection pool is involved, once a session dies, the last idle connection
5175 it was attached to is deleted at the same time. This ensures that connections
5176 may not last after all sessions are closed.
5177
5178 Note: connection reuse improves the accuracy of the "server maxconn" setting,
5179 because almost no new connection will be established while idle connections
5180 remain available. This is particularly true with the "always" strategy.
5181
5182 See also : "option http-keep-alive", "server maxconn"
5183
5184
Mark Lamourinec2247f02012-01-04 13:02:01 -05005185http-send-name-header [<header>]
5186 Add the server name to a request. Use the header string given by <header>
5187
5188 May be used in sections: defaults | frontend | listen | backend
5189 yes | no | yes | yes
5190
5191 Arguments :
5192
5193 <header> The header string to use to send the server name
5194
5195 The "http-send-name-header" statement causes the name of the target
Davor Ocelice9ed2812017-12-25 17:49:28 +01005196 server to be added to the headers of an HTTP request. The name
Mark Lamourinec2247f02012-01-04 13:02:01 -05005197 is added with the header string proved.
5198
5199 See also : "server"
5200
Krzysztof Piotr Oledzkif58a9622008-02-23 01:19:10 +01005201id <value>
Willy Tarreau53fb4ae2009-10-04 23:04:08 +02005202 Set a persistent ID to a proxy.
5203 May be used in sections : defaults | frontend | listen | backend
5204 no | yes | yes | yes
5205 Arguments : none
5206
5207 Set a persistent ID for the proxy. This ID must be unique and positive.
5208 An unused ID will automatically be assigned if unset. The first assigned
5209 value will be 1. This ID is currently only returned in statistics.
Krzysztof Piotr Oledzkif58a9622008-02-23 01:19:10 +01005210
5211
Cyril Bonté0d4bf012010-04-25 23:21:46 +02005212ignore-persist { if | unless } <condition>
5213 Declare a condition to ignore persistence
5214 May be used in sections: defaults | frontend | listen | backend
Cyril Bonté4288c5a2018-03-12 22:02:59 +01005215 no | no | yes | yes
Cyril Bonté0d4bf012010-04-25 23:21:46 +02005216
5217 By default, when cookie persistence is enabled, every requests containing
5218 the cookie are unconditionally persistent (assuming the target server is up
5219 and running).
5220
5221 The "ignore-persist" statement allows one to declare various ACL-based
5222 conditions which, when met, will cause a request to ignore persistence.
5223 This is sometimes useful to load balance requests for static files, which
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03005224 often don't require persistence. This can also be used to fully disable
Cyril Bonté0d4bf012010-04-25 23:21:46 +02005225 persistence for a specific User-Agent (for example, some web crawler bots).
5226
Cyril Bonté0d4bf012010-04-25 23:21:46 +02005227 The persistence is ignored when an "if" condition is met, or unless an
5228 "unless" condition is met.
5229
Jarno Huuskonene5ae7022017-04-03 14:36:21 +03005230 Example:
5231 acl url_static path_beg /static /images /img /css
5232 acl url_static path_end .gif .png .jpg .css .js
5233 ignore-persist if url_static
5234
Cyril Bonté0d4bf012010-04-25 23:21:46 +02005235 See also : "force-persist", "cookie", and section 7 about ACL usage.
5236
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02005237load-server-state-from-file { global | local | none }
5238 Allow seamless reload of HAProxy
5239 May be used in sections: defaults | frontend | listen | backend
5240 yes | no | yes | yes
5241
5242 This directive points HAProxy to a file where server state from previous
5243 running process has been saved. That way, when starting up, before handling
5244 traffic, the new process can apply old states to servers exactly has if no
Davor Ocelice9ed2812017-12-25 17:49:28 +01005245 reload occurred. The purpose of the "load-server-state-from-file" directive is
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02005246 to tell haproxy which file to use. For now, only 2 arguments to either prevent
5247 loading state or load states from a file containing all backends and servers.
5248 The state file can be generated by running the command "show servers state"
5249 over the stats socket and redirect output.
5250
Davor Ocelice9ed2812017-12-25 17:49:28 +01005251 The format of the file is versioned and is very specific. To understand it,
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02005252 please read the documentation of the "show servers state" command (chapter
Willy Tarreau1af20c72017-06-23 16:01:14 +02005253 9.3 of Management Guide).
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02005254
5255 Arguments:
5256 global load the content of the file pointed by the global directive
5257 named "server-state-file".
5258
5259 local load the content of the file pointed by the directive
5260 "server-state-file-name" if set. If not set, then the backend
5261 name is used as a file name.
5262
5263 none don't load any stat for this backend
5264
5265 Notes:
Willy Tarreaue5a60682016-11-09 14:54:53 +01005266 - server's IP address is preserved across reloads by default, but the
5267 order can be changed thanks to the server's "init-addr" setting. This
5268 means that an IP address change performed on the CLI at run time will
Davor Ocelice9ed2812017-12-25 17:49:28 +01005269 be preserved, and that any change to the local resolver (e.g. /etc/hosts)
Willy Tarreaue5a60682016-11-09 14:54:53 +01005270 will possibly not have any effect if the state file is in use.
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02005271
5272 - server's weight is applied from previous running process unless it has
5273 has changed between previous and new configuration files.
5274
Olivier Doucetaa1ea8a2016-08-05 17:15:20 +02005275 Example: Minimal configuration
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02005276
Olivier Doucetaa1ea8a2016-08-05 17:15:20 +02005277 global
5278 stats socket /tmp/socket
5279 server-state-file /tmp/server_state
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02005280
Olivier Doucetaa1ea8a2016-08-05 17:15:20 +02005281 defaults
5282 load-server-state-from-file global
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02005283
Olivier Doucetaa1ea8a2016-08-05 17:15:20 +02005284 backend bk
5285 server s1 127.0.0.1:22 check weight 11
5286 server s2 127.0.0.1:22 check weight 12
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02005287
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02005288
5289 Then one can run :
5290
5291 socat /tmp/socket - <<< "show servers state" > /tmp/server_state
5292
5293 Content of the file /tmp/server_state would be like this:
5294
5295 1
5296 # <field names skipped for the doc example>
5297 1 bk 1 s1 127.0.0.1 2 0 11 11 4 6 3 4 6 0 0
5298 1 bk 2 s2 127.0.0.1 2 0 12 12 4 6 3 4 6 0 0
5299
Olivier Doucetaa1ea8a2016-08-05 17:15:20 +02005300 Example: Minimal configuration
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02005301
5302 global
5303 stats socket /tmp/socket
5304 server-state-base /etc/haproxy/states
5305
5306 defaults
5307 load-server-state-from-file local
5308
5309 backend bk
5310 server s1 127.0.0.1:22 check weight 11
5311 server s2 127.0.0.1:22 check weight 12
5312
Olivier Doucetaa1ea8a2016-08-05 17:15:20 +02005313
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02005314 Then one can run :
5315
5316 socat /tmp/socket - <<< "show servers state bk" > /etc/haproxy/states/bk
5317
5318 Content of the file /etc/haproxy/states/bk would be like this:
5319
5320 1
5321 # <field names skipped for the doc example>
5322 1 bk 1 s1 127.0.0.1 2 0 11 11 4 6 3 4 6 0 0
5323 1 bk 2 s2 127.0.0.1 2 0 12 12 4 6 3 4 6 0 0
5324
5325 See also: "server-state-file", "server-state-file-name", and
5326 "show servers state"
5327
Cyril Bonté0d4bf012010-04-25 23:21:46 +02005328
Willy Tarreau2769aa02007-12-27 18:26:09 +01005329log global
Frédéric Lécailled690dfa2019-04-25 10:52:17 +02005330log <address> [len <length>] [format <format>] [sample <ranges>:<smp_size>]
5331 <facility> [<level> [<minlevel>]]
William Lallemand0f99e342011-10-12 17:50:54 +02005332no log
Willy Tarreau2769aa02007-12-27 18:26:09 +01005333 Enable per-instance logging of events and traffic.
5334 May be used in sections : defaults | frontend | listen | backend
5335 yes | yes | yes | yes
William Lallemand0f99e342011-10-12 17:50:54 +02005336
5337 Prefix :
5338 no should be used when the logger list must be flushed. For example,
5339 if you don't want to inherit from the default logger list. This
5340 prefix does not allow arguments.
5341
Willy Tarreau2769aa02007-12-27 18:26:09 +01005342 Arguments :
5343 global should be used when the instance's logging parameters are the
5344 same as the global ones. This is the most common usage. "global"
5345 replaces <address>, <facility> and <level> with those of the log
5346 entries found in the "global" section. Only one "log global"
5347 statement may be used per instance, and this form takes no other
5348 parameter.
5349
5350 <address> indicates where to send the logs. It takes the same format as
5351 for the "global" section's logs, and can be one of :
5352
5353 - An IPv4 address optionally followed by a colon (':') and a UDP
5354 port. If no port is specified, 514 is used by default (the
5355 standard syslog port).
5356
David du Colombier24bb5f52011-03-17 10:40:23 +01005357 - An IPv6 address followed by a colon (':') and optionally a UDP
5358 port. If no port is specified, 514 is used by default (the
5359 standard syslog port).
5360
Willy Tarreau2769aa02007-12-27 18:26:09 +01005361 - A filesystem path to a UNIX domain socket, keeping in mind
5362 considerations for chroot (be sure the path is accessible
5363 inside the chroot) and uid/gid (be sure the path is
Davor Ocelice9ed2812017-12-25 17:49:28 +01005364 appropriately writable).
Willy Tarreau2769aa02007-12-27 18:26:09 +01005365
Willy Tarreau5a32ecc2018-11-12 07:34:59 +01005366 - A file descriptor number in the form "fd@<number>", which may
5367 point to a pipe, terminal, or socket. In this case unbuffered
5368 logs are used and one writev() call per log is performed. This
5369 is a bit expensive but acceptable for most workloads. Messages
5370 sent this way will not be truncated but may be dropped, in
5371 which case the DroppedLogs counter will be incremented. The
5372 writev() call is atomic even on pipes for messages up to
5373 PIPE_BUF size, which POSIX recommends to be at least 512 and
5374 which is 4096 bytes on most modern operating systems. Any
5375 larger message may be interleaved with messages from other
5376 processes. Exceptionally for debugging purposes the file
5377 descriptor may also be directed to a file, but doing so will
5378 significantly slow haproxy down as non-blocking calls will be
5379 ignored. Also there will be no way to purge nor rotate this
5380 file without restarting the process. Note that the configured
5381 syslog format is preserved, so the output is suitable for use
Willy Tarreauc1b06452018-11-12 11:57:56 +01005382 with a TCP syslog server. See also the "short" and "raw"
5383 formats below.
Willy Tarreau5a32ecc2018-11-12 07:34:59 +01005384
5385 - "stdout" / "stderr", which are respectively aliases for "fd@1"
5386 and "fd@2", see above.
5387
5388 You may want to reference some environment variables in the
5389 address parameter, see section 2.3 about environment variables.
Willy Tarreaudad36a32013-03-11 01:20:04 +01005390
Willy Tarreau18324f52014-06-27 18:10:07 +02005391 <length> is an optional maximum line length. Log lines larger than this
5392 value will be truncated before being sent. The reason is that
5393 syslog servers act differently on log line length. All servers
5394 support the default value of 1024, but some servers simply drop
5395 larger lines while others do log them. If a server supports long
5396 lines, it may make sense to set this value here in order to avoid
5397 truncating long lines. Similarly, if a server drops long lines,
5398 it is preferable to truncate them before sending them. Accepted
5399 values are 80 to 65535 inclusive. The default value of 1024 is
5400 generally fine for all standard usages. Some specific cases of
Davor Ocelice9ed2812017-12-25 17:49:28 +01005401 long captures or JSON-formatted logs may require larger values.
Willy Tarreau18324f52014-06-27 18:10:07 +02005402
Frédéric Lécailled690dfa2019-04-25 10:52:17 +02005403 <ranges> A list of comma-separated ranges to identify the logs to sample.
5404 This is used to balance the load of the logs to send to the log
5405 server. The limits of the ranges cannot be null. They are numbered
5406 from 1. The size or period (in number of logs) of the sample must
5407 be set with <sample_size> parameter.
5408
5409 <sample_size>
5410 The size of the sample in number of logs to consider when balancing
5411 their logging loads. It is used to balance the load of the logs to
5412 send to the syslog server. This size must be greater or equal to the
5413 maximum of the high limits of the ranges.
5414 (see also <ranges> parameter).
5415
Willy Tarreauadb345d2018-11-12 07:56:13 +01005416 <format> is the log format used when generating syslog messages. It may be
5417 one of the following :
5418
5419 rfc3164 The RFC3164 syslog message format. This is the default.
5420 (https://tools.ietf.org/html/rfc3164)
5421
5422 rfc5424 The RFC5424 syslog message format.
5423 (https://tools.ietf.org/html/rfc5424)
5424
Willy Tarreaue8746a02018-11-12 08:45:00 +01005425 short A message containing only a level between angle brackets such as
5426 '<3>', followed by the text. The PID, date, time, process name
5427 and system name are omitted. This is designed to be used with a
5428 local log server. This format is compatible with what the
5429 systemd logger consumes.
5430
Willy Tarreauc1b06452018-11-12 11:57:56 +01005431 raw A message containing only the text. The level, PID, date, time,
5432 process name and system name are omitted. This is designed to
5433 be used in containers or during development, where the severity
5434 only depends on the file descriptor used (stdout/stderr).
5435
Willy Tarreau2769aa02007-12-27 18:26:09 +01005436 <facility> must be one of the 24 standard syslog facilities :
5437
Willy Tarreaue8746a02018-11-12 08:45:00 +01005438 kern user mail daemon auth syslog lpr news
5439 uucp cron auth2 ftp ntp audit alert cron2
5440 local0 local1 local2 local3 local4 local5 local6 local7
5441
Willy Tarreauc1b06452018-11-12 11:57:56 +01005442 Note that the facility is ignored for the "short" and "raw"
5443 formats, but still required as a positional field. It is
5444 recommended to use "daemon" in this case to make it clear that
5445 it's only supposed to be used locally.
Willy Tarreau2769aa02007-12-27 18:26:09 +01005446
5447 <level> is optional and can be specified to filter outgoing messages. By
5448 default, all messages are sent. If a level is specified, only
5449 messages with a severity at least as important as this level
Willy Tarreauf7edefa2009-05-10 17:20:05 +02005450 will be sent. An optional minimum level can be specified. If it
5451 is set, logs emitted with a more severe level than this one will
5452 be capped to this level. This is used to avoid sending "emerg"
5453 messages on all terminals on some default syslog configurations.
5454 Eight levels are known :
Willy Tarreau2769aa02007-12-27 18:26:09 +01005455
5456 emerg alert crit err warning notice info debug
5457
William Lallemand0f99e342011-10-12 17:50:54 +02005458 It is important to keep in mind that it is the frontend which decides what to
5459 log from a connection, and that in case of content switching, the log entries
5460 from the backend will be ignored. Connections are logged at level "info".
Willy Tarreaucc6c8912009-02-22 10:53:55 +01005461
5462 However, backend log declaration define how and where servers status changes
5463 will be logged. Level "notice" will be used to indicate a server going up,
5464 "warning" will be used for termination signals and definitive service
5465 termination, and "alert" will be used for when a server goes down.
5466
5467 Note : According to RFC3164, messages are truncated to 1024 bytes before
5468 being emitted.
Willy Tarreau2769aa02007-12-27 18:26:09 +01005469
5470 Example :
5471 log global
Willy Tarreauc1b06452018-11-12 11:57:56 +01005472 log stdout format short daemon # send log to systemd
5473 log stdout format raw daemon # send everything to stdout
5474 log stderr format raw daemon notice # send important events to stderr
Willy Tarreauf7edefa2009-05-10 17:20:05 +02005475 log 127.0.0.1:514 local0 notice # only send important events
5476 log 127.0.0.1:514 local0 notice notice # same but limit output level
William Lallemandb2f07452015-05-12 14:27:13 +02005477 log "${LOCAL_SYSLOG}:514" local0 notice # send to local server
Willy Tarreaudad36a32013-03-11 01:20:04 +01005478
Willy Tarreau2769aa02007-12-27 18:26:09 +01005479
William Lallemand48940402012-01-30 16:47:22 +01005480log-format <string>
Willy Tarreaufb4e7ea2015-01-07 14:55:17 +01005481 Specifies the log format string to use for traffic logs
5482 May be used in sections: defaults | frontend | listen | backend
5483 yes | yes | yes | no
William Lallemand48940402012-01-30 16:47:22 +01005484
Willy Tarreaufb4e7ea2015-01-07 14:55:17 +01005485 This directive specifies the log format string that will be used for all logs
5486 resulting from traffic passing through the frontend using this line. If the
5487 directive is used in a defaults section, all subsequent frontends will use
5488 the same log format. Please see section 8.2.4 which covers the log format
5489 string in depth.
William Lallemand48940402012-01-30 16:47:22 +01005490
Guillaume de Lafond29f45602017-03-31 19:52:15 +02005491 "log-format" directive overrides previous "option tcplog", "log-format" and
5492 "option httplog" directives.
5493
Dragan Dosen7ad31542015-09-28 17:16:47 +02005494log-format-sd <string>
5495 Specifies the RFC5424 structured-data log format string
5496 May be used in sections: defaults | frontend | listen | backend
5497 yes | yes | yes | no
5498
5499 This directive specifies the RFC5424 structured-data log format string that
5500 will be used for all logs resulting from traffic passing through the frontend
5501 using this line. If the directive is used in a defaults section, all
5502 subsequent frontends will use the same log format. Please see section 8.2.4
5503 which covers the log format string in depth.
5504
5505 See https://tools.ietf.org/html/rfc5424#section-6.3 for more information
5506 about the RFC5424 structured-data part.
5507
5508 Note : This log format string will be used only for loggers that have set
5509 log format to "rfc5424".
5510
5511 Example :
5512 log-format-sd [exampleSDID@1234\ bytes=\"%B\"\ status=\"%ST\"]
5513
5514
Willy Tarreau094af4e2015-01-07 15:03:42 +01005515log-tag <string>
5516 Specifies the log tag to use for all outgoing logs
5517 May be used in sections: defaults | frontend | listen | backend
5518 yes | yes | yes | yes
5519
5520 Sets the tag field in the syslog header to this string. It defaults to the
5521 log-tag set in the global section, otherwise the program name as launched
5522 from the command line, which usually is "haproxy". Sometimes it can be useful
5523 to differentiate between multiple processes running on the same host, or to
5524 differentiate customer instances running in the same process. In the backend,
5525 logs about servers up/down will use this tag. As a hint, it can be convenient
5526 to set a log-tag related to a hosted customer in a defaults section then put
5527 all the frontends and backends for that customer, then start another customer
5528 in a new defaults section. See also the global "log-tag" directive.
Willy Tarreau2769aa02007-12-27 18:26:09 +01005529
Willy Tarreauc35362a2014-04-25 13:58:37 +02005530max-keep-alive-queue <value>
5531 Set the maximum server queue size for maintaining keep-alive connections
5532 May be used in sections: defaults | frontend | listen | backend
5533 yes | no | yes | yes
5534
5535 HTTP keep-alive tries to reuse the same server connection whenever possible,
5536 but sometimes it can be counter-productive, for example if a server has a lot
5537 of connections while other ones are idle. This is especially true for static
5538 servers.
5539
5540 The purpose of this setting is to set a threshold on the number of queued
5541 connections at which haproxy stops trying to reuse the same server and prefers
5542 to find another one. The default value, -1, means there is no limit. A value
5543 of zero means that keep-alive requests will never be queued. For very close
5544 servers which can be reached with a low latency and which are not sensible to
Davor Ocelice9ed2812017-12-25 17:49:28 +01005545 breaking keep-alive, a low value is recommended (e.g. local static server can
Willy Tarreauc35362a2014-04-25 13:58:37 +02005546 use a value of 10 or less). For remote servers suffering from a high latency,
5547 higher values might be needed to cover for the latency and/or the cost of
5548 picking a different server.
5549
5550 Note that this has no impact on responses which are maintained to the same
5551 server consecutively to a 401 response. They will still go to the same server
5552 even if they have to be queued.
5553
5554 See also : "option http-server-close", "option prefer-last-server", server
5555 "maxconn" and cookie persistence.
5556
Olivier Houcharda4d4fdf2018-12-14 19:27:06 +01005557max-session-srv-conns <nb>
5558 Set the maximum number of outgoing connections we can keep idling for a given
5559 client session. The default is 5 (it precisely equals MAX_SRV_LIST which is
5560 defined at build time).
Willy Tarreauc35362a2014-04-25 13:58:37 +02005561
Willy Tarreau2769aa02007-12-27 18:26:09 +01005562maxconn <conns>
5563 Fix the maximum number of concurrent connections on a frontend
5564 May be used in sections : defaults | frontend | listen | backend
5565 yes | yes | yes | no
5566 Arguments :
5567 <conns> is the maximum number of concurrent connections the frontend will
5568 accept to serve. Excess connections will be queued by the system
5569 in the socket's listen queue and will be served once a connection
5570 closes.
5571
5572 If the system supports it, it can be useful on big sites to raise this limit
5573 very high so that haproxy manages connection queues, instead of leaving the
5574 clients with unanswered connection attempts. This value should not exceed the
5575 global maxconn. Also, keep in mind that a connection contains two buffers
Baptiste Assmann79fb45d2016-03-06 23:34:31 +01005576 of tune.bufsize (16kB by default) each, as well as some other data resulting
5577 in about 33 kB of RAM being consumed per established connection. That means
5578 that a medium system equipped with 1GB of RAM can withstand around
5579 20000-25000 concurrent connections if properly tuned.
Willy Tarreau2769aa02007-12-27 18:26:09 +01005580
5581 Also, when <conns> is set to large values, it is possible that the servers
5582 are not sized to accept such loads, and for this reason it is generally wise
5583 to assign them some reasonable connection limits.
5584
Willy Tarreauc8d5b952019-02-27 17:25:52 +01005585 When this value is set to zero, which is the default, the global "maxconn"
5586 value is used.
Vincent Bernat6341be52012-06-27 17:18:30 +02005587
Willy Tarreau2769aa02007-12-27 18:26:09 +01005588 See also : "server", global section's "maxconn", "fullconn"
5589
5590
5591mode { tcp|http|health }
5592 Set the running mode or protocol of the instance
5593 May be used in sections : defaults | frontend | listen | backend
5594 yes | yes | yes | yes
5595 Arguments :
5596 tcp The instance will work in pure TCP mode. A full-duplex connection
5597 will be established between clients and servers, and no layer 7
5598 examination will be performed. This is the default mode. It
5599 should be used for SSL, SSH, SMTP, ...
5600
5601 http The instance will work in HTTP mode. The client request will be
5602 analyzed in depth before connecting to any server. Any request
5603 which is not RFC-compliant will be rejected. Layer 7 filtering,
5604 processing and switching will be possible. This is the mode which
5605 brings HAProxy most of its value.
5606
5607 health The instance will work in "health" mode. It will just reply "OK"
Willy Tarreau82569f92012-09-27 23:48:56 +02005608 to incoming connections and close the connection. Alternatively,
5609 If the "httpchk" option is set, "HTTP/1.0 200 OK" will be sent
5610 instead. Nothing will be logged in either case. This mode is used
5611 to reply to external components health checks. This mode is
5612 deprecated and should not be used anymore as it is possible to do
5613 the same and even better by combining TCP or HTTP modes with the
5614 "monitor" keyword.
Willy Tarreau2769aa02007-12-27 18:26:09 +01005615
Cyril Bonté108cf6e2012-04-21 23:30:29 +02005616 When doing content switching, it is mandatory that the frontend and the
5617 backend are in the same mode (generally HTTP), otherwise the configuration
5618 will be refused.
Willy Tarreau2769aa02007-12-27 18:26:09 +01005619
Cyril Bonté108cf6e2012-04-21 23:30:29 +02005620 Example :
Willy Tarreau2769aa02007-12-27 18:26:09 +01005621 defaults http_instances
5622 mode http
5623
Cyril Bonté108cf6e2012-04-21 23:30:29 +02005624 See also : "monitor", "monitor-net"
Willy Tarreau2769aa02007-12-27 18:26:09 +01005625
Willy Tarreau0ba27502007-12-24 16:55:16 +01005626
Cyril Bontéf0c60612010-02-06 14:44:47 +01005627monitor fail { if | unless } <condition>
Willy Tarreau2769aa02007-12-27 18:26:09 +01005628 Add a condition to report a failure to a monitor HTTP request.
Willy Tarreau0ba27502007-12-24 16:55:16 +01005629 May be used in sections : defaults | frontend | listen | backend
5630 no | yes | yes | no
Willy Tarreau0ba27502007-12-24 16:55:16 +01005631 Arguments :
5632 if <cond> the monitor request will fail if the condition is satisfied,
5633 and will succeed otherwise. The condition should describe a
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01005634 combined test which must induce a failure if all conditions
Willy Tarreau0ba27502007-12-24 16:55:16 +01005635 are met, for instance a low number of servers both in a
5636 backend and its backup.
5637
5638 unless <cond> the monitor request will succeed only if the condition is
5639 satisfied, and will fail otherwise. Such a condition may be
5640 based on a test on the presence of a minimum number of active
5641 servers in a list of backends.
5642
5643 This statement adds a condition which can force the response to a monitor
5644 request to report a failure. By default, when an external component queries
5645 the URI dedicated to monitoring, a 200 response is returned. When one of the
5646 conditions above is met, haproxy will return 503 instead of 200. This is
5647 very useful to report a site failure to an external component which may base
5648 routing advertisements between multiple sites on the availability reported by
5649 haproxy. In this case, one would rely on an ACL involving the "nbsrv"
Willy Tarreauae94d4d2011-05-11 16:28:49 +02005650 criterion. Note that "monitor fail" only works in HTTP mode. Both status
5651 messages may be tweaked using "errorfile" or "errorloc" if needed.
Willy Tarreau0ba27502007-12-24 16:55:16 +01005652
5653 Example:
5654 frontend www
Willy Tarreau2769aa02007-12-27 18:26:09 +01005655 mode http
Willy Tarreau0ba27502007-12-24 16:55:16 +01005656 acl site_dead nbsrv(dynamic) lt 2
5657 acl site_dead nbsrv(static) lt 2
5658 monitor-uri /site_alive
5659 monitor fail if site_dead
5660
Willy Tarreauae94d4d2011-05-11 16:28:49 +02005661 See also : "monitor-net", "monitor-uri", "errorfile", "errorloc"
Willy Tarreau2769aa02007-12-27 18:26:09 +01005662
5663
5664monitor-net <source>
5665 Declare a source network which is limited to monitor requests
5666 May be used in sections : defaults | frontend | listen | backend
5667 yes | yes | yes | no
5668 Arguments :
5669 <source> is the source IPv4 address or network which will only be able to
5670 get monitor responses to any request. It can be either an IPv4
5671 address, a host name, or an address followed by a slash ('/')
5672 followed by a mask.
5673
5674 In TCP mode, any connection coming from a source matching <source> will cause
5675 the connection to be immediately closed without any log. This allows another
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01005676 equipment to probe the port and verify that it is still listening, without
Willy Tarreau2769aa02007-12-27 18:26:09 +01005677 forwarding the connection to a remote server.
5678
5679 In HTTP mode, a connection coming from a source matching <source> will be
5680 accepted, the following response will be sent without waiting for a request,
5681 then the connection will be closed : "HTTP/1.0 200 OK". This is normally
5682 enough for any front-end HTTP probe to detect that the service is UP and
Willy Tarreau82569f92012-09-27 23:48:56 +02005683 running without forwarding the request to a backend server. Note that this
5684 response is sent in raw format, without any transformation. This is important
5685 as it means that it will not be SSL-encrypted on SSL listeners.
Willy Tarreau2769aa02007-12-27 18:26:09 +01005686
Willy Tarreau82569f92012-09-27 23:48:56 +02005687 Monitor requests are processed very early, just after tcp-request connection
5688 ACLs which are the only ones able to block them. These connections are short
5689 lived and never wait for any data from the client. They cannot be logged, and
5690 it is the intended purpose. They are only used to report HAProxy's health to
5691 an upper component, nothing more. Please note that "monitor fail" rules do
5692 not apply to connections intercepted by "monitor-net".
Willy Tarreau2769aa02007-12-27 18:26:09 +01005693
Willy Tarreau95cd2832010-03-04 23:36:33 +01005694 Last, please note that only one "monitor-net" statement can be specified in
5695 a frontend. If more than one is found, only the last one will be considered.
Cyril Bonté108cf6e2012-04-21 23:30:29 +02005696
Willy Tarreau2769aa02007-12-27 18:26:09 +01005697 Example :
5698 # addresses .252 and .253 are just probing us.
5699 frontend www
5700 monitor-net 192.168.0.252/31
5701
5702 See also : "monitor fail", "monitor-uri"
5703
5704
5705monitor-uri <uri>
5706 Intercept a URI used by external components' monitor requests
5707 May be used in sections : defaults | frontend | listen | backend
5708 yes | yes | yes | no
5709 Arguments :
5710 <uri> is the exact URI which we want to intercept to return HAProxy's
5711 health status instead of forwarding the request.
5712
5713 When an HTTP request referencing <uri> will be received on a frontend,
5714 HAProxy will not forward it nor log it, but instead will return either
5715 "HTTP/1.0 200 OK" or "HTTP/1.0 503 Service unavailable", depending on failure
5716 conditions defined with "monitor fail". This is normally enough for any
5717 front-end HTTP probe to detect that the service is UP and running without
5718 forwarding the request to a backend server. Note that the HTTP method, the
5719 version and all headers are ignored, but the request must at least be valid
5720 at the HTTP level. This keyword may only be used with an HTTP-mode frontend.
5721
Willy Tarreau721d8e02017-12-01 18:25:08 +01005722 Monitor requests are processed very early, just after the request is parsed
5723 and even before any "http-request" or "block" rulesets. The only rulesets
5724 applied before are the tcp-request ones. They cannot be logged either, and it
5725 is the intended purpose. They are only used to report HAProxy's health to an
5726 upper component, nothing more. However, it is possible to add any number of
5727 conditions using "monitor fail" and ACLs so that the result can be adjusted
5728 to whatever check can be imagined (most often the number of available servers
5729 in a backend).
Willy Tarreau2769aa02007-12-27 18:26:09 +01005730
5731 Example :
5732 # Use /haproxy_test to report haproxy's status
5733 frontend www
5734 mode http
5735 monitor-uri /haproxy_test
5736
5737 See also : "monitor fail", "monitor-net"
5738
Willy Tarreau0ba27502007-12-24 16:55:16 +01005739
Willy Tarreaubf1f8162007-12-28 17:42:56 +01005740option abortonclose
5741no option abortonclose
5742 Enable or disable early dropping of aborted requests pending in queues.
5743 May be used in sections : defaults | frontend | listen | backend
5744 yes | no | yes | yes
5745 Arguments : none
5746
5747 In presence of very high loads, the servers will take some time to respond.
5748 The per-instance connection queue will inflate, and the response time will
5749 increase respective to the size of the queue times the average per-session
5750 response time. When clients will wait for more than a few seconds, they will
Willy Tarreau198a7442008-01-17 12:05:32 +01005751 often hit the "STOP" button on their browser, leaving a useless request in
Willy Tarreaubf1f8162007-12-28 17:42:56 +01005752 the queue, and slowing down other users, and the servers as well, because the
5753 request will eventually be served, then aborted at the first error
5754 encountered while delivering the response.
5755
5756 As there is no way to distinguish between a full STOP and a simple output
5757 close on the client side, HTTP agents should be conservative and consider
5758 that the client might only have closed its output channel while waiting for
5759 the response. However, this introduces risks of congestion when lots of users
5760 do the same, and is completely useless nowadays because probably no client at
5761 all will close the session while waiting for the response. Some HTTP agents
Davor Ocelice9ed2812017-12-25 17:49:28 +01005762 support this behavior (Squid, Apache, HAProxy), and others do not (TUX, most
Willy Tarreaubf1f8162007-12-28 17:42:56 +01005763 hardware-based load balancers). So the probability for a closed input channel
Willy Tarreau198a7442008-01-17 12:05:32 +01005764 to represent a user hitting the "STOP" button is close to 100%, and the risk
Willy Tarreaubf1f8162007-12-28 17:42:56 +01005765 of being the single component to break rare but valid traffic is extremely
5766 low, which adds to the temptation to be able to abort a session early while
5767 still not served and not pollute the servers.
5768
Davor Ocelice9ed2812017-12-25 17:49:28 +01005769 In HAProxy, the user can choose the desired behavior using the option
5770 "abortonclose". By default (without the option) the behavior is HTTP
Willy Tarreaubf1f8162007-12-28 17:42:56 +01005771 compliant and aborted requests will be served. But when the option is
5772 specified, a session with an incoming channel closed will be aborted while
5773 it is still possible, either pending in the queue for a connection slot, or
5774 during the connection establishment if the server has not yet acknowledged
5775 the connection request. This considerably reduces the queue size and the load
5776 on saturated servers when users are tempted to click on STOP, which in turn
Willy Tarreaud72758d2010-01-12 10:42:19 +01005777 reduces the response time for other users.
Willy Tarreaubf1f8162007-12-28 17:42:56 +01005778
5779 If this option has been enabled in a "defaults" section, it can be disabled
5780 in a specific instance by prepending the "no" keyword before it.
5781
5782 See also : "timeout queue" and server's "maxconn" and "maxqueue" parameters
5783
5784
Willy Tarreau4076a152009-04-02 15:18:36 +02005785option accept-invalid-http-request
5786no option accept-invalid-http-request
5787 Enable or disable relaxing of HTTP request parsing
5788 May be used in sections : defaults | frontend | listen | backend
5789 yes | yes | yes | no
5790 Arguments : none
5791
Willy Tarreau91852eb2015-05-01 13:26:00 +02005792 By default, HAProxy complies with RFC7230 in terms of message parsing. This
Willy Tarreau4076a152009-04-02 15:18:36 +02005793 means that invalid characters in header names are not permitted and cause an
Davor Ocelice9ed2812017-12-25 17:49:28 +01005794 error to be returned to the client. This is the desired behavior as such
Willy Tarreau4076a152009-04-02 15:18:36 +02005795 forbidden characters are essentially used to build attacks exploiting server
5796 weaknesses, and bypass security filtering. Sometimes, a buggy browser or
5797 server will emit invalid header names for whatever reason (configuration,
5798 implementation) and the issue will not be immediately fixed. In such a case,
5799 it is possible to relax HAProxy's header name parser to accept any character
Willy Tarreau422246e2012-01-07 23:54:13 +01005800 even if that does not make sense, by specifying this option. Similarly, the
5801 list of characters allowed to appear in a URI is well defined by RFC3986, and
5802 chars 0-31, 32 (space), 34 ('"'), 60 ('<'), 62 ('>'), 92 ('\'), 94 ('^'), 96
5803 ('`'), 123 ('{'), 124 ('|'), 125 ('}'), 127 (delete) and anything above are
Davor Ocelice9ed2812017-12-25 17:49:28 +01005804 not allowed at all. HAProxy always blocks a number of them (0..32, 127). The
Willy Tarreau91852eb2015-05-01 13:26:00 +02005805 remaining ones are blocked by default unless this option is enabled. This
Willy Tarreau13317662015-05-01 13:47:08 +02005806 option also relaxes the test on the HTTP version, it allows HTTP/0.9 requests
5807 to pass through (no version specified) and multiple digits for both the major
5808 and the minor version.
Willy Tarreau4076a152009-04-02 15:18:36 +02005809
5810 This option should never be enabled by default as it hides application bugs
5811 and open security breaches. It should only be deployed after a problem has
5812 been confirmed.
5813
5814 When this option is enabled, erroneous header names will still be accepted in
5815 requests, but the complete request will be captured in order to permit later
Willy Tarreau422246e2012-01-07 23:54:13 +01005816 analysis using the "show errors" request on the UNIX stats socket. Similarly,
5817 requests containing invalid chars in the URI part will be logged. Doing this
Willy Tarreau4076a152009-04-02 15:18:36 +02005818 also helps confirming that the issue has been solved.
5819
5820 If this option has been enabled in a "defaults" section, it can be disabled
5821 in a specific instance by prepending the "no" keyword before it.
5822
5823 See also : "option accept-invalid-http-response" and "show errors" on the
5824 stats socket.
5825
5826
5827option accept-invalid-http-response
5828no option accept-invalid-http-response
5829 Enable or disable relaxing of HTTP response parsing
5830 May be used in sections : defaults | frontend | listen | backend
5831 yes | no | yes | yes
5832 Arguments : none
5833
Willy Tarreau91852eb2015-05-01 13:26:00 +02005834 By default, HAProxy complies with RFC7230 in terms of message parsing. This
Willy Tarreau4076a152009-04-02 15:18:36 +02005835 means that invalid characters in header names are not permitted and cause an
Davor Ocelice9ed2812017-12-25 17:49:28 +01005836 error to be returned to the client. This is the desired behavior as such
Willy Tarreau4076a152009-04-02 15:18:36 +02005837 forbidden characters are essentially used to build attacks exploiting server
5838 weaknesses, and bypass security filtering. Sometimes, a buggy browser or
5839 server will emit invalid header names for whatever reason (configuration,
5840 implementation) and the issue will not be immediately fixed. In such a case,
5841 it is possible to relax HAProxy's header name parser to accept any character
Willy Tarreau91852eb2015-05-01 13:26:00 +02005842 even if that does not make sense, by specifying this option. This option also
5843 relaxes the test on the HTTP version format, it allows multiple digits for
5844 both the major and the minor version.
Willy Tarreau4076a152009-04-02 15:18:36 +02005845
5846 This option should never be enabled by default as it hides application bugs
5847 and open security breaches. It should only be deployed after a problem has
5848 been confirmed.
5849
5850 When this option is enabled, erroneous header names will still be accepted in
5851 responses, but the complete response will be captured in order to permit
5852 later analysis using the "show errors" request on the UNIX stats socket.
5853 Doing this also helps confirming that the issue has been solved.
5854
5855 If this option has been enabled in a "defaults" section, it can be disabled
5856 in a specific instance by prepending the "no" keyword before it.
5857
5858 See also : "option accept-invalid-http-request" and "show errors" on the
5859 stats socket.
5860
5861
Willy Tarreaubf1f8162007-12-28 17:42:56 +01005862option allbackups
5863no option allbackups
5864 Use either all backup servers at a time or only the first one
5865 May be used in sections : defaults | frontend | listen | backend
5866 yes | no | yes | yes
5867 Arguments : none
5868
5869 By default, the first operational backup server gets all traffic when normal
5870 servers are all down. Sometimes, it may be preferred to use multiple backups
5871 at once, because one will not be enough. When "option allbackups" is enabled,
5872 the load balancing will be performed among all backup servers when all normal
5873 ones are unavailable. The same load balancing algorithm will be used and the
5874 servers' weights will be respected. Thus, there will not be any priority
5875 order between the backup servers anymore.
5876
5877 This option is mostly used with static server farms dedicated to return a
5878 "sorry" page when an application is completely offline.
5879
5880 If this option has been enabled in a "defaults" section, it can be disabled
5881 in a specific instance by prepending the "no" keyword before it.
5882
5883
5884option checkcache
5885no option checkcache
Godbach7056a352013-12-11 20:01:07 +08005886 Analyze all server responses and block responses with cacheable cookies
Willy Tarreaubf1f8162007-12-28 17:42:56 +01005887 May be used in sections : defaults | frontend | listen | backend
5888 yes | no | yes | yes
5889 Arguments : none
5890
5891 Some high-level frameworks set application cookies everywhere and do not
5892 always let enough control to the developer to manage how the responses should
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01005893 be cached. When a session cookie is returned on a cacheable object, there is a
Willy Tarreaubf1f8162007-12-28 17:42:56 +01005894 high risk of session crossing or stealing between users traversing the same
5895 caches. In some situations, it is better to block the response than to let
Willy Tarreau3c92c5f2011-08-28 09:45:47 +02005896 some sensitive session information go in the wild.
Willy Tarreaubf1f8162007-12-28 17:42:56 +01005897
5898 The option "checkcache" enables deep inspection of all server responses for
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01005899 strict compliance with HTTP specification in terms of cacheability. It
Willy Tarreau198a7442008-01-17 12:05:32 +01005900 carefully checks "Cache-control", "Pragma" and "Set-cookie" headers in server
Willy Tarreaubf1f8162007-12-28 17:42:56 +01005901 response to check if there's a risk of caching a cookie on a client-side
5902 proxy. When this option is enabled, the only responses which can be delivered
Willy Tarreau198a7442008-01-17 12:05:32 +01005903 to the client are :
Davor Ocelice9ed2812017-12-25 17:49:28 +01005904 - all those without "Set-Cookie" header;
Willy Tarreauc55ddce2017-12-21 11:41:38 +01005905 - all those with a return code other than 200, 203, 204, 206, 300, 301,
5906 404, 405, 410, 414, 501, provided that the server has not set a
Davor Ocelice9ed2812017-12-25 17:49:28 +01005907 "Cache-control: public" header field;
Willy Tarreau24ea0bc2017-12-21 11:32:55 +01005908 - all those that result from a request using a method other than GET, HEAD,
5909 OPTIONS, TRACE, provided that the server has not set a 'Cache-Control:
Davor Ocelice9ed2812017-12-25 17:49:28 +01005910 public' header field;
Willy Tarreaubf1f8162007-12-28 17:42:56 +01005911 - those with a 'Pragma: no-cache' header
5912 - those with a 'Cache-control: private' header
5913 - those with a 'Cache-control: no-store' header
5914 - those with a 'Cache-control: max-age=0' header
5915 - those with a 'Cache-control: s-maxage=0' header
5916 - those with a 'Cache-control: no-cache' header
5917 - those with a 'Cache-control: no-cache="set-cookie"' header
5918 - those with a 'Cache-control: no-cache="set-cookie,' header
5919 (allowing other fields after set-cookie)
5920
5921 If a response doesn't respect these requirements, then it will be blocked
Willy Tarreau198a7442008-01-17 12:05:32 +01005922 just as if it was from an "rspdeny" filter, with an "HTTP 502 bad gateway".
Willy Tarreaubf1f8162007-12-28 17:42:56 +01005923 The session state shows "PH--" meaning that the proxy blocked the response
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01005924 during headers processing. Additionally, an alert will be sent in the logs so
Willy Tarreaubf1f8162007-12-28 17:42:56 +01005925 that admins are informed that there's something to be fixed.
5926
5927 Due to the high impact on the application, the application should be tested
5928 in depth with the option enabled before going to production. It is also a
Willy Tarreaud2a4aa22008-01-31 15:28:22 +01005929 good practice to always activate it during tests, even if it is not used in
Davor Ocelice9ed2812017-12-25 17:49:28 +01005930 production, as it will report potentially dangerous application behaviors.
Willy Tarreaubf1f8162007-12-28 17:42:56 +01005931
5932 If this option has been enabled in a "defaults" section, it can be disabled
5933 in a specific instance by prepending the "no" keyword before it.
5934
5935
5936option clitcpka
5937no option clitcpka
5938 Enable or disable the sending of TCP keepalive packets on the client side
5939 May be used in sections : defaults | frontend | listen | backend
5940 yes | yes | yes | no
5941 Arguments : none
5942
5943 When there is a firewall or any session-aware component between a client and
5944 a server, and when the protocol involves very long sessions with long idle
Davor Ocelice9ed2812017-12-25 17:49:28 +01005945 periods (e.g. remote desktops), there is a risk that one of the intermediate
Willy Tarreaubf1f8162007-12-28 17:42:56 +01005946 components decides to expire a session which has remained idle for too long.
5947
5948 Enabling socket-level TCP keep-alives makes the system regularly send packets
5949 to the other end of the connection, leaving it active. The delay between
5950 keep-alive probes is controlled by the system only and depends both on the
5951 operating system and its tuning parameters.
5952
5953 It is important to understand that keep-alive packets are neither emitted nor
5954 received at the application level. It is only the network stacks which sees
5955 them. For this reason, even if one side of the proxy already uses keep-alives
5956 to maintain its connection alive, those keep-alive packets will not be
5957 forwarded to the other side of the proxy.
5958
5959 Please note that this has nothing to do with HTTP keep-alive.
5960
5961 Using option "clitcpka" enables the emission of TCP keep-alive probes on the
5962 client side of a connection, which should help when session expirations are
5963 noticed between HAProxy and a client.
5964
5965 If this option has been enabled in a "defaults" section, it can be disabled
5966 in a specific instance by prepending the "no" keyword before it.
5967
5968 See also : "option srvtcpka", "option tcpka"
5969
5970
Willy Tarreau0ba27502007-12-24 16:55:16 +01005971option contstats
5972 Enable continuous traffic statistics updates
5973 May be used in sections : defaults | frontend | listen | backend
5974 yes | yes | yes | no
5975 Arguments : none
5976
5977 By default, counters used for statistics calculation are incremented
5978 only when a session finishes. It works quite well when serving small
5979 objects, but with big ones (for example large images or archives) or
5980 with A/V streaming, a graph generated from haproxy counters looks like
Willy Tarreaudef0d222016-11-08 22:03:00 +01005981 a hedgehog. With this option enabled counters get incremented frequently
5982 along the session, typically every 5 seconds, which is often enough to
5983 produce clean graphs. Recounting touches a hotpath directly so it is not
5984 not enabled by default, as it can cause a lot of wakeups for very large
5985 session counts and cause a small performance drop.
Willy Tarreau0ba27502007-12-24 16:55:16 +01005986
5987
Willy Tarreauc9bd0cc2009-05-10 11:57:02 +02005988option dontlog-normal
5989no option dontlog-normal
5990 Enable or disable logging of normal, successful connections
5991 May be used in sections : defaults | frontend | listen | backend
5992 yes | yes | yes | no
5993 Arguments : none
5994
5995 There are large sites dealing with several thousand connections per second
5996 and for which logging is a major pain. Some of them are even forced to turn
5997 logs off and cannot debug production issues. Setting this option ensures that
5998 normal connections, those which experience no error, no timeout, no retry nor
5999 redispatch, will not be logged. This leaves disk space for anomalies. In HTTP
6000 mode, the response status code is checked and return codes 5xx will still be
6001 logged.
6002
6003 It is strongly discouraged to use this option as most of the time, the key to
6004 complex issues is in the normal logs which will not be logged here. If you
6005 need to separate logs, see the "log-separate-errors" option instead.
6006
Willy Tarreauc57f0e22009-05-10 13:12:33 +02006007 See also : "log", "dontlognull", "log-separate-errors" and section 8 about
Willy Tarreauc9bd0cc2009-05-10 11:57:02 +02006008 logging.
6009
6010
Willy Tarreaubf1f8162007-12-28 17:42:56 +01006011option dontlognull
6012no option dontlognull
6013 Enable or disable logging of null connections
6014 May be used in sections : defaults | frontend | listen | backend
6015 yes | yes | yes | no
6016 Arguments : none
6017
6018 In certain environments, there are components which will regularly connect to
6019 various systems to ensure that they are still alive. It can be the case from
6020 another load balancer as well as from monitoring systems. By default, even a
6021 simple port probe or scan will produce a log. If those connections pollute
6022 the logs too much, it is possible to enable option "dontlognull" to indicate
6023 that a connection on which no data has been transferred will not be logged,
Willy Tarreau0f228a02015-05-01 15:37:53 +02006024 which typically corresponds to those probes. Note that errors will still be
6025 returned to the client and accounted for in the stats. If this is not what is
6026 desired, option http-ignore-probes can be used instead.
Willy Tarreaubf1f8162007-12-28 17:42:56 +01006027
6028 It is generally recommended not to use this option in uncontrolled
Davor Ocelice9ed2812017-12-25 17:49:28 +01006029 environments (e.g. internet), otherwise scans and other malicious activities
Willy Tarreaubf1f8162007-12-28 17:42:56 +01006030 would not be logged.
6031
6032 If this option has been enabled in a "defaults" section, it can be disabled
6033 in a specific instance by prepending the "no" keyword before it.
6034
Willy Tarreau0f228a02015-05-01 15:37:53 +02006035 See also : "log", "http-ignore-probes", "monitor-net", "monitor-uri", and
6036 section 8 about logging.
Willy Tarreaubf1f8162007-12-28 17:42:56 +01006037
Willy Tarreaubf1f8162007-12-28 17:42:56 +01006038
Willy Tarreau87cf5142011-08-19 22:57:24 +02006039option forwardfor [ except <network> ] [ header <name> ] [ if-none ]
Willy Tarreauc27debf2008-01-06 08:57:02 +01006040 Enable insertion of the X-Forwarded-For header to requests sent to servers
6041 May be used in sections : defaults | frontend | listen | backend
6042 yes | yes | yes | yes
6043 Arguments :
6044 <network> is an optional argument used to disable this option for sources
6045 matching <network>
Ross Westaf72a1d2008-08-03 10:51:45 +02006046 <name> an optional argument to specify a different "X-Forwarded-For"
Willy Tarreaud72758d2010-01-12 10:42:19 +01006047 header name.
Willy Tarreauc27debf2008-01-06 08:57:02 +01006048
6049 Since HAProxy works in reverse-proxy mode, the servers see its IP address as
6050 their client address. This is sometimes annoying when the client's IP address
6051 is expected in server logs. To solve this problem, the well-known HTTP header
6052 "X-Forwarded-For" may be added by HAProxy to all requests sent to the server.
6053 This header contains a value representing the client's IP address. Since this
6054 header is always appended at the end of the existing header list, the server
6055 must be configured to always use the last occurrence of this header only. See
Ross Westaf72a1d2008-08-03 10:51:45 +02006056 the server's manual to find how to enable use of this standard header. Note
6057 that only the last occurrence of the header must be used, since it is really
6058 possible that the client has already brought one.
6059
Willy Tarreaud72758d2010-01-12 10:42:19 +01006060 The keyword "header" may be used to supply a different header name to replace
Ross Westaf72a1d2008-08-03 10:51:45 +02006061 the default "X-Forwarded-For". This can be useful where you might already
Davor Ocelice9ed2812017-12-25 17:49:28 +01006062 have a "X-Forwarded-For" header from a different application (e.g. stunnel),
Willy Tarreaud72758d2010-01-12 10:42:19 +01006063 and you need preserve it. Also if your backend server doesn't use the
Davor Ocelice9ed2812017-12-25 17:49:28 +01006064 "X-Forwarded-For" header and requires different one (e.g. Zeus Web Servers
Ross Westaf72a1d2008-08-03 10:51:45 +02006065 require "X-Cluster-Client-IP").
Willy Tarreauc27debf2008-01-06 08:57:02 +01006066
6067 Sometimes, a same HAProxy instance may be shared between a direct client
6068 access and a reverse-proxy access (for instance when an SSL reverse-proxy is
6069 used to decrypt HTTPS traffic). It is possible to disable the addition of the
6070 header for a known source address or network by adding the "except" keyword
6071 followed by the network address. In this case, any source IP matching the
6072 network will not cause an addition of this header. Most common uses are with
6073 private networks or 127.0.0.1.
6074
Willy Tarreau87cf5142011-08-19 22:57:24 +02006075 Alternatively, the keyword "if-none" states that the header will only be
6076 added if it is not present. This should only be used in perfectly trusted
6077 environment, as this might cause a security issue if headers reaching haproxy
6078 are under the control of the end-user.
6079
Willy Tarreauc27debf2008-01-06 08:57:02 +01006080 This option may be specified either in the frontend or in the backend. If at
Ross Westaf72a1d2008-08-03 10:51:45 +02006081 least one of them uses it, the header will be added. Note that the backend's
6082 setting of the header subargument takes precedence over the frontend's if
Willy Tarreau87cf5142011-08-19 22:57:24 +02006083 both are defined. In the case of the "if-none" argument, if at least one of
6084 the frontend or the backend does not specify it, it wants the addition to be
6085 mandatory, so it wins.
Willy Tarreauc27debf2008-01-06 08:57:02 +01006086
Olivier Doucetaa1ea8a2016-08-05 17:15:20 +02006087 Example :
Willy Tarreauc27debf2008-01-06 08:57:02 +01006088 # Public HTTP address also used by stunnel on the same machine
6089 frontend www
6090 mode http
6091 option forwardfor except 127.0.0.1 # stunnel already adds the header
6092
Ross Westaf72a1d2008-08-03 10:51:45 +02006093 # Those servers want the IP Address in X-Client
6094 backend www
6095 mode http
6096 option forwardfor header X-Client
6097
Willy Tarreau87cf5142011-08-19 22:57:24 +02006098 See also : "option httpclose", "option http-server-close",
Christopher Faulet315b39c2018-09-21 16:26:19 +02006099 "option http-keep-alive"
Willy Tarreauc27debf2008-01-06 08:57:02 +01006100
Willy Tarreau8a8e1d92010-04-05 16:15:16 +02006101
Willy Tarreau9fbe18e2015-05-01 22:42:08 +02006102option http-buffer-request
6103no option http-buffer-request
6104 Enable or disable waiting for whole HTTP request body before proceeding
6105 May be used in sections : defaults | frontend | listen | backend
6106 yes | yes | yes | yes
6107 Arguments : none
6108
6109 It is sometimes desirable to wait for the body of an HTTP request before
6110 taking a decision. This is what is being done by "balance url_param" for
6111 example. The first use case is to buffer requests from slow clients before
6112 connecting to the server. Another use case consists in taking the routing
6113 decision based on the request body's contents. This option placed in a
6114 frontend or backend forces the HTTP processing to wait until either the whole
6115 body is received, or the request buffer is full, or the first chunk is
6116 complete in case of chunked encoding. It can have undesired side effects with
Tim Düsterhus4896c442016-11-29 02:15:19 +01006117 some applications abusing HTTP by expecting unbuffered transmissions between
Willy Tarreau9fbe18e2015-05-01 22:42:08 +02006118 the frontend and the backend, so this should definitely not be used by
6119 default.
6120
Baptiste Assmanneccdf432015-10-28 13:49:01 +01006121 See also : "option http-no-delay", "timeout http-request"
Willy Tarreau9fbe18e2015-05-01 22:42:08 +02006122
6123
Willy Tarreau0f228a02015-05-01 15:37:53 +02006124option http-ignore-probes
6125no option http-ignore-probes
6126 Enable or disable logging of null connections and request timeouts
6127 May be used in sections : defaults | frontend | listen | backend
6128 yes | yes | yes | no
6129 Arguments : none
6130
6131 Recently some browsers started to implement a "pre-connect" feature
6132 consisting in speculatively connecting to some recently visited web sites
6133 just in case the user would like to visit them. This results in many
6134 connections being established to web sites, which end up in 408 Request
6135 Timeout if the timeout strikes first, or 400 Bad Request when the browser
6136 decides to close them first. These ones pollute the log and feed the error
6137 counters. There was already "option dontlognull" but it's insufficient in
6138 this case. Instead, this option does the following things :
6139 - prevent any 400/408 message from being sent to the client if nothing
Davor Ocelice9ed2812017-12-25 17:49:28 +01006140 was received over a connection before it was closed;
6141 - prevent any log from being emitted in this situation;
Willy Tarreau0f228a02015-05-01 15:37:53 +02006142 - prevent any error counter from being incremented
6143
6144 That way the empty connection is silently ignored. Note that it is better
6145 not to use this unless it is clear that it is needed, because it will hide
6146 real problems. The most common reason for not receiving a request and seeing
6147 a 408 is due to an MTU inconsistency between the client and an intermediary
6148 element such as a VPN, which blocks too large packets. These issues are
6149 generally seen with POST requests as well as GET with large cookies. The logs
6150 are often the only way to detect them.
6151
6152 If this option has been enabled in a "defaults" section, it can be disabled
6153 in a specific instance by prepending the "no" keyword before it.
6154
6155 See also : "log", "dontlognull", "errorfile", and section 8 about logging.
6156
6157
Willy Tarreau16bfb022010-01-16 19:48:41 +01006158option http-keep-alive
6159no option http-keep-alive
6160 Enable or disable HTTP keep-alive from client to server
6161 May be used in sections : defaults | frontend | listen | backend
6162 yes | yes | yes | yes
6163 Arguments : none
6164
Willy Tarreau70dffda2014-01-30 03:07:23 +01006165 By default HAProxy operates in keep-alive mode with regards to persistent
6166 connections: for each connection it processes each request and response, and
Christopher Faulet315b39c2018-09-21 16:26:19 +02006167 leaves the connection idle on both sides between the end of a response and
6168 the start of a new request. This mode may be changed by several options such
6169 as "option http-server-close", "option httpclose" or "option http-tunnel".
6170 This option allows to set back the keep-alive mode, which can be useful when
6171 another mode was used in a defaults section.
Willy Tarreau70dffda2014-01-30 03:07:23 +01006172
6173 Setting "option http-keep-alive" enables HTTP keep-alive mode on the client-
6174 and server- sides. This provides the lowest latency on the client side (slow
Willy Tarreau16bfb022010-01-16 19:48:41 +01006175 network) and the fastest session reuse on the server side at the expense
6176 of maintaining idle connections to the servers. In general, it is possible
6177 with this option to achieve approximately twice the request rate that the
6178 "http-server-close" option achieves on small objects. There are mainly two
6179 situations where this option may be useful :
6180
6181 - when the server is non-HTTP compliant and authenticates the connection
Davor Ocelice9ed2812017-12-25 17:49:28 +01006182 instead of requests (e.g. NTLM authentication)
Willy Tarreau16bfb022010-01-16 19:48:41 +01006183
6184 - when the cost of establishing the connection to the server is significant
6185 compared to the cost of retrieving the associated object from the server.
6186
6187 This last case can happen when the server is a fast static server of cache.
6188 In this case, the server will need to be properly tuned to support high enough
6189 connection counts because connections will last until the client sends another
6190 request.
6191
6192 If the client request has to go to another backend or another server due to
6193 content switching or the load balancing algorithm, the idle connection will
Willy Tarreau9420b122013-12-15 18:58:25 +01006194 immediately be closed and a new one re-opened. Option "prefer-last-server" is
6195 available to try optimize server selection so that if the server currently
6196 attached to an idle connection is usable, it will be used.
Willy Tarreau16bfb022010-01-16 19:48:41 +01006197
Willy Tarreau16bfb022010-01-16 19:48:41 +01006198 At the moment, logs will not indicate whether requests came from the same
6199 session or not. The accept date reported in the logs corresponds to the end
6200 of the previous request, and the request time corresponds to the time spent
6201 waiting for a new request. The keep-alive request time is still bound to the
6202 timeout defined by "timeout http-keep-alive" or "timeout http-request" if
6203 not set.
6204
Cyril Bonté653dcd62014-02-20 00:13:15 +01006205 This option disables and replaces any previous "option httpclose", "option
Christopher Faulet315b39c2018-09-21 16:26:19 +02006206 http-server-close" or "option http-tunnel". When backend and frontend options
6207 differ, all of these 4 options have precedence over "option http-keep-alive".
Willy Tarreau16bfb022010-01-16 19:48:41 +01006208
Christopher Faulet315b39c2018-09-21 16:26:19 +02006209 See also : "option httpclose",, "option http-server-close",
Willy Tarreau9420b122013-12-15 18:58:25 +01006210 "option prefer-last-server", "option http-pretend-keepalive",
Frédéric Lécaille93d33162019-03-06 09:35:59 +01006211 and "1.1. The HTTP transaction model".
Willy Tarreau16bfb022010-01-16 19:48:41 +01006212
6213
Willy Tarreau96e31212011-05-30 18:10:30 +02006214option http-no-delay
6215no option http-no-delay
6216 Instruct the system to favor low interactive delays over performance in HTTP
6217 May be used in sections : defaults | frontend | listen | backend
6218 yes | yes | yes | yes
6219 Arguments : none
6220
6221 In HTTP, each payload is unidirectional and has no notion of interactivity.
6222 Any agent is expected to queue data somewhat for a reasonably low delay.
6223 There are some very rare server-to-server applications that abuse the HTTP
6224 protocol and expect the payload phase to be highly interactive, with many
6225 interleaved data chunks in both directions within a single request. This is
6226 absolutely not supported by the HTTP specification and will not work across
6227 most proxies or servers. When such applications attempt to do this through
6228 haproxy, it works but they will experience high delays due to the network
6229 optimizations which favor performance by instructing the system to wait for
6230 enough data to be available in order to only send full packets. Typical
6231 delays are around 200 ms per round trip. Note that this only happens with
6232 abnormal uses. Normal uses such as CONNECT requests nor WebSockets are not
6233 affected.
6234
6235 When "option http-no-delay" is present in either the frontend or the backend
6236 used by a connection, all such optimizations will be disabled in order to
6237 make the exchanges as fast as possible. Of course this offers no guarantee on
6238 the functionality, as it may break at any other place. But if it works via
6239 HAProxy, it will work as fast as possible. This option should never be used
6240 by default, and should never be used at all unless such a buggy application
6241 is discovered. The impact of using this option is an increase of bandwidth
6242 usage and CPU usage, which may significantly lower performance in high
6243 latency environments.
6244
Willy Tarreau9fbe18e2015-05-01 22:42:08 +02006245 See also : "option http-buffer-request"
6246
Willy Tarreau96e31212011-05-30 18:10:30 +02006247
Willy Tarreau8a8e1d92010-04-05 16:15:16 +02006248option http-pretend-keepalive
6249no option http-pretend-keepalive
6250 Define whether haproxy will announce keepalive to the server or not
6251 May be used in sections : defaults | frontend | listen | backend
Christopher Faulet98db9762018-09-21 10:25:19 +02006252 yes | no | yes | yes
Willy Tarreau8a8e1d92010-04-05 16:15:16 +02006253 Arguments : none
6254
Christopher Faulet315b39c2018-09-21 16:26:19 +02006255 When running with "option http-server-close" or "option httpclose", haproxy
Willy Tarreau8a8e1d92010-04-05 16:15:16 +02006256 adds a "Connection: close" header to the request forwarded to the server.
6257 Unfortunately, when some servers see this header, they automatically refrain
6258 from using the chunked encoding for responses of unknown length, while this
6259 is totally unrelated. The immediate effect is that this prevents haproxy from
6260 maintaining the client connection alive. A second effect is that a client or
6261 a cache could receive an incomplete response without being aware of it, and
6262 consider the response complete.
6263
6264 By setting "option http-pretend-keepalive", haproxy will make the server
6265 believe it will keep the connection alive. The server will then not fall back
6266 to the abnormal undesired above. When haproxy gets the whole response, it
6267 will close the connection with the server just as it would do with the
Christopher Faulet315b39c2018-09-21 16:26:19 +02006268 "option httpclose". That way the client gets a normal response and the
Willy Tarreau8a8e1d92010-04-05 16:15:16 +02006269 connection is correctly closed on the server side.
6270
6271 It is recommended not to enable this option by default, because most servers
6272 will more efficiently close the connection themselves after the last packet,
6273 and release its buffers slightly earlier. Also, the added packet on the
6274 network could slightly reduce the overall peak performance. However it is
6275 worth noting that when this option is enabled, haproxy will have slightly
6276 less work to do. So if haproxy is the bottleneck on the whole architecture,
6277 enabling this option might save a few CPU cycles.
6278
Christopher Faulet98db9762018-09-21 10:25:19 +02006279 This option may be set in backend and listen sections. Using it in a frontend
6280 section will be ignored and a warning will be reported during startup. It is
6281 a backend related option, so there is no real reason to set it on a
6282 frontend. This option may be combined with "option httpclose", which will
6283 cause keepalive to be announced to the server and close to be announced to
6284 the client. This practice is discouraged though.
Willy Tarreau8a8e1d92010-04-05 16:15:16 +02006285
6286 If this option has been enabled in a "defaults" section, it can be disabled
6287 in a specific instance by prepending the "no" keyword before it.
6288
Christopher Faulet315b39c2018-09-21 16:26:19 +02006289 See also : "option httpclose", "option http-server-close", and
Willy Tarreau16bfb022010-01-16 19:48:41 +01006290 "option http-keep-alive"
Willy Tarreau8a8e1d92010-04-05 16:15:16 +02006291
Willy Tarreauc27debf2008-01-06 08:57:02 +01006292
Willy Tarreaub608feb2010-01-02 22:47:18 +01006293option http-server-close
6294no option http-server-close
6295 Enable or disable HTTP connection closing on the server side
6296 May be used in sections : defaults | frontend | listen | backend
6297 yes | yes | yes | yes
6298 Arguments : none
6299
Willy Tarreau70dffda2014-01-30 03:07:23 +01006300 By default HAProxy operates in keep-alive mode with regards to persistent
6301 connections: for each connection it processes each request and response, and
6302 leaves the connection idle on both sides between the end of a response and
6303 the start of a new request. This mode may be changed by several options such
Christopher Faulet315b39c2018-09-21 16:26:19 +02006304 as "option http-server-close", "option httpclose" or "option http-tunnel".
6305 Setting "option http-server-close" enables HTTP connection-close mode on the
6306 server side while keeping the ability to support HTTP keep-alive and
6307 pipelining on the client side. This provides the lowest latency on the client
6308 side (slow network) and the fastest session reuse on the server side to save
6309 server resources, similarly to "option httpclose". It also permits
6310 non-keepalive capable servers to be served in keep-alive mode to the clients
6311 if they conform to the requirements of RFC7230. Please note that some servers
6312 do not always conform to those requirements when they see "Connection: close"
6313 in the request. The effect will be that keep-alive will never be used. A
6314 workaround consists in enabling "option http-pretend-keepalive".
Willy Tarreaub608feb2010-01-02 22:47:18 +01006315
6316 At the moment, logs will not indicate whether requests came from the same
6317 session or not. The accept date reported in the logs corresponds to the end
6318 of the previous request, and the request time corresponds to the time spent
6319 waiting for a new request. The keep-alive request time is still bound to the
Willy Tarreaub16a5742010-01-10 14:46:16 +01006320 timeout defined by "timeout http-keep-alive" or "timeout http-request" if
6321 not set.
Willy Tarreaub608feb2010-01-02 22:47:18 +01006322
6323 This option may be set both in a frontend and in a backend. It is enabled if
6324 at least one of the frontend or backend holding a connection has it enabled.
Christopher Faulet315b39c2018-09-21 16:26:19 +02006325 It disables and replaces any previous "option httpclose", "option http-tunnel"
6326 or "option http-keep-alive". Please check section 4 ("Proxies") to see how
6327 this option combines with others when frontend and backend options differ.
Willy Tarreaub608feb2010-01-02 22:47:18 +01006328
6329 If this option has been enabled in a "defaults" section, it can be disabled
6330 in a specific instance by prepending the "no" keyword before it.
6331
Christopher Faulet315b39c2018-09-21 16:26:19 +02006332 See also : "option httpclose", "option http-pretend-keepalive",
6333 "option http-keep-alive", and "1.1. The HTTP transaction model".
Willy Tarreaub608feb2010-01-02 22:47:18 +01006334
6335
Christopher Faulet6c9bbb22019-03-26 21:37:23 +01006336option http-tunnel (deprecated)
6337no option http-tunnel (deprecated)
6338 Disable or enable HTTP connection processing after first transaction.
Willy Tarreau02bce8b2014-01-30 00:15:28 +01006339 May be used in sections : defaults | frontend | listen | backend
Christopher Faulet4212a302018-09-21 10:42:19 +02006340 yes | yes | yes | no
Willy Tarreau02bce8b2014-01-30 00:15:28 +01006341 Arguments : none
6342
Christopher Faulet6c9bbb22019-03-26 21:37:23 +01006343 Warning : Because it cannot work in HTTP/2, this option is deprecated and it
6344 is only supported on legacy HTTP frontends. In HTX, it is ignored and a
6345 warning is emitted during HAProxy startup.
6346
Willy Tarreau70dffda2014-01-30 03:07:23 +01006347 By default HAProxy operates in keep-alive mode with regards to persistent
6348 connections: for each connection it processes each request and response, and
6349 leaves the connection idle on both sides between the end of a response and
6350 the start of a new request. This mode may be changed by several options such
Christopher Faulet315b39c2018-09-21 16:26:19 +02006351 as "option http-server-close", "option httpclose" or "option http-tunnel".
Willy Tarreau70dffda2014-01-30 03:07:23 +01006352
6353 Option "http-tunnel" disables any HTTP processing past the first request and
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03006354 the first response. This is the mode which was used by default in versions
Willy Tarreau70dffda2014-01-30 03:07:23 +01006355 1.0 to 1.5-dev21. It is the mode with the lowest processing overhead, which
6356 is normally not needed anymore unless in very specific cases such as when
6357 using an in-house protocol that looks like HTTP but is not compatible, or
6358 just to log one request per client in order to reduce log size. Note that
6359 everything which works at the HTTP level, including header parsing/addition,
6360 cookie processing or content switching will only work for the first request
6361 and will be ignored after the first response.
Willy Tarreau02bce8b2014-01-30 00:15:28 +01006362
Christopher Faulet4212a302018-09-21 10:42:19 +02006363 This option may be set on frontend and listen sections. Using it on a backend
6364 section will be ignored and a warning will be reported during the startup. It
6365 is a frontend related option, so there is no real reason to set it on a
6366 backend.
6367
Willy Tarreau02bce8b2014-01-30 00:15:28 +01006368 If this option has been enabled in a "defaults" section, it can be disabled
6369 in a specific instance by prepending the "no" keyword before it.
6370
Christopher Faulet315b39c2018-09-21 16:26:19 +02006371 See also : "option httpclose", "option http-server-close",
6372 "option http-keep-alive", and "1.1. The HTTP transaction model".
Willy Tarreau02bce8b2014-01-30 00:15:28 +01006373
6374
Willy Tarreau88d349d2010-01-25 12:15:43 +01006375option http-use-proxy-header
Cyril Bontéf0c60612010-02-06 14:44:47 +01006376no option http-use-proxy-header
Willy Tarreau88d349d2010-01-25 12:15:43 +01006377 Make use of non-standard Proxy-Connection header instead of Connection
6378 May be used in sections : defaults | frontend | listen | backend
6379 yes | yes | yes | no
6380 Arguments : none
6381
Lukas Tribus23953682017-04-28 13:24:30 +00006382 While RFC7230 explicitly states that HTTP/1.1 agents must use the
Willy Tarreau88d349d2010-01-25 12:15:43 +01006383 Connection header to indicate their wish of persistent or non-persistent
6384 connections, both browsers and proxies ignore this header for proxied
6385 connections and make use of the undocumented, non-standard Proxy-Connection
6386 header instead. The issue begins when trying to put a load balancer between
6387 browsers and such proxies, because there will be a difference between what
6388 haproxy understands and what the client and the proxy agree on.
6389
6390 By setting this option in a frontend, haproxy can automatically switch to use
6391 that non-standard header if it sees proxied requests. A proxied request is
Lukas Tribusf01a9cd2016-02-03 18:09:37 +01006392 defined here as one where the URI begins with neither a '/' nor a '*'. This
6393 is incompatible with the HTTP tunnel mode. Note that this option can only be
6394 specified in a frontend and will affect the request along its whole life.
Willy Tarreau88d349d2010-01-25 12:15:43 +01006395
Willy Tarreau844a7e72010-01-31 21:46:18 +01006396 Also, when this option is set, a request which requires authentication will
6397 automatically switch to use proxy authentication headers if it is itself a
6398 proxied request. That makes it possible to check or enforce authentication in
6399 front of an existing proxy.
6400
Willy Tarreau88d349d2010-01-25 12:15:43 +01006401 This option should normally never be used, except in front of a proxy.
6402
Christopher Faulet315b39c2018-09-21 16:26:19 +02006403 See also : "option httpclose", and "option http-server-close".
Willy Tarreau88d349d2010-01-25 12:15:43 +01006404
6405
Willy Tarreau68ad3a42018-10-22 11:49:15 +02006406option http-use-htx
6407no option http-use-htx
6408 Switch to the new HTX internal representation for HTTP protocol elements
6409 May be used in sections : defaults | frontend | listen | backend
6410 yes | yes | yes | yes
6411 Arguments : none
6412
Christopher Faulet1d2b5862019-04-12 16:10:51 +02006413 Historically, the HTTP protocol is processed as-is. Inserting, deleting, or
Willy Tarreau68ad3a42018-10-22 11:49:15 +02006414 modifying a header field requires to rewrite the affected part in the buffer
Christopher Faulet1d2b5862019-04-12 16:10:51 +02006415 and to move the buffer's tail accordingly. This mode is known as the legacy
6416 HTTP mode. Since this principle has deep roots in haproxy, the HTTP/2
6417 protocol is converted to HTTP/1.1 before being processed this way. It also
6418 results in the inability to establish HTTP/2 connections to servers because
6419 of the loss of HTTP/2 semantics in the HTTP/1 representation.
Willy Tarreau68ad3a42018-10-22 11:49:15 +02006420
6421 HTX is the name of a totally new native internal representation for the HTTP
6422 protocol, that is agnostic to the version and aims at preserving semantics
6423 all along the chain. It relies on a fast parsing, tokenizing and indexing of
6424 the protocol elements so that no more memory moves are necessary and that
Christopher Faulet1d2b5862019-04-12 16:10:51 +02006425 most elements are directly accessed. It supports using either HTTP/1 or
6426 HTTP/2 on any side regardless of the other side's version. It also supports
6427 upgrades from TCP to HTTP and implicit ones from HTTP/1 to HTTP/2 (matching
6428 the HTTP/2 preface).
Willy Tarreau68ad3a42018-10-22 11:49:15 +02006429
Christopher Faulet1d2b5862019-04-12 16:10:51 +02006430 This option indicates that HTX needs to be used. Since the version 2.0-dev3,
6431 the HTX is the default mode. To switch back on the legacy HTTP mode, the
6432 option must be explicitly disabled using the "no" prefix. For prior versions,
6433 the feature has incomplete functional coverage, so it is not enabled by
6434 default.
Willy Tarreau68ad3a42018-10-22 11:49:15 +02006435
6436 See also : "mode http"
6437
6438
Willy Tarreaud63335a2010-02-26 12:56:52 +01006439option httpchk
6440option httpchk <uri>
6441option httpchk <method> <uri>
6442option httpchk <method> <uri> <version>
6443 Enable HTTP protocol to check on the servers health
6444 May be used in sections : defaults | frontend | listen | backend
6445 yes | no | yes | yes
6446 Arguments :
6447 <method> is the optional HTTP method used with the requests. When not set,
6448 the "OPTIONS" method is used, as it generally requires low server
6449 processing and is easy to filter out from the logs. Any method
6450 may be used, though it is not recommended to invent non-standard
6451 ones.
6452
6453 <uri> is the URI referenced in the HTTP requests. It defaults to " / "
6454 which is accessible by default on almost any server, but may be
6455 changed to any other URI. Query strings are permitted.
6456
6457 <version> is the optional HTTP version string. It defaults to "HTTP/1.0"
6458 but some servers might behave incorrectly in HTTP 1.0, so turning
6459 it to HTTP/1.1 may sometimes help. Note that the Host field is
6460 mandatory in HTTP/1.1, and as a trick, it is possible to pass it
6461 after "\r\n" following the version string.
6462
6463 By default, server health checks only consist in trying to establish a TCP
6464 connection. When "option httpchk" is specified, a complete HTTP request is
6465 sent once the TCP connection is established, and responses 2xx and 3xx are
6466 considered valid, while all other ones indicate a server failure, including
6467 the lack of any response.
6468
6469 The port and interval are specified in the server configuration.
6470
6471 This option does not necessarily require an HTTP backend, it also works with
6472 plain TCP backends. This is particularly useful to check simple scripts bound
6473 to some dedicated ports using the inetd daemon.
6474
6475 Examples :
6476 # Relay HTTPS traffic to Apache instance and check service availability
6477 # using HTTP request "OPTIONS * HTTP/1.1" on port 80.
6478 backend https_relay
6479 mode tcp
6480 option httpchk OPTIONS * HTTP/1.1\r\nHost:\ www
6481 server apache1 192.168.1.1:443 check port 80
6482
Simon Hormanafc47ee2013-11-25 10:46:35 +09006483 See also : "option ssl-hello-chk", "option smtpchk", "option mysql-check",
6484 "option pgsql-check", "http-check" and the "check", "port" and
6485 "inter" server options.
Willy Tarreaud63335a2010-02-26 12:56:52 +01006486
6487
Willy Tarreauc27debf2008-01-06 08:57:02 +01006488option httpclose
6489no option httpclose
Christopher Faulet315b39c2018-09-21 16:26:19 +02006490 Enable or disable HTTP connection closing
Willy Tarreauc27debf2008-01-06 08:57:02 +01006491 May be used in sections : defaults | frontend | listen | backend
6492 yes | yes | yes | yes
6493 Arguments : none
6494
Willy Tarreau70dffda2014-01-30 03:07:23 +01006495 By default HAProxy operates in keep-alive mode with regards to persistent
6496 connections: for each connection it processes each request and response, and
6497 leaves the connection idle on both sides between the end of a response and
6498 the start of a new request. This mode may be changed by several options such
Christopher Faulet315b39c2018-09-21 16:26:19 +02006499 as "option http-server-close", "option httpclose" or "option http-tunnel".
Willy Tarreau70dffda2014-01-30 03:07:23 +01006500
Christopher Faulet315b39c2018-09-21 16:26:19 +02006501 If "option httpclose" is set, HAProxy will close connections with the server
6502 and the client as soon as the request and the response are received. It will
John Roeslerfb2fce12019-07-10 15:45:51 -05006503 also check if a "Connection: close" header is already set in each direction,
Christopher Faulet315b39c2018-09-21 16:26:19 +02006504 and will add one if missing. Any "Connection" header different from "close"
6505 will also be removed.
Willy Tarreauc27debf2008-01-06 08:57:02 +01006506
Christopher Faulet315b39c2018-09-21 16:26:19 +02006507 This option may also be combined with "option http-pretend-keepalive", which
6508 will disable sending of the "Connection: close" header, but will still cause
6509 the connection to be closed once the whole response is received.
Willy Tarreauc27debf2008-01-06 08:57:02 +01006510
6511 This option may be set both in a frontend and in a backend. It is enabled if
6512 at least one of the frontend or backend holding a connection has it enabled.
Cyril Bonté653dcd62014-02-20 00:13:15 +01006513 It disables and replaces any previous "option http-server-close",
Christopher Faulet315b39c2018-09-21 16:26:19 +02006514 "option http-keep-alive" or "option http-tunnel". Please check section 4
6515 ("Proxies") to see how this option combines with others when frontend and
6516 backend options differ.
Willy Tarreauc27debf2008-01-06 08:57:02 +01006517
6518 If this option has been enabled in a "defaults" section, it can be disabled
6519 in a specific instance by prepending the "no" keyword before it.
6520
Christopher Faulet315b39c2018-09-21 16:26:19 +02006521 See also : "option http-server-close" and "1.1. The HTTP transaction model".
Willy Tarreauc27debf2008-01-06 08:57:02 +01006522
6523
Emeric Brun3a058f32009-06-30 18:26:00 +02006524option httplog [ clf ]
Willy Tarreauc27debf2008-01-06 08:57:02 +01006525 Enable logging of HTTP request, session state and timers
6526 May be used in sections : defaults | frontend | listen | backend
Tim Duesterhus9ad9f352018-02-05 20:52:27 +01006527 yes | yes | yes | no
Emeric Brun3a058f32009-06-30 18:26:00 +02006528 Arguments :
6529 clf if the "clf" argument is added, then the output format will be
6530 the CLF format instead of HAProxy's default HTTP format. You can
6531 use this when you need to feed HAProxy's logs through a specific
Davor Ocelice9ed2812017-12-25 17:49:28 +01006532 log analyzer which only support the CLF format and which is not
Emeric Brun3a058f32009-06-30 18:26:00 +02006533 extensible.
Willy Tarreauc27debf2008-01-06 08:57:02 +01006534
6535 By default, the log output format is very poor, as it only contains the
6536 source and destination addresses, and the instance name. By specifying
6537 "option httplog", each log line turns into a much richer format including,
6538 but not limited to, the HTTP request, the connection timers, the session
6539 status, the connections numbers, the captured headers and cookies, the
6540 frontend, backend and server name, and of course the source address and
6541 ports.
6542
PiBa-NLbd556bf2014-12-11 21:31:54 +01006543 Specifying only "option httplog" will automatically clear the 'clf' mode
6544 if it was set by default.
Emeric Brun3a058f32009-06-30 18:26:00 +02006545
Guillaume de Lafond29f45602017-03-31 19:52:15 +02006546 "option httplog" overrides any previous "log-format" directive.
6547
Willy Tarreauc57f0e22009-05-10 13:12:33 +02006548 See also : section 8 about logging.
Willy Tarreauc27debf2008-01-06 08:57:02 +01006549
Willy Tarreau55165fe2009-05-10 12:02:55 +02006550
6551option http_proxy
6552no option http_proxy
6553 Enable or disable plain HTTP proxy mode
6554 May be used in sections : defaults | frontend | listen | backend
6555 yes | yes | yes | yes
6556 Arguments : none
6557
6558 It sometimes happens that people need a pure HTTP proxy which understands
6559 basic proxy requests without caching nor any fancy feature. In this case,
6560 it may be worth setting up an HAProxy instance with the "option http_proxy"
6561 set. In this mode, no server is declared, and the connection is forwarded to
6562 the IP address and port found in the URL after the "http://" scheme.
6563
6564 No host address resolution is performed, so this only works when pure IP
6565 addresses are passed. Since this option's usage perimeter is rather limited,
Lukas Tribusf01a9cd2016-02-03 18:09:37 +01006566 it will probably be used only by experts who know they need exactly it. This
6567 is incompatible with the HTTP tunnel mode.
Willy Tarreau55165fe2009-05-10 12:02:55 +02006568
6569 If this option has been enabled in a "defaults" section, it can be disabled
6570 in a specific instance by prepending the "no" keyword before it.
6571
6572 Example :
6573 # this backend understands HTTP proxy requests and forwards them directly.
6574 backend direct_forward
6575 option httpclose
6576 option http_proxy
6577
6578 See also : "option httpclose"
6579
Willy Tarreau211ad242009-10-03 21:45:07 +02006580
Jamie Gloudon801a0a32012-08-25 00:18:33 -04006581option independent-streams
6582no option independent-streams
6583 Enable or disable independent timeout processing for both directions
Willy Tarreauf27b5ea2009-10-03 22:01:18 +02006584 May be used in sections : defaults | frontend | listen | backend
6585 yes | yes | yes | yes
6586 Arguments : none
6587
6588 By default, when data is sent over a socket, both the write timeout and the
6589 read timeout for that socket are refreshed, because we consider that there is
6590 activity on that socket, and we have no other means of guessing if we should
6591 receive data or not.
6592
Davor Ocelice9ed2812017-12-25 17:49:28 +01006593 While this default behavior is desirable for almost all applications, there
Willy Tarreauf27b5ea2009-10-03 22:01:18 +02006594 exists a situation where it is desirable to disable it, and only refresh the
6595 read timeout if there are incoming data. This happens on sessions with large
6596 timeouts and low amounts of exchanged data such as telnet session. If the
6597 server suddenly disappears, the output data accumulates in the system's
6598 socket buffers, both timeouts are correctly refreshed, and there is no way
6599 to know the server does not receive them, so we don't timeout. However, when
6600 the underlying protocol always echoes sent data, it would be enough by itself
6601 to detect the issue using the read timeout. Note that this problem does not
6602 happen with more verbose protocols because data won't accumulate long in the
6603 socket buffers.
6604
6605 When this option is set on the frontend, it will disable read timeout updates
6606 on data sent to the client. There probably is little use of this case. When
6607 the option is set on the backend, it will disable read timeout updates on
6608 data sent to the server. Doing so will typically break large HTTP posts from
6609 slow lines, so use it with caution.
6610
Willy Tarreauce887fd2012-05-12 12:50:00 +02006611 See also : "timeout client", "timeout server" and "timeout tunnel"
Willy Tarreauf27b5ea2009-10-03 22:01:18 +02006612
6613
Gabor Lekenyb4c81e42010-09-29 18:17:05 +02006614option ldap-check
6615 Use LDAPv3 health checks for server testing
6616 May be used in sections : defaults | frontend | listen | backend
6617 yes | no | yes | yes
6618 Arguments : none
6619
6620 It is possible to test that the server correctly talks LDAPv3 instead of just
6621 testing that it accepts the TCP connection. When this option is set, an
6622 LDAPv3 anonymous simple bind message is sent to the server, and the response
6623 is analyzed to find an LDAPv3 bind response message.
6624
6625 The server is considered valid only when the LDAP response contains success
6626 resultCode (http://tools.ietf.org/html/rfc4511#section-4.1.9).
6627
6628 Logging of bind requests is server dependent see your documentation how to
6629 configure it.
6630
6631 Example :
6632 option ldap-check
6633
6634 See also : "option httpchk"
6635
6636
Simon Horman98637e52014-06-20 12:30:16 +09006637option external-check
6638 Use external processes for server health checks
6639 May be used in sections : defaults | frontend | listen | backend
6640 yes | no | yes | yes
6641
6642 It is possible to test the health of a server using an external command.
6643 This is achieved by running the executable set using "external-check
6644 command".
6645
6646 Requires the "external-check" global to be set.
6647
6648 See also : "external-check", "external-check command", "external-check path"
6649
6650
Willy Tarreau211ad242009-10-03 21:45:07 +02006651option log-health-checks
6652no option log-health-checks
Willy Tarreaubef1b322014-05-13 21:01:39 +02006653 Enable or disable logging of health checks status updates
Willy Tarreau211ad242009-10-03 21:45:07 +02006654 May be used in sections : defaults | frontend | listen | backend
6655 yes | no | yes | yes
6656 Arguments : none
6657
Willy Tarreaubef1b322014-05-13 21:01:39 +02006658 By default, failed health check are logged if server is UP and successful
6659 health checks are logged if server is DOWN, so the amount of additional
6660 information is limited.
Willy Tarreau211ad242009-10-03 21:45:07 +02006661
Willy Tarreaubef1b322014-05-13 21:01:39 +02006662 When this option is enabled, any change of the health check status or to
6663 the server's health will be logged, so that it becomes possible to know
6664 that a server was failing occasional checks before crashing, or exactly when
6665 it failed to respond a valid HTTP status, then when the port started to
6666 reject connections, then when the server stopped responding at all.
6667
Davor Ocelice9ed2812017-12-25 17:49:28 +01006668 Note that status changes not caused by health checks (e.g. enable/disable on
Willy Tarreaubef1b322014-05-13 21:01:39 +02006669 the CLI) are intentionally not logged by this option.
Willy Tarreau211ad242009-10-03 21:45:07 +02006670
Willy Tarreaubef1b322014-05-13 21:01:39 +02006671 See also: "option httpchk", "option ldap-check", "option mysql-check",
6672 "option pgsql-check", "option redis-check", "option smtpchk",
6673 "option tcp-check", "log" and section 8 about logging.
Willy Tarreau211ad242009-10-03 21:45:07 +02006674
Willy Tarreauc9bd0cc2009-05-10 11:57:02 +02006675
6676option log-separate-errors
6677no option log-separate-errors
6678 Change log level for non-completely successful connections
6679 May be used in sections : defaults | frontend | listen | backend
6680 yes | yes | yes | no
6681 Arguments : none
6682
6683 Sometimes looking for errors in logs is not easy. This option makes haproxy
6684 raise the level of logs containing potentially interesting information such
6685 as errors, timeouts, retries, redispatches, or HTTP status codes 5xx. The
6686 level changes from "info" to "err". This makes it possible to log them
6687 separately to a different file with most syslog daemons. Be careful not to
6688 remove them from the original file, otherwise you would lose ordering which
6689 provides very important information.
6690
6691 Using this option, large sites dealing with several thousand connections per
6692 second may log normal traffic to a rotating buffer and only archive smaller
6693 error logs.
6694
Willy Tarreauc57f0e22009-05-10 13:12:33 +02006695 See also : "log", "dontlognull", "dontlog-normal" and section 8 about
Willy Tarreauc9bd0cc2009-05-10 11:57:02 +02006696 logging.
6697
Willy Tarreauc27debf2008-01-06 08:57:02 +01006698
6699option logasap
6700no option logasap
6701 Enable or disable early logging of HTTP requests
6702 May be used in sections : defaults | frontend | listen | backend
6703 yes | yes | yes | no
6704 Arguments : none
6705
6706 By default, HTTP requests are logged upon termination so that the total
6707 transfer time and the number of bytes appear in the logs. When large objects
6708 are being transferred, it may take a while before the request appears in the
6709 logs. Using "option logasap", the request gets logged as soon as the server
6710 sends the complete headers. The only missing information in the logs will be
6711 the total number of bytes which will indicate everything except the amount
6712 of data transferred, and the total time which will not take the transfer
Willy Tarreaud2a4aa22008-01-31 15:28:22 +01006713 time into account. In such a situation, it's a good practice to capture the
Willy Tarreauc27debf2008-01-06 08:57:02 +01006714 "Content-Length" response header so that the logs at least indicate how many
6715 bytes are expected to be transferred.
6716
Willy Tarreaucc6c8912009-02-22 10:53:55 +01006717 Examples :
6718 listen http_proxy 0.0.0.0:80
6719 mode http
6720 option httplog
6721 option logasap
6722 log 192.168.2.200 local3
6723
6724 >>> Feb 6 12:14:14 localhost \
6725 haproxy[14389]: 10.0.1.2:33317 [06/Feb/2009:12:14:14.655] http-in \
6726 static/srv1 9/10/7/14/+30 200 +243 - - ---- 3/1/1/1/0 1/0 \
6727 "GET /image.iso HTTP/1.0"
6728
Willy Tarreauc57f0e22009-05-10 13:12:33 +02006729 See also : "option httplog", "capture response header", and section 8 about
Willy Tarreauc27debf2008-01-06 08:57:02 +01006730 logging.
6731
6732
Nenad Merdanovic6639a7c2014-05-30 14:26:32 +02006733option mysql-check [ user <username> [ post-41 ] ]
Hervé COMMOWICK8776f1b2010-10-18 15:58:36 +02006734 Use MySQL health checks for server testing
Hervé COMMOWICK698ae002010-01-12 09:25:13 +01006735 May be used in sections : defaults | frontend | listen | backend
6736 yes | no | yes | yes
Hervé COMMOWICK8776f1b2010-10-18 15:58:36 +02006737 Arguments :
Cyril Bonté108cf6e2012-04-21 23:30:29 +02006738 <username> This is the username which will be used when connecting to MySQL
6739 server.
Nenad Merdanovic6639a7c2014-05-30 14:26:32 +02006740 post-41 Send post v4.1 client compatible checks
Hervé COMMOWICK8776f1b2010-10-18 15:58:36 +02006741
6742 If you specify a username, the check consists of sending two MySQL packet,
6743 one Client Authentication packet, and one QUIT packet, to correctly close
Davor Ocelice9ed2812017-12-25 17:49:28 +01006744 MySQL session. We then parse the MySQL Handshake Initialization packet and/or
Hervé COMMOWICK8776f1b2010-10-18 15:58:36 +02006745 Error packet. It is a basic but useful test which does not produce error nor
6746 aborted connect on the server. However, it requires adding an authorization
6747 in the MySQL table, like this :
6748
6749 USE mysql;
6750 INSERT INTO user (Host,User) values ('<ip_of_haproxy>','<username>');
6751 FLUSH PRIVILEGES;
6752
6753 If you don't specify a username (it is deprecated and not recommended), the
Davor Ocelice9ed2812017-12-25 17:49:28 +01006754 check only consists in parsing the Mysql Handshake Initialization packet or
Hervé COMMOWICK8776f1b2010-10-18 15:58:36 +02006755 Error packet, we don't send anything in this mode. It was reported that it
6756 can generate lockout if check is too frequent and/or if there is not enough
6757 traffic. In fact, you need in this case to check MySQL "max_connect_errors"
6758 value as if a connection is established successfully within fewer than MySQL
6759 "max_connect_errors" attempts after a previous connection was interrupted,
6760 the error count for the host is cleared to zero. If HAProxy's server get
6761 blocked, the "FLUSH HOSTS" statement is the only way to unblock it.
6762
6763 Remember that this does not check database presence nor database consistency.
6764 To do this, you can use an external check with xinetd for example.
Hervé COMMOWICK698ae002010-01-12 09:25:13 +01006765
Hervé COMMOWICK212f7782011-06-10 14:05:59 +02006766 The check requires MySQL >=3.22, for older version, please use TCP check.
Hervé COMMOWICK698ae002010-01-12 09:25:13 +01006767
6768 Most often, an incoming MySQL server needs to see the client's IP address for
6769 various purposes, including IP privilege matching and connection logging.
6770 When possible, it is often wise to masquerade the client's IP address when
6771 connecting to the server using the "usesrc" argument of the "source" keyword,
Willy Tarreau29fbe512015-08-20 19:35:14 +02006772 which requires the transparent proxy feature to be compiled in, and the MySQL
6773 server to route the client via the machine hosting haproxy.
Hervé COMMOWICK698ae002010-01-12 09:25:13 +01006774
6775 See also: "option httpchk"
6776
6777
Willy Tarreaua453bdd2008-01-08 19:50:52 +01006778option nolinger
6779no option nolinger
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01006780 Enable or disable immediate session resource cleaning after close
Willy Tarreaua453bdd2008-01-08 19:50:52 +01006781 May be used in sections: defaults | frontend | listen | backend
6782 yes | yes | yes | yes
Willy Tarreaueabeafa2008-01-16 16:17:06 +01006783 Arguments : none
Willy Tarreaua453bdd2008-01-08 19:50:52 +01006784
Davor Ocelice9ed2812017-12-25 17:49:28 +01006785 When clients or servers abort connections in a dirty way (e.g. they are
Willy Tarreaua453bdd2008-01-08 19:50:52 +01006786 physically disconnected), the session timeouts triggers and the session is
6787 closed. But it will remain in FIN_WAIT1 state for some time in the system,
6788 using some resources and possibly limiting the ability to establish newer
6789 connections.
6790
6791 When this happens, it is possible to activate "option nolinger" which forces
6792 the system to immediately remove any socket's pending data on close. Thus,
6793 the session is instantly purged from the system's tables. This usually has
6794 side effects such as increased number of TCP resets due to old retransmits
6795 getting immediately rejected. Some firewalls may sometimes complain about
6796 this too.
6797
6798 For this reason, it is not recommended to use this option when not absolutely
6799 needed. You know that you need it when you have thousands of FIN_WAIT1
6800 sessions on your system (TIME_WAIT ones do not count).
6801
6802 This option may be used both on frontends and backends, depending on the side
6803 where it is required. Use it on the frontend for clients, and on the backend
6804 for servers.
6805
6806 If this option has been enabled in a "defaults" section, it can be disabled
6807 in a specific instance by prepending the "no" keyword before it.
6808
6809
Willy Tarreau55165fe2009-05-10 12:02:55 +02006810option originalto [ except <network> ] [ header <name> ]
6811 Enable insertion of the X-Original-To header to requests sent to servers
6812 May be used in sections : defaults | frontend | listen | backend
6813 yes | yes | yes | yes
6814 Arguments :
6815 <network> is an optional argument used to disable this option for sources
6816 matching <network>
6817 <name> an optional argument to specify a different "X-Original-To"
6818 header name.
6819
6820 Since HAProxy can work in transparent mode, every request from a client can
6821 be redirected to the proxy and HAProxy itself can proxy every request to a
6822 complex SQUID environment and the destination host from SO_ORIGINAL_DST will
6823 be lost. This is annoying when you want access rules based on destination ip
6824 addresses. To solve this problem, a new HTTP header "X-Original-To" may be
6825 added by HAProxy to all requests sent to the server. This header contains a
6826 value representing the original destination IP address. Since this must be
6827 configured to always use the last occurrence of this header only. Note that
6828 only the last occurrence of the header must be used, since it is really
6829 possible that the client has already brought one.
6830
6831 The keyword "header" may be used to supply a different header name to replace
6832 the default "X-Original-To". This can be useful where you might already
6833 have a "X-Original-To" header from a different application, and you need
6834 preserve it. Also if your backend server doesn't use the "X-Original-To"
6835 header and requires different one.
6836
6837 Sometimes, a same HAProxy instance may be shared between a direct client
6838 access and a reverse-proxy access (for instance when an SSL reverse-proxy is
6839 used to decrypt HTTPS traffic). It is possible to disable the addition of the
6840 header for a known source address or network by adding the "except" keyword
6841 followed by the network address. In this case, any source IP matching the
6842 network will not cause an addition of this header. Most common uses are with
6843 private networks or 127.0.0.1.
6844
6845 This option may be specified either in the frontend or in the backend. If at
6846 least one of them uses it, the header will be added. Note that the backend's
6847 setting of the header subargument takes precedence over the frontend's if
6848 both are defined.
6849
Willy Tarreau55165fe2009-05-10 12:02:55 +02006850 Examples :
6851 # Original Destination address
6852 frontend www
6853 mode http
6854 option originalto except 127.0.0.1
6855
6856 # Those servers want the IP Address in X-Client-Dst
6857 backend www
6858 mode http
6859 option originalto header X-Client-Dst
6860
Christopher Faulet315b39c2018-09-21 16:26:19 +02006861 See also : "option httpclose", "option http-server-close".
Willy Tarreau55165fe2009-05-10 12:02:55 +02006862
6863
Willy Tarreaua453bdd2008-01-08 19:50:52 +01006864option persist
6865no option persist
6866 Enable or disable forced persistence on down servers
6867 May be used in sections: defaults | frontend | listen | backend
6868 yes | no | yes | yes
Willy Tarreaueabeafa2008-01-16 16:17:06 +01006869 Arguments : none
Willy Tarreaua453bdd2008-01-08 19:50:52 +01006870
6871 When an HTTP request reaches a backend with a cookie which references a dead
6872 server, by default it is redispatched to another server. It is possible to
6873 force the request to be sent to the dead server first using "option persist"
6874 if absolutely needed. A common use case is when servers are under extreme
6875 load and spend their time flapping. In this case, the users would still be
6876 directed to the server they opened the session on, in the hope they would be
6877 correctly served. It is recommended to use "option redispatch" in conjunction
6878 with this option so that in the event it would not be possible to connect to
6879 the server at all (server definitely dead), the client would finally be
6880 redirected to another valid server.
6881
6882 If this option has been enabled in a "defaults" section, it can be disabled
6883 in a specific instance by prepending the "no" keyword before it.
6884
Willy Tarreau4de91492010-01-22 19:10:05 +01006885 See also : "option redispatch", "retries", "force-persist"
Willy Tarreaua453bdd2008-01-08 19:50:52 +01006886
6887
Willy Tarreau0c122822013-12-15 18:49:01 +01006888option pgsql-check [ user <username> ]
6889 Use PostgreSQL health checks for server testing
6890 May be used in sections : defaults | frontend | listen | backend
6891 yes | no | yes | yes
6892 Arguments :
6893 <username> This is the username which will be used when connecting to
6894 PostgreSQL server.
6895
6896 The check sends a PostgreSQL StartupMessage and waits for either
6897 Authentication request or ErrorResponse message. It is a basic but useful
6898 test which does not produce error nor aborted connect on the server.
6899 This check is identical with the "mysql-check".
6900
6901 See also: "option httpchk"
6902
6903
Willy Tarreau9420b122013-12-15 18:58:25 +01006904option prefer-last-server
6905no option prefer-last-server
6906 Allow multiple load balanced requests to remain on the same server
6907 May be used in sections: defaults | frontend | listen | backend
6908 yes | no | yes | yes
6909 Arguments : none
6910
6911 When the load balancing algorithm in use is not deterministic, and a previous
6912 request was sent to a server to which haproxy still holds a connection, it is
6913 sometimes desirable that subsequent requests on a same session go to the same
6914 server as much as possible. Note that this is different from persistence, as
6915 we only indicate a preference which haproxy tries to apply without any form
6916 of warranty. The real use is for keep-alive connections sent to servers. When
6917 this option is used, haproxy will try to reuse the same connection that is
6918 attached to the server instead of rebalancing to another server, causing a
6919 close of the connection. This can make sense for static file servers. It does
Willy Tarreau068621e2013-12-23 15:11:25 +01006920 not make much sense to use this in combination with hashing algorithms. Note,
6921 haproxy already automatically tries to stick to a server which sends a 401 or
Lukas Tribus80512b12018-10-27 20:07:40 +02006922 to a proxy which sends a 407 (authentication required), when the load
6923 balancing algorithm is not deterministic. This is mandatory for use with the
6924 broken NTLM authentication challenge, and significantly helps in
Willy Tarreau068621e2013-12-23 15:11:25 +01006925 troubleshooting some faulty applications. Option prefer-last-server might be
6926 desirable in these environments as well, to avoid redistributing the traffic
6927 after every other response.
Willy Tarreau9420b122013-12-15 18:58:25 +01006928
6929 If this option has been enabled in a "defaults" section, it can be disabled
6930 in a specific instance by prepending the "no" keyword before it.
6931
6932 See also: "option http-keep-alive"
6933
6934
Krzysztof Piotr Oledzki25b501a2008-01-06 16:36:16 +01006935option redispatch
Joseph Lynch726ab712015-05-11 23:25:34 -07006936option redispatch <interval>
Krzysztof Piotr Oledzki25b501a2008-01-06 16:36:16 +01006937no option redispatch
6938 Enable or disable session redistribution in case of connection failure
6939 May be used in sections: defaults | frontend | listen | backend
6940 yes | no | yes | yes
Joseph Lynch726ab712015-05-11 23:25:34 -07006941 Arguments :
6942 <interval> The optional integer value that controls how often redispatches
6943 occur when retrying connections. Positive value P indicates a
6944 redispatch is desired on every Pth retry, and negative value
Davor Ocelice9ed2812017-12-25 17:49:28 +01006945 N indicate a redispatch is desired on the Nth retry prior to the
Joseph Lynch726ab712015-05-11 23:25:34 -07006946 last retry. For example, the default of -1 preserves the
Davor Ocelice9ed2812017-12-25 17:49:28 +01006947 historical behavior of redispatching on the last retry, a
Joseph Lynch726ab712015-05-11 23:25:34 -07006948 positive value of 1 would indicate a redispatch on every retry,
6949 and a positive value of 3 would indicate a redispatch on every
6950 third retry. You can disable redispatches with a value of 0.
6951
Krzysztof Piotr Oledzki25b501a2008-01-06 16:36:16 +01006952
6953 In HTTP mode, if a server designated by a cookie is down, clients may
6954 definitely stick to it because they cannot flush the cookie, so they will not
6955 be able to access the service anymore.
6956
Willy Tarreau59884a62019-01-02 14:48:31 +01006957 Specifying "option redispatch" will allow the proxy to break cookie or
6958 consistent hash based persistence and redistribute them to a working server.
Krzysztof Piotr Oledzki25b501a2008-01-06 16:36:16 +01006959
Joseph Lynch726ab712015-05-11 23:25:34 -07006960 It also allows to retry connections to another server in case of multiple
Krzysztof Piotr Oledzki25b501a2008-01-06 16:36:16 +01006961 connection failures. Of course, it requires having "retries" set to a nonzero
6962 value.
Willy Tarreaud72758d2010-01-12 10:42:19 +01006963
Krzysztof Piotr Oledzki25b501a2008-01-06 16:36:16 +01006964 This form is the preferred form, which replaces both the "redispatch" and
6965 "redisp" keywords.
6966
6967 If this option has been enabled in a "defaults" section, it can be disabled
6968 in a specific instance by prepending the "no" keyword before it.
6969
Willy Tarreau4de91492010-01-22 19:10:05 +01006970 See also : "redispatch", "retries", "force-persist"
Krzysztof Piotr Oledzki25b501a2008-01-06 16:36:16 +01006971
Willy Tarreaua453bdd2008-01-08 19:50:52 +01006972
Hervé COMMOWICKec032d62011-08-05 16:23:48 +02006973option redis-check
6974 Use redis health checks for server testing
6975 May be used in sections : defaults | frontend | listen | backend
6976 yes | no | yes | yes
6977 Arguments : none
6978
6979 It is possible to test that the server correctly talks REDIS protocol instead
6980 of just testing that it accepts the TCP connection. When this option is set,
6981 a PING redis command is sent to the server, and the response is analyzed to
6982 find the "+PONG" response message.
6983
6984 Example :
6985 option redis-check
6986
Jarno Huuskonene5ae7022017-04-03 14:36:21 +03006987 See also : "option httpchk", "option tcp-check", "tcp-check expect"
Hervé COMMOWICKec032d62011-08-05 16:23:48 +02006988
6989
Willy Tarreaua453bdd2008-01-08 19:50:52 +01006990option smtpchk
6991option smtpchk <hello> <domain>
6992 Use SMTP health checks for server testing
6993 May be used in sections : defaults | frontend | listen | backend
6994 yes | no | yes | yes
Willy Tarreaud72758d2010-01-12 10:42:19 +01006995 Arguments :
Willy Tarreaua453bdd2008-01-08 19:50:52 +01006996 <hello> is an optional argument. It is the "hello" command to use. It can
Lukas Tribus27935782018-10-01 02:00:16 +02006997 be either "HELO" (for SMTP) or "EHLO" (for ESMTP). All other
Willy Tarreaua453bdd2008-01-08 19:50:52 +01006998 values will be turned into the default command ("HELO").
6999
7000 <domain> is the domain name to present to the server. It may only be
7001 specified (and is mandatory) if the hello command has been
7002 specified. By default, "localhost" is used.
7003
7004 When "option smtpchk" is set, the health checks will consist in TCP
7005 connections followed by an SMTP command. By default, this command is
7006 "HELO localhost". The server's return code is analyzed and only return codes
7007 starting with a "2" will be considered as valid. All other responses,
7008 including a lack of response will constitute an error and will indicate a
7009 dead server.
7010
7011 This test is meant to be used with SMTP servers or relays. Depending on the
7012 request, it is possible that some servers do not log each connection attempt,
Davor Ocelice9ed2812017-12-25 17:49:28 +01007013 so you may want to experiment to improve the behavior. Using telnet on port
Willy Tarreaua453bdd2008-01-08 19:50:52 +01007014 25 is often easier than adjusting the configuration.
7015
7016 Most often, an incoming SMTP server needs to see the client's IP address for
7017 various purposes, including spam filtering, anti-spoofing and logging. When
7018 possible, it is often wise to masquerade the client's IP address when
7019 connecting to the server using the "usesrc" argument of the "source" keyword,
Willy Tarreau29fbe512015-08-20 19:35:14 +02007020 which requires the transparent proxy feature to be compiled in.
Willy Tarreaua453bdd2008-01-08 19:50:52 +01007021
7022 Example :
7023 option smtpchk HELO mydomain.org
7024
7025 See also : "option httpchk", "source"
7026
Krzysztof Piotr Oledzki25b501a2008-01-06 16:36:16 +01007027
Krzysztof Piotr Oledzkiaeebf9b2009-10-04 15:43:17 +02007028option socket-stats
7029no option socket-stats
7030
7031 Enable or disable collecting & providing separate statistics for each socket.
7032 May be used in sections : defaults | frontend | listen | backend
7033 yes | yes | yes | no
7034
7035 Arguments : none
7036
7037
Willy Tarreauff4f82d2009-02-06 11:28:13 +01007038option splice-auto
7039no option splice-auto
7040 Enable or disable automatic kernel acceleration on sockets in both directions
7041 May be used in sections : defaults | frontend | listen | backend
7042 yes | yes | yes | yes
7043 Arguments : none
7044
7045 When this option is enabled either on a frontend or on a backend, haproxy
7046 will automatically evaluate the opportunity to use kernel tcp splicing to
Davor Ocelice9ed2812017-12-25 17:49:28 +01007047 forward data between the client and the server, in either direction. HAProxy
Willy Tarreauff4f82d2009-02-06 11:28:13 +01007048 uses heuristics to estimate if kernel splicing might improve performance or
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01007049 not. Both directions are handled independently. Note that the heuristics used
Willy Tarreauff4f82d2009-02-06 11:28:13 +01007050 are not much aggressive in order to limit excessive use of splicing. This
7051 option requires splicing to be enabled at compile time, and may be globally
7052 disabled with the global option "nosplice". Since splice uses pipes, using it
7053 requires that there are enough spare pipes.
7054
7055 Important note: kernel-based TCP splicing is a Linux-specific feature which
7056 first appeared in kernel 2.6.25. It offers kernel-based acceleration to
7057 transfer data between sockets without copying these data to user-space, thus
7058 providing noticeable performance gains and CPU cycles savings. Since many
7059 early implementations are buggy, corrupt data and/or are inefficient, this
7060 feature is not enabled by default, and it should be used with extreme care.
7061 While it is not possible to detect the correctness of an implementation,
7062 2.6.29 is the first version offering a properly working implementation. In
7063 case of doubt, splicing may be globally disabled using the global "nosplice"
7064 keyword.
7065
7066 Example :
7067 option splice-auto
7068
7069 If this option has been enabled in a "defaults" section, it can be disabled
7070 in a specific instance by prepending the "no" keyword before it.
7071
7072 See also : "option splice-request", "option splice-response", and global
7073 options "nosplice" and "maxpipes"
7074
7075
7076option splice-request
7077no option splice-request
7078 Enable or disable automatic kernel acceleration on sockets for requests
7079 May be used in sections : defaults | frontend | listen | backend
7080 yes | yes | yes | yes
7081 Arguments : none
7082
7083 When this option is enabled either on a frontend or on a backend, haproxy
Jamie Gloudonaaa21002012-08-25 00:18:33 -04007084 will use kernel tcp splicing whenever possible to forward data going from
Willy Tarreauff4f82d2009-02-06 11:28:13 +01007085 the client to the server. It might still use the recv/send scheme if there
7086 are no spare pipes left. This option requires splicing to be enabled at
7087 compile time, and may be globally disabled with the global option "nosplice".
7088 Since splice uses pipes, using it requires that there are enough spare pipes.
7089
7090 Important note: see "option splice-auto" for usage limitations.
7091
7092 Example :
7093 option splice-request
7094
7095 If this option has been enabled in a "defaults" section, it can be disabled
7096 in a specific instance by prepending the "no" keyword before it.
7097
7098 See also : "option splice-auto", "option splice-response", and global options
7099 "nosplice" and "maxpipes"
7100
7101
7102option splice-response
7103no option splice-response
7104 Enable or disable automatic kernel acceleration on sockets for responses
7105 May be used in sections : defaults | frontend | listen | backend
7106 yes | yes | yes | yes
7107 Arguments : none
7108
7109 When this option is enabled either on a frontend or on a backend, haproxy
Jamie Gloudonaaa21002012-08-25 00:18:33 -04007110 will use kernel tcp splicing whenever possible to forward data going from
Willy Tarreauff4f82d2009-02-06 11:28:13 +01007111 the server to the client. It might still use the recv/send scheme if there
7112 are no spare pipes left. This option requires splicing to be enabled at
7113 compile time, and may be globally disabled with the global option "nosplice".
7114 Since splice uses pipes, using it requires that there are enough spare pipes.
7115
7116 Important note: see "option splice-auto" for usage limitations.
7117
7118 Example :
7119 option splice-response
7120
7121 If this option has been enabled in a "defaults" section, it can be disabled
7122 in a specific instance by prepending the "no" keyword before it.
7123
7124 See also : "option splice-auto", "option splice-request", and global options
7125 "nosplice" and "maxpipes"
7126
7127
Christopher Fauletba7bc162016-11-07 21:07:38 +01007128option spop-check
7129 Use SPOP health checks for server testing
7130 May be used in sections : defaults | frontend | listen | backend
7131 no | no | no | yes
7132 Arguments : none
7133
7134 It is possible to test that the server correctly talks SPOP protocol instead
7135 of just testing that it accepts the TCP connection. When this option is set,
7136 a HELLO handshake is performed between HAProxy and the server, and the
7137 response is analyzed to check no error is reported.
7138
7139 Example :
7140 option spop-check
7141
7142 See also : "option httpchk"
7143
7144
Willy Tarreaubf1f8162007-12-28 17:42:56 +01007145option srvtcpka
7146no option srvtcpka
7147 Enable or disable the sending of TCP keepalive packets on the server side
7148 May be used in sections : defaults | frontend | listen | backend
7149 yes | no | yes | yes
7150 Arguments : none
7151
7152 When there is a firewall or any session-aware component between a client and
7153 a server, and when the protocol involves very long sessions with long idle
Davor Ocelice9ed2812017-12-25 17:49:28 +01007154 periods (e.g. remote desktops), there is a risk that one of the intermediate
Willy Tarreaubf1f8162007-12-28 17:42:56 +01007155 components decides to expire a session which has remained idle for too long.
7156
7157 Enabling socket-level TCP keep-alives makes the system regularly send packets
7158 to the other end of the connection, leaving it active. The delay between
7159 keep-alive probes is controlled by the system only and depends both on the
7160 operating system and its tuning parameters.
7161
7162 It is important to understand that keep-alive packets are neither emitted nor
7163 received at the application level. It is only the network stacks which sees
7164 them. For this reason, even if one side of the proxy already uses keep-alives
7165 to maintain its connection alive, those keep-alive packets will not be
7166 forwarded to the other side of the proxy.
7167
7168 Please note that this has nothing to do with HTTP keep-alive.
7169
7170 Using option "srvtcpka" enables the emission of TCP keep-alive probes on the
7171 server side of a connection, which should help when session expirations are
7172 noticed between HAProxy and a server.
7173
7174 If this option has been enabled in a "defaults" section, it can be disabled
7175 in a specific instance by prepending the "no" keyword before it.
7176
7177 See also : "option clitcpka", "option tcpka"
7178
7179
Willy Tarreaua453bdd2008-01-08 19:50:52 +01007180option ssl-hello-chk
7181 Use SSLv3 client hello health checks for server testing
7182 May be used in sections : defaults | frontend | listen | backend
7183 yes | no | yes | yes
7184 Arguments : none
7185
7186 When some SSL-based protocols are relayed in TCP mode through HAProxy, it is
7187 possible to test that the server correctly talks SSL instead of just testing
7188 that it accepts the TCP connection. When "option ssl-hello-chk" is set, pure
7189 SSLv3 client hello messages are sent once the connection is established to
7190 the server, and the response is analyzed to find an SSL server hello message.
7191 The server is considered valid only when the response contains this server
7192 hello message.
7193
7194 All servers tested till there correctly reply to SSLv3 client hello messages,
7195 and most servers tested do not even log the requests containing only hello
7196 messages, which is appreciable.
7197
Willy Tarreau763a95b2012-10-04 23:15:39 +02007198 Note that this check works even when SSL support was not built into haproxy
7199 because it forges the SSL message. When SSL support is available, it is best
7200 to use native SSL health checks instead of this one.
Willy Tarreaua453bdd2008-01-08 19:50:52 +01007201
Willy Tarreau763a95b2012-10-04 23:15:39 +02007202 See also: "option httpchk", "check-ssl"
7203
Willy Tarreaua453bdd2008-01-08 19:50:52 +01007204
Willy Tarreaued179852013-12-16 01:07:00 +01007205option tcp-check
7206 Perform health checks using tcp-check send/expect sequences
7207 May be used in sections: defaults | frontend | listen | backend
7208 yes | no | yes | yes
7209
7210 This health check method is intended to be combined with "tcp-check" command
7211 lists in order to support send/expect types of health check sequences.
7212
7213 TCP checks currently support 4 modes of operations :
7214 - no "tcp-check" directive : the health check only consists in a connection
7215 attempt, which remains the default mode.
7216
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03007217 - "tcp-check send" or "tcp-check send-binary" only is mentioned : this is
Willy Tarreaued179852013-12-16 01:07:00 +01007218 used to send a string along with a connection opening. With some
7219 protocols, it helps sending a "QUIT" message for example that prevents
7220 the server from logging a connection error for each health check. The
7221 check result will still be based on the ability to open the connection
7222 only.
7223
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03007224 - "tcp-check expect" only is mentioned : this is used to test a banner.
Willy Tarreaued179852013-12-16 01:07:00 +01007225 The connection is opened and haproxy waits for the server to present some
7226 contents which must validate some rules. The check result will be based
7227 on the matching between the contents and the rules. This is suited for
7228 POP, IMAP, SMTP, FTP, SSH, TELNET.
7229
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03007230 - both "tcp-check send" and "tcp-check expect" are mentioned : this is
Davor Ocelice9ed2812017-12-25 17:49:28 +01007231 used to test a hello-type protocol. HAProxy sends a message, the server
7232 responds and its response is analyzed. the check result will be based on
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03007233 the matching between the response contents and the rules. This is often
Willy Tarreaued179852013-12-16 01:07:00 +01007234 suited for protocols which require a binding or a request/response model.
7235 LDAP, MySQL, Redis and SSL are example of such protocols, though they
7236 already all have their dedicated checks with a deeper understanding of
7237 the respective protocols.
7238 In this mode, many questions may be sent and many answers may be
Davor Ocelice9ed2812017-12-25 17:49:28 +01007239 analyzed.
Willy Tarreaued179852013-12-16 01:07:00 +01007240
Baptiste Assmannd60a9e52015-04-25 16:27:23 +02007241 A fifth mode can be used to insert comments in different steps of the
7242 script.
7243
7244 For each tcp-check rule you create, you can add a "comment" directive,
7245 followed by a string. This string will be reported in the log and stderr
7246 in debug mode. It is useful to make user-friendly error reporting.
7247 The "comment" is of course optional.
7248
7249
Willy Tarreaued179852013-12-16 01:07:00 +01007250 Examples :
Davor Ocelice9ed2812017-12-25 17:49:28 +01007251 # perform a POP check (analyze only server's banner)
Willy Tarreaued179852013-12-16 01:07:00 +01007252 option tcp-check
Baptiste Assmannd60a9e52015-04-25 16:27:23 +02007253 tcp-check expect string +OK\ POP3\ ready comment POP\ protocol
Willy Tarreaued179852013-12-16 01:07:00 +01007254
Davor Ocelice9ed2812017-12-25 17:49:28 +01007255 # perform an IMAP check (analyze only server's banner)
Willy Tarreaued179852013-12-16 01:07:00 +01007256 option tcp-check
Baptiste Assmannd60a9e52015-04-25 16:27:23 +02007257 tcp-check expect string *\ OK\ IMAP4\ ready comment IMAP\ protocol
Willy Tarreaued179852013-12-16 01:07:00 +01007258
7259 # look for the redis master server after ensuring it speaks well
7260 # redis protocol, then it exits properly.
Davor Ocelice9ed2812017-12-25 17:49:28 +01007261 # (send a command then analyze the response 3 times)
Willy Tarreaued179852013-12-16 01:07:00 +01007262 option tcp-check
Baptiste Assmannd60a9e52015-04-25 16:27:23 +02007263 tcp-check comment PING\ phase
Willy Tarreaued179852013-12-16 01:07:00 +01007264 tcp-check send PING\r\n
Baptiste Assmanna3322992015-08-04 10:12:18 +02007265 tcp-check expect string +PONG
Baptiste Assmannd60a9e52015-04-25 16:27:23 +02007266 tcp-check comment role\ check
Willy Tarreaued179852013-12-16 01:07:00 +01007267 tcp-check send info\ replication\r\n
7268 tcp-check expect string role:master
Baptiste Assmannd60a9e52015-04-25 16:27:23 +02007269 tcp-check comment QUIT\ phase
Willy Tarreaued179852013-12-16 01:07:00 +01007270 tcp-check send QUIT\r\n
7271 tcp-check expect string +OK
7272
Davor Ocelice9ed2812017-12-25 17:49:28 +01007273 forge a HTTP request, then analyze the response
Willy Tarreaued179852013-12-16 01:07:00 +01007274 (send many headers before analyzing)
7275 option tcp-check
Baptiste Assmannd60a9e52015-04-25 16:27:23 +02007276 tcp-check comment forge\ and\ send\ HTTP\ request
Willy Tarreaued179852013-12-16 01:07:00 +01007277 tcp-check send HEAD\ /\ HTTP/1.1\r\n
7278 tcp-check send Host:\ www.mydomain.com\r\n
7279 tcp-check send User-Agent:\ HAProxy\ tcpcheck\r\n
7280 tcp-check send \r\n
Baptiste Assmannd60a9e52015-04-25 16:27:23 +02007281 tcp-check expect rstring HTTP/1\..\ (2..|3..) comment check\ HTTP\ response
Willy Tarreaued179852013-12-16 01:07:00 +01007282
7283
7284 See also : "tcp-check expect", "tcp-check send"
7285
7286
Willy Tarreau9ea05a72009-06-14 12:07:01 +02007287option tcp-smart-accept
7288no option tcp-smart-accept
7289 Enable or disable the saving of one ACK packet during the accept sequence
7290 May be used in sections : defaults | frontend | listen | backend
7291 yes | yes | yes | no
7292 Arguments : none
7293
7294 When an HTTP connection request comes in, the system acknowledges it on
7295 behalf of HAProxy, then the client immediately sends its request, and the
7296 system acknowledges it too while it is notifying HAProxy about the new
7297 connection. HAProxy then reads the request and responds. This means that we
7298 have one TCP ACK sent by the system for nothing, because the request could
7299 very well be acknowledged by HAProxy when it sends its response.
7300
7301 For this reason, in HTTP mode, HAProxy automatically asks the system to avoid
7302 sending this useless ACK on platforms which support it (currently at least
7303 Linux). It must not cause any problem, because the system will send it anyway
7304 after 40 ms if the response takes more time than expected to come.
7305
7306 During complex network debugging sessions, it may be desirable to disable
7307 this optimization because delayed ACKs can make troubleshooting more complex
7308 when trying to identify where packets are delayed. It is then possible to
Davor Ocelice9ed2812017-12-25 17:49:28 +01007309 fall back to normal behavior by specifying "no option tcp-smart-accept".
Willy Tarreau9ea05a72009-06-14 12:07:01 +02007310
7311 It is also possible to force it for non-HTTP proxies by simply specifying
7312 "option tcp-smart-accept". For instance, it can make sense with some services
7313 such as SMTP where the server speaks first.
7314
7315 It is recommended to avoid forcing this option in a defaults section. In case
7316 of doubt, consider setting it back to automatic values by prepending the
7317 "default" keyword before it, or disabling it using the "no" keyword.
7318
Willy Tarreaud88edf22009-06-14 15:48:17 +02007319 See also : "option tcp-smart-connect"
7320
7321
7322option tcp-smart-connect
7323no option tcp-smart-connect
7324 Enable or disable the saving of one ACK packet during the connect sequence
7325 May be used in sections : defaults | frontend | listen | backend
7326 yes | no | yes | yes
7327 Arguments : none
7328
7329 On certain systems (at least Linux), HAProxy can ask the kernel not to
7330 immediately send an empty ACK upon a connection request, but to directly
7331 send the buffer request instead. This saves one packet on the network and
7332 thus boosts performance. It can also be useful for some servers, because they
7333 immediately get the request along with the incoming connection.
7334
7335 This feature is enabled when "option tcp-smart-connect" is set in a backend.
7336 It is not enabled by default because it makes network troubleshooting more
7337 complex.
7338
7339 It only makes sense to enable it with protocols where the client speaks first
7340 such as HTTP. In other situations, if there is no data to send in place of
7341 the ACK, a normal ACK is sent.
7342
7343 If this option has been enabled in a "defaults" section, it can be disabled
7344 in a specific instance by prepending the "no" keyword before it.
7345
7346 See also : "option tcp-smart-accept"
7347
Willy Tarreau9ea05a72009-06-14 12:07:01 +02007348
Willy Tarreaubf1f8162007-12-28 17:42:56 +01007349option tcpka
7350 Enable or disable the sending of TCP keepalive packets on both sides
7351 May be used in sections : defaults | frontend | listen | backend
7352 yes | yes | yes | yes
7353 Arguments : none
7354
7355 When there is a firewall or any session-aware component between a client and
7356 a server, and when the protocol involves very long sessions with long idle
Davor Ocelice9ed2812017-12-25 17:49:28 +01007357 periods (e.g. remote desktops), there is a risk that one of the intermediate
Willy Tarreaubf1f8162007-12-28 17:42:56 +01007358 components decides to expire a session which has remained idle for too long.
7359
7360 Enabling socket-level TCP keep-alives makes the system regularly send packets
7361 to the other end of the connection, leaving it active. The delay between
7362 keep-alive probes is controlled by the system only and depends both on the
7363 operating system and its tuning parameters.
7364
7365 It is important to understand that keep-alive packets are neither emitted nor
7366 received at the application level. It is only the network stacks which sees
7367 them. For this reason, even if one side of the proxy already uses keep-alives
7368 to maintain its connection alive, those keep-alive packets will not be
7369 forwarded to the other side of the proxy.
7370
7371 Please note that this has nothing to do with HTTP keep-alive.
7372
7373 Using option "tcpka" enables the emission of TCP keep-alive probes on both
7374 the client and server sides of a connection. Note that this is meaningful
7375 only in "defaults" or "listen" sections. If this option is used in a
7376 frontend, only the client side will get keep-alives, and if this option is
7377 used in a backend, only the server side will get keep-alives. For this
7378 reason, it is strongly recommended to explicitly use "option clitcpka" and
7379 "option srvtcpka" when the configuration is split between frontends and
7380 backends.
7381
7382 See also : "option clitcpka", "option srvtcpka"
7383
Willy Tarreau844e3c52008-01-11 16:28:18 +01007384
7385option tcplog
7386 Enable advanced logging of TCP connections with session state and timers
7387 May be used in sections : defaults | frontend | listen | backend
Tim Duesterhus9ad9f352018-02-05 20:52:27 +01007388 yes | yes | yes | no
Willy Tarreau844e3c52008-01-11 16:28:18 +01007389 Arguments : none
7390
7391 By default, the log output format is very poor, as it only contains the
7392 source and destination addresses, and the instance name. By specifying
7393 "option tcplog", each log line turns into a much richer format including, but
7394 not limited to, the connection timers, the session status, the connections
7395 numbers, the frontend, backend and server name, and of course the source
7396 address and ports. This option is useful for pure TCP proxies in order to
7397 find which of the client or server disconnects or times out. For normal HTTP
7398 proxies, it's better to use "option httplog" which is even more complete.
7399
Guillaume de Lafond29f45602017-03-31 19:52:15 +02007400 "option tcplog" overrides any previous "log-format" directive.
7401
Willy Tarreauc57f0e22009-05-10 13:12:33 +02007402 See also : "option httplog", and section 8 about logging.
Willy Tarreau844e3c52008-01-11 16:28:18 +01007403
7404
Willy Tarreau844e3c52008-01-11 16:28:18 +01007405option transparent
7406no option transparent
7407 Enable client-side transparent proxying
7408 May be used in sections : defaults | frontend | listen | backend
Willy Tarreau4b1f8592008-12-23 23:13:55 +01007409 yes | no | yes | yes
Willy Tarreau844e3c52008-01-11 16:28:18 +01007410 Arguments : none
7411
7412 This option was introduced in order to provide layer 7 persistence to layer 3
7413 load balancers. The idea is to use the OS's ability to redirect an incoming
7414 connection for a remote address to a local process (here HAProxy), and let
7415 this process know what address was initially requested. When this option is
7416 used, sessions without cookies will be forwarded to the original destination
7417 IP address of the incoming request (which should match that of another
7418 equipment), while requests with cookies will still be forwarded to the
7419 appropriate server.
7420
7421 Note that contrary to a common belief, this option does NOT make HAProxy
7422 present the client's IP to the server when establishing the connection.
7423
Willy Tarreaua1146052011-03-01 09:51:54 +01007424 See also: the "usesrc" argument of the "source" keyword, and the
Willy Tarreaueabeafa2008-01-16 16:17:06 +01007425 "transparent" option of the "bind" keyword.
Willy Tarreau844e3c52008-01-11 16:28:18 +01007426
Willy Tarreaubf1f8162007-12-28 17:42:56 +01007427
Simon Horman98637e52014-06-20 12:30:16 +09007428external-check command <command>
7429 Executable to run when performing an external-check
7430 May be used in sections : defaults | frontend | listen | backend
7431 yes | no | yes | yes
7432
7433 Arguments :
7434 <command> is the external command to run
7435
Simon Horman98637e52014-06-20 12:30:16 +09007436 The arguments passed to the to the command are:
7437
Cyril Bonté777be862014-12-02 21:21:35 +01007438 <proxy_address> <proxy_port> <server_address> <server_port>
Simon Horman98637e52014-06-20 12:30:16 +09007439
Cyril Bonté777be862014-12-02 21:21:35 +01007440 The <proxy_address> and <proxy_port> are derived from the first listener
7441 that is either IPv4, IPv6 or a UNIX socket. In the case of a UNIX socket
7442 listener the proxy_address will be the path of the socket and the
7443 <proxy_port> will be the string "NOT_USED". In a backend section, it's not
7444 possible to determine a listener, and both <proxy_address> and <proxy_port>
7445 will have the string value "NOT_USED".
Simon Horman98637e52014-06-20 12:30:16 +09007446
Cyril Bonté72cda2a2014-12-27 22:28:39 +01007447 Some values are also provided through environment variables.
7448
7449 Environment variables :
7450 HAPROXY_PROXY_ADDR The first bind address if available (or empty if not
7451 applicable, for example in a "backend" section).
7452
7453 HAPROXY_PROXY_ID The backend id.
7454
7455 HAPROXY_PROXY_NAME The backend name.
7456
7457 HAPROXY_PROXY_PORT The first bind port if available (or empty if not
7458 applicable, for example in a "backend" section or
7459 for a UNIX socket).
7460
7461 HAPROXY_SERVER_ADDR The server address.
7462
7463 HAPROXY_SERVER_CURCONN The current number of connections on the server.
7464
7465 HAPROXY_SERVER_ID The server id.
7466
7467 HAPROXY_SERVER_MAXCONN The server max connections.
7468
7469 HAPROXY_SERVER_NAME The server name.
7470
7471 HAPROXY_SERVER_PORT The server port if available (or empty for a UNIX
7472 socket).
7473
7474 PATH The PATH environment variable used when executing
7475 the command may be set using "external-check path".
7476
William Lallemand4d03e432019-06-14 15:35:37 +02007477 See also "2.3. Environment variables" for other variables.
7478
Simon Horman98637e52014-06-20 12:30:16 +09007479 If the command executed and exits with a zero status then the check is
7480 considered to have passed, otherwise the check is considered to have
7481 failed.
7482
7483 Example :
7484 external-check command /bin/true
7485
7486 See also : "external-check", "option external-check", "external-check path"
7487
7488
7489external-check path <path>
7490 The value of the PATH environment variable used when running an external-check
7491 May be used in sections : defaults | frontend | listen | backend
7492 yes | no | yes | yes
7493
7494 Arguments :
7495 <path> is the path used when executing external command to run
7496
7497 The default path is "".
7498
7499 Example :
7500 external-check path "/usr/bin:/bin"
7501
7502 See also : "external-check", "option external-check",
7503 "external-check command"
7504
7505
Emeric Brun647caf12009-06-30 17:57:00 +02007506persist rdp-cookie
Hervé COMMOWICKa3eb39c2011-08-05 18:48:51 +02007507persist rdp-cookie(<name>)
Emeric Brun647caf12009-06-30 17:57:00 +02007508 Enable RDP cookie-based persistence
7509 May be used in sections : defaults | frontend | listen | backend
7510 yes | no | yes | yes
7511 Arguments :
7512 <name> is the optional name of the RDP cookie to check. If omitted, the
Willy Tarreau61e28f22010-05-16 22:31:05 +02007513 default cookie name "msts" will be used. There currently is no
7514 valid reason to change this name.
Emeric Brun647caf12009-06-30 17:57:00 +02007515
7516 This statement enables persistence based on an RDP cookie. The RDP cookie
7517 contains all information required to find the server in the list of known
Davor Ocelice9ed2812017-12-25 17:49:28 +01007518 servers. So when this option is set in the backend, the request is analyzed
Emeric Brun647caf12009-06-30 17:57:00 +02007519 and if an RDP cookie is found, it is decoded. If it matches a known server
7520 which is still UP (or if "option persist" is set), then the connection is
7521 forwarded to this server.
7522
7523 Note that this only makes sense in a TCP backend, but for this to work, the
7524 frontend must have waited long enough to ensure that an RDP cookie is present
7525 in the request buffer. This is the same requirement as with the "rdp-cookie"
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01007526 load-balancing method. Thus it is highly recommended to put all statements in
Emeric Brun647caf12009-06-30 17:57:00 +02007527 a single "listen" section.
7528
Willy Tarreau61e28f22010-05-16 22:31:05 +02007529 Also, it is important to understand that the terminal server will emit this
7530 RDP cookie only if it is configured for "token redirection mode", which means
7531 that the "IP address redirection" option is disabled.
7532
Emeric Brun647caf12009-06-30 17:57:00 +02007533 Example :
7534 listen tse-farm
7535 bind :3389
7536 # wait up to 5s for an RDP cookie in the request
7537 tcp-request inspect-delay 5s
7538 tcp-request content accept if RDP_COOKIE
7539 # apply RDP cookie persistence
7540 persist rdp-cookie
7541 # if server is unknown, let's balance on the same cookie.
Cyril Bontédc4d9032012-04-08 21:57:39 +02007542 # alternatively, "balance leastconn" may be useful too.
Emeric Brun647caf12009-06-30 17:57:00 +02007543 balance rdp-cookie
7544 server srv1 1.1.1.1:3389
7545 server srv2 1.1.1.2:3389
7546
Simon Hormanab814e02011-06-24 14:50:20 +09007547 See also : "balance rdp-cookie", "tcp-request", the "req_rdp_cookie" ACL and
7548 the rdp_cookie pattern fetch function.
Emeric Brun647caf12009-06-30 17:57:00 +02007549
7550
Willy Tarreau3a7d2072009-03-05 23:48:25 +01007551rate-limit sessions <rate>
7552 Set a limit on the number of new sessions accepted per second on a frontend
7553 May be used in sections : defaults | frontend | listen | backend
7554 yes | yes | yes | no
7555 Arguments :
7556 <rate> The <rate> parameter is an integer designating the maximum number
7557 of new sessions per second to accept on the frontend.
7558
7559 When the frontend reaches the specified number of new sessions per second, it
7560 stops accepting new connections until the rate drops below the limit again.
7561 During this time, the pending sessions will be kept in the socket's backlog
7562 (in system buffers) and haproxy will not even be aware that sessions are
7563 pending. When applying very low limit on a highly loaded service, it may make
7564 sense to increase the socket's backlog using the "backlog" keyword.
7565
7566 This feature is particularly efficient at blocking connection-based attacks
7567 or service abuse on fragile servers. Since the session rate is measured every
7568 millisecond, it is extremely accurate. Also, the limit applies immediately,
7569 no delay is needed at all to detect the threshold.
7570
7571 Example : limit the connection rate on SMTP to 10 per second max
7572 listen smtp
7573 mode tcp
7574 bind :25
7575 rate-limit sessions 10
Panagiotis Panagiotopoulos7282d8e2016-02-11 16:37:15 +02007576 server smtp1 127.0.0.1:1025
Willy Tarreau3a7d2072009-03-05 23:48:25 +01007577
Willy Tarreaua17c2d92011-07-25 08:16:20 +02007578 Note : when the maximum rate is reached, the frontend's status is not changed
7579 but its sockets appear as "WAITING" in the statistics if the
7580 "socket-stats" option is enabled.
Willy Tarreau3a7d2072009-03-05 23:48:25 +01007581
7582 See also : the "backlog" keyword and the "fe_sess_rate" ACL criterion.
7583
7584
Willy Tarreau2e1dca82012-09-12 08:43:15 +02007585redirect location <loc> [code <code>] <option> [{if | unless} <condition>]
7586redirect prefix <pfx> [code <code>] <option> [{if | unless} <condition>]
7587redirect scheme <sch> [code <code>] <option> [{if | unless} <condition>]
Willy Tarreaub463dfb2008-06-07 23:08:56 +02007588 Return an HTTP redirection if/unless a condition is matched
7589 May be used in sections : defaults | frontend | listen | backend
7590 no | yes | yes | yes
7591
7592 If/unless the condition is matched, the HTTP request will lead to a redirect
Willy Tarreauf285f542010-01-03 20:03:03 +01007593 response. If no condition is specified, the redirect applies unconditionally.
Willy Tarreaub463dfb2008-06-07 23:08:56 +02007594
Willy Tarreau0140f252008-11-19 21:07:09 +01007595 Arguments :
Willy Tarreau2e1dca82012-09-12 08:43:15 +02007596 <loc> With "redirect location", the exact value in <loc> is placed into
Thierry FOURNIERd18cd0f2013-11-29 12:15:45 +01007597 the HTTP "Location" header. When used in an "http-request" rule,
7598 <loc> value follows the log-format rules and can include some
7599 dynamic values (see Custom Log Format in section 8.2.4).
Willy Tarreau2e1dca82012-09-12 08:43:15 +02007600
7601 <pfx> With "redirect prefix", the "Location" header is built from the
7602 concatenation of <pfx> and the complete URI path, including the
7603 query string, unless the "drop-query" option is specified (see
7604 below). As a special case, if <pfx> equals exactly "/", then
7605 nothing is inserted before the original URI. It allows one to
Thierry FOURNIERd18cd0f2013-11-29 12:15:45 +01007606 redirect to the same URL (for instance, to insert a cookie). When
7607 used in an "http-request" rule, <pfx> value follows the log-format
7608 rules and can include some dynamic values (see Custom Log Format
7609 in section 8.2.4).
Willy Tarreau2e1dca82012-09-12 08:43:15 +02007610
7611 <sch> With "redirect scheme", then the "Location" header is built by
7612 concatenating <sch> with "://" then the first occurrence of the
7613 "Host" header, and then the URI path, including the query string
7614 unless the "drop-query" option is specified (see below). If no
7615 path is found or if the path is "*", then "/" is used instead. If
7616 no "Host" header is found, then an empty host component will be
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03007617 returned, which most recent browsers interpret as redirecting to
Willy Tarreau2e1dca82012-09-12 08:43:15 +02007618 the same host. This directive is mostly used to redirect HTTP to
Thierry FOURNIERd18cd0f2013-11-29 12:15:45 +01007619 HTTPS. When used in an "http-request" rule, <sch> value follows
7620 the log-format rules and can include some dynamic values (see
7621 Custom Log Format in section 8.2.4).
Willy Tarreau0140f252008-11-19 21:07:09 +01007622
7623 <code> The code is optional. It indicates which type of HTTP redirection
Willy Tarreaub67fdc42013-03-29 19:28:11 +01007624 is desired. Only codes 301, 302, 303, 307 and 308 are supported,
7625 with 302 used by default if no code is specified. 301 means
7626 "Moved permanently", and a browser may cache the Location. 302
Baptiste Assmannea849c02015-08-03 11:42:50 +02007627 means "Moved temporarily" and means that the browser should not
Willy Tarreaub67fdc42013-03-29 19:28:11 +01007628 cache the redirection. 303 is equivalent to 302 except that the
7629 browser will fetch the location with a GET method. 307 is just
7630 like 302 but makes it clear that the same method must be reused.
7631 Likewise, 308 replaces 301 if the same method must be used.
Willy Tarreau0140f252008-11-19 21:07:09 +01007632
7633 <option> There are several options which can be specified to adjust the
Davor Ocelice9ed2812017-12-25 17:49:28 +01007634 expected behavior of a redirection :
Willy Tarreau0140f252008-11-19 21:07:09 +01007635
7636 - "drop-query"
7637 When this keyword is used in a prefix-based redirection, then the
7638 location will be set without any possible query-string, which is useful
7639 for directing users to a non-secure page for instance. It has no effect
7640 with a location-type redirect.
7641
Willy Tarreau81e3b4f2010-01-10 00:42:19 +01007642 - "append-slash"
7643 This keyword may be used in conjunction with "drop-query" to redirect
7644 users who use a URL not ending with a '/' to the same one with the '/'.
7645 It can be useful to ensure that search engines will only see one URL.
7646 For this, a return code 301 is preferred.
7647
Willy Tarreau0140f252008-11-19 21:07:09 +01007648 - "set-cookie NAME[=value]"
7649 A "Set-Cookie" header will be added with NAME (and optionally "=value")
7650 to the response. This is sometimes used to indicate that a user has
7651 been seen, for instance to protect against some types of DoS. No other
7652 cookie option is added, so the cookie will be a session cookie. Note
7653 that for a browser, a sole cookie name without an equal sign is
7654 different from a cookie with an equal sign.
7655
7656 - "clear-cookie NAME[=]"
7657 A "Set-Cookie" header will be added with NAME (and optionally "="), but
7658 with the "Max-Age" attribute set to zero. This will tell the browser to
7659 delete this cookie. It is useful for instance on logout pages. It is
7660 important to note that clearing the cookie "NAME" will not remove a
7661 cookie set with "NAME=value". You have to clear the cookie "NAME=" for
7662 that, because the browser makes the difference.
Willy Tarreaub463dfb2008-06-07 23:08:56 +02007663
7664 Example: move the login URL only to HTTPS.
7665 acl clear dst_port 80
7666 acl secure dst_port 8080
7667 acl login_page url_beg /login
Willy Tarreau0140f252008-11-19 21:07:09 +01007668 acl logout url_beg /logout
Willy Tarreau79da4692008-11-19 20:03:04 +01007669 acl uid_given url_reg /login?userid=[^&]+
Willy Tarreau0140f252008-11-19 21:07:09 +01007670 acl cookie_set hdr_sub(cookie) SEEN=1
7671
7672 redirect prefix https://mysite.com set-cookie SEEN=1 if !cookie_set
Willy Tarreau79da4692008-11-19 20:03:04 +01007673 redirect prefix https://mysite.com if login_page !secure
7674 redirect prefix http://mysite.com drop-query if login_page !uid_given
7675 redirect location http://mysite.com/ if !login_page secure
Willy Tarreau0140f252008-11-19 21:07:09 +01007676 redirect location / clear-cookie USERID= if logout
Willy Tarreaub463dfb2008-06-07 23:08:56 +02007677
Willy Tarreau81e3b4f2010-01-10 00:42:19 +01007678 Example: send redirects for request for articles without a '/'.
7679 acl missing_slash path_reg ^/article/[^/]*$
7680 redirect code 301 prefix / drop-query append-slash if missing_slash
7681
Willy Tarreau2e1dca82012-09-12 08:43:15 +02007682 Example: redirect all HTTP traffic to HTTPS when SSL is handled by haproxy.
David BERARDe7153042012-11-03 00:11:31 +01007683 redirect scheme https if !{ ssl_fc }
Willy Tarreau2e1dca82012-09-12 08:43:15 +02007684
Thierry FOURNIERd18cd0f2013-11-29 12:15:45 +01007685 Example: append 'www.' prefix in front of all hosts not having it
Coen Rosdorff596659b2016-04-11 11:33:49 +02007686 http-request redirect code 301 location \
7687 http://www.%[hdr(host)]%[capture.req.uri] \
7688 unless { hdr_beg(host) -i www }
Thierry FOURNIERd18cd0f2013-11-29 12:15:45 +01007689
Willy Tarreauc57f0e22009-05-10 13:12:33 +02007690 See section 7 about ACL usage.
Willy Tarreaub463dfb2008-06-07 23:08:56 +02007691
7692
Willy Tarreau96d51952019-05-22 20:34:35 +02007693reqadd <string> [{if | unless} <cond>] (deprecated)
Willy Tarreau303c0352008-01-17 19:01:39 +01007694 Add a header at the end of the HTTP request
7695 May be used in sections : defaults | frontend | listen | backend
7696 no | yes | yes | yes
7697 Arguments :
7698 <string> is the complete line to be added. Any space or known delimiter
7699 must be escaped using a backslash ('\'). Please refer to section
Willy Tarreauc57f0e22009-05-10 13:12:33 +02007700 6 about HTTP header manipulation for more information.
Willy Tarreau303c0352008-01-17 19:01:39 +01007701
Willy Tarreau8abd4cd2010-01-31 14:30:44 +01007702 <cond> is an optional matching condition built from ACLs. It makes it
7703 possible to ignore this rule when other conditions are not met.
7704
Willy Tarreau303c0352008-01-17 19:01:39 +01007705 A new line consisting in <string> followed by a line feed will be added after
7706 the last header of an HTTP request.
7707
7708 Header transformations only apply to traffic which passes through HAProxy,
7709 and not to traffic generated by HAProxy, such as health-checks or error
7710 responses.
7711
Willy Tarreau8abd4cd2010-01-31 14:30:44 +01007712 Example : add "X-Proto: SSL" to requests coming via port 81
7713 acl is-ssl dst_port 81
7714 reqadd X-Proto:\ SSL if is-ssl
7715
Ruoshan Huangeb5a3632015-12-08 21:00:23 +08007716 See also: "rspadd", "http-request", section 6 about HTTP header manipulation,
7717 and section 7 about ACLs.
Willy Tarreau303c0352008-01-17 19:01:39 +01007718
7719
Willy Tarreau96d51952019-05-22 20:34:35 +02007720reqallow <search> [{if | unless} <cond>] (deprecated)
7721reqiallow <search> [{if | unless} <cond>] (ignore case) (deprecated)
Willy Tarreau303c0352008-01-17 19:01:39 +01007722 Definitely allow an HTTP request if a line matches a regular expression
7723 May be used in sections : defaults | frontend | listen | backend
7724 no | yes | yes | yes
7725 Arguments :
7726 <search> is the regular expression applied to HTTP headers and to the
7727 request line. This is an extended regular expression. Parenthesis
7728 grouping is supported and no preliminary backslash is required.
7729 Any space or known delimiter must be escaped using a backslash
7730 ('\'). The pattern applies to a full line at a time. The
7731 "reqallow" keyword strictly matches case while "reqiallow"
7732 ignores case.
7733
Willy Tarreau5321c422010-01-28 20:35:13 +01007734 <cond> is an optional matching condition built from ACLs. It makes it
7735 possible to ignore this rule when other conditions are not met.
7736
Willy Tarreau303c0352008-01-17 19:01:39 +01007737 A request containing any line which matches extended regular expression
7738 <search> will mark the request as allowed, even if any later test would
7739 result in a deny. The test applies both to the request line and to request
7740 headers. Keep in mind that URLs in request line are case-sensitive while
Willy Tarreaud72758d2010-01-12 10:42:19 +01007741 header names are not.
Willy Tarreau303c0352008-01-17 19:01:39 +01007742
7743 It is easier, faster and more powerful to use ACLs to write access policies.
7744 Reqdeny, reqallow and reqpass should be avoided in new designs.
7745
7746 Example :
7747 # allow www.* but refuse *.local
7748 reqiallow ^Host:\ www\.
7749 reqideny ^Host:\ .*\.local
7750
Ruoshan Huangeb5a3632015-12-08 21:00:23 +08007751 See also: "reqdeny", "block", "http-request", section 6 about HTTP header
7752 manipulation, and section 7 about ACLs.
Willy Tarreau303c0352008-01-17 19:01:39 +01007753
7754
Willy Tarreau96d51952019-05-22 20:34:35 +02007755reqdel <search> [{if | unless} <cond>] (deprecated)
7756reqidel <search> [{if | unless} <cond>] (ignore case) (deprecated)
Willy Tarreau303c0352008-01-17 19:01:39 +01007757 Delete all headers matching a regular expression in an HTTP request
7758 May be used in sections : defaults | frontend | listen | backend
7759 no | yes | yes | yes
7760 Arguments :
7761 <search> is the regular expression applied to HTTP headers and to the
7762 request line. This is an extended regular expression. Parenthesis
7763 grouping is supported and no preliminary backslash is required.
7764 Any space or known delimiter must be escaped using a backslash
7765 ('\'). The pattern applies to a full line at a time. The "reqdel"
7766 keyword strictly matches case while "reqidel" ignores case.
7767
Willy Tarreau5321c422010-01-28 20:35:13 +01007768 <cond> is an optional matching condition built from ACLs. It makes it
7769 possible to ignore this rule when other conditions are not met.
7770
Willy Tarreau303c0352008-01-17 19:01:39 +01007771 Any header line matching extended regular expression <search> in the request
7772 will be completely deleted. Most common use of this is to remove unwanted
7773 and/or dangerous headers or cookies from a request before passing it to the
7774 next servers.
7775
7776 Header transformations only apply to traffic which passes through HAProxy,
7777 and not to traffic generated by HAProxy, such as health-checks or error
7778 responses. Keep in mind that header names are not case-sensitive.
7779
7780 Example :
7781 # remove X-Forwarded-For header and SERVER cookie
7782 reqidel ^X-Forwarded-For:.*
7783 reqidel ^Cookie:.*SERVER=
7784
Ruoshan Huangeb5a3632015-12-08 21:00:23 +08007785 See also: "reqadd", "reqrep", "rspdel", "http-request", section 6 about
7786 HTTP header manipulation, and section 7 about ACLs.
Willy Tarreau303c0352008-01-17 19:01:39 +01007787
7788
Willy Tarreau96d51952019-05-22 20:34:35 +02007789reqdeny <search> [{if | unless} <cond>] (deprecated)
7790reqideny <search> [{if | unless} <cond>] (ignore case) (deprecated)
Willy Tarreau303c0352008-01-17 19:01:39 +01007791 Deny an HTTP request if a line matches a regular expression
7792 May be used in sections : defaults | frontend | listen | backend
7793 no | yes | yes | yes
7794 Arguments :
7795 <search> is the regular expression applied to HTTP headers and to the
7796 request line. This is an extended regular expression. Parenthesis
7797 grouping is supported and no preliminary backslash is required.
7798 Any space or known delimiter must be escaped using a backslash
7799 ('\'). The pattern applies to a full line at a time. The
7800 "reqdeny" keyword strictly matches case while "reqideny" ignores
7801 case.
7802
Willy Tarreau5321c422010-01-28 20:35:13 +01007803 <cond> is an optional matching condition built from ACLs. It makes it
7804 possible to ignore this rule when other conditions are not met.
7805
Willy Tarreau303c0352008-01-17 19:01:39 +01007806 A request containing any line which matches extended regular expression
7807 <search> will mark the request as denied, even if any later test would
7808 result in an allow. The test applies both to the request line and to request
7809 headers. Keep in mind that URLs in request line are case-sensitive while
Willy Tarreaud72758d2010-01-12 10:42:19 +01007810 header names are not.
Willy Tarreau303c0352008-01-17 19:01:39 +01007811
Willy Tarreauced27012008-01-17 20:35:34 +01007812 A denied request will generate an "HTTP 403 forbidden" response once the
Willy Tarreaud2a4aa22008-01-31 15:28:22 +01007813 complete request has been parsed. This is consistent with what is practiced
Willy Tarreaud72758d2010-01-12 10:42:19 +01007814 using ACLs.
Willy Tarreauced27012008-01-17 20:35:34 +01007815
Willy Tarreau303c0352008-01-17 19:01:39 +01007816 It is easier, faster and more powerful to use ACLs to write access policies.
7817 Reqdeny, reqallow and reqpass should be avoided in new designs.
7818
7819 Example :
7820 # refuse *.local, then allow www.*
7821 reqideny ^Host:\ .*\.local
7822 reqiallow ^Host:\ www\.
7823
Ruoshan Huangeb5a3632015-12-08 21:00:23 +08007824 See also: "reqallow", "rspdeny", "block", "http-request", section 6 about
7825 HTTP header manipulation, and section 7 about ACLs.
Willy Tarreau303c0352008-01-17 19:01:39 +01007826
7827
Willy Tarreau96d51952019-05-22 20:34:35 +02007828reqpass <search> [{if | unless} <cond>] (deprecated)
7829reqipass <search> [{if | unless} <cond>] (ignore case) (deprecated)
Willy Tarreau303c0352008-01-17 19:01:39 +01007830 Ignore any HTTP request line matching a regular expression in next rules
7831 May be used in sections : defaults | frontend | listen | backend
7832 no | yes | yes | yes
7833 Arguments :
7834 <search> is the regular expression applied to HTTP headers and to the
7835 request line. This is an extended regular expression. Parenthesis
7836 grouping is supported and no preliminary backslash is required.
7837 Any space or known delimiter must be escaped using a backslash
7838 ('\'). The pattern applies to a full line at a time. The
7839 "reqpass" keyword strictly matches case while "reqipass" ignores
7840 case.
7841
Willy Tarreau5321c422010-01-28 20:35:13 +01007842 <cond> is an optional matching condition built from ACLs. It makes it
7843 possible to ignore this rule when other conditions are not met.
7844
Willy Tarreau303c0352008-01-17 19:01:39 +01007845 A request containing any line which matches extended regular expression
7846 <search> will skip next rules, without assigning any deny or allow verdict.
7847 The test applies both to the request line and to request headers. Keep in
7848 mind that URLs in request line are case-sensitive while header names are not.
7849
7850 It is easier, faster and more powerful to use ACLs to write access policies.
7851 Reqdeny, reqallow and reqpass should be avoided in new designs.
7852
7853 Example :
7854 # refuse *.local, then allow www.*, but ignore "www.private.local"
7855 reqipass ^Host:\ www.private\.local
7856 reqideny ^Host:\ .*\.local
7857 reqiallow ^Host:\ www\.
7858
Ruoshan Huangeb5a3632015-12-08 21:00:23 +08007859 See also: "reqallow", "reqdeny", "block", "http-request", section 6 about
7860 HTTP header manipulation, and section 7 about ACLs.
Willy Tarreau303c0352008-01-17 19:01:39 +01007861
7862
Willy Tarreau96d51952019-05-22 20:34:35 +02007863reqrep <search> <string> [{if | unless} <cond>] (deprecated)
7864reqirep <search> <string> [{if | unless} <cond>] (ignore case) (deprecated)
Willy Tarreau303c0352008-01-17 19:01:39 +01007865 Replace a regular expression with a string in an HTTP request line
7866 May be used in sections : defaults | frontend | listen | backend
7867 no | yes | yes | yes
7868 Arguments :
7869 <search> is the regular expression applied to HTTP headers and to the
7870 request line. This is an extended regular expression. Parenthesis
7871 grouping is supported and no preliminary backslash is required.
7872 Any space or known delimiter must be escaped using a backslash
7873 ('\'). The pattern applies to a full line at a time. The "reqrep"
7874 keyword strictly matches case while "reqirep" ignores case.
7875
7876 <string> is the complete line to be added. Any space or known delimiter
7877 must be escaped using a backslash ('\'). References to matched
7878 pattern groups are possible using the common \N form, with N
7879 being a single digit between 0 and 9. Please refer to section
Willy Tarreauc57f0e22009-05-10 13:12:33 +02007880 6 about HTTP header manipulation for more information.
Willy Tarreau303c0352008-01-17 19:01:39 +01007881
Willy Tarreau5321c422010-01-28 20:35:13 +01007882 <cond> is an optional matching condition built from ACLs. It makes it
7883 possible to ignore this rule when other conditions are not met.
7884
Willy Tarreau303c0352008-01-17 19:01:39 +01007885 Any line matching extended regular expression <search> in the request (both
7886 the request line and header lines) will be completely replaced with <string>.
7887 Most common use of this is to rewrite URLs or domain names in "Host" headers.
7888
7889 Header transformations only apply to traffic which passes through HAProxy,
7890 and not to traffic generated by HAProxy, such as health-checks or error
7891 responses. Note that for increased readability, it is suggested to add enough
7892 spaces between the request and the response. Keep in mind that URLs in
7893 request line are case-sensitive while header names are not.
7894
7895 Example :
7896 # replace "/static/" with "/" at the beginning of any request path.
Dmitry Sivachenko7823de32012-05-16 14:00:26 +04007897 reqrep ^([^\ :]*)\ /static/(.*) \1\ /\2
Willy Tarreau303c0352008-01-17 19:01:39 +01007898 # replace "www.mydomain.com" with "www" in the host name.
7899 reqirep ^Host:\ www.mydomain.com Host:\ www
7900
Ruoshan Huangeb5a3632015-12-08 21:00:23 +08007901 See also: "reqadd", "reqdel", "rsprep", "tune.bufsize", "http-request",
7902 section 6 about HTTP header manipulation, and section 7 about ACLs.
Willy Tarreau303c0352008-01-17 19:01:39 +01007903
7904
Willy Tarreau96d51952019-05-22 20:34:35 +02007905reqtarpit <search> [{if | unless} <cond>] (deprecated)
7906reqitarpit <search> [{if | unless} <cond>] (ignore case) (deprecated)
Willy Tarreau303c0352008-01-17 19:01:39 +01007907 Tarpit an HTTP request containing a line matching a regular expression
7908 May be used in sections : defaults | frontend | listen | backend
7909 no | yes | yes | yes
7910 Arguments :
7911 <search> is the regular expression applied to HTTP headers and to the
7912 request line. This is an extended regular expression. Parenthesis
7913 grouping is supported and no preliminary backslash is required.
7914 Any space or known delimiter must be escaped using a backslash
7915 ('\'). The pattern applies to a full line at a time. The
7916 "reqtarpit" keyword strictly matches case while "reqitarpit"
7917 ignores case.
7918
Willy Tarreau5321c422010-01-28 20:35:13 +01007919 <cond> is an optional matching condition built from ACLs. It makes it
7920 possible to ignore this rule when other conditions are not met.
7921
Willy Tarreau303c0352008-01-17 19:01:39 +01007922 A request containing any line which matches extended regular expression
7923 <search> will be tarpitted, which means that it will connect to nowhere, will
Willy Tarreauced27012008-01-17 20:35:34 +01007924 be kept open for a pre-defined time, then will return an HTTP error 500 so
7925 that the attacker does not suspect it has been tarpitted. The status 500 will
7926 be reported in the logs, but the completion flags will indicate "PT". The
Willy Tarreau303c0352008-01-17 19:01:39 +01007927 delay is defined by "timeout tarpit", or "timeout connect" if the former is
7928 not set.
7929
7930 The goal of the tarpit is to slow down robots attacking servers with
7931 identifiable requests. Many robots limit their outgoing number of connections
7932 and stay connected waiting for a reply which can take several minutes to
7933 come. Depending on the environment and attack, it may be particularly
7934 efficient at reducing the load on the network and firewalls.
7935
Willy Tarreau5321c422010-01-28 20:35:13 +01007936 Examples :
Davor Ocelice9ed2812017-12-25 17:49:28 +01007937 # ignore user-agents reporting any flavor of "Mozilla" or "MSIE", but
Willy Tarreau303c0352008-01-17 19:01:39 +01007938 # block all others.
7939 reqipass ^User-Agent:\.*(Mozilla|MSIE)
7940 reqitarpit ^User-Agent:
7941
Willy Tarreau5321c422010-01-28 20:35:13 +01007942 # block bad guys
7943 acl badguys src 10.1.0.3 172.16.13.20/28
7944 reqitarpit . if badguys
7945
Ruoshan Huangeb5a3632015-12-08 21:00:23 +08007946 See also: "reqallow", "reqdeny", "reqpass", "http-request", section 6
7947 about HTTP header manipulation, and section 7 about ACLs.
Willy Tarreau303c0352008-01-17 19:01:39 +01007948
7949
Willy Tarreaue5c5ce92008-06-20 17:27:19 +02007950retries <value>
7951 Set the number of retries to perform on a server after a connection failure
7952 May be used in sections: defaults | frontend | listen | backend
7953 yes | no | yes | yes
7954 Arguments :
7955 <value> is the number of times a connection attempt should be retried on
7956 a server when a connection either is refused or times out. The
7957 default value is 3.
7958
7959 It is important to understand that this value applies to the number of
7960 connection attempts, not full requests. When a connection has effectively
7961 been established to a server, there will be no more retry.
7962
7963 In order to avoid immediate reconnections to a server which is restarting,
Joseph Lynch726ab712015-05-11 23:25:34 -07007964 a turn-around timer of min("timeout connect", one second) is applied before
7965 a retry occurs.
Willy Tarreaue5c5ce92008-06-20 17:27:19 +02007966
7967 When "option redispatch" is set, the last retry may be performed on another
7968 server even if a cookie references a different server.
7969
7970 See also : "option redispatch"
7971
7972
Olivier Houcharda254a372019-04-05 15:30:12 +02007973retry-on [list of keywords]
7974 Specify when to attempt to automatically retry a failed request
7975 May be used in sections: defaults | frontend | listen | backend
7976 yes | no | yes | yes
7977 Arguments :
7978 <keywords> is a list of keywords or HTTP status codes, each representing a
7979 type of failure event on which an attempt to retry the request
7980 is desired. Please read the notes at the bottom before changing
7981 this setting. The following keywords are supported :
7982
7983 none never retry
7984
7985 conn-failure retry when the connection or the SSL handshake failed
7986 and the request could not be sent. This is the default.
7987
7988 empty-response retry when the server connection was closed after part
7989 of the request was sent, and nothing was received from
7990 the server. This type of failure may be caused by the
7991 request timeout on the server side, poor network
7992 condition, or a server crash or restart while
7993 processing the request.
7994
Olivier Houcharde3249a92019-05-03 23:01:47 +02007995 junk-response retry when the server returned something not looking
7996 like a complete HTTP response. This includes partial
7997 responses headers as well as non-HTTP contents. It
7998 usually is a bad idea to retry on such events, which
7999 may be caused a configuration issue (wrong server port)
8000 or by the request being harmful to the server (buffer
8001 overflow attack for example).
8002
Olivier Houcharda254a372019-04-05 15:30:12 +02008003 response-timeout the server timeout stroke while waiting for the server
8004 to respond to the request. This may be caused by poor
8005 network condition, the reuse of an idle connection
8006 which has expired on the path, or by the request being
8007 extremely expensive to process. It generally is a bad
8008 idea to retry on such events on servers dealing with
8009 heavy database processing (full scans, etc) as it may
8010 amplify denial of service attacks.
8011
Olivier Houchard865d8392019-05-03 22:46:27 +02008012 0rtt-rejected retry requests which were sent over early data and were
8013 rejected by the server. These requests are generally
8014 considered to be safe to retry.
8015
Olivier Houcharda254a372019-04-05 15:30:12 +02008016 <status> any HTTP status code among "404" (Not Found), "408"
8017 (Request Timeout), "425" (Too Early), "500" (Server
8018 Error), "501" (Not Implemented), "502" (Bad Gateway),
8019 "503" (Service Unavailable), "504" (Gateway Timeout).
8020
Olivier Houchardddf0e032019-05-10 18:05:40 +02008021 all-retryable-errors
8022 retry request for any error that are considered
8023 retryable. This currently activates "conn-failure",
8024 "empty-response", "junk-response", "response-timeout",
8025 "0rtt-rejected", "500", "502", "503", and "504".
8026
Olivier Houcharda254a372019-04-05 15:30:12 +02008027 Using this directive replaces any previous settings with the new ones; it is
8028 not cumulative.
8029
8030 Please note that using anything other than "none" and "conn-failure" requires
8031 to allocate a buffer and copy the whole request into it, so it has memory and
8032 performance impacts. Requests not fitting in a single buffer will never be
8033 retried (see the global tune.bufsize setting).
8034
8035 You have to make sure the application has a replay protection mechanism built
8036 in such as a unique transaction IDs passed in requests, or that replaying the
8037 same request has no consequence, or it is very dangerous to use any retry-on
8038 value beside "conn-failure" and "none". Static file servers and caches are
8039 generally considered safe against any type of retry. Using a status code can
8040 be useful to quickly leave a server showing an abnormal behavior (out of
8041 memory, file system issues, etc), but in this case it may be a good idea to
8042 immediately redispatch the connection to another server (please see "option
8043 redispatch" for this). Last, it is important to understand that most causes
8044 of failures are the requests themselves and that retrying a request causing a
8045 server to misbehave will often make the situation even worse for this server,
8046 or for the whole service in case of redispatch.
8047
8048 Unless you know exactly how the application deals with replayed requests, you
8049 should not use this directive.
8050
8051 The default is "conn-failure".
8052
8053 See also: "retries", "option redispatch", "tune.bufsize"
8054
Willy Tarreau96d51952019-05-22 20:34:35 +02008055rspadd <string> [{if | unless} <cond>] (deprecated)
Willy Tarreau303c0352008-01-17 19:01:39 +01008056 Add a header at the end of the HTTP response
8057 May be used in sections : defaults | frontend | listen | backend
8058 no | yes | yes | yes
8059 Arguments :
8060 <string> is the complete line to be added. Any space or known delimiter
8061 must be escaped using a backslash ('\'). Please refer to section
Willy Tarreauc57f0e22009-05-10 13:12:33 +02008062 6 about HTTP header manipulation for more information.
Willy Tarreau303c0352008-01-17 19:01:39 +01008063
Willy Tarreaufdb563c2010-01-31 15:43:27 +01008064 <cond> is an optional matching condition built from ACLs. It makes it
8065 possible to ignore this rule when other conditions are not met.
8066
Willy Tarreau303c0352008-01-17 19:01:39 +01008067 A new line consisting in <string> followed by a line feed will be added after
8068 the last header of an HTTP response.
8069
8070 Header transformations only apply to traffic which passes through HAProxy,
8071 and not to traffic generated by HAProxy, such as health-checks or error
8072 responses.
8073
Ruoshan Huangeb5a3632015-12-08 21:00:23 +08008074 See also: "rspdel" "reqadd", "http-response", section 6 about HTTP header
8075 manipulation, and section 7 about ACLs.
Willy Tarreau303c0352008-01-17 19:01:39 +01008076
8077
Willy Tarreau96d51952019-05-22 20:34:35 +02008078rspdel <search> [{if | unless} <cond>] (deprecated)
8079rspidel <search> [{if | unless} <cond>] (ignore case) (deprecated)
Willy Tarreau303c0352008-01-17 19:01:39 +01008080 Delete all headers matching a regular expression in an HTTP response
8081 May be used in sections : defaults | frontend | listen | backend
8082 no | yes | yes | yes
8083 Arguments :
8084 <search> is the regular expression applied to HTTP headers and to the
8085 response line. This is an extended regular expression, so
8086 parenthesis grouping is supported and no preliminary backslash
8087 is required. Any space or known delimiter must be escaped using
8088 a backslash ('\'). The pattern applies to a full line at a time.
8089 The "rspdel" keyword strictly matches case while "rspidel"
8090 ignores case.
8091
Willy Tarreaufdb563c2010-01-31 15:43:27 +01008092 <cond> is an optional matching condition built from ACLs. It makes it
8093 possible to ignore this rule when other conditions are not met.
8094
Willy Tarreau303c0352008-01-17 19:01:39 +01008095 Any header line matching extended regular expression <search> in the response
8096 will be completely deleted. Most common use of this is to remove unwanted
Willy Tarreau3c92c5f2011-08-28 09:45:47 +02008097 and/or sensitive headers or cookies from a response before passing it to the
Willy Tarreau303c0352008-01-17 19:01:39 +01008098 client.
8099
8100 Header transformations only apply to traffic which passes through HAProxy,
8101 and not to traffic generated by HAProxy, such as health-checks or error
8102 responses. Keep in mind that header names are not case-sensitive.
8103
8104 Example :
8105 # remove the Server header from responses
Willy Tarreau5e80e022013-05-25 08:31:25 +02008106 rspidel ^Server:.*
Willy Tarreau303c0352008-01-17 19:01:39 +01008107
Ruoshan Huangeb5a3632015-12-08 21:00:23 +08008108 See also: "rspadd", "rsprep", "reqdel", "http-response", section 6 about
8109 HTTP header manipulation, and section 7 about ACLs.
Willy Tarreau303c0352008-01-17 19:01:39 +01008110
8111
Willy Tarreau96d51952019-05-22 20:34:35 +02008112rspdeny <search> [{if | unless} <cond>] (deprecated)
8113rspideny <search> [{if | unless} <cond>] (ignore case) (deprecated)
Willy Tarreau303c0352008-01-17 19:01:39 +01008114 Block an HTTP response if a line matches a regular expression
8115 May be used in sections : defaults | frontend | listen | backend
8116 no | yes | yes | yes
8117 Arguments :
8118 <search> is the regular expression applied to HTTP headers and to the
8119 response line. This is an extended regular expression, so
8120 parenthesis grouping is supported and no preliminary backslash
8121 is required. Any space or known delimiter must be escaped using
8122 a backslash ('\'). The pattern applies to a full line at a time.
8123 The "rspdeny" keyword strictly matches case while "rspideny"
8124 ignores case.
8125
Willy Tarreaufdb563c2010-01-31 15:43:27 +01008126 <cond> is an optional matching condition built from ACLs. It makes it
8127 possible to ignore this rule when other conditions are not met.
8128
Willy Tarreau303c0352008-01-17 19:01:39 +01008129 A response containing any line which matches extended regular expression
8130 <search> will mark the request as denied. The test applies both to the
8131 response line and to response headers. Keep in mind that header names are not
8132 case-sensitive.
8133
8134 Main use of this keyword is to prevent sensitive information leak and to
Willy Tarreauced27012008-01-17 20:35:34 +01008135 block the response before it reaches the client. If a response is denied, it
8136 will be replaced with an HTTP 502 error so that the client never retrieves
8137 any sensitive data.
Willy Tarreau303c0352008-01-17 19:01:39 +01008138
8139 It is easier, faster and more powerful to use ACLs to write access policies.
8140 Rspdeny should be avoided in new designs.
8141
8142 Example :
8143 # Ensure that no content type matching ms-word will leak
8144 rspideny ^Content-type:\.*/ms-word
8145
Ruoshan Huangeb5a3632015-12-08 21:00:23 +08008146 See also: "reqdeny", "acl", "block", "http-response", section 6 about
8147 HTTP header manipulation and section 7 about ACLs.
Willy Tarreau303c0352008-01-17 19:01:39 +01008148
8149
Willy Tarreau96d51952019-05-22 20:34:35 +02008150rsprep <search> <string> [{if | unless} <cond>] (deprecated)
8151rspirep <search> <string> [{if | unless} <cond>] (ignore case) (deprecated)
Willy Tarreau303c0352008-01-17 19:01:39 +01008152 Replace a regular expression with a string in an HTTP response line
8153 May be used in sections : defaults | frontend | listen | backend
8154 no | yes | yes | yes
8155 Arguments :
8156 <search> is the regular expression applied to HTTP headers and to the
8157 response line. This is an extended regular expression, so
8158 parenthesis grouping is supported and no preliminary backslash
8159 is required. Any space or known delimiter must be escaped using
8160 a backslash ('\'). The pattern applies to a full line at a time.
8161 The "rsprep" keyword strictly matches case while "rspirep"
8162 ignores case.
8163
8164 <string> is the complete line to be added. Any space or known delimiter
8165 must be escaped using a backslash ('\'). References to matched
8166 pattern groups are possible using the common \N form, with N
8167 being a single digit between 0 and 9. Please refer to section
Willy Tarreauc57f0e22009-05-10 13:12:33 +02008168 6 about HTTP header manipulation for more information.
Willy Tarreau303c0352008-01-17 19:01:39 +01008169
Willy Tarreaufdb563c2010-01-31 15:43:27 +01008170 <cond> is an optional matching condition built from ACLs. It makes it
8171 possible to ignore this rule when other conditions are not met.
8172
Willy Tarreau303c0352008-01-17 19:01:39 +01008173 Any line matching extended regular expression <search> in the response (both
8174 the response line and header lines) will be completely replaced with
8175 <string>. Most common use of this is to rewrite Location headers.
8176
8177 Header transformations only apply to traffic which passes through HAProxy,
8178 and not to traffic generated by HAProxy, such as health-checks or error
8179 responses. Note that for increased readability, it is suggested to add enough
8180 spaces between the request and the response. Keep in mind that header names
8181 are not case-sensitive.
8182
8183 Example :
8184 # replace "Location: 127.0.0.1:8080" with "Location: www.mydomain.com"
8185 rspirep ^Location:\ 127.0.0.1:8080 Location:\ www.mydomain.com
8186
Ruoshan Huangeb5a3632015-12-08 21:00:23 +08008187 See also: "rspadd", "rspdel", "reqrep", "http-response", section 6 about
8188 HTTP header manipulation, and section 7 about ACLs.
Willy Tarreau303c0352008-01-17 19:01:39 +01008189
8190
David du Colombier486df472011-03-17 10:40:26 +01008191server <name> <address>[:[port]] [param*]
Willy Tarreaueabeafa2008-01-16 16:17:06 +01008192 Declare a server in a backend
8193 May be used in sections : defaults | frontend | listen | backend
8194 no | no | yes | yes
8195 Arguments :
8196 <name> is the internal name assigned to this server. This name will
Davor Ocelice9ed2812017-12-25 17:49:28 +01008197 appear in logs and alerts. If "http-send-name-header" is
Mark Lamourinec2247f02012-01-04 13:02:01 -05008198 set, it will be added to the request header sent to the server.
Willy Tarreaueabeafa2008-01-16 16:17:06 +01008199
David du Colombier486df472011-03-17 10:40:26 +01008200 <address> is the IPv4 or IPv6 address of the server. Alternatively, a
8201 resolvable hostname is supported, but this name will be resolved
8202 during start-up. Address "0.0.0.0" or "*" has a special meaning.
8203 It indicates that the connection will be forwarded to the same IP
Willy Tarreaud669a4f2010-07-13 14:49:50 +02008204 address as the one from the client connection. This is useful in
8205 transparent proxy architectures where the client's connection is
8206 intercepted and haproxy must forward to the original destination
8207 address. This is more or less what the "transparent" keyword does
8208 except that with a server it's possible to limit concurrency and
Willy Tarreau24709282013-03-10 21:32:12 +01008209 to report statistics. Optionally, an address family prefix may be
8210 used before the address to force the family regardless of the
8211 address format, which can be useful to specify a path to a unix
8212 socket with no slash ('/'). Currently supported prefixes are :
8213 - 'ipv4@' -> address is always IPv4
8214 - 'ipv6@' -> address is always IPv6
8215 - 'unix@' -> address is a path to a local unix socket
Willy Tarreauccfccef2014-05-10 01:49:15 +02008216 - 'abns@' -> address is in abstract namespace (Linux only)
William Lallemand2fe7dd02018-09-11 16:51:29 +02008217 - 'sockpair@' -> address is the FD of a connected unix
8218 socket or of a socketpair. During a connection, the
8219 backend creates a pair of connected sockets, and passes
8220 one of them over the FD. The bind part will use the
8221 received socket as the client FD. Should be used
8222 carefully.
William Lallemandb2f07452015-05-12 14:27:13 +02008223 You may want to reference some environment variables in the
8224 address parameter, see section 2.3 about environment
Willy Tarreau6a031d12016-11-07 19:42:35 +01008225 variables. The "init-addr" setting can be used to modify the way
8226 IP addresses should be resolved upon startup.
Willy Tarreaueabeafa2008-01-16 16:17:06 +01008227
Willy Tarreaub6205fd2012-09-24 12:27:33 +02008228 <port> is an optional port specification. If set, all connections will
Willy Tarreaueabeafa2008-01-16 16:17:06 +01008229 be sent to this port. If unset, the same port the client
8230 connected to will be used. The port may also be prefixed by a "+"
8231 or a "-". In this case, the server's port will be determined by
8232 adding this value to the client's port.
8233
8234 <param*> is a list of parameters for this server. The "server" keywords
8235 accepts an important number of options and has a complete section
Willy Tarreauc57f0e22009-05-10 13:12:33 +02008236 dedicated to it. Please refer to section 5 for more details.
Willy Tarreaueabeafa2008-01-16 16:17:06 +01008237
8238 Examples :
8239 server first 10.1.1.1:1080 cookie first check inter 1000
8240 server second 10.1.1.2:1080 cookie second check inter 1000
Willy Tarreau24709282013-03-10 21:32:12 +01008241 server transp ipv4@
William Lallemandb2f07452015-05-12 14:27:13 +02008242 server backup "${SRV_BACKUP}:1080" backup
8243 server www1_dc1 "${LAN_DC1}.101:80"
8244 server www1_dc2 "${LAN_DC2}.101:80"
Willy Tarreaueabeafa2008-01-16 16:17:06 +01008245
Willy Tarreau55dcaf62015-09-27 15:03:15 +02008246 Note: regarding Linux's abstract namespace sockets, HAProxy uses the whole
8247 sun_path length is used for the address length. Some other programs
8248 such as socat use the string length only by default. Pass the option
8249 ",unix-tightsocklen=0" to any abstract socket definition in socat to
8250 make it compatible with HAProxy's.
8251
Mark Lamourinec2247f02012-01-04 13:02:01 -05008252 See also: "default-server", "http-send-name-header" and section 5 about
8253 server options
Willy Tarreaueabeafa2008-01-16 16:17:06 +01008254
Baptiste Assmann01c6cc32015-08-23 11:45:29 +02008255server-state-file-name [<file>]
8256 Set the server state file to read, load and apply to servers available in
8257 this backend. It only applies when the directive "load-server-state-from-file"
8258 is set to "local". When <file> is not provided or if this directive is not
8259 set, then backend name is used. If <file> starts with a slash '/', then it is
8260 considered as an absolute path. Otherwise, <file> is concatenated to the
8261 global directive "server-state-file-base".
8262
8263 Example: the minimal configuration below would make HAProxy look for the
8264 state server file '/etc/haproxy/states/bk':
8265
8266 global
8267 server-state-file-base /etc/haproxy/states
8268
8269 backend bk
8270 load-server-state-from-file
8271
8272 See also: "server-state-file-base", "load-server-state-from-file", and
8273 "show servers state"
Willy Tarreaueabeafa2008-01-16 16:17:06 +01008274
Frédéric Lécaillecb4502e2017-04-20 13:36:25 +02008275server-template <prefix> <num | range> <fqdn>[:<port>] [params*]
8276 Set a template to initialize servers with shared parameters.
8277 The names of these servers are built from <prefix> and <num | range> parameters.
8278 May be used in sections : defaults | frontend | listen | backend
8279 no | no | yes | yes
8280
8281 Arguments:
8282 <prefix> A prefix for the server names to be built.
8283
8284 <num | range>
8285 If <num> is provided, this template initializes <num> servers
8286 with 1 up to <num> as server name suffixes. A range of numbers
8287 <num_low>-<num_high> may also be used to use <num_low> up to
8288 <num_high> as server name suffixes.
8289
8290 <fqdn> A FQDN for all the servers this template initializes.
8291
8292 <port> Same meaning as "server" <port> argument (see "server" keyword).
8293
8294 <params*>
8295 Remaining server parameters among all those supported by "server"
8296 keyword.
8297
8298 Examples:
8299 # Initializes 3 servers with srv1, srv2 and srv3 as names,
8300 # google.com as FQDN, and health-check enabled.
8301 server-template srv 1-3 google.com:80 check
8302
8303 # or
8304 server-template srv 3 google.com:80 check
8305
8306 # would be equivalent to:
8307 server srv1 google.com:80 check
8308 server srv2 google.com:80 check
8309 server srv3 google.com:80 check
8310
8311
8312
Willy Tarreaueabeafa2008-01-16 16:17:06 +01008313source <addr>[:<port>] [usesrc { <addr2>[:<port2>] | client | clientip } ]
Willy Tarreaubce70882009-09-07 11:51:47 +02008314source <addr>[:<port>] [usesrc { <addr2>[:<port2>] | hdr_ip(<hdr>[,<occ>]) } ]
Willy Tarreaud53f96b2009-02-04 18:46:54 +01008315source <addr>[:<port>] [interface <name>]
Willy Tarreaueabeafa2008-01-16 16:17:06 +01008316 Set the source address for outgoing connections
8317 May be used in sections : defaults | frontend | listen | backend
8318 yes | no | yes | yes
8319 Arguments :
8320 <addr> is the IPv4 address HAProxy will bind to before connecting to a
8321 server. This address is also used as a source for health checks.
Willy Tarreau24709282013-03-10 21:32:12 +01008322
Willy Tarreaueabeafa2008-01-16 16:17:06 +01008323 The default value of 0.0.0.0 means that the system will select
Willy Tarreau24709282013-03-10 21:32:12 +01008324 the most appropriate address to reach its destination. Optionally
8325 an address family prefix may be used before the address to force
8326 the family regardless of the address format, which can be useful
8327 to specify a path to a unix socket with no slash ('/'). Currently
8328 supported prefixes are :
8329 - 'ipv4@' -> address is always IPv4
8330 - 'ipv6@' -> address is always IPv6
8331 - 'unix@' -> address is a path to a local unix socket
Willy Tarreauccfccef2014-05-10 01:49:15 +02008332 - 'abns@' -> address is in abstract namespace (Linux only)
Cyril Bonté307ee1e2015-09-28 23:16:06 +02008333 You may want to reference some environment variables in the
8334 address parameter, see section 2.3 about environment variables.
Willy Tarreaueabeafa2008-01-16 16:17:06 +01008335
8336 <port> is an optional port. It is normally not needed but may be useful
8337 in some very specific contexts. The default value of zero means
Willy Tarreauc6f4ce82009-06-10 11:09:37 +02008338 the system will select a free port. Note that port ranges are not
8339 supported in the backend. If you want to force port ranges, you
8340 have to specify them on each "server" line.
Willy Tarreaueabeafa2008-01-16 16:17:06 +01008341
8342 <addr2> is the IP address to present to the server when connections are
8343 forwarded in full transparent proxy mode. This is currently only
8344 supported on some patched Linux kernels. When this address is
8345 specified, clients connecting to the server will be presented
8346 with this address, while health checks will still use the address
8347 <addr>.
8348
8349 <port2> is the optional port to present to the server when connections
8350 are forwarded in full transparent proxy mode (see <addr2> above).
8351 The default value of zero means the system will select a free
8352 port.
8353
Willy Tarreaubce70882009-09-07 11:51:47 +02008354 <hdr> is the name of a HTTP header in which to fetch the IP to bind to.
8355 This is the name of a comma-separated header list which can
8356 contain multiple IP addresses. By default, the last occurrence is
8357 used. This is designed to work with the X-Forwarded-For header
Baptiste Assmannea3e73b2013-02-02 23:47:49 +01008358 and to automatically bind to the client's IP address as seen
Willy Tarreaubce70882009-09-07 11:51:47 +02008359 by previous proxy, typically Stunnel. In order to use another
8360 occurrence from the last one, please see the <occ> parameter
8361 below. When the header (or occurrence) is not found, no binding
8362 is performed so that the proxy's default IP address is used. Also
8363 keep in mind that the header name is case insensitive, as for any
8364 HTTP header.
8365
8366 <occ> is the occurrence number of a value to be used in a multi-value
8367 header. This is to be used in conjunction with "hdr_ip(<hdr>)",
Jamie Gloudonaaa21002012-08-25 00:18:33 -04008368 in order to specify which occurrence to use for the source IP
Willy Tarreaubce70882009-09-07 11:51:47 +02008369 address. Positive values indicate a position from the first
8370 occurrence, 1 being the first one. Negative values indicate
8371 positions relative to the last one, -1 being the last one. This
8372 is helpful for situations where an X-Forwarded-For header is set
8373 at the entry point of an infrastructure and must be used several
8374 proxy layers away. When this value is not specified, -1 is
8375 assumed. Passing a zero here disables the feature.
8376
Willy Tarreaud53f96b2009-02-04 18:46:54 +01008377 <name> is an optional interface name to which to bind to for outgoing
8378 traffic. On systems supporting this features (currently, only
8379 Linux), this allows one to bind all traffic to the server to
8380 this interface even if it is not the one the system would select
8381 based on routing tables. This should be used with extreme care.
8382 Note that using this option requires root privileges.
8383
Willy Tarreaueabeafa2008-01-16 16:17:06 +01008384 The "source" keyword is useful in complex environments where a specific
8385 address only is allowed to connect to the servers. It may be needed when a
8386 private address must be used through a public gateway for instance, and it is
8387 known that the system cannot determine the adequate source address by itself.
8388
8389 An extension which is available on certain patched Linux kernels may be used
8390 through the "usesrc" optional keyword. It makes it possible to connect to the
8391 servers with an IP address which does not belong to the system itself. This
8392 is called "full transparent proxy mode". For this to work, the destination
8393 servers have to route their traffic back to this address through the machine
8394 running HAProxy, and IP forwarding must generally be enabled on this machine.
8395
8396 In this "full transparent proxy" mode, it is possible to force a specific IP
8397 address to be presented to the servers. This is not much used in fact. A more
8398 common use is to tell HAProxy to present the client's IP address. For this,
8399 there are two methods :
8400
8401 - present the client's IP and port addresses. This is the most transparent
8402 mode, but it can cause problems when IP connection tracking is enabled on
8403 the machine, because a same connection may be seen twice with different
8404 states. However, this solution presents the huge advantage of not
8405 limiting the system to the 64k outgoing address+port couples, because all
8406 of the client ranges may be used.
8407
8408 - present only the client's IP address and select a spare port. This
8409 solution is still quite elegant but slightly less transparent (downstream
8410 firewalls logs will not match upstream's). It also presents the downside
8411 of limiting the number of concurrent connections to the usual 64k ports.
8412 However, since the upstream and downstream ports are different, local IP
8413 connection tracking on the machine will not be upset by the reuse of the
8414 same session.
8415
Willy Tarreaueabeafa2008-01-16 16:17:06 +01008416 This option sets the default source for all servers in the backend. It may
8417 also be specified in a "defaults" section. Finer source address specification
8418 is possible at the server level using the "source" server option. Refer to
Willy Tarreauc57f0e22009-05-10 13:12:33 +02008419 section 5 for more information.
Willy Tarreaueabeafa2008-01-16 16:17:06 +01008420
Baptiste Assmann91bd3372015-07-17 21:59:42 +02008421 In order to work, "usesrc" requires root privileges.
8422
Willy Tarreaueabeafa2008-01-16 16:17:06 +01008423 Examples :
8424 backend private
8425 # Connect to the servers using our 192.168.1.200 source address
8426 source 192.168.1.200
8427
8428 backend transparent_ssl1
8429 # Connect to the SSL farm from the client's source address
8430 source 192.168.1.200 usesrc clientip
8431
8432 backend transparent_ssl2
8433 # Connect to the SSL farm from the client's source address and port
8434 # not recommended if IP conntrack is present on the local machine.
8435 source 192.168.1.200 usesrc client
8436
8437 backend transparent_ssl3
8438 # Connect to the SSL farm from the client's source address. It
8439 # is more conntrack-friendly.
8440 source 192.168.1.200 usesrc clientip
8441
8442 backend transparent_smtp
8443 # Connect to the SMTP farm from the client's source address/port
8444 # with Tproxy version 4.
8445 source 0.0.0.0 usesrc clientip
8446
Willy Tarreaubce70882009-09-07 11:51:47 +02008447 backend transparent_http
8448 # Connect to the servers using the client's IP as seen by previous
8449 # proxy.
8450 source 0.0.0.0 usesrc hdr_ip(x-forwarded-for,-1)
8451
Willy Tarreauc57f0e22009-05-10 13:12:33 +02008452 See also : the "source" server option in section 5, the Tproxy patches for
Willy Tarreaueabeafa2008-01-16 16:17:06 +01008453 the Linux kernel on www.balabit.com, the "bind" keyword.
8454
Willy Tarreau844e3c52008-01-11 16:28:18 +01008455
Cyril Bonté66c327d2010-10-12 00:14:37 +02008456stats admin { if | unless } <cond>
8457 Enable statistics admin level if/unless a condition is matched
8458 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaued2119c2014-04-24 22:10:39 +02008459 no | yes | yes | yes
Cyril Bonté66c327d2010-10-12 00:14:37 +02008460
8461 This statement enables the statistics admin level if/unless a condition is
8462 matched.
8463
8464 The admin level allows to enable/disable servers from the web interface. By
8465 default, statistics page is read-only for security reasons.
8466
Cyril Bonté02ff8ef2010-12-14 22:48:49 +01008467 Note : Consider not using this feature in multi-process mode (nbproc > 1)
8468 unless you know what you do : memory is not shared between the
Davor Ocelice9ed2812017-12-25 17:49:28 +01008469 processes, which can result in random behaviors.
Cyril Bonté02ff8ef2010-12-14 22:48:49 +01008470
Cyril Bonté23b39d92011-02-10 22:54:44 +01008471 Currently, the POST request is limited to the buffer size minus the reserved
8472 buffer space, which means that if the list of servers is too long, the
8473 request won't be processed. It is recommended to alter few servers at a
8474 time.
Cyril Bonté66c327d2010-10-12 00:14:37 +02008475
8476 Example :
8477 # statistics admin level only for localhost
8478 backend stats_localhost
8479 stats enable
8480 stats admin if LOCALHOST
8481
8482 Example :
8483 # statistics admin level always enabled because of the authentication
8484 backend stats_auth
8485 stats enable
8486 stats auth admin:AdMiN123
8487 stats admin if TRUE
8488
8489 Example :
8490 # statistics admin level depends on the authenticated user
8491 userlist stats-auth
8492 group admin users admin
8493 user admin insecure-password AdMiN123
8494 group readonly users haproxy
8495 user haproxy insecure-password haproxy
8496
8497 backend stats_auth
8498 stats enable
8499 acl AUTH http_auth(stats-auth)
8500 acl AUTH_ADMIN http_auth_group(stats-auth) admin
8501 stats http-request auth unless AUTH
8502 stats admin if AUTH_ADMIN
8503
Cyril Bonté02ff8ef2010-12-14 22:48:49 +01008504 See also : "stats enable", "stats auth", "stats http-request", "nbproc",
8505 "bind-process", section 3.4 about userlists and section 7 about
8506 ACL usage.
Cyril Bonté66c327d2010-10-12 00:14:37 +02008507
8508
Willy Tarreaueabeafa2008-01-16 16:17:06 +01008509stats auth <user>:<passwd>
8510 Enable statistics with authentication and grant access to an account
8511 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaued2119c2014-04-24 22:10:39 +02008512 yes | yes | yes | yes
Willy Tarreaueabeafa2008-01-16 16:17:06 +01008513 Arguments :
8514 <user> is a user name to grant access to
8515
8516 <passwd> is the cleartext password associated to this user
8517
8518 This statement enables statistics with default settings, and restricts access
8519 to declared users only. It may be repeated as many times as necessary to
8520 allow as many users as desired. When a user tries to access the statistics
8521 without a valid account, a "401 Forbidden" response will be returned so that
8522 the browser asks the user to provide a valid user and password. The real
8523 which will be returned to the browser is configurable using "stats realm".
8524
8525 Since the authentication method is HTTP Basic Authentication, the passwords
8526 circulate in cleartext on the network. Thus, it was decided that the
8527 configuration file would also use cleartext passwords to remind the users
Willy Tarreau3c92c5f2011-08-28 09:45:47 +02008528 that those ones should not be sensitive and not shared with any other account.
Willy Tarreaueabeafa2008-01-16 16:17:06 +01008529
8530 It is also possible to reduce the scope of the proxies which appear in the
8531 report using "stats scope".
8532
8533 Though this statement alone is enough to enable statistics reporting, it is
8534 recommended to set all other settings in order to avoid relying on default
8535 unobvious parameters.
8536
8537 Example :
8538 # public access (limited to this backend only)
8539 backend public_www
8540 server srv1 192.168.0.1:80
8541 stats enable
8542 stats hide-version
8543 stats scope .
8544 stats uri /admin?stats
Davor Ocelice9ed2812017-12-25 17:49:28 +01008545 stats realm HAProxy\ Statistics
Willy Tarreaueabeafa2008-01-16 16:17:06 +01008546 stats auth admin1:AdMiN123
8547 stats auth admin2:AdMiN321
8548
8549 # internal monitoring access (unlimited)
8550 backend private_monitoring
8551 stats enable
8552 stats uri /admin?stats
8553 stats refresh 5s
8554
8555 See also : "stats enable", "stats realm", "stats scope", "stats uri"
8556
8557
8558stats enable
8559 Enable statistics reporting with default settings
8560 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaued2119c2014-04-24 22:10:39 +02008561 yes | yes | yes | yes
Willy Tarreaueabeafa2008-01-16 16:17:06 +01008562 Arguments : none
8563
8564 This statement enables statistics reporting with default settings defined
8565 at build time. Unless stated otherwise, these settings are used :
8566 - stats uri : /haproxy?stats
8567 - stats realm : "HAProxy Statistics"
8568 - stats auth : no authentication
8569 - stats scope : no restriction
8570
8571 Though this statement alone is enough to enable statistics reporting, it is
8572 recommended to set all other settings in order to avoid relying on default
8573 unobvious parameters.
8574
8575 Example :
8576 # public access (limited to this backend only)
8577 backend public_www
8578 server srv1 192.168.0.1:80
8579 stats enable
8580 stats hide-version
8581 stats scope .
8582 stats uri /admin?stats
Davor Ocelice9ed2812017-12-25 17:49:28 +01008583 stats realm HAProxy\ Statistics
Willy Tarreaueabeafa2008-01-16 16:17:06 +01008584 stats auth admin1:AdMiN123
8585 stats auth admin2:AdMiN321
8586
8587 # internal monitoring access (unlimited)
8588 backend private_monitoring
8589 stats enable
8590 stats uri /admin?stats
8591 stats refresh 5s
8592
8593 See also : "stats auth", "stats realm", "stats uri"
8594
8595
Willy Tarreaud63335a2010-02-26 12:56:52 +01008596stats hide-version
8597 Enable statistics and hide HAProxy version reporting
Willy Tarreau1d45b7c2009-08-16 10:29:18 +02008598 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaued2119c2014-04-24 22:10:39 +02008599 yes | yes | yes | yes
Willy Tarreaud63335a2010-02-26 12:56:52 +01008600 Arguments : none
Willy Tarreau1d45b7c2009-08-16 10:29:18 +02008601
Willy Tarreaud63335a2010-02-26 12:56:52 +01008602 By default, the stats page reports some useful status information along with
8603 the statistics. Among them is HAProxy's version. However, it is generally
8604 considered dangerous to report precise version to anyone, as it can help them
8605 target known weaknesses with specific attacks. The "stats hide-version"
8606 statement removes the version from the statistics report. This is recommended
8607 for public sites or any site with a weak login/password.
Willy Tarreau1d45b7c2009-08-16 10:29:18 +02008608
Krzysztof Piotr Oledzki48cb2ae2009-10-02 22:51:14 +02008609 Though this statement alone is enough to enable statistics reporting, it is
8610 recommended to set all other settings in order to avoid relying on default
8611 unobvious parameters.
8612
Willy Tarreaud63335a2010-02-26 12:56:52 +01008613 Example :
8614 # public access (limited to this backend only)
8615 backend public_www
8616 server srv1 192.168.0.1:80
Krzysztof Piotr Oledzki48cb2ae2009-10-02 22:51:14 +02008617 stats enable
Willy Tarreaud63335a2010-02-26 12:56:52 +01008618 stats hide-version
8619 stats scope .
8620 stats uri /admin?stats
Davor Ocelice9ed2812017-12-25 17:49:28 +01008621 stats realm HAProxy\ Statistics
Willy Tarreaud63335a2010-02-26 12:56:52 +01008622 stats auth admin1:AdMiN123
8623 stats auth admin2:AdMiN321
Willy Tarreau1d45b7c2009-08-16 10:29:18 +02008624
Willy Tarreau1d45b7c2009-08-16 10:29:18 +02008625 # internal monitoring access (unlimited)
8626 backend private_monitoring
8627 stats enable
Willy Tarreaud63335a2010-02-26 12:56:52 +01008628 stats uri /admin?stats
8629 stats refresh 5s
Krzysztof Piotr Oledzki15514c22010-01-04 16:03:09 +01008630
Willy Tarreaud63335a2010-02-26 12:56:52 +01008631 See also : "stats auth", "stats enable", "stats realm", "stats uri"
Willy Tarreau1d45b7c2009-08-16 10:29:18 +02008632
Willy Tarreau983e01e2010-01-11 18:42:06 +01008633
Cyril Bonté2be1b3f2010-09-30 23:46:30 +02008634stats http-request { allow | deny | auth [realm <realm>] }
8635 [ { if | unless } <condition> ]
8636 Access control for statistics
8637
8638 May be used in sections: defaults | frontend | listen | backend
8639 no | no | yes | yes
8640
8641 As "http-request", these set of options allow to fine control access to
8642 statistics. Each option may be followed by if/unless and acl.
8643 First option with matched condition (or option without condition) is final.
8644 For "deny" a 403 error will be returned, for "allow" normal processing is
8645 performed, for "auth" a 401/407 error code is returned so the client
8646 should be asked to enter a username and password.
8647
8648 There is no fixed limit to the number of http-request statements per
8649 instance.
8650
8651 See also : "http-request", section 3.4 about userlists and section 7
8652 about ACL usage.
8653
8654
Willy Tarreaueabeafa2008-01-16 16:17:06 +01008655stats realm <realm>
8656 Enable statistics and set authentication realm
8657 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaued2119c2014-04-24 22:10:39 +02008658 yes | yes | yes | yes
Willy Tarreaueabeafa2008-01-16 16:17:06 +01008659 Arguments :
8660 <realm> is the name of the HTTP Basic Authentication realm reported to
8661 the browser. The browser uses it to display it in the pop-up
8662 inviting the user to enter a valid username and password.
8663
8664 The realm is read as a single word, so any spaces in it should be escaped
8665 using a backslash ('\').
8666
8667 This statement is useful only in conjunction with "stats auth" since it is
8668 only related to authentication.
8669
8670 Though this statement alone is enough to enable statistics reporting, it is
8671 recommended to set all other settings in order to avoid relying on default
8672 unobvious parameters.
8673
8674 Example :
8675 # public access (limited to this backend only)
8676 backend public_www
8677 server srv1 192.168.0.1:80
8678 stats enable
8679 stats hide-version
8680 stats scope .
8681 stats uri /admin?stats
Davor Ocelice9ed2812017-12-25 17:49:28 +01008682 stats realm HAProxy\ Statistics
Willy Tarreaueabeafa2008-01-16 16:17:06 +01008683 stats auth admin1:AdMiN123
8684 stats auth admin2:AdMiN321
8685
8686 # internal monitoring access (unlimited)
8687 backend private_monitoring
8688 stats enable
8689 stats uri /admin?stats
8690 stats refresh 5s
8691
8692 See also : "stats auth", "stats enable", "stats uri"
8693
8694
8695stats refresh <delay>
8696 Enable statistics with automatic refresh
8697 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaued2119c2014-04-24 22:10:39 +02008698 yes | yes | yes | yes
Willy Tarreaueabeafa2008-01-16 16:17:06 +01008699 Arguments :
8700 <delay> is the suggested refresh delay, specified in seconds, which will
8701 be returned to the browser consulting the report page. While the
8702 browser is free to apply any delay, it will generally respect it
8703 and refresh the page this every seconds. The refresh interval may
8704 be specified in any other non-default time unit, by suffixing the
8705 unit after the value, as explained at the top of this document.
8706
8707 This statement is useful on monitoring displays with a permanent page
8708 reporting the load balancer's activity. When set, the HTML report page will
8709 include a link "refresh"/"stop refresh" so that the user can select whether
8710 he wants automatic refresh of the page or not.
8711
8712 Though this statement alone is enough to enable statistics reporting, it is
8713 recommended to set all other settings in order to avoid relying on default
8714 unobvious parameters.
8715
8716 Example :
8717 # public access (limited to this backend only)
8718 backend public_www
8719 server srv1 192.168.0.1:80
8720 stats enable
8721 stats hide-version
8722 stats scope .
8723 stats uri /admin?stats
Davor Ocelice9ed2812017-12-25 17:49:28 +01008724 stats realm HAProxy\ Statistics
Willy Tarreaueabeafa2008-01-16 16:17:06 +01008725 stats auth admin1:AdMiN123
8726 stats auth admin2:AdMiN321
8727
8728 # internal monitoring access (unlimited)
8729 backend private_monitoring
8730 stats enable
8731 stats uri /admin?stats
8732 stats refresh 5s
8733
8734 See also : "stats auth", "stats enable", "stats realm", "stats uri"
8735
8736
8737stats scope { <name> | "." }
8738 Enable statistics and limit access scope
8739 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaued2119c2014-04-24 22:10:39 +02008740 yes | yes | yes | yes
Willy Tarreaueabeafa2008-01-16 16:17:06 +01008741 Arguments :
8742 <name> is the name of a listen, frontend or backend section to be
8743 reported. The special name "." (a single dot) designates the
8744 section in which the statement appears.
8745
8746 When this statement is specified, only the sections enumerated with this
8747 statement will appear in the report. All other ones will be hidden. This
8748 statement may appear as many times as needed if multiple sections need to be
8749 reported. Please note that the name checking is performed as simple string
8750 comparisons, and that it is never checked that a give section name really
8751 exists.
8752
8753 Though this statement alone is enough to enable statistics reporting, it is
8754 recommended to set all other settings in order to avoid relying on default
8755 unobvious parameters.
8756
8757 Example :
8758 # public access (limited to this backend only)
8759 backend public_www
8760 server srv1 192.168.0.1:80
8761 stats enable
8762 stats hide-version
8763 stats scope .
8764 stats uri /admin?stats
Davor Ocelice9ed2812017-12-25 17:49:28 +01008765 stats realm HAProxy\ Statistics
Willy Tarreaueabeafa2008-01-16 16:17:06 +01008766 stats auth admin1:AdMiN123
8767 stats auth admin2:AdMiN321
8768
8769 # internal monitoring access (unlimited)
8770 backend private_monitoring
8771 stats enable
8772 stats uri /admin?stats
8773 stats refresh 5s
8774
8775 See also : "stats auth", "stats enable", "stats realm", "stats uri"
8776
Willy Tarreaud63335a2010-02-26 12:56:52 +01008777
Willy Tarreauc9705a12010-07-27 20:05:50 +02008778stats show-desc [ <desc> ]
Willy Tarreaud63335a2010-02-26 12:56:52 +01008779 Enable reporting of a description on the statistics page.
8780 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaued2119c2014-04-24 22:10:39 +02008781 yes | yes | yes | yes
Willy Tarreaud63335a2010-02-26 12:56:52 +01008782
Willy Tarreauc9705a12010-07-27 20:05:50 +02008783 <desc> is an optional description to be reported. If unspecified, the
Willy Tarreaud63335a2010-02-26 12:56:52 +01008784 description from global section is automatically used instead.
8785
8786 This statement is useful for users that offer shared services to their
8787 customers, where node or description should be different for each customer.
8788
8789 Though this statement alone is enough to enable statistics reporting, it is
8790 recommended to set all other settings in order to avoid relying on default
Davor Ocelice9ed2812017-12-25 17:49:28 +01008791 unobvious parameters. By default description is not shown.
Willy Tarreaud63335a2010-02-26 12:56:52 +01008792
8793 Example :
8794 # internal monitoring access (unlimited)
8795 backend private_monitoring
8796 stats enable
8797 stats show-desc Master node for Europe, Asia, Africa
8798 stats uri /admin?stats
8799 stats refresh 5s
8800
8801 See also: "show-node", "stats enable", "stats uri" and "description" in
8802 global section.
8803
8804
8805stats show-legends
Willy Tarreaued2119c2014-04-24 22:10:39 +02008806 Enable reporting additional information on the statistics page
8807 May be used in sections : defaults | frontend | listen | backend
8808 yes | yes | yes | yes
8809 Arguments : none
8810
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03008811 Enable reporting additional information on the statistics page :
Willy Tarreaud63335a2010-02-26 12:56:52 +01008812 - cap: capabilities (proxy)
8813 - mode: one of tcp, http or health (proxy)
8814 - id: SNMP ID (proxy, socket, server)
8815 - IP (socket, server)
8816 - cookie (backend, server)
8817
8818 Though this statement alone is enough to enable statistics reporting, it is
8819 recommended to set all other settings in order to avoid relying on default
Davor Ocelice9ed2812017-12-25 17:49:28 +01008820 unobvious parameters. Default behavior is not to show this information.
Willy Tarreaud63335a2010-02-26 12:56:52 +01008821
8822 See also: "stats enable", "stats uri".
8823
8824
8825stats show-node [ <name> ]
8826 Enable reporting of a host name on the statistics page.
8827 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaued2119c2014-04-24 22:10:39 +02008828 yes | yes | yes | yes
Willy Tarreaud63335a2010-02-26 12:56:52 +01008829 Arguments:
8830 <name> is an optional name to be reported. If unspecified, the
8831 node name from global section is automatically used instead.
8832
8833 This statement is useful for users that offer shared services to their
8834 customers, where node or description might be different on a stats page
Davor Ocelice9ed2812017-12-25 17:49:28 +01008835 provided for each customer. Default behavior is not to show host name.
Willy Tarreaud63335a2010-02-26 12:56:52 +01008836
8837 Though this statement alone is enough to enable statistics reporting, it is
8838 recommended to set all other settings in order to avoid relying on default
8839 unobvious parameters.
8840
8841 Example:
8842 # internal monitoring access (unlimited)
8843 backend private_monitoring
8844 stats enable
8845 stats show-node Europe-1
8846 stats uri /admin?stats
8847 stats refresh 5s
8848
8849 See also: "show-desc", "stats enable", "stats uri", and "node" in global
8850 section.
8851
Willy Tarreaueabeafa2008-01-16 16:17:06 +01008852
8853stats uri <prefix>
8854 Enable statistics and define the URI prefix to access them
8855 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaued2119c2014-04-24 22:10:39 +02008856 yes | yes | yes | yes
Willy Tarreaueabeafa2008-01-16 16:17:06 +01008857 Arguments :
8858 <prefix> is the prefix of any URI which will be redirected to stats. This
8859 prefix may contain a question mark ('?') to indicate part of a
8860 query string.
8861
8862 The statistics URI is intercepted on the relayed traffic, so it appears as a
8863 page within the normal application. It is strongly advised to ensure that the
8864 selected URI will never appear in the application, otherwise it will never be
8865 possible to reach it in the application.
8866
8867 The default URI compiled in haproxy is "/haproxy?stats", but this may be
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01008868 changed at build time, so it's better to always explicitly specify it here.
Willy Tarreaueabeafa2008-01-16 16:17:06 +01008869 It is generally a good idea to include a question mark in the URI so that
8870 intermediate proxies refrain from caching the results. Also, since any string
8871 beginning with the prefix will be accepted as a stats request, the question
8872 mark helps ensuring that no valid URI will begin with the same words.
8873
8874 It is sometimes very convenient to use "/" as the URI prefix, and put that
8875 statement in a "listen" instance of its own. That makes it easy to dedicate
8876 an address or a port to statistics only.
8877
8878 Though this statement alone is enough to enable statistics reporting, it is
8879 recommended to set all other settings in order to avoid relying on default
8880 unobvious parameters.
8881
8882 Example :
8883 # public access (limited to this backend only)
8884 backend public_www
8885 server srv1 192.168.0.1:80
8886 stats enable
8887 stats hide-version
8888 stats scope .
8889 stats uri /admin?stats
Davor Ocelice9ed2812017-12-25 17:49:28 +01008890 stats realm HAProxy\ Statistics
Willy Tarreaueabeafa2008-01-16 16:17:06 +01008891 stats auth admin1:AdMiN123
8892 stats auth admin2:AdMiN321
8893
8894 # internal monitoring access (unlimited)
8895 backend private_monitoring
8896 stats enable
8897 stats uri /admin?stats
8898 stats refresh 5s
8899
8900 See also : "stats auth", "stats enable", "stats realm"
8901
8902
Willy Tarreaud63335a2010-02-26 12:56:52 +01008903stick match <pattern> [table <table>] [{if | unless} <cond>]
8904 Define a request pattern matching condition to stick a user to a server
Willy Tarreaueabeafa2008-01-16 16:17:06 +01008905 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaud63335a2010-02-26 12:56:52 +01008906 no | no | yes | yes
Willy Tarreaub937b7e2010-01-12 15:27:54 +01008907
8908 Arguments :
Willy Tarreaube722a22014-06-13 16:31:59 +02008909 <pattern> is a sample expression rule as described in section 7.3. It
Willy Tarreaub937b7e2010-01-12 15:27:54 +01008910 describes what elements of the incoming request or connection
Davor Ocelice9ed2812017-12-25 17:49:28 +01008911 will be analyzed in the hope to find a matching entry in a
Willy Tarreaub937b7e2010-01-12 15:27:54 +01008912 stickiness table. This rule is mandatory.
8913
8914 <table> is an optional stickiness table name. If unspecified, the same
8915 backend's table is used. A stickiness table is declared using
8916 the "stick-table" statement.
8917
8918 <cond> is an optional matching condition. It makes it possible to match
8919 on a certain criterion only when other conditions are met (or
8920 not met). For instance, it could be used to match on a source IP
8921 address except when a request passes through a known proxy, in
8922 which case we'd match on a header containing that IP address.
8923
8924 Some protocols or applications require complex stickiness rules and cannot
8925 always simply rely on cookies nor hashing. The "stick match" statement
8926 describes a rule to extract the stickiness criterion from an incoming request
8927 or connection. See section 7 for a complete list of possible patterns and
8928 transformation rules.
8929
8930 The table has to be declared using the "stick-table" statement. It must be of
8931 a type compatible with the pattern. By default it is the one which is present
8932 in the same backend. It is possible to share a table with other backends by
8933 referencing it using the "table" keyword. If another table is referenced,
8934 the server's ID inside the backends are used. By default, all server IDs
8935 start at 1 in each backend, so the server ordering is enough. But in case of
8936 doubt, it is highly recommended to force server IDs using their "id" setting.
8937
8938 It is possible to restrict the conditions where a "stick match" statement
8939 will apply, using "if" or "unless" followed by a condition. See section 7 for
8940 ACL based conditions.
8941
8942 There is no limit on the number of "stick match" statements. The first that
8943 applies and matches will cause the request to be directed to the same server
8944 as was used for the request which created the entry. That way, multiple
8945 matches can be used as fallbacks.
8946
8947 The stick rules are checked after the persistence cookies, so they will not
8948 affect stickiness if a cookie has already been used to select a server. That
8949 way, it becomes very easy to insert cookies and match on IP addresses in
8950 order to maintain stickiness between HTTP and HTTPS.
8951
Cyril Bonté02ff8ef2010-12-14 22:48:49 +01008952 Note : Consider not using this feature in multi-process mode (nbproc > 1)
8953 unless you know what you do : memory is not shared between the
Davor Ocelice9ed2812017-12-25 17:49:28 +01008954 processes, which can result in random behaviors.
Cyril Bonté02ff8ef2010-12-14 22:48:49 +01008955
Willy Tarreaub937b7e2010-01-12 15:27:54 +01008956 Example :
8957 # forward SMTP users to the same server they just used for POP in the
8958 # last 30 minutes
8959 backend pop
8960 mode tcp
8961 balance roundrobin
8962 stick store-request src
8963 stick-table type ip size 200k expire 30m
8964 server s1 192.168.1.1:110
8965 server s2 192.168.1.1:110
8966
8967 backend smtp
8968 mode tcp
8969 balance roundrobin
8970 stick match src table pop
8971 server s1 192.168.1.1:25
8972 server s2 192.168.1.1:25
8973
Cyril Bonté02ff8ef2010-12-14 22:48:49 +01008974 See also : "stick-table", "stick on", "nbproc", "bind-process" and section 7
Willy Tarreaube722a22014-06-13 16:31:59 +02008975 about ACLs and samples fetching.
Willy Tarreaub937b7e2010-01-12 15:27:54 +01008976
8977
8978stick on <pattern> [table <table>] [{if | unless} <condition>]
8979 Define a request pattern to associate a user to a server
8980 May be used in sections : defaults | frontend | listen | backend
8981 no | no | yes | yes
8982
8983 Note : This form is exactly equivalent to "stick match" followed by
8984 "stick store-request", all with the same arguments. Please refer
8985 to both keywords for details. It is only provided as a convenience
8986 for writing more maintainable configurations.
8987
Cyril Bonté02ff8ef2010-12-14 22:48:49 +01008988 Note : Consider not using this feature in multi-process mode (nbproc > 1)
8989 unless you know what you do : memory is not shared between the
Davor Ocelice9ed2812017-12-25 17:49:28 +01008990 processes, which can result in random behaviors.
Cyril Bonté02ff8ef2010-12-14 22:48:49 +01008991
Willy Tarreaub937b7e2010-01-12 15:27:54 +01008992 Examples :
8993 # The following form ...
Willy Tarreauec579d82010-02-26 19:15:04 +01008994 stick on src table pop if !localhost
Willy Tarreaub937b7e2010-01-12 15:27:54 +01008995
8996 # ...is strictly equivalent to this one :
8997 stick match src table pop if !localhost
8998 stick store-request src table pop if !localhost
8999
9000
9001 # Use cookie persistence for HTTP, and stick on source address for HTTPS as
9002 # well as HTTP without cookie. Share the same table between both accesses.
9003 backend http
9004 mode http
9005 balance roundrobin
9006 stick on src table https
9007 cookie SRV insert indirect nocache
9008 server s1 192.168.1.1:80 cookie s1
9009 server s2 192.168.1.1:80 cookie s2
9010
9011 backend https
9012 mode tcp
9013 balance roundrobin
9014 stick-table type ip size 200k expire 30m
9015 stick on src
9016 server s1 192.168.1.1:443
9017 server s2 192.168.1.1:443
9018
Cyril Bonté02ff8ef2010-12-14 22:48:49 +01009019 See also : "stick match", "stick store-request", "nbproc" and "bind-process".
Willy Tarreaub937b7e2010-01-12 15:27:54 +01009020
9021
9022stick store-request <pattern> [table <table>] [{if | unless} <condition>]
9023 Define a request pattern used to create an entry in a stickiness table
9024 May be used in sections : defaults | frontend | listen | backend
9025 no | no | yes | yes
9026
9027 Arguments :
Willy Tarreaube722a22014-06-13 16:31:59 +02009028 <pattern> is a sample expression rule as described in section 7.3. It
Willy Tarreaub937b7e2010-01-12 15:27:54 +01009029 describes what elements of the incoming request or connection
Davor Ocelice9ed2812017-12-25 17:49:28 +01009030 will be analyzed, extracted and stored in the table once a
Willy Tarreaub937b7e2010-01-12 15:27:54 +01009031 server is selected.
9032
9033 <table> is an optional stickiness table name. If unspecified, the same
9034 backend's table is used. A stickiness table is declared using
9035 the "stick-table" statement.
9036
9037 <cond> is an optional storage condition. It makes it possible to store
9038 certain criteria only when some conditions are met (or not met).
9039 For instance, it could be used to store the source IP address
9040 except when the request passes through a known proxy, in which
9041 case we'd store a converted form of a header containing that IP
9042 address.
9043
9044 Some protocols or applications require complex stickiness rules and cannot
9045 always simply rely on cookies nor hashing. The "stick store-request" statement
9046 describes a rule to decide what to extract from the request and when to do
9047 it, in order to store it into a stickiness table for further requests to
9048 match it using the "stick match" statement. Obviously the extracted part must
9049 make sense and have a chance to be matched in a further request. Storing a
9050 client's IP address for instance often makes sense. Storing an ID found in a
9051 URL parameter also makes sense. Storing a source port will almost never make
9052 any sense because it will be randomly matched. See section 7 for a complete
9053 list of possible patterns and transformation rules.
9054
9055 The table has to be declared using the "stick-table" statement. It must be of
9056 a type compatible with the pattern. By default it is the one which is present
9057 in the same backend. It is possible to share a table with other backends by
9058 referencing it using the "table" keyword. If another table is referenced,
9059 the server's ID inside the backends are used. By default, all server IDs
9060 start at 1 in each backend, so the server ordering is enough. But in case of
9061 doubt, it is highly recommended to force server IDs using their "id" setting.
9062
9063 It is possible to restrict the conditions where a "stick store-request"
9064 statement will apply, using "if" or "unless" followed by a condition. This
9065 condition will be evaluated while parsing the request, so any criteria can be
9066 used. See section 7 for ACL based conditions.
9067
9068 There is no limit on the number of "stick store-request" statements, but
9069 there is a limit of 8 simultaneous stores per request or response. This
9070 makes it possible to store up to 8 criteria, all extracted from either the
9071 request or the response, regardless of the number of rules. Only the 8 first
9072 ones which match will be kept. Using this, it is possible to feed multiple
9073 tables at once in the hope to increase the chance to recognize a user on
Willy Tarreau9667a802013-12-09 12:52:13 +01009074 another protocol or access method. Using multiple store-request rules with
9075 the same table is possible and may be used to find the best criterion to rely
9076 on, by arranging the rules by decreasing preference order. Only the first
9077 extracted criterion for a given table will be stored. All subsequent store-
9078 request rules referencing the same table will be skipped and their ACLs will
9079 not be evaluated.
Willy Tarreaub937b7e2010-01-12 15:27:54 +01009080
9081 The "store-request" rules are evaluated once the server connection has been
9082 established, so that the table will contain the real server that processed
9083 the request.
9084
Cyril Bonté02ff8ef2010-12-14 22:48:49 +01009085 Note : Consider not using this feature in multi-process mode (nbproc > 1)
9086 unless you know what you do : memory is not shared between the
Davor Ocelice9ed2812017-12-25 17:49:28 +01009087 processes, which can result in random behaviors.
Cyril Bonté02ff8ef2010-12-14 22:48:49 +01009088
Willy Tarreaub937b7e2010-01-12 15:27:54 +01009089 Example :
9090 # forward SMTP users to the same server they just used for POP in the
9091 # last 30 minutes
9092 backend pop
9093 mode tcp
9094 balance roundrobin
9095 stick store-request src
9096 stick-table type ip size 200k expire 30m
9097 server s1 192.168.1.1:110
9098 server s2 192.168.1.1:110
9099
9100 backend smtp
9101 mode tcp
9102 balance roundrobin
9103 stick match src table pop
9104 server s1 192.168.1.1:25
9105 server s2 192.168.1.1:25
9106
Cyril Bonté02ff8ef2010-12-14 22:48:49 +01009107 See also : "stick-table", "stick on", "nbproc", "bind-process" and section 7
Willy Tarreaube722a22014-06-13 16:31:59 +02009108 about ACLs and sample fetching.
Willy Tarreaub937b7e2010-01-12 15:27:54 +01009109
9110
Emeric Brun7c6b82e2010-09-24 16:34:28 +02009111stick-table type {ip | integer | string [len <length>] | binary [len <length>]}
Emeric Brunf099e792010-09-27 12:05:28 +02009112 size <size> [expire <expire>] [nopurge] [peers <peersect>]
9113 [store <data_type>]*
Godbach64cef792013-12-04 16:08:22 +08009114 Configure the stickiness table for the current section
Willy Tarreaub937b7e2010-01-12 15:27:54 +01009115 May be used in sections : defaults | frontend | listen | backend
Willy Tarreauc00cdc22010-06-06 16:48:26 +02009116 no | yes | yes | yes
Willy Tarreaub937b7e2010-01-12 15:27:54 +01009117
9118 Arguments :
9119 ip a table declared with "type ip" will only store IPv4 addresses.
9120 This form is very compact (about 50 bytes per entry) and allows
9121 very fast entry lookup and stores with almost no overhead. This
9122 is mainly used to store client source IP addresses.
9123
David du Colombier9a6d3c92011-03-17 10:40:24 +01009124 ipv6 a table declared with "type ipv6" will only store IPv6 addresses.
9125 This form is very compact (about 60 bytes per entry) and allows
9126 very fast entry lookup and stores with almost no overhead. This
9127 is mainly used to store client source IP addresses.
9128
Willy Tarreaub937b7e2010-01-12 15:27:54 +01009129 integer a table declared with "type integer" will store 32bit integers
9130 which can represent a client identifier found in a request for
9131 instance.
9132
9133 string a table declared with "type string" will store substrings of up
9134 to <len> characters. If the string provided by the pattern
9135 extractor is larger than <len>, it will be truncated before
9136 being stored. During matching, at most <len> characters will be
9137 compared between the string in the table and the extracted
9138 pattern. When not specified, the string is automatically limited
Emeric Brun7c6b82e2010-09-24 16:34:28 +02009139 to 32 characters.
9140
9141 binary a table declared with "type binary" will store binary blocks
9142 of <len> bytes. If the block provided by the pattern
9143 extractor is larger than <len>, it will be truncated before
Willy Tarreaube722a22014-06-13 16:31:59 +02009144 being stored. If the block provided by the sample expression
Emeric Brun7c6b82e2010-09-24 16:34:28 +02009145 is shorter than <len>, it will be padded by 0. When not
9146 specified, the block is automatically limited to 32 bytes.
Willy Tarreaub937b7e2010-01-12 15:27:54 +01009147
9148 <length> is the maximum number of characters that will be stored in a
Emeric Brun7c6b82e2010-09-24 16:34:28 +02009149 "string" type table (See type "string" above). Or the number
9150 of bytes of the block in "binary" type table. Be careful when
Willy Tarreaub937b7e2010-01-12 15:27:54 +01009151 changing this parameter as memory usage will proportionally
9152 increase.
9153
9154 <size> is the maximum number of entries that can fit in the table. This
Cyril Bonté78caf842010-03-10 22:41:43 +01009155 value directly impacts memory usage. Count approximately
9156 50 bytes per entry, plus the size of a string if any. The size
9157 supports suffixes "k", "m", "g" for 2^10, 2^20 and 2^30 factors.
Willy Tarreaub937b7e2010-01-12 15:27:54 +01009158
9159 [nopurge] indicates that we refuse to purge older entries when the table
9160 is full. When not specified and the table is full when haproxy
9161 wants to store an entry in it, it will flush a few of the oldest
9162 entries in order to release some space for the new ones. This is
Davor Ocelice9ed2812017-12-25 17:49:28 +01009163 most often the desired behavior. In some specific cases, it
Willy Tarreaub937b7e2010-01-12 15:27:54 +01009164 be desirable to refuse new entries instead of purging the older
9165 ones. That may be the case when the amount of data to store is
9166 far above the hardware limits and we prefer not to offer access
9167 to new clients than to reject the ones already connected. When
9168 using this parameter, be sure to properly set the "expire"
9169 parameter (see below).
9170
Emeric Brunf099e792010-09-27 12:05:28 +02009171 <peersect> is the name of the peers section to use for replication. Entries
9172 which associate keys to server IDs are kept synchronized with
9173 the remote peers declared in this section. All entries are also
9174 automatically learned from the local peer (old process) during a
9175 soft restart.
9176
Willy Tarreau1abc6732015-05-01 19:21:02 +02009177 NOTE : each peers section may be referenced only by tables
9178 belonging to the same unique process.
Cyril Bonté02ff8ef2010-12-14 22:48:49 +01009179
Willy Tarreaub937b7e2010-01-12 15:27:54 +01009180 <expire> defines the maximum duration of an entry in the table since it
9181 was last created, refreshed or matched. The expiration delay is
9182 defined using the standard time format, similarly as the various
9183 timeouts. The maximum duration is slightly above 24 days. See
Jarno Huuskonene0ee0be2017-07-04 10:35:12 +03009184 section 2.4 for more information. If this delay is not specified,
Cyril Bontédc4d9032012-04-08 21:57:39 +02009185 the session won't automatically expire, but older entries will
Willy Tarreaub937b7e2010-01-12 15:27:54 +01009186 be removed once full. Be sure not to use the "nopurge" parameter
9187 if not expiration delay is specified.
9188
Willy Tarreau08d5f982010-06-06 13:34:54 +02009189 <data_type> is used to store additional information in the stick-table. This
9190 may be used by ACLs in order to control various criteria related
9191 to the activity of the client matching the stick-table. For each
9192 item specified here, the size of each entry will be inflated so
Willy Tarreauc9705a12010-07-27 20:05:50 +02009193 that the additional data can fit. Several data types may be
9194 stored with an entry. Multiple data types may be specified after
9195 the "store" keyword, as a comma-separated list. Alternatively,
9196 it is possible to repeat the "store" keyword followed by one or
9197 several data types. Except for the "server_id" type which is
9198 automatically detected and enabled, all data types must be
9199 explicitly declared to be stored. If an ACL references a data
9200 type which is not stored, the ACL will simply not match. Some
9201 data types require an argument which must be passed just after
9202 the type between parenthesis. See below for the supported data
9203 types and their arguments.
9204
9205 The data types that can be stored with an entry are the following :
9206 - server_id : this is an integer which holds the numeric ID of the server a
9207 request was assigned to. It is used by the "stick match", "stick store",
9208 and "stick on" rules. It is automatically enabled when referenced.
9209
9210 - gpc0 : first General Purpose Counter. It is a positive 32-bit integer
9211 integer which may be used for anything. Most of the time it will be used
9212 to put a special tag on some entries, for instance to note that a
Davor Ocelice9ed2812017-12-25 17:49:28 +01009213 specific behavior was detected and must be known for future matches.
Willy Tarreauc9705a12010-07-27 20:05:50 +02009214
Willy Tarreauba2ffd12013-05-29 15:54:14 +02009215 - gpc0_rate(<period>) : increment rate of the first General Purpose Counter
9216 over a period. It is a positive 32-bit integer integer which may be used
9217 for anything. Just like <gpc0>, it counts events, but instead of keeping
Davor Ocelice9ed2812017-12-25 17:49:28 +01009218 a cumulative number, it maintains the rate at which the counter is
Willy Tarreauba2ffd12013-05-29 15:54:14 +02009219 incremented. Most of the time it will be used to measure the frequency of
Davor Ocelice9ed2812017-12-25 17:49:28 +01009220 occurrence of certain events (e.g. requests to a specific URL).
Willy Tarreauba2ffd12013-05-29 15:54:14 +02009221
Frédéric Lécaille6778b272018-01-29 15:22:53 +01009222 - gpc1 : second General Purpose Counter. It is a positive 32-bit integer
9223 integer which may be used for anything. Most of the time it will be used
9224 to put a special tag on some entries, for instance to note that a
9225 specific behavior was detected and must be known for future matches.
9226
9227 - gpc1_rate(<period>) : increment rate of the second General Purpose Counter
9228 over a period. It is a positive 32-bit integer integer which may be used
9229 for anything. Just like <gpc1>, it counts events, but instead of keeping
9230 a cumulative number, it maintains the rate at which the counter is
9231 incremented. Most of the time it will be used to measure the frequency of
9232 occurrence of certain events (e.g. requests to a specific URL).
9233
Willy Tarreauc9705a12010-07-27 20:05:50 +02009234 - conn_cnt : Connection Count. It is a positive 32-bit integer which counts
9235 the absolute number of connections received from clients which matched
9236 this entry. It does not mean the connections were accepted, just that
9237 they were received.
9238
9239 - conn_cur : Current Connections. It is a positive 32-bit integer which
9240 stores the concurrent connection counts for the entry. It is incremented
9241 once an incoming connection matches the entry, and decremented once the
9242 connection leaves. That way it is possible to know at any time the exact
9243 number of concurrent connections for an entry.
9244
9245 - conn_rate(<period>) : frequency counter (takes 12 bytes). It takes an
9246 integer parameter <period> which indicates in milliseconds the length
9247 of the period over which the average is measured. It reports the average
9248 incoming connection rate over that period, in connections per period. The
9249 result is an integer which can be matched using ACLs.
9250
9251 - sess_cnt : Session Count. It is a positive 32-bit integer which counts
9252 the absolute number of sessions received from clients which matched this
9253 entry. A session is a connection that was accepted by the layer 4 rules.
9254
9255 - sess_rate(<period>) : frequency counter (takes 12 bytes). It takes an
9256 integer parameter <period> which indicates in milliseconds the length
9257 of the period over which the average is measured. It reports the average
9258 incoming session rate over that period, in sessions per period. The
9259 result is an integer which can be matched using ACLs.
9260
9261 - http_req_cnt : HTTP request Count. It is a positive 32-bit integer which
9262 counts the absolute number of HTTP requests received from clients which
9263 matched this entry. It does not matter whether they are valid requests or
9264 not. Note that this is different from sessions when keep-alive is used on
9265 the client side.
9266
9267 - http_req_rate(<period>) : frequency counter (takes 12 bytes). It takes an
9268 integer parameter <period> which indicates in milliseconds the length
9269 of the period over which the average is measured. It reports the average
9270 HTTP request rate over that period, in requests per period. The result is
9271 an integer which can be matched using ACLs. It does not matter whether
9272 they are valid requests or not. Note that this is different from sessions
9273 when keep-alive is used on the client side.
9274
9275 - http_err_cnt : HTTP Error Count. It is a positive 32-bit integer which
9276 counts the absolute number of HTTP requests errors induced by clients
9277 which matched this entry. Errors are counted on invalid and truncated
9278 requests, as well as on denied or tarpitted requests, and on failed
9279 authentications. If the server responds with 4xx, then the request is
9280 also counted as an error since it's an error triggered by the client
Davor Ocelice9ed2812017-12-25 17:49:28 +01009281 (e.g. vulnerability scan).
Willy Tarreauc9705a12010-07-27 20:05:50 +02009282
9283 - http_err_rate(<period>) : frequency counter (takes 12 bytes). It takes an
9284 integer parameter <period> which indicates in milliseconds the length
9285 of the period over which the average is measured. It reports the average
9286 HTTP request error rate over that period, in requests per period (see
9287 http_err_cnt above for what is accounted as an error). The result is an
9288 integer which can be matched using ACLs.
9289
9290 - bytes_in_cnt : client to server byte count. It is a positive 64-bit
Davor Ocelice9ed2812017-12-25 17:49:28 +01009291 integer which counts the cumulative number of bytes received from clients
Willy Tarreauc9705a12010-07-27 20:05:50 +02009292 which matched this entry. Headers are included in the count. This may be
9293 used to limit abuse of upload features on photo or video servers.
9294
9295 - bytes_in_rate(<period>) : frequency counter (takes 12 bytes). It takes an
9296 integer parameter <period> which indicates in milliseconds the length
9297 of the period over which the average is measured. It reports the average
9298 incoming bytes rate over that period, in bytes per period. It may be used
9299 to detect users which upload too much and too fast. Warning: with large
9300 uploads, it is possible that the amount of uploaded data will be counted
9301 once upon termination, thus causing spikes in the average transfer speed
9302 instead of having a smooth one. This may partially be smoothed with
9303 "option contstats" though this is not perfect yet. Use of byte_in_cnt is
9304 recommended for better fairness.
9305
9306 - bytes_out_cnt : server to client byte count. It is a positive 64-bit
Davor Ocelice9ed2812017-12-25 17:49:28 +01009307 integer which counts the cumulative number of bytes sent to clients which
Willy Tarreauc9705a12010-07-27 20:05:50 +02009308 matched this entry. Headers are included in the count. This may be used
9309 to limit abuse of bots sucking the whole site.
9310
9311 - bytes_out_rate(<period>) : frequency counter (takes 12 bytes). It takes
9312 an integer parameter <period> which indicates in milliseconds the length
9313 of the period over which the average is measured. It reports the average
9314 outgoing bytes rate over that period, in bytes per period. It may be used
9315 to detect users which download too much and too fast. Warning: with large
9316 transfers, it is possible that the amount of transferred data will be
9317 counted once upon termination, thus causing spikes in the average
9318 transfer speed instead of having a smooth one. This may partially be
9319 smoothed with "option contstats" though this is not perfect yet. Use of
9320 byte_out_cnt is recommended for better fairness.
Willy Tarreau08d5f982010-06-06 13:34:54 +02009321
Willy Tarreauc00cdc22010-06-06 16:48:26 +02009322 There is only one stick-table per proxy. At the moment of writing this doc,
9323 it does not seem useful to have multiple tables per proxy. If this happens
Willy Tarreaub937b7e2010-01-12 15:27:54 +01009324 to be required, simply create a dummy backend with a stick-table in it and
9325 reference it.
9326
9327 It is important to understand that stickiness based on learning information
9328 has some limitations, including the fact that all learned associations are
Baptiste Assmann123ff042016-03-06 23:29:28 +01009329 lost upon restart unless peers are properly configured to transfer such
9330 information upon restart (recommended). In general it can be good as a
9331 complement but not always as an exclusive stickiness.
Willy Tarreaub937b7e2010-01-12 15:27:54 +01009332
Willy Tarreauc9705a12010-07-27 20:05:50 +02009333 Last, memory requirements may be important when storing many data types.
9334 Indeed, storing all indicators above at once in each entry requires 116 bytes
9335 per entry, or 116 MB for a 1-million entries table. This is definitely not
9336 something that can be ignored.
9337
9338 Example:
9339 # Keep track of counters of up to 1 million IP addresses over 5 minutes
9340 # and store a general purpose counter and the average connection rate
9341 # computed over a sliding window of 30 seconds.
9342 stick-table type ip size 1m expire 5m store gpc0,conn_rate(30s)
9343
Jarno Huuskonene0ee0be2017-07-04 10:35:12 +03009344 See also : "stick match", "stick on", "stick store-request", section 2.4
David du Colombiera13d1b92011-03-17 10:40:22 +01009345 about time format and section 7 about ACLs.
Willy Tarreaub937b7e2010-01-12 15:27:54 +01009346
9347
Emeric Brun6a1cefa2010-09-24 18:15:17 +02009348stick store-response <pattern> [table <table>] [{if | unless} <condition>]
Baptiste Assmann2f2d2ec2016-03-06 23:27:24 +01009349 Define a response pattern used to create an entry in a stickiness table
Emeric Brun6a1cefa2010-09-24 18:15:17 +02009350 May be used in sections : defaults | frontend | listen | backend
9351 no | no | yes | yes
9352
9353 Arguments :
Willy Tarreaube722a22014-06-13 16:31:59 +02009354 <pattern> is a sample expression rule as described in section 7.3. It
Emeric Brun6a1cefa2010-09-24 18:15:17 +02009355 describes what elements of the response or connection will
Davor Ocelice9ed2812017-12-25 17:49:28 +01009356 be analyzed, extracted and stored in the table once a
Emeric Brun6a1cefa2010-09-24 18:15:17 +02009357 server is selected.
9358
9359 <table> is an optional stickiness table name. If unspecified, the same
9360 backend's table is used. A stickiness table is declared using
9361 the "stick-table" statement.
9362
9363 <cond> is an optional storage condition. It makes it possible to store
9364 certain criteria only when some conditions are met (or not met).
9365 For instance, it could be used to store the SSL session ID only
9366 when the response is a SSL server hello.
9367
9368 Some protocols or applications require complex stickiness rules and cannot
9369 always simply rely on cookies nor hashing. The "stick store-response"
9370 statement describes a rule to decide what to extract from the response and
9371 when to do it, in order to store it into a stickiness table for further
9372 requests to match it using the "stick match" statement. Obviously the
9373 extracted part must make sense and have a chance to be matched in a further
Cyril Bonté108cf6e2012-04-21 23:30:29 +02009374 request. Storing an ID found in a header of a response makes sense.
Emeric Brun6a1cefa2010-09-24 18:15:17 +02009375 See section 7 for a complete list of possible patterns and transformation
9376 rules.
9377
9378 The table has to be declared using the "stick-table" statement. It must be of
9379 a type compatible with the pattern. By default it is the one which is present
9380 in the same backend. It is possible to share a table with other backends by
9381 referencing it using the "table" keyword. If another table is referenced,
9382 the server's ID inside the backends are used. By default, all server IDs
9383 start at 1 in each backend, so the server ordering is enough. But in case of
9384 doubt, it is highly recommended to force server IDs using their "id" setting.
9385
9386 It is possible to restrict the conditions where a "stick store-response"
9387 statement will apply, using "if" or "unless" followed by a condition. This
9388 condition will be evaluated while parsing the response, so any criteria can
9389 be used. See section 7 for ACL based conditions.
9390
9391 There is no limit on the number of "stick store-response" statements, but
9392 there is a limit of 8 simultaneous stores per request or response. This
9393 makes it possible to store up to 8 criteria, all extracted from either the
9394 request or the response, regardless of the number of rules. Only the 8 first
9395 ones which match will be kept. Using this, it is possible to feed multiple
9396 tables at once in the hope to increase the chance to recognize a user on
Willy Tarreau9667a802013-12-09 12:52:13 +01009397 another protocol or access method. Using multiple store-response rules with
9398 the same table is possible and may be used to find the best criterion to rely
9399 on, by arranging the rules by decreasing preference order. Only the first
9400 extracted criterion for a given table will be stored. All subsequent store-
9401 response rules referencing the same table will be skipped and their ACLs will
9402 not be evaluated. However, even if a store-request rule references a table, a
9403 store-response rule may also use the same table. This means that each table
9404 may learn exactly one element from the request and one element from the
9405 response at once.
Emeric Brun6a1cefa2010-09-24 18:15:17 +02009406
9407 The table will contain the real server that processed the request.
9408
9409 Example :
9410 # Learn SSL session ID from both request and response and create affinity.
9411 backend https
9412 mode tcp
9413 balance roundrobin
Cyril Bontédc4d9032012-04-08 21:57:39 +02009414 # maximum SSL session ID length is 32 bytes.
Emeric Brun6a1cefa2010-09-24 18:15:17 +02009415 stick-table type binary len 32 size 30k expire 30m
Cyril Bonté108cf6e2012-04-21 23:30:29 +02009416
Emeric Brun6a1cefa2010-09-24 18:15:17 +02009417 acl clienthello req_ssl_hello_type 1
9418 acl serverhello rep_ssl_hello_type 2
9419
9420 # use tcp content accepts to detects ssl client and server hello.
9421 tcp-request inspect-delay 5s
9422 tcp-request content accept if clienthello
9423
9424 # no timeout on response inspect delay by default.
9425 tcp-response content accept if serverhello
Cyril Bonté108cf6e2012-04-21 23:30:29 +02009426
Emeric Brun6a1cefa2010-09-24 18:15:17 +02009427 # SSL session ID (SSLID) may be present on a client or server hello.
9428 # Its length is coded on 1 byte at offset 43 and its value starts
9429 # at offset 44.
9430
9431 # Match and learn on request if client hello.
9432 stick on payload_lv(43,1) if clienthello
9433
9434 # Learn on response if server hello.
9435 stick store-response payload_lv(43,1) if serverhello
Cyril Bontédc4d9032012-04-08 21:57:39 +02009436
Emeric Brun6a1cefa2010-09-24 18:15:17 +02009437 server s1 192.168.1.1:443
9438 server s2 192.168.1.1:443
9439
9440 See also : "stick-table", "stick on", and section 7 about ACLs and pattern
9441 extraction.
9442
9443
Willy Tarreau938c7fe2014-04-25 14:21:39 +02009444tcp-check connect [params*]
9445 Opens a new connection
9446 May be used in sections: defaults | frontend | listen | backend
9447 no | no | yes | yes
9448
9449 When an application lies on more than a single TCP port or when HAProxy
9450 load-balance many services in a single backend, it makes sense to probe all
9451 the services individually before considering a server as operational.
9452
9453 When there are no TCP port configured on the server line neither server port
9454 directive, then the 'tcp-check connect port <port>' must be the first step
9455 of the sequence.
9456
9457 In a tcp-check ruleset a 'connect' is required, it is also mandatory to start
9458 the ruleset with a 'connect' rule. Purpose is to ensure admin know what they
9459 do.
9460
9461 Parameters :
9462 They are optional and can be used to describe how HAProxy should open and
9463 use the TCP connection.
9464
9465 port if not set, check port or server port is used.
9466 It tells HAProxy where to open the connection to.
9467 <port> must be a valid TCP port source integer, from 1 to 65535.
9468
9469 send-proxy send a PROXY protocol string
9470
9471 ssl opens a ciphered connection
9472
9473 Examples:
9474 # check HTTP and HTTPs services on a server.
9475 # first open port 80 thanks to server line port directive, then
9476 # tcp-check opens port 443, ciphered and run a request on it:
9477 option tcp-check
9478 tcp-check connect
9479 tcp-check send GET\ /\ HTTP/1.0\r\n
9480 tcp-check send Host:\ haproxy.1wt.eu\r\n
9481 tcp-check send \r\n
9482 tcp-check expect rstring (2..|3..)
9483 tcp-check connect port 443 ssl
9484 tcp-check send GET\ /\ HTTP/1.0\r\n
9485 tcp-check send Host:\ haproxy.1wt.eu\r\n
9486 tcp-check send \r\n
9487 tcp-check expect rstring (2..|3..)
9488 server www 10.0.0.1 check port 80
9489
9490 # check both POP and IMAP from a single server:
9491 option tcp-check
9492 tcp-check connect port 110
9493 tcp-check expect string +OK\ POP3\ ready
9494 tcp-check connect port 143
9495 tcp-check expect string *\ OK\ IMAP4\ ready
9496 server mail 10.0.0.1 check
9497
9498 See also : "option tcp-check", "tcp-check send", "tcp-check expect"
9499
9500
9501tcp-check expect [!] <match> <pattern>
Davor Ocelice9ed2812017-12-25 17:49:28 +01009502 Specify data to be collected and analyzed during a generic health check
Willy Tarreau938c7fe2014-04-25 14:21:39 +02009503 May be used in sections: defaults | frontend | listen | backend
9504 no | no | yes | yes
9505
9506 Arguments :
9507 <match> is a keyword indicating how to look for a specific pattern in the
9508 response. The keyword may be one of "string", "rstring" or
9509 binary.
9510 The keyword may be preceded by an exclamation mark ("!") to negate
9511 the match. Spaces are allowed between the exclamation mark and the
9512 keyword. See below for more details on the supported keywords.
9513
9514 <pattern> is the pattern to look for. It may be a string or a regular
9515 expression. If the pattern contains spaces, they must be escaped
9516 with the usual backslash ('\').
9517 If the match is set to binary, then the pattern must be passed as
Davor Ocelice9ed2812017-12-25 17:49:28 +01009518 a series of hexadecimal digits in an even number. Each sequence of
Willy Tarreau938c7fe2014-04-25 14:21:39 +02009519 two digits will represent a byte. The hexadecimal digits may be
9520 used upper or lower case.
9521
9522
9523 The available matches are intentionally similar to their http-check cousins :
9524
9525 string <string> : test the exact string matches in the response buffer.
9526 A health check response will be considered valid if the
9527 response's buffer contains this exact string. If the
9528 "string" keyword is prefixed with "!", then the response
9529 will be considered invalid if the body contains this
9530 string. This can be used to look for a mandatory pattern
9531 in a protocol response, or to detect a failure when a
9532 specific error appears in a protocol banner.
9533
9534 rstring <regex> : test a regular expression on the response buffer.
9535 A health check response will be considered valid if the
9536 response's buffer matches this expression. If the
9537 "rstring" keyword is prefixed with "!", then the response
9538 will be considered invalid if the body matches the
9539 expression.
9540
9541 binary <hexstring> : test the exact string in its hexadecimal form matches
9542 in the response buffer. A health check response will
9543 be considered valid if the response's buffer contains
9544 this exact hexadecimal string.
9545 Purpose is to match data on binary protocols.
9546
9547 It is important to note that the responses will be limited to a certain size
9548 defined by the global "tune.chksize" option, which defaults to 16384 bytes.
9549 Thus, too large responses may not contain the mandatory pattern when using
9550 "string", "rstring" or binary. If a large response is absolutely required, it
9551 is possible to change the default max size by setting the global variable.
9552 However, it is worth keeping in mind that parsing very large responses can
9553 waste some CPU cycles, especially when regular expressions are used, and that
9554 it is always better to focus the checks on smaller resources. Also, in its
9555 current state, the check will not find any string nor regex past a null
9556 character in the response. Similarly it is not possible to request matching
9557 the null character.
9558
9559 Examples :
9560 # perform a POP check
9561 option tcp-check
9562 tcp-check expect string +OK\ POP3\ ready
9563
9564 # perform an IMAP check
9565 option tcp-check
9566 tcp-check expect string *\ OK\ IMAP4\ ready
9567
9568 # look for the redis master server
9569 option tcp-check
9570 tcp-check send PING\r\n
Baptiste Assmanna3322992015-08-04 10:12:18 +02009571 tcp-check expect string +PONG
Willy Tarreau938c7fe2014-04-25 14:21:39 +02009572 tcp-check send info\ replication\r\n
9573 tcp-check expect string role:master
9574 tcp-check send QUIT\r\n
9575 tcp-check expect string +OK
9576
9577
9578 See also : "option tcp-check", "tcp-check connect", "tcp-check send",
9579 "tcp-check send-binary", "http-check expect", tune.chksize
9580
9581
9582tcp-check send <data>
9583 Specify a string to be sent as a question during a generic health check
9584 May be used in sections: defaults | frontend | listen | backend
9585 no | no | yes | yes
9586
9587 <data> : the data to be sent as a question during a generic health check
9588 session. For now, <data> must be a string.
9589
9590 Examples :
9591 # look for the redis master server
9592 option tcp-check
9593 tcp-check send info\ replication\r\n
9594 tcp-check expect string role:master
9595
9596 See also : "option tcp-check", "tcp-check connect", "tcp-check expect",
9597 "tcp-check send-binary", tune.chksize
9598
9599
Davor Ocelice9ed2812017-12-25 17:49:28 +01009600tcp-check send-binary <hexstring>
9601 Specify a hex digits string to be sent as a binary question during a raw
Willy Tarreau938c7fe2014-04-25 14:21:39 +02009602 tcp health check
9603 May be used in sections: defaults | frontend | listen | backend
9604 no | no | yes | yes
9605
9606 <data> : the data to be sent as a question during a generic health check
9607 session. For now, <data> must be a string.
Davor Ocelice9ed2812017-12-25 17:49:28 +01009608 <hexstring> : test the exact string in its hexadecimal form matches in the
Willy Tarreau938c7fe2014-04-25 14:21:39 +02009609 response buffer. A health check response will be considered
9610 valid if the response's buffer contains this exact
9611 hexadecimal string.
9612 Purpose is to send binary data to ask on binary protocols.
9613
9614 Examples :
9615 # redis check in binary
9616 option tcp-check
9617 tcp-check send-binary 50494e470d0a # PING\r\n
9618 tcp-check expect binary 2b504F4e47 # +PONG
9619
9620
9621 See also : "option tcp-check", "tcp-check connect", "tcp-check expect",
9622 "tcp-check send", tune.chksize
9623
9624
Willy Tarreaue9656522010-08-17 15:40:09 +02009625tcp-request connection <action> [{if | unless} <condition>]
9626 Perform an action on an incoming connection depending on a layer 4 condition
Willy Tarreau1a687942010-05-23 22:40:30 +02009627 May be used in sections : defaults | frontend | listen | backend
9628 no | yes | yes | no
Willy Tarreaue9656522010-08-17 15:40:09 +02009629 Arguments :
Willy Tarreauc870bfd2015-09-28 18:47:38 +02009630 <action> defines the action to perform if the condition applies. See
9631 below.
Willy Tarreau1a687942010-05-23 22:40:30 +02009632
Willy Tarreaue9656522010-08-17 15:40:09 +02009633 <condition> is a standard layer4-only ACL-based condition (see section 7).
Willy Tarreau68c03ab2010-08-06 15:08:45 +02009634
9635 Immediately after acceptance of a new incoming connection, it is possible to
9636 evaluate some conditions to decide whether this connection must be accepted
Willy Tarreaue9656522010-08-17 15:40:09 +02009637 or dropped or have its counters tracked. Those conditions cannot make use of
9638 any data contents because the connection has not been read from yet, and the
9639 buffers are not yet allocated. This is used to selectively and very quickly
9640 accept or drop connections from various sources with a very low overhead. If
9641 some contents need to be inspected in order to take the decision, the
9642 "tcp-request content" statements must be used instead.
Willy Tarreau68c03ab2010-08-06 15:08:45 +02009643
Willy Tarreaue9656522010-08-17 15:40:09 +02009644 The "tcp-request connection" rules are evaluated in their exact declaration
9645 order. If no rule matches or if there is no rule, the default action is to
9646 accept the incoming connection. There is no specific limit to the number of
9647 rules which may be inserted.
Willy Tarreau68c03ab2010-08-06 15:08:45 +02009648
Willy Tarreaua9083d02015-05-08 15:27:59 +02009649 Four types of actions are supported :
Willy Tarreaue9656522010-08-17 15:40:09 +02009650 - accept :
9651 accepts the connection if the condition is true (when used with "if")
9652 or false (when used with "unless"). The first such rule executed ends
9653 the rules evaluation.
Willy Tarreau68c03ab2010-08-06 15:08:45 +02009654
Willy Tarreaue9656522010-08-17 15:40:09 +02009655 - reject :
9656 rejects the connection if the condition is true (when used with "if")
9657 or false (when used with "unless"). The first such rule executed ends
9658 the rules evaluation. Rejected connections do not even become a
9659 session, which is why they are accounted separately for in the stats,
9660 as "denied connections". They are not considered for the session
9661 rate-limit and are not logged either. The reason is that these rules
9662 should only be used to filter extremely high connection rates such as
9663 the ones encountered during a massive DDoS attack. Under these extreme
9664 conditions, the simple action of logging each event would make the
9665 system collapse and would considerably lower the filtering capacity. If
9666 logging is absolutely desired, then "tcp-request content" rules should
Willy Tarreau4f614292016-10-21 17:49:36 +02009667 be used instead, as "tcp-request session" rules will not log either.
Willy Tarreau68c03ab2010-08-06 15:08:45 +02009668
Willy Tarreau4f0d9192013-06-11 20:40:55 +02009669 - expect-proxy layer4 :
9670 configures the client-facing connection to receive a PROXY protocol
9671 header before any byte is read from the socket. This is equivalent to
9672 having the "accept-proxy" keyword on the "bind" line, except that using
9673 the TCP rule allows the PROXY protocol to be accepted only for certain
9674 IP address ranges using an ACL. This is convenient when multiple layers
9675 of load balancers are passed through by traffic coming from public
9676 hosts.
9677
Bertrand Jacquin90759682016-06-06 15:35:39 +01009678 - expect-netscaler-cip layer4 :
9679 configures the client-facing connection to receive a NetScaler Client
9680 IP insertion protocol header before any byte is read from the socket.
9681 This is equivalent to having the "accept-netscaler-cip" keyword on the
9682 "bind" line, except that using the TCP rule allows the PROXY protocol
9683 to be accepted only for certain IP address ranges using an ACL. This
9684 is convenient when multiple layers of load balancers are passed
9685 through by traffic coming from public hosts.
9686
Willy Tarreau18bf01e2014-06-13 16:18:52 +02009687 - capture <sample> len <length> :
9688 This only applies to "tcp-request content" rules. It captures sample
9689 expression <sample> from the request buffer, and converts it to a
9690 string of at most <len> characters. The resulting string is stored into
9691 the next request "capture" slot, so it will possibly appear next to
9692 some captured HTTP headers. It will then automatically appear in the
9693 logs, and it will be possible to extract it using sample fetch rules to
9694 feed it into headers or anything. The length should be limited given
9695 that this size will be allocated for each capture during the whole
Willy Tarreaua9083d02015-05-08 15:27:59 +02009696 session life. Please check section 7.3 (Fetching samples) and "capture
9697 request header" for more information.
Willy Tarreau18bf01e2014-06-13 16:18:52 +02009698
Willy Tarreaube4a3ef2013-06-17 15:04:07 +02009699 - { track-sc0 | track-sc1 | track-sc2 } <key> [table <table>] :
Willy Tarreaue9656522010-08-17 15:40:09 +02009700 enables tracking of sticky counters from current connection. These
Moemen MHEDHBI9cf46342018-09-25 17:50:53 +02009701 rules do not stop evaluation and do not change default action. The
9702 number of counters that may be simultaneously tracked by the same
9703 connection is set in MAX_SESS_STKCTR at build time (reported in
John Roeslerfb2fce12019-07-10 15:45:51 -05009704 haproxy -vv) which defaults to 3, so the track-sc number is between 0
Moemen MHEDHBI9cf46342018-09-25 17:50:53 +02009705 and (MAX_SESS_STCKTR-1). The first "track-sc0" rule executed enables
9706 tracking of the counters of the specified table as the first set. The
9707 first "track-sc1" rule executed enables tracking of the counters of the
9708 specified table as the second set. The first "track-sc2" rule executed
9709 enables tracking of the counters of the specified table as the third
9710 set. It is a recommended practice to use the first set of counters for
9711 the per-frontend counters and the second set for the per-backend ones.
9712 But this is just a guideline, all may be used everywhere.
Willy Tarreau68c03ab2010-08-06 15:08:45 +02009713
Willy Tarreaue9656522010-08-17 15:40:09 +02009714 These actions take one or two arguments :
Willy Tarreaube722a22014-06-13 16:31:59 +02009715 <key> is mandatory, and is a sample expression rule as described
Willy Tarreau74ca5042013-06-11 23:12:07 +02009716 in section 7.3. It describes what elements of the incoming
Davor Ocelice9ed2812017-12-25 17:49:28 +01009717 request or connection will be analyzed, extracted, combined,
Willy Tarreau5d5b5d82012-12-09 12:00:04 +01009718 and used to select which table entry to update the counters.
9719 Note that "tcp-request connection" cannot use content-based
9720 fetches.
Willy Tarreau68c03ab2010-08-06 15:08:45 +02009721
Willy Tarreaue9656522010-08-17 15:40:09 +02009722 <table> is an optional table to be used instead of the default one,
9723 which is the stick-table declared in the current proxy. All
9724 the counters for the matches and updates for the key will
9725 then be performed in that table until the session ends.
Willy Tarreau68c03ab2010-08-06 15:08:45 +02009726
Willy Tarreaue9656522010-08-17 15:40:09 +02009727 Once a "track-sc*" rule is executed, the key is looked up in the table
9728 and if it is not found, an entry is allocated for it. Then a pointer to
9729 that entry is kept during all the session's life, and this entry's
9730 counters are updated as often as possible, every time the session's
9731 counters are updated, and also systematically when the session ends.
Willy Tarreau5d5b5d82012-12-09 12:00:04 +01009732 Counters are only updated for events that happen after the tracking has
9733 been started. For example, connection counters will not be updated when
9734 tracking layer 7 information, since the connection event happens before
9735 layer7 information is extracted.
9736
Willy Tarreaue9656522010-08-17 15:40:09 +02009737 If the entry tracks concurrent connection counters, one connection is
9738 counted for as long as the entry is tracked, and the entry will not
9739 expire during that time. Tracking counters also provides a performance
9740 advantage over just checking the keys, because only one table lookup is
9741 performed for all ACL checks that make use of it.
Willy Tarreau68c03ab2010-08-06 15:08:45 +02009742
Thierry FOURNIERe0627bd2015-08-04 08:20:33 +02009743 - sc-inc-gpc0(<sc-id>):
9744 The "sc-inc-gpc0" increments the GPC0 counter according to the sticky
9745 counter designated by <sc-id>. If an error occurs, this action silently
9746 fails and the actions evaluation continues.
9747
Frédéric Lécaille6778b272018-01-29 15:22:53 +01009748 - sc-inc-gpc1(<sc-id>):
9749 The "sc-inc-gpc1" increments the GPC1 counter according to the sticky
9750 counter designated by <sc-id>. If an error occurs, this action silently
9751 fails and the actions evaluation continues.
9752
Thierry FOURNIER236657b2015-08-19 08:25:14 +02009753 - sc-set-gpt0(<sc-id>) <int>:
9754 This action sets the GPT0 tag according to the sticky counter designated
9755 by <sc-id> and the value of <int>. The expected result is a boolean. If
9756 an error occurs, this action silently fails and the actions evaluation
9757 continues.
9758
William Lallemand2e785f22016-05-25 01:48:42 +02009759 - set-src <expr> :
9760 Is used to set the source IP address to the value of specified
9761 expression. Useful if you want to mask source IP for privacy.
9762 If you want to provide an IP from a HTTP header use "http-request
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02009763 set-src".
William Lallemand2e785f22016-05-25 01:48:42 +02009764
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02009765 Arguments:
9766 <expr> Is a standard HAProxy expression formed by a sample-fetch
9767 followed by some converters.
William Lallemand2e785f22016-05-25 01:48:42 +02009768
9769 Example:
William Lallemand2e785f22016-05-25 01:48:42 +02009770 tcp-request connection set-src src,ipmask(24)
9771
Willy Tarreau0c630532016-10-21 17:52:58 +02009772 When possible, set-src preserves the original source port as long as the
9773 address family allows it, otherwise the source port is set to 0.
William Lallemand2e785f22016-05-25 01:48:42 +02009774
William Lallemand44be6402016-05-25 01:51:35 +02009775 - set-src-port <expr> :
9776 Is used to set the source port address to the value of specified
9777 expression.
9778
Cyril Bonté6c81d5f2018-10-17 00:14:51 +02009779 Arguments:
9780 <expr> Is a standard HAProxy expression formed by a sample-fetch
9781 followed by some converters.
William Lallemand44be6402016-05-25 01:51:35 +02009782
9783 Example:
William Lallemand44be6402016-05-25 01:51:35 +02009784 tcp-request connection set-src-port int(4000)
9785
Willy Tarreau0c630532016-10-21 17:52:58 +02009786 When possible, set-src-port preserves the original source address as long
9787 as the address family supports a port, otherwise it forces the source
9788 address to IPv4 "0.0.0.0" before rewriting the port.
William Lallemand44be6402016-05-25 01:51:35 +02009789
William Lallemand13e9b0c2016-05-25 02:34:07 +02009790 - set-dst <expr> :
9791 Is used to set the destination IP address to the value of specified
9792 expression. Useful if you want to mask IP for privacy in log.
9793 If you want to provide an IP from a HTTP header use "http-request
9794 set-dst". If you want to connect to the new address/port, use
9795 '0.0.0.0:0' as a server address in the backend.
9796
9797 <expr> Is a standard HAProxy expression formed by a sample-fetch
9798 followed by some converters.
9799
9800 Example:
9801
9802 tcp-request connection set-dst dst,ipmask(24)
9803 tcp-request connection set-dst ipv4(10.0.0.1)
9804
Willy Tarreau0c630532016-10-21 17:52:58 +02009805 When possible, set-dst preserves the original destination port as long as
9806 the address family allows it, otherwise the destination port is set to 0.
9807
William Lallemand13e9b0c2016-05-25 02:34:07 +02009808 - set-dst-port <expr> :
9809 Is used to set the destination port address to the value of specified
9810 expression. If you want to connect to the new address/port, use
9811 '0.0.0.0:0' as a server address in the backend.
9812
9813
9814 <expr> Is a standard HAProxy expression formed by a sample-fetch
9815 followed by some converters.
9816
9817 Example:
9818
9819 tcp-request connection set-dst-port int(4000)
9820
Willy Tarreau0c630532016-10-21 17:52:58 +02009821 When possible, set-dst-port preserves the original destination address as
9822 long as the address family supports a port, otherwise it forces the
9823 destination address to IPv4 "0.0.0.0" before rewriting the port.
9824
Willy Tarreau2d392c22015-08-24 01:43:45 +02009825 - "silent-drop" :
9826 This stops the evaluation of the rules and makes the client-facing
Davor Ocelice9ed2812017-12-25 17:49:28 +01009827 connection suddenly disappear using a system-dependent way that tries
Willy Tarreau2d392c22015-08-24 01:43:45 +02009828 to prevent the client from being notified. The effect it then that the
9829 client still sees an established connection while there's none on
9830 HAProxy. The purpose is to achieve a comparable effect to "tarpit"
9831 except that it doesn't use any local resource at all on the machine
9832 running HAProxy. It can resist much higher loads than "tarpit", and
Davor Ocelice9ed2812017-12-25 17:49:28 +01009833 slow down stronger attackers. It is important to understand the impact
9834 of using this mechanism. All stateful equipment placed between the
Willy Tarreau2d392c22015-08-24 01:43:45 +02009835 client and HAProxy (firewalls, proxies, load balancers) will also keep
9836 the established connection for a long time and may suffer from this
Davor Ocelice9ed2812017-12-25 17:49:28 +01009837 action. On modern Linux systems running with enough privileges, the
Willy Tarreau2d392c22015-08-24 01:43:45 +02009838 TCP_REPAIR socket option is used to block the emission of a TCP
9839 reset. On other systems, the socket's TTL is reduced to 1 so that the
9840 TCP reset doesn't pass the first router, though it's still delivered to
9841 local networks. Do not use it unless you fully understand how it works.
9842
Willy Tarreaue9656522010-08-17 15:40:09 +02009843 Note that the "if/unless" condition is optional. If no condition is set on
9844 the action, it is simply performed unconditionally. That can be useful for
9845 "track-sc*" actions as well as for changing the default action to a reject.
Willy Tarreau68c03ab2010-08-06 15:08:45 +02009846
Willy Tarreaue9656522010-08-17 15:40:09 +02009847 Example: accept all connections from white-listed hosts, reject too fast
9848 connection without counting them, and track accepted connections.
9849 This results in connection rate being capped from abusive sources.
Willy Tarreau68c03ab2010-08-06 15:08:45 +02009850
Willy Tarreaue9656522010-08-17 15:40:09 +02009851 tcp-request connection accept if { src -f /etc/haproxy/whitelist.lst }
Willy Tarreau68c03ab2010-08-06 15:08:45 +02009852 tcp-request connection reject if { src_conn_rate gt 10 }
Willy Tarreaube4a3ef2013-06-17 15:04:07 +02009853 tcp-request connection track-sc0 src
Willy Tarreau68c03ab2010-08-06 15:08:45 +02009854
Willy Tarreaue9656522010-08-17 15:40:09 +02009855 Example: accept all connections from white-listed hosts, count all other
9856 connections and reject too fast ones. This results in abusive ones
9857 being blocked as long as they don't slow down.
Willy Tarreau68c03ab2010-08-06 15:08:45 +02009858
Willy Tarreaue9656522010-08-17 15:40:09 +02009859 tcp-request connection accept if { src -f /etc/haproxy/whitelist.lst }
Willy Tarreaube4a3ef2013-06-17 15:04:07 +02009860 tcp-request connection track-sc0 src
9861 tcp-request connection reject if { sc0_conn_rate gt 10 }
Willy Tarreau68c03ab2010-08-06 15:08:45 +02009862
Willy Tarreau4f0d9192013-06-11 20:40:55 +02009863 Example: enable the PROXY protocol for traffic coming from all known proxies.
9864
9865 tcp-request connection expect-proxy layer4 if { src -f proxies.lst }
9866
Willy Tarreau68c03ab2010-08-06 15:08:45 +02009867 See section 7 about ACL usage.
9868
Willy Tarreau4f614292016-10-21 17:49:36 +02009869 See also : "tcp-request session", "tcp-request content", "stick-table"
Willy Tarreau68c03ab2010-08-06 15:08:45 +02009870
9871
Willy Tarreaue9656522010-08-17 15:40:09 +02009872tcp-request content <action> [{if | unless} <condition>]
9873 Perform an action on a new session depending on a layer 4-7 condition
Willy Tarreau62644772008-07-16 18:36:06 +02009874 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaufb356202010-08-03 14:02:05 +02009875 no | yes | yes | yes
Willy Tarreaue9656522010-08-17 15:40:09 +02009876 Arguments :
Willy Tarreauc870bfd2015-09-28 18:47:38 +02009877 <action> defines the action to perform if the condition applies. See
9878 below.
Willy Tarreau62644772008-07-16 18:36:06 +02009879
Willy Tarreaue9656522010-08-17 15:40:09 +02009880 <condition> is a standard layer 4-7 ACL-based condition (see section 7).
Willy Tarreau62644772008-07-16 18:36:06 +02009881
Davor Ocelice9ed2812017-12-25 17:49:28 +01009882 A request's contents can be analyzed at an early stage of request processing
Willy Tarreaue9656522010-08-17 15:40:09 +02009883 called "TCP content inspection". During this stage, ACL-based rules are
9884 evaluated every time the request contents are updated, until either an
9885 "accept" or a "reject" rule matches, or the TCP request inspection delay
9886 expires with no matching rule.
Willy Tarreau62644772008-07-16 18:36:06 +02009887
Willy Tarreaue9656522010-08-17 15:40:09 +02009888 The first difference between these rules and "tcp-request connection" rules
9889 is that "tcp-request content" rules can make use of contents to take a
9890 decision. Most often, these decisions will consider a protocol recognition or
9891 validity. The second difference is that content-based rules can be used in
Willy Tarreauf3338342014-01-28 21:40:28 +01009892 both frontends and backends. In case of HTTP keep-alive with the client, all
9893 tcp-request content rules are evaluated again, so haproxy keeps a record of
9894 what sticky counters were assigned by a "tcp-request connection" versus a
9895 "tcp-request content" rule, and flushes all the content-related ones after
9896 processing an HTTP request, so that they may be evaluated again by the rules
9897 being evaluated again for the next request. This is of particular importance
Jarno Huuskonen0e82b922014-04-12 18:22:19 +03009898 when the rule tracks some L7 information or when it is conditioned by an
Willy Tarreauf3338342014-01-28 21:40:28 +01009899 L7-based ACL, since tracking may change between requests.
Willy Tarreau62644772008-07-16 18:36:06 +02009900
Willy Tarreaue9656522010-08-17 15:40:09 +02009901 Content-based rules are evaluated in their exact declaration order. If no
9902 rule matches or if there is no rule, the default action is to accept the
9903 contents. There is no specific limit to the number of rules which may be
9904 inserted.
Willy Tarreau62644772008-07-16 18:36:06 +02009905
Thierry FOURNIER236657b2015-08-19 08:25:14 +02009906 Several types of actions are supported :
Willy Tarreau18bf01e2014-06-13 16:18:52 +02009907 - accept : the request is accepted
Baptiste Assmann333939c2019-01-21 08:34:50 +01009908 - do-resolve: perform a DNS resolution
Willy Tarreau18bf01e2014-06-13 16:18:52 +02009909 - reject : the request is rejected and the connection is closed
9910 - capture : the specified sample expression is captured
Patrick Hemmer268a7072018-05-11 12:52:31 -04009911 - set-priority-class <expr> | set-priority-offset <expr>
Willy Tarreaube4a3ef2013-06-17 15:04:07 +02009912 - { track-sc0 | track-sc1 | track-sc2 } <key> [table <table>]
Thierry FOURNIERe0627bd2015-08-04 08:20:33 +02009913 - sc-inc-gpc0(<sc-id>)
Frédéric Lécaille6778b272018-01-29 15:22:53 +01009914 - sc-inc-gpc1(<sc-id>)
Thierry Fournierb9125672016-03-29 19:34:37 +02009915 - sc-set-gpt0(<sc-id>) <int>
Baptiste Assmanne1afd4f2019-04-18 16:21:13 +02009916 - set-dst <expr>
9917 - set-dst-port <expr>
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02009918 - set-var(<var-name>) <expr>
Christopher Faulet85d79c92016-11-09 16:54:56 +01009919 - unset-var(<var-name>)
Willy Tarreau2d392c22015-08-24 01:43:45 +02009920 - silent-drop
Davor Ocelice9ed2812017-12-25 17:49:28 +01009921 - send-spoe-group <engine-name> <group-name>
Willy Tarreau62644772008-07-16 18:36:06 +02009922
Willy Tarreaue9656522010-08-17 15:40:09 +02009923 They have the same meaning as their counter-parts in "tcp-request connection"
9924 so please refer to that section for a complete description.
Baptiste Assmann333939c2019-01-21 08:34:50 +01009925 For "do-resolve" action, please check the "http-request do-resolve"
9926 configuration section.
Willy Tarreau62644772008-07-16 18:36:06 +02009927
Willy Tarreauf3338342014-01-28 21:40:28 +01009928 While there is nothing mandatory about it, it is recommended to use the
9929 track-sc0 in "tcp-request connection" rules, track-sc1 for "tcp-request
9930 content" rules in the frontend, and track-sc2 for "tcp-request content"
9931 rules in the backend, because that makes the configuration more readable
9932 and easier to troubleshoot, but this is just a guideline and all counters
9933 may be used everywhere.
Willy Tarreau62644772008-07-16 18:36:06 +02009934
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +01009935 Note that the "if/unless" condition is optional. If no condition is set on
Willy Tarreaue9656522010-08-17 15:40:09 +02009936 the action, it is simply performed unconditionally. That can be useful for
9937 "track-sc*" actions as well as for changing the default action to a reject.
Willy Tarreau62644772008-07-16 18:36:06 +02009938
Willy Tarreaue9656522010-08-17 15:40:09 +02009939 It is perfectly possible to match layer 7 contents with "tcp-request content"
Willy Tarreauc0239e02012-04-16 14:42:55 +02009940 rules, since HTTP-specific ACL matches are able to preliminarily parse the
9941 contents of a buffer before extracting the required data. If the buffered
9942 contents do not parse as a valid HTTP message, then the ACL does not match.
9943 The parser which is involved there is exactly the same as for all other HTTP
Willy Tarreauf3338342014-01-28 21:40:28 +01009944 processing, so there is no risk of parsing something differently. In an HTTP
9945 backend connected to from an HTTP frontend, it is guaranteed that HTTP
9946 contents will always be immediately present when the rule is evaluated first.
Willy Tarreau62644772008-07-16 18:36:06 +02009947
Willy Tarreau5d5b5d82012-12-09 12:00:04 +01009948 Tracking layer7 information is also possible provided that the information
Willy Tarreau4d54c7c2014-09-16 15:48:15 +02009949 are present when the rule is processed. The rule processing engine is able to
9950 wait until the inspect delay expires when the data to be tracked is not yet
9951 available.
Willy Tarreau5d5b5d82012-12-09 12:00:04 +01009952
Baptiste Assmanne1afd4f2019-04-18 16:21:13 +02009953 The "set-dst" and "set-dst-port" are used to set respectively the destination
9954 IP and port. More information on how to use it at "http-request set-dst".
9955
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02009956 The "set-var" is used to set the content of a variable. The variable is
Willy Tarreau4f614292016-10-21 17:49:36 +02009957 declared inline. For "tcp-request session" rules, only session-level
9958 variables can be used, without any layer7 contents.
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02009959
Daniel Schneller0b547052016-03-21 20:46:57 +01009960 <var-name> The name of the variable starts with an indication about
9961 its scope. The scopes allowed are:
Christopher Fauletff2613e2016-11-09 11:36:17 +01009962 "proc" : the variable is shared with the whole process
Daniel Schneller0b547052016-03-21 20:46:57 +01009963 "sess" : the variable is shared with the whole session
9964 "txn" : the variable is shared with the transaction
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02009965 (request and response)
Daniel Schneller0b547052016-03-21 20:46:57 +01009966 "req" : the variable is shared only during request
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02009967 processing
Daniel Schneller0b547052016-03-21 20:46:57 +01009968 "res" : the variable is shared only during response
9969 processing
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02009970 This prefix is followed by a name. The separator is a '.'.
Christopher Fauletb71557a2016-10-31 10:49:03 +01009971 The name may only contain characters 'a-z', 'A-Z', '0-9',
9972 '.' and '_'.
Thierry FOURNIER4834bc72015-06-06 19:29:07 +02009973
9974 <expr> Is a standard HAProxy expression formed by a sample-fetch
9975 followed by some converters.
9976
Christopher Faulet85d79c92016-11-09 16:54:56 +01009977 The "unset-var" is used to unset a variable. See above for details about
9978 <var-name>.
9979
Patrick Hemmer268a7072018-05-11 12:52:31 -04009980 The "set-priority-class" is used to set the queue priority class of the
9981 current request. The value must be a sample expression which converts to an
9982 integer in the range -2047..2047. Results outside this range will be
9983 truncated. The priority class determines the order in which queued requests
9984 are processed. Lower values have higher priority.
9985
9986 The "set-priority-offset" is used to set the queue priority timestamp offset
9987 of the current request. The value must be a sample expression which converts
9988 to an integer in the range -524287..524287. Results outside this range will be
9989 truncated. When a request is queued, it is ordered first by the priority
9990 class, then by the current timestamp adjusted by the given offset in
9991 milliseconds. Lower values have higher priority.
9992 Note that the resulting timestamp is is only tracked with enough precision for
9993 524,287ms (8m44s287ms). If the request is queued long enough to where the
9994 adjusted timestamp exceeds this value, it will be misidentified as highest
9995 priority. Thus it is important to set "timeout queue" to a value, where when
9996 combined with the offset, does not exceed this limit.
9997
Christopher Faulet76c09ef2017-09-21 11:03:52 +02009998 The "send-spoe-group" is used to trigger sending of a group of SPOE
9999 messages. To do so, the SPOE engine used to send messages must be defined, as
10000 well as the SPOE group to send. Of course, the SPOE engine must refer to an
10001 existing SPOE filter. If not engine name is provided on the SPOE filter line,
10002 the SPOE agent name must be used.
10003
10004 <engine-name> The SPOE engine name.
10005
10006 <group-name> The SPOE group name as specified in the engine configuration.
10007
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020010008 Example:
10009
10010 tcp-request content set-var(sess.my_var) src
Christopher Faulet85d79c92016-11-09 16:54:56 +010010011 tcp-request content unset-var(sess.my_var2)
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020010012
Willy Tarreau62644772008-07-16 18:36:06 +020010013 Example:
Willy Tarreaue9656522010-08-17 15:40:09 +020010014 # Accept HTTP requests containing a Host header saying "example.com"
10015 # and reject everything else.
10016 acl is_host_com hdr(Host) -i example.com
10017 tcp-request inspect-delay 30s
Willy Tarreauc0239e02012-04-16 14:42:55 +020010018 tcp-request content accept if is_host_com
Willy Tarreaue9656522010-08-17 15:40:09 +020010019 tcp-request content reject
10020
10021 Example:
Willy Tarreau62644772008-07-16 18:36:06 +020010022 # reject SMTP connection if client speaks first
10023 tcp-request inspect-delay 30s
10024 acl content_present req_len gt 0
Willy Tarreau68c03ab2010-08-06 15:08:45 +020010025 tcp-request content reject if content_present
Willy Tarreau62644772008-07-16 18:36:06 +020010026
10027 # Forward HTTPS connection only if client speaks
10028 tcp-request inspect-delay 30s
10029 acl content_present req_len gt 0
Willy Tarreau68c03ab2010-08-06 15:08:45 +020010030 tcp-request content accept if content_present
Willy Tarreaue9656522010-08-17 15:40:09 +020010031 tcp-request content reject
10032
Willy Tarreau5d5b5d82012-12-09 12:00:04 +010010033 Example:
Jarno Huuskonene5ae7022017-04-03 14:36:21 +030010034 # Track the last IP(stick-table type string) from X-Forwarded-For
Willy Tarreau5d5b5d82012-12-09 12:00:04 +010010035 tcp-request inspect-delay 10s
Willy Tarreau4d54c7c2014-09-16 15:48:15 +020010036 tcp-request content track-sc0 hdr(x-forwarded-for,-1)
Jarno Huuskonene5ae7022017-04-03 14:36:21 +030010037 # Or track the last IP(stick-table type ip|ipv6) from X-Forwarded-For
10038 tcp-request content track-sc0 req.hdr_ip(x-forwarded-for,-1)
Willy Tarreau5d5b5d82012-12-09 12:00:04 +010010039
10040 Example:
10041 # track request counts per "base" (concatenation of Host+URL)
10042 tcp-request inspect-delay 10s
Willy Tarreau4d54c7c2014-09-16 15:48:15 +020010043 tcp-request content track-sc0 base table req-rate
Willy Tarreau5d5b5d82012-12-09 12:00:04 +010010044
Willy Tarreaue9656522010-08-17 15:40:09 +020010045 Example: track per-frontend and per-backend counters, block abusers at the
Jarno Huuskonene5ae7022017-04-03 14:36:21 +030010046 frontend when the backend detects abuse(and marks gpc0).
Willy Tarreaue9656522010-08-17 15:40:09 +020010047
10048 frontend http
Davor Ocelice9ed2812017-12-25 17:49:28 +010010049 # Use General Purpose Counter 0 in SC0 as a global abuse counter
Willy Tarreaue9656522010-08-17 15:40:09 +020010050 # protecting all our sites
10051 stick-table type ip size 1m expire 5m store gpc0
Willy Tarreaube4a3ef2013-06-17 15:04:07 +020010052 tcp-request connection track-sc0 src
10053 tcp-request connection reject if { sc0_get_gpc0 gt 0 }
Willy Tarreaue9656522010-08-17 15:40:09 +020010054 ...
10055 use_backend http_dynamic if { path_end .php }
10056
10057 backend http_dynamic
10058 # if a source makes too fast requests to this dynamic site (tracked
Willy Tarreaube4a3ef2013-06-17 15:04:07 +020010059 # by SC1), block it globally in the frontend.
Willy Tarreaue9656522010-08-17 15:40:09 +020010060 stick-table type ip size 1m expire 5m store http_req_rate(10s)
Willy Tarreaube4a3ef2013-06-17 15:04:07 +020010061 acl click_too_fast sc1_http_req_rate gt 10
Jarno Huuskonene5ae7022017-04-03 14:36:21 +030010062 acl mark_as_abuser sc0_inc_gpc0(http) gt 0
Willy Tarreaube4a3ef2013-06-17 15:04:07 +020010063 tcp-request content track-sc1 src
Willy Tarreaue9656522010-08-17 15:40:09 +020010064 tcp-request content reject if click_too_fast mark_as_abuser
Willy Tarreau62644772008-07-16 18:36:06 +020010065
Willy Tarreauc57f0e22009-05-10 13:12:33 +020010066 See section 7 about ACL usage.
Willy Tarreau62644772008-07-16 18:36:06 +020010067
Jarno Huuskonen95b012b2017-04-06 13:59:14 +030010068 See also : "tcp-request connection", "tcp-request session",
10069 "tcp-request inspect-delay", and "http-request".
Willy Tarreau62644772008-07-16 18:36:06 +020010070
10071
10072tcp-request inspect-delay <timeout>
10073 Set the maximum allowed time to wait for data during content inspection
10074 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaufb356202010-08-03 14:02:05 +020010075 no | yes | yes | yes
Willy Tarreau62644772008-07-16 18:36:06 +020010076 Arguments :
10077 <timeout> is the timeout value specified in milliseconds by default, but
10078 can be in any other unit if the number is suffixed by the unit,
10079 as explained at the top of this document.
10080
10081 People using haproxy primarily as a TCP relay are often worried about the
10082 risk of passing any type of protocol to a server without any analysis. In
10083 order to be able to analyze the request contents, we must first withhold
10084 the data then analyze them. This statement simply enables withholding of
10085 data for at most the specified amount of time.
10086
Willy Tarreaufb356202010-08-03 14:02:05 +020010087 TCP content inspection applies very early when a connection reaches a
10088 frontend, then very early when the connection is forwarded to a backend. This
10089 means that a connection may experience a first delay in the frontend and a
10090 second delay in the backend if both have tcp-request rules.
10091
Willy Tarreau62644772008-07-16 18:36:06 +020010092 Note that when performing content inspection, haproxy will evaluate the whole
10093 rules for every new chunk which gets in, taking into account the fact that
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +010010094 those data are partial. If no rule matches before the aforementioned delay,
Willy Tarreau62644772008-07-16 18:36:06 +020010095 a last check is performed upon expiration, this time considering that the
Willy Tarreaud869b242009-03-15 14:43:58 +010010096 contents are definitive. If no delay is set, haproxy will not wait at all
10097 and will immediately apply a verdict based on the available information.
10098 Obviously this is unlikely to be very useful and might even be racy, so such
10099 setups are not recommended.
Willy Tarreau62644772008-07-16 18:36:06 +020010100
10101 As soon as a rule matches, the request is released and continues as usual. If
10102 the timeout is reached and no rule matches, the default policy will be to let
10103 it pass through unaffected.
10104
10105 For most protocols, it is enough to set it to a few seconds, as most clients
10106 send the full request immediately upon connection. Add 3 or more seconds to
10107 cover TCP retransmits but that's all. For some protocols, it may make sense
Willy Tarreaud72758d2010-01-12 10:42:19 +010010108 to use large values, for instance to ensure that the client never talks
Davor Ocelice9ed2812017-12-25 17:49:28 +010010109 before the server (e.g. SMTP), or to wait for a client to talk before passing
10110 data to the server (e.g. SSL). Note that the client timeout must cover at
Willy Tarreaub824b002010-09-29 16:36:16 +020010111 least the inspection delay, otherwise it will expire first. If the client
10112 closes the connection or if the buffer is full, the delay immediately expires
10113 since the contents will not be able to change anymore.
Willy Tarreau62644772008-07-16 18:36:06 +020010114
Willy Tarreau55165fe2009-05-10 12:02:55 +020010115 See also : "tcp-request content accept", "tcp-request content reject",
Willy Tarreau62644772008-07-16 18:36:06 +020010116 "timeout client".
10117
10118
Emeric Brun0a3b67f2010-09-24 15:34:53 +020010119tcp-response content <action> [{if | unless} <condition>]
10120 Perform an action on a session response depending on a layer 4-7 condition
10121 May be used in sections : defaults | frontend | listen | backend
10122 no | no | yes | yes
10123 Arguments :
Willy Tarreauc870bfd2015-09-28 18:47:38 +020010124 <action> defines the action to perform if the condition applies. See
10125 below.
Emeric Brun0a3b67f2010-09-24 15:34:53 +020010126
10127 <condition> is a standard layer 4-7 ACL-based condition (see section 7).
10128
Davor Ocelice9ed2812017-12-25 17:49:28 +010010129 Response contents can be analyzed at an early stage of response processing
Emeric Brun0a3b67f2010-09-24 15:34:53 +020010130 called "TCP content inspection". During this stage, ACL-based rules are
10131 evaluated every time the response contents are updated, until either an
Willy Tarreaucc1e04b2013-09-11 23:20:29 +020010132 "accept", "close" or a "reject" rule matches, or a TCP response inspection
10133 delay is set and expires with no matching rule.
Emeric Brun0a3b67f2010-09-24 15:34:53 +020010134
10135 Most often, these decisions will consider a protocol recognition or validity.
10136
10137 Content-based rules are evaluated in their exact declaration order. If no
10138 rule matches or if there is no rule, the default action is to accept the
10139 contents. There is no specific limit to the number of rules which may be
10140 inserted.
10141
Thierry FOURNIER236657b2015-08-19 08:25:14 +020010142 Several types of actions are supported :
Emeric Brun0a3b67f2010-09-24 15:34:53 +020010143 - accept :
10144 accepts the response if the condition is true (when used with "if")
10145 or false (when used with "unless"). The first such rule executed ends
10146 the rules evaluation.
10147
Willy Tarreaucc1e04b2013-09-11 23:20:29 +020010148 - close :
10149 immediately closes the connection with the server if the condition is
10150 true (when used with "if"), or false (when used with "unless"). The
10151 first such rule executed ends the rules evaluation. The main purpose of
10152 this action is to force a connection to be finished between a client
10153 and a server after an exchange when the application protocol expects
10154 some long time outs to elapse first. The goal is to eliminate idle
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030010155 connections which take significant resources on servers with certain
Willy Tarreaucc1e04b2013-09-11 23:20:29 +020010156 protocols.
10157
Emeric Brun0a3b67f2010-09-24 15:34:53 +020010158 - reject :
10159 rejects the response if the condition is true (when used with "if")
10160 or false (when used with "unless"). The first such rule executed ends
Jamie Gloudonaaa21002012-08-25 00:18:33 -040010161 the rules evaluation. Rejected session are immediately closed.
Emeric Brun0a3b67f2010-09-24 15:34:53 +020010162
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020010163 - set-var(<var-name>) <expr>
10164 Sets a variable.
10165
Christopher Faulet85d79c92016-11-09 16:54:56 +010010166 - unset-var(<var-name>)
10167 Unsets a variable.
10168
Thierry FOURNIERe0627bd2015-08-04 08:20:33 +020010169 - sc-inc-gpc0(<sc-id>):
10170 This action increments the GPC0 counter according to the sticky
10171 counter designated by <sc-id>. If an error occurs, this action fails
10172 silently and the actions evaluation continues.
10173
Frédéric Lécaille6778b272018-01-29 15:22:53 +010010174 - sc-inc-gpc1(<sc-id>):
10175 This action increments the GPC1 counter according to the sticky
10176 counter designated by <sc-id>. If an error occurs, this action fails
10177 silently and the actions evaluation continues.
10178
Thierry FOURNIER236657b2015-08-19 08:25:14 +020010179 - sc-set-gpt0(<sc-id>) <int> :
10180 This action sets the GPT0 tag according to the sticky counter designated
10181 by <sc-id> and the value of <int>. The expected result is a boolean. If
10182 an error occurs, this action silently fails and the actions evaluation
10183 continues.
10184
Willy Tarreau2d392c22015-08-24 01:43:45 +020010185 - "silent-drop" :
10186 This stops the evaluation of the rules and makes the client-facing
Davor Ocelice9ed2812017-12-25 17:49:28 +010010187 connection suddenly disappear using a system-dependent way that tries
Willy Tarreau2d392c22015-08-24 01:43:45 +020010188 to prevent the client from being notified. The effect it then that the
10189 client still sees an established connection while there's none on
10190 HAProxy. The purpose is to achieve a comparable effect to "tarpit"
10191 except that it doesn't use any local resource at all on the machine
10192 running HAProxy. It can resist much higher loads than "tarpit", and
Davor Ocelice9ed2812017-12-25 17:49:28 +010010193 slow down stronger attackers. It is important to understand the impact
10194 of using this mechanism. All stateful equipment placed between the
Willy Tarreau2d392c22015-08-24 01:43:45 +020010195 client and HAProxy (firewalls, proxies, load balancers) will also keep
10196 the established connection for a long time and may suffer from this
Davor Ocelice9ed2812017-12-25 17:49:28 +010010197 action. On modern Linux systems running with enough privileges, the
Willy Tarreau2d392c22015-08-24 01:43:45 +020010198 TCP_REPAIR socket option is used to block the emission of a TCP
10199 reset. On other systems, the socket's TTL is reduced to 1 so that the
10200 TCP reset doesn't pass the first router, though it's still delivered to
10201 local networks. Do not use it unless you fully understand how it works.
10202
Christopher Faulet76c09ef2017-09-21 11:03:52 +020010203 - send-spoe-group <engine-name> <group-name>
10204 Send a group of SPOE messages.
10205
Emeric Brun0a3b67f2010-09-24 15:34:53 +020010206 Note that the "if/unless" condition is optional. If no condition is set on
10207 the action, it is simply performed unconditionally. That can be useful for
10208 for changing the default action to a reject.
10209
Jamie Gloudonaaa21002012-08-25 00:18:33 -040010210 It is perfectly possible to match layer 7 contents with "tcp-response
10211 content" rules, but then it is important to ensure that a full response has
10212 been buffered, otherwise no contents will match. In order to achieve this,
10213 the best solution involves detecting the HTTP protocol during the inspection
Emeric Brun0a3b67f2010-09-24 15:34:53 +020010214 period.
10215
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020010216 The "set-var" is used to set the content of a variable. The variable is
10217 declared inline.
10218
Daniel Schneller0b547052016-03-21 20:46:57 +010010219 <var-name> The name of the variable starts with an indication about
10220 its scope. The scopes allowed are:
Christopher Fauletff2613e2016-11-09 11:36:17 +010010221 "proc" : the variable is shared with the whole process
Daniel Schneller0b547052016-03-21 20:46:57 +010010222 "sess" : the variable is shared with the whole session
10223 "txn" : the variable is shared with the transaction
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020010224 (request and response)
Daniel Schneller0b547052016-03-21 20:46:57 +010010225 "req" : the variable is shared only during request
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020010226 processing
Daniel Schneller0b547052016-03-21 20:46:57 +010010227 "res" : the variable is shared only during response
10228 processing
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020010229 This prefix is followed by a name. The separator is a '.'.
Christopher Fauletb71557a2016-10-31 10:49:03 +010010230 The name may only contain characters 'a-z', 'A-Z', '0-9',
10231 '.' and '_'.
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020010232
10233 <expr> Is a standard HAProxy expression formed by a sample-fetch
10234 followed by some converters.
10235
10236 Example:
10237
10238 tcp-request content set-var(sess.my_var) src
10239
Christopher Faulet85d79c92016-11-09 16:54:56 +010010240 The "unset-var" is used to unset a variable. See above for details about
10241 <var-name>.
10242
10243 Example:
10244
10245 tcp-request content unset-var(sess.my_var)
10246
Christopher Faulet76c09ef2017-09-21 11:03:52 +020010247 The "send-spoe-group" is used to trigger sending of a group of SPOE
10248 messages. To do so, the SPOE engine used to send messages must be defined, as
10249 well as the SPOE group to send. Of course, the SPOE engine must refer to an
10250 existing SPOE filter. If not engine name is provided on the SPOE filter line,
10251 the SPOE agent name must be used.
10252
10253 <engine-name> The SPOE engine name.
10254
10255 <group-name> The SPOE group name as specified in the engine configuration.
10256
Emeric Brun0a3b67f2010-09-24 15:34:53 +020010257 See section 7 about ACL usage.
10258
10259 See also : "tcp-request content", "tcp-response inspect-delay"
10260
10261
Willy Tarreau4f614292016-10-21 17:49:36 +020010262tcp-request session <action> [{if | unless} <condition>]
10263 Perform an action on a validated session depending on a layer 5 condition
10264 May be used in sections : defaults | frontend | listen | backend
10265 no | yes | yes | no
10266 Arguments :
10267 <action> defines the action to perform if the condition applies. See
10268 below.
10269
10270 <condition> is a standard layer5-only ACL-based condition (see section 7).
10271
Davor Ocelice9ed2812017-12-25 17:49:28 +010010272 Once a session is validated, (i.e. after all handshakes have been completed),
Willy Tarreau4f614292016-10-21 17:49:36 +020010273 it is possible to evaluate some conditions to decide whether this session
10274 must be accepted or dropped or have its counters tracked. Those conditions
10275 cannot make use of any data contents because no buffers are allocated yet and
10276 the processing cannot wait at this stage. The main use case it to copy some
10277 early information into variables (since variables are accessible in the
10278 session), or to keep track of some information collected after the handshake,
10279 such as SSL-level elements (SNI, ciphers, client cert's CN) or information
Davor Ocelice9ed2812017-12-25 17:49:28 +010010280 from the PROXY protocol header (e.g. track a source forwarded this way). The
Willy Tarreau4f614292016-10-21 17:49:36 +020010281 extracted information can thus be copied to a variable or tracked using
10282 "track-sc" rules. Of course it is also possible to decide to accept/reject as
10283 with other rulesets. Most operations performed here could also be performed
10284 in "tcp-request content" rules, except that in HTTP these rules are evaluated
10285 for each new request, and that might not always be acceptable. For example a
10286 rule might increment a counter on each evaluation. It would also be possible
10287 that a country is resolved by geolocation from the source IP address,
10288 assigned to a session-wide variable, then the source address rewritten from
10289 an HTTP header for all requests. If some contents need to be inspected in
10290 order to take the decision, the "tcp-request content" statements must be used
10291 instead.
10292
10293 The "tcp-request session" rules are evaluated in their exact declaration
10294 order. If no rule matches or if there is no rule, the default action is to
10295 accept the incoming session. There is no specific limit to the number of
10296 rules which may be inserted.
10297
10298 Several types of actions are supported :
10299 - accept : the request is accepted
10300 - reject : the request is rejected and the connection is closed
10301 - { track-sc0 | track-sc1 | track-sc2 } <key> [table <table>]
10302 - sc-inc-gpc0(<sc-id>)
Frédéric Lécaille6778b272018-01-29 15:22:53 +010010303 - sc-inc-gpc1(<sc-id>)
Willy Tarreau4f614292016-10-21 17:49:36 +020010304 - sc-set-gpt0(<sc-id>) <int>
10305 - set-var(<var-name>) <expr>
Christopher Faulet85d79c92016-11-09 16:54:56 +010010306 - unset-var(<var-name>)
Willy Tarreau4f614292016-10-21 17:49:36 +020010307 - silent-drop
10308
10309 These actions have the same meaning as their respective counter-parts in
10310 "tcp-request connection" and "tcp-request content", so please refer to these
10311 sections for a complete description.
10312
10313 Note that the "if/unless" condition is optional. If no condition is set on
10314 the action, it is simply performed unconditionally. That can be useful for
10315 "track-sc*" actions as well as for changing the default action to a reject.
10316
10317 Example: track the original source address by default, or the one advertised
10318 in the PROXY protocol header for connection coming from the local
10319 proxies. The first connection-level rule enables receipt of the
10320 PROXY protocol for these ones, the second rule tracks whatever
10321 address we decide to keep after optional decoding.
10322
10323 tcp-request connection expect-proxy layer4 if { src -f proxies.lst }
10324 tcp-request session track-sc0 src
10325
10326 Example: accept all sessions from white-listed hosts, reject too fast
10327 sessions without counting them, and track accepted sessions.
10328 This results in session rate being capped from abusive sources.
10329
10330 tcp-request session accept if { src -f /etc/haproxy/whitelist.lst }
10331 tcp-request session reject if { src_sess_rate gt 10 }
10332 tcp-request session track-sc0 src
10333
10334 Example: accept all sessions from white-listed hosts, count all other
10335 sessions and reject too fast ones. This results in abusive ones
10336 being blocked as long as they don't slow down.
10337
10338 tcp-request session accept if { src -f /etc/haproxy/whitelist.lst }
10339 tcp-request session track-sc0 src
10340 tcp-request session reject if { sc0_sess_rate gt 10 }
10341
10342 See section 7 about ACL usage.
10343
10344 See also : "tcp-request connection", "tcp-request content", "stick-table"
10345
10346
Emeric Brun0a3b67f2010-09-24 15:34:53 +020010347tcp-response inspect-delay <timeout>
10348 Set the maximum allowed time to wait for a response during content inspection
10349 May be used in sections : defaults | frontend | listen | backend
10350 no | no | yes | yes
10351 Arguments :
10352 <timeout> is the timeout value specified in milliseconds by default, but
10353 can be in any other unit if the number is suffixed by the unit,
10354 as explained at the top of this document.
10355
10356 See also : "tcp-response content", "tcp-request inspect-delay".
10357
10358
Krzysztof Piotr Oledzki5259dfe2008-01-21 01:54:06 +010010359timeout check <timeout>
10360 Set additional check timeout, but only after a connection has been already
10361 established.
10362
10363 May be used in sections: defaults | frontend | listen | backend
10364 yes | no | yes | yes
10365 Arguments:
10366 <timeout> is the timeout value specified in milliseconds by default, but
10367 can be in any other unit if the number is suffixed by the unit,
10368 as explained at the top of this document.
10369
10370 If set, haproxy uses min("timeout connect", "inter") as a connect timeout
10371 for check and "timeout check" as an additional read timeout. The "min" is
Davor Ocelice9ed2812017-12-25 17:49:28 +010010372 used so that people running with *very* long "timeout connect" (e.g. those
Krzysztof Piotr Oledzki5259dfe2008-01-21 01:54:06 +010010373 who needed this due to the queue or tarpit) do not slow down their checks.
Willy Tarreaud7550a22010-02-10 05:10:19 +010010374 (Please also note that there is no valid reason to have such long connect
10375 timeouts, because "timeout queue" and "timeout tarpit" can always be used to
10376 avoid that).
Krzysztof Piotr Oledzki5259dfe2008-01-21 01:54:06 +010010377
10378 If "timeout check" is not set haproxy uses "inter" for complete check
10379 timeout (connect + read) exactly like all <1.3.15 version.
10380
10381 In most cases check request is much simpler and faster to handle than normal
10382 requests and people may want to kick out laggy servers so this timeout should
Willy Tarreau41a340d2008-01-22 12:25:31 +010010383 be smaller than "timeout server".
Krzysztof Piotr Oledzki5259dfe2008-01-21 01:54:06 +010010384
10385 This parameter is specific to backends, but can be specified once for all in
10386 "defaults" sections. This is in fact one of the easiest solutions not to
10387 forget about it.
10388
Willy Tarreau41a340d2008-01-22 12:25:31 +010010389 See also: "timeout connect", "timeout queue", "timeout server",
10390 "timeout tarpit".
Krzysztof Piotr Oledzki5259dfe2008-01-21 01:54:06 +010010391
10392
Willy Tarreau0ba27502007-12-24 16:55:16 +010010393timeout client <timeout>
Willy Tarreau0ba27502007-12-24 16:55:16 +010010394 Set the maximum inactivity time on the client side.
10395 May be used in sections : defaults | frontend | listen | backend
10396 yes | yes | yes | no
10397 Arguments :
Willy Tarreau844e3c52008-01-11 16:28:18 +010010398 <timeout> is the timeout value specified in milliseconds by default, but
Willy Tarreau0ba27502007-12-24 16:55:16 +010010399 can be in any other unit if the number is suffixed by the unit,
10400 as explained at the top of this document.
10401
10402 The inactivity timeout applies when the client is expected to acknowledge or
10403 send data. In HTTP mode, this timeout is particularly important to consider
10404 during the first phase, when the client sends the request, and during the
Baptiste Assmann2e1941e2016-03-06 23:24:12 +010010405 response while it is reading data sent by the server. That said, for the
10406 first phase, it is preferable to set the "timeout http-request" to better
10407 protect HAProxy from Slowloris like attacks. The value is specified in
10408 milliseconds by default, but can be in any other unit if the number is
Willy Tarreau0ba27502007-12-24 16:55:16 +010010409 suffixed by the unit, as specified at the top of this document. In TCP mode
10410 (and to a lesser extent, in HTTP mode), it is highly recommended that the
10411 client timeout remains equal to the server timeout in order to avoid complex
Willy Tarreaud2a4aa22008-01-31 15:28:22 +010010412 situations to debug. It is a good practice to cover one or several TCP packet
Willy Tarreau0ba27502007-12-24 16:55:16 +010010413 losses by specifying timeouts that are slightly above multiples of 3 seconds
Davor Ocelice9ed2812017-12-25 17:49:28 +010010414 (e.g. 4 or 5 seconds). If some long-lived sessions are mixed with short-lived
10415 sessions (e.g. WebSocket and HTTP), it's worth considering "timeout tunnel",
Willy Tarreau05cdd962014-05-10 14:30:07 +020010416 which overrides "timeout client" and "timeout server" for tunnels, as well as
10417 "timeout client-fin" for half-closed connections.
Willy Tarreau0ba27502007-12-24 16:55:16 +010010418
10419 This parameter is specific to frontends, but can be specified once for all in
10420 "defaults" sections. This is in fact one of the easiest solutions not to
10421 forget about it. An unspecified timeout results in an infinite timeout, which
10422 is not recommended. Such a usage is accepted and works but reports a warning
John Roeslerfb2fce12019-07-10 15:45:51 -050010423 during startup because it may result in accumulation of expired sessions in
Willy Tarreau0ba27502007-12-24 16:55:16 +010010424 the system if the system's timeouts are not configured either.
10425
Willy Tarreau95c4e142017-11-26 12:18:55 +010010426 This also applies to HTTP/2 connections, which will be closed with GOAWAY.
Lukas Tribus75df9d72017-11-24 19:05:12 +010010427
Tim Duesterhus86e6b6e2019-05-14 20:57:59 +020010428 See also : "timeout server", "timeout tunnel", "timeout http-request".
Willy Tarreau0ba27502007-12-24 16:55:16 +010010429
Willy Tarreau0ba27502007-12-24 16:55:16 +010010430
Willy Tarreau05cdd962014-05-10 14:30:07 +020010431timeout client-fin <timeout>
10432 Set the inactivity timeout on the client side for half-closed connections.
10433 May be used in sections : defaults | frontend | listen | backend
10434 yes | yes | yes | no
10435 Arguments :
10436 <timeout> is the timeout value specified in milliseconds by default, but
10437 can be in any other unit if the number is suffixed by the unit,
10438 as explained at the top of this document.
10439
10440 The inactivity timeout applies when the client is expected to acknowledge or
10441 send data while one direction is already shut down. This timeout is different
10442 from "timeout client" in that it only applies to connections which are closed
10443 in one direction. This is particularly useful to avoid keeping connections in
10444 FIN_WAIT state for too long when clients do not disconnect cleanly. This
10445 problem is particularly common long connections such as RDP or WebSocket.
10446 Note that this timeout can override "timeout tunnel" when a connection shuts
Willy Tarreau599391a2017-11-24 10:16:00 +010010447 down in one direction. It is applied to idle HTTP/2 connections once a GOAWAY
10448 frame was sent, often indicating an expectation that the connection quickly
10449 ends.
Willy Tarreau05cdd962014-05-10 14:30:07 +020010450
10451 This parameter is specific to frontends, but can be specified once for all in
10452 "defaults" sections. By default it is not set, so half-closed connections
10453 will use the other timeouts (timeout.client or timeout.tunnel).
10454
10455 See also : "timeout client", "timeout server-fin", and "timeout tunnel".
10456
10457
Willy Tarreau0ba27502007-12-24 16:55:16 +010010458timeout connect <timeout>
Willy Tarreau0ba27502007-12-24 16:55:16 +010010459 Set the maximum time to wait for a connection attempt to a server to succeed.
10460 May be used in sections : defaults | frontend | listen | backend
10461 yes | no | yes | yes
10462 Arguments :
Willy Tarreau844e3c52008-01-11 16:28:18 +010010463 <timeout> is the timeout value specified in milliseconds by default, but
Willy Tarreau0ba27502007-12-24 16:55:16 +010010464 can be in any other unit if the number is suffixed by the unit,
10465 as explained at the top of this document.
10466
10467 If the server is located on the same LAN as haproxy, the connection should be
Willy Tarreaud2a4aa22008-01-31 15:28:22 +010010468 immediate (less than a few milliseconds). Anyway, it is a good practice to
Willy Tarreaud72758d2010-01-12 10:42:19 +010010469 cover one or several TCP packet losses by specifying timeouts that are
Davor Ocelice9ed2812017-12-25 17:49:28 +010010470 slightly above multiples of 3 seconds (e.g. 4 or 5 seconds). By default, the
Krzysztof Piotr Oledzki5259dfe2008-01-21 01:54:06 +010010471 connect timeout also presets both queue and tarpit timeouts to the same value
10472 if these have not been specified.
Willy Tarreau0ba27502007-12-24 16:55:16 +010010473
10474 This parameter is specific to backends, but can be specified once for all in
10475 "defaults" sections. This is in fact one of the easiest solutions not to
10476 forget about it. An unspecified timeout results in an infinite timeout, which
10477 is not recommended. Such a usage is accepted and works but reports a warning
John Roeslerfb2fce12019-07-10 15:45:51 -050010478 during startup because it may result in accumulation of failed sessions in
Willy Tarreau0ba27502007-12-24 16:55:16 +010010479 the system if the system's timeouts are not configured either.
10480
Tim Duesterhus86e6b6e2019-05-14 20:57:59 +020010481 See also: "timeout check", "timeout queue", "timeout server", "timeout tarpit".
Willy Tarreau0ba27502007-12-24 16:55:16 +010010482
Willy Tarreau0ba27502007-12-24 16:55:16 +010010483
Willy Tarreaub16a5742010-01-10 14:46:16 +010010484timeout http-keep-alive <timeout>
10485 Set the maximum allowed time to wait for a new HTTP request to appear
10486 May be used in sections : defaults | frontend | listen | backend
10487 yes | yes | yes | yes
10488 Arguments :
10489 <timeout> is the timeout value specified in milliseconds by default, but
10490 can be in any other unit if the number is suffixed by the unit,
10491 as explained at the top of this document.
10492
10493 By default, the time to wait for a new request in case of keep-alive is set
10494 by "timeout http-request". However this is not always convenient because some
10495 people want very short keep-alive timeouts in order to release connections
10496 faster, and others prefer to have larger ones but still have short timeouts
10497 once the request has started to present itself.
10498
10499 The "http-keep-alive" timeout covers these needs. It will define how long to
10500 wait for a new HTTP request to start coming after a response was sent. Once
10501 the first byte of request has been seen, the "http-request" timeout is used
10502 to wait for the complete request to come. Note that empty lines prior to a
10503 new request do not refresh the timeout and are not counted as a new request.
10504
10505 There is also another difference between the two timeouts : when a connection
10506 expires during timeout http-keep-alive, no error is returned, the connection
10507 just closes. If the connection expires in "http-request" while waiting for a
10508 connection to complete, a HTTP 408 error is returned.
10509
10510 In general it is optimal to set this value to a few tens to hundreds of
10511 milliseconds, to allow users to fetch all objects of a page at once but
Davor Ocelice9ed2812017-12-25 17:49:28 +010010512 without waiting for further clicks. Also, if set to a very small value (e.g.
Willy Tarreaub16a5742010-01-10 14:46:16 +010010513 1 millisecond) it will probably only accept pipelined requests but not the
10514 non-pipelined ones. It may be a nice trade-off for very large sites running
Patrick Mézard2382ad62010-05-09 10:43:32 +020010515 with tens to hundreds of thousands of clients.
Willy Tarreaub16a5742010-01-10 14:46:16 +010010516
10517 If this parameter is not set, the "http-request" timeout applies, and if both
10518 are not set, "timeout client" still applies at the lower level. It should be
10519 set in the frontend to take effect, unless the frontend is in TCP mode, in
10520 which case the HTTP backend's timeout will be used.
10521
Willy Tarreau95c4e142017-11-26 12:18:55 +010010522 When using HTTP/2 "timeout client" is applied instead. This is so we can keep
10523 using short keep-alive timeouts in HTTP/1.1 while using longer ones in HTTP/2
Lukas Tribus75df9d72017-11-24 19:05:12 +010010524 (where we only have one connection per client and a connection setup).
10525
Willy Tarreaub16a5742010-01-10 14:46:16 +010010526 See also : "timeout http-request", "timeout client".
10527
10528
Willy Tarreau036fae02008-01-06 13:24:40 +010010529timeout http-request <timeout>
10530 Set the maximum allowed time to wait for a complete HTTP request
10531 May be used in sections : defaults | frontend | listen | backend
Willy Tarreaucd7afc02009-07-12 10:03:17 +020010532 yes | yes | yes | yes
Willy Tarreau036fae02008-01-06 13:24:40 +010010533 Arguments :
Willy Tarreau844e3c52008-01-11 16:28:18 +010010534 <timeout> is the timeout value specified in milliseconds by default, but
Willy Tarreau036fae02008-01-06 13:24:40 +010010535 can be in any other unit if the number is suffixed by the unit,
10536 as explained at the top of this document.
10537
10538 In order to offer DoS protection, it may be required to lower the maximum
10539 accepted time to receive a complete HTTP request without affecting the client
10540 timeout. This helps protecting against established connections on which
10541 nothing is sent. The client timeout cannot offer a good protection against
10542 this abuse because it is an inactivity timeout, which means that if the
10543 attacker sends one character every now and then, the timeout will not
10544 trigger. With the HTTP request timeout, no matter what speed the client
Willy Tarreau2705a612014-05-23 17:38:34 +020010545 types, the request will be aborted if it does not complete in time. When the
10546 timeout expires, an HTTP 408 response is sent to the client to inform it
10547 about the problem, and the connection is closed. The logs will report
10548 termination codes "cR". Some recent browsers are having problems with this
Davor Ocelice9ed2812017-12-25 17:49:28 +010010549 standard, well-documented behavior, so it might be needed to hide the 408
Willy Tarreau0f228a02015-05-01 15:37:53 +020010550 code using "option http-ignore-probes" or "errorfile 408 /dev/null". See
10551 more details in the explanations of the "cR" termination code in section 8.5.
Willy Tarreau036fae02008-01-06 13:24:40 +010010552
Baptiste Assmanneccdf432015-10-28 13:49:01 +010010553 By default, this timeout only applies to the header part of the request,
10554 and not to any data. As soon as the empty line is received, this timeout is
10555 not used anymore. When combined with "option http-buffer-request", this
10556 timeout also applies to the body of the request..
10557 It is used again on keep-alive connections to wait for a second
Willy Tarreaub16a5742010-01-10 14:46:16 +010010558 request if "timeout http-keep-alive" is not set.
Willy Tarreau036fae02008-01-06 13:24:40 +010010559
10560 Generally it is enough to set it to a few seconds, as most clients send the
10561 full request immediately upon connection. Add 3 or more seconds to cover TCP
Davor Ocelice9ed2812017-12-25 17:49:28 +010010562 retransmits but that's all. Setting it to very low values (e.g. 50 ms) will
Willy Tarreau036fae02008-01-06 13:24:40 +010010563 generally work on local networks as long as there are no packet losses. This
10564 will prevent people from sending bare HTTP requests using telnet.
10565
10566 If this parameter is not set, the client timeout still applies between each
Willy Tarreaucd7afc02009-07-12 10:03:17 +020010567 chunk of the incoming request. It should be set in the frontend to take
10568 effect, unless the frontend is in TCP mode, in which case the HTTP backend's
10569 timeout will be used.
Willy Tarreau036fae02008-01-06 13:24:40 +010010570
Willy Tarreau0f228a02015-05-01 15:37:53 +020010571 See also : "errorfile", "http-ignore-probes", "timeout http-keep-alive", and
Baptiste Assmanneccdf432015-10-28 13:49:01 +010010572 "timeout client", "option http-buffer-request".
Willy Tarreau036fae02008-01-06 13:24:40 +010010573
Willy Tarreau844e3c52008-01-11 16:28:18 +010010574
10575timeout queue <timeout>
10576 Set the maximum time to wait in the queue for a connection slot to be free
10577 May be used in sections : defaults | frontend | listen | backend
10578 yes | no | yes | yes
10579 Arguments :
10580 <timeout> is the timeout value specified in milliseconds by default, but
10581 can be in any other unit if the number is suffixed by the unit,
10582 as explained at the top of this document.
10583
10584 When a server's maxconn is reached, connections are left pending in a queue
10585 which may be server-specific or global to the backend. In order not to wait
10586 indefinitely, a timeout is applied to requests pending in the queue. If the
10587 timeout is reached, it is considered that the request will almost never be
10588 served, so it is dropped and a 503 error is returned to the client.
10589
10590 The "timeout queue" statement allows to fix the maximum time for a request to
10591 be left pending in a queue. If unspecified, the same value as the backend's
10592 connection timeout ("timeout connect") is used, for backwards compatibility
10593 with older versions with no "timeout queue" parameter.
10594
Tim Duesterhus86e6b6e2019-05-14 20:57:59 +020010595 See also : "timeout connect".
Willy Tarreau844e3c52008-01-11 16:28:18 +010010596
10597
10598timeout server <timeout>
Willy Tarreau844e3c52008-01-11 16:28:18 +010010599 Set the maximum inactivity time on the server side.
10600 May be used in sections : defaults | frontend | listen | backend
10601 yes | no | yes | yes
10602 Arguments :
10603 <timeout> is the timeout value specified in milliseconds by default, but
10604 can be in any other unit if the number is suffixed by the unit,
10605 as explained at the top of this document.
10606
10607 The inactivity timeout applies when the server is expected to acknowledge or
10608 send data. In HTTP mode, this timeout is particularly important to consider
10609 during the first phase of the server's response, when it has to send the
10610 headers, as it directly represents the server's processing time for the
10611 request. To find out what value to put there, it's often good to start with
10612 what would be considered as unacceptable response times, then check the logs
10613 to observe the response time distribution, and adjust the value accordingly.
10614
10615 The value is specified in milliseconds by default, but can be in any other
10616 unit if the number is suffixed by the unit, as specified at the top of this
10617 document. In TCP mode (and to a lesser extent, in HTTP mode), it is highly
10618 recommended that the client timeout remains equal to the server timeout in
10619 order to avoid complex situations to debug. Whatever the expected server
Willy Tarreaud2a4aa22008-01-31 15:28:22 +010010620 response times, it is a good practice to cover at least one or several TCP
Willy Tarreau844e3c52008-01-11 16:28:18 +010010621 packet losses by specifying timeouts that are slightly above multiples of 3
Davor Ocelice9ed2812017-12-25 17:49:28 +010010622 seconds (e.g. 4 or 5 seconds minimum). If some long-lived sessions are mixed
10623 with short-lived sessions (e.g. WebSocket and HTTP), it's worth considering
Willy Tarreauce887fd2012-05-12 12:50:00 +020010624 "timeout tunnel", which overrides "timeout client" and "timeout server" for
10625 tunnels.
Willy Tarreau844e3c52008-01-11 16:28:18 +010010626
10627 This parameter is specific to backends, but can be specified once for all in
10628 "defaults" sections. This is in fact one of the easiest solutions not to
10629 forget about it. An unspecified timeout results in an infinite timeout, which
10630 is not recommended. Such a usage is accepted and works but reports a warning
John Roeslerfb2fce12019-07-10 15:45:51 -050010631 during startup because it may result in accumulation of expired sessions in
Willy Tarreau844e3c52008-01-11 16:28:18 +010010632 the system if the system's timeouts are not configured either.
10633
Tim Duesterhus86e6b6e2019-05-14 20:57:59 +020010634 See also : "timeout client" and "timeout tunnel".
Willy Tarreau844e3c52008-01-11 16:28:18 +010010635
Willy Tarreau05cdd962014-05-10 14:30:07 +020010636
10637timeout server-fin <timeout>
10638 Set the inactivity timeout on the server side for half-closed connections.
10639 May be used in sections : defaults | frontend | listen | backend
10640 yes | no | yes | yes
10641 Arguments :
10642 <timeout> is the timeout value specified in milliseconds by default, but
10643 can be in any other unit if the number is suffixed by the unit,
10644 as explained at the top of this document.
10645
10646 The inactivity timeout applies when the server is expected to acknowledge or
10647 send data while one direction is already shut down. This timeout is different
10648 from "timeout server" in that it only applies to connections which are closed
10649 in one direction. This is particularly useful to avoid keeping connections in
10650 FIN_WAIT state for too long when a remote server does not disconnect cleanly.
10651 This problem is particularly common long connections such as RDP or WebSocket.
10652 Note that this timeout can override "timeout tunnel" when a connection shuts
10653 down in one direction. This setting was provided for completeness, but in most
10654 situations, it should not be needed.
10655
10656 This parameter is specific to backends, but can be specified once for all in
10657 "defaults" sections. By default it is not set, so half-closed connections
10658 will use the other timeouts (timeout.server or timeout.tunnel).
10659
10660 See also : "timeout client-fin", "timeout server", and "timeout tunnel".
10661
Willy Tarreau844e3c52008-01-11 16:28:18 +010010662
10663timeout tarpit <timeout>
Cyril Bonté78caf842010-03-10 22:41:43 +010010664 Set the duration for which tarpitted connections will be maintained
Willy Tarreau844e3c52008-01-11 16:28:18 +010010665 May be used in sections : defaults | frontend | listen | backend
10666 yes | yes | yes | yes
10667 Arguments :
10668 <timeout> is the tarpit duration specified in milliseconds by default, but
10669 can be in any other unit if the number is suffixed by the unit,
10670 as explained at the top of this document.
10671
Jarno Huuskonene5ae7022017-04-03 14:36:21 +030010672 When a connection is tarpitted using "http-request tarpit" or
10673 "reqtarpit", it is maintained open with no activity for a certain
10674 amount of time, then closed. "timeout tarpit" defines how long it will
10675 be maintained open.
Willy Tarreau844e3c52008-01-11 16:28:18 +010010676
10677 The value is specified in milliseconds by default, but can be in any other
10678 unit if the number is suffixed by the unit, as specified at the top of this
10679 document. If unspecified, the same value as the backend's connection timeout
10680 ("timeout connect") is used, for backwards compatibility with older versions
Cyril Bonté78caf842010-03-10 22:41:43 +010010681 with no "timeout tarpit" parameter.
Willy Tarreau844e3c52008-01-11 16:28:18 +010010682
Tim Duesterhus86e6b6e2019-05-14 20:57:59 +020010683 See also : "timeout connect".
Willy Tarreau844e3c52008-01-11 16:28:18 +010010684
10685
Willy Tarreauce887fd2012-05-12 12:50:00 +020010686timeout tunnel <timeout>
10687 Set the maximum inactivity time on the client and server side for tunnels.
10688 May be used in sections : defaults | frontend | listen | backend
10689 yes | no | yes | yes
10690 Arguments :
10691 <timeout> is the timeout value specified in milliseconds by default, but
10692 can be in any other unit if the number is suffixed by the unit,
10693 as explained at the top of this document.
10694
Jamie Gloudonaaa21002012-08-25 00:18:33 -040010695 The tunnel timeout applies when a bidirectional connection is established
Willy Tarreauce887fd2012-05-12 12:50:00 +020010696 between a client and a server, and the connection remains inactive in both
10697 directions. This timeout supersedes both the client and server timeouts once
10698 the connection becomes a tunnel. In TCP, this timeout is used as soon as no
Davor Ocelice9ed2812017-12-25 17:49:28 +010010699 analyzer remains attached to either connection (e.g. tcp content rules are
10700 accepted). In HTTP, this timeout is used when a connection is upgraded (e.g.
Willy Tarreauce887fd2012-05-12 12:50:00 +020010701 when switching to the WebSocket protocol, or forwarding a CONNECT request
10702 to a proxy), or after the first response when no keepalive/close option is
10703 specified.
10704
Willy Tarreau05cdd962014-05-10 14:30:07 +020010705 Since this timeout is usually used in conjunction with long-lived connections,
10706 it usually is a good idea to also set "timeout client-fin" to handle the
10707 situation where a client suddenly disappears from the net and does not
10708 acknowledge a close, or sends a shutdown and does not acknowledge pending
10709 data anymore. This can happen in lossy networks where firewalls are present,
10710 and is detected by the presence of large amounts of sessions in a FIN_WAIT
10711 state.
10712
Willy Tarreauce887fd2012-05-12 12:50:00 +020010713 The value is specified in milliseconds by default, but can be in any other
10714 unit if the number is suffixed by the unit, as specified at the top of this
10715 document. Whatever the expected normal idle time, it is a good practice to
10716 cover at least one or several TCP packet losses by specifying timeouts that
Davor Ocelice9ed2812017-12-25 17:49:28 +010010717 are slightly above multiples of 3 seconds (e.g. 4 or 5 seconds minimum).
Willy Tarreauce887fd2012-05-12 12:50:00 +020010718
10719 This parameter is specific to backends, but can be specified once for all in
10720 "defaults" sections. This is in fact one of the easiest solutions not to
10721 forget about it.
10722
10723 Example :
10724 defaults http
10725 option http-server-close
10726 timeout connect 5s
10727 timeout client 30s
Willy Tarreau05cdd962014-05-10 14:30:07 +020010728 timeout client-fin 30s
Willy Tarreauce887fd2012-05-12 12:50:00 +020010729 timeout server 30s
10730 timeout tunnel 1h # timeout to use with WebSocket and CONNECT
10731
Willy Tarreau05cdd962014-05-10 14:30:07 +020010732 See also : "timeout client", "timeout client-fin", "timeout server".
Willy Tarreauce887fd2012-05-12 12:50:00 +020010733
10734
Willy Tarreau844e3c52008-01-11 16:28:18 +010010735transparent (deprecated)
10736 Enable client-side transparent proxying
10737 May be used in sections : defaults | frontend | listen | backend
Willy Tarreau4b1f8592008-12-23 23:13:55 +010010738 yes | no | yes | yes
Willy Tarreau844e3c52008-01-11 16:28:18 +010010739 Arguments : none
10740
10741 This keyword was introduced in order to provide layer 7 persistence to layer
10742 3 load balancers. The idea is to use the OS's ability to redirect an incoming
10743 connection for a remote address to a local process (here HAProxy), and let
10744 this process know what address was initially requested. When this option is
10745 used, sessions without cookies will be forwarded to the original destination
10746 IP address of the incoming request (which should match that of another
10747 equipment), while requests with cookies will still be forwarded to the
10748 appropriate server.
10749
10750 The "transparent" keyword is deprecated, use "option transparent" instead.
10751
10752 Note that contrary to a common belief, this option does NOT make HAProxy
10753 present the client's IP to the server when establishing the connection.
10754
Willy Tarreau844e3c52008-01-11 16:28:18 +010010755 See also: "option transparent"
10756
William Lallemanda73203e2012-03-12 12:48:57 +010010757unique-id-format <string>
10758 Generate a unique ID for each request.
10759 May be used in sections : defaults | frontend | listen | backend
10760 yes | yes | yes | no
10761 Arguments :
10762 <string> is a log-format string.
10763
Cyril Bonté108cf6e2012-04-21 23:30:29 +020010764 This keyword creates a ID for each request using the custom log format. A
10765 unique ID is useful to trace a request passing through many components of
10766 a complex infrastructure. The newly created ID may also be logged using the
10767 %ID tag the log-format string.
William Lallemanda73203e2012-03-12 12:48:57 +010010768
Cyril Bonté108cf6e2012-04-21 23:30:29 +020010769 The format should be composed from elements that are guaranteed to be
10770 unique when combined together. For instance, if multiple haproxy instances
10771 are involved, it might be important to include the node name. It is often
10772 needed to log the incoming connection's source and destination addresses
10773 and ports. Note that since multiple requests may be performed over the same
10774 connection, including a request counter may help differentiate them.
10775 Similarly, a timestamp may protect against a rollover of the counter.
10776 Logging the process ID will avoid collisions after a service restart.
William Lallemanda73203e2012-03-12 12:48:57 +010010777
Cyril Bonté108cf6e2012-04-21 23:30:29 +020010778 It is recommended to use hexadecimal notation for many fields since it
10779 makes them more compact and saves space in logs.
William Lallemanda73203e2012-03-12 12:48:57 +010010780
Cyril Bonté108cf6e2012-04-21 23:30:29 +020010781 Example:
William Lallemanda73203e2012-03-12 12:48:57 +010010782
Julien Vehentf21be322014-03-07 08:27:34 -050010783 unique-id-format %{+X}o\ %ci:%cp_%fi:%fp_%Ts_%rt:%pid
William Lallemanda73203e2012-03-12 12:48:57 +010010784
10785 will generate:
10786
10787 7F000001:8296_7F00001E:1F90_4F7B0A69_0003:790A
10788
10789 See also: "unique-id-header"
10790
10791unique-id-header <name>
10792 Add a unique ID header in the HTTP request.
10793 May be used in sections : defaults | frontend | listen | backend
10794 yes | yes | yes | no
10795 Arguments :
10796 <name> is the name of the header.
10797
Cyril Bonté108cf6e2012-04-21 23:30:29 +020010798 Add a unique-id header in the HTTP request sent to the server, using the
10799 unique-id-format. It can't work if the unique-id-format doesn't exist.
William Lallemanda73203e2012-03-12 12:48:57 +010010800
Cyril Bonté108cf6e2012-04-21 23:30:29 +020010801 Example:
William Lallemanda73203e2012-03-12 12:48:57 +010010802
Julien Vehentf21be322014-03-07 08:27:34 -050010803 unique-id-format %{+X}o\ %ci:%cp_%fi:%fp_%Ts_%rt:%pid
William Lallemanda73203e2012-03-12 12:48:57 +010010804 unique-id-header X-Unique-ID
10805
10806 will generate:
10807
10808 X-Unique-ID: 7F000001:8296_7F00001E:1F90_4F7B0A69_0003:790A
10809
10810 See also: "unique-id-format"
Willy Tarreau844e3c52008-01-11 16:28:18 +010010811
Willy Tarreauf51658d2014-04-23 01:21:56 +020010812use_backend <backend> [{if | unless} <condition>]
Willy Tarreau1d0dfb12009-07-07 15:10:31 +020010813 Switch to a specific backend if/unless an ACL-based condition is matched.
Willy Tarreau844e3c52008-01-11 16:28:18 +010010814 May be used in sections : defaults | frontend | listen | backend
10815 no | yes | yes | no
10816 Arguments :
Bertrand Jacquin702d44f2013-11-19 11:43:06 +010010817 <backend> is the name of a valid backend or "listen" section, or a
10818 "log-format" string resolving to a backend name.
Willy Tarreau844e3c52008-01-11 16:28:18 +010010819
Willy Tarreauf51658d2014-04-23 01:21:56 +020010820 <condition> is a condition composed of ACLs, as described in section 7. If
10821 it is omitted, the rule is unconditionally applied.
Willy Tarreau844e3c52008-01-11 16:28:18 +010010822
10823 When doing content-switching, connections arrive on a frontend and are then
10824 dispatched to various backends depending on a number of conditions. The
10825 relation between the conditions and the backends is described with the
Willy Tarreau1d0dfb12009-07-07 15:10:31 +020010826 "use_backend" keyword. While it is normally used with HTTP processing, it can
Davor Ocelice9ed2812017-12-25 17:49:28 +010010827 also be used in pure TCP, either without content using stateless ACLs (e.g.
Willy Tarreau1d0dfb12009-07-07 15:10:31 +020010828 source address validation) or combined with a "tcp-request" rule to wait for
10829 some payload.
Willy Tarreau844e3c52008-01-11 16:28:18 +010010830
10831 There may be as many "use_backend" rules as desired. All of these rules are
10832 evaluated in their declaration order, and the first one which matches will
10833 assign the backend.
10834
10835 In the first form, the backend will be used if the condition is met. In the
10836 second form, the backend will be used if the condition is not met. If no
10837 condition is valid, the backend defined with "default_backend" will be used.
10838 If no default backend is defined, either the servers in the same section are
10839 used (in case of a "listen" section) or, in case of a frontend, no server is
10840 used and a 503 service unavailable response is returned.
10841
Willy Tarreau51aecc72009-07-12 09:47:04 +020010842 Note that it is possible to switch from a TCP frontend to an HTTP backend. In
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +010010843 this case, either the frontend has already checked that the protocol is HTTP,
Willy Tarreau51aecc72009-07-12 09:47:04 +020010844 and backend processing will immediately follow, or the backend will wait for
10845 a complete HTTP request to get in. This feature is useful when a frontend
10846 must decode several protocols on a unique port, one of them being HTTP.
10847
Bertrand Jacquin702d44f2013-11-19 11:43:06 +010010848 When <backend> is a simple name, it is resolved at configuration time, and an
10849 error is reported if the specified backend does not exist. If <backend> is
10850 a log-format string instead, no check may be done at configuration time, so
10851 the backend name is resolved dynamically at run time. If the resulting
10852 backend name does not correspond to any valid backend, no other rule is
10853 evaluated, and the default_backend directive is applied instead. Note that
10854 when using dynamic backend names, it is highly recommended to use a prefix
10855 that no other backend uses in order to ensure that an unauthorized backend
10856 cannot be forced from the request.
10857
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030010858 It is worth mentioning that "use_backend" rules with an explicit name are
Bertrand Jacquin702d44f2013-11-19 11:43:06 +010010859 used to detect the association between frontends and backends to compute the
10860 backend's "fullconn" setting. This cannot be done for dynamic names.
10861
10862 See also: "default_backend", "tcp-request", "fullconn", "log-format", and
10863 section 7 about ACLs.
Willy Tarreaud72758d2010-01-12 10:42:19 +010010864
Willy Tarreau036fae02008-01-06 13:24:40 +010010865
Willy Tarreau4a5cade2012-04-05 21:09:48 +020010866use-server <server> if <condition>
10867use-server <server> unless <condition>
10868 Only use a specific server if/unless an ACL-based condition is matched.
10869 May be used in sections : defaults | frontend | listen | backend
10870 no | no | yes | yes
10871 Arguments :
Cyril Bonté108cf6e2012-04-21 23:30:29 +020010872 <server> is the name of a valid server in the same backend section.
Willy Tarreau4a5cade2012-04-05 21:09:48 +020010873
10874 <condition> is a condition composed of ACLs, as described in section 7.
10875
10876 By default, connections which arrive to a backend are load-balanced across
10877 the available servers according to the configured algorithm, unless a
10878 persistence mechanism such as a cookie is used and found in the request.
10879
10880 Sometimes it is desirable to forward a particular request to a specific
10881 server without having to declare a dedicated backend for this server. This
10882 can be achieved using the "use-server" rules. These rules are evaluated after
10883 the "redirect" rules and before evaluating cookies, and they have precedence
10884 on them. There may be as many "use-server" rules as desired. All of these
10885 rules are evaluated in their declaration order, and the first one which
10886 matches will assign the server.
10887
10888 If a rule designates a server which is down, and "option persist" is not used
10889 and no force-persist rule was validated, it is ignored and evaluation goes on
10890 with the next rules until one matches.
10891
10892 In the first form, the server will be used if the condition is met. In the
10893 second form, the server will be used if the condition is not met. If no
10894 condition is valid, the processing continues and the server will be assigned
10895 according to other persistence mechanisms.
10896
10897 Note that even if a rule is matched, cookie processing is still performed but
10898 does not assign the server. This allows prefixed cookies to have their prefix
10899 stripped.
10900
10901 The "use-server" statement works both in HTTP and TCP mode. This makes it
10902 suitable for use with content-based inspection. For instance, a server could
10903 be selected in a farm according to the TLS SNI field. And if these servers
10904 have their weight set to zero, they will not be used for other traffic.
10905
10906 Example :
10907 # intercept incoming TLS requests based on the SNI field
10908 use-server www if { req_ssl_sni -i www.example.com }
10909 server www 192.168.0.1:443 weight 0
10910 use-server mail if { req_ssl_sni -i mail.example.com }
10911 server mail 192.168.0.1:587 weight 0
10912 use-server imap if { req_ssl_sni -i imap.example.com }
Lukas Tribus98a3e3f2017-03-26 12:55:35 +000010913 server imap 192.168.0.1:993 weight 0
Willy Tarreau4a5cade2012-04-05 21:09:48 +020010914 # all the rest is forwarded to this server
10915 server default 192.168.0.2:443 check
10916
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030010917 See also: "use_backend", section 5 about server and section 7 about ACLs.
Willy Tarreau4a5cade2012-04-05 21:09:48 +020010918
Willy Tarreaub6205fd2012-09-24 12:27:33 +020010919
Davor Ocelice9ed2812017-12-25 17:49:28 +0100109205. Bind and server options
Willy Tarreaub6205fd2012-09-24 12:27:33 +020010921--------------------------
10922
10923The "bind", "server" and "default-server" keywords support a number of settings
10924depending on some build options and on the system HAProxy was built on. These
10925settings generally each consist in one word sometimes followed by a value,
10926written on the same line as the "bind" or "server" line. All these options are
10927described in this section.
10928
10929
109305.1. Bind options
10931-----------------
10932
10933The "bind" keyword supports a certain number of settings which are all passed
10934as arguments on the same line. The order in which those arguments appear makes
10935no importance, provided that they appear after the bind address. All of these
10936parameters are optional. Some of them consist in a single words (booleans),
10937while other ones expect a value after them. In this case, the value must be
10938provided immediately after the setting name.
10939
10940The currently supported settings are the following ones.
10941
Bertrand Jacquin93b227d2016-06-04 15:11:10 +010010942accept-netscaler-cip <magic number>
10943 Enforces the use of the NetScaler Client IP insertion protocol over any
10944 connection accepted by any of the TCP sockets declared on the same line. The
10945 NetScaler Client IP insertion protocol dictates the layer 3/4 addresses of
10946 the incoming connection to be used everywhere an address is used, with the
10947 only exception of "tcp-request connection" rules which will only see the
10948 real connection address. Logs will reflect the addresses indicated in the
10949 protocol, unless it is violated, in which case the real address will still
10950 be used. This keyword combined with support from external components can be
10951 used as an efficient and reliable alternative to the X-Forwarded-For
Bertrand Jacquin90759682016-06-06 15:35:39 +010010952 mechanism which is not always reliable and not even always usable. See also
10953 "tcp-request connection expect-netscaler-cip" for a finer-grained setting of
10954 which client is allowed to use the protocol.
Bertrand Jacquin93b227d2016-06-04 15:11:10 +010010955
Willy Tarreaub6205fd2012-09-24 12:27:33 +020010956accept-proxy
10957 Enforces the use of the PROXY protocol over any connection accepted by any of
Willy Tarreau77992672014-06-14 11:06:17 +020010958 the sockets declared on the same line. Versions 1 and 2 of the PROXY protocol
10959 are supported and correctly detected. The PROXY protocol dictates the layer
Willy Tarreaub6205fd2012-09-24 12:27:33 +020010960 3/4 addresses of the incoming connection to be used everywhere an address is
10961 used, with the only exception of "tcp-request connection" rules which will
10962 only see the real connection address. Logs will reflect the addresses
10963 indicated in the protocol, unless it is violated, in which case the real
Davor Ocelice9ed2812017-12-25 17:49:28 +010010964 address will still be used. This keyword combined with support from external
Willy Tarreaub6205fd2012-09-24 12:27:33 +020010965 components can be used as an efficient and reliable alternative to the
10966 X-Forwarded-For mechanism which is not always reliable and not even always
Willy Tarreau4f0d9192013-06-11 20:40:55 +020010967 usable. See also "tcp-request connection expect-proxy" for a finer-grained
10968 setting of which client is allowed to use the protocol.
Willy Tarreaub6205fd2012-09-24 12:27:33 +020010969
Olivier Houchardc2aae742017-09-22 18:26:28 +020010970allow-0rtt
Bertrand Jacquina25282b2018-08-14 00:56:13 +010010971 Allow receiving early data when using TLSv1.3. This is disabled by default,
Olivier Houchard69752962019-01-08 15:35:32 +010010972 due to security considerations. Because it is vulnerable to replay attacks,
John Roeslerfb2fce12019-07-10 15:45:51 -050010973 you should only allow if for requests that are safe to replay, i.e. requests
Olivier Houchard69752962019-01-08 15:35:32 +010010974 that are idempotent. You can use the "wait-for-handshake" action for any
10975 request that wouldn't be safe with early data.
Olivier Houchardc2aae742017-09-22 18:26:28 +020010976
Willy Tarreauab861d32013-04-02 02:30:41 +020010977alpn <protocols>
10978 This enables the TLS ALPN extension and advertises the specified protocol
10979 list as supported on top of ALPN. The protocol list consists in a comma-
10980 delimited list of protocol names, for instance: "http/1.1,http/1.0" (without
John Roeslerfb2fce12019-07-10 15:45:51 -050010981 quotes). This requires that the SSL library is built with support for TLS
Willy Tarreauab861d32013-04-02 02:30:41 +020010982 extensions enabled (check with haproxy -vv). The ALPN extension replaces the
Willy Tarreau95c4e142017-11-26 12:18:55 +010010983 initial NPN extension. ALPN is required to enable HTTP/2 on an HTTP frontend.
10984 Versions of OpenSSL prior to 1.0.2 didn't support ALPN and only supposed the
10985 now obsolete NPN extension. At the time of writing this, most browsers still
10986 support both ALPN and NPN for HTTP/2 so a fallback to NPN may still work for
10987 a while. But ALPN must be used whenever possible. If both HTTP/2 and HTTP/1.1
10988 are expected to be supported, both versions can be advertised, in order of
10989 preference, like below :
10990
10991 bind :443 ssl crt pub.pem alpn h2,http/1.1
Willy Tarreauab861d32013-04-02 02:30:41 +020010992
Willy Tarreaub6205fd2012-09-24 12:27:33 +020010993backlog <backlog>
Willy Tarreaue2711c72019-02-27 15:39:41 +010010994 Sets the socket's backlog to this value. If unspecified or 0, the frontend's
Willy Tarreaub6205fd2012-09-24 12:27:33 +020010995 backlog is used instead, which generally defaults to the maxconn value.
10996
Emmanuel Hocdete7f2b732017-01-09 16:15:54 +010010997curves <curves>
10998 This setting is only available when support for OpenSSL was built in. It sets
10999 the string describing the list of elliptic curves algorithms ("curve suite")
11000 that are negotiated during the SSL/TLS handshake with ECDHE. The format of the
11001 string is a colon-delimited list of curve name.
11002 Example: "X25519:P-256" (without quote)
11003 When "curves" is set, "ecdhe" parameter is ignored.
11004
Emeric Brun7fb34422012-09-28 15:26:15 +020011005ecdhe <named curve>
11006 This setting is only available when support for OpenSSL was built in. It sets
Emeric Brun6924ef82013-03-06 14:08:53 +010011007 the named curve (RFC 4492) used to generate ECDH ephemeral keys. By default,
11008 used named curve is prime256v1.
Emeric Brun7fb34422012-09-28 15:26:15 +020011009
Emeric Brunfd33a262012-10-11 16:28:27 +020011010ca-file <cafile>
Emeric Brun1a073b42012-09-28 17:07:34 +020011011 This setting is only available when support for OpenSSL was built in. It
11012 designates a PEM file from which to load CA certificates used to verify
11013 client's certificate.
11014
Emeric Brunb6dc9342012-09-28 17:55:37 +020011015ca-ignore-err [all|<errorID>,...]
11016 This setting is only available when support for OpenSSL was built in.
11017 Sets a comma separated list of errorIDs to ignore during verify at depth > 0.
11018 If set to 'all', all errors are ignored. SSL handshake is not aborted if an
11019 error is ignored.
11020
Christopher Faulet31af49d2015-06-09 17:29:50 +020011021ca-sign-file <cafile>
11022 This setting is only available when support for OpenSSL was built in. It
11023 designates a PEM file containing both the CA certificate and the CA private
11024 key used to create and sign server's certificates. This is a mandatory
11025 setting when the dynamic generation of certificates is enabled. See
11026 'generate-certificates' for details.
11027
Bertrand Jacquind4d0a232016-11-13 16:37:12 +000011028ca-sign-pass <passphrase>
Christopher Faulet31af49d2015-06-09 17:29:50 +020011029 This setting is only available when support for OpenSSL was built in. It is
11030 the CA private key passphrase. This setting is optional and used only when
11031 the dynamic generation of certificates is enabled. See
11032 'generate-certificates' for details.
11033
Willy Tarreaub6205fd2012-09-24 12:27:33 +020011034ciphers <ciphers>
11035 This setting is only available when support for OpenSSL was built in. It sets
11036 the string describing the list of cipher algorithms ("cipher suite") that are
Bertrand Jacquin8cf7c1e2019-02-03 18:35:25 +000011037 negotiated during the SSL/TLS handshake up to TLSv1.2. The format of the
Bertrand Jacquin4f03ab02019-02-03 18:48:49 +000011038 string is defined in "man 1 ciphers" from OpenSSL man pages. For background
Dirkjan Bussink415150f2018-09-14 11:14:21 +020011039 information and recommendations see e.g.
11040 (https://wiki.mozilla.org/Security/Server_Side_TLS) and
11041 (https://mozilla.github.io/server-side-tls/ssl-config-generator/). For TLSv1.3
11042 cipher configuration, please check the "ciphersuites" keyword.
11043
11044ciphersuites <ciphersuites>
11045 This setting is only available when support for OpenSSL was built in and
11046 OpenSSL 1.1.1 or later was used to build HAProxy. It sets the string describing
11047 the list of cipher algorithms ("cipher suite") that are negotiated during the
11048 TLSv1.3 handshake. The format of the string is defined in "man 1 ciphers" from
Bertrand Jacquin4f03ab02019-02-03 18:48:49 +000011049 OpenSSL man pages under the "ciphersuites" section. For cipher configuration
11050 for TLSv1.2 and earlier, please check the "ciphers" keyword.
Willy Tarreaub6205fd2012-09-24 12:27:33 +020011051
Emeric Brunfd33a262012-10-11 16:28:27 +020011052crl-file <crlfile>
Emeric Brun1a073b42012-09-28 17:07:34 +020011053 This setting is only available when support for OpenSSL was built in. It
11054 designates a PEM file from which to load certificate revocation list used
11055 to verify client's certificate.
11056
Willy Tarreaub6205fd2012-09-24 12:27:33 +020011057crt <cert>
Alex Davies0fbf0162013-03-02 16:04:50 +000011058 This setting is only available when support for OpenSSL was built in. It
11059 designates a PEM file containing both the required certificates and any
11060 associated private keys. This file can be built by concatenating multiple
11061 PEM files into one (e.g. cat cert.pem key.pem > combined.pem). If your CA
11062 requires an intermediate certificate, this can also be concatenated into this
11063 file.
11064
11065 If the OpenSSL used supports Diffie-Hellman, parameters present in this file
11066 are loaded.
11067
11068 If a directory name is used instead of a PEM file, then all files found in
Cyril Bonté3180f7b2015-01-25 00:16:08 +010011069 that directory will be loaded in alphabetic order unless their name ends with
Janusz Dziemidowicz2c701b52015-03-07 23:03:59 +010011070 '.issuer', '.ocsp' or '.sctl' (reserved extensions). This directive may be
11071 specified multiple times in order to load certificates from multiple files or
11072 directories. The certificates will be presented to clients who provide a
11073 valid TLS Server Name Indication field matching one of their CN or alt
Davor Ocelice9ed2812017-12-25 17:49:28 +010011074 subjects. Wildcards are supported, where a wildcard character '*' is used
11075 instead of the first hostname component (e.g. *.example.org matches
Janusz Dziemidowicz2c701b52015-03-07 23:03:59 +010011076 www.example.org but not www.sub.example.org).
Alex Davies0fbf0162013-03-02 16:04:50 +000011077
11078 If no SNI is provided by the client or if the SSL library does not support
11079 TLS extensions, or if the client provides an SNI hostname which does not
11080 match any certificate, then the first loaded certificate will be presented.
11081 This means that when loading certificates from a directory, it is highly
Cyril Bonté3180f7b2015-01-25 00:16:08 +010011082 recommended to load the default one first as a file or to ensure that it will
11083 always be the first one in the directory.
Alex Davies0fbf0162013-03-02 16:04:50 +000011084
Emeric Brune032bfa2012-09-28 13:01:45 +020011085 Note that the same cert may be loaded multiple times without side effects.
Willy Tarreaub6205fd2012-09-24 12:27:33 +020011086
Davor Ocelice9ed2812017-12-25 17:49:28 +010011087 Some CAs (such as GoDaddy) offer a drop down list of server types that do not
Alex Davies0fbf0162013-03-02 16:04:50 +000011088 include HAProxy when obtaining a certificate. If this happens be sure to
Davor Ocelice9ed2812017-12-25 17:49:28 +010011089 choose a web server that the CA believes requires an intermediate CA (for
11090 GoDaddy, selection Apache Tomcat will get the correct bundle, but many
Alex Davies0fbf0162013-03-02 16:04:50 +000011091 others, e.g. nginx, result in a wrong bundle that will not work for some
11092 clients).
11093
Emeric Brun4147b2e2014-06-16 18:36:30 +020011094 For each PEM file, haproxy checks for the presence of file at the same path
11095 suffixed by ".ocsp". If such file is found, support for the TLS Certificate
11096 Status Request extension (also known as "OCSP stapling") is automatically
11097 enabled. The content of this file is optional. If not empty, it must contain
11098 a valid OCSP Response in DER format. In order to be valid an OCSP Response
11099 must comply with the following rules: it has to indicate a good status,
11100 it has to be a single response for the certificate of the PEM file, and it
11101 has to be valid at the moment of addition. If these rules are not respected
11102 the OCSP Response is ignored and a warning is emitted. In order to identify
11103 which certificate an OCSP Response applies to, the issuer's certificate is
11104 necessary. If the issuer's certificate is not found in the PEM file, it will
11105 be loaded from a file at the same path as the PEM file suffixed by ".issuer"
11106 if it exists otherwise it will fail with an error.
11107
Janusz Dziemidowicz2c701b52015-03-07 23:03:59 +010011108 For each PEM file, haproxy also checks for the presence of file at the same
11109 path suffixed by ".sctl". If such file is found, support for Certificate
11110 Transparency (RFC6962) TLS extension is enabled. The file must contain a
11111 valid Signed Certificate Timestamp List, as described in RFC. File is parsed
11112 to check basic syntax, but no signatures are verified.
11113
yanbzhu6c25e9e2016-01-05 12:52:02 -050011114 There are cases where it is desirable to support multiple key types, e.g. RSA
11115 and ECDSA in the cipher suites offered to the clients. This allows clients
11116 that support EC certificates to be able to use EC ciphers, while
11117 simultaneously supporting older, RSA only clients.
yanbzhud19630c2015-12-14 15:10:25 -050011118
11119 In order to provide this functionality, multiple PEM files, each with a
11120 different key type, are required. To associate these PEM files into a
11121 "cert bundle" that is recognized by haproxy, they must be named in the
11122 following way: All PEM files that are to be bundled must have the same base
11123 name, with a suffix indicating the key type. Currently, three suffixes are
11124 supported: rsa, dsa and ecdsa. For example, if www.example.com has two PEM
11125 files, an RSA file and an ECDSA file, they must be named: "example.pem.rsa"
11126 and "example.pem.ecdsa". The first part of the filename is arbitrary; only the
11127 suffix matters. To load this bundle into haproxy, specify the base name only:
11128
11129 Example : bind :8443 ssl crt example.pem
11130
yanbzhu6c25e9e2016-01-05 12:52:02 -050011131 Note that the suffix is not given to haproxy; this tells haproxy to look for
yanbzhud19630c2015-12-14 15:10:25 -050011132 a cert bundle.
11133
Davor Ocelice9ed2812017-12-25 17:49:28 +010011134 HAProxy will load all PEM files in the bundle at the same time to try to
yanbzhud19630c2015-12-14 15:10:25 -050011135 support multiple key types. PEM files are combined based on Common Name
11136 (CN) and Subject Alternative Name (SAN) to support SNI lookups. This means
11137 that even if you give haproxy a cert bundle, if there are no shared CN/SAN
11138 entries in the certificates in that bundle, haproxy will not be able to
11139 provide multi-cert support.
11140
11141 Assuming bundle in the example above contained the following:
11142
11143 Filename | CN | SAN
11144 -------------------+-----------------+-------------------
11145 example.pem.rsa | www.example.com | rsa.example.com
yanbzhu6c25e9e2016-01-05 12:52:02 -050011146 -------------------+-----------------+-------------------
yanbzhud19630c2015-12-14 15:10:25 -050011147 example.pem.ecdsa | www.example.com | ecdsa.example.com
11148 -------------------+-----------------+-------------------
11149
11150 Users connecting with an SNI of "www.example.com" will be able
11151 to use both RSA and ECDSA cipher suites. Users connecting with an SNI of
11152 "rsa.example.com" will only be able to use RSA cipher suites, and users
11153 connecting with "ecdsa.example.com" will only be able to use ECDSA cipher
Emmanuel Hocdet84e417d2017-08-16 11:33:17 +020011154 suites. With BoringSSL and Openssl >= 1.1.1 multi-cert is natively supported,
11155 no need to bundle certificates. ECDSA certificate will be preferred if client
11156 support it.
yanbzhud19630c2015-12-14 15:10:25 -050011157
11158 If a directory name is given as the <cert> argument, haproxy will
11159 automatically search and load bundled files in that directory.
11160
11161 OSCP files (.ocsp) and issuer files (.issuer) are supported with multi-cert
11162 bundling. Each certificate can have its own .ocsp and .issuer file. At this
11163 time, sctl is not supported in multi-certificate bundling.
11164
Emeric Brunb6dc9342012-09-28 17:55:37 +020011165crt-ignore-err <errors>
Alex Davies0fbf0162013-03-02 16:04:50 +000011166 This setting is only available when support for OpenSSL was built in. Sets a
Davor Ocelice9ed2812017-12-25 17:49:28 +010011167 comma separated list of errorIDs to ignore during verify at depth == 0. If
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030011168 set to 'all', all errors are ignored. SSL handshake is not aborted if an error
Alex Davies0fbf0162013-03-02 16:04:50 +000011169 is ignored.
Emeric Brunb6dc9342012-09-28 17:55:37 +020011170
Emmanuel Hocdetfe616562013-01-22 15:31:15 +010011171crt-list <file>
11172 This setting is only available when support for OpenSSL was built in. It
Emmanuel Hocdet98263292016-12-29 18:26:15 +010011173 designates a list of PEM file with an optional ssl configuration and a SNI
11174 filter per certificate, with the following format for each line :
Emmanuel Hocdetfe616562013-01-22 15:31:15 +010011175
Emmanuel Hocdet98263292016-12-29 18:26:15 +010011176 <crtfile> [\[<sslbindconf> ...\]] [[!]<snifilter> ...]
11177
Emmanuel Hocdet174dfe52017-07-28 15:01:05 +020011178 sslbindconf support "npn", "alpn", "verify", "ca-file", "no-ca-names",
11179 crl-file", "ecdhe", "curves", "ciphers" configuration. With BoringSSL
Emmanuel Hocdet84e417d2017-08-16 11:33:17 +020011180 and Openssl >= 1.1.1 "ssl-min-ver" and "ssl-max-ver" are also supported.
Emmanuel Hocdet98263292016-12-29 18:26:15 +010011181 It override the configuration set in bind line for the certificate.
Emmanuel Hocdetfe616562013-01-22 15:31:15 +010011182
Emmanuel Hocdet7c41a1b2013-05-07 20:20:06 +020011183 Wildcards are supported in the SNI filter. Negative filter are also supported,
11184 only useful in combination with a wildcard filter to exclude a particular SNI.
11185 The certificates will be presented to clients who provide a valid TLS Server
11186 Name Indication field matching one of the SNI filters. If no SNI filter is
11187 specified, the CN and alt subjects are used. This directive may be specified
11188 multiple times. See the "crt" option for more information. The default
11189 certificate is still needed to meet OpenSSL expectations. If it is not used,
11190 the 'strict-sni' option may be used.
Emmanuel Hocdetfe616562013-01-22 15:31:15 +010011191
yanbzhu6c25e9e2016-01-05 12:52:02 -050011192 Multi-cert bundling (see "crt") is supported with crt-list, as long as only
Emmanuel Hocdetd294aea2016-05-13 11:14:06 +020011193 the base name is given in the crt-list. SNI filter will do the same work on
Emmanuel Hocdet84e417d2017-08-16 11:33:17 +020011194 all bundled certificates. With BoringSSL and Openssl >= 1.1.1 multi-cert is
11195 natively supported, avoid multi-cert bundling. RSA and ECDSA certificates can
11196 be declared in a row, and set different ssl and filter parameter.
yanbzhud19630c2015-12-14 15:10:25 -050011197
Emmanuel Hocdet98263292016-12-29 18:26:15 +010011198 crt-list file example:
11199 cert1.pem
Emmanuel Hocdet05942112017-02-20 16:11:50 +010011200 cert2.pem [alpn h2,http/1.1]
Emmanuel Hocdet98263292016-12-29 18:26:15 +010011201 certW.pem *.domain.tld !secure.domain.tld
Emmanuel Hocdet05942112017-02-20 16:11:50 +010011202 certS.pem [curves X25519:P-256 ciphers ECDHE-ECDSA-AES256-GCM-SHA384] secure.domain.tld
Emmanuel Hocdet98263292016-12-29 18:26:15 +010011203
Willy Tarreaub6205fd2012-09-24 12:27:33 +020011204defer-accept
11205 Is an optional keyword which is supported only on certain Linux kernels. It
11206 states that a connection will only be accepted once some data arrive on it,
11207 or at worst after the first retransmit. This should be used only on protocols
Davor Ocelice9ed2812017-12-25 17:49:28 +010011208 for which the client talks first (e.g. HTTP). It can slightly improve
Willy Tarreaub6205fd2012-09-24 12:27:33 +020011209 performance by ensuring that most of the request is already available when
11210 the connection is accepted. On the other hand, it will not be able to detect
11211 connections which don't talk. It is important to note that this option is
11212 broken in all kernels up to 2.6.31, as the connection is never accepted until
11213 the client talks. This can cause issues with front firewalls which would see
11214 an established connection while the proxy will only see it in SYN_RECV. This
11215 option is only supported on TCPv4/TCPv6 sockets and ignored by other ones.
11216
William Lallemandf6975e92017-05-26 17:42:10 +020011217expose-fd listeners
11218 This option is only usable with the stats socket. It gives your stats socket
11219 the capability to pass listeners FD to another HAProxy process.
William Lallemande202b1e2017-06-01 17:38:56 +020011220 During a reload with the master-worker mode, the process is automatically
11221 reexecuted adding -x and one of the stats socket with this option.
Davor Ocelice9ed2812017-12-25 17:49:28 +010011222 See also "-x" in the management guide.
William Lallemandf6975e92017-05-26 17:42:10 +020011223
Emeric Brun2cb7ae52012-10-05 14:14:21 +020011224force-sslv3
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030011225 This option enforces use of SSLv3 only on SSL connections instantiated from
Emeric Brun2cb7ae52012-10-05 14:14:21 +020011226 this listener. SSLv3 is generally less expensive than the TLS counterparts
Emeric Brun2c86cbf2014-10-30 15:56:50 +010011227 for high connection rates. This option is also available on global statement
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020011228 "ssl-default-bind-options". See also "ssl-min-ver" and "ssl-max-ver".
Emeric Brun2cb7ae52012-10-05 14:14:21 +020011229
11230force-tlsv10
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030011231 This option enforces use of TLSv1.0 only on SSL connections instantiated from
Emeric Brun2c86cbf2014-10-30 15:56:50 +010011232 this listener. This option is also available on global statement
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020011233 "ssl-default-bind-options". See also "ssl-min-ver" and "ssl-max-ver".
Emeric Brun2cb7ae52012-10-05 14:14:21 +020011234
11235force-tlsv11
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030011236 This option enforces use of TLSv1.1 only on SSL connections instantiated from
Emeric Brun2c86cbf2014-10-30 15:56:50 +010011237 this listener. This option is also available on global statement
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020011238 "ssl-default-bind-options". See also "ssl-min-ver" and "ssl-max-ver".
Emeric Brun2cb7ae52012-10-05 14:14:21 +020011239
11240force-tlsv12
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030011241 This option enforces use of TLSv1.2 only on SSL connections instantiated from
Emeric Brun2c86cbf2014-10-30 15:56:50 +010011242 this listener. This option is also available on global statement
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020011243 "ssl-default-bind-options". See also "ssl-min-ver" and "ssl-max-ver".
Emeric Brun2cb7ae52012-10-05 14:14:21 +020011244
Emmanuel Hocdet42fb9802017-03-30 19:29:39 +020011245force-tlsv13
11246 This option enforces use of TLSv1.3 only on SSL connections instantiated from
11247 this listener. This option is also available on global statement
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020011248 "ssl-default-bind-options". See also "ssl-min-ver" and "ssl-max-ver".
Emmanuel Hocdet42fb9802017-03-30 19:29:39 +020011249
Christopher Faulet31af49d2015-06-09 17:29:50 +020011250generate-certificates
11251 This setting is only available when support for OpenSSL was built in. It
11252 enables the dynamic SSL certificates generation. A CA certificate and its
11253 private key are necessary (see 'ca-sign-file'). When HAProxy is configured as
11254 a transparent forward proxy, SSL requests generate errors because of a common
11255 name mismatch on the certificate presented to the client. With this option
11256 enabled, HAProxy will try to forge a certificate using the SNI hostname
11257 indicated by the client. This is done only if no certificate matches the SNI
11258 hostname (see 'crt-list'). If an error occurs, the default certificate is
11259 used, else the 'strict-sni' option is set.
11260 It can also be used when HAProxy is configured as a reverse proxy to ease the
11261 deployment of an architecture with many backends.
11262
11263 Creating a SSL certificate is an expensive operation, so a LRU cache is used
11264 to store forged certificates (see 'tune.ssl.ssl-ctx-cache-size'). It
Davor Ocelice9ed2812017-12-25 17:49:28 +010011265 increases the HAProxy's memory footprint to reduce latency when the same
Christopher Faulet31af49d2015-06-09 17:29:50 +020011266 certificate is used many times.
11267
Willy Tarreaub6205fd2012-09-24 12:27:33 +020011268gid <gid>
11269 Sets the group of the UNIX sockets to the designated system gid. It can also
11270 be set by default in the global section's "unix-bind" statement. Note that
11271 some platforms simply ignore this. This setting is equivalent to the "group"
11272 setting except that the group ID is used instead of its name. This setting is
11273 ignored by non UNIX sockets.
11274
11275group <group>
11276 Sets the group of the UNIX sockets to the designated system group. It can
11277 also be set by default in the global section's "unix-bind" statement. Note
11278 that some platforms simply ignore this. This setting is equivalent to the
11279 "gid" setting except that the group name is used instead of its gid. This
11280 setting is ignored by non UNIX sockets.
11281
11282id <id>
11283 Fixes the socket ID. By default, socket IDs are automatically assigned, but
11284 sometimes it is more convenient to fix them to ease monitoring. This value
11285 must be strictly positive and unique within the listener/frontend. This
11286 option can only be used when defining only a single socket.
11287
11288interface <interface>
Lukas Tribusfce2e962013-02-12 22:13:19 +010011289 Restricts the socket to a specific interface. When specified, only packets
11290 received from that particular interface are processed by the socket. This is
11291 currently only supported on Linux. The interface must be a primary system
11292 interface, not an aliased interface. It is also possible to bind multiple
11293 frontends to the same address if they are bound to different interfaces. Note
11294 that binding to a network interface requires root privileges. This parameter
Jérôme Magnin61275192018-02-07 11:39:58 +010011295 is only compatible with TCPv4/TCPv6 sockets. When specified, return traffic
11296 uses the same interface as inbound traffic, and its associated routing table,
11297 even if there are explicit routes through different interfaces configured.
11298 This can prove useful to address asymmetric routing issues when the same
11299 client IP addresses need to be able to reach frontends hosted on different
11300 interfaces.
Willy Tarreaub6205fd2012-09-24 12:27:33 +020011301
Willy Tarreauabb175f2012-09-24 12:43:26 +020011302level <level>
11303 This setting is used with the stats sockets only to restrict the nature of
11304 the commands that can be issued on the socket. It is ignored by other
11305 sockets. <level> can be one of :
Davor Ocelice9ed2812017-12-25 17:49:28 +010011306 - "user" is the least privileged level; only non-sensitive stats can be
Willy Tarreauabb175f2012-09-24 12:43:26 +020011307 read, and no change is allowed. It would make sense on systems where it
11308 is not easy to restrict access to the socket.
11309 - "operator" is the default level and fits most common uses. All data can
Davor Ocelice9ed2812017-12-25 17:49:28 +010011310 be read, and only non-sensitive changes are permitted (e.g. clear max
Willy Tarreauabb175f2012-09-24 12:43:26 +020011311 counters).
Davor Ocelice9ed2812017-12-25 17:49:28 +010011312 - "admin" should be used with care, as everything is permitted (e.g. clear
Willy Tarreauabb175f2012-09-24 12:43:26 +020011313 all counters).
11314
Andjelko Iharosc4df59e2017-07-20 11:59:48 +020011315severity-output <format>
11316 This setting is used with the stats sockets only to configure severity
11317 level output prepended to informational feedback messages. Severity
11318 level of messages can range between 0 and 7, conforming to syslog
11319 rfc5424. Valid and successful socket commands requesting data
11320 (i.e. "show map", "get acl foo" etc.) will never have a severity level
11321 prepended. It is ignored by other sockets. <format> can be one of :
11322 - "none" (default) no severity level is prepended to feedback messages.
11323 - "number" severity level is prepended as a number.
11324 - "string" severity level is prepended as a string following the
11325 rfc5424 convention.
11326
Willy Tarreaub6205fd2012-09-24 12:27:33 +020011327maxconn <maxconn>
11328 Limits the sockets to this number of concurrent connections. Extraneous
11329 connections will remain in the system's backlog until a connection is
11330 released. If unspecified, the limit will be the same as the frontend's
11331 maxconn. Note that in case of port ranges or multiple addresses, the same
11332 value will be applied to each socket. This setting enables different
11333 limitations on expensive sockets, for instance SSL entries which may easily
11334 eat all memory.
11335
11336mode <mode>
11337 Sets the octal mode used to define access permissions on the UNIX socket. It
11338 can also be set by default in the global section's "unix-bind" statement.
11339 Note that some platforms simply ignore this. This setting is ignored by non
11340 UNIX sockets.
11341
11342mss <maxseg>
11343 Sets the TCP Maximum Segment Size (MSS) value to be advertised on incoming
11344 connections. This can be used to force a lower MSS for certain specific
11345 ports, for instance for connections passing through a VPN. Note that this
11346 relies on a kernel feature which is theoretically supported under Linux but
11347 was buggy in all versions prior to 2.6.28. It may or may not work on other
11348 operating systems. It may also not change the advertised value but change the
11349 effective size of outgoing segments. The commonly advertised value for TCPv4
11350 over Ethernet networks is 1460 = 1500(MTU) - 40(IP+TCP). If this value is
11351 positive, it will be used as the advertised MSS. If it is negative, it will
11352 indicate by how much to reduce the incoming connection's advertised MSS for
11353 outgoing segments. This parameter is only compatible with TCP v4/v6 sockets.
11354
11355name <name>
11356 Sets an optional name for these sockets, which will be reported on the stats
11357 page.
11358
Willy Tarreaud72f0f32015-10-13 14:50:22 +020011359namespace <name>
11360 On Linux, it is possible to specify which network namespace a socket will
11361 belong to. This directive makes it possible to explicitly bind a listener to
11362 a namespace different from the default one. Please refer to your operating
11363 system's documentation to find more details about network namespaces.
11364
Willy Tarreaub6205fd2012-09-24 12:27:33 +020011365nice <nice>
11366 Sets the 'niceness' of connections initiated from the socket. Value must be
11367 in the range -1024..1024 inclusive, and defaults to zero. Positive values
11368 means that such connections are more friendly to others and easily offer
11369 their place in the scheduler. On the opposite, negative values mean that
11370 connections want to run with a higher priority than others. The difference
11371 only happens under high loads when the system is close to saturation.
11372 Negative values are appropriate for low-latency or administration services,
11373 and high values are generally recommended for CPU intensive tasks such as SSL
11374 processing or bulk transfers which are less sensible to latency. For example,
11375 it may make sense to use a positive value for an SMTP socket and a negative
11376 one for an RDP socket.
11377
Emmanuel Hocdet174dfe52017-07-28 15:01:05 +020011378no-ca-names
11379 This setting is only available when support for OpenSSL was built in. It
11380 prevents from send CA names in server hello message when ca-file is used.
11381
Emeric Brun9b3009b2012-10-05 11:55:06 +020011382no-sslv3
Willy Tarreaub6205fd2012-09-24 12:27:33 +020011383 This setting is only available when support for OpenSSL was built in. It
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030011384 disables support for SSLv3 on any sockets instantiated from the listener when
Willy Tarreaub6205fd2012-09-24 12:27:33 +020011385 SSL is supported. Note that SSLv2 is forced disabled in the code and cannot
Emeric Brun2c86cbf2014-10-30 15:56:50 +010011386 be enabled using any configuration option. This option is also available on
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020011387 global statement "ssl-default-bind-options". Use "ssl-min-ver" and
11388 "ssl-max-ver" instead.
Willy Tarreaub6205fd2012-09-24 12:27:33 +020011389
Emeric Brun90ad8722012-10-02 14:00:59 +020011390no-tls-tickets
11391 This setting is only available when support for OpenSSL was built in. It
11392 disables the stateless session resumption (RFC 5077 TLS Ticket
11393 extension) and force to use stateful session resumption. Stateless
Emeric Brun2c86cbf2014-10-30 15:56:50 +010011394 session resumption is more expensive in CPU usage. This option is also
11395 available on global statement "ssl-default-bind-options".
Emeric Brun90ad8722012-10-02 14:00:59 +020011396
Emeric Brun9b3009b2012-10-05 11:55:06 +020011397no-tlsv10
Willy Tarreaub6205fd2012-09-24 12:27:33 +020011398 This setting is only available when support for OpenSSL was built in. It
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030011399 disables support for TLSv1.0 on any sockets instantiated from the listener
Emeric Brun2cb7ae52012-10-05 14:14:21 +020011400 when SSL is supported. Note that SSLv2 is forced disabled in the code and
Emeric Brun2c86cbf2014-10-30 15:56:50 +010011401 cannot be enabled using any configuration option. This option is also
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020011402 available on global statement "ssl-default-bind-options". Use "ssl-min-ver"
11403 and "ssl-max-ver" instead.
Willy Tarreaub6205fd2012-09-24 12:27:33 +020011404
Emeric Brun9b3009b2012-10-05 11:55:06 +020011405no-tlsv11
Emeric Brunf5da4932012-09-28 19:42:54 +020011406 This setting is only available when support for OpenSSL was built in. It
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030011407 disables support for TLSv1.1 on any sockets instantiated from the listener
Emeric Brun2cb7ae52012-10-05 14:14:21 +020011408 when SSL is supported. Note that SSLv2 is forced disabled in the code and
Emeric Brun2c86cbf2014-10-30 15:56:50 +010011409 cannot be enabled using any configuration option. This option is also
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020011410 available on global statement "ssl-default-bind-options". Use "ssl-min-ver"
11411 and "ssl-max-ver" instead.
Emeric Brunf5da4932012-09-28 19:42:54 +020011412
Emeric Brun9b3009b2012-10-05 11:55:06 +020011413no-tlsv12
Emeric Brunf5da4932012-09-28 19:42:54 +020011414 This setting is only available when support for OpenSSL was built in. It
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030011415 disables support for TLSv1.2 on any sockets instantiated from the listener
Emeric Brun2cb7ae52012-10-05 14:14:21 +020011416 when SSL is supported. Note that SSLv2 is forced disabled in the code and
Emeric Brun2c86cbf2014-10-30 15:56:50 +010011417 cannot be enabled using any configuration option. This option is also
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020011418 available on global statement "ssl-default-bind-options". Use "ssl-min-ver"
11419 and "ssl-max-ver" instead.
Emeric Brunf5da4932012-09-28 19:42:54 +020011420
Emmanuel Hocdet42fb9802017-03-30 19:29:39 +020011421no-tlsv13
11422 This setting is only available when support for OpenSSL was built in. It
11423 disables support for TLSv1.3 on any sockets instantiated from the listener
11424 when SSL is supported. Note that SSLv2 is forced disabled in the code and
11425 cannot be enabled using any configuration option. This option is also
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020011426 available on global statement "ssl-default-bind-options". Use "ssl-min-ver"
11427 and "ssl-max-ver" instead.
Emmanuel Hocdet42fb9802017-03-30 19:29:39 +020011428
Willy Tarreau6c9a3d52012-10-18 18:57:14 +020011429npn <protocols>
11430 This enables the NPN TLS extension and advertises the specified protocol list
11431 as supported on top of NPN. The protocol list consists in a comma-delimited
11432 list of protocol names, for instance: "http/1.1,http/1.0" (without quotes).
John Roeslerfb2fce12019-07-10 15:45:51 -050011433 This requires that the SSL library is built with support for TLS extensions
Willy Tarreauab861d32013-04-02 02:30:41 +020011434 enabled (check with haproxy -vv). Note that the NPN extension has been
Willy Tarreau95c4e142017-11-26 12:18:55 +010011435 replaced with the ALPN extension (see the "alpn" keyword), though this one is
11436 only available starting with OpenSSL 1.0.2. If HTTP/2 is desired on an older
11437 version of OpenSSL, NPN might still be used as most clients still support it
11438 at the time of writing this. It is possible to enable both NPN and ALPN
11439 though it probably doesn't make any sense out of testing.
Willy Tarreau6c9a3d52012-10-18 18:57:14 +020011440
Lukas Tribus53ae85c2017-05-04 15:45:40 +000011441prefer-client-ciphers
11442 Use the client's preference when selecting the cipher suite, by default
11443 the server's preference is enforced. This option is also available on
11444 global statement "ssl-default-bind-options".
Lukas Tribus926594f2018-05-18 17:55:57 +020011445 Note that with OpenSSL >= 1.1.1 ChaCha20-Poly1305 is reprioritized anyway
11446 (without setting this option), if a ChaCha20-Poly1305 cipher is at the top of
11447 the client cipher list.
Lukas Tribus53ae85c2017-05-04 15:45:40 +000011448
Christopher Fauletc644fa92017-11-23 22:44:11 +010011449process <process-set>[/<thread-set>]
Willy Tarreaua36b3242019-02-02 13:14:34 +010011450 This restricts the list of processes or threads on which this listener is
Christopher Fauletc644fa92017-11-23 22:44:11 +010011451 allowed to run. It does not enforce any process but eliminates those which do
Davor Ocelice9ed2812017-12-25 17:49:28 +010011452 not match. If the frontend uses a "bind-process" setting, the intersection
Christopher Fauletc644fa92017-11-23 22:44:11 +010011453 between the two is applied. If in the end the listener is not allowed to run
11454 on any remaining process, a warning is emitted, and the listener will either
11455 run on the first process of the listener if a single process was specified,
11456 or on all of its processes if multiple processes were specified. If a thread
Davor Ocelice9ed2812017-12-25 17:49:28 +010011457 set is specified, it limits the threads allowed to process incoming
Willy Tarreaua36b3242019-02-02 13:14:34 +010011458 connections for this listener, for the the process set. If multiple processes
11459 and threads are configured, a warning is emitted, as it either results from a
11460 configuration error or a misunderstanding of these models. For the unlikely
11461 case where several ranges are needed, this directive may be repeated.
11462 <process-set> and <thread-set> must use the format
Christopher Fauletc644fa92017-11-23 22:44:11 +010011463
11464 all | odd | even | number[-[number]]
11465
11466 Ranges can be partially defined. The higher bound can be omitted. In such
11467 case, it is replaced by the corresponding maximum value. The main purpose of
11468 this directive is to be used with the stats sockets and have one different
11469 socket per process. The second purpose is to have multiple bind lines sharing
11470 the same IP:port but not the same process in a listener, so that the system
11471 can distribute the incoming connections into multiple queues and allow a
11472 smoother inter-process load balancing. Currently Linux 3.9 and above is known
11473 for supporting this. See also "bind-process" and "nbproc".
Willy Tarreau6ae1ba62014-05-07 19:01:58 +020011474
Christopher Fauleta717b992018-04-10 14:43:00 +020011475proto <name>
11476 Forces the multiplexer's protocol to use for the incoming connections. It
11477 must be compatible with the mode of the frontend (TCP or HTTP). It must also
11478 be usable on the frontend side. The list of available protocols is reported
11479 in haproxy -vv.
11480 Idea behind this optipon is to bypass the selection of the best multiplexer's
11481 protocol for all connections instantiated from this listening socket. For
Joseph Herlant71b4b152018-11-13 16:55:16 -080011482 instance, it is possible to force the http/2 on clear TCP by specifying "proto
Christopher Fauleta717b992018-04-10 14:43:00 +020011483 h2" on the bind line.
11484
Willy Tarreaub6205fd2012-09-24 12:27:33 +020011485ssl
11486 This setting is only available when support for OpenSSL was built in. It
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030011487 enables SSL deciphering on connections instantiated from this listener. A
Willy Tarreaub6205fd2012-09-24 12:27:33 +020011488 certificate is necessary (see "crt" above). All contents in the buffers will
11489 appear in clear text, so that ACLs and HTTP processing will only have access
Emmanuel Hocdetbd695fe2017-05-15 15:53:41 +020011490 to deciphered contents. SSLv3 is disabled per default, use "ssl-min-ver SSLv3"
11491 to enable it.
Willy Tarreaub6205fd2012-09-24 12:27:33 +020011492
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020011493ssl-max-ver [ SSLv3 | TLSv1.0 | TLSv1.1 | TLSv1.2 | TLSv1.3 ]
11494 This option enforces use of <version> or lower on SSL connections instantiated
11495 from this listener. This option is also available on global statement
11496 "ssl-default-bind-options". See also "ssl-min-ver".
11497
11498ssl-min-ver [ SSLv3 | TLSv1.0 | TLSv1.1 | TLSv1.2 | TLSv1.3 ]
11499 This option enforces use of <version> or upper on SSL connections instantiated
11500 from this listener. This option is also available on global statement
11501 "ssl-default-bind-options". See also "ssl-max-ver".
11502
Emmanuel Hocdet65623372013-01-24 17:17:15 +010011503strict-sni
11504 This setting is only available when support for OpenSSL was built in. The
11505 SSL/TLS negotiation is allow only if the client provided an SNI which match
11506 a certificate. The default certificate is not used.
11507 See the "crt" option for more information.
11508
Willy Tarreau2af207a2015-02-04 00:45:58 +010011509tcp-ut <delay>
Tim Düsterhus4896c442016-11-29 02:15:19 +010011510 Sets the TCP User Timeout for all incoming connections instantiated from this
Willy Tarreau2af207a2015-02-04 00:45:58 +010011511 listening socket. This option is available on Linux since version 2.6.37. It
11512 allows haproxy to configure a timeout for sockets which contain data not
Davor Ocelice9ed2812017-12-25 17:49:28 +010011513 receiving an acknowledgment for the configured delay. This is especially
Willy Tarreau2af207a2015-02-04 00:45:58 +010011514 useful on long-lived connections experiencing long idle periods such as
11515 remote terminals or database connection pools, where the client and server
11516 timeouts must remain high to allow a long period of idle, but where it is
11517 important to detect that the client has disappeared in order to release all
11518 resources associated with its connection (and the server's session). The
11519 argument is a delay expressed in milliseconds by default. This only works
11520 for regular TCP connections, and is ignored for other protocols.
11521
Willy Tarreau1c862c52012-10-05 16:21:00 +020011522tfo
Lukas Tribus0defb902013-02-13 23:35:39 +010011523 Is an optional keyword which is supported only on Linux kernels >= 3.7. It
Willy Tarreau1c862c52012-10-05 16:21:00 +020011524 enables TCP Fast Open on the listening socket, which means that clients which
11525 support this feature will be able to send a request and receive a response
11526 during the 3-way handshake starting from second connection, thus saving one
11527 round-trip after the first connection. This only makes sense with protocols
11528 that use high connection rates and where each round trip matters. This can
11529 possibly cause issues with many firewalls which do not accept data on SYN
11530 packets, so this option should only be enabled once well tested. This option
Lukas Tribus0999f762013-04-02 16:43:24 +020011531 is only supported on TCPv4/TCPv6 sockets and ignored by other ones. You may
11532 need to build HAProxy with USE_TFO=1 if your libc doesn't define
11533 TCP_FASTOPEN.
Willy Tarreau1c862c52012-10-05 16:21:00 +020011534
Nenad Merdanovic188ad3e2015-02-27 19:56:50 +010011535tls-ticket-keys <keyfile>
11536 Sets the TLS ticket keys file to load the keys from. The keys need to be 48
Emeric Brun9e754772019-01-10 17:51:55 +010011537 or 80 bytes long, depending if aes128 or aes256 is used, encoded with base64
11538 with one line per key (ex. openssl rand 80 | openssl base64 -A | xargs echo).
11539 The first key determines the key length used for next keys: you can't mix
11540 aes128 and aes256 keys. Number of keys is specified by the TLS_TICKETS_NO
11541 build option (default 3) and at least as many keys need to be present in
11542 the file. Last TLS_TICKETS_NO keys will be used for decryption and the
11543 penultimate one for encryption. This enables easy key rotation by just
11544 appending new key to the file and reloading the process. Keys must be
11545 periodically rotated (ex. every 12h) or Perfect Forward Secrecy is
11546 compromised. It is also a good idea to keep the keys off any permanent
Nenad Merdanovic188ad3e2015-02-27 19:56:50 +010011547 storage such as hard drives (hint: use tmpfs and don't swap those files).
11548 Lifetime hint can be changed using tune.ssl.timeout.
11549
Willy Tarreaub6205fd2012-09-24 12:27:33 +020011550transparent
11551 Is an optional keyword which is supported only on certain Linux kernels. It
11552 indicates that the addresses will be bound even if they do not belong to the
11553 local machine, and that packets targeting any of these addresses will be
11554 intercepted just as if the addresses were locally configured. This normally
11555 requires that IP forwarding is enabled. Caution! do not use this with the
11556 default address '*', as it would redirect any traffic for the specified port.
11557 This keyword is available only when HAProxy is built with USE_LINUX_TPROXY=1.
11558 This parameter is only compatible with TCPv4 and TCPv6 sockets, depending on
11559 kernel version. Some distribution kernels include backports of the feature,
11560 so check for support with your vendor.
11561
Willy Tarreau77e3af92012-11-24 15:07:23 +010011562v4v6
11563 Is an optional keyword which is supported only on most recent systems
11564 including Linux kernels >= 2.4.21. It is used to bind a socket to both IPv4
11565 and IPv6 when it uses the default address. Doing so is sometimes necessary
11566 on systems which bind to IPv6 only by default. It has no effect on non-IPv6
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030011567 sockets, and is overridden by the "v6only" option.
Willy Tarreau77e3af92012-11-24 15:07:23 +010011568
Willy Tarreau9b6700f2012-11-24 11:55:28 +010011569v6only
11570 Is an optional keyword which is supported only on most recent systems
11571 including Linux kernels >= 2.4.21. It is used to bind a socket to IPv6 only
11572 when it uses the default address. Doing so is sometimes preferred to doing it
Willy Tarreau77e3af92012-11-24 15:07:23 +010011573 system-wide as it is per-listener. It has no effect on non-IPv6 sockets and
11574 has precedence over the "v4v6" option.
Willy Tarreau9b6700f2012-11-24 11:55:28 +010011575
Willy Tarreaub6205fd2012-09-24 12:27:33 +020011576uid <uid>
11577 Sets the owner of the UNIX sockets to the designated system uid. It can also
11578 be set by default in the global section's "unix-bind" statement. Note that
11579 some platforms simply ignore this. This setting is equivalent to the "user"
11580 setting except that the user numeric ID is used instead of its name. This
11581 setting is ignored by non UNIX sockets.
11582
11583user <user>
11584 Sets the owner of the UNIX sockets to the designated system user. It can also
11585 be set by default in the global section's "unix-bind" statement. Note that
11586 some platforms simply ignore this. This setting is equivalent to the "uid"
11587 setting except that the user name is used instead of its uid. This setting is
11588 ignored by non UNIX sockets.
11589
Emeric Brun1a073b42012-09-28 17:07:34 +020011590verify [none|optional|required]
11591 This setting is only available when support for OpenSSL was built in. If set
11592 to 'none', client certificate is not requested. This is the default. In other
11593 cases, a client certificate is requested. If the client does not provide a
11594 certificate after the request and if 'verify' is set to 'required', then the
11595 handshake is aborted, while it would have succeeded if set to 'optional'. The
Emeric Brunfd33a262012-10-11 16:28:27 +020011596 certificate provided by the client is always verified using CAs from
11597 'ca-file' and optional CRLs from 'crl-file'. On verify failure the handshake
11598 is aborted, regardless of the 'verify' option, unless the error code exactly
11599 matches one of those listed with 'ca-ignore-err' or 'crt-ignore-err'.
Willy Tarreau4a5cade2012-04-05 21:09:48 +020011600
Willy Tarreaub6205fd2012-09-24 12:27:33 +0200116015.2. Server and default-server options
Cyril Bontéf0c60612010-02-06 14:44:47 +010011602------------------------------------
Willy Tarreau6a06a402007-07-15 20:15:28 +020011603
Krzysztof Piotr Oledzkic6df0662010-01-05 16:38:49 +010011604The "server" and "default-server" keywords support a certain number of settings
11605which are all passed as arguments on the server line. The order in which those
11606arguments appear does not count, and they are all optional. Some of those
11607settings are single words (booleans) while others expect one or several values
11608after them. In this case, the values must immediately follow the setting name.
11609Except default-server, all those settings must be specified after the server's
11610address if they are used:
Willy Tarreau6a06a402007-07-15 20:15:28 +020011611
Willy Tarreauc57f0e22009-05-10 13:12:33 +020011612 server <name> <address>[:port] [settings ...]
Krzysztof Piotr Oledzkic6df0662010-01-05 16:38:49 +010011613 default-server [settings ...]
Willy Tarreau6a06a402007-07-15 20:15:28 +020011614
Frédéric Lécailled2376272017-03-21 18:52:12 +010011615Note that all these settings are supported both by "server" and "default-server"
11616keywords, except "id" which is only supported by "server".
11617
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010011618The currently supported settings are the following ones.
Willy Tarreau0ba27502007-12-24 16:55:16 +010011619
Willy Tarreauceb4ac92012-04-28 00:41:46 +020011620addr <ipv4|ipv6>
Willy Tarreauc57f0e22009-05-10 13:12:33 +020011621 Using the "addr" parameter, it becomes possible to use a different IP address
Baptiste Assmann13f83532016-03-06 23:14:36 +010011622 to send health-checks or to probe the agent-check. On some servers, it may be
11623 desirable to dedicate an IP address to specific component able to perform
11624 complex tests which are more suitable to health-checks than the application.
11625 This parameter is ignored if the "check" parameter is not set. See also the
11626 "port" parameter.
Willy Tarreau6a06a402007-07-15 20:15:28 +020011627
Simon Hormand60d6912013-11-25 10:46:36 +090011628agent-check
11629 Enable an auxiliary agent check which is run independently of a regular
Willy Tarreau81f5d942013-12-09 20:51:51 +010011630 health check. An agent health check is performed by making a TCP connection
Willy Tarreau7a0139e2018-12-16 08:42:56 +010011631 to the port set by the "agent-port" parameter and reading an ASCII string
11632 terminated by the first '\r' or '\n' met. The string is made of a series of
11633 words delimited by spaces, tabs or commas in any order, each consisting of :
Simon Hormand60d6912013-11-25 10:46:36 +090011634
Willy Tarreau81f5d942013-12-09 20:51:51 +010011635 - An ASCII representation of a positive integer percentage, e.g. "75%".
Simon Hormand60d6912013-11-25 10:46:36 +090011636 Values in this format will set the weight proportional to the initial
Willy Tarreauc5af3a62014-10-07 15:27:33 +020011637 weight of a server as configured when haproxy starts. Note that a zero
11638 weight is reported on the stats page as "DRAIN" since it has the same
11639 effect on the server (it's removed from the LB farm).
Simon Hormand60d6912013-11-25 10:46:36 +090011640
Davor Ocelice9ed2812017-12-25 17:49:28 +010011641 - The string "maxconn:" followed by an integer (no space between). Values
11642 in this format will set the maxconn of a server. The maximum number of
11643 connections advertised needs to be multiplied by the number of load
11644 balancers and different backends that use this health check to get the
11645 total number of connections the server might receive. Example: maxconn:30
Nenad Merdanovic174dd372016-04-24 23:10:06 +020011646
Willy Tarreau81f5d942013-12-09 20:51:51 +010011647 - The word "ready". This will turn the server's administrative state to the
Davor Ocelice9ed2812017-12-25 17:49:28 +010011648 READY mode, thus canceling any DRAIN or MAINT state
Simon Hormand60d6912013-11-25 10:46:36 +090011649
Willy Tarreau81f5d942013-12-09 20:51:51 +010011650 - The word "drain". This will turn the server's administrative state to the
11651 DRAIN mode, thus it will not accept any new connections other than those
11652 that are accepted via persistence.
Simon Hormand60d6912013-11-25 10:46:36 +090011653
Willy Tarreau81f5d942013-12-09 20:51:51 +010011654 - The word "maint". This will turn the server's administrative state to the
11655 MAINT mode, thus it will not accept any new connections at all, and health
11656 checks will be stopped.
Simon Hormand60d6912013-11-25 10:46:36 +090011657
Willy Tarreau81f5d942013-12-09 20:51:51 +010011658 - The words "down", "failed", or "stopped", optionally followed by a
11659 description string after a sharp ('#'). All of these mark the server's
11660 operating state as DOWN, but since the word itself is reported on the stats
11661 page, the difference allows an administrator to know if the situation was
11662 expected or not : the service may intentionally be stopped, may appear up
Davor Ocelice9ed2812017-12-25 17:49:28 +010011663 but fail some validity tests, or may be seen as down (e.g. missing process,
Willy Tarreau81f5d942013-12-09 20:51:51 +010011664 or port not responding).
Simon Hormand60d6912013-11-25 10:46:36 +090011665
Willy Tarreau81f5d942013-12-09 20:51:51 +010011666 - The word "up" sets back the server's operating state as UP if health checks
11667 also report that the service is accessible.
Simon Hormand60d6912013-11-25 10:46:36 +090011668
Willy Tarreau81f5d942013-12-09 20:51:51 +010011669 Parameters which are not advertised by the agent are not changed. For
11670 example, an agent might be designed to monitor CPU usage and only report a
11671 relative weight and never interact with the operating status. Similarly, an
11672 agent could be designed as an end-user interface with 3 radio buttons
11673 allowing an administrator to change only the administrative state. However,
11674 it is important to consider that only the agent may revert its own actions,
11675 so if a server is set to DRAIN mode or to DOWN state using the agent, the
11676 agent must implement the other equivalent actions to bring the service into
11677 operations again.
Simon Hormand60d6912013-11-25 10:46:36 +090011678
Simon Horman2f1f9552013-11-25 10:46:37 +090011679 Failure to connect to the agent is not considered an error as connectivity
11680 is tested by the regular health check which is enabled by the "check"
Willy Tarreau81f5d942013-12-09 20:51:51 +010011681 parameter. Warning though, it is not a good idea to stop an agent after it
11682 reports "down", since only an agent reporting "up" will be able to turn the
11683 server up again. Note that the CLI on the Unix stats socket is also able to
Willy Tarreau989222a2016-01-15 10:26:26 +010011684 force an agent's result in order to work around a bogus agent if needed.
Simon Horman2f1f9552013-11-25 10:46:37 +090011685
Willy Tarreau81f5d942013-12-09 20:51:51 +010011686 Requires the "agent-port" parameter to be set. See also the "agent-inter"
Frédéric Lécailled2376272017-03-21 18:52:12 +010011687 and "no-agent-check" parameters.
Simon Hormand60d6912013-11-25 10:46:36 +090011688
James Brown55f9ff12015-10-21 18:19:05 -070011689agent-send <string>
11690 If this option is specified, haproxy will send the given string (verbatim)
11691 to the agent server upon connection. You could, for example, encode
11692 the backend name into this string, which would enable your agent to send
11693 different responses based on the backend. Make sure to include a '\n' if
11694 you want to terminate your request with a newline.
11695
Simon Hormand60d6912013-11-25 10:46:36 +090011696agent-inter <delay>
11697 The "agent-inter" parameter sets the interval between two agent checks
11698 to <delay> milliseconds. If left unspecified, the delay defaults to 2000 ms.
11699
11700 Just as with every other time-based parameter, it may be entered in any
11701 other explicit unit among { us, ms, s, m, h, d }. The "agent-inter"
11702 parameter also serves as a timeout for agent checks "timeout check" is
11703 not set. In order to reduce "resonance" effects when multiple servers are
11704 hosted on the same hardware, the agent and health checks of all servers
11705 are started with a small time offset between them. It is also possible to
11706 add some random noise in the agent and health checks interval using the
11707 global "spread-checks" keyword. This makes sense for instance when a lot
11708 of backends use the same servers.
11709
11710 See also the "agent-check" and "agent-port" parameters.
11711
Misiek768d8602017-01-09 09:52:43 +010011712agent-addr <addr>
11713 The "agent-addr" parameter sets address for agent check.
11714
11715 You can offload agent-check to another target, so you can make single place
11716 managing status and weights of servers defined in haproxy in case you can't
11717 make self-aware and self-managing services. You can specify both IP or
11718 hostname, it will be resolved.
11719
Simon Hormand60d6912013-11-25 10:46:36 +090011720agent-port <port>
11721 The "agent-port" parameter sets the TCP port used for agent checks.
11722
11723 See also the "agent-check" and "agent-inter" parameters.
11724
Olivier Houchard8cb2d2e2019-05-06 18:58:48 +020011725allow-0rtt
11726 Allow sending early data to the server when using TLS 1.3.
Olivier Houchard22c9b442019-05-06 19:01:04 +020011727 Note that early data will be sent only if the client used early data, or
11728 if the backend uses "retry-on" with the "0rtt-rejected" keyword.
Olivier Houchard8cb2d2e2019-05-06 18:58:48 +020011729
Olivier Houchardc7566002018-11-20 23:33:50 +010011730alpn <protocols>
11731 This enables the TLS ALPN extension and advertises the specified protocol
11732 list as supported on top of ALPN. The protocol list consists in a comma-
11733 delimited list of protocol names, for instance: "http/1.1,http/1.0" (without
John Roeslerfb2fce12019-07-10 15:45:51 -050011734 quotes). This requires that the SSL library is built with support for TLS
Olivier Houchardc7566002018-11-20 23:33:50 +010011735 extensions enabled (check with haproxy -vv). The ALPN extension replaces the
11736 initial NPN extension. ALPN is required to connect to HTTP/2 servers.
11737 Versions of OpenSSL prior to 1.0.2 didn't support ALPN and only supposed the
11738 now obsolete NPN extension.
11739 If both HTTP/2 and HTTP/1.1 are expected to be supported, both versions can
11740 be advertised, in order of preference, like below :
11741
11742 server 127.0.0.1:443 ssl crt pub.pem alpn h2,http/1.1
11743
Willy Tarreauc57f0e22009-05-10 13:12:33 +020011744backup
11745 When "backup" is present on a server line, the server is only used in load
11746 balancing when all other non-backup servers are unavailable. Requests coming
11747 with a persistence cookie referencing the server will always be served
11748 though. By default, only the first operational backup server is used, unless
Frédéric Lécailled2376272017-03-21 18:52:12 +010011749 the "allbackups" option is set in the backend. See also the "no-backup" and
11750 "allbackups" options.
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010011751
Emeric Brunef42d922012-10-11 16:11:36 +020011752ca-file <cafile>
11753 This setting is only available when support for OpenSSL was built in. It
11754 designates a PEM file from which to load CA certificates used to verify
11755 server's certificate.
11756
Willy Tarreauc57f0e22009-05-10 13:12:33 +020011757check
11758 This option enables health checks on the server. By default, a server is
Patrick Mézardb7aeec62012-01-22 16:01:22 +010011759 always considered available. If "check" is set, the server is available when
11760 accepting periodic TCP connections, to ensure that it is really able to serve
11761 requests. The default address and port to send the tests to are those of the
11762 server, and the default source is the same as the one defined in the
11763 backend. It is possible to change the address using the "addr" parameter, the
11764 port using the "port" parameter, the source address using the "source"
11765 address, and the interval and timers using the "inter", "rise" and "fall"
Simon Hormanafc47ee2013-11-25 10:46:35 +090011766 parameters. The request method is define in the backend using the "httpchk",
11767 "smtpchk", "mysql-check", "pgsql-check" and "ssl-hello-chk" options. Please
Frédéric Lécailled2376272017-03-21 18:52:12 +010011768 refer to those options and parameters for more information. See also
11769 "no-check" option.
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010011770
Willy Tarreau6c16adc2012-10-05 00:04:16 +020011771check-send-proxy
11772 This option forces emission of a PROXY protocol line with outgoing health
11773 checks, regardless of whether the server uses send-proxy or not for the
11774 normal traffic. By default, the PROXY protocol is enabled for health checks
11775 if it is already enabled for normal traffic and if no "port" nor "addr"
11776 directive is present. However, if such a directive is present, the
11777 "check-send-proxy" option needs to be used to force the use of the
11778 protocol. See also the "send-proxy" option for more information.
11779
Olivier Houchard92150142018-12-21 19:47:01 +010011780check-alpn <protocols>
11781 Defines which protocols to advertise with ALPN. The protocol list consists in
11782 a comma-delimited list of protocol names, for instance: "http/1.1,http/1.0"
11783 (without quotes). If it is not set, the server ALPN is used.
11784
Jérôme Magninae9bb762018-12-09 16:08:26 +010011785check-sni <sni>
Olivier Houchard9130a962017-10-17 17:33:43 +020011786 This option allows you to specify the SNI to be used when doing health checks
Jérôme Magninae9bb762018-12-09 16:08:26 +010011787 over SSL. It is only possible to use a string to set <sni>. If you want to
11788 set a SNI for proxied traffic, see "sni".
Olivier Houchard9130a962017-10-17 17:33:43 +020011789
Willy Tarreau763a95b2012-10-04 23:15:39 +020011790check-ssl
11791 This option forces encryption of all health checks over SSL, regardless of
11792 whether the server uses SSL or not for the normal traffic. This is generally
11793 used when an explicit "port" or "addr" directive is specified and SSL health
11794 checks are not inherited. It is important to understand that this option
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030011795 inserts an SSL transport layer below the checks, so that a simple TCP connect
Willy Tarreau763a95b2012-10-04 23:15:39 +020011796 check becomes an SSL connect, which replaces the old ssl-hello-chk. The most
11797 common use is to send HTTPS checks by combining "httpchk" with SSL checks.
Davor Ocelice9ed2812017-12-25 17:49:28 +010011798 All SSL settings are common to health checks and traffic (e.g. ciphers).
Frédéric Lécailled2376272017-03-21 18:52:12 +010011799 See the "ssl" option for more information and "no-check-ssl" to disable
11800 this option.
Willy Tarreau763a95b2012-10-04 23:15:39 +020011801
Alexander Liu2a54bb72019-05-22 19:44:48 +080011802check-via-socks4
John Roeslerfb2fce12019-07-10 15:45:51 -050011803 This option enables outgoing health checks using upstream socks4 proxy. By
Alexander Liu2a54bb72019-05-22 19:44:48 +080011804 default, the health checks won't go through socks tunnel even it was enabled
11805 for normal traffic.
11806
Willy Tarreaua0ee1d02012-09-10 09:01:23 +020011807ciphers <ciphers>
Dirkjan Bussink415150f2018-09-14 11:14:21 +020011808 This setting is only available when support for OpenSSL was built in. This
11809 option sets the string describing the list of cipher algorithms that is
11810 negotiated during the SSL/TLS handshake with the server. The format of the
Bertrand Jacquin4f03ab02019-02-03 18:48:49 +000011811 string is defined in "man 1 ciphers" from OpenSSL man pages. For background
11812 information and recommendations see e.g.
11813 (https://wiki.mozilla.org/Security/Server_Side_TLS) and
11814 (https://mozilla.github.io/server-side-tls/ssl-config-generator/). For TLSv1.3
11815 cipher configuration, please check the "ciphersuites" keyword.
Willy Tarreaua0ee1d02012-09-10 09:01:23 +020011816
Dirkjan Bussink415150f2018-09-14 11:14:21 +020011817ciphersuites <ciphersuites>
11818 This setting is only available when support for OpenSSL was built in and
11819 OpenSSL 1.1.1 or later was used to build HAProxy. This option sets the string
11820 describing the list of cipher algorithms that is negotiated during the TLS
11821 1.3 handshake with the server. The format of the string is defined in
Bertrand Jacquin4f03ab02019-02-03 18:48:49 +000011822 "man 1 ciphers" from OpenSSL man pages under the "ciphersuites" section.
11823 For cipher configuration for TLSv1.2 and earlier, please check the "ciphers"
11824 keyword.
Dirkjan Bussink415150f2018-09-14 11:14:21 +020011825
Willy Tarreauc57f0e22009-05-10 13:12:33 +020011826cookie <value>
11827 The "cookie" parameter sets the cookie value assigned to the server to
11828 <value>. This value will be checked in incoming requests, and the first
11829 operational server possessing the same value will be selected. In return, in
11830 cookie insertion or rewrite modes, this value will be assigned to the cookie
11831 sent to the client. There is nothing wrong in having several servers sharing
11832 the same cookie value, and it is in fact somewhat common between normal and
11833 backup servers. See also the "cookie" keyword in backend section.
11834
Emeric Brunef42d922012-10-11 16:11:36 +020011835crl-file <crlfile>
11836 This setting is only available when support for OpenSSL was built in. It
11837 designates a PEM file from which to load certificate revocation list used
11838 to verify server's certificate.
11839
Emeric Bruna7aa3092012-10-26 12:58:00 +020011840crt <cert>
11841 This setting is only available when support for OpenSSL was built in.
11842 It designates a PEM file from which to load both a certificate and the
11843 associated private key. This file can be built by concatenating both PEM
11844 files into one. This certificate will be sent if the server send a client
11845 certificate request.
11846
Willy Tarreau96839092010-03-29 10:02:24 +020011847disabled
11848 The "disabled" keyword starts the server in the "disabled" state. That means
11849 that it is marked down in maintenance mode, and no connection other than the
11850 ones allowed by persist mode will reach it. It is very well suited to setup
11851 new servers, because normal traffic will never reach them, while it is still
11852 possible to test the service by making use of the force-persist mechanism.
Frédéric Lécailled2376272017-03-21 18:52:12 +010011853 See also "enabled" setting.
Willy Tarreau96839092010-03-29 10:02:24 +020011854
Frédéric Lécailled2376272017-03-21 18:52:12 +010011855enabled
11856 This option may be used as 'server' setting to reset any 'disabled'
11857 setting which would have been inherited from 'default-server' directive as
11858 default value.
11859 It may also be used as 'default-server' setting to reset any previous
11860 'default-server' 'disabled' setting.
Willy Tarreau96839092010-03-29 10:02:24 +020011861
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010011862error-limit <count>
Willy Tarreau983e01e2010-01-11 18:42:06 +010011863 If health observing is enabled, the "error-limit" parameter specifies the
11864 number of consecutive errors that triggers event selected by the "on-error"
11865 option. By default it is set to 10 consecutive errors.
Krzysztof Piotr Oledzki97f07b82009-12-15 22:31:24 +010011866
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010011867 See also the "check", "error-limit" and "on-error".
Krzysztof Piotr Oledzki97f07b82009-12-15 22:31:24 +010011868
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010011869fall <count>
Willy Tarreauc57f0e22009-05-10 13:12:33 +020011870 The "fall" parameter states that a server will be considered as dead after
11871 <count> consecutive unsuccessful health checks. This value defaults to 3 if
11872 unspecified. See also the "check", "inter" and "rise" parameters.
11873
Emeric Brun8694b9a2012-10-05 14:39:07 +020011874force-sslv3
11875 This option enforces use of SSLv3 only when SSL is used to communicate with
11876 the server. SSLv3 is generally less expensive than the TLS counterparts for
Emeric Brun2c86cbf2014-10-30 15:56:50 +010011877 high connection rates. This option is also available on global statement
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020011878 "ssl-default-server-options". See also "ssl-min-ver" and ssl-max-ver".
Emeric Brun8694b9a2012-10-05 14:39:07 +020011879
11880force-tlsv10
11881 This option enforces use of TLSv1.0 only when SSL is used to communicate with
Emeric Brun2c86cbf2014-10-30 15:56:50 +010011882 the server. This option is also available on global statement
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020011883 "ssl-default-server-options". See also "ssl-min-ver" and ssl-max-ver".
Emeric Brun8694b9a2012-10-05 14:39:07 +020011884
11885force-tlsv11
11886 This option enforces use of TLSv1.1 only when SSL is used to communicate with
Emeric Brun2c86cbf2014-10-30 15:56:50 +010011887 the server. This option is also available on global statement
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020011888 "ssl-default-server-options". See also "ssl-min-ver" and ssl-max-ver".
Emeric Brun8694b9a2012-10-05 14:39:07 +020011889
11890force-tlsv12
11891 This option enforces use of TLSv1.2 only when SSL is used to communicate with
Emeric Brun2c86cbf2014-10-30 15:56:50 +010011892 the server. This option is also available on global statement
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020011893 "ssl-default-server-options". See also "ssl-min-ver" and ssl-max-ver".
Emeric Brun8694b9a2012-10-05 14:39:07 +020011894
Emmanuel Hocdet42fb9802017-03-30 19:29:39 +020011895force-tlsv13
11896 This option enforces use of TLSv1.3 only when SSL is used to communicate with
11897 the server. This option is also available on global statement
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020011898 "ssl-default-server-options". See also "ssl-min-ver" and ssl-max-ver".
Emmanuel Hocdet42fb9802017-03-30 19:29:39 +020011899
Willy Tarreauc57f0e22009-05-10 13:12:33 +020011900id <value>
Willy Tarreau53fb4ae2009-10-04 23:04:08 +020011901 Set a persistent ID for the server. This ID must be positive and unique for
11902 the proxy. An unused ID will automatically be assigned if unset. The first
11903 assigned value will be 1. This ID is currently only returned in statistics.
Willy Tarreauc57f0e22009-05-10 13:12:33 +020011904
Willy Tarreau6a031d12016-11-07 19:42:35 +010011905init-addr {last | libc | none | <ip>},[...]*
11906 Indicate in what order the server's address should be resolved upon startup
11907 if it uses an FQDN. Attempts are made to resolve the address by applying in
Davor Ocelice9ed2812017-12-25 17:49:28 +010011908 turn each of the methods mentioned in the comma-delimited list. The first
Willy Tarreau6a031d12016-11-07 19:42:35 +010011909 method which succeeds is used. If the end of the list is reached without
11910 finding a working method, an error is thrown. Method "last" suggests to pick
11911 the address which appears in the state file (see "server-state-file"). Method
11912 "libc" uses the libc's internal resolver (gethostbyname() or getaddrinfo()
11913 depending on the operating system and build options). Method "none"
11914 specifically indicates that the server should start without any valid IP
11915 address in a down state. It can be useful to ignore some DNS issues upon
11916 startup, waiting for the situation to get fixed later. Finally, an IP address
11917 (IPv4 or IPv6) may be provided. It can be the currently known address of the
Davor Ocelice9ed2812017-12-25 17:49:28 +010011918 server (e.g. filled by a configuration generator), or the address of a dummy
Willy Tarreau6a031d12016-11-07 19:42:35 +010011919 server used to catch old sessions and present them with a decent error
11920 message for example. When the "first" load balancing algorithm is used, this
11921 IP address could point to a fake server used to trigger the creation of new
11922 instances on the fly. This option defaults to "last,libc" indicating that the
11923 previous address found in the state file (if any) is used first, otherwise
11924 the libc's resolver is used. This ensures continued compatibility with the
Davor Ocelice9ed2812017-12-25 17:49:28 +010011925 historic behavior.
Willy Tarreau6a031d12016-11-07 19:42:35 +010011926
11927 Example:
11928 defaults
11929 # never fail on address resolution
11930 default-server init-addr last,libc,none
11931
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010011932inter <delay>
11933fastinter <delay>
11934downinter <delay>
Willy Tarreauc57f0e22009-05-10 13:12:33 +020011935 The "inter" parameter sets the interval between two consecutive health checks
11936 to <delay> milliseconds. If left unspecified, the delay defaults to 2000 ms.
11937 It is also possible to use "fastinter" and "downinter" to optimize delays
11938 between checks depending on the server state :
11939
Pieter Baauw44fc9df2015-09-17 21:30:46 +020011940 Server state | Interval used
11941 ----------------------------------------+----------------------------------
11942 UP 100% (non-transitional) | "inter"
11943 ----------------------------------------+----------------------------------
11944 Transitionally UP (going down "fall"), | "fastinter" if set,
11945 Transitionally DOWN (going up "rise"), | "inter" otherwise.
11946 or yet unchecked. |
11947 ----------------------------------------+----------------------------------
11948 DOWN 100% (non-transitional) | "downinter" if set,
11949 | "inter" otherwise.
11950 ----------------------------------------+----------------------------------
Willy Tarreaud72758d2010-01-12 10:42:19 +010011951
Willy Tarreauc57f0e22009-05-10 13:12:33 +020011952 Just as with every other time-based parameter, they can be entered in any
11953 other explicit unit among { us, ms, s, m, h, d }. The "inter" parameter also
11954 serves as a timeout for health checks sent to servers if "timeout check" is
11955 not set. In order to reduce "resonance" effects when multiple servers are
Simon Hormand60d6912013-11-25 10:46:36 +090011956 hosted on the same hardware, the agent and health checks of all servers
11957 are started with a small time offset between them. It is also possible to
11958 add some random noise in the agent and health checks interval using the
11959 global "spread-checks" keyword. This makes sense for instance when a lot
11960 of backends use the same servers.
Willy Tarreauc57f0e22009-05-10 13:12:33 +020011961
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010011962maxconn <maxconn>
Willy Tarreauc57f0e22009-05-10 13:12:33 +020011963 The "maxconn" parameter specifies the maximal number of concurrent
11964 connections that will be sent to this server. If the number of incoming
11965 concurrent requests goes higher than this value, they will be queued, waiting
11966 for a connection to be released. This parameter is very important as it can
11967 save fragile servers from going down under extreme loads. If a "minconn"
11968 parameter is specified, the limit becomes dynamic. The default value is "0"
11969 which means unlimited. See also the "minconn" and "maxqueue" parameters, and
11970 the backend's "fullconn" keyword.
11971
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010011972maxqueue <maxqueue>
Willy Tarreauc57f0e22009-05-10 13:12:33 +020011973 The "maxqueue" parameter specifies the maximal number of connections which
11974 will wait in the queue for this server. If this limit is reached, next
11975 requests will be redispatched to other servers instead of indefinitely
11976 waiting to be served. This will break persistence but may allow people to
11977 quickly re-log in when the server they try to connect to is dying. The
11978 default value is "0" which means the queue is unlimited. See also the
11979 "maxconn" and "minconn" parameters.
11980
Willy Tarreau9c538e02019-01-23 10:21:49 +010011981max-reuse <count>
11982 The "max-reuse" argument indicates the HTTP connection processors that they
11983 should not reuse a server connection more than this number of times to send
11984 new requests. Permitted values are -1 (the default), which disables this
11985 limit, or any positive value. Value zero will effectively disable keep-alive.
11986 This is only used to work around certain server bugs which cause them to leak
11987 resources over time. The argument is not necessarily respected by the lower
11988 layers as there might be technical limitations making it impossible to
11989 enforce. At least HTTP/2 connections to servers will respect it.
11990
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010011991minconn <minconn>
Willy Tarreauc57f0e22009-05-10 13:12:33 +020011992 When the "minconn" parameter is set, the maxconn limit becomes a dynamic
11993 limit following the backend's load. The server will always accept at least
11994 <minconn> connections, never more than <maxconn>, and the limit will be on
11995 the ramp between both values when the backend has less than <fullconn>
11996 concurrent connections. This makes it possible to limit the load on the
11997 server during normal loads, but push it further for important loads without
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +010011998 overloading the server during exceptional loads. See also the "maxconn"
Willy Tarreauc57f0e22009-05-10 13:12:33 +020011999 and "maxqueue" parameters, as well as the "fullconn" backend keyword.
Krzysztof Piotr Oledzki97f07b82009-12-15 22:31:24 +010012000
Willy Tarreaud72f0f32015-10-13 14:50:22 +020012001namespace <name>
12002 On Linux, it is possible to specify which network namespace a socket will
12003 belong to. This directive makes it possible to explicitly bind a server to
12004 a namespace different from the default one. Please refer to your operating
12005 system's documentation to find more details about network namespaces.
12006
Frédéric Lécailled2376272017-03-21 18:52:12 +010012007no-agent-check
12008 This option may be used as "server" setting to reset any "agent-check"
12009 setting which would have been inherited from "default-server" directive as
12010 default value.
12011 It may also be used as "default-server" setting to reset any previous
12012 "default-server" "agent-check" setting.
12013
12014no-backup
12015 This option may be used as "server" setting to reset any "backup"
12016 setting which would have been inherited from "default-server" directive as
12017 default value.
12018 It may also be used as "default-server" setting to reset any previous
12019 "default-server" "backup" setting.
12020
12021no-check
12022 This option may be used as "server" setting to reset any "check"
12023 setting which would have been inherited from "default-server" directive as
12024 default value.
12025 It may also be used as "default-server" setting to reset any previous
12026 "default-server" "check" setting.
12027
12028no-check-ssl
12029 This option may be used as "server" setting to reset any "check-ssl"
12030 setting which would have been inherited from "default-server" directive as
12031 default value.
12032 It may also be used as "default-server" setting to reset any previous
12033 "default-server" "check-ssl" setting.
12034
Frédéric Lécailled2376272017-03-21 18:52:12 +010012035no-send-proxy
12036 This option may be used as "server" setting to reset any "send-proxy"
12037 setting which would have been inherited from "default-server" directive as
12038 default value.
12039 It may also be used as "default-server" setting to reset any previous
12040 "default-server" "send-proxy" setting.
12041
12042no-send-proxy-v2
12043 This option may be used as "server" setting to reset any "send-proxy-v2"
12044 setting which would have been inherited from "default-server" directive as
12045 default value.
12046 It may also be used as "default-server" setting to reset any previous
12047 "default-server" "send-proxy-v2" setting.
12048
12049no-send-proxy-v2-ssl
12050 This option may be used as "server" setting to reset any "send-proxy-v2-ssl"
12051 setting which would have been inherited from "default-server" directive as
12052 default value.
12053 It may also be used as "default-server" setting to reset any previous
12054 "default-server" "send-proxy-v2-ssl" setting.
12055
12056no-send-proxy-v2-ssl-cn
12057 This option may be used as "server" setting to reset any "send-proxy-v2-ssl-cn"
12058 setting which would have been inherited from "default-server" directive as
12059 default value.
12060 It may also be used as "default-server" setting to reset any previous
12061 "default-server" "send-proxy-v2-ssl-cn" setting.
12062
12063no-ssl
12064 This option may be used as "server" setting to reset any "ssl"
12065 setting which would have been inherited from "default-server" directive as
12066 default value.
12067 It may also be used as "default-server" setting to reset any previous
12068 "default-server" "ssl" setting.
12069
Willy Tarreau2a3fb1c2015-02-05 16:47:07 +010012070no-ssl-reuse
12071 This option disables SSL session reuse when SSL is used to communicate with
12072 the server. It will force the server to perform a full handshake for every
12073 new connection. It's probably only useful for benchmarking, troubleshooting,
12074 and for paranoid users.
12075
Emeric Brun9b3009b2012-10-05 11:55:06 +020012076no-sslv3
Willy Tarreaua0ee1d02012-09-10 09:01:23 +020012077 This option disables support for SSLv3 when SSL is used to communicate with
12078 the server. Note that SSLv2 is disabled in the code and cannot be enabled
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020012079 using any configuration option. Use "ssl-min-ver" and "ssl-max-ver" instead.
Willy Tarreaua0ee1d02012-09-10 09:01:23 +020012080
Emmanuel Hocdet6cb2d1e2017-03-30 14:43:31 +020012081 Supported in default-server: No
12082
Emeric Brunf9c5c472012-10-11 15:28:34 +020012083no-tls-tickets
12084 This setting is only available when support for OpenSSL was built in. It
12085 disables the stateless session resumption (RFC 5077 TLS Ticket
12086 extension) and force to use stateful session resumption. Stateless
Emeric Brun2c86cbf2014-10-30 15:56:50 +010012087 session resumption is more expensive in CPU usage for servers. This option
12088 is also available on global statement "ssl-default-server-options".
Frédéric Lécailled2376272017-03-21 18:52:12 +010012089 See also "tls-tickets".
Emeric Brunf9c5c472012-10-11 15:28:34 +020012090
Emeric Brun9b3009b2012-10-05 11:55:06 +020012091no-tlsv10
Emeric Brun8694b9a2012-10-05 14:39:07 +020012092 This option disables support for TLSv1.0 when SSL is used to communicate with
Emeric Brunf5da4932012-09-28 19:42:54 +020012093 the server. Note that SSLv2 is disabled in the code and cannot be enabled
12094 using any configuration option. TLSv1 is more expensive than SSLv3 so it
Emeric Brun2c86cbf2014-10-30 15:56:50 +010012095 often makes sense to disable it when communicating with local servers. This
12096 option is also available on global statement "ssl-default-server-options".
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020012097 Use "ssl-min-ver" and "ssl-max-ver" instead.
Willy Tarreau763a95b2012-10-04 23:15:39 +020012098
Emmanuel Hocdet6cb2d1e2017-03-30 14:43:31 +020012099 Supported in default-server: No
12100
Emeric Brun9b3009b2012-10-05 11:55:06 +020012101no-tlsv11
Emeric Brun8694b9a2012-10-05 14:39:07 +020012102 This option disables support for TLSv1.1 when SSL is used to communicate with
Emeric Brunf5da4932012-09-28 19:42:54 +020012103 the server. Note that SSLv2 is disabled in the code and cannot be enabled
12104 using any configuration option. TLSv1 is more expensive than SSLv3 so it
Emeric Brun2c86cbf2014-10-30 15:56:50 +010012105 often makes sense to disable it when communicating with local servers. This
12106 option is also available on global statement "ssl-default-server-options".
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020012107 Use "ssl-min-ver" and "ssl-max-ver" instead.
Willy Tarreau763a95b2012-10-04 23:15:39 +020012108
Emmanuel Hocdet6cb2d1e2017-03-30 14:43:31 +020012109 Supported in default-server: No
12110
Emeric Brun9b3009b2012-10-05 11:55:06 +020012111no-tlsv12
Emeric Brun8694b9a2012-10-05 14:39:07 +020012112 This option disables support for TLSv1.2 when SSL is used to communicate with
Willy Tarreaua0ee1d02012-09-10 09:01:23 +020012113 the server. Note that SSLv2 is disabled in the code and cannot be enabled
12114 using any configuration option. TLSv1 is more expensive than SSLv3 so it
Emeric Brun2c86cbf2014-10-30 15:56:50 +010012115 often makes sense to disable it when communicating with local servers. This
12116 option is also available on global statement "ssl-default-server-options".
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020012117 Use "ssl-min-ver" and "ssl-max-ver" instead.
Emmanuel Hocdet42fb9802017-03-30 19:29:39 +020012118
12119 Supported in default-server: No
12120
12121no-tlsv13
12122 This option disables support for TLSv1.3 when SSL is used to communicate with
12123 the server. Note that SSLv2 is disabled in the code and cannot be enabled
12124 using any configuration option. TLSv1 is more expensive than SSLv3 so it
12125 often makes sense to disable it when communicating with local servers. This
12126 option is also available on global statement "ssl-default-server-options".
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020012127 Use "ssl-min-ver" and "ssl-max-ver" instead.
Willy Tarreaua0ee1d02012-09-10 09:01:23 +020012128
Emmanuel Hocdet6cb2d1e2017-03-30 14:43:31 +020012129 Supported in default-server: No
12130
Frédéric Lécailled2376272017-03-21 18:52:12 +010012131no-verifyhost
12132 This option may be used as "server" setting to reset any "verifyhost"
12133 setting which would have been inherited from "default-server" directive as
12134 default value.
12135 It may also be used as "default-server" setting to reset any previous
12136 "default-server" "verifyhost" setting.
Willy Tarreau763a95b2012-10-04 23:15:39 +020012137
Frédéric Lécaille1b9423d2019-07-04 14:19:06 +020012138no-tfo
12139 This option may be used as "server" setting to reset any "tfo"
12140 setting which would have been inherited from "default-server" directive as
12141 default value.
12142 It may also be used as "default-server" setting to reset any previous
12143 "default-server" "tfo" setting.
12144
Simon Hormanfa461682011-06-25 09:39:49 +090012145non-stick
12146 Never add connections allocated to this sever to a stick-table.
12147 This may be used in conjunction with backup to ensure that
12148 stick-table persistence is disabled for backup servers.
12149
Olivier Houchardc7566002018-11-20 23:33:50 +010012150npn <protocols>
12151 This enables the NPN TLS extension and advertises the specified protocol list
12152 as supported on top of NPN. The protocol list consists in a comma-delimited
12153 list of protocol names, for instance: "http/1.1,http/1.0" (without quotes).
John Roeslerfb2fce12019-07-10 15:45:51 -050012154 This requires that the SSL library is built with support for TLS extensions
Olivier Houchardc7566002018-11-20 23:33:50 +010012155 enabled (check with haproxy -vv). Note that the NPN extension has been
12156 replaced with the ALPN extension (see the "alpn" keyword), though this one is
12157 only available starting with OpenSSL 1.0.2.
12158
Krzysztof Piotr Oledzki97f07b82009-12-15 22:31:24 +010012159observe <mode>
12160 This option enables health adjusting based on observing communication with
12161 the server. By default this functionality is disabled and enabling it also
12162 requires to enable health checks. There are two supported modes: "layer4" and
12163 "layer7". In layer4 mode, only successful/unsuccessful tcp connections are
12164 significant. In layer7, which is only allowed for http proxies, responses
12165 received from server are verified, like valid/wrong http code, unparsable
Willy Tarreau150d1462012-03-10 08:19:02 +010012166 headers, a timeout, etc. Valid status codes include 100 to 499, 501 and 505.
Krzysztof Piotr Oledzki97f07b82009-12-15 22:31:24 +010012167
12168 See also the "check", "on-error" and "error-limit".
12169
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010012170on-error <mode>
Krzysztof Piotr Oledzki97f07b82009-12-15 22:31:24 +010012171 Select what should happen when enough consecutive errors are detected.
12172 Currently, four modes are available:
12173 - fastinter: force fastinter
12174 - fail-check: simulate a failed check, also forces fastinter (default)
12175 - sudden-death: simulate a pre-fatal failed health check, one more failed
12176 check will mark a server down, forces fastinter
12177 - mark-down: mark the server immediately down and force fastinter
12178
12179 See also the "check", "observe" and "error-limit".
12180
Simon Hormane0d1bfb2011-06-21 14:34:58 +090012181on-marked-down <action>
12182 Modify what occurs when a server is marked down.
12183 Currently one action is available:
Justin Karnegeseb2c24a2012-05-24 15:28:52 -070012184 - shutdown-sessions: Shutdown peer sessions. When this setting is enabled,
12185 all connections to the server are immediately terminated when the server
12186 goes down. It might be used if the health check detects more complex cases
12187 than a simple connection status, and long timeouts would cause the service
12188 to remain unresponsive for too long a time. For instance, a health check
12189 might detect that a database is stuck and that there's no chance to reuse
12190 existing connections anymore. Connections killed this way are logged with
12191 a 'D' termination code (for "Down").
Simon Hormane0d1bfb2011-06-21 14:34:58 +090012192
12193 Actions are disabled by default
12194
Justin Karnegeseb2c24a2012-05-24 15:28:52 -070012195on-marked-up <action>
12196 Modify what occurs when a server is marked up.
12197 Currently one action is available:
12198 - shutdown-backup-sessions: Shutdown sessions on all backup servers. This is
12199 done only if the server is not in backup state and if it is not disabled
12200 (it must have an effective weight > 0). This can be used sometimes to force
12201 an active server to take all the traffic back after recovery when dealing
Davor Ocelice9ed2812017-12-25 17:49:28 +010012202 with long sessions (e.g. LDAP, SQL, ...). Doing this can cause more trouble
12203 than it tries to solve (e.g. incomplete transactions), so use this feature
Justin Karnegeseb2c24a2012-05-24 15:28:52 -070012204 with extreme care. Sessions killed because a server comes up are logged
12205 with an 'U' termination code (for "Up").
12206
12207 Actions are disabled by default
12208
Olivier Houchard006e3102018-12-10 18:30:32 +010012209pool-max-conn <max>
12210 Set the maximum number of idling connections for a server. -1 means unlimited
12211 connections, 0 means no idle connections. The default is -1. When idle
12212 connections are enabled, orphaned idle connections which do not belong to any
12213 client session anymore are moved to a dedicated pool so that they remain
12214 usable by future clients. This only applies to connections that can be shared
12215 according to the same principles as those applying to "http-reuse".
12216
Olivier Houchardb7b3faa2018-12-14 18:15:36 +010012217pool-purge-delay <delay>
12218 Sets the delay to start purging idle connections. Each <delay> interval, half
Olivier Houcharda56eebf2019-03-19 16:44:02 +010012219 of the idle connections are closed. 0 means we don't keep any idle connection.
Willy Tarreaufb553652019-06-04 14:06:31 +020012220 The default is 5s.
Olivier Houchardb7b3faa2018-12-14 18:15:36 +010012221
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010012222port <port>
Willy Tarreauc57f0e22009-05-10 13:12:33 +020012223 Using the "port" parameter, it becomes possible to use a different port to
12224 send health-checks. On some servers, it may be desirable to dedicate a port
12225 to a specific component able to perform complex tests which are more suitable
12226 to health-checks than the application. It is common to run a simple script in
12227 inetd for instance. This parameter is ignored if the "check" parameter is not
12228 set. See also the "addr" parameter.
12229
Christopher Faulet8ed0a3e2018-04-10 14:45:45 +020012230proto <name>
12231
12232 Forces the multiplexer's protocol to use for the outgoing connections to this
12233 server. It must be compatible with the mode of the backend (TCP or HTTP). It
12234 must also be usable on the backend side. The list of available protocols is
12235 reported in haproxy -vv.
12236 Idea behind this optipon is to bypass the selection of the best multiplexer's
12237 protocol for all connections established to this server.
12238
Willy Tarreauc57f0e22009-05-10 13:12:33 +020012239redir <prefix>
12240 The "redir" parameter enables the redirection mode for all GET and HEAD
12241 requests addressing this server. This means that instead of having HAProxy
12242 forward the request to the server, it will send an "HTTP 302" response with
12243 the "Location" header composed of this prefix immediately followed by the
12244 requested URI beginning at the leading '/' of the path component. That means
12245 that no trailing slash should be used after <prefix>. All invalid requests
12246 will be rejected, and all non-GET or HEAD requests will be normally served by
12247 the server. Note that since the response is completely forged, no header
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +010012248 mangling nor cookie insertion is possible in the response. However, cookies in
Davor Ocelice9ed2812017-12-25 17:49:28 +010012249 requests are still analyzed, making this solution completely usable to direct
Willy Tarreauc57f0e22009-05-10 13:12:33 +020012250 users to a remote location in case of local disaster. Main use consists in
12251 increasing bandwidth for static servers by having the clients directly
12252 connect to them. Note: never use a relative location here, it would cause a
12253 loop between the client and HAProxy!
12254
12255 Example : server srv1 192.168.1.1:80 redir http://image1.mydomain.com check
12256
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010012257rise <count>
Willy Tarreauc57f0e22009-05-10 13:12:33 +020012258 The "rise" parameter states that a server will be considered as operational
12259 after <count> consecutive successful health checks. This value defaults to 2
12260 if unspecified. See also the "check", "inter" and "fall" parameters.
12261
Baptiste Assmann8e2d9432018-06-22 15:04:43 +020012262resolve-opts <option>,<option>,...
12263 Comma separated list of options to apply to DNS resolution linked to this
12264 server.
12265
12266 Available options:
12267
12268 * allow-dup-ip
12269 By default, HAProxy prevents IP address duplication in a backend when DNS
12270 resolution at runtime is in operation.
12271 That said, for some cases, it makes sense that two servers (in the same
12272 backend, being resolved by the same FQDN) have the same IP address.
12273 For such case, simply enable this option.
12274 This is the opposite of prevent-dup-ip.
12275
12276 * prevent-dup-ip
12277 Ensure HAProxy's default behavior is enforced on a server: prevent re-using
12278 an IP address already set to a server in the same backend and sharing the
12279 same fqdn.
12280 This is the opposite of allow-dup-ip.
12281
12282 Example:
12283 backend b_myapp
12284 default-server init-addr none resolvers dns
12285 server s1 myapp.example.com:80 check resolve-opts allow-dup-ip
12286 server s2 myapp.example.com:81 check resolve-opts allow-dup-ip
12287
12288 With the option allow-dup-ip set:
12289 * if the nameserver returns a single IP address, then both servers will use
12290 it
12291 * If the nameserver returns 2 IP addresses, then each server will pick up a
12292 different address
12293
12294 Default value: not set
12295
Baptiste Assmann1fa66662015-04-14 00:28:47 +020012296resolve-prefer <family>
12297 When DNS resolution is enabled for a server and multiple IP addresses from
12298 different families are returned, HAProxy will prefer using an IP address
12299 from the family mentioned in the "resolve-prefer" parameter.
12300 Available families: "ipv4" and "ipv6"
12301
Baptiste Assmannc4aabae2015-08-04 22:43:06 +020012302 Default value: ipv6
12303
Olivier Doucetaa1ea8a2016-08-05 17:15:20 +020012304 Example:
12305
12306 server s1 app1.domain.com:80 resolvers mydns resolve-prefer ipv6
Baptiste Assmann1fa66662015-04-14 00:28:47 +020012307
Thierry Fournierac88cfe2016-02-17 22:05:30 +010012308resolve-net <network>[,<network[,...]]
John Roeslerfb2fce12019-07-10 15:45:51 -050012309 This option prioritizes the choice of an ip address matching a network. This is
Thierry Fournierac88cfe2016-02-17 22:05:30 +010012310 useful with clouds to prefer a local ip. In some cases, a cloud high
Tim Düsterhus4896c442016-11-29 02:15:19 +010012311 availability service can be announced with many ip addresses on many
Davor Ocelice9ed2812017-12-25 17:49:28 +010012312 different datacenters. The latency between datacenter is not negligible, so
12313 this patch permits to prefer a local datacenter. If no address matches the
Thierry Fournierac88cfe2016-02-17 22:05:30 +010012314 configured network, another address is selected.
12315
Olivier Doucetaa1ea8a2016-08-05 17:15:20 +020012316 Example:
12317
12318 server s1 app1.domain.com:80 resolvers mydns resolve-net 10.0.0.0/8
Thierry Fournierac88cfe2016-02-17 22:05:30 +010012319
Baptiste Assmann1fa66662015-04-14 00:28:47 +020012320resolvers <id>
12321 Points to an existing "resolvers" section to resolve current server's
12322 hostname.
12323
Olivier Doucetaa1ea8a2016-08-05 17:15:20 +020012324 Example:
12325
12326 server s1 app1.domain.com:80 check resolvers mydns
Baptiste Assmann1fa66662015-04-14 00:28:47 +020012327
Olivier Doucetaa1ea8a2016-08-05 17:15:20 +020012328 See also section 5.3
Baptiste Assmann1fa66662015-04-14 00:28:47 +020012329
Willy Tarreau5ab04ec2011-03-20 10:32:26 +010012330send-proxy
12331 The "send-proxy" parameter enforces use of the PROXY protocol over any
12332 connection established to this server. The PROXY protocol informs the other
12333 end about the layer 3/4 addresses of the incoming connection, so that it can
12334 know the client's address or the public address it accessed to, whatever the
Bertrand Jacquin93b227d2016-06-04 15:11:10 +010012335 upper layer protocol. For connections accepted by an "accept-proxy" or
12336 "accept-netscaler-cip" listener, the advertised address will be used. Only
12337 TCPv4 and TCPv6 address families are supported. Other families such as
12338 Unix sockets, will report an UNKNOWN family. Servers using this option can
12339 fully be chained to another instance of haproxy listening with an
12340 "accept-proxy" setting. This setting must not be used if the server isn't
12341 aware of the protocol. When health checks are sent to the server, the PROXY
12342 protocol is automatically used when this option is set, unless there is an
12343 explicit "port" or "addr" directive, in which case an explicit
12344 "check-send-proxy" directive would also be needed to use the PROXY protocol.
Frédéric Lécailled2376272017-03-21 18:52:12 +010012345 See also the "no-send-proxy" option of this section and "accept-proxy" and
12346 "accept-netscaler-cip" option of the "bind" keyword.
Willy Tarreau5ab04ec2011-03-20 10:32:26 +010012347
David Safb76832014-05-08 23:42:08 -040012348send-proxy-v2
12349 The "send-proxy-v2" parameter enforces use of the PROXY protocol version 2
12350 over any connection established to this server. The PROXY protocol informs
12351 the other end about the layer 3/4 addresses of the incoming connection, so
12352 that it can know the client's address or the public address it accessed to,
Emmanuel Hocdet404d9782017-10-24 10:55:14 +020012353 whatever the upper layer protocol. It also send ALPN information if an alpn
12354 have been negotiated. This setting must not be used if the server isn't aware
12355 of this version of the protocol. See also the "no-send-proxy-v2" option of
12356 this section and send-proxy" option of the "bind" keyword.
David Safb76832014-05-08 23:42:08 -040012357
Emmanuel Hocdetf643b802018-02-01 15:20:32 +010012358proxy-v2-options <option>[,<option>]*
12359 The "proxy-v2-options" parameter add option to send in PROXY protocol version
12360 2 when "send-proxy-v2" is used. Options available are "ssl" (see also
Emmanuel Hocdetfa8d0f12018-02-01 15:53:52 +010012361 send-proxy-v2-ssl), "cert-cn" (see also "send-proxy-v2-ssl-cn"), "ssl-cipher":
12362 name of the used cipher, "cert-sig": signature algorithm of the used
Emmanuel Hocdet253c3b72018-02-01 18:29:59 +010012363 certificate, "cert-key": key algorithm of the used certificate), "authority":
12364 host name value passed by the client (only sni from a tls connection is
Emmanuel Hocdet4399c752018-02-05 15:26:43 +010012365 supported), "crc32c": checksum of the proxy protocol v2 header.
Emmanuel Hocdetf643b802018-02-01 15:20:32 +010012366
David Safb76832014-05-08 23:42:08 -040012367send-proxy-v2-ssl
12368 The "send-proxy-v2-ssl" parameter enforces use of the PROXY protocol version
12369 2 over any connection established to this server. The PROXY protocol informs
12370 the other end about the layer 3/4 addresses of the incoming connection, so
12371 that it can know the client's address or the public address it accessed to,
12372 whatever the upper layer protocol. In addition, the SSL information extension
12373 of the PROXY protocol is added to the PROXY protocol header. This setting
12374 must not be used if the server isn't aware of this version of the protocol.
Frédéric Lécailled2376272017-03-21 18:52:12 +010012375 See also the "no-send-proxy-v2-ssl" option of this section and the
12376 "send-proxy-v2" option of the "bind" keyword.
David Safb76832014-05-08 23:42:08 -040012377
12378send-proxy-v2-ssl-cn
12379 The "send-proxy-v2-ssl" parameter enforces use of the PROXY protocol version
12380 2 over any connection established to this server. The PROXY protocol informs
12381 the other end about the layer 3/4 addresses of the incoming connection, so
12382 that it can know the client's address or the public address it accessed to,
12383 whatever the upper layer protocol. In addition, the SSL information extension
12384 of the PROXY protocol, along along with the Common Name from the subject of
12385 the client certificate (if any), is added to the PROXY protocol header. This
12386 setting must not be used if the server isn't aware of this version of the
Davor Ocelice9ed2812017-12-25 17:49:28 +010012387 protocol. See also the "no-send-proxy-v2-ssl-cn" option of this section and
12388 the "send-proxy-v2" option of the "bind" keyword.
David Safb76832014-05-08 23:42:08 -040012389
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010012390slowstart <start_time_in_ms>
Willy Tarreauc57f0e22009-05-10 13:12:33 +020012391 The "slowstart" parameter for a server accepts a value in milliseconds which
12392 indicates after how long a server which has just come back up will run at
12393 full speed. Just as with every other time-based parameter, it can be entered
12394 in any other explicit unit among { us, ms, s, m, h, d }. The speed grows
12395 linearly from 0 to 100% during this time. The limitation applies to two
12396 parameters :
12397
12398 - maxconn: the number of connections accepted by the server will grow from 1
12399 to 100% of the usual dynamic limit defined by (minconn,maxconn,fullconn).
12400
12401 - weight: when the backend uses a dynamic weighted algorithm, the weight
12402 grows linearly from 1 to 100%. In this case, the weight is updated at every
12403 health-check. For this reason, it is important that the "inter" parameter
12404 is smaller than the "slowstart", in order to maximize the number of steps.
12405
12406 The slowstart never applies when haproxy starts, otherwise it would cause
12407 trouble to running servers. It only applies when a server has been previously
12408 seen as failed.
12409
Willy Tarreau732eac42015-07-09 11:40:25 +020012410sni <expression>
12411 The "sni" parameter evaluates the sample fetch expression, converts it to a
12412 string and uses the result as the host name sent in the SNI TLS extension to
12413 the server. A typical use case is to send the SNI received from the client in
12414 a bridged HTTPS scenario, using the "ssl_fc_sni" sample fetch for the
Willy Tarreau2ab88672017-07-05 18:23:03 +020012415 expression, though alternatives such as req.hdr(host) can also make sense. If
12416 "verify required" is set (which is the recommended setting), the resulting
Willy Tarreauad92a9a2017-07-28 11:38:41 +020012417 name will also be matched against the server certificate's names. See the
Jérôme Magninb36a6d22018-12-09 16:03:40 +010012418 "verify" directive for more details. If you want to set a SNI for health
12419 checks, see the "check-sni" directive for more details.
Willy Tarreau732eac42015-07-09 11:40:25 +020012420
Willy Tarreauc6f4ce82009-06-10 11:09:37 +020012421source <addr>[:<pl>[-<ph>]] [usesrc { <addr2>[:<port2>] | client | clientip } ]
Willy Tarreaubce70882009-09-07 11:51:47 +020012422source <addr>[:<port>] [usesrc { <addr2>[:<port2>] | hdr_ip(<hdr>[,<occ>]) } ]
Willy Tarreauc6f4ce82009-06-10 11:09:37 +020012423source <addr>[:<pl>[-<ph>]] [interface <name>] ...
Willy Tarreauc57f0e22009-05-10 13:12:33 +020012424 The "source" parameter sets the source address which will be used when
12425 connecting to the server. It follows the exact same parameters and principle
12426 as the backend "source" keyword, except that it only applies to the server
12427 referencing it. Please consult the "source" keyword for details.
12428
Willy Tarreauc6f4ce82009-06-10 11:09:37 +020012429 Additionally, the "source" statement on a server line allows one to specify a
12430 source port range by indicating the lower and higher bounds delimited by a
12431 dash ('-'). Some operating systems might require a valid IP address when a
12432 source port range is specified. It is permitted to have the same IP/range for
12433 several servers. Doing so makes it possible to bypass the maximum of 64k
12434 total concurrent connections. The limit will then reach 64k connections per
12435 server.
12436
Lukas Tribus7d56c6d2016-09-13 09:51:15 +000012437 Since Linux 4.2/libc 2.23 IP_BIND_ADDRESS_NO_PORT is set for connections
12438 specifying the source address without port(s).
12439
Willy Tarreaua0ee1d02012-09-10 09:01:23 +020012440ssl
Willy Tarreau44f65392013-06-25 07:56:20 +020012441 This option enables SSL ciphering on outgoing connections to the server. It
12442 is critical to verify server certificates using "verify" when using SSL to
12443 connect to servers, otherwise the communication is prone to trivial man in
12444 the-middle attacks rendering SSL useless. When this option is used, health
12445 checks are automatically sent in SSL too unless there is a "port" or an
12446 "addr" directive indicating the check should be sent to a different location.
Frédéric Lécailled2376272017-03-21 18:52:12 +010012447 See the "no-ssl" to disable "ssl" option and "check-ssl" option to force
12448 SSL health checks.
Willy Tarreau763a95b2012-10-04 23:15:39 +020012449
Emmanuel Hocdete1c722b2017-03-31 15:02:54 +020012450ssl-max-ver [ SSLv3 | TLSv1.0 | TLSv1.1 | TLSv1.2 | TLSv1.3 ]
12451 This option enforces use of <version> or lower when SSL is used to communicate
12452 with the server. This option is also available on global statement
12453 "ssl-default-server-options". See also "ssl-min-ver".
12454
12455ssl-min-ver [ SSLv3 | TLSv1.0 | TLSv1.1 | TLSv1.2 | TLSv1.3 ]
12456 This option enforces use of <version> or upper when SSL is used to communicate
12457 with the server. This option is also available on global statement
12458 "ssl-default-server-options". See also "ssl-max-ver".
12459
Frédéric Lécailled2376272017-03-21 18:52:12 +010012460ssl-reuse
12461 This option may be used as "server" setting to reset any "no-ssl-reuse"
12462 setting which would have been inherited from "default-server" directive as
12463 default value.
12464 It may also be used as "default-server" setting to reset any previous
12465 "default-server" "no-ssl-reuse" setting.
12466
12467stick
12468 This option may be used as "server" setting to reset any "non-stick"
12469 setting which would have been inherited from "default-server" directive as
12470 default value.
12471 It may also be used as "default-server" setting to reset any previous
12472 "default-server" "non-stick" setting.
Willy Tarreaua0ee1d02012-09-10 09:01:23 +020012473
Alexander Liu2a54bb72019-05-22 19:44:48 +080012474socks4 <addr>:<port>
John Roeslerfb2fce12019-07-10 15:45:51 -050012475 This option enables upstream socks4 tunnel for outgoing connections to the
Alexander Liu2a54bb72019-05-22 19:44:48 +080012476 server. Using this option won't force the health check to go via socks4 by
12477 default. You will have to use the keyword "check-via-socks4" to enable it.
12478
Willy Tarreau163d4622015-10-13 16:16:41 +020012479tcp-ut <delay>
12480 Sets the TCP User Timeout for all outgoing connections to this server. This
12481 option is available on Linux since version 2.6.37. It allows haproxy to
12482 configure a timeout for sockets which contain data not receiving an
Davor Ocelice9ed2812017-12-25 17:49:28 +010012483 acknowledgment for the configured delay. This is especially useful on
Willy Tarreau163d4622015-10-13 16:16:41 +020012484 long-lived connections experiencing long idle periods such as remote
12485 terminals or database connection pools, where the client and server timeouts
12486 must remain high to allow a long period of idle, but where it is important to
12487 detect that the server has disappeared in order to release all resources
12488 associated with its connection (and the client's session). One typical use
12489 case is also to force dead server connections to die when health checks are
12490 too slow or during a soft reload since health checks are then disabled. The
12491 argument is a delay expressed in milliseconds by default. This only works for
12492 regular TCP connections, and is ignored for other protocols.
12493
Willy Tarreau034c88c2017-01-23 23:36:45 +010012494tfo
12495 This option enables using TCP fast open when connecting to servers, on
12496 systems that support it (currently only the Linux kernel >= 4.11).
12497 See the "tfo" bind option for more information about TCP fast open.
12498 Please note that when using tfo, you should also use the "conn-failure",
12499 "empty-response" and "response-timeout" keywords for "retry-on", or haproxy
Frédéric Lécaille1b9423d2019-07-04 14:19:06 +020012500 won't be able to retry the connection on failure. See also "no-tfo".
Willy Tarreau034c88c2017-01-23 23:36:45 +010012501
Willy Tarreauc57f0e22009-05-10 13:12:33 +020012502track [<proxy>/]<server>
Willy Tarreau32091232014-05-16 13:52:00 +020012503 This option enables ability to set the current state of the server by tracking
12504 another one. It is possible to track a server which itself tracks another
12505 server, provided that at the end of the chain, a server has health checks
12506 enabled. If <proxy> is omitted the current one is used. If disable-on-404 is
Willy Tarreauc57f0e22009-05-10 13:12:33 +020012507 used, it has to be enabled on both proxies.
12508
Frédéric Lécailled2376272017-03-21 18:52:12 +010012509tls-tickets
12510 This option may be used as "server" setting to reset any "no-tls-tickets"
12511 setting which would have been inherited from "default-server" directive as
12512 default value.
12513 It may also be used as "default-server" setting to reset any previous
12514 "default-server" "no-tlsv-tickets" setting.
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010012515
Emeric Brunef42d922012-10-11 16:11:36 +020012516verify [none|required]
12517 This setting is only available when support for OpenSSL was built in. If set
Emeric Brun850efd52014-01-29 12:24:34 +010012518 to 'none', server certificate is not verified. In the other case, The
Willy Tarreauad92a9a2017-07-28 11:38:41 +020012519 certificate provided by the server is verified using CAs from 'ca-file' and
12520 optional CRLs from 'crl-file' after having checked that the names provided in
Davor Ocelice9ed2812017-12-25 17:49:28 +010012521 the certificate's subject and subjectAlternateNames attributes match either
Willy Tarreauad92a9a2017-07-28 11:38:41 +020012522 the name passed using the "sni" directive, or if not provided, the static
12523 host name passed using the "verifyhost" directive. When no name is found, the
12524 certificate's names are ignored. For this reason, without SNI it's important
12525 to use "verifyhost". On verification failure the handshake is aborted. It is
12526 critically important to verify server certificates when using SSL to connect
12527 to servers, otherwise the communication is prone to trivial man-in-the-middle
12528 attacks rendering SSL totally useless. Unless "ssl_server_verify" appears in
12529 the global section, "verify" is set to "required" by default.
Emeric Brunef42d922012-10-11 16:11:36 +020012530
Evan Broderbe554312013-06-27 00:05:25 -070012531verifyhost <hostname>
12532 This setting is only available when support for OpenSSL was built in, and
Willy Tarreauad92a9a2017-07-28 11:38:41 +020012533 only takes effect if 'verify required' is also specified. This directive sets
12534 a default static hostname to check the server's certificate against when no
12535 SNI was used to connect to the server. If SNI is not used, this is the only
12536 way to enable hostname verification. This static hostname, when set, will
12537 also be used for health checks (which cannot provide an SNI value). If none
12538 of the hostnames in the certificate match the specified hostname, the
12539 handshake is aborted. The hostnames in the server-provided certificate may
12540 include wildcards. See also "verify", "sni" and "no-verifyhost" options.
Evan Broderbe554312013-06-27 00:05:25 -070012541
Krzysztof Piotr Oledzkic53601c2010-01-06 10:50:42 +010012542weight <weight>
Willy Tarreauc57f0e22009-05-10 13:12:33 +020012543 The "weight" parameter is used to adjust the server's weight relative to
12544 other servers. All servers will receive a load proportional to their weight
12545 relative to the sum of all weights, so the higher the weight, the higher the
Willy Tarreau6704d672009-06-15 10:56:05 +020012546 load. The default weight is 1, and the maximal value is 256. A value of 0
12547 means the server will not participate in load-balancing but will still accept
12548 persistent connections. If this parameter is used to distribute the load
12549 according to server's capacity, it is recommended to start with values which
12550 can both grow and shrink, for instance between 10 and 100 to leave enough
12551 room above and below for later adjustments.
Willy Tarreauc57f0e22009-05-10 13:12:33 +020012552
12553
Cyril Bonté46175dd2015-07-02 22:45:32 +0200125545.3. Server IP address resolution using DNS
12555-------------------------------------------
Baptiste Assmann1fa66662015-04-14 00:28:47 +020012556
Baptiste Assmann62b75b42015-09-09 01:11:36 +020012557HAProxy allows using a host name on the server line to retrieve its IP address
12558using name servers. By default, HAProxy resolves the name when parsing the
12559configuration file, at startup and cache the result for the process' life.
Baptiste Assmann1fa66662015-04-14 00:28:47 +020012560This is not sufficient in some cases, such as in Amazon where a server's IP
12561can change after a reboot or an ELB Virtual IP can change based on current
12562workload.
12563This chapter describes how HAProxy can be configured to process server's name
12564resolution at run time.
12565Whether run time server name resolution has been enable or not, HAProxy will
12566carry on doing the first resolution when parsing the configuration.
12567
12568
Cyril Bonté46175dd2015-07-02 22:45:32 +0200125695.3.1. Global overview
12570----------------------
Baptiste Assmann1fa66662015-04-14 00:28:47 +020012571
12572As we've seen in introduction, name resolution in HAProxy occurs at two
12573different steps of the process life:
12574
12575 1. when starting up, HAProxy parses the server line definition and matches a
12576 host name. It uses libc functions to get the host name resolved. This
12577 resolution relies on /etc/resolv.conf file.
12578
Christopher Faulet67957bd2017-09-27 11:00:59 +020012579 2. at run time, HAProxy performs periodically name resolutions for servers
12580 requiring DNS resolutions.
Baptiste Assmann1fa66662015-04-14 00:28:47 +020012581
12582A few other events can trigger a name resolution at run time:
12583 - when a server's health check ends up in a connection timeout: this may be
12584 because the server has a new IP address. So we need to trigger a name
12585 resolution to know this new IP.
12586
Christopher Faulet67957bd2017-09-27 11:00:59 +020012587When using resolvers, the server name can either be a hostname, or a SRV label.
Davor Ocelice9ed2812017-12-25 17:49:28 +010012588HAProxy considers anything that starts with an underscore as a SRV label. If a
Christopher Faulet67957bd2017-09-27 11:00:59 +020012589SRV label is specified, then the corresponding SRV records will be retrieved
12590from the DNS server, and the provided hostnames will be used. The SRV label
12591will be checked periodically, and if any server are added or removed, haproxy
12592will automatically do the same.
Olivier Houchardecfa18d2017-08-07 17:30:03 +020012593
Baptiste Assmann1fa66662015-04-14 00:28:47 +020012594A few things important to notice:
John Roeslerfb2fce12019-07-10 15:45:51 -050012595 - all the name servers are queried in the meantime. HAProxy will process the
Baptiste Assmann1fa66662015-04-14 00:28:47 +020012596 first valid response.
12597
12598 - a resolution is considered as invalid (NX, timeout, refused), when all the
12599 servers return an error.
12600
12601
Cyril Bonté46175dd2015-07-02 22:45:32 +0200126025.3.2. The resolvers section
12603----------------------------
Baptiste Assmann1fa66662015-04-14 00:28:47 +020012604
12605This section is dedicated to host information related to name resolution in
Christopher Faulet67957bd2017-09-27 11:00:59 +020012606HAProxy. There can be as many as resolvers section as needed. Each section can
12607contain many name servers.
Baptiste Assmann1fa66662015-04-14 00:28:47 +020012608
Baptiste Assmann62b75b42015-09-09 01:11:36 +020012609When multiple name servers are configured in a resolvers section, then HAProxy
12610uses the first valid response. In case of invalid responses, only the last one
12611is treated. Purpose is to give the chance to a slow server to deliver a valid
12612answer after a fast faulty or outdated server.
12613
12614When each server returns a different error type, then only the last error is
Christopher Faulet67957bd2017-09-27 11:00:59 +020012615used by HAProxy. The following processing is applied on this error:
Baptiste Assmann62b75b42015-09-09 01:11:36 +020012616
Christopher Faulet67957bd2017-09-27 11:00:59 +020012617 1. HAProxy retries the same DNS query with a new query type. The A queries are
12618 switch to AAAA or the opposite. SRV queries are not concerned here. Timeout
12619 errors are also excluded.
Baptiste Assmann62b75b42015-09-09 01:11:36 +020012620
Christopher Faulet67957bd2017-09-27 11:00:59 +020012621 2. When the fallback on the query type was done (or not applicable), HAProxy
12622 retries the original DNS query, with the preferred query type.
Baptiste Assmann62b75b42015-09-09 01:11:36 +020012623
Christopher Faulet67957bd2017-09-27 11:00:59 +020012624 3. HAProxy retries previous steps <resolve_retires> times. If no valid
12625 response is received after that, it stops the DNS resolution and reports
12626 the error.
Baptiste Assmann62b75b42015-09-09 01:11:36 +020012627
Christopher Faulet67957bd2017-09-27 11:00:59 +020012628For example, with 2 name servers configured in a resolvers section, the
12629following scenarios are possible:
12630
12631 - First response is valid and is applied directly, second response is
12632 ignored
12633
12634 - First response is invalid and second one is valid, then second response is
12635 applied
12636
12637 - First response is a NX domain and second one a truncated response, then
12638 HAProxy retries the query with a new type
12639
12640 - First response is a NX domain and second one is a timeout, then HAProxy
12641 retries the query with a new type
12642
12643 - Query timed out for both name servers, then HAProxy retries it with the
12644 same query type
Baptiste Assmann62b75b42015-09-09 01:11:36 +020012645
Olivier Houcharda8c6db82017-07-06 18:46:47 +020012646As a DNS server may not answer all the IPs in one DNS request, haproxy keeps
12647a cache of previous answers, an answer will be considered obsolete after
Christopher Faulet67957bd2017-09-27 11:00:59 +020012648<hold obsolete> seconds without the IP returned.
Olivier Houcharda8c6db82017-07-06 18:46:47 +020012649
Baptiste Assmann62b75b42015-09-09 01:11:36 +020012650
Baptiste Assmann1fa66662015-04-14 00:28:47 +020012651resolvers <resolvers id>
Davor Ocelice9ed2812017-12-25 17:49:28 +010012652 Creates a new name server list labeled <resolvers id>
Baptiste Assmann1fa66662015-04-14 00:28:47 +020012653
12654A resolvers section accept the following parameters:
12655
Baptiste Assmann2af08fe2017-08-14 00:13:01 +020012656accepted_payload_size <nb>
Davor Ocelice9ed2812017-12-25 17:49:28 +010012657 Defines the maximum payload size accepted by HAProxy and announced to all the
Christopher Faulet67957bd2017-09-27 11:00:59 +020012658 name servers configured in this resolvers section.
Baptiste Assmann2af08fe2017-08-14 00:13:01 +020012659 <nb> is in bytes. If not set, HAProxy announces 512. (minimal value defined
12660 by RFC 6891)
12661
Baptiste Assmann9d8dbbc2017-08-18 23:35:08 +020012662 Note: the maximum allowed value is 8192.
12663
Baptiste Assmann1fa66662015-04-14 00:28:47 +020012664nameserver <id> <ip>:<port>
12665 DNS server description:
12666 <id> : label of the server, should be unique
12667 <ip> : IP address of the server
12668 <port> : port where the DNS service actually runs
12669
Ben Draut44e609b2018-05-29 15:40:08 -060012670parse-resolv-conf
12671 Adds all nameservers found in /etc/resolv.conf to this resolvers nameservers
12672 list. Ordered as if each nameserver in /etc/resolv.conf was individually
12673 placed in the resolvers section in place of this directive.
12674
Baptiste Assmann1fa66662015-04-14 00:28:47 +020012675hold <status> <period>
12676 Defines <period> during which the last name resolution should be kept based
12677 on last resolution <status>
Baptiste Assmann987e16d2016-11-02 22:23:31 +010012678 <status> : last name resolution status. Acceptable values are "nx",
Olivier Houcharda8c6db82017-07-06 18:46:47 +020012679 "other", "refused", "timeout", "valid", "obsolete".
Baptiste Assmann1fa66662015-04-14 00:28:47 +020012680 <period> : interval between two successive name resolution when the last
12681 answer was in <status>. It follows the HAProxy time format.
12682 <period> is in milliseconds by default.
12683
Baptiste Assmann686408b2017-08-18 10:15:42 +020012684 Default value is 10s for "valid", 0s for "obsolete" and 30s for others.
Baptiste Assmann1fa66662015-04-14 00:28:47 +020012685
Baptiste Assmann1fa66662015-04-14 00:28:47 +020012686resolve_retries <nb>
12687 Defines the number <nb> of queries to send to resolve a server name before
12688 giving up.
12689 Default value: 3
12690
Baptiste Assmann62b75b42015-09-09 01:11:36 +020012691 A retry occurs on name server timeout or when the full sequence of DNS query
12692 type failover is over and we need to start up from the default ANY query
12693 type.
12694
Baptiste Assmann1fa66662015-04-14 00:28:47 +020012695timeout <event> <time>
12696 Defines timeouts related to name resolution
12697 <event> : the event on which the <time> timeout period applies to.
12698 events available are:
Frédéric Lécaille93d33162019-03-06 09:35:59 +010012699 - resolve : default time to trigger name resolutions when no
12700 other time applied.
Christopher Faulet67957bd2017-09-27 11:00:59 +020012701 Default value: 1s
12702 - retry : time between two DNS queries, when no valid response
Frédéric Lécaille93d33162019-03-06 09:35:59 +010012703 have been received.
Christopher Faulet67957bd2017-09-27 11:00:59 +020012704 Default value: 1s
Baptiste Assmann1fa66662015-04-14 00:28:47 +020012705 <time> : time related to the event. It follows the HAProxy time format.
12706 <time> is expressed in milliseconds.
12707
Olivier Doucetaa1ea8a2016-08-05 17:15:20 +020012708 Example:
Baptiste Assmann1fa66662015-04-14 00:28:47 +020012709
12710 resolvers mydns
12711 nameserver dns1 10.0.0.1:53
12712 nameserver dns2 10.0.0.2:53
Ben Draut44e609b2018-05-29 15:40:08 -060012713 parse-resolv-conf
Baptiste Assmann1fa66662015-04-14 00:28:47 +020012714 resolve_retries 3
Christopher Faulet67957bd2017-09-27 11:00:59 +020012715 timeout resolve 1s
Baptiste Assmann1fa66662015-04-14 00:28:47 +020012716 timeout retry 1s
Baptiste Assmann987e16d2016-11-02 22:23:31 +010012717 hold other 30s
12718 hold refused 30s
12719 hold nx 30s
12720 hold timeout 30s
Baptiste Assmann1fa66662015-04-14 00:28:47 +020012721 hold valid 10s
Olivier Houcharda8c6db82017-07-06 18:46:47 +020012722 hold obsolete 30s
Baptiste Assmann1fa66662015-04-14 00:28:47 +020012723
12724
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200127256. HTTP header manipulation
12726---------------------------
12727
12728In HTTP mode, it is possible to rewrite, add or delete some of the request and
12729response headers based on regular expressions. It is also possible to block a
12730request or a response if a particular header matches a regular expression,
12731which is enough to stop most elementary protocol attacks, and to protect
Willy Tarreau70dffda2014-01-30 03:07:23 +010012732against information leak from the internal network.
Willy Tarreauc57f0e22009-05-10 13:12:33 +020012733
Willy Tarreau70dffda2014-01-30 03:07:23 +010012734If HAProxy encounters an "Informational Response" (status code 1xx), it is able
12735to process all rsp* rules which can allow, deny, rewrite or delete a header,
12736but it will refuse to add a header to any such messages as this is not
12737HTTP-compliant. The reason for still processing headers in such responses is to
12738stop and/or fix any possible information leak which may happen, for instance
12739because another downstream equipment would unconditionally add a header, or if
12740a server name appears there. When such messages are seen, normal processing
12741still occurs on the next non-informational messages.
Willy Tarreau816b9792009-09-15 21:25:21 +020012742
Willy Tarreauc57f0e22009-05-10 13:12:33 +020012743This section covers common usage of the following keywords, described in detail
12744in section 4.2 :
12745
12746 - reqadd <string>
12747 - reqallow <search>
12748 - reqiallow <search>
12749 - reqdel <search>
12750 - reqidel <search>
12751 - reqdeny <search>
12752 - reqideny <search>
12753 - reqpass <search>
12754 - reqipass <search>
12755 - reqrep <search> <replace>
12756 - reqirep <search> <replace>
12757 - reqtarpit <search>
12758 - reqitarpit <search>
12759 - rspadd <string>
12760 - rspdel <search>
12761 - rspidel <search>
12762 - rspdeny <search>
12763 - rspideny <search>
12764 - rsprep <search> <replace>
12765 - rspirep <search> <replace>
12766
12767With all these keywords, the same conventions are used. The <search> parameter
12768is a POSIX extended regular expression (regex) which supports grouping through
12769parenthesis (without the backslash). Spaces and other delimiters must be
12770prefixed with a backslash ('\') to avoid confusion with a field delimiter.
12771Other characters may be prefixed with a backslash to change their meaning :
12772
12773 \t for a tab
12774 \r for a carriage return (CR)
12775 \n for a new line (LF)
12776 \ to mark a space and differentiate it from a delimiter
12777 \# to mark a sharp and differentiate it from a comment
12778 \\ to use a backslash in a regex
12779 \\\\ to use a backslash in the text (*2 for regex, *2 for haproxy)
12780 \xXX to write the ASCII hex code XX as in the C language
12781
12782The <replace> parameter contains the string to be used to replace the largest
12783portion of text matching the regex. It can make use of the special characters
12784above, and can reference a substring which is delimited by parenthesis in the
12785regex, by writing a backslash ('\') immediately followed by one digit from 0 to
127869 indicating the group position (0 designating the entire line). This practice
12787is very common to users of the "sed" program.
12788
12789The <string> parameter represents the string which will systematically be added
12790after the last header line. It can also use special character sequences above.
12791
12792Notes related to these keywords :
12793---------------------------------
12794 - these keywords are not always convenient to allow/deny based on header
12795 contents. It is strongly recommended to use ACLs with the "block" keyword
12796 instead, resulting in far more flexible and manageable rules.
12797
12798 - lines are always considered as a whole. It is not possible to reference
12799 a header name only or a value only. This is important because of the way
12800 headers are written (notably the number of spaces after the colon).
12801
12802 - the first line is always considered as a header, which makes it possible to
12803 rewrite or filter HTTP requests URIs or response codes, but in turn makes
12804 it harder to distinguish between headers and request line. The regex prefix
12805 ^[^\ \t]*[\ \t] matches any HTTP method followed by a space, and the prefix
12806 ^[^ \t:]*: matches any header name followed by a colon.
12807
12808 - for performances reasons, the number of characters added to a request or to
12809 a response is limited at build time to values between 1 and 4 kB. This
12810 should normally be far more than enough for most usages. If it is too short
12811 on occasional usages, it is possible to gain some space by removing some
12812 useless headers before adding new ones.
12813
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +010012814 - keywords beginning with "reqi" and "rspi" are the same as their counterpart
Willy Tarreauc57f0e22009-05-10 13:12:33 +020012815 without the 'i' letter except that they ignore case when matching patterns.
12816
12817 - when a request passes through a frontend then a backend, all req* rules
12818 from the frontend will be evaluated, then all req* rules from the backend
12819 will be evaluated. The reverse path is applied to responses.
12820
12821 - req* statements are applied after "block" statements, so that "block" is
12822 always the first one, but before "use_backend" in order to permit rewriting
Willy Tarreaud72758d2010-01-12 10:42:19 +010012823 before switching.
Willy Tarreauc57f0e22009-05-10 13:12:33 +020012824
12825
Willy Tarreau74ca5042013-06-11 23:12:07 +0200128267. Using ACLs and fetching samples
12827----------------------------------
12828
Davor Ocelice9ed2812017-12-25 17:49:28 +010012829HAProxy is capable of extracting data from request or response streams, from
Willy Tarreau74ca5042013-06-11 23:12:07 +020012830client or server information, from tables, environmental information etc...
12831The action of extracting such data is called fetching a sample. Once retrieved,
12832these samples may be used for various purposes such as a key to a stick-table,
12833but most common usages consist in matching them against predefined constant
12834data called patterns.
12835
12836
128377.1. ACL basics
12838---------------
Willy Tarreauc57f0e22009-05-10 13:12:33 +020012839
12840The use of Access Control Lists (ACL) provides a flexible solution to perform
12841content switching and generally to take decisions based on content extracted
12842from the request, the response or any environmental status. The principle is
12843simple :
12844
Willy Tarreau74ca5042013-06-11 23:12:07 +020012845 - extract a data sample from a stream, table or the environment
Willy Tarreaue6b11e42013-11-26 19:02:32 +010012846 - optionally apply some format conversion to the extracted sample
Willy Tarreau74ca5042013-06-11 23:12:07 +020012847 - apply one or multiple pattern matching methods on this sample
12848 - perform actions only when a pattern matches the sample
Willy Tarreauc57f0e22009-05-10 13:12:33 +020012849
Willy Tarreau74ca5042013-06-11 23:12:07 +020012850The actions generally consist in blocking a request, selecting a backend, or
12851adding a header.
Willy Tarreauc57f0e22009-05-10 13:12:33 +020012852
12853In order to define a test, the "acl" keyword is used. The syntax is :
12854
Willy Tarreau74ca5042013-06-11 23:12:07 +020012855 acl <aclname> <criterion> [flags] [operator] [<value>] ...
Willy Tarreauc57f0e22009-05-10 13:12:33 +020012856
12857This creates a new ACL <aclname> or completes an existing one with new tests.
12858Those tests apply to the portion of request/response specified in <criterion>
12859and may be adjusted with optional flags [flags]. Some criteria also support
Willy Tarreaue6b11e42013-11-26 19:02:32 +010012860an operator which may be specified before the set of values. Optionally some
12861conversion operators may be applied to the sample, and they will be specified
12862as a comma-delimited list of keywords just after the first keyword. The values
12863are of the type supported by the criterion, and are separated by spaces.
Willy Tarreauc57f0e22009-05-10 13:12:33 +020012864
12865ACL names must be formed from upper and lower case letters, digits, '-' (dash),
12866'_' (underscore) , '.' (dot) and ':' (colon). ACL names are case-sensitive,
12867which means that "my_acl" and "My_Acl" are two different ACLs.
12868
12869There is no enforced limit to the number of ACLs. The unused ones do not affect
12870performance, they just consume a small amount of memory.
12871
Willy Tarreau74ca5042013-06-11 23:12:07 +020012872The criterion generally is the name of a sample fetch method, or one of its ACL
12873specific declinations. The default test method is implied by the output type of
12874this sample fetch method. The ACL declinations can describe alternate matching
Willy Tarreaue6b11e42013-11-26 19:02:32 +010012875methods of a same sample fetch method. The sample fetch methods are the only
12876ones supporting a conversion.
Willy Tarreau74ca5042013-06-11 23:12:07 +020012877
12878Sample fetch methods return data which can be of the following types :
12879 - boolean
12880 - integer (signed or unsigned)
12881 - IPv4 or IPv6 address
12882 - string
12883 - data block
12884
Willy Tarreaue6b11e42013-11-26 19:02:32 +010012885Converters transform any of these data into any of these. For example, some
12886converters might convert a string to a lower-case string while other ones
12887would turn a string to an IPv4 address, or apply a netmask to an IP address.
12888The resulting sample is of the type of the last converter applied to the list,
12889which defaults to the type of the sample fetch method.
12890
Thierry FOURNIER2a06e392014-05-11 15:49:55 +020012891Each sample or converter returns data of a specific type, specified with its
12892keyword in this documentation. When an ACL is declared using a standard sample
12893fetch method, certain types automatically involved a default matching method
12894which are summarized in the table below :
12895
12896 +---------------------+-----------------+
12897 | Sample or converter | Default |
12898 | output type | matching method |
12899 +---------------------+-----------------+
12900 | boolean | bool |
12901 +---------------------+-----------------+
12902 | integer | int |
12903 +---------------------+-----------------+
12904 | ip | ip |
12905 +---------------------+-----------------+
12906 | string | str |
12907 +---------------------+-----------------+
12908 | binary | none, use "-m" |
12909 +---------------------+-----------------+
12910
12911Note that in order to match a binary samples, it is mandatory to specify a
12912matching method, see below.
12913
Willy Tarreau74ca5042013-06-11 23:12:07 +020012914The ACL engine can match these types against patterns of the following types :
12915 - boolean
12916 - integer or integer range
12917 - IP address / network
12918 - string (exact, substring, suffix, prefix, subdir, domain)
12919 - regular expression
12920 - hex block
12921
Willy Tarreauc57f0e22009-05-10 13:12:33 +020012922The following ACL flags are currently supported :
12923
Willy Tarreau2b5285d2010-05-09 23:45:24 +020012924 -i : ignore case during matching of all subsequent patterns.
12925 -f : load patterns from a file.
Willy Tarreau74ca5042013-06-11 23:12:07 +020012926 -m : use a specific pattern matching method
Thierry FOURNIERb7729c92014-02-11 16:24:41 +010012927 -n : forbid the DNS resolutions
Thierry FOURNIER9860c412014-01-29 14:23:29 +010012928 -M : load the file pointed by -f like a map file.
Thierry FOURNIER3534d882014-01-20 17:01:44 +010012929 -u : force the unique id of the ACL
Willy Tarreau6a06a402007-07-15 20:15:28 +020012930 -- : force end of flags. Useful when a string looks like one of the flags.
12931
Willy Tarreau74ca5042013-06-11 23:12:07 +020012932The "-f" flag is followed by the name of a file from which all lines will be
12933read as individual values. It is even possible to pass multiple "-f" arguments
12934if the patterns are to be loaded from multiple files. Empty lines as well as
12935lines beginning with a sharp ('#') will be ignored. All leading spaces and tabs
12936will be stripped. If it is absolutely necessary to insert a valid pattern
12937beginning with a sharp, just prefix it with a space so that it is not taken for
12938a comment. Depending on the data type and match method, haproxy may load the
12939lines into a binary tree, allowing very fast lookups. This is true for IPv4 and
12940exact string matching. In this case, duplicates will automatically be removed.
12941
Thierry FOURNIER9860c412014-01-29 14:23:29 +010012942The "-M" flag allows an ACL to use a map file. If this flag is set, the file is
12943parsed as two column file. The first column contains the patterns used by the
12944ACL, and the second column contain the samples. The sample can be used later by
12945a map. This can be useful in some rare cases where an ACL would just be used to
12946check for the existence of a pattern in a map before a mapping is applied.
12947
Thierry FOURNIER3534d882014-01-20 17:01:44 +010012948The "-u" flag forces the unique id of the ACL. This unique id is used with the
12949socket interface to identify ACL and dynamically change its values. Note that a
12950file is always identified by its name even if an id is set.
12951
Willy Tarreau74ca5042013-06-11 23:12:07 +020012952Also, note that the "-i" flag applies to subsequent entries and not to entries
12953loaded from files preceding it. For instance :
12954
12955 acl valid-ua hdr(user-agent) -f exact-ua.lst -i -f generic-ua.lst test
12956
12957In this example, each line of "exact-ua.lst" will be exactly matched against
12958the "user-agent" header of the request. Then each line of "generic-ua" will be
12959case-insensitively matched. Then the word "test" will be insensitively matched
12960as well.
12961
12962The "-m" flag is used to select a specific pattern matching method on the input
12963sample. All ACL-specific criteria imply a pattern matching method and generally
12964do not need this flag. However, this flag is useful with generic sample fetch
12965methods to describe how they're going to be matched against the patterns. This
12966is required for sample fetches which return data type for which there is no
Davor Ocelice9ed2812017-12-25 17:49:28 +010012967obvious matching method (e.g. string or binary). When "-m" is specified and
Willy Tarreau74ca5042013-06-11 23:12:07 +020012968followed by a pattern matching method name, this method is used instead of the
12969default one for the criterion. This makes it possible to match contents in ways
12970that were not initially planned, or with sample fetch methods which return a
12971string. The matching method also affects the way the patterns are parsed.
12972
Thierry FOURNIERb7729c92014-02-11 16:24:41 +010012973The "-n" flag forbids the dns resolutions. It is used with the load of ip files.
12974By default, if the parser cannot parse ip address it considers that the parsed
12975string is maybe a domain name and try dns resolution. The flag "-n" disable this
12976resolution. It is useful for detecting malformed ip lists. Note that if the DNS
12977server is not reachable, the haproxy configuration parsing may last many minutes
John Roeslerfb2fce12019-07-10 15:45:51 -050012978waiting for the timeout. During this time no error messages are displayed. The
Thierry FOURNIERb7729c92014-02-11 16:24:41 +010012979flag "-n" disable this behavior. Note also that during the runtime, this
12980function is disabled for the dynamic acl modifications.
12981
Willy Tarreau74ca5042013-06-11 23:12:07 +020012982There are some restrictions however. Not all methods can be used with all
12983sample fetch methods. Also, if "-m" is used in conjunction with "-f", it must
12984be placed first. The pattern matching method must be one of the following :
Willy Tarreau5adeda12013-03-31 22:13:34 +020012985
12986 - "found" : only check if the requested sample could be found in the stream,
12987 but do not compare it against any pattern. It is recommended not
Willy Tarreau74ca5042013-06-11 23:12:07 +020012988 to pass any pattern to avoid confusion. This matching method is
12989 particularly useful to detect presence of certain contents such
12990 as headers, cookies, etc... even if they are empty and without
12991 comparing them to anything nor counting them.
Willy Tarreau5adeda12013-03-31 22:13:34 +020012992
12993 - "bool" : check the value as a boolean. It can only be applied to fetches
12994 which return a boolean or integer value, and takes no pattern.
Willy Tarreau74ca5042013-06-11 23:12:07 +020012995 Value zero or false does not match, all other values do match.
Willy Tarreau5adeda12013-03-31 22:13:34 +020012996
12997 - "int" : match the value as an integer. It can be used with integer and
Willy Tarreau74ca5042013-06-11 23:12:07 +020012998 boolean samples. Boolean false is integer 0, true is integer 1.
Willy Tarreau5adeda12013-03-31 22:13:34 +020012999
13000 - "ip" : match the value as an IPv4 or IPv6 address. It is compatible
Willy Tarreau74ca5042013-06-11 23:12:07 +020013001 with IP address samples only, so it is implied and never needed.
Willy Tarreau5adeda12013-03-31 22:13:34 +020013002
Davor Ocelice9ed2812017-12-25 17:49:28 +010013003 - "bin" : match the contents against a hexadecimal string representing a
Willy Tarreau5adeda12013-03-31 22:13:34 +020013004 binary sequence. This may be used with binary or string samples.
13005
13006 - "len" : match the sample's length as an integer. This may be used with
13007 binary or string samples.
13008
Willy Tarreau74ca5042013-06-11 23:12:07 +020013009 - "str" : exact match : match the contents against a string. This may be
13010 used with binary or string samples.
Willy Tarreau5adeda12013-03-31 22:13:34 +020013011
Willy Tarreau74ca5042013-06-11 23:12:07 +020013012 - "sub" : substring match : check that the contents contain at least one of
13013 the provided string patterns. This may be used with binary or
13014 string samples.
Willy Tarreau5adeda12013-03-31 22:13:34 +020013015
Willy Tarreau74ca5042013-06-11 23:12:07 +020013016 - "reg" : regex match : match the contents against a list of regular
13017 expressions. This may be used with binary or string samples.
Willy Tarreau5adeda12013-03-31 22:13:34 +020013018
Willy Tarreau74ca5042013-06-11 23:12:07 +020013019 - "beg" : prefix match : check that the contents begin like the provided
13020 string patterns. This may be used with binary or string samples.
Willy Tarreau5adeda12013-03-31 22:13:34 +020013021
Willy Tarreau74ca5042013-06-11 23:12:07 +020013022 - "end" : suffix match : check that the contents end like the provided
13023 string patterns. This may be used with binary or string samples.
Willy Tarreau5adeda12013-03-31 22:13:34 +020013024
Willy Tarreau74ca5042013-06-11 23:12:07 +020013025 - "dir" : subdir match : check that a slash-delimited portion of the
13026 contents exactly matches one of the provided string patterns.
Willy Tarreau5adeda12013-03-31 22:13:34 +020013027 This may be used with binary or string samples.
13028
Willy Tarreau74ca5042013-06-11 23:12:07 +020013029 - "dom" : domain match : check that a dot-delimited portion of the contents
13030 exactly match one of the provided string patterns. This may be
13031 used with binary or string samples.
Willy Tarreau5adeda12013-03-31 22:13:34 +020013032
13033For example, to quickly detect the presence of cookie "JSESSIONID" in an HTTP
13034request, it is possible to do :
13035
13036 acl jsess_present cook(JSESSIONID) -m found
13037
13038In order to apply a regular expression on the 500 first bytes of data in the
13039buffer, one would use the following acl :
13040
13041 acl script_tag payload(0,500) -m reg -i <script>
13042
Willy Tarreaue6b11e42013-11-26 19:02:32 +010013043On systems where the regex library is much slower when using "-i", it is
13044possible to convert the sample to lowercase before matching, like this :
13045
13046 acl script_tag payload(0,500),lower -m reg <script>
13047
Willy Tarreau74ca5042013-06-11 23:12:07 +020013048All ACL-specific criteria imply a default matching method. Most often, these
13049criteria are composed by concatenating the name of the original sample fetch
13050method and the matching method. For example, "hdr_beg" applies the "beg" match
13051to samples retrieved using the "hdr" fetch method. Since all ACL-specific
13052criteria rely on a sample fetch method, it is always possible instead to use
13053the original sample fetch method and the explicit matching method using "-m".
Willy Tarreau2b5285d2010-05-09 23:45:24 +020013054
Willy Tarreau74ca5042013-06-11 23:12:07 +020013055If an alternate match is specified using "-m" on an ACL-specific criterion,
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030013056the matching method is simply applied to the underlying sample fetch method.
13057For example, all ACLs below are exact equivalent :
Willy Tarreau2b5285d2010-05-09 23:45:24 +020013058
Willy Tarreau74ca5042013-06-11 23:12:07 +020013059 acl short_form hdr_beg(host) www.
13060 acl alternate1 hdr_beg(host) -m beg www.
13061 acl alternate2 hdr_dom(host) -m beg www.
13062 acl alternate3 hdr(host) -m beg www.
Willy Tarreau2b5285d2010-05-09 23:45:24 +020013063
Willy Tarreau2b5285d2010-05-09 23:45:24 +020013064
Thierry FOURNIER2a06e392014-05-11 15:49:55 +020013065The table below summarizes the compatibility matrix between sample or converter
13066types and the pattern types to fetch against. It indicates for each compatible
13067combination the name of the matching method to be used, surrounded with angle
13068brackets ">" and "<" when the method is the default one and will work by
13069default without "-m".
Willy Tarreau0ba27502007-12-24 16:55:16 +010013070
Willy Tarreau74ca5042013-06-11 23:12:07 +020013071 +-------------------------------------------------+
13072 | Input sample type |
13073 +----------------------+---------+---------+---------+---------+---------+
Thierry FOURNIER2a06e392014-05-11 15:49:55 +020013074 | pattern type | boolean | integer | ip | string | binary |
Willy Tarreau74ca5042013-06-11 23:12:07 +020013075 +----------------------+---------+---------+---------+---------+---------+
13076 | none (presence only) | found | found | found | found | found |
13077 +----------------------+---------+---------+---------+---------+---------+
Thierry FOURNIER2a06e392014-05-11 15:49:55 +020013078 | none (boolean value) |> bool <| bool | | bool | |
Willy Tarreau74ca5042013-06-11 23:12:07 +020013079 +----------------------+---------+---------+---------+---------+---------+
Thierry FOURNIER2a06e392014-05-11 15:49:55 +020013080 | integer (value) | int |> int <| int | int | |
Willy Tarreau74ca5042013-06-11 23:12:07 +020013081 +----------------------+---------+---------+---------+---------+---------+
Thierry FOURNIERe3ded592013-12-06 15:36:54 +010013082 | integer (length) | len | len | len | len | len |
Willy Tarreau74ca5042013-06-11 23:12:07 +020013083 +----------------------+---------+---------+---------+---------+---------+
Thierry FOURNIER2a06e392014-05-11 15:49:55 +020013084 | IP address | | |> ip <| ip | ip |
Willy Tarreau74ca5042013-06-11 23:12:07 +020013085 +----------------------+---------+---------+---------+---------+---------+
Thierry FOURNIER2a06e392014-05-11 15:49:55 +020013086 | exact string | str | str | str |> str <| str |
Willy Tarreau74ca5042013-06-11 23:12:07 +020013087 +----------------------+---------+---------+---------+---------+---------+
Thierry FOURNIERe3ded592013-12-06 15:36:54 +010013088 | prefix | beg | beg | beg | beg | beg |
Willy Tarreau74ca5042013-06-11 23:12:07 +020013089 +----------------------+---------+---------+---------+---------+---------+
Thierry FOURNIERe3ded592013-12-06 15:36:54 +010013090 | suffix | end | end | end | end | end |
Willy Tarreau74ca5042013-06-11 23:12:07 +020013091 +----------------------+---------+---------+---------+---------+---------+
Thierry FOURNIERe3ded592013-12-06 15:36:54 +010013092 | substring | sub | sub | sub | sub | sub |
Willy Tarreau74ca5042013-06-11 23:12:07 +020013093 +----------------------+---------+---------+---------+---------+---------+
Thierry FOURNIERe3ded592013-12-06 15:36:54 +010013094 | subdir | dir | dir | dir | dir | dir |
Willy Tarreau74ca5042013-06-11 23:12:07 +020013095 +----------------------+---------+---------+---------+---------+---------+
Thierry FOURNIERe3ded592013-12-06 15:36:54 +010013096 | domain | dom | dom | dom | dom | dom |
Willy Tarreau74ca5042013-06-11 23:12:07 +020013097 +----------------------+---------+---------+---------+---------+---------+
Thierry FOURNIERe3ded592013-12-06 15:36:54 +010013098 | regex | reg | reg | reg | reg | reg |
Willy Tarreau74ca5042013-06-11 23:12:07 +020013099 +----------------------+---------+---------+---------+---------+---------+
13100 | hex block | | | | bin | bin |
13101 +----------------------+---------+---------+---------+---------+---------+
Willy Tarreau6a06a402007-07-15 20:15:28 +020013102
13103
Willy Tarreau74ca5042013-06-11 23:12:07 +0200131047.1.1. Matching booleans
13105------------------------
13106
13107In order to match a boolean, no value is needed and all values are ignored.
13108Boolean matching is used by default for all fetch methods of type "boolean".
13109When boolean matching is used, the fetched value is returned as-is, which means
13110that a boolean "true" will always match and a boolean "false" will never match.
13111
13112Boolean matching may also be enforced using "-m bool" on fetch methods which
13113return an integer value. Then, integer value 0 is converted to the boolean
13114"false" and all other values are converted to "true".
13115
Willy Tarreau6a06a402007-07-15 20:15:28 +020013116
Willy Tarreau74ca5042013-06-11 23:12:07 +0200131177.1.2. Matching integers
13118------------------------
13119
13120Integer matching applies by default to integer fetch methods. It can also be
13121enforced on boolean fetches using "-m int". In this case, "false" is converted
13122to the integer 0, and "true" is converted to the integer 1.
13123
13124Integer matching also supports integer ranges and operators. Note that integer
13125matching only applies to positive values. A range is a value expressed with a
13126lower and an upper bound separated with a colon, both of which may be omitted.
Willy Tarreau6a06a402007-07-15 20:15:28 +020013127
13128For instance, "1024:65535" is a valid range to represent a range of
13129unprivileged ports, and "1024:" would also work. "0:1023" is a valid
13130representation of privileged ports, and ":1023" would also work.
13131
Willy Tarreau62644772008-07-16 18:36:06 +020013132As a special case, some ACL functions support decimal numbers which are in fact
13133two integers separated by a dot. This is used with some version checks for
13134instance. All integer properties apply to those decimal numbers, including
13135ranges and operators.
13136
Willy Tarreau6a06a402007-07-15 20:15:28 +020013137For an easier usage, comparison operators are also supported. Note that using
Willy Tarreau0ba27502007-12-24 16:55:16 +010013138operators with ranges does not make much sense and is strongly discouraged.
13139Similarly, it does not make much sense to perform order comparisons with a set
13140of values.
Willy Tarreau6a06a402007-07-15 20:15:28 +020013141
Willy Tarreau0ba27502007-12-24 16:55:16 +010013142Available operators for integer matching are :
Willy Tarreau6a06a402007-07-15 20:15:28 +020013143
13144 eq : true if the tested value equals at least one value
13145 ge : true if the tested value is greater than or equal to at least one value
13146 gt : true if the tested value is greater than at least one value
13147 le : true if the tested value is less than or equal to at least one value
13148 lt : true if the tested value is less than at least one value
13149
Willy Tarreau0ba27502007-12-24 16:55:16 +010013150For instance, the following ACL matches any negative Content-Length header :
Willy Tarreau6a06a402007-07-15 20:15:28 +020013151
13152 acl negative-length hdr_val(content-length) lt 0
13153
Willy Tarreau62644772008-07-16 18:36:06 +020013154This one matches SSL versions between 3.0 and 3.1 (inclusive) :
13155
13156 acl sslv3 req_ssl_ver 3:3.1
13157
Willy Tarreau6a06a402007-07-15 20:15:28 +020013158
Willy Tarreau74ca5042013-06-11 23:12:07 +0200131597.1.3. Matching strings
13160-----------------------
13161
13162String matching applies to string or binary fetch methods, and exists in 6
13163different forms :
13164
13165 - exact match (-m str) : the extracted string must exactly match the
Davor Ocelice9ed2812017-12-25 17:49:28 +010013166 patterns;
Willy Tarreau74ca5042013-06-11 23:12:07 +020013167
13168 - substring match (-m sub) : the patterns are looked up inside the
Davor Ocelice9ed2812017-12-25 17:49:28 +010013169 extracted string, and the ACL matches if any of them is found inside;
Willy Tarreau74ca5042013-06-11 23:12:07 +020013170
13171 - prefix match (-m beg) : the patterns are compared with the beginning of
13172 the extracted string, and the ACL matches if any of them matches.
13173
13174 - suffix match (-m end) : the patterns are compared with the end of the
13175 extracted string, and the ACL matches if any of them matches.
13176
Baptiste Assmann33db6002016-03-06 23:32:10 +010013177 - subdir match (-m dir) : the patterns are looked up inside the extracted
Willy Tarreau74ca5042013-06-11 23:12:07 +020013178 string, delimited with slashes ("/"), and the ACL matches if any of them
13179 matches.
13180
13181 - domain match (-m dom) : the patterns are looked up inside the extracted
13182 string, delimited with dots ("."), and the ACL matches if any of them
13183 matches.
Willy Tarreau6a06a402007-07-15 20:15:28 +020013184
13185String matching applies to verbatim strings as they are passed, with the
13186exception of the backslash ("\") which makes it possible to escape some
13187characters such as the space. If the "-i" flag is passed before the first
13188string, then the matching will be performed ignoring the case. In order
13189to match the string "-i", either set it second, or pass the "--" flag
Willy Tarreau0ba27502007-12-24 16:55:16 +010013190before the first string. Same applies of course to match the string "--".
Willy Tarreau6a06a402007-07-15 20:15:28 +020013191
13192
Willy Tarreau74ca5042013-06-11 23:12:07 +0200131937.1.4. Matching regular expressions (regexes)
13194---------------------------------------------
Willy Tarreau6a06a402007-07-15 20:15:28 +020013195
13196Just like with string matching, regex matching applies to verbatim strings as
13197they are passed, with the exception of the backslash ("\") which makes it
13198possible to escape some characters such as the space. If the "-i" flag is
13199passed before the first regex, then the matching will be performed ignoring
13200the case. In order to match the string "-i", either set it second, or pass
Willy Tarreau0ba27502007-12-24 16:55:16 +010013201the "--" flag before the first string. Same principle applies of course to
13202match the string "--".
Willy Tarreau6a06a402007-07-15 20:15:28 +020013203
13204
Willy Tarreau74ca5042013-06-11 23:12:07 +0200132057.1.5. Matching arbitrary data blocks
13206-------------------------------------
13207
13208It is possible to match some extracted samples against a binary block which may
13209not safely be represented as a string. For this, the patterns must be passed as
13210a series of hexadecimal digits in an even number, when the match method is set
13211to binary. Each sequence of two digits will represent a byte. The hexadecimal
13212digits may be used upper or lower case.
13213
13214Example :
13215 # match "Hello\n" in the input stream (\x48 \x65 \x6c \x6c \x6f \x0a)
13216 acl hello payload(0,6) -m bin 48656c6c6f0a
13217
13218
132197.1.6. Matching IPv4 and IPv6 addresses
13220---------------------------------------
Willy Tarreau6a06a402007-07-15 20:15:28 +020013221
13222IPv4 addresses values can be specified either as plain addresses or with a
13223netmask appended, in which case the IPv4 address matches whenever it is
13224within the network. Plain addresses may also be replaced with a resolvable
Willy Tarreaud2a4aa22008-01-31 15:28:22 +010013225host name, but this practice is generally discouraged as it makes it more
Willy Tarreau0ba27502007-12-24 16:55:16 +010013226difficult to read and debug configurations. If hostnames are used, you should
13227at least ensure that they are present in /etc/hosts so that the configuration
13228does not depend on any random DNS match at the moment the configuration is
13229parsed.
Willy Tarreau6a06a402007-07-15 20:15:28 +020013230
Daniel Schnellereba56342016-04-13 00:26:52 +020013231The dotted IPv4 address notation is supported in both regular as well as the
13232abbreviated form with all-0-octets omitted:
13233
13234 +------------------+------------------+------------------+
13235 | Example 1 | Example 2 | Example 3 |
13236 +------------------+------------------+------------------+
13237 | 192.168.0.1 | 10.0.0.12 | 127.0.0.1 |
13238 | 192.168.1 | 10.12 | 127.1 |
13239 | 192.168.0.1/22 | 10.0.0.12/8 | 127.0.0.1/8 |
13240 | 192.168.1/22 | 10.12/8 | 127.1/8 |
13241 +------------------+------------------+------------------+
13242
13243Notice that this is different from RFC 4632 CIDR address notation in which
13244192.168.42/24 would be equivalent to 192.168.42.0/24.
13245
Willy Tarreauceb4ac92012-04-28 00:41:46 +020013246IPv6 may be entered in their usual form, with or without a netmask appended.
13247Only bit counts are accepted for IPv6 netmasks. In order to avoid any risk of
13248trouble with randomly resolved IP addresses, host names are never allowed in
13249IPv6 patterns.
13250
13251HAProxy is also able to match IPv4 addresses with IPv6 addresses in the
13252following situations :
13253 - tested address is IPv4, pattern address is IPv4, the match applies
13254 in IPv4 using the supplied mask if any.
13255 - tested address is IPv6, pattern address is IPv6, the match applies
13256 in IPv6 using the supplied mask if any.
13257 - tested address is IPv6, pattern address is IPv4, the match applies in IPv4
13258 using the pattern's mask if the IPv6 address matches with 2002:IPV4::,
13259 ::IPV4 or ::ffff:IPV4, otherwise it fails.
13260 - tested address is IPv4, pattern address is IPv6, the IPv4 address is first
13261 converted to IPv6 by prefixing ::ffff: in front of it, then the match is
13262 applied in IPv6 using the supplied IPv6 mask.
13263
Willy Tarreau74ca5042013-06-11 23:12:07 +020013264
132657.2. Using ACLs to form conditions
13266----------------------------------
13267
13268Some actions are only performed upon a valid condition. A condition is a
13269combination of ACLs with operators. 3 operators are supported :
13270
13271 - AND (implicit)
13272 - OR (explicit with the "or" keyword or the "||" operator)
13273 - Negation with the exclamation mark ("!")
Willy Tarreau6a06a402007-07-15 20:15:28 +020013274
Willy Tarreau74ca5042013-06-11 23:12:07 +020013275A condition is formed as a disjunctive form:
Willy Tarreau6a06a402007-07-15 20:15:28 +020013276
Willy Tarreau74ca5042013-06-11 23:12:07 +020013277 [!]acl1 [!]acl2 ... [!]acln { or [!]acl1 [!]acl2 ... [!]acln } ...
Willy Tarreaubef91e72013-03-31 23:14:46 +020013278
Willy Tarreau74ca5042013-06-11 23:12:07 +020013279Such conditions are generally used after an "if" or "unless" statement,
13280indicating when the condition will trigger the action.
Willy Tarreaubef91e72013-03-31 23:14:46 +020013281
Willy Tarreau74ca5042013-06-11 23:12:07 +020013282For instance, to block HTTP requests to the "*" URL with methods other than
13283"OPTIONS", as well as POST requests without content-length, and GET or HEAD
13284requests with a content-length greater than 0, and finally every request which
13285is not either GET/HEAD/POST/OPTIONS !
13286
13287 acl missing_cl hdr_cnt(Content-length) eq 0
Jarno Huuskonen84c51ec2017-04-03 14:20:34 +030013288 http-request deny if HTTP_URL_STAR !METH_OPTIONS || METH_POST missing_cl
13289 http-request deny if METH_GET HTTP_CONTENT
13290 http-request deny unless METH_GET or METH_POST or METH_OPTIONS
Willy Tarreau74ca5042013-06-11 23:12:07 +020013291
13292To select a different backend for requests to static contents on the "www" site
13293and to every request on the "img", "video", "download" and "ftp" hosts :
13294
13295 acl url_static path_beg /static /images /img /css
13296 acl url_static path_end .gif .png .jpg .css .js
13297 acl host_www hdr_beg(host) -i www
13298 acl host_static hdr_beg(host) -i img. video. download. ftp.
13299
Davor Ocelice9ed2812017-12-25 17:49:28 +010013300 # now use backend "static" for all static-only hosts, and for static URLs
Willy Tarreau74ca5042013-06-11 23:12:07 +020013301 # of host "www". Use backend "www" for the rest.
13302 use_backend static if host_static or host_www url_static
13303 use_backend www if host_www
13304
13305It is also possible to form rules using "anonymous ACLs". Those are unnamed ACL
13306expressions that are built on the fly without needing to be declared. They must
13307be enclosed between braces, with a space before and after each brace (because
13308the braces must be seen as independent words). Example :
13309
13310 The following rule :
13311
13312 acl missing_cl hdr_cnt(Content-length) eq 0
Jarno Huuskonen84c51ec2017-04-03 14:20:34 +030013313 http-request deny if METH_POST missing_cl
Willy Tarreau74ca5042013-06-11 23:12:07 +020013314
13315 Can also be written that way :
13316
Jarno Huuskonen84c51ec2017-04-03 14:20:34 +030013317 http-request deny if METH_POST { hdr_cnt(Content-length) eq 0 }
Willy Tarreau74ca5042013-06-11 23:12:07 +020013318
13319It is generally not recommended to use this construct because it's a lot easier
13320to leave errors in the configuration when written that way. However, for very
13321simple rules matching only one source IP address for instance, it can make more
13322sense to use them than to declare ACLs with random names. Another example of
13323good use is the following :
13324
13325 With named ACLs :
13326
13327 acl site_dead nbsrv(dynamic) lt 2
13328 acl site_dead nbsrv(static) lt 2
13329 monitor fail if site_dead
13330
13331 With anonymous ACLs :
13332
13333 monitor fail if { nbsrv(dynamic) lt 2 } || { nbsrv(static) lt 2 }
13334
Jarno Huuskonen84c51ec2017-04-03 14:20:34 +030013335See section 4.2 for detailed help on the "http-request deny" and "use_backend"
13336keywords.
Willy Tarreau74ca5042013-06-11 23:12:07 +020013337
13338
133397.3. Fetching samples
13340---------------------
13341
13342Historically, sample fetch methods were only used to retrieve data to match
13343against patterns using ACLs. With the arrival of stick-tables, a new class of
13344sample fetch methods was created, most often sharing the same syntax as their
13345ACL counterpart. These sample fetch methods are also known as "fetches". As
13346of now, ACLs and fetches have converged. All ACL fetch methods have been made
13347available as fetch methods, and ACLs may use any sample fetch method as well.
13348
13349This section details all available sample fetch methods and their output type.
13350Some sample fetch methods have deprecated aliases that are used to maintain
13351compatibility with existing configurations. They are then explicitly marked as
13352deprecated and should not be used in new setups.
13353
13354The ACL derivatives are also indicated when available, with their respective
13355matching methods. These ones all have a well defined default pattern matching
13356method, so it is never necessary (though allowed) to pass the "-m" option to
13357indicate how the sample will be matched using ACLs.
13358
13359As indicated in the sample type versus matching compatibility matrix above,
13360when using a generic sample fetch method in an ACL, the "-m" option is
13361mandatory unless the sample type is one of boolean, integer, IPv4 or IPv6. When
13362the same keyword exists as an ACL keyword and as a standard fetch method, the
13363ACL engine will automatically pick the ACL-only one by default.
13364
13365Some of these keywords support one or multiple mandatory arguments, and one or
13366multiple optional arguments. These arguments are strongly typed and are checked
13367when the configuration is parsed so that there is no risk of running with an
Davor Ocelice9ed2812017-12-25 17:49:28 +010013368incorrect argument (e.g. an unresolved backend name). Fetch function arguments
13369are passed between parenthesis and are delimited by commas. When an argument
Willy Tarreau74ca5042013-06-11 23:12:07 +020013370is optional, it will be indicated below between square brackets ('[ ]'). When
13371all arguments are optional, the parenthesis may be omitted.
13372
13373Thus, the syntax of a standard sample fetch method is one of the following :
13374 - name
13375 - name(arg1)
13376 - name(arg1,arg2)
13377
Thierry FOURNIER060762e2014-04-23 13:29:15 +020013378
133797.3.1. Converters
13380-----------------
13381
Willy Tarreaue6b11e42013-11-26 19:02:32 +010013382Sample fetch methods may be combined with transformations to be applied on top
13383of the fetched sample (also called "converters"). These combinations form what
13384is called "sample expressions" and the result is a "sample". Initially this
13385was only supported by "stick on" and "stick store-request" directives but this
Davor Ocelice9ed2812017-12-25 17:49:28 +010013386has now be extended to all places where samples may be used (ACLs, log-format,
Willy Tarreaue6b11e42013-11-26 19:02:32 +010013387unique-id-format, add-header, ...).
13388
13389These transformations are enumerated as a series of specific keywords after the
13390sample fetch method. These keywords may equally be appended immediately after
13391the fetch keyword's argument, delimited by a comma. These keywords can also
Davor Ocelice9ed2812017-12-25 17:49:28 +010013392support some arguments (e.g. a netmask) which must be passed in parenthesis.
Willy Tarreau0ba27502007-12-24 16:55:16 +010013393
Willy Tarreau97707872015-01-27 15:12:13 +010013394A certain category of converters are bitwise and arithmetic operators which
13395support performing basic operations on integers. Some bitwise operations are
13396supported (and, or, xor, cpl) and some arithmetic operations are supported
13397(add, sub, mul, div, mod, neg). Some comparators are provided (odd, even, not,
13398bool) which make it possible to report a match without having to write an ACL.
13399
Willy Tarreau74ca5042013-06-11 23:12:07 +020013400The currently available list of transformation keywords include :
Willy Tarreau0ba27502007-12-24 16:55:16 +010013401
Ben Shillitof25e8e52016-12-02 14:25:37 +00001340251d.single(<prop>[,<prop>*])
13403 Returns values for the properties requested as a string, where values are
13404 separated by the delimiter specified with "51degrees-property-separator".
13405 The device is identified using the User-Agent header passed to the
13406 converter. The function can be passed up to five property names, and if a
13407 property name can't be found, the value "NoData" is returned.
13408
13409 Example :
Davor Ocelice9ed2812017-12-25 17:49:28 +010013410 # Here the header "X-51D-DeviceTypeMobileTablet" is added to the request,
13411 # containing values for the three properties requested by using the
Ben Shillitof25e8e52016-12-02 14:25:37 +000013412 # User-Agent passed to the converter.
13413 frontend http-in
13414 bind *:8081
13415 default_backend servers
13416 http-request set-header X-51D-DeviceTypeMobileTablet \
13417 %[req.fhdr(User-Agent),51d.single(DeviceType,IsMobile,IsTablet)]
13418
Willy Tarreau97707872015-01-27 15:12:13 +010013419add(<value>)
Thierry FOURNIER07ee64e2015-07-06 23:43:03 +020013420 Adds <value> to the input value of type signed integer, and returns the
Thierry FOURNIER5d86fae2015-07-07 21:10:16 +020013421 result as a signed integer. <value> can be a numeric value or a variable
Daniel Schneller0b547052016-03-21 20:46:57 +010013422 name. The name of the variable starts with an indication about its scope. The
13423 scopes allowed are:
Christopher Fauletff2613e2016-11-09 11:36:17 +010013424 "proc" : the variable is shared with the whole process
Daniel Schneller0b547052016-03-21 20:46:57 +010013425 "sess" : the variable is shared with the whole session
13426 "txn" : the variable is shared with the transaction (request and response)
13427 "req" : the variable is shared only during request processing
13428 "res" : the variable is shared only during response processing
Davor Ocelice9ed2812017-12-25 17:49:28 +010013429 This prefix is followed by a name. The separator is a '.'. The name may only
Christopher Fauletb71557a2016-10-31 10:49:03 +010013430 contain characters 'a-z', 'A-Z', '0-9', '.' and '_'.
Willy Tarreau97707872015-01-27 15:12:13 +010013431
Nenad Merdanovicc31499d2019-03-23 11:00:32 +010013432aes_gcm_dec(<bits>,<nonce>,<key>,<aead_tag>)
13433 Decrypts the raw byte input using the AES128-GCM, AES192-GCM or
13434 AES256-GCM algorithm, depending on the <bits> parameter. All other parameters
13435 need to be base64 encoded and the returned result is in raw byte format.
13436 If the <aead_tag> validation fails, the converter doesn't return any data.
13437 The <nonce>, <key> and <aead_tag> can either be strings or variables. This
13438 converter requires at least OpenSSL 1.0.1.
13439
13440 Example:
13441 http-response set-header X-Decrypted-Text %[var(txn.enc),\
13442 aes_gcm_dec(128,txn.nonce,Zm9vb2Zvb29mb29wZm9vbw==,txn.aead_tag)]
13443
Willy Tarreau97707872015-01-27 15:12:13 +010013444and(<value>)
Thierry FOURNIER07ee64e2015-07-06 23:43:03 +020013445 Performs a bitwise "AND" between <value> and the input value of type signed
Thierry FOURNIER5d86fae2015-07-07 21:10:16 +020013446 integer, and returns the result as an signed integer. <value> can be a
Daniel Schneller0b547052016-03-21 20:46:57 +010013447 numeric value or a variable name. The name of the variable starts with an
13448 indication about its scope. The scopes allowed are:
Christopher Fauletff2613e2016-11-09 11:36:17 +010013449 "proc" : the variable is shared with the whole process
Daniel Schneller0b547052016-03-21 20:46:57 +010013450 "sess" : the variable is shared with the whole session
13451 "txn" : the variable is shared with the transaction (request and response)
13452 "req" : the variable is shared only during request processing
13453 "res" : the variable is shared only during response processing
Davor Ocelice9ed2812017-12-25 17:49:28 +010013454 This prefix is followed by a name. The separator is a '.'. The name may only
Christopher Fauletb71557a2016-10-31 10:49:03 +010013455 contain characters 'a-z', 'A-Z', '0-9', '.' and '_'.
Willy Tarreau97707872015-01-27 15:12:13 +010013456
Holger Just1bfc24b2017-05-06 00:56:53 +020013457b64dec
13458 Converts (decodes) a base64 encoded input string to its binary
13459 representation. It performs the inverse operation of base64().
13460
Emeric Brun53d1a982014-04-30 18:21:37 +020013461base64
13462 Converts a binary input sample to a base64 string. It is used to log or
Davor Ocelice9ed2812017-12-25 17:49:28 +010013463 transfer binary content in a way that can be reliably transferred (e.g.
Emeric Brun53d1a982014-04-30 18:21:37 +020013464 an SSL ID can be copied in a header).
13465
Willy Tarreau97707872015-01-27 15:12:13 +010013466bool
Thierry FOURNIER07ee64e2015-07-06 23:43:03 +020013467 Returns a boolean TRUE if the input value of type signed integer is
Willy Tarreau97707872015-01-27 15:12:13 +010013468 non-null, otherwise returns FALSE. Used in conjunction with and(), it can be
Davor Ocelice9ed2812017-12-25 17:49:28 +010013469 used to report true/false for bit testing on input values (e.g. verify the
Willy Tarreau97707872015-01-27 15:12:13 +010013470 presence of a flag).
13471
Emeric Brun54c4ac82014-11-03 15:32:43 +010013472bytes(<offset>[,<length>])
13473 Extracts some bytes from an input binary sample. The result is a binary
13474 sample starting at an offset (in bytes) of the original sample and
Tim Düsterhus4896c442016-11-29 02:15:19 +010013475 optionally truncated at the given length.
Emeric Brun54c4ac82014-11-03 15:32:43 +010013476
Willy Tarreau280f42b2018-02-19 15:34:12 +010013477concat([<start>],[<var>],[<end>])
13478 Concatenates up to 3 fields after the current sample which is then turned to
13479 a string. The first one, <start>, is a constant string, that will be appended
13480 immediately after the existing sample. It may be omitted if not used. The
13481 second one, <var>, is a variable name. The variable will be looked up, its
13482 contents converted to a string, and it will be appended immediately after the
13483 <first> part. If the variable is not found, nothing is appended. It may be
13484 omitted as well. The third field, <end> is a constant string that will be
13485 appended after the variable. It may also be omitted. Together, these elements
13486 allow to concatenate variables with delimiters to an existing set of
13487 variables. This can be used to build new variables made of a succession of
John Roeslerfb2fce12019-07-10 15:45:51 -050013488 other variables, such as colon-delimited values. Note that due to the config
Willy Tarreau280f42b2018-02-19 15:34:12 +010013489 parser, it is not possible to use a comma nor a closing parenthesis as
John Roeslerfb2fce12019-07-10 15:45:51 -050013490 delimiters.
Willy Tarreau280f42b2018-02-19 15:34:12 +010013491
13492 Example:
13493 tcp-request session set-var(sess.src) src
13494 tcp-request session set-var(sess.dn) ssl_c_s_dn
13495 tcp-request session set-var(txn.sig) str(),concat(<ip=,sess.ip,>),concat(<dn=,sess.dn,>)
13496 http-request set-header x-hap-sig %[var(txn.sig)]
13497
Willy Tarreau97707872015-01-27 15:12:13 +010013498cpl
Thierry FOURNIER07ee64e2015-07-06 23:43:03 +020013499 Takes the input value of type signed integer, applies a ones-complement
13500 (flips all bits) and returns the result as an signed integer.
Willy Tarreau97707872015-01-27 15:12:13 +010013501
Willy Tarreau80599772015-01-20 19:35:24 +010013502crc32([<avalanche>])
13503 Hashes a binary input sample into an unsigned 32-bit quantity using the CRC32
13504 hash function. Optionally, it is possible to apply a full avalanche hash
13505 function to the output if the optional <avalanche> argument equals 1. This
13506 converter uses the same functions as used by the various hash-based load
13507 balancing algorithms, so it will provide exactly the same results. It is
13508 provided for compatibility with other software which want a CRC32 to be
13509 computed on some input keys, so it follows the most common implementation as
13510 found in Ethernet, Gzip, PNG, etc... It is slower than the other algorithms
13511 but may provide a better or at least less predictable distribution. It must
13512 not be used for security purposes as a 32-bit hash is trivial to break. See
Emmanuel Hocdet50791a72018-03-21 11:19:01 +010013513 also "djb2", "sdbm", "wt6", "crc32c" and the "hash-type" directive.
13514
13515crc32c([<avalanche>])
13516 Hashes a binary input sample into an unsigned 32-bit quantity using the CRC32C
13517 hash function. Optionally, it is possible to apply a full avalanche hash
13518 function to the output if the optional <avalanche> argument equals 1. This
13519 converter uses the same functions as described in RFC4960, Appendix B [8].
13520 It is provided for compatibility with other software which want a CRC32C to be
13521 computed on some input keys. It is slower than the other algorithms and it must
13522 not be used for security purposes as a 32-bit hash is trivial to break. See
13523 also "djb2", "sdbm", "wt6", "crc32" and the "hash-type" directive.
Willy Tarreau80599772015-01-20 19:35:24 +010013524
David Carlier29b3ca32015-09-25 14:09:21 +010013525da-csv-conv(<prop>[,<prop>*])
David Carlier4542b102015-06-01 13:54:29 +020013526 Asks the DeviceAtlas converter to identify the User Agent string passed on
13527 input, and to emit a string made of the concatenation of the properties
13528 enumerated in argument, delimited by the separator defined by the global
13529 keyword "deviceatlas-property-separator", or by default the pipe character
David Carlier840b0242016-03-16 10:09:55 +000013530 ('|'). There's a limit of 12 different properties imposed by the haproxy
David Carlier4542b102015-06-01 13:54:29 +020013531 configuration language.
13532
13533 Example:
13534 frontend www
Cyril Bonté307ee1e2015-09-28 23:16:06 +020013535 bind *:8881
13536 default_backend servers
David Carlier840b0242016-03-16 10:09:55 +000013537 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 +020013538
Thierry FOURNIER9687c772015-05-07 15:46:29 +020013539debug
13540 This converter is used as debug tool. It dumps on screen the content and the
13541 type of the input sample. The sample is returned as is on its output. This
13542 converter only exists when haproxy was built with debugging enabled.
13543
Willy Tarreau97707872015-01-27 15:12:13 +010013544div(<value>)
Thierry FOURNIER07ee64e2015-07-06 23:43:03 +020013545 Divides the input value of type signed integer by <value>, and returns the
13546 result as an signed integer. If <value> is null, the largest unsigned
Thierry FOURNIER5d86fae2015-07-07 21:10:16 +020013547 integer is returned (typically 2^63-1). <value> can be a numeric value or a
Daniel Schneller0b547052016-03-21 20:46:57 +010013548 variable name. The name of the variable starts with an indication about its
13549 scope. The scopes allowed are:
Christopher Fauletff2613e2016-11-09 11:36:17 +010013550 "proc" : the variable is shared with the whole process
Daniel Schneller0b547052016-03-21 20:46:57 +010013551 "sess" : the variable is shared with the whole session
13552 "txn" : the variable is shared with the transaction (request and response)
13553 "req" : the variable is shared only during request processing
13554 "res" : the variable is shared only during response processing
Davor Ocelice9ed2812017-12-25 17:49:28 +010013555 This prefix is followed by a name. The separator is a '.'. The name may only
Christopher Fauletb71557a2016-10-31 10:49:03 +010013556 contain characters 'a-z', 'A-Z', '0-9', '.' and '_'.
Willy Tarreau97707872015-01-27 15:12:13 +010013557
Willy Tarreau23ec4ca2014-07-15 20:15:37 +020013558djb2([<avalanche>])
13559 Hashes a binary input sample into an unsigned 32-bit quantity using the DJB2
13560 hash function. Optionally, it is possible to apply a full avalanche hash
13561 function to the output if the optional <avalanche> argument equals 1. This
13562 converter uses the same functions as used by the various hash-based load
13563 balancing algorithms, so it will provide exactly the same results. It is
13564 mostly intended for debugging, but can be used as a stick-table entry to
13565 collect rough statistics. It must not be used for security purposes as a
Emmanuel Hocdet50791a72018-03-21 11:19:01 +010013566 32-bit hash is trivial to break. See also "crc32", "sdbm", "wt6", "crc32c",
13567 and the "hash-type" directive.
Willy Tarreau23ec4ca2014-07-15 20:15:37 +020013568
Willy Tarreau97707872015-01-27 15:12:13 +010013569even
Thierry FOURNIER07ee64e2015-07-06 23:43:03 +020013570 Returns a boolean TRUE if the input value of type signed integer is even
Willy Tarreau97707872015-01-27 15:12:13 +010013571 otherwise returns FALSE. It is functionally equivalent to "not,and(1),bool".
13572
Marcin Deranek9631a282018-04-16 14:30:46 +020013573field(<index>,<delimiters>[,<count>])
13574 Extracts the substring at the given index counting from the beginning
13575 (positive index) or from the end (negative index) considering given delimiters
13576 from an input string. Indexes start at 1 or -1 and delimiters are a string
13577 formatted list of chars. Optionally you can specify <count> of fields to
13578 extract (default: 1). Value of 0 indicates extraction of all remaining
13579 fields.
13580
13581 Example :
13582 str(f1_f2_f3__f5),field(5,_) # f5
13583 str(f1_f2_f3__f5),field(2,_,0) # f2_f3__f5
13584 str(f1_f2_f3__f5),field(2,_,2) # f2_f3
13585 str(f1_f2_f3__f5),field(-2,_,3) # f2_f3_
13586 str(f1_f2_f3__f5),field(-3,_,0) # f1_f2_f3
Emeric Brunf399b0d2014-11-03 17:07:03 +010013587
Thierry FOURNIER060762e2014-04-23 13:29:15 +020013588hex
Davor Ocelice9ed2812017-12-25 17:49:28 +010013589 Converts a binary input sample to a hex string containing two hex digits per
Thierry FOURNIER060762e2014-04-23 13:29:15 +020013590 input byte. It is used to log or transfer hex dumps of some binary input data
Davor Ocelice9ed2812017-12-25 17:49:28 +010013591 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 +020013592 header).
Thierry FOURNIER2f49d6d2014-03-12 15:01:52 +010013593
Dragan Dosen3f957b22017-10-24 09:27:34 +020013594hex2i
13595 Converts a hex string containing two hex digits per input byte to an
John Roeslerfb2fce12019-07-10 15:45:51 -050013596 integer. If the input value cannot be converted, then zero is returned.
Dragan Dosen3f957b22017-10-24 09:27:34 +020013597
Thierry FOURNIER060762e2014-04-23 13:29:15 +020013598http_date([<offset>])
13599 Converts an integer supposed to contain a date since epoch to a string
13600 representing this date in a format suitable for use in HTTP header fields. If
13601 an offset value is specified, then it is a number of seconds that is added to
13602 the date before the conversion is operated. This is particularly useful to
13603 emit Date header fields, Expires values in responses when combined with a
13604 positive offset, or Last-Modified values when the offset is negative.
Willy Tarreau74ca5042013-06-11 23:12:07 +020013605
Willy Tarreaud9f316a2014-07-10 14:03:38 +020013606in_table(<table>)
13607 Uses the string representation of the input sample to perform a look up in
13608 the specified table. If the key is not found in the table, a boolean false
13609 is returned. Otherwise a boolean true is returned. This can be used to verify
Davor Ocelice9ed2812017-12-25 17:49:28 +010013610 the presence of a certain key in a table tracking some elements (e.g. whether
Willy Tarreaud9f316a2014-07-10 14:03:38 +020013611 or not a source IP address or an Authorization header was already seen).
13612
Tim Duesterhus1478aa72018-01-25 16:24:51 +010013613ipmask(<mask4>, [<mask6>])
13614 Apply a mask to an IP address, and use the result for lookups and storage.
Willy Tarreauffcb2e42014-07-10 16:29:08 +020013615 This can be used to make all hosts within a certain mask to share the same
Tim Duesterhus1478aa72018-01-25 16:24:51 +010013616 table entries and as such use the same server. The mask4 can be passed in
13617 dotted form (e.g. 255.255.255.0) or in CIDR form (e.g. 24). The mask6 can
13618 be passed in quadruplet form (e.g. ffff:ffff::) or in CIDR form (e.g. 64).
13619 If no mask6 is given IPv6 addresses will fail to convert for backwards
13620 compatibility reasons.
Willy Tarreauffcb2e42014-07-10 16:29:08 +020013621
Thierry FOURNIER317e1c42014-08-12 10:20:47 +020013622json([<input-code>])
Davor Ocelice9ed2812017-12-25 17:49:28 +010013623 Escapes the input string and produces an ASCII output string ready to use as a
Thierry FOURNIER317e1c42014-08-12 10:20:47 +020013624 JSON string. The converter tries to decode the input string according to the
Herve COMMOWICK8dfe8632016-08-05 12:01:20 +020013625 <input-code> parameter. It can be "ascii", "utf8", "utf8s", "utf8p" or
Thierry FOURNIER317e1c42014-08-12 10:20:47 +020013626 "utf8ps". The "ascii" decoder never fails. The "utf8" decoder detects 3 types
13627 of errors:
13628 - bad UTF-8 sequence (lone continuation byte, bad number of continuation
13629 bytes, ...)
13630 - invalid range (the decoded value is within a UTF-8 prohibited range),
13631 - code overlong (the value is encoded with more bytes than necessary).
13632
13633 The UTF-8 JSON encoding can produce a "too long value" error when the UTF-8
13634 character is greater than 0xffff because the JSON string escape specification
13635 only authorizes 4 hex digits for the value encoding. The UTF-8 decoder exists
13636 in 4 variants designated by a combination of two suffix letters : "p" for
13637 "permissive" and "s" for "silently ignore". The behaviors of the decoders
13638 are :
Davor Ocelice9ed2812017-12-25 17:49:28 +010013639 - "ascii" : never fails;
13640 - "utf8" : fails on any detected errors;
13641 - "utf8s" : never fails, but removes characters corresponding to errors;
Thierry FOURNIER317e1c42014-08-12 10:20:47 +020013642 - "utf8p" : accepts and fixes the overlong errors, but fails on any other
Davor Ocelice9ed2812017-12-25 17:49:28 +010013643 error;
Thierry FOURNIER317e1c42014-08-12 10:20:47 +020013644 - "utf8ps" : never fails, accepts and fixes the overlong errors, but removes
13645 characters corresponding to the other errors.
13646
13647 This converter is particularly useful for building properly escaped JSON for
Davor Ocelice9ed2812017-12-25 17:49:28 +010013648 logging to servers which consume JSON-formatted traffic logs.
Thierry FOURNIER317e1c42014-08-12 10:20:47 +020013649
13650 Example:
Thierry FOURNIER317e1c42014-08-12 10:20:47 +020013651 capture request header Host len 15
Herve COMMOWICK8dfe8632016-08-05 12:01:20 +020013652 capture request header user-agent len 150
13653 log-format '{"ip":"%[src]","user-agent":"%[capture.req.hdr(1),json(utf8s)]"}'
Thierry FOURNIER317e1c42014-08-12 10:20:47 +020013654
13655 Input request from client 127.0.0.1:
13656 GET / HTTP/1.0
13657 User-Agent: Very "Ugly" UA 1/2
13658
13659 Output log:
13660 {"ip":"127.0.0.1","user-agent":"Very \"Ugly\" UA 1\/2"}
13661
Thierry FOURNIER060762e2014-04-23 13:29:15 +020013662language(<value>[,<default>])
13663 Returns the value with the highest q-factor from a list as extracted from the
13664 "accept-language" header using "req.fhdr". Values with no q-factor have a
13665 q-factor of 1. Values with a q-factor of 0 are dropped. Only values which
13666 belong to the list of semi-colon delimited <values> will be considered. The
13667 argument <value> syntax is "lang[;lang[;lang[;...]]]". If no value matches the
13668 given list and a default value is provided, it is returned. Note that language
13669 names may have a variant after a dash ('-'). If this variant is present in the
13670 list, it will be matched, but if it is not, only the base language is checked.
13671 The match is case-sensitive, and the output string is always one of those
Davor Ocelice9ed2812017-12-25 17:49:28 +010013672 provided in arguments. The ordering of arguments is meaningless, only the
Thierry FOURNIER060762e2014-04-23 13:29:15 +020013673 ordering of the values in the request counts, as the first value among
13674 multiple sharing the same q-factor is used.
Thierry FOURNIERad903512014-04-11 17:51:01 +020013675
Thierry FOURNIER060762e2014-04-23 13:29:15 +020013676 Example :
Thierry FOURNIERad903512014-04-11 17:51:01 +020013677
Thierry FOURNIER060762e2014-04-23 13:29:15 +020013678 # this configuration switches to the backend matching a
13679 # given language based on the request :
Thierry FOURNIERad903512014-04-11 17:51:01 +020013680
Thierry FOURNIER060762e2014-04-23 13:29:15 +020013681 acl es req.fhdr(accept-language),language(es;fr;en) -m str es
13682 acl fr req.fhdr(accept-language),language(es;fr;en) -m str fr
13683 acl en req.fhdr(accept-language),language(es;fr;en) -m str en
13684 use_backend spanish if es
13685 use_backend french if fr
13686 use_backend english if en
13687 default_backend choose_your_language
Thierry FOURNIERad903512014-04-11 17:51:01 +020013688
Willy Tarreau60a2ee72017-12-15 07:13:48 +010013689length
Etienne Carriereed0d24e2017-12-13 13:41:34 +010013690 Get the length of the string. This can only be placed after a string
13691 sample fetch function or after a transformation keyword returning a string
13692 type. The result is of type integer.
13693
Willy Tarreauffcb2e42014-07-10 16:29:08 +020013694lower
13695 Convert a string sample to lower case. This can only be placed after a string
13696 sample fetch function or after a transformation keyword returning a string
13697 type. The result is of type string.
13698
Willy Tarreau0dbfdba2014-07-10 16:37:47 +020013699ltime(<format>[,<offset>])
13700 Converts an integer supposed to contain a date since epoch to a string
13701 representing this date in local time using a format defined by the <format>
13702 string using strftime(3). The purpose is to allow any date format to be used
13703 in logs. An optional <offset> in seconds may be applied to the input date
13704 (positive or negative). See the strftime() man page for the format supported
13705 by your operating system. See also the utime converter.
13706
13707 Example :
13708
13709 # Emit two colons, one with the local time and another with ip:port
Davor Ocelice9ed2812017-12-25 17:49:28 +010013710 # e.g. 20140710162350 127.0.0.1:57325
Willy Tarreau0dbfdba2014-07-10 16:37:47 +020013711 log-format %[date,ltime(%Y%m%d%H%M%S)]\ %ci:%cp
13712
Thierry FOURNIER060762e2014-04-23 13:29:15 +020013713map(<map_file>[,<default_value>])
13714map_<match_type>(<map_file>[,<default_value>])
13715map_<match_type>_<output_type>(<map_file>[,<default_value>])
13716 Search the input value from <map_file> using the <match_type> matching method,
13717 and return the associated value converted to the type <output_type>. If the
13718 input value cannot be found in the <map_file>, the converter returns the
13719 <default_value>. If the <default_value> is not set, the converter fails and
13720 acts as if no input value could be fetched. If the <match_type> is not set, it
13721 defaults to "str". Likewise, if the <output_type> is not set, it defaults to
13722 "str". For convenience, the "map" keyword is an alias for "map_str" and maps a
13723 string to another string.
Thierry FOURNIERd5f624d2013-11-26 11:52:33 +010013724
Thierry FOURNIER060762e2014-04-23 13:29:15 +020013725 It is important to avoid overlapping between the keys : IP addresses and
13726 strings are stored in trees, so the first of the finest match will be used.
13727 Other keys are stored in lists, so the first matching occurrence will be used.
Thierry FOURNIERd5f624d2013-11-26 11:52:33 +010013728
Tim Düsterhus4896c442016-11-29 02:15:19 +010013729 The following array contains the list of all map functions available sorted by
Thierry FOURNIER060762e2014-04-23 13:29:15 +020013730 input type, match type and output type.
Thierry FOURNIERd5f624d2013-11-26 11:52:33 +010013731
Thierry FOURNIER060762e2014-04-23 13:29:15 +020013732 input type | match method | output type str | output type int | output type ip
13733 -----------+--------------+-----------------+-----------------+---------------
13734 str | str | map_str | map_str_int | map_str_ip
13735 -----------+--------------+-----------------+-----------------+---------------
Willy Tarreau787a4c02014-05-10 07:55:30 +020013736 str | beg | map_beg | map_beg_int | map_end_ip
13737 -----------+--------------+-----------------+-----------------+---------------
Thierry FOURNIER060762e2014-04-23 13:29:15 +020013738 str | sub | map_sub | map_sub_int | map_sub_ip
13739 -----------+--------------+-----------------+-----------------+---------------
13740 str | dir | map_dir | map_dir_int | map_dir_ip
13741 -----------+--------------+-----------------+-----------------+---------------
13742 str | dom | map_dom | map_dom_int | map_dom_ip
13743 -----------+--------------+-----------------+-----------------+---------------
13744 str | end | map_end | map_end_int | map_end_ip
13745 -----------+--------------+-----------------+-----------------+---------------
Ruoshan Huang3c5e3742016-12-02 16:25:31 +080013746 str | reg | map_reg | map_reg_int | map_reg_ip
13747 -----------+--------------+-----------------+-----------------+---------------
13748 str | reg | map_regm | map_reg_int | map_reg_ip
Thierry FOURNIER060762e2014-04-23 13:29:15 +020013749 -----------+--------------+-----------------+-----------------+---------------
13750 int | int | map_int | map_int_int | map_int_ip
13751 -----------+--------------+-----------------+-----------------+---------------
13752 ip | ip | map_ip | map_ip_int | map_ip_ip
13753 -----------+--------------+-----------------+-----------------+---------------
Thierry FOURNIERd5f624d2013-11-26 11:52:33 +010013754
Thierry Fournier8feaa662016-02-10 22:55:20 +010013755 The special map called "map_regm" expect matching zone in the regular
13756 expression and modify the output replacing back reference (like "\1") by
13757 the corresponding match text.
13758
Thierry FOURNIER060762e2014-04-23 13:29:15 +020013759 The file contains one key + value per line. Lines which start with '#' are
13760 ignored, just like empty lines. Leading tabs and spaces are stripped. The key
13761 is then the first "word" (series of non-space/tabs characters), and the value
13762 is what follows this series of space/tab till the end of the line excluding
13763 trailing spaces/tabs.
Thierry FOURNIERd5f624d2013-11-26 11:52:33 +010013764
Thierry FOURNIER060762e2014-04-23 13:29:15 +020013765 Example :
13766
13767 # this is a comment and is ignored
13768 2.22.246.0/23 United Kingdom \n
13769 <-><-----------><--><------------><---->
13770 | | | | `- trailing spaces ignored
13771 | | | `---------- value
13772 | | `-------------------- middle spaces ignored
13773 | `---------------------------- key
13774 `------------------------------------ leading spaces ignored
13775
Willy Tarreau97707872015-01-27 15:12:13 +010013776mod(<value>)
Thierry FOURNIER07ee64e2015-07-06 23:43:03 +020013777 Divides the input value of type signed integer by <value>, and returns the
13778 remainder as an signed integer. If <value> is null, then zero is returned.
Thierry FOURNIER5d86fae2015-07-07 21:10:16 +020013779 <value> can be a numeric value or a variable name. The name of the variable
Daniel Schneller0b547052016-03-21 20:46:57 +010013780 starts with an indication about its scope. The scopes allowed are:
Christopher Fauletff2613e2016-11-09 11:36:17 +010013781 "proc" : the variable is shared with the whole process
Daniel Schneller0b547052016-03-21 20:46:57 +010013782 "sess" : the variable is shared with the whole session
13783 "txn" : the variable is shared with the transaction (request and response)
13784 "req" : the variable is shared only during request processing
13785 "res" : the variable is shared only during response processing
Davor Ocelice9ed2812017-12-25 17:49:28 +010013786 This prefix is followed by a name. The separator is a '.'. The name may only
Christopher Fauletb71557a2016-10-31 10:49:03 +010013787 contain characters 'a-z', 'A-Z', '0-9', '.' and '_'.
Willy Tarreau97707872015-01-27 15:12:13 +010013788
13789mul(<value>)
Thierry FOURNIER07ee64e2015-07-06 23:43:03 +020013790 Multiplies the input value of type signed integer by <value>, and returns
Thierry FOURNIER00c005c2015-07-08 01:10:21 +020013791 the product as an signed integer. In case of overflow, the largest possible
13792 value for the sign is returned so that the operation doesn't wrap around.
Thierry FOURNIER5d86fae2015-07-07 21:10:16 +020013793 <value> can be a numeric value or a variable name. The name of the variable
Daniel Schneller0b547052016-03-21 20:46:57 +010013794 starts with an indication about its scope. The scopes allowed are:
Christopher Fauletff2613e2016-11-09 11:36:17 +010013795 "proc" : the variable is shared with the whole process
Daniel Schneller0b547052016-03-21 20:46:57 +010013796 "sess" : the variable is shared with the whole session
13797 "txn" : the variable is shared with the transaction (request and response)
13798 "req" : the variable is shared only during request processing
13799 "res" : the variable is shared only during response processing
Davor Ocelice9ed2812017-12-25 17:49:28 +010013800 This prefix is followed by a name. The separator is a '.'. The name may only
Christopher Fauletb71557a2016-10-31 10:49:03 +010013801 contain characters 'a-z', 'A-Z', '0-9', '.' and '_'.
Willy Tarreau97707872015-01-27 15:12:13 +010013802
Nenad Merdanovicb7e7c472017-03-12 21:56:55 +010013803nbsrv
13804 Takes an input value of type string, interprets it as a backend name and
13805 returns the number of usable servers in that backend. Can be used in places
13806 where we want to look up a backend from a dynamic name, like a result of a
13807 map lookup.
13808
Willy Tarreau97707872015-01-27 15:12:13 +010013809neg
Thierry FOURNIER07ee64e2015-07-06 23:43:03 +020013810 Takes the input value of type signed integer, computes the opposite value,
13811 and returns the remainder as an signed integer. 0 is identity. This operator
13812 is provided for reversed subtracts : in order to subtract the input from a
13813 constant, simply perform a "neg,add(value)".
Willy Tarreau97707872015-01-27 15:12:13 +010013814
13815not
Thierry FOURNIER07ee64e2015-07-06 23:43:03 +020013816 Returns a boolean FALSE if the input value of type signed integer is
Willy Tarreau97707872015-01-27 15:12:13 +010013817 non-null, otherwise returns TRUE. Used in conjunction with and(), it can be
Davor Ocelice9ed2812017-12-25 17:49:28 +010013818 used to report true/false for bit testing on input values (e.g. verify the
Willy Tarreau97707872015-01-27 15:12:13 +010013819 absence of a flag).
13820
13821odd
Thierry FOURNIER07ee64e2015-07-06 23:43:03 +020013822 Returns a boolean TRUE if the input value of type signed integer is odd
Willy Tarreau97707872015-01-27 15:12:13 +010013823 otherwise returns FALSE. It is functionally equivalent to "and(1),bool".
13824
13825or(<value>)
Thierry FOURNIER07ee64e2015-07-06 23:43:03 +020013826 Performs a bitwise "OR" between <value> and the input value of type signed
Thierry FOURNIER5d86fae2015-07-07 21:10:16 +020013827 integer, and returns the result as an signed integer. <value> can be a
Daniel Schneller0b547052016-03-21 20:46:57 +010013828 numeric value or a variable name. The name of the variable starts with an
13829 indication about its scope. The scopes allowed are:
Christopher Fauletff2613e2016-11-09 11:36:17 +010013830 "proc" : the variable is shared with the whole process
Daniel Schneller0b547052016-03-21 20:46:57 +010013831 "sess" : the variable is shared with the whole session
13832 "txn" : the variable is shared with the transaction (request and response)
13833 "req" : the variable is shared only during request processing
13834 "res" : the variable is shared only during response processing
Davor Ocelice9ed2812017-12-25 17:49:28 +010013835 This prefix is followed by a name. The separator is a '.'. The name may only
Christopher Fauletb71557a2016-10-31 10:49:03 +010013836 contain characters 'a-z', 'A-Z', '0-9', '.' and '_'.
Willy Tarreau97707872015-01-27 15:12:13 +010013837
Frédéric Lécaillebfe61382019-03-06 14:34:36 +010013838protobuf(<field_number>,[<field_type>])
13839 This extracts the protocol buffers message field in raw mode of an input binary
13840 sample representation of a protocol buffer message with <field_number> as field
13841 number (dotted notation) if <field_type> is not present, or as an integer sample
13842 if this field is present (see also "ungrpc" below).
13843 The list of the authorized types is the following one: "int32", "int64", "uint32",
13844 "uint64", "sint32", "sint64", "bool", "enum" for the "varint" wire type 0
13845 "fixed64", "sfixed64", "double" for the 64bit wire type 1, "fixed32", "sfixed32",
13846 "float" for the wire type 5. Note that "string" is considered as a length-delimited
13847 type, so it does not require any <field_type> argument to be extracted.
13848 More information may be found here about the protocol buffers message field types:
13849 https://developers.google.com/protocol-buffers/docs/encoding
13850
Willy Tarreauc4dc3502015-01-23 20:39:28 +010013851regsub(<regex>,<subst>[,<flags>])
Willy Tarreau7eda8492015-01-20 19:47:06 +010013852 Applies a regex-based substitution to the input string. It does the same
13853 operation as the well-known "sed" utility with "s/<regex>/<subst>/". By
13854 default it will replace in the input string the first occurrence of the
13855 largest part matching the regular expression <regex> with the substitution
13856 string <subst>. It is possible to replace all occurrences instead by adding
13857 the flag "g" in the third argument <flags>. It is also possible to make the
13858 regex case insensitive by adding the flag "i" in <flags>. Since <flags> is a
13859 string, it is made up from the concatenation of all desired flags. Thus if
13860 both "i" and "g" are desired, using "gi" or "ig" will have the same effect.
13861 It is important to note that due to the current limitations of the
Baptiste Assmann66025d82016-03-06 23:36:48 +010013862 configuration parser, some characters such as closing parenthesis, closing
13863 square brackets or comma are not possible to use in the arguments. The first
13864 use of this converter is to replace certain characters or sequence of
13865 characters with other ones.
Willy Tarreau7eda8492015-01-20 19:47:06 +010013866
13867 Example :
13868
13869 # de-duplicate "/" in header "x-path".
13870 # input: x-path: /////a///b/c/xzxyz/
13871 # output: x-path: /a/b/c/xzxyz/
13872 http-request set-header x-path %[hdr(x-path),regsub(/+,/,g)]
13873
Thierry FOURNIER35ab2752015-05-28 13:22:03 +020013874capture-req(<id>)
13875 Capture the string entry in the request slot <id> and returns the entry as
13876 is. If the slot doesn't exist, the capture fails silently.
13877
13878 See also: "declare capture", "http-request capture",
Baptiste Assmann5ac425c2015-10-21 23:13:46 +020013879 "http-response capture", "capture.req.hdr" and
13880 "capture.res.hdr" (sample fetches).
Thierry FOURNIER35ab2752015-05-28 13:22:03 +020013881
13882capture-res(<id>)
13883 Capture the string entry in the response slot <id> and returns the entry as
13884 is. If the slot doesn't exist, the capture fails silently.
13885
13886 See also: "declare capture", "http-request capture",
Baptiste Assmann5ac425c2015-10-21 23:13:46 +020013887 "http-response capture", "capture.req.hdr" and
13888 "capture.res.hdr" (sample fetches).
Thierry FOURNIER35ab2752015-05-28 13:22:03 +020013889
Willy Tarreau23ec4ca2014-07-15 20:15:37 +020013890sdbm([<avalanche>])
13891 Hashes a binary input sample into an unsigned 32-bit quantity using the SDBM
13892 hash function. Optionally, it is possible to apply a full avalanche hash
13893 function to the output if the optional <avalanche> argument equals 1. This
13894 converter uses the same functions as used by the various hash-based load
13895 balancing algorithms, so it will provide exactly the same results. It is
13896 mostly intended for debugging, but can be used as a stick-table entry to
13897 collect rough statistics. It must not be used for security purposes as a
Emmanuel Hocdet50791a72018-03-21 11:19:01 +010013898 32-bit hash is trivial to break. See also "crc32", "djb2", "wt6", "crc32c",
13899 and the "hash-type" directive.
Willy Tarreau23ec4ca2014-07-15 20:15:37 +020013900
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020013901set-var(<var name>)
Davor Ocelice9ed2812017-12-25 17:49:28 +010013902 Sets a variable with the input content and returns the content on the output
13903 as-is. The variable keeps the value and the associated input type. The name of
13904 the variable starts with an indication about its scope. The scopes allowed are:
Christopher Fauletff2613e2016-11-09 11:36:17 +010013905 "proc" : the variable is shared with the whole process
Daniel Schneller0b547052016-03-21 20:46:57 +010013906 "sess" : the variable is shared with the whole session
13907 "txn" : the variable is shared with the transaction (request and
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020013908 response),
Daniel Schneller0b547052016-03-21 20:46:57 +010013909 "req" : the variable is shared only during request processing,
13910 "res" : the variable is shared only during response processing.
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020013911 This prefix is followed by a name. The separator is a '.'. The name may only
Christopher Fauletb71557a2016-10-31 10:49:03 +010013912 contain characters 'a-z', 'A-Z', '0-9', '.' and '_'.
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020013913
Dragan Dosen6e5a9ca2017-10-24 09:18:23 +020013914sha1
Tim Duesterhusd4376302019-06-17 12:41:44 +020013915 Converts a binary input sample to a SHA-1 digest. The result is a binary
Dragan Dosen6e5a9ca2017-10-24 09:18:23 +020013916 sample with length of 20 bytes.
13917
Tim Duesterhusd4376302019-06-17 12:41:44 +020013918sha2([<bits>])
13919 Converts a binary input sample to a digest in the SHA-2 family. The result
13920 is a binary sample with length of <bits>/8 bytes.
13921
13922 Valid values for <bits> are 224, 256, 384, 512, each corresponding to
13923 SHA-<bits>. The default value is 256.
13924
13925 Please note that this converter is only available when haproxy has been
13926 compiled with USE_OPENSSL.
13927
Tim Duesterhusca097c12018-04-27 21:18:45 +020013928strcmp(<var>)
13929 Compares the contents of <var> with the input value of type string. Returns
13930 the result as a signed integer compatible with strcmp(3): 0 if both strings
13931 are identical. A value less than 0 if the left string is lexicographically
13932 smaller than the right string or if the left string is shorter. A value greater
13933 than 0 otherwise (right string greater than left string or the right string is
13934 shorter).
13935
13936 Example :
13937
13938 http-request set-var(txn.host) hdr(host)
13939 # Check whether the client is attempting domain fronting.
13940 acl ssl_sni_http_host_match ssl_fc_sni,strcmp(txn.host) eq 0
13941
13942
Willy Tarreau97707872015-01-27 15:12:13 +010013943sub(<value>)
Thierry FOURNIER07ee64e2015-07-06 23:43:03 +020013944 Subtracts <value> from the input value of type signed integer, and returns
13945 the result as an signed integer. Note: in order to subtract the input from
Thierry FOURNIER5d86fae2015-07-07 21:10:16 +020013946 a constant, simply perform a "neg,add(value)". <value> can be a numeric value
Daniel Schneller0b547052016-03-21 20:46:57 +010013947 or a variable name. The name of the variable starts with an indication about
13948 its scope. The scopes allowed are:
Christopher Fauletff2613e2016-11-09 11:36:17 +010013949 "proc" : the variable is shared with the whole process
Daniel Schneller0b547052016-03-21 20:46:57 +010013950 "sess" : the variable is shared with the whole session
13951 "txn" : the variable is shared with the transaction (request and
Thierry FOURNIER5d86fae2015-07-07 21:10:16 +020013952 response),
Daniel Schneller0b547052016-03-21 20:46:57 +010013953 "req" : the variable is shared only during request processing,
13954 "res" : the variable is shared only during response processing.
Thierry FOURNIER5d86fae2015-07-07 21:10:16 +020013955 This prefix is followed by a name. The separator is a '.'. The name may only
Christopher Fauletb71557a2016-10-31 10:49:03 +010013956 contain characters 'a-z', 'A-Z', '0-9', '.' and '_'.
Willy Tarreau97707872015-01-27 15:12:13 +010013957
Willy Tarreaud9f316a2014-07-10 14:03:38 +020013958table_bytes_in_rate(<table>)
13959 Uses the string representation of the input sample to perform a look up in
13960 the specified table. If the key is not found in the table, integer value zero
13961 is returned. Otherwise the converter returns the average client-to-server
13962 bytes rate associated with the input sample in the designated table, measured
13963 in amount of bytes over the period configured in the table. See also the
13964 sc_bytes_in_rate sample fetch keyword.
13965
13966
13967table_bytes_out_rate(<table>)
13968 Uses the string representation of the input sample to perform a look up in
13969 the specified table. If the key is not found in the table, integer value zero
13970 is returned. Otherwise the converter returns the average server-to-client
13971 bytes rate associated with the input sample in the designated table, measured
13972 in amount of bytes over the period configured in the table. See also the
13973 sc_bytes_out_rate sample fetch keyword.
13974
13975table_conn_cnt(<table>)
13976 Uses the string representation of the input sample to perform a look up in
13977 the specified table. If the key is not found in the table, integer value zero
Davor Ocelice9ed2812017-12-25 17:49:28 +010013978 is returned. Otherwise the converter returns the cumulative number of incoming
Willy Tarreaud9f316a2014-07-10 14:03:38 +020013979 connections associated with the input sample in the designated table. See
13980 also the sc_conn_cnt sample fetch keyword.
13981
13982table_conn_cur(<table>)
13983 Uses the string representation of the input sample to perform a look up in
13984 the specified table. If the key is not found in the table, integer value zero
13985 is returned. Otherwise the converter returns the current amount of concurrent
13986 tracked connections associated with the input sample in the designated table.
13987 See also the sc_conn_cur sample fetch keyword.
13988
13989table_conn_rate(<table>)
13990 Uses the string representation of the input sample to perform a look up in
13991 the specified table. If the key is not found in the table, integer value zero
13992 is returned. Otherwise the converter returns the average incoming connection
13993 rate associated with the input sample in the designated table. See also the
13994 sc_conn_rate sample fetch keyword.
13995
Thierry FOURNIER236657b2015-08-19 08:25:14 +020013996table_gpt0(<table>)
13997 Uses the string representation of the input sample to perform a look up in
13998 the specified table. If the key is not found in the table, boolean value zero
13999 is returned. Otherwise the converter returns the current value of the first
14000 general purpose tag associated with the input sample in the designated table.
14001 See also the sc_get_gpt0 sample fetch keyword.
14002
Willy Tarreaud9f316a2014-07-10 14:03:38 +020014003table_gpc0(<table>)
14004 Uses the string representation of the input sample to perform a look up in
14005 the specified table. If the key is not found in the table, integer value zero
14006 is returned. Otherwise the converter returns the current value of the first
14007 general purpose counter associated with the input sample in the designated
14008 table. See also the sc_get_gpc0 sample fetch keyword.
14009
14010table_gpc0_rate(<table>)
14011 Uses the string representation of the input sample to perform a look up in
14012 the specified table. If the key is not found in the table, integer value zero
14013 is returned. Otherwise the converter returns the frequency which the gpc0
14014 counter was incremented over the configured period in the table, associated
14015 with the input sample in the designated table. See also the sc_get_gpc0_rate
14016 sample fetch keyword.
14017
Frédéric Lécaille6778b272018-01-29 15:22:53 +010014018table_gpc1(<table>)
14019 Uses the string representation of the input sample to perform a look up in
14020 the specified table. If the key is not found in the table, integer value zero
14021 is returned. Otherwise the converter returns the current value of the second
14022 general purpose counter associated with the input sample in the designated
14023 table. See also the sc_get_gpc1 sample fetch keyword.
14024
14025table_gpc1_rate(<table>)
14026 Uses the string representation of the input sample to perform a look up in
14027 the specified table. If the key is not found in the table, integer value zero
14028 is returned. Otherwise the converter returns the frequency which the gpc1
14029 counter was incremented over the configured period in the table, associated
14030 with the input sample in the designated table. See also the sc_get_gpc1_rate
14031 sample fetch keyword.
14032
Willy Tarreaud9f316a2014-07-10 14:03:38 +020014033table_http_err_cnt(<table>)
14034 Uses the string representation of the input sample to perform a look up in
14035 the specified table. If the key is not found in the table, integer value zero
Davor Ocelice9ed2812017-12-25 17:49:28 +010014036 is returned. Otherwise the converter returns the cumulative number of HTTP
Willy Tarreaud9f316a2014-07-10 14:03:38 +020014037 errors associated with the input sample in the designated table. See also the
14038 sc_http_err_cnt sample fetch keyword.
14039
14040table_http_err_rate(<table>)
14041 Uses the string representation of the input sample to perform a look up in
14042 the specified table. If the key is not found in the table, integer value zero
14043 is returned. Otherwise the average rate of HTTP errors associated with the
14044 input sample in the designated table, measured in amount of errors over the
14045 period configured in the table. See also the sc_http_err_rate sample fetch
14046 keyword.
14047
14048table_http_req_cnt(<table>)
14049 Uses the string representation of the input sample to perform a look up in
14050 the specified table. If the key is not found in the table, integer value zero
Davor Ocelice9ed2812017-12-25 17:49:28 +010014051 is returned. Otherwise the converter returns the cumulative number of HTTP
Willy Tarreaud9f316a2014-07-10 14:03:38 +020014052 requests associated with the input sample in the designated table. See also
14053 the sc_http_req_cnt sample fetch keyword.
14054
14055table_http_req_rate(<table>)
14056 Uses the string representation of the input sample to perform a look up in
14057 the specified table. If the key is not found in the table, integer value zero
14058 is returned. Otherwise the average rate of HTTP requests associated with the
14059 input sample in the designated table, measured in amount of requests over the
14060 period configured in the table. See also the sc_http_req_rate sample fetch
14061 keyword.
14062
14063table_kbytes_in(<table>)
14064 Uses the string representation of the input sample to perform a look up in
14065 the specified table. If the key is not found in the table, integer value zero
Davor Ocelice9ed2812017-12-25 17:49:28 +010014066 is returned. Otherwise the converter returns the cumulative number of client-
Willy Tarreaud9f316a2014-07-10 14:03:38 +020014067 to-server data associated with the input sample in the designated table,
14068 measured in kilobytes. The test is currently performed on 32-bit integers,
14069 which limits values to 4 terabytes. See also the sc_kbytes_in sample fetch
14070 keyword.
14071
14072table_kbytes_out(<table>)
14073 Uses the string representation of the input sample to perform a look up in
14074 the specified table. If the key is not found in the table, integer value zero
Davor Ocelice9ed2812017-12-25 17:49:28 +010014075 is returned. Otherwise the converter returns the cumulative number of server-
Willy Tarreaud9f316a2014-07-10 14:03:38 +020014076 to-client data associated with the input sample in the designated table,
14077 measured in kilobytes. The test is currently performed on 32-bit integers,
14078 which limits values to 4 terabytes. See also the sc_kbytes_out sample fetch
14079 keyword.
14080
14081table_server_id(<table>)
14082 Uses the string representation of the input sample to perform a look up in
14083 the specified table. If the key is not found in the table, integer value zero
14084 is returned. Otherwise the converter returns the server ID associated with
14085 the input sample in the designated table. A server ID is associated to a
14086 sample by a "stick" rule when a connection to a server succeeds. A server ID
14087 zero means that no server is associated with this key.
14088
14089table_sess_cnt(<table>)
14090 Uses the string representation of the input sample to perform a look up in
14091 the specified table. If the key is not found in the table, integer value zero
Davor Ocelice9ed2812017-12-25 17:49:28 +010014092 is returned. Otherwise the converter returns the cumulative number of incoming
Willy Tarreaud9f316a2014-07-10 14:03:38 +020014093 sessions associated with the input sample in the designated table. Note that
14094 a session here refers to an incoming connection being accepted by the
14095 "tcp-request connection" rulesets. See also the sc_sess_cnt sample fetch
14096 keyword.
14097
14098table_sess_rate(<table>)
14099 Uses the string representation of the input sample to perform a look up in
14100 the specified table. If the key is not found in the table, integer value zero
14101 is returned. Otherwise the converter returns the average incoming session
14102 rate associated with the input sample in the designated table. Note that a
14103 session here refers to an incoming connection being accepted by the
14104 "tcp-request connection" rulesets. See also the sc_sess_rate sample fetch
14105 keyword.
14106
14107table_trackers(<table>)
14108 Uses the string representation of the input sample to perform a look up in
14109 the specified table. If the key is not found in the table, integer value zero
14110 is returned. Otherwise the converter returns the current amount of concurrent
14111 connections tracking the same key as the input sample in the designated
14112 table. It differs from table_conn_cur in that it does not rely on any stored
14113 information but on the table's reference count (the "use" value which is
14114 returned by "show table" on the CLI). This may sometimes be more suited for
14115 layer7 tracking. It can be used to tell a server how many concurrent
14116 connections there are from a given address for example. See also the
14117 sc_trackers sample fetch keyword.
14118
Willy Tarreauffcb2e42014-07-10 16:29:08 +020014119upper
14120 Convert a string sample to upper case. This can only be placed after a string
14121 sample fetch function or after a transformation keyword returning a string
14122 type. The result is of type string.
14123
Thierry FOURNIER82ff3c92015-05-07 15:46:20 +020014124url_dec
14125 Takes an url-encoded string provided as input and returns the decoded
14126 version as output. The input and the output are of type string.
14127
Frédéric Lécaille756d97f2019-03-04 19:03:48 +010014128ungrpc(<field_number>,[<field_type>])
Frédéric Lécaille50290fb2019-02-27 14:34:51 +010014129 This extracts the protocol buffers message field in raw mode of an input binary
Frédéric Lécaillebfe61382019-03-06 14:34:36 +010014130 sample representation of a gRPC message with <field_number> as field number
14131 (dotted notation) if <field_type> is not present, or as an integer sample if this
14132 field is present.
Frédéric Lécaille756d97f2019-03-04 19:03:48 +010014133 The list of the authorized types is the following one: "int32", "int64", "uint32",
14134 "uint64", "sint32", "sint64", "bool", "enum" for the "varint" wire type 0
14135 "fixed64", "sfixed64", "double" for the 64bit wire type 1, "fixed32", "sfixed32",
14136 "float" for the wire type 5. Note that "string" is considered as a length-delimited
Frédéric Lécaille93d33162019-03-06 09:35:59 +010014137 type, so it does not require any <field_type> argument to be extracted.
Frédéric Lécaille756d97f2019-03-04 19:03:48 +010014138 More information may be found here about the protocol buffers message field types:
14139 https://developers.google.com/protocol-buffers/docs/encoding
Frédéric Lécaille50290fb2019-02-27 14:34:51 +010014140
14141 Example:
14142 // with such a protocol buffer .proto file content adapted from
14143 // https://github.com/grpc/grpc/blob/master/examples/protos/route_guide.proto
14144
14145 message Point {
14146 int32 latitude = 1;
14147 int32 longitude = 2;
14148 }
14149
14150 message PPoint {
14151 Point point = 59;
14152 }
14153
14154 message Rectangle {
14155 // One corner of the rectangle.
14156 PPoint lo = 48;
14157 // The other corner of the rectangle.
14158 PPoint hi = 49;
14159 }
14160
14161 let's say a body request is made of a "Rectangle" object value (two PPoint
14162 protocol buffers messages), the four protocol buffers fields could be
14163 extracted with these "ungrpc" directives:
14164
Frédéric Lécaille756d97f2019-03-04 19:03:48 +010014165 req.body,ungrpc(48.59.1,int32) # "latitude" of "lo" first PPoint
14166 req.body,ungrpc(48.59.2,int32) # "longitude" of "lo" first PPoint
John Roeslerfb2fce12019-07-10 15:45:51 -050014167 req.body,ungrpc(49.59.1,int32) # "latitude" of "hi" second PPoint
Frédéric Lécaille756d97f2019-03-04 19:03:48 +010014168 req.body,ungrpc(49.59.2,int32) # "longitude" of "hi" second PPoint
14169
Frédéric Lécaille93d33162019-03-06 09:35:59 +010014170 We could also extract the intermediary 48.59 field as a binary sample as follows:
Frédéric Lécaille756d97f2019-03-04 19:03:48 +010014171
Frédéric Lécaille93d33162019-03-06 09:35:59 +010014172 req.body,ungrpc(48.59)
Frédéric Lécaille756d97f2019-03-04 19:03:48 +010014173
John Roeslerfb2fce12019-07-10 15:45:51 -050014174 As a gRPC message is always made of a gRPC header followed by protocol buffers
Frédéric Lécaillebfe61382019-03-06 14:34:36 +010014175 messages, in the previous example the "latitude" of "lo" first PPoint
14176 could be extracted with these equivalent directives:
14177
14178 req.body,ungrpc(48.59),protobuf(1,int32)
14179 req.body,ungrpc(48),protobuf(59.1,int32)
14180 req.body,ungrpc(48),protobuf(59),protobuf(1,int32)
14181
14182 Note that the first convert must be "ungrpc", the remaining ones must be
14183 "protobuf" and only the last one may have or not a second argument to
14184 interpret the previous binary sample.
14185
Frédéric Lécaille50290fb2019-02-27 14:34:51 +010014186
Christopher Faulet85d79c92016-11-09 16:54:56 +010014187unset-var(<var name>)
14188 Unsets a variable if the input content is defined. The name of the variable
14189 starts with an indication about its scope. The scopes allowed are:
14190 "proc" : the variable is shared with the whole process
14191 "sess" : the variable is shared with the whole session
14192 "txn" : the variable is shared with the transaction (request and
14193 response),
14194 "req" : the variable is shared only during request processing,
14195 "res" : the variable is shared only during response processing.
14196 This prefix is followed by a name. The separator is a '.'. The name may only
14197 contain characters 'a-z', 'A-Z', '0-9', '.' and '_'.
14198
Willy Tarreau0dbfdba2014-07-10 16:37:47 +020014199utime(<format>[,<offset>])
14200 Converts an integer supposed to contain a date since epoch to a string
14201 representing this date in UTC time using a format defined by the <format>
14202 string using strftime(3). The purpose is to allow any date format to be used
14203 in logs. An optional <offset> in seconds may be applied to the input date
14204 (positive or negative). See the strftime() man page for the format supported
14205 by your operating system. See also the ltime converter.
14206
14207 Example :
14208
14209 # Emit two colons, one with the UTC time and another with ip:port
Davor Ocelice9ed2812017-12-25 17:49:28 +010014210 # e.g. 20140710162350 127.0.0.1:57325
Willy Tarreau0dbfdba2014-07-10 16:37:47 +020014211 log-format %[date,utime(%Y%m%d%H%M%S)]\ %ci:%cp
14212
Marcin Deranek9631a282018-04-16 14:30:46 +020014213word(<index>,<delimiters>[,<count>])
14214 Extracts the nth word counting from the beginning (positive index) or from
14215 the end (negative index) considering given delimiters from an input string.
14216 Indexes start at 1 or -1 and delimiters are a string formatted list of chars.
14217 Optionally you can specify <count> of words to extract (default: 1).
14218 Value of 0 indicates extraction of all remaining words.
14219
14220 Example :
14221 str(f1_f2_f3__f5),word(4,_) # f5
14222 str(f1_f2_f3__f5),word(2,_,0) # f2_f3__f5
14223 str(f1_f2_f3__f5),word(3,_,2) # f3__f5
14224 str(f1_f2_f3__f5),word(-2,_,3) # f1_f2_f3
14225 str(f1_f2_f3__f5),word(-3,_,0) # f1_f2
Emeric Brunc9a0f6d2014-11-25 14:09:01 +010014226
Willy Tarreau23ec4ca2014-07-15 20:15:37 +020014227wt6([<avalanche>])
14228 Hashes a binary input sample into an unsigned 32-bit quantity using the WT6
14229 hash function. Optionally, it is possible to apply a full avalanche hash
14230 function to the output if the optional <avalanche> argument equals 1. This
14231 converter uses the same functions as used by the various hash-based load
14232 balancing algorithms, so it will provide exactly the same results. It is
14233 mostly intended for debugging, but can be used as a stick-table entry to
14234 collect rough statistics. It must not be used for security purposes as a
Emmanuel Hocdet50791a72018-03-21 11:19:01 +010014235 32-bit hash is trivial to break. See also "crc32", "djb2", "sdbm", "crc32c",
14236 and the "hash-type" directive.
Willy Tarreau23ec4ca2014-07-15 20:15:37 +020014237
Willy Tarreau97707872015-01-27 15:12:13 +010014238xor(<value>)
14239 Performs a bitwise "XOR" (exclusive OR) between <value> and the input value
Thierry FOURNIER07ee64e2015-07-06 23:43:03 +020014240 of type signed integer, and returns the result as an signed integer.
Thierry FOURNIER5d86fae2015-07-07 21:10:16 +020014241 <value> can be a numeric value or a variable name. The name of the variable
Daniel Schneller0b547052016-03-21 20:46:57 +010014242 starts with an indication about its scope. The scopes allowed are:
Christopher Fauletff2613e2016-11-09 11:36:17 +010014243 "proc" : the variable is shared with the whole process
Daniel Schneller0b547052016-03-21 20:46:57 +010014244 "sess" : the variable is shared with the whole session
14245 "txn" : the variable is shared with the transaction (request and
Thierry FOURNIER5d86fae2015-07-07 21:10:16 +020014246 response),
Daniel Schneller0b547052016-03-21 20:46:57 +010014247 "req" : the variable is shared only during request processing,
14248 "res" : the variable is shared only during response processing.
Thierry FOURNIER5d86fae2015-07-07 21:10:16 +020014249 This prefix is followed by a name. The separator is a '.'. The name may only
Christopher Fauletb71557a2016-10-31 10:49:03 +010014250 contain characters 'a-z', 'A-Z', '0-9', '.' and '_'.
Willy Tarreau97707872015-01-27 15:12:13 +010014251
Thierry FOURNIER01e09742016-12-26 11:46:11 +010014252xxh32([<seed>])
14253 Hashes a binary input sample into an unsigned 32-bit quantity using the 32-bit
14254 variant of the XXHash hash function. This hash supports a seed which defaults
14255 to zero but a different value maybe passed as the <seed> argument. This hash
14256 is known to be very good and very fast so it can be used to hash URLs and/or
14257 URL parameters for use as stick-table keys to collect statistics with a low
14258 collision rate, though care must be taken as the algorithm is not considered
14259 as cryptographically secure.
14260
14261xxh64([<seed>])
14262 Hashes a binary input sample into a signed 64-bit quantity using the 64-bit
14263 variant of the XXHash hash function. This hash supports a seed which defaults
14264 to zero but a different value maybe passed as the <seed> argument. This hash
14265 is known to be very good and very fast so it can be used to hash URLs and/or
14266 URL parameters for use as stick-table keys to collect statistics with a low
14267 collision rate, though care must be taken as the algorithm is not considered
14268 as cryptographically secure.
14269
Thierry FOURNIERd5f624d2013-11-26 11:52:33 +010014270
Thierry FOURNIER060762e2014-04-23 13:29:15 +0200142717.3.2. Fetching samples from internal states
Willy Tarreau74ca5042013-06-11 23:12:07 +020014272--------------------------------------------
14273
14274A first set of sample fetch methods applies to internal information which does
14275not even relate to any client information. These ones are sometimes used with
14276"monitor-fail" directives to report an internal status to external watchers.
14277The sample fetch methods described in this section are usable anywhere.
14278
14279always_false : boolean
14280 Always returns the boolean "false" value. It may be used with ACLs as a
14281 temporary replacement for another one when adjusting configurations.
14282
14283always_true : boolean
14284 Always returns the boolean "true" value. It may be used with ACLs as a
14285 temporary replacement for another one when adjusting configurations.
14286
14287avg_queue([<backend>]) : integer
Willy Tarreaud63335a2010-02-26 12:56:52 +010014288 Returns the total number of queued connections of the designated backend
Willy Tarreau74ca5042013-06-11 23:12:07 +020014289 divided by the number of active servers. The current backend is used if no
14290 backend is specified. This is very similar to "queue" except that the size of
14291 the farm is considered, in order to give a more accurate measurement of the
14292 time it may take for a new connection to be processed. The main usage is with
14293 ACL to return a sorry page to new users when it becomes certain they will get
14294 a degraded service, or to pass to the backend servers in a header so that
14295 they decide to work in degraded mode or to disable some functions to speed up
14296 the processing a bit. Note that in the event there would not be any active
14297 server anymore, twice the number of queued connections would be considered as
14298 the measured value. This is a fair estimate, as we expect one server to get
14299 back soon anyway, but we still prefer to send new traffic to another backend
14300 if in better shape. See also the "queue", "be_conn", and "be_sess_rate"
14301 sample fetches.
Krzysztof Piotr Oledzki346f76d2010-01-12 21:59:30 +010014302
Willy Tarreau74ca5042013-06-11 23:12:07 +020014303be_conn([<backend>]) : integer
Willy Tarreaua36af912009-10-10 12:02:45 +020014304 Applies to the number of currently established connections on the backend,
14305 possibly including the connection being evaluated. If no backend name is
14306 specified, the current one is used. But it is also possible to check another
14307 backend. It can be used to use a specific farm when the nominal one is full.
Patrick Hemmer4cdf3ab2018-06-14 17:10:27 -040014308 See also the "fe_conn", "queue", "be_conn_free", and "be_sess_rate" criteria.
14309
14310be_conn_free([<backend>]) : integer
14311 Returns an integer value corresponding to the number of available connections
14312 across available servers in the backend. Queue slots are not included. Backup
14313 servers are also not included, unless all other servers are down. If no
14314 backend name is specified, the current one is used. But it is also possible
14315 to check another backend. It can be used to use a specific farm when the
Patrick Hemmer155e93e2018-06-14 18:01:35 -040014316 nominal one is full. See also the "be_conn", "connslots", and "srv_conn_free"
14317 criteria.
Patrick Hemmer4cdf3ab2018-06-14 17:10:27 -040014318
14319 OTHER CAVEATS AND NOTES: if any of the server maxconn, or maxqueue is 0
14320 (meaning unlimited), then this fetch clearly does not make sense, in which
14321 case the value returned will be -1.
Willy Tarreau6a06a402007-07-15 20:15:28 +020014322
Willy Tarreau74ca5042013-06-11 23:12:07 +020014323be_sess_rate([<backend>]) : integer
14324 Returns an integer value corresponding to the sessions creation rate on the
14325 backend, in number of new sessions per second. This is used with ACLs to
14326 switch to an alternate backend when an expensive or fragile one reaches too
Davor Ocelice9ed2812017-12-25 17:49:28 +010014327 high a session rate, or to limit abuse of service (e.g. prevent sucking of an
Willy Tarreau74ca5042013-06-11 23:12:07 +020014328 online dictionary). It can also be useful to add this element to logs using a
14329 log-format directive.
Willy Tarreaud63335a2010-02-26 12:56:52 +010014330
14331 Example :
14332 # Redirect to an error page if the dictionary is requested too often
14333 backend dynamic
14334 mode http
14335 acl being_scanned be_sess_rate gt 100
14336 redirect location /denied.html if being_scanned
Willy Tarreau0ba27502007-12-24 16:55:16 +010014337
Davor Ocelice9ed2812017-12-25 17:49:28 +010014338bin(<hex>) : bin
Thierry FOURNIERcc103292015-06-06 19:30:17 +020014339 Returns a binary chain. The input is the hexadecimal representation
14340 of the string.
14341
14342bool(<bool>) : bool
14343 Returns a boolean value. <bool> can be 'true', 'false', '1' or '0'.
14344 'false' and '0' are the same. 'true' and '1' are the same.
14345
Willy Tarreau74ca5042013-06-11 23:12:07 +020014346connslots([<backend>]) : integer
14347 Returns an integer value corresponding to the number of connection slots
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030014348 still available in the backend, by totaling the maximum amount of
Willy Tarreau74ca5042013-06-11 23:12:07 +020014349 connections on all servers and the maximum queue size. This is probably only
14350 used with ACLs.
Tait Clarridge7896d522012-12-05 21:39:31 -050014351
Jeffrey 'jf' Lim5051d7b2008-09-04 01:03:03 +080014352 The basic idea here is to be able to measure the number of connection "slots"
Willy Tarreau55165fe2009-05-10 12:02:55 +020014353 still available (connection + queue), so that anything beyond that (intended
Jeffrey 'jf' Lim5051d7b2008-09-04 01:03:03 +080014354 usage; see "use_backend" keyword) can be redirected to a different backend.
14355
Willy Tarreau55165fe2009-05-10 12:02:55 +020014356 'connslots' = number of available server connection slots, + number of
14357 available server queue slots.
Jeffrey 'jf' Lim5051d7b2008-09-04 01:03:03 +080014358
Willy Tarreaua36af912009-10-10 12:02:45 +020014359 Note that while "fe_conn" may be used, "connslots" comes in especially
Willy Tarreau55165fe2009-05-10 12:02:55 +020014360 useful when you have a case of traffic going to one single ip, splitting into
Willy Tarreau74ca5042013-06-11 23:12:07 +020014361 multiple backends (perhaps using ACLs to do name-based load balancing) and
Willy Tarreau55165fe2009-05-10 12:02:55 +020014362 you want to be able to differentiate between different backends, and their
Davor Ocelice9ed2812017-12-25 17:49:28 +010014363 available "connslots". Also, whereas "nbsrv" only measures servers that are
Willy Tarreau74ca5042013-06-11 23:12:07 +020014364 actually *down*, this fetch is more fine-grained and looks into the number of
Willy Tarreaua36af912009-10-10 12:02:45 +020014365 available connection slots as well. See also "queue" and "avg_queue".
Jeffrey 'jf' Lim5051d7b2008-09-04 01:03:03 +080014366
Willy Tarreau55165fe2009-05-10 12:02:55 +020014367 OTHER CAVEATS AND NOTES: at this point in time, the code does not take care
14368 of dynamic connections. Also, if any of the server maxconn, or maxqueue is 0,
Willy Tarreau74ca5042013-06-11 23:12:07 +020014369 then this fetch clearly does not make sense, in which case the value returned
Willy Tarreau55165fe2009-05-10 12:02:55 +020014370 will be -1.
Jeffrey 'jf' Lim5051d7b2008-09-04 01:03:03 +080014371
Willy Tarreau70fe9442018-11-22 16:07:39 +010014372cpu_calls : integer
14373 Returns the number of calls to the task processing the stream or current
14374 request since it was allocated. This number is reset for each new request on
14375 the same connections in case of HTTP keep-alive. This value should usually be
14376 low and stable (around 2 calls for a typically simple request) but may become
14377 high if some processing (compression, caching or analysis) is performed. This
14378 is purely for performance monitoring purposes.
14379
14380cpu_ns_avg : integer
14381 Returns the average number of nanoseconds spent in each call to the task
14382 processing the stream or current request. This number is reset for each new
14383 request on the same connections in case of HTTP keep-alive. This value
14384 indicates the overall cost of processing the request or the connection for
14385 each call. There is no good nor bad value but the time spent in a call
14386 automatically causes latency for other processing (see lat_ns_avg below),
14387 and may affect other connection's apparent response time. Certain operations
14388 like compression, complex regex matching or heavy Lua operations may directly
14389 affect this value, and having it in the logs will make it easier to spot the
14390 faulty processing that needs to be fixed to recover decent performance.
14391 Note: this value is exactly cpu_ns_tot divided by cpu_calls.
14392
14393cpu_ns_tot : integer
14394 Returns the total number of nanoseconds spent in each call to the task
14395 processing the stream or current request. This number is reset for each new
14396 request on the same connections in case of HTTP keep-alive. This value
14397 indicates the overall cost of processing the request or the connection for
14398 each call. There is no good nor bad value but the time spent in a call
14399 automatically causes latency for other processing (see lat_ns_avg below),
14400 induces CPU costs on the machine, and may affect other connection's apparent
14401 response time. Certain operations like compression, complex regex matching or
14402 heavy Lua operations may directly affect this value, and having it in the
14403 logs will make it easier to spot the faulty processing that needs to be fixed
14404 to recover decent performance. The value may be artificially high due to a
14405 high cpu_calls count, for example when processing many HTTP chunks, and for
14406 this reason it is often preferred to log cpu_ns_avg instead.
14407
Willy Tarreau6236d3a2013-07-25 14:28:25 +020014408date([<offset>]) : integer
14409 Returns the current date as the epoch (number of seconds since 01/01/1970).
14410 If an offset value is specified, then it is a number of seconds that is added
14411 to the current date before returning the value. This is particularly useful
14412 to compute relative dates, as both positive and negative offsets are allowed.
Willy Tarreau276fae92013-07-25 14:36:01 +020014413 It is useful combined with the http_date converter.
14414
14415 Example :
14416
14417 # set an expires header to now+1 hour in every response
14418 http-response set-header Expires %[date(3600),http_date]
Willy Tarreau6236d3a2013-07-25 14:28:25 +020014419
Etienne Carrierea792a0a2018-01-17 13:43:24 +010014420date_us : integer
14421 Return the microseconds part of the date (the "second" part is returned by
14422 date sample). This sample is coherent with the date sample as it is comes
14423 from the same timeval structure.
14424
Willy Tarreaud716f9b2017-10-13 11:03:15 +020014425distcc_body(<token>[,<occ>]) : binary
14426 Parses a distcc message and returns the body associated to occurrence #<occ>
14427 of the token <token>. Occurrences start at 1, and when unspecified, any may
14428 match though in practice only the first one is checked for now. This can be
14429 used to extract file names or arguments in files built using distcc through
14430 haproxy. Please refer to distcc's protocol documentation for the complete
14431 list of supported tokens.
14432
14433distcc_param(<token>[,<occ>]) : integer
14434 Parses a distcc message and returns the parameter associated to occurrence
14435 #<occ> of the token <token>. Occurrences start at 1, and when unspecified,
14436 any may match though in practice only the first one is checked for now. This
14437 can be used to extract certain information such as the protocol version, the
14438 file size or the argument in files built using distcc through haproxy.
14439 Another use case consists in waiting for the start of the preprocessed file
14440 contents before connecting to the server to avoid keeping idle connections.
14441 Please refer to distcc's protocol documentation for the complete list of
14442 supported tokens.
14443
14444 Example :
14445 # wait up to 20s for the pre-processed file to be uploaded
14446 tcp-request inspect-delay 20s
14447 tcp-request content accept if { distcc_param(DOTI) -m found }
14448 # send large files to the big farm
14449 use_backend big_farm if { distcc_param(DOTI) gt 1000000 }
14450
Willy Tarreau595ec542013-06-12 21:34:28 +020014451env(<name>) : string
14452 Returns a string containing the value of environment variable <name>. As a
14453 reminder, environment variables are per-process and are sampled when the
14454 process starts. This can be useful to pass some information to a next hop
14455 server, or with ACLs to take specific action when the process is started a
14456 certain way.
14457
14458 Examples :
14459 # Pass the Via header to next hop with the local hostname in it
14460 http-request add-header Via 1.1\ %[env(HOSTNAME)]
14461
14462 # reject cookie-less requests when the STOP environment variable is set
14463 http-request deny if !{ cook(SESSIONID) -m found } { env(STOP) -m found }
14464
Willy Tarreau74ca5042013-06-11 23:12:07 +020014465fe_conn([<frontend>]) : integer
14466 Returns the number of currently established connections on the frontend,
Willy Tarreaud63335a2010-02-26 12:56:52 +010014467 possibly including the connection being evaluated. If no frontend name is
14468 specified, the current one is used. But it is also possible to check another
Willy Tarreau74ca5042013-06-11 23:12:07 +020014469 frontend. It can be used to return a sorry page before hard-blocking, or to
14470 use a specific backend to drain new requests when the farm is considered
Davor Ocelice9ed2812017-12-25 17:49:28 +010014471 full. This is mostly used with ACLs but can also be used to pass some
Willy Tarreau74ca5042013-06-11 23:12:07 +020014472 statistics to servers in HTTP headers. See also the "dst_conn", "be_conn",
14473 "fe_sess_rate" fetches.
Willy Tarreaua36af912009-10-10 12:02:45 +020014474
Nenad Merdanovicad9a7e92016-10-03 04:57:37 +020014475fe_req_rate([<frontend>]) : integer
14476 Returns an integer value corresponding to the number of HTTP requests per
14477 second sent to a frontend. This number can differ from "fe_sess_rate" in
14478 situations where client-side keep-alive is enabled.
14479
Willy Tarreau74ca5042013-06-11 23:12:07 +020014480fe_sess_rate([<frontend>]) : integer
14481 Returns an integer value corresponding to the sessions creation rate on the
14482 frontend, in number of new sessions per second. This is used with ACLs to
14483 limit the incoming session rate to an acceptable range in order to prevent
14484 abuse of service at the earliest moment, for example when combined with other
14485 layer 4 ACLs in order to force the clients to wait a bit for the rate to go
14486 down below the limit. It can also be useful to add this element to logs using
14487 a log-format directive. See also the "rate-limit sessions" directive for use
14488 in frontends.
Willy Tarreau079ff0a2009-03-05 21:34:28 +010014489
14490 Example :
14491 # This frontend limits incoming mails to 10/s with a max of 100
14492 # concurrent connections. We accept any connection below 10/s, and
14493 # force excess clients to wait for 100 ms. Since clients are limited to
14494 # 100 max, there cannot be more than 10 incoming mails per second.
14495 frontend mail
14496 bind :25
14497 mode tcp
14498 maxconn 100
14499 acl too_fast fe_sess_rate ge 10
14500 tcp-request inspect-delay 100ms
14501 tcp-request content accept if ! too_fast
14502 tcp-request content accept if WAIT_END
Willy Tarreaud72758d2010-01-12 10:42:19 +010014503
Nenad Merdanovic807a6e72017-03-12 22:00:00 +010014504hostname : string
14505 Returns the system hostname.
14506
Thierry FOURNIER07ee64e2015-07-06 23:43:03 +020014507int(<integer>) : signed integer
14508 Returns a signed integer.
14509
Thierry FOURNIERcc103292015-06-06 19:30:17 +020014510ipv4(<ipv4>) : ipv4
14511 Returns an ipv4.
14512
14513ipv6(<ipv6>) : ipv6
14514 Returns an ipv6.
14515
Willy Tarreau70fe9442018-11-22 16:07:39 +010014516lat_ns_avg : integer
14517 Returns the average number of nanoseconds spent between the moment the task
14518 handling the stream is woken up and the moment it is effectively called. This
14519 number is reset for each new request on the same connections in case of HTTP
14520 keep-alive. This value indicates the overall latency inflicted to the current
14521 request by all other requests being processed in parallel, and is a direct
14522 indicator of perceived performance due to noisy neighbours. In order to keep
14523 the value low, it is possible to reduce the scheduler's run queue depth using
14524 "tune.runqueue-depth", to reduce the number of concurrent events processed at
14525 once using "tune.maxpollevents", to decrease the stream's nice value using
14526 the "nice" option on the "bind" lines or in the frontend, or to look for
14527 other heavy requests in logs (those exhibiting large values of "cpu_ns_avg"),
14528 whose processing needs to be adjusted or fixed. Compression of large buffers
14529 could be a culprit, like heavy regex or long lists of regex.
14530 Note: this value is exactly lat_ns_tot divided by cpu_calls.
14531
14532lat_ns_tot : integer
14533 Returns the total number of nanoseconds spent between the moment the task
14534 handling the stream is woken up and the moment it is effectively called. This
14535 number is reset for each new request on the same connections in case of HTTP
14536 keep-alive. This value indicates the overall latency inflicted to the current
14537 request by all other requests being processed in parallel, and is a direct
14538 indicator of perceived performance due to noisy neighbours. In order to keep
14539 the value low, it is possible to reduce the scheduler's run queue depth using
14540 "tune.runqueue-depth", to reduce the number of concurrent events processed at
14541 once using "tune.maxpollevents", to decrease the stream's nice value using
14542 the "nice" option on the "bind" lines or in the frontend, or to look for
14543 other heavy requests in logs (those exhibiting large values of "cpu_ns_avg"),
14544 whose processing needs to be adjusted or fixed. Compression of large buffers
14545 could be a culprit, like heavy regex or long lists of regex. Note: while it
14546 may intuitively seem that the total latency adds to a transfer time, it is
14547 almost never true because while a task waits for the CPU, network buffers
14548 continue to fill up and the next call will process more at once. The value
14549 may be artificially high due to a high cpu_calls count, for example when
14550 processing many HTTP chunks, and for this reason it is often preferred to log
14551 lat_ns_avg instead, which is a more relevant performance indicator.
14552
Thierry FOURNIERcc103292015-06-06 19:30:17 +020014553meth(<method>) : method
14554 Returns a method.
14555
Willy Tarreau0f30d262014-11-24 16:02:05 +010014556nbproc : integer
14557 Returns an integer value corresponding to the number of processes that were
14558 started (it equals the global "nbproc" setting). This is useful for logging
14559 and debugging purposes.
14560
Willy Tarreau74ca5042013-06-11 23:12:07 +020014561nbsrv([<backend>]) : integer
14562 Returns an integer value corresponding to the number of usable servers of
14563 either the current backend or the named backend. This is mostly used with
14564 ACLs but can also be useful when added to logs. This is normally used to
Willy Tarreaud63335a2010-02-26 12:56:52 +010014565 switch to an alternate backend when the number of servers is too low to
14566 to handle some load. It is useful to report a failure when combined with
14567 "monitor fail".
Willy Tarreau079ff0a2009-03-05 21:34:28 +010014568
Patrick Hemmerfabb24f2018-08-13 14:07:57 -040014569prio_class : integer
14570 Returns the priority class of the current session for http mode or connection
14571 for tcp mode. The value will be that set by the last call to "http-request
14572 set-priority-class" or "tcp-request content set-priority-class".
14573
14574prio_offset : integer
14575 Returns the priority offset of the current session for http mode or
14576 connection for tcp mode. The value will be that set by the last call to
14577 "http-request set-priority-offset" or "tcp-request content
14578 set-priority-offset".
14579
Willy Tarreau0f30d262014-11-24 16:02:05 +010014580proc : integer
14581 Returns an integer value corresponding to the position of the process calling
14582 the function, between 1 and global.nbproc. This is useful for logging and
14583 debugging purposes.
14584
Willy Tarreau74ca5042013-06-11 23:12:07 +020014585queue([<backend>]) : integer
Willy Tarreaud63335a2010-02-26 12:56:52 +010014586 Returns the total number of queued connections of the designated backend,
14587 including all the connections in server queues. If no backend name is
14588 specified, the current one is used, but it is also possible to check another
Willy Tarreau74ca5042013-06-11 23:12:07 +020014589 one. This is useful with ACLs or to pass statistics to backend servers. This
14590 can be used to take actions when queuing goes above a known level, generally
14591 indicating a surge of traffic or a massive slowdown on the servers. One
14592 possible action could be to reject new users but still accept old ones. See
14593 also the "avg_queue", "be_conn", and "be_sess_rate" fetches.
14594
Willy Tarreau84310e22014-02-14 11:59:04 +010014595rand([<range>]) : integer
14596 Returns a random integer value within a range of <range> possible values,
14597 starting at zero. If the range is not specified, it defaults to 2^32, which
14598 gives numbers between 0 and 4294967295. It can be useful to pass some values
14599 needed to take some routing decisions for example, or just for debugging
14600 purposes. This random must not be used for security purposes.
14601
Willy Tarreau74ca5042013-06-11 23:12:07 +020014602srv_conn([<backend>/]<server>) : integer
14603 Returns an integer value corresponding to the number of currently established
14604 connections on the designated server, possibly including the connection being
14605 evaluated. If <backend> is omitted, then the server is looked up in the
14606 current backend. It can be used to use a specific farm when one server is
14607 full, or to inform the server about our view of the number of active
Patrick Hemmer155e93e2018-06-14 18:01:35 -040014608 connections with it. See also the "fe_conn", "be_conn", "queue", and
14609 "srv_conn_free" fetch methods.
14610
14611srv_conn_free([<backend>/]<server>) : integer
14612 Returns an integer value corresponding to the number of available connections
14613 on the designated server, possibly including the connection being evaluated.
14614 The value does not include queue slots. If <backend> is omitted, then the
14615 server is looked up in the current backend. It can be used to use a specific
14616 farm when one server is full, or to inform the server about our view of the
14617 number of active connections with it. See also the "be_conn_free" and
14618 "srv_conn" fetch methods.
14619
14620 OTHER CAVEATS AND NOTES: If the server maxconn is 0, then this fetch clearly
14621 does not make sense, in which case the value returned will be -1.
Willy Tarreau74ca5042013-06-11 23:12:07 +020014622
14623srv_is_up([<backend>/]<server>) : boolean
14624 Returns true when the designated server is UP, and false when it is either
14625 DOWN or in maintenance mode. If <backend> is omitted, then the server is
14626 looked up in the current backend. It is mainly used to take action based on
Davor Ocelice9ed2812017-12-25 17:49:28 +010014627 an external status reported via a health check (e.g. a geographical site's
Willy Tarreau74ca5042013-06-11 23:12:07 +020014628 availability). Another possible use which is more of a hack consists in
14629 using dummy servers as boolean variables that can be enabled or disabled from
14630 the CLI, so that rules depending on those ACLs can be tweaked in realtime.
14631
Willy Tarreauff2b7af2017-10-13 11:46:26 +020014632srv_queue([<backend>/]<server>) : integer
14633 Returns an integer value corresponding to the number of connections currently
14634 pending in the designated server's queue. If <backend> is omitted, then the
14635 server is looked up in the current backend. It can sometimes be used together
14636 with the "use-server" directive to force to use a known faster server when it
14637 is not much loaded. See also the "srv_conn", "avg_queue" and "queue" sample
14638 fetch methods.
14639
Willy Tarreau74ca5042013-06-11 23:12:07 +020014640srv_sess_rate([<backend>/]<server>) : integer
14641 Returns an integer corresponding to the sessions creation rate on the
14642 designated server, in number of new sessions per second. If <backend> is
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030014643 omitted, then the server is looked up in the current backend. This is mostly
Willy Tarreau74ca5042013-06-11 23:12:07 +020014644 used with ACLs but can make sense with logs too. This is used to switch to an
14645 alternate backend when an expensive or fragile one reaches too high a session
Davor Ocelice9ed2812017-12-25 17:49:28 +010014646 rate, or to limit abuse of service (e.g. prevent latent requests from
Willy Tarreau74ca5042013-06-11 23:12:07 +020014647 overloading servers).
14648
14649 Example :
14650 # Redirect to a separate back
14651 acl srv1_full srv_sess_rate(be1/srv1) gt 50
14652 acl srv2_full srv_sess_rate(be1/srv2) gt 50
14653 use_backend be2 if srv1_full or srv2_full
14654
Willy Tarreau0f30d262014-11-24 16:02:05 +010014655stopping : boolean
14656 Returns TRUE if the process calling the function is currently stopping. This
14657 can be useful for logging, or for relaxing certain checks or helping close
14658 certain connections upon graceful shutdown.
14659
Thierry FOURNIERcc103292015-06-06 19:30:17 +020014660str(<string>) : string
14661 Returns a string.
14662
Willy Tarreau74ca5042013-06-11 23:12:07 +020014663table_avl([<table>]) : integer
14664 Returns the total number of available entries in the current proxy's
14665 stick-table or in the designated stick-table. See also table_cnt.
14666
14667table_cnt([<table>]) : integer
14668 Returns the total number of entries currently in use in the current proxy's
14669 stick-table or in the designated stick-table. See also src_conn_cnt and
14670 table_avl for other entry counting methods.
14671
Christopher Faulet34adb2a2017-11-21 21:45:38 +010014672thread : integer
14673 Returns an integer value corresponding to the position of the thread calling
14674 the function, between 0 and (global.nbthread-1). This is useful for logging
14675 and debugging purposes.
14676
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020014677var(<var-name>) : undefined
14678 Returns a variable with the stored type. If the variable is not set, the
Daniel Schneller0b547052016-03-21 20:46:57 +010014679 sample fetch fails. The name of the variable starts with an indication
14680 about its scope. The scopes allowed are:
Christopher Fauletff2613e2016-11-09 11:36:17 +010014681 "proc" : the variable is shared with the whole process
Daniel Schneller0b547052016-03-21 20:46:57 +010014682 "sess" : the variable is shared with the whole session
14683 "txn" : the variable is shared with the transaction (request and
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020014684 response),
Daniel Schneller0b547052016-03-21 20:46:57 +010014685 "req" : the variable is shared only during request processing,
14686 "res" : the variable is shared only during response processing.
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020014687 This prefix is followed by a name. The separator is a '.'. The name may only
Christopher Fauletb71557a2016-10-31 10:49:03 +010014688 contain characters 'a-z', 'A-Z', '0-9', '.' and '_'.
Thierry FOURNIER4834bc72015-06-06 19:29:07 +020014689
Thierry FOURNIER060762e2014-04-23 13:29:15 +0200146907.3.3. Fetching samples at Layer 4
Willy Tarreau74ca5042013-06-11 23:12:07 +020014691----------------------------------
14692
14693The layer 4 usually describes just the transport layer which in haproxy is
14694closest to the connection, where no content is yet made available. The fetch
14695methods described here are usable as low as the "tcp-request connection" rule
14696sets unless they require some future information. Those generally include
14697TCP/IP addresses and ports, as well as elements from stick-tables related to
Willy Tarreau4d4149c2013-07-23 19:33:46 +020014698the incoming connection. For retrieving a value from a sticky counters, the
14699counter number can be explicitly set as 0, 1, or 2 using the pre-defined
Moemen MHEDHBI9cf46342018-09-25 17:50:53 +020014700"sc0_", "sc1_", or "sc2_" prefix. These three pre-defined prefixes can only be
14701used if MAX_SESS_STKCTR value does not exceed 3, otherwise the counter number
14702can be specified as the first integer argument when using the "sc_" prefix.
14703Starting from "sc_0" to "sc_N" where N is (MAX_SESS_STKCTR-1). An optional
14704table may be specified with the "sc*" form, in which case the currently
14705tracked key will be looked up into this alternate table instead of the table
14706currently being tracked.
Willy Tarreau74ca5042013-06-11 23:12:07 +020014707
Jérôme Magnin35e53a62019-01-16 14:38:37 +010014708bc_http_major : integer
Jérôme Magnin86577422018-12-07 09:03:11 +010014709 Returns the backend connection's HTTP major version encoding, which may be 1
14710 for HTTP/0.9 to HTTP/1.1 or 2 for HTTP/2. Note, this is based on the on-wire
14711 encoding and not the version present in the request header.
14712
Willy Tarreau74ca5042013-06-11 23:12:07 +020014713be_id : integer
14714 Returns an integer containing the current backend's id. It can be used in
14715 frontends with responses to check which backend processed the request.
14716
Marcin Deranekd2471c22016-12-12 14:08:05 +010014717be_name : string
14718 Returns a string containing the current backend's name. It can be used in
14719 frontends with responses to check which backend processed the request.
14720
Willy Tarreau74ca5042013-06-11 23:12:07 +020014721dst : ip
14722 This is the destination IPv4 address of the connection on the client side,
14723 which is the address the client connected to. It can be useful when running
14724 in transparent mode. It is of type IP and works on both IPv4 and IPv6 tables.
14725 On IPv6 tables, IPv4 address is mapped to its IPv6 equivalent, according to
Willy Tarreau64ded3d2019-01-23 10:02:15 +010014726 RFC 4291. When the incoming connection passed through address translation or
14727 redirection involving connection tracking, the original destination address
14728 before the redirection will be reported. On Linux systems, the source and
14729 destination may seldom appear reversed if the nf_conntrack_tcp_loose sysctl
14730 is set, because a late response may reopen a timed out connection and switch
14731 what is believed to be the source and the destination.
Willy Tarreau74ca5042013-06-11 23:12:07 +020014732
14733dst_conn : integer
14734 Returns an integer value corresponding to the number of currently established
14735 connections on the same socket including the one being evaluated. It is
14736 normally used with ACLs but can as well be used to pass the information to
14737 servers in an HTTP header or in logs. It can be used to either return a sorry
14738 page before hard-blocking, or to use a specific backend to drain new requests
14739 when the socket is considered saturated. This offers the ability to assign
14740 different limits to different listening ports or addresses. See also the
14741 "fe_conn" and "be_conn" fetches.
Willy Tarreaud63335a2010-02-26 12:56:52 +010014742
Willy Tarreau16e01562016-08-09 16:46:18 +020014743dst_is_local : boolean
14744 Returns true if the destination address of the incoming connection is local
14745 to the system, or false if the address doesn't exist on the system, meaning
14746 that it was intercepted in transparent mode. It can be useful to apply
14747 certain rules by default to forwarded traffic and other rules to the traffic
Davor Ocelice9ed2812017-12-25 17:49:28 +010014748 targeting the real address of the machine. For example the stats page could
Willy Tarreau16e01562016-08-09 16:46:18 +020014749 be delivered only on this address, or SSH access could be locally redirected.
14750 Please note that the check involves a few system calls, so it's better to do
14751 it only once per connection.
14752
Willy Tarreau74ca5042013-06-11 23:12:07 +020014753dst_port : integer
14754 Returns an integer value corresponding to the destination TCP port of the
14755 connection on the client side, which is the port the client connected to.
14756 This might be used when running in transparent mode, when assigning dynamic
14757 ports to some clients for a whole application session, to stick all users to
14758 a same server, or to pass the destination port information to a server using
14759 an HTTP header.
14760
Willy Tarreau60ca10a2017-08-18 15:26:54 +020014761fc_http_major : integer
14762 Reports the front connection's HTTP major version encoding, which may be 1
14763 for HTTP/0.9 to HTTP/1.1 or 2 for HTTP/2. Note, this is based on the on-wire
14764 encoding and not on the version present in the request header.
14765
Emeric Brun4f603012017-01-05 15:11:44 +010014766fc_rcvd_proxy : boolean
14767 Returns true if the client initiated the connection with a PROXY protocol
14768 header.
14769
Thierry Fournier / OZON.IO6310bef2016-07-24 20:16:50 +020014770fc_rtt(<unit>) : integer
14771 Returns the Round Trip Time (RTT) measured by the kernel for the client
14772 connection. <unit> is facultative, by default the unit is milliseconds. <unit>
14773 can be set to "ms" for milliseconds or "us" for microseconds. If the server
14774 connection is not established, if the connection is not TCP or if the
14775 operating system does not support TCP_INFO, for example Linux kernels before
14776 2.4, the sample fetch fails.
14777
14778fc_rttvar(<unit>) : integer
14779 Returns the Round Trip Time (RTT) variance measured by the kernel for the
14780 client connection. <unit> is facultative, by default the unit is milliseconds.
14781 <unit> can be set to "ms" for milliseconds or "us" for microseconds. If the
14782 server connection is not established, if the connection is not TCP or if the
14783 operating system does not support TCP_INFO, for example Linux kernels before
14784 2.4, the sample fetch fails.
14785
Joe Williams30fcd392016-08-10 07:06:44 -070014786fc_unacked(<unit>) : integer
14787 Returns the unacked counter measured by the kernel for the client connection.
14788 If the server connection is not established, if the connection is not TCP or
14789 if the operating system does not support TCP_INFO, for example Linux kernels
14790 before 2.4, the sample fetch fails.
14791
14792fc_sacked(<unit>) : integer
14793 Returns the sacked counter measured by the kernel for the client connection.
14794 If the server connection is not established, if the connection is not TCP or
14795 if the operating system does not support TCP_INFO, for example Linux kernels
14796 before 2.4, the sample fetch fails.
14797
14798fc_retrans(<unit>) : integer
14799 Returns the retransmits counter measured by the kernel for the client
14800 connection. If the server connection is not established, if the connection is
14801 not TCP or if the operating system does not support TCP_INFO, for example
14802 Linux kernels before 2.4, the sample fetch fails.
14803
14804fc_fackets(<unit>) : integer
14805 Returns the fack counter measured by the kernel for the client
14806 connection. If the server connection is not established, if the connection is
14807 not TCP or if the operating system does not support TCP_INFO, for example
14808 Linux kernels before 2.4, the sample fetch fails.
14809
14810fc_lost(<unit>) : integer
14811 Returns the lost counter measured by the kernel for the client
14812 connection. If the server connection is not established, if the connection is
14813 not TCP or if the operating system does not support TCP_INFO, for example
14814 Linux kernels before 2.4, the sample fetch fails.
14815
14816fc_reordering(<unit>) : integer
14817 Returns the reordering counter measured by the kernel for the client
14818 connection. If the server connection is not established, if the connection is
14819 not TCP or if the operating system does not support TCP_INFO, for example
14820 Linux kernels before 2.4, the sample fetch fails.
14821
Marcin Deranek9a66dfb2018-04-13 14:37:50 +020014822fe_defbe : string
14823 Returns a string containing the frontend's default backend name. It can be
14824 used in frontends to check which backend will handle requests by default.
14825
Willy Tarreau74ca5042013-06-11 23:12:07 +020014826fe_id : integer
14827 Returns an integer containing the current frontend's id. It can be used in
Marcin Deranek6e413ed2016-12-13 12:40:01 +010014828 backends to check from which frontend it was called, or to stick all users
Willy Tarreau74ca5042013-06-11 23:12:07 +020014829 coming via a same frontend to the same server.
14830
Marcin Deranekd2471c22016-12-12 14:08:05 +010014831fe_name : string
14832 Returns a string containing the current frontend's name. It can be used in
14833 backends to check from which frontend it was called, or to stick all users
14834 coming via a same frontend to the same server.
14835
Cyril Bonté62ba8702014-04-22 23:52:25 +020014836sc_bytes_in_rate(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020014837sc0_bytes_in_rate([<table>]) : integer
14838sc1_bytes_in_rate([<table>]) : integer
14839sc2_bytes_in_rate([<table>]) : integer
Willy Tarreaue9656522010-08-17 15:40:09 +020014840 Returns the average client-to-server bytes rate from the currently tracked
14841 counters, measured in amount of bytes over the period configured in the
14842 table. See also src_bytes_in_rate.
14843
Cyril Bonté62ba8702014-04-22 23:52:25 +020014844sc_bytes_out_rate(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020014845sc0_bytes_out_rate([<table>]) : integer
14846sc1_bytes_out_rate([<table>]) : integer
14847sc2_bytes_out_rate([<table>]) : integer
Willy Tarreaue9656522010-08-17 15:40:09 +020014848 Returns the average server-to-client bytes rate from the currently tracked
14849 counters, measured in amount of bytes over the period configured in the
14850 table. See also src_bytes_out_rate.
14851
Cyril Bonté62ba8702014-04-22 23:52:25 +020014852sc_clr_gpc0(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020014853sc0_clr_gpc0([<table>]) : integer
14854sc1_clr_gpc0([<table>]) : integer
14855sc2_clr_gpc0([<table>]) : integer
Willy Tarreauf73cd112011-08-13 01:45:16 +020014856 Clears the first General Purpose Counter associated to the currently tracked
14857 counters, and returns its previous value. Before the first invocation, the
Willy Tarreau869948b2013-01-04 14:14:57 +010014858 stored value is zero, so first invocation will always return zero. This is
14859 typically used as a second ACL in an expression in order to mark a connection
14860 when a first ACL was verified :
Willy Tarreauf73cd112011-08-13 01:45:16 +020014861
Jarno Huuskonen676f6222017-03-30 09:19:45 +030014862 Example:
Willy Tarreauf73cd112011-08-13 01:45:16 +020014863 # block if 5 consecutive requests continue to come faster than 10 sess
14864 # per second, and reset the counter as soon as the traffic slows down.
Willy Tarreaube4a3ef2013-06-17 15:04:07 +020014865 acl abuse sc0_http_req_rate gt 10
14866 acl kill sc0_inc_gpc0 gt 5
14867 acl save sc0_clr_gpc0 ge 0
Willy Tarreauf73cd112011-08-13 01:45:16 +020014868 tcp-request connection accept if !abuse save
14869 tcp-request connection reject if abuse kill
14870
Frédéric Lécaille6778b272018-01-29 15:22:53 +010014871sc_clr_gpc1(<ctr>[,<table>]) : integer
14872sc0_clr_gpc1([<table>]) : integer
14873sc1_clr_gpc1([<table>]) : integer
14874sc2_clr_gpc1([<table>]) : integer
14875 Clears the second General Purpose Counter associated to the currently tracked
14876 counters, and returns its previous value. Before the first invocation, the
14877 stored value is zero, so first invocation will always return zero. This is
14878 typically used as a second ACL in an expression in order to mark a connection
14879 when a first ACL was verified.
14880
Cyril Bonté62ba8702014-04-22 23:52:25 +020014881sc_conn_cnt(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020014882sc0_conn_cnt([<table>]) : integer
14883sc1_conn_cnt([<table>]) : integer
14884sc2_conn_cnt([<table>]) : integer
Davor Ocelice9ed2812017-12-25 17:49:28 +010014885 Returns the cumulative number of incoming connections from currently tracked
Willy Tarreaue9656522010-08-17 15:40:09 +020014886 counters. See also src_conn_cnt.
14887
Cyril Bonté62ba8702014-04-22 23:52:25 +020014888sc_conn_cur(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020014889sc0_conn_cur([<table>]) : integer
14890sc1_conn_cur([<table>]) : integer
14891sc2_conn_cur([<table>]) : integer
Willy Tarreaue9656522010-08-17 15:40:09 +020014892 Returns the current amount of concurrent connections tracking the same
14893 tracked counters. This number is automatically incremented when tracking
14894 begins and decremented when tracking stops. See also src_conn_cur.
14895
Cyril Bonté62ba8702014-04-22 23:52:25 +020014896sc_conn_rate(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020014897sc0_conn_rate([<table>]) : integer
14898sc1_conn_rate([<table>]) : integer
14899sc2_conn_rate([<table>]) : integer
Willy Tarreaue9656522010-08-17 15:40:09 +020014900 Returns the average connection rate from the currently tracked counters,
14901 measured in amount of connections over the period configured in the table.
14902 See also src_conn_rate.
14903
Cyril Bonté62ba8702014-04-22 23:52:25 +020014904sc_get_gpc0(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020014905sc0_get_gpc0([<table>]) : integer
14906sc1_get_gpc0([<table>]) : integer
14907sc2_get_gpc0([<table>]) : integer
Willy Tarreaue9656522010-08-17 15:40:09 +020014908 Returns the value of the first General Purpose Counter associated to the
Willy Tarreau4d4149c2013-07-23 19:33:46 +020014909 currently tracked counters. See also src_get_gpc0 and sc/sc0/sc1/sc2_inc_gpc0.
Willy Tarreauba2ffd12013-05-29 15:54:14 +020014910
Frédéric Lécaille6778b272018-01-29 15:22:53 +010014911sc_get_gpc1(<ctr>[,<table>]) : integer
14912sc0_get_gpc1([<table>]) : integer
14913sc1_get_gpc1([<table>]) : integer
14914sc2_get_gpc1([<table>]) : integer
14915 Returns the value of the second General Purpose Counter associated to the
14916 currently tracked counters. See also src_get_gpc1 and sc/sc0/sc1/sc2_inc_gpc1.
14917
Thierry FOURNIER236657b2015-08-19 08:25:14 +020014918sc_get_gpt0(<ctr>[,<table>]) : integer
14919sc0_get_gpt0([<table>]) : integer
14920sc1_get_gpt0([<table>]) : integer
14921sc2_get_gpt0([<table>]) : integer
14922 Returns the value of the first General Purpose Tag associated to the
14923 currently tracked counters. See also src_get_gpt0.
14924
Cyril Bonté62ba8702014-04-22 23:52:25 +020014925sc_gpc0_rate(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020014926sc0_gpc0_rate([<table>]) : integer
14927sc1_gpc0_rate([<table>]) : integer
14928sc2_gpc0_rate([<table>]) : integer
Willy Tarreauba2ffd12013-05-29 15:54:14 +020014929 Returns the average increment rate of the first General Purpose Counter
14930 associated to the currently tracked counters. It reports the frequency
14931 which the gpc0 counter was incremented over the configured period. See also
Willy Tarreau4d4149c2013-07-23 19:33:46 +020014932 src_gpc0_rate, sc/sc0/sc1/sc2_get_gpc0, and sc/sc0/sc1/sc2_inc_gpc0. Note
14933 that the "gpc0_rate" counter must be stored in the stick-table for a value to
14934 be returned, as "gpc0" only holds the event count.
Willy Tarreaue9656522010-08-17 15:40:09 +020014935
Frédéric Lécaille6778b272018-01-29 15:22:53 +010014936sc_gpc1_rate(<ctr>[,<table>]) : integer
14937sc0_gpc1_rate([<table>]) : integer
14938sc1_gpc1_rate([<table>]) : integer
14939sc2_gpc1_rate([<table>]) : integer
14940 Returns the average increment rate of the second General Purpose Counter
14941 associated to the currently tracked counters. It reports the frequency
14942 which the gpc1 counter was incremented over the configured period. See also
14943 src_gpcA_rate, sc/sc0/sc1/sc2_get_gpc1, and sc/sc0/sc1/sc2_inc_gpc1. Note
14944 that the "gpc1_rate" counter must be stored in the stick-table for a value to
14945 be returned, as "gpc1" only holds the event count.
14946
Cyril Bonté62ba8702014-04-22 23:52:25 +020014947sc_http_err_cnt(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020014948sc0_http_err_cnt([<table>]) : integer
14949sc1_http_err_cnt([<table>]) : integer
14950sc2_http_err_cnt([<table>]) : integer
Davor Ocelice9ed2812017-12-25 17:49:28 +010014951 Returns the cumulative number of HTTP errors from the currently tracked
Willy Tarreaue9656522010-08-17 15:40:09 +020014952 counters. This includes the both request errors and 4xx error responses.
14953 See also src_http_err_cnt.
14954
Cyril Bonté62ba8702014-04-22 23:52:25 +020014955sc_http_err_rate(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020014956sc0_http_err_rate([<table>]) : integer
14957sc1_http_err_rate([<table>]) : integer
14958sc2_http_err_rate([<table>]) : integer
Willy Tarreaue9656522010-08-17 15:40:09 +020014959 Returns the average rate of HTTP errors from the currently tracked counters,
14960 measured in amount of errors over the period configured in the table. This
14961 includes the both request errors and 4xx error responses. See also
14962 src_http_err_rate.
14963
Cyril Bonté62ba8702014-04-22 23:52:25 +020014964sc_http_req_cnt(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020014965sc0_http_req_cnt([<table>]) : integer
14966sc1_http_req_cnt([<table>]) : integer
14967sc2_http_req_cnt([<table>]) : integer
Davor Ocelice9ed2812017-12-25 17:49:28 +010014968 Returns the cumulative number of HTTP requests from the currently tracked
Willy Tarreaue9656522010-08-17 15:40:09 +020014969 counters. This includes every started request, valid or not. See also
14970 src_http_req_cnt.
14971
Cyril Bonté62ba8702014-04-22 23:52:25 +020014972sc_http_req_rate(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020014973sc0_http_req_rate([<table>]) : integer
14974sc1_http_req_rate([<table>]) : integer
14975sc2_http_req_rate([<table>]) : integer
Willy Tarreaue9656522010-08-17 15:40:09 +020014976 Returns the average rate of HTTP requests from the currently tracked
14977 counters, measured in amount of requests over the period configured in
14978 the table. This includes every started request, valid or not. See also
14979 src_http_req_rate.
14980
Cyril Bonté62ba8702014-04-22 23:52:25 +020014981sc_inc_gpc0(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020014982sc0_inc_gpc0([<table>]) : integer
14983sc1_inc_gpc0([<table>]) : integer
14984sc2_inc_gpc0([<table>]) : integer
Willy Tarreaue9656522010-08-17 15:40:09 +020014985 Increments the first General Purpose Counter associated to the currently
Willy Tarreau869948b2013-01-04 14:14:57 +010014986 tracked counters, and returns its new value. Before the first invocation,
14987 the stored value is zero, so first invocation will increase it to 1 and will
14988 return 1. This is typically used as a second ACL in an expression in order
14989 to mark a connection when a first ACL was verified :
Willy Tarreaue9656522010-08-17 15:40:09 +020014990
Jarno Huuskonen676f6222017-03-30 09:19:45 +030014991 Example:
Willy Tarreaube4a3ef2013-06-17 15:04:07 +020014992 acl abuse sc0_http_req_rate gt 10
14993 acl kill sc0_inc_gpc0 gt 0
Willy Tarreaue9656522010-08-17 15:40:09 +020014994 tcp-request connection reject if abuse kill
14995
Frédéric Lécaille6778b272018-01-29 15:22:53 +010014996sc_inc_gpc1(<ctr>[,<table>]) : integer
14997sc0_inc_gpc1([<table>]) : integer
14998sc1_inc_gpc1([<table>]) : integer
14999sc2_inc_gpc1([<table>]) : integer
15000 Increments the second General Purpose Counter associated to the currently
15001 tracked counters, and returns its new value. Before the first invocation,
15002 the stored value is zero, so first invocation will increase it to 1 and will
15003 return 1. This is typically used as a second ACL in an expression in order
15004 to mark a connection when a first ACL was verified.
15005
Cyril Bonté62ba8702014-04-22 23:52:25 +020015006sc_kbytes_in(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020015007sc0_kbytes_in([<table>]) : integer
15008sc1_kbytes_in([<table>]) : integer
15009sc2_kbytes_in([<table>]) : integer
Willy Tarreaua01b9742014-07-10 15:29:24 +020015010 Returns the total amount of client-to-server data from the currently tracked
15011 counters, measured in kilobytes. The test is currently performed on 32-bit
15012 integers, which limits values to 4 terabytes. See also src_kbytes_in.
Willy Tarreaue9656522010-08-17 15:40:09 +020015013
Cyril Bonté62ba8702014-04-22 23:52:25 +020015014sc_kbytes_out(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020015015sc0_kbytes_out([<table>]) : integer
15016sc1_kbytes_out([<table>]) : integer
15017sc2_kbytes_out([<table>]) : integer
Willy Tarreaua01b9742014-07-10 15:29:24 +020015018 Returns the total amount of server-to-client data from the currently tracked
15019 counters, measured in kilobytes. The test is currently performed on 32-bit
15020 integers, which limits values to 4 terabytes. See also src_kbytes_out.
Willy Tarreaue9656522010-08-17 15:40:09 +020015021
Cyril Bonté62ba8702014-04-22 23:52:25 +020015022sc_sess_cnt(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020015023sc0_sess_cnt([<table>]) : integer
15024sc1_sess_cnt([<table>]) : integer
15025sc2_sess_cnt([<table>]) : integer
Davor Ocelice9ed2812017-12-25 17:49:28 +010015026 Returns the cumulative number of incoming connections that were transformed
Willy Tarreaue9656522010-08-17 15:40:09 +020015027 into sessions, which means that they were accepted by a "tcp-request
15028 connection" rule, from the currently tracked counters. A backend may count
15029 more sessions than connections because each connection could result in many
Jamie Gloudonaaa21002012-08-25 00:18:33 -040015030 backend sessions if some HTTP keep-alive is performed over the connection
Willy Tarreaue9656522010-08-17 15:40:09 +020015031 with the client. See also src_sess_cnt.
15032
Cyril Bonté62ba8702014-04-22 23:52:25 +020015033sc_sess_rate(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020015034sc0_sess_rate([<table>]) : integer
15035sc1_sess_rate([<table>]) : integer
15036sc2_sess_rate([<table>]) : integer
Willy Tarreaue9656522010-08-17 15:40:09 +020015037 Returns the average session rate from the currently tracked counters,
15038 measured in amount of sessions over the period configured in the table. A
15039 session is a connection that got past the early "tcp-request connection"
15040 rules. A backend may count more sessions than connections because each
15041 connection could result in many backend sessions if some HTTP keep-alive is
Jamie Gloudonaaa21002012-08-25 00:18:33 -040015042 performed over the connection with the client. See also src_sess_rate.
Willy Tarreaue9656522010-08-17 15:40:09 +020015043
Cyril Bonté62ba8702014-04-22 23:52:25 +020015044sc_tracked(<ctr>[,<table>]) : boolean
Willy Tarreau0f791d42013-07-23 19:56:43 +020015045sc0_tracked([<table>]) : boolean
15046sc1_tracked([<table>]) : boolean
15047sc2_tracked([<table>]) : boolean
Willy Tarreau6f1615f2013-06-03 15:15:22 +020015048 Returns true if the designated session counter is currently being tracked by
15049 the current session. This can be useful when deciding whether or not we want
15050 to set some values in a header passed to the server.
15051
Cyril Bonté62ba8702014-04-22 23:52:25 +020015052sc_trackers(<ctr>[,<table>]) : integer
Willy Tarreau0f791d42013-07-23 19:56:43 +020015053sc0_trackers([<table>]) : integer
15054sc1_trackers([<table>]) : integer
15055sc2_trackers([<table>]) : integer
Willy Tarreau2406db42012-12-09 12:16:43 +010015056 Returns the current amount of concurrent connections tracking the same
15057 tracked counters. This number is automatically incremented when tracking
Willy Tarreaube4a3ef2013-06-17 15:04:07 +020015058 begins and decremented when tracking stops. It differs from sc0_conn_cur in
Willy Tarreau2406db42012-12-09 12:16:43 +010015059 that it does not rely on any stored information but on the table's reference
15060 count (the "use" value which is returned by "show table" on the CLI). This
Willy Tarreau74ca5042013-06-11 23:12:07 +020015061 may sometimes be more suited for layer7 tracking. It can be used to tell a
15062 server how many concurrent connections there are from a given address for
15063 example.
Willy Tarreau2406db42012-12-09 12:16:43 +010015064
Willy Tarreau74ca5042013-06-11 23:12:07 +020015065so_id : integer
15066 Returns an integer containing the current listening socket's id. It is useful
15067 in frontends involving many "bind" lines, or to stick all users coming via a
15068 same socket to the same server.
Willy Tarreaud63335a2010-02-26 12:56:52 +010015069
Willy Tarreau74ca5042013-06-11 23:12:07 +020015070src : ip
Davor Ocelice9ed2812017-12-25 17:49:28 +010015071 This is the source IPv4 address of the client of the session. It is of type
Willy Tarreau74ca5042013-06-11 23:12:07 +020015072 IP and works on both IPv4 and IPv6 tables. On IPv6 tables, IPv4 addresses are
15073 mapped to their IPv6 equivalent, according to RFC 4291. Note that it is the
15074 TCP-level source address which is used, and not the address of a client
Bertrand Jacquin93b227d2016-06-04 15:11:10 +010015075 behind a proxy. However if the "accept-proxy" or "accept-netscaler-cip" bind
15076 directive is used, it can be the address of a client behind another
15077 PROXY-protocol compatible component for all rule sets except
Willy Tarreau64ded3d2019-01-23 10:02:15 +010015078 "tcp-request connection" which sees the real address. When the incoming
15079 connection passed through address translation or redirection involving
15080 connection tracking, the original destination address before the redirection
15081 will be reported. On Linux systems, the source and destination may seldom
15082 appear reversed if the nf_conntrack_tcp_loose sysctl is set, because a late
15083 response may reopen a timed out connection and switch what is believed to be
15084 the source and the destination.
Willy Tarreaud63335a2010-02-26 12:56:52 +010015085
Thierry FOURNIERd5f624d2013-11-26 11:52:33 +010015086 Example:
15087 # add an HTTP header in requests with the originating address' country
15088 http-request set-header X-Country %[src,map_ip(geoip.lst)]
15089
Willy Tarreau74ca5042013-06-11 23:12:07 +020015090src_bytes_in_rate([<table>]) : integer
15091 Returns the average bytes rate from the incoming connection's source address
15092 in the current proxy's stick-table or in the designated stick-table, measured
15093 in amount of bytes over the period configured in the table. If the address is
Willy Tarreau4d4149c2013-07-23 19:33:46 +020015094 not found, zero is returned. See also sc/sc0/sc1/sc2_bytes_in_rate.
Willy Tarreauc9705a12010-07-27 20:05:50 +020015095
Willy Tarreau74ca5042013-06-11 23:12:07 +020015096src_bytes_out_rate([<table>]) : integer
15097 Returns the average bytes rate to the incoming connection's source address in
15098 the current proxy's stick-table or in the designated stick-table, measured in
Willy Tarreauc9705a12010-07-27 20:05:50 +020015099 amount of bytes over the period configured in the table. If the address is
Willy Tarreau4d4149c2013-07-23 19:33:46 +020015100 not found, zero is returned. See also sc/sc0/sc1/sc2_bytes_out_rate.
Willy Tarreauc9705a12010-07-27 20:05:50 +020015101
Willy Tarreau74ca5042013-06-11 23:12:07 +020015102src_clr_gpc0([<table>]) : integer
15103 Clears the first General Purpose Counter associated to the incoming
15104 connection's source address in the current proxy's stick-table or in the
15105 designated stick-table, and returns its previous value. If the address is not
15106 found, an entry is created and 0 is returned. This is typically used as a
15107 second ACL in an expression in order to mark a connection when a first ACL
15108 was verified :
Willy Tarreauf73cd112011-08-13 01:45:16 +020015109
Jarno Huuskonen676f6222017-03-30 09:19:45 +030015110 Example:
Willy Tarreauf73cd112011-08-13 01:45:16 +020015111 # block if 5 consecutive requests continue to come faster than 10 sess
15112 # per second, and reset the counter as soon as the traffic slows down.
15113 acl abuse src_http_req_rate gt 10
15114 acl kill src_inc_gpc0 gt 5
Willy Tarreau869948b2013-01-04 14:14:57 +010015115 acl save src_clr_gpc0 ge 0
Willy Tarreauf73cd112011-08-13 01:45:16 +020015116 tcp-request connection accept if !abuse save
15117 tcp-request connection reject if abuse kill
15118
Frédéric Lécaille6778b272018-01-29 15:22:53 +010015119src_clr_gpc1([<table>]) : integer
15120 Clears the second General Purpose Counter associated to the incoming
15121 connection's source address in the current proxy's stick-table or in the
15122 designated stick-table, and returns its previous value. If the address is not
15123 found, an entry is created and 0 is returned. This is typically used as a
15124 second ACL in an expression in order to mark a connection when a first ACL
15125 was verified.
15126
Willy Tarreau74ca5042013-06-11 23:12:07 +020015127src_conn_cnt([<table>]) : integer
Davor Ocelice9ed2812017-12-25 17:49:28 +010015128 Returns the cumulative number of connections initiated from the current
Willy Tarreau74ca5042013-06-11 23:12:07 +020015129 incoming connection's source address in the current proxy's stick-table or in
Willy Tarreauc9705a12010-07-27 20:05:50 +020015130 the designated stick-table. If the address is not found, zero is returned.
Willy Tarreau4d4149c2013-07-23 19:33:46 +020015131 See also sc/sc0/sc1/sc2_conn_cnt.
Willy Tarreauc9705a12010-07-27 20:05:50 +020015132
Willy Tarreau74ca5042013-06-11 23:12:07 +020015133src_conn_cur([<table>]) : integer
Willy Tarreauc9705a12010-07-27 20:05:50 +020015134 Returns the current amount of concurrent connections initiated from the
Willy Tarreau74ca5042013-06-11 23:12:07 +020015135 current incoming connection's source address in the current proxy's
15136 stick-table or in the designated stick-table. If the address is not found,
Willy Tarreau4d4149c2013-07-23 19:33:46 +020015137 zero is returned. See also sc/sc0/sc1/sc2_conn_cur.
Willy Tarreauc9705a12010-07-27 20:05:50 +020015138
Willy Tarreau74ca5042013-06-11 23:12:07 +020015139src_conn_rate([<table>]) : integer
15140 Returns the average connection rate from the incoming connection's source
15141 address in the current proxy's stick-table or in the designated stick-table,
15142 measured in amount of connections over the period configured in the table. If
Willy Tarreau4d4149c2013-07-23 19:33:46 +020015143 the address is not found, zero is returned. See also sc/sc0/sc1/sc2_conn_rate.
Willy Tarreauc9705a12010-07-27 20:05:50 +020015144
Willy Tarreau74ca5042013-06-11 23:12:07 +020015145src_get_gpc0([<table>]) : integer
Willy Tarreauc9705a12010-07-27 20:05:50 +020015146 Returns the value of the first General Purpose Counter associated to the
Willy Tarreau74ca5042013-06-11 23:12:07 +020015147 incoming connection's source address in the current proxy's stick-table or in
Willy Tarreauc9705a12010-07-27 20:05:50 +020015148 the designated stick-table. If the address is not found, zero is returned.
Willy Tarreau4d4149c2013-07-23 19:33:46 +020015149 See also sc/sc0/sc1/sc2_get_gpc0 and src_inc_gpc0.
Willy Tarreauc9705a12010-07-27 20:05:50 +020015150
Frédéric Lécaille6778b272018-01-29 15:22:53 +010015151src_get_gpc1([<table>]) : integer
15152 Returns the value of the second General Purpose Counter associated to the
15153 incoming connection's source address in the current proxy's stick-table or in
15154 the designated stick-table. If the address is not found, zero is returned.
15155 See also sc/sc0/sc1/sc2_get_gpc1 and src_inc_gpc1.
15156
Thierry FOURNIER236657b2015-08-19 08:25:14 +020015157src_get_gpt0([<table>]) : integer
15158 Returns the value of the first General Purpose Tag associated to the
15159 incoming connection's source address in the current proxy's stick-table or in
15160 the designated stick-table. If the address is not found, zero is returned.
15161 See also sc/sc0/sc1/sc2_get_gpt0.
15162
Willy Tarreau74ca5042013-06-11 23:12:07 +020015163src_gpc0_rate([<table>]) : integer
Willy Tarreauba2ffd12013-05-29 15:54:14 +020015164 Returns the average increment rate of the first General Purpose Counter
Willy Tarreau74ca5042013-06-11 23:12:07 +020015165 associated to the incoming connection's source address in the current proxy's
Willy Tarreauba2ffd12013-05-29 15:54:14 +020015166 stick-table or in the designated stick-table. It reports the frequency
15167 which the gpc0 counter was incremented over the configured period. See also
Willy Tarreau4d4149c2013-07-23 19:33:46 +020015168 sc/sc0/sc1/sc2_gpc0_rate, src_get_gpc0, and sc/sc0/sc1/sc2_inc_gpc0. Note
15169 that the "gpc0_rate" counter must be stored in the stick-table for a value to
15170 be returned, as "gpc0" only holds the event count.
Willy Tarreauba2ffd12013-05-29 15:54:14 +020015171
Frédéric Lécaille6778b272018-01-29 15:22:53 +010015172src_gpc1_rate([<table>]) : integer
15173 Returns the average increment rate of the second General Purpose Counter
15174 associated to the incoming connection's source address in the current proxy's
15175 stick-table or in the designated stick-table. It reports the frequency
15176 which the gpc1 counter was incremented over the configured period. See also
15177 sc/sc0/sc1/sc2_gpc1_rate, src_get_gpc1, and sc/sc0/sc1/sc2_inc_gpc1. Note
15178 that the "gpc1_rate" counter must be stored in the stick-table for a value to
15179 be returned, as "gpc1" only holds the event count.
15180
Willy Tarreau74ca5042013-06-11 23:12:07 +020015181src_http_err_cnt([<table>]) : integer
Davor Ocelice9ed2812017-12-25 17:49:28 +010015182 Returns the cumulative number of HTTP errors from the incoming connection's
Willy Tarreau74ca5042013-06-11 23:12:07 +020015183 source address in the current proxy's stick-table or in the designated
Willy Tarreauc9705a12010-07-27 20:05:50 +020015184 stick-table. This includes the both request errors and 4xx error responses.
Willy Tarreau4d4149c2013-07-23 19:33:46 +020015185 See also sc/sc0/sc1/sc2_http_err_cnt. If the address is not found, zero is
Willy Tarreau74ca5042013-06-11 23:12:07 +020015186 returned.
Willy Tarreauc9705a12010-07-27 20:05:50 +020015187
Willy Tarreau74ca5042013-06-11 23:12:07 +020015188src_http_err_rate([<table>]) : integer
15189 Returns the average rate of HTTP errors from the incoming connection's source
15190 address in the current proxy's stick-table or in the designated stick-table,
15191 measured in amount of errors over the period configured in the table. This
15192 includes the both request errors and 4xx error responses. If the address is
Willy Tarreau4d4149c2013-07-23 19:33:46 +020015193 not found, zero is returned. See also sc/sc0/sc1/sc2_http_err_rate.
Willy Tarreauc9705a12010-07-27 20:05:50 +020015194
Willy Tarreau74ca5042013-06-11 23:12:07 +020015195src_http_req_cnt([<table>]) : integer
Davor Ocelice9ed2812017-12-25 17:49:28 +010015196 Returns the cumulative number of HTTP requests from the incoming connection's
Willy Tarreau74ca5042013-06-11 23:12:07 +020015197 source address in the current proxy's stick-table or in the designated stick-
15198 table. This includes every started request, valid or not. If the address is
Willy Tarreau4d4149c2013-07-23 19:33:46 +020015199 not found, zero is returned. See also sc/sc0/sc1/sc2_http_req_cnt.
Willy Tarreauc9705a12010-07-27 20:05:50 +020015200
Willy Tarreau74ca5042013-06-11 23:12:07 +020015201src_http_req_rate([<table>]) : integer
15202 Returns the average rate of HTTP requests from the incoming connection's
15203 source address in the current proxy's stick-table or in the designated stick-
15204 table, measured in amount of requests over the period configured in the
Willy Tarreauc9705a12010-07-27 20:05:50 +020015205 table. This includes every started request, valid or not. If the address is
Willy Tarreau4d4149c2013-07-23 19:33:46 +020015206 not found, zero is returned. See also sc/sc0/sc1/sc2_http_req_rate.
Willy Tarreauc9705a12010-07-27 20:05:50 +020015207
Willy Tarreau74ca5042013-06-11 23:12:07 +020015208src_inc_gpc0([<table>]) : integer
15209 Increments the first General Purpose Counter associated to the incoming
15210 connection's source address in the current proxy's stick-table or in the
15211 designated stick-table, and returns its new value. If the address is not
Willy Tarreaube4a3ef2013-06-17 15:04:07 +020015212 found, an entry is created and 1 is returned. See also sc0/sc2/sc2_inc_gpc0.
Willy Tarreau74ca5042013-06-11 23:12:07 +020015213 This is typically used as a second ACL in an expression in order to mark a
15214 connection when a first ACL was verified :
Willy Tarreauc9705a12010-07-27 20:05:50 +020015215
Jarno Huuskonen676f6222017-03-30 09:19:45 +030015216 Example:
Willy Tarreauc9705a12010-07-27 20:05:50 +020015217 acl abuse src_http_req_rate gt 10
Willy Tarreau869948b2013-01-04 14:14:57 +010015218 acl kill src_inc_gpc0 gt 0
Willy Tarreaue9656522010-08-17 15:40:09 +020015219 tcp-request connection reject if abuse kill
Willy Tarreauc9705a12010-07-27 20:05:50 +020015220
Frédéric Lécaille6778b272018-01-29 15:22:53 +010015221src_inc_gpc1([<table>]) : integer
15222 Increments the second General Purpose Counter associated to the incoming
15223 connection's source address in the current proxy's stick-table or in the
15224 designated stick-table, and returns its new value. If the address is not
15225 found, an entry is created and 1 is returned. See also sc0/sc2/sc2_inc_gpc1.
15226 This is typically used as a second ACL in an expression in order to mark a
15227 connection when a first ACL was verified.
15228
Willy Tarreau16e01562016-08-09 16:46:18 +020015229src_is_local : boolean
15230 Returns true if the source address of the incoming connection is local to the
15231 system, or false if the address doesn't exist on the system, meaning that it
15232 comes from a remote machine. Note that UNIX addresses are considered local.
15233 It can be useful to apply certain access restrictions based on where the
Davor Ocelice9ed2812017-12-25 17:49:28 +010015234 client comes from (e.g. require auth or https for remote machines). Please
Willy Tarreau16e01562016-08-09 16:46:18 +020015235 note that the check involves a few system calls, so it's better to do it only
15236 once per connection.
15237
Willy Tarreau74ca5042013-06-11 23:12:07 +020015238src_kbytes_in([<table>]) : integer
Willy Tarreaua01b9742014-07-10 15:29:24 +020015239 Returns the total amount of data received from the incoming connection's
15240 source address in the current proxy's stick-table or in the designated
15241 stick-table, measured in kilobytes. If the address is not found, zero is
15242 returned. The test is currently performed on 32-bit integers, which limits
15243 values to 4 terabytes. See also sc/sc0/sc1/sc2_kbytes_in.
Willy Tarreauc9705a12010-07-27 20:05:50 +020015244
Willy Tarreau74ca5042013-06-11 23:12:07 +020015245src_kbytes_out([<table>]) : integer
Willy Tarreaua01b9742014-07-10 15:29:24 +020015246 Returns the total amount of data sent to the incoming connection's source
15247 address in the current proxy's stick-table or in the designated stick-table,
15248 measured in kilobytes. If the address is not found, zero is returned. The
15249 test is currently performed on 32-bit integers, which limits values to 4
15250 terabytes. See also sc/sc0/sc1/sc2_kbytes_out.
Willy Tarreaua975b8f2010-06-05 19:13:27 +020015251
Willy Tarreau74ca5042013-06-11 23:12:07 +020015252src_port : integer
15253 Returns an integer value corresponding to the TCP source port of the
15254 connection on the client side, which is the port the client connected from.
15255 Usage of this function is very limited as modern protocols do not care much
15256 about source ports nowadays.
Willy Tarreau079ff0a2009-03-05 21:34:28 +010015257
Willy Tarreau74ca5042013-06-11 23:12:07 +020015258src_sess_cnt([<table>]) : integer
Davor Ocelice9ed2812017-12-25 17:49:28 +010015259 Returns the cumulative number of connections initiated from the incoming
Willy Tarreauc9705a12010-07-27 20:05:50 +020015260 connection's source IPv4 address in the current proxy's stick-table or in the
15261 designated stick-table, that were transformed into sessions, which means that
15262 they were accepted by "tcp-request" rules. If the address is not found, zero
Willy Tarreau4d4149c2013-07-23 19:33:46 +020015263 is returned. See also sc/sc0/sc1/sc2_sess_cnt.
Willy Tarreauc9705a12010-07-27 20:05:50 +020015264
Willy Tarreau74ca5042013-06-11 23:12:07 +020015265src_sess_rate([<table>]) : integer
15266 Returns the average session rate from the incoming connection's source
15267 address in the current proxy's stick-table or in the designated stick-table,
15268 measured in amount of sessions over the period configured in the table. A
15269 session is a connection that went past the early "tcp-request" rules. If the
Willy Tarreau4d4149c2013-07-23 19:33:46 +020015270 address is not found, zero is returned. See also sc/sc0/sc1/sc2_sess_rate.
Willy Tarreauc9705a12010-07-27 20:05:50 +020015271
Willy Tarreau74ca5042013-06-11 23:12:07 +020015272src_updt_conn_cnt([<table>]) : integer
15273 Creates or updates the entry associated to the incoming connection's source
15274 address in the current proxy's stick-table or in the designated stick-table.
15275 This table must be configured to store the "conn_cnt" data type, otherwise
15276 the match will be ignored. The current count is incremented by one, and the
15277 expiration timer refreshed. The updated count is returned, so this match
15278 can't return zero. This was used to reject service abusers based on their
15279 source address. Note: it is recommended to use the more complete "track-sc*"
15280 actions in "tcp-request" rules instead.
Willy Tarreaua975b8f2010-06-05 19:13:27 +020015281
15282 Example :
15283 # This frontend limits incoming SSH connections to 3 per 10 second for
15284 # each source address, and rejects excess connections until a 10 second
15285 # silence is observed. At most 20 addresses are tracked.
15286 listen ssh
15287 bind :22
15288 mode tcp
15289 maxconn 100
Willy Tarreauc9705a12010-07-27 20:05:50 +020015290 stick-table type ip size 20 expire 10s store conn_cnt
Willy Tarreau74ca5042013-06-11 23:12:07 +020015291 tcp-request content reject if { src_updt_conn_cnt gt 3 }
Willy Tarreaua975b8f2010-06-05 19:13:27 +020015292 server local 127.0.0.1:22
15293
Willy Tarreau74ca5042013-06-11 23:12:07 +020015294srv_id : integer
15295 Returns an integer containing the server's id when processing the response.
15296 While it's almost only used with ACLs, it may be used for logging or
15297 debugging.
Hervé COMMOWICKdaa824e2011-08-05 12:09:44 +020015298
Thierry FOURNIER060762e2014-04-23 13:29:15 +0200152997.3.4. Fetching samples at Layer 5
Willy Tarreau74ca5042013-06-11 23:12:07 +020015300----------------------------------
Willy Tarreau0b1cd942010-05-16 22:18:27 +020015301
Willy Tarreau74ca5042013-06-11 23:12:07 +020015302The layer 5 usually describes just the session layer which in haproxy is
15303closest to the session once all the connection handshakes are finished, but
15304when no content is yet made available. The fetch methods described here are
15305usable as low as the "tcp-request content" rule sets unless they require some
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030015306future information. Those generally include the results of SSL negotiations.
Willy Tarreauc735a072011-03-29 00:57:02 +020015307
Ben Shillitof25e8e52016-12-02 14:25:37 +00001530851d.all(<prop>[,<prop>*]) : string
15309 Returns values for the properties requested as a string, where values are
15310 separated by the delimiter specified with "51degrees-property-separator".
15311 The device is identified using all the important HTTP headers from the
15312 request. The function can be passed up to five property names, and if a
15313 property name can't be found, the value "NoData" is returned.
15314
15315 Example :
15316 # Here the header "X-51D-DeviceTypeMobileTablet" is added to the request
15317 # containing the three properties requested using all relevant headers from
15318 # the request.
15319 frontend http-in
15320 bind *:8081
15321 default_backend servers
15322 http-request set-header X-51D-DeviceTypeMobileTablet \
15323 %[51d.all(DeviceType,IsMobile,IsTablet)]
15324
Emeric Brun645ae792014-04-30 14:21:06 +020015325ssl_bc : boolean
15326 Returns true when the back connection was made via an SSL/TLS transport
15327 layer and is locally deciphered. This means the outgoing connection was made
15328 other a server with the "ssl" option.
15329
15330ssl_bc_alg_keysize : integer
15331 Returns the symmetric cipher key size supported in bits when the outgoing
15332 connection was made over an SSL/TLS transport layer.
15333
Olivier Houchard6b77f492018-11-22 18:18:29 +010015334ssl_bc_alpn : string
15335 This extracts the Application Layer Protocol Negotiation field from an
15336 outgoing connection made via a TLS transport layer.
Michael Prokop4438c602019-05-24 10:25:45 +020015337 The result is a string containing the protocol name negotiated with the
Olivier Houchard6b77f492018-11-22 18:18:29 +010015338 server. The SSL library must have been built with support for TLS
15339 extensions enabled (check haproxy -vv). Note that the TLS ALPN extension is
15340 not advertised unless the "alpn" keyword on the "server" line specifies a
15341 protocol list. Also, nothing forces the server to pick a protocol from this
15342 list, any other one may be requested. The TLS ALPN extension is meant to
15343 replace the TLS NPN extension. See also "ssl_bc_npn".
15344
Emeric Brun645ae792014-04-30 14:21:06 +020015345ssl_bc_cipher : string
15346 Returns the name of the used cipher when the outgoing connection was made
15347 over an SSL/TLS transport layer.
15348
Patrick Hemmer65674662019-06-04 08:13:03 -040015349ssl_bc_client_random : binary
15350 Returns the client random of the back connection when the incoming connection
15351 was made over an SSL/TLS transport layer. It is useful to to decrypt traffic
15352 sent using ephemeral ciphers. This requires OpenSSL >= 1.1.0, or BoringSSL.
15353
Emeric Brun74f7ffa2018-02-19 16:14:12 +010015354ssl_bc_is_resumed : boolean
15355 Returns true when the back connection was made over an SSL/TLS transport
15356 layer and the newly created SSL session was resumed using a cached
15357 session or a TLS ticket.
15358
Olivier Houchard6b77f492018-11-22 18:18:29 +010015359ssl_bc_npn : string
15360 This extracts the Next Protocol Negotiation field from an outgoing connection
15361 made via a TLS transport layer. The result is a string containing the
Michael Prokop4438c602019-05-24 10:25:45 +020015362 protocol name negotiated with the server . The SSL library must have been
Olivier Houchard6b77f492018-11-22 18:18:29 +010015363 built with support for TLS extensions enabled (check haproxy -vv). Note that
15364 the TLS NPN extension is not advertised unless the "npn" keyword on the
15365 "server" line specifies a protocol list. Also, nothing forces the server to
15366 pick a protocol from this list, any other one may be used. Please note that
15367 the TLS NPN extension was replaced with ALPN.
15368
Emeric Brun645ae792014-04-30 14:21:06 +020015369ssl_bc_protocol : string
15370 Returns the name of the used protocol when the outgoing connection was made
15371 over an SSL/TLS transport layer.
15372
Emeric Brunb73a9b02014-04-30 18:49:19 +020015373ssl_bc_unique_id : binary
Emeric Brun645ae792014-04-30 14:21:06 +020015374 When the outgoing connection was made over an SSL/TLS transport layer,
Emeric Brunb73a9b02014-04-30 18:49:19 +020015375 returns the TLS unique ID as defined in RFC5929 section 3. The unique id
15376 can be encoded to base64 using the converter: "ssl_bc_unique_id,base64".
Emeric Brun645ae792014-04-30 14:21:06 +020015377
Patrick Hemmer65674662019-06-04 08:13:03 -040015378ssl_bc_server_random : binary
15379 Returns the server random of the back connection when the incoming connection
15380 was made over an SSL/TLS transport layer. It is useful to to decrypt traffic
15381 sent using ephemeral ciphers. This requires OpenSSL >= 1.1.0, or BoringSSL.
15382
Emeric Brun645ae792014-04-30 14:21:06 +020015383ssl_bc_session_id : binary
15384 Returns the SSL ID of the back connection when the outgoing connection was
15385 made over an SSL/TLS transport layer. It is useful to log if we want to know
15386 if session was reused or not.
15387
Patrick Hemmere0275472018-04-28 19:15:51 -040015388ssl_bc_session_key : binary
15389 Returns the SSL session master key of the back connection when the outgoing
15390 connection was made over an SSL/TLS transport layer. It is useful to decrypt
15391 traffic sent using ephemeral ciphers. This requires OpenSSL >= 1.1.0, or
15392 BoringSSL.
15393
Emeric Brun645ae792014-04-30 14:21:06 +020015394ssl_bc_use_keysize : integer
15395 Returns the symmetric cipher key size used in bits when the outgoing
15396 connection was made over an SSL/TLS transport layer.
15397
Willy Tarreau74ca5042013-06-11 23:12:07 +020015398ssl_c_ca_err : integer
15399 When the incoming connection was made over an SSL/TLS transport layer,
15400 returns the ID of the first error detected during verification of the client
15401 certificate at depth > 0, or 0 if no error was encountered during this
15402 verification process. Please refer to your SSL library's documentation to
15403 find the exhaustive list of error codes.
Willy Tarreauc735a072011-03-29 00:57:02 +020015404
Willy Tarreau74ca5042013-06-11 23:12:07 +020015405ssl_c_ca_err_depth : integer
15406 When the incoming connection was made over an SSL/TLS transport layer,
15407 returns the depth in the CA chain of the first error detected during the
15408 verification of the client certificate. If no error is encountered, 0 is
15409 returned.
Willy Tarreau0ba27502007-12-24 16:55:16 +010015410
Emeric Brun43e79582014-10-29 19:03:26 +010015411ssl_c_der : binary
15412 Returns the DER formatted certificate presented by the client when the
15413 incoming connection was made over an SSL/TLS transport layer. When used for
15414 an ACL, the value(s) to match against can be passed in hexadecimal form.
15415
Willy Tarreau74ca5042013-06-11 23:12:07 +020015416ssl_c_err : integer
15417 When the incoming connection was made over an SSL/TLS transport layer,
15418 returns the ID of the first error detected during verification at depth 0, or
15419 0 if no error was encountered during this verification process. Please refer
15420 to your SSL library's documentation to find the exhaustive list of error
15421 codes.
Willy Tarreau62644772008-07-16 18:36:06 +020015422
Willy Tarreau74ca5042013-06-11 23:12:07 +020015423ssl_c_i_dn([<entry>[,<occ>]]) : string
15424 When the incoming connection was made over an SSL/TLS transport layer,
15425 returns the full distinguished name of the issuer of the certificate
15426 presented by the client when no <entry> is specified, or the value of the
15427 first given entry found from the beginning of the DN. If a positive/negative
15428 occurrence number is specified as the optional second argument, it returns
15429 the value of the nth given entry value from the beginning/end of the DN.
15430 For instance, "ssl_c_i_dn(OU,2)" the second organization unit, and
15431 "ssl_c_i_dn(CN)" retrieves the common name.
Willy Tarreau62644772008-07-16 18:36:06 +020015432
Willy Tarreau74ca5042013-06-11 23:12:07 +020015433ssl_c_key_alg : string
15434 Returns the name of the algorithm used to generate the key of the certificate
15435 presented by the client when the incoming connection was made over an SSL/TLS
15436 transport layer.
Willy Tarreau62644772008-07-16 18:36:06 +020015437
Willy Tarreau74ca5042013-06-11 23:12:07 +020015438ssl_c_notafter : string
15439 Returns the end date presented by the client as a formatted string
15440 YYMMDDhhmmss[Z] when the incoming connection was made over an SSL/TLS
15441 transport layer.
Emeric Brunbede3d02009-06-30 17:54:00 +020015442
Willy Tarreau74ca5042013-06-11 23:12:07 +020015443ssl_c_notbefore : string
15444 Returns the start date presented by the client as a formatted string
15445 YYMMDDhhmmss[Z] when the incoming connection was made over an SSL/TLS
15446 transport layer.
Willy Tarreaub6672b52011-12-12 17:23:41 +010015447
Willy Tarreau74ca5042013-06-11 23:12:07 +020015448ssl_c_s_dn([<entry>[,<occ>]]) : string
15449 When the incoming connection was made over an SSL/TLS transport layer,
15450 returns the full distinguished name of the subject of the certificate
15451 presented by the client when no <entry> is specified, or the value of the
15452 first given entry found from the beginning of the DN. If a positive/negative
15453 occurrence number is specified as the optional second argument, it returns
15454 the value of the nth given entry value from the beginning/end of the DN.
15455 For instance, "ssl_c_s_dn(OU,2)" the second organization unit, and
15456 "ssl_c_s_dn(CN)" retrieves the common name.
Willy Tarreaub6672b52011-12-12 17:23:41 +010015457
Willy Tarreau74ca5042013-06-11 23:12:07 +020015458ssl_c_serial : binary
15459 Returns the serial of the certificate presented by the client when the
15460 incoming connection was made over an SSL/TLS transport layer. When used for
15461 an ACL, the value(s) to match against can be passed in hexadecimal form.
Emeric Brun2525b6b2012-10-18 15:59:43 +020015462
Willy Tarreau74ca5042013-06-11 23:12:07 +020015463ssl_c_sha1 : binary
15464 Returns the SHA-1 fingerprint of the certificate presented by the client when
15465 the incoming connection was made over an SSL/TLS transport layer. This can be
15466 used to stick a client to a server, or to pass this information to a server.
Willy Tarreau2d0caa32014-07-02 19:01:22 +020015467 Note that the output is binary, so if you want to pass that signature to the
15468 server, you need to encode it in hex or base64, such as in the example below:
15469
Jarno Huuskonen676f6222017-03-30 09:19:45 +030015470 Example:
Willy Tarreau2d0caa32014-07-02 19:01:22 +020015471 http-request set-header X-SSL-Client-SHA1 %[ssl_c_sha1,hex]
Emeric Brun2525b6b2012-10-18 15:59:43 +020015472
Willy Tarreau74ca5042013-06-11 23:12:07 +020015473ssl_c_sig_alg : string
15474 Returns the name of the algorithm used to sign the certificate presented by
15475 the client when the incoming connection was made over an SSL/TLS transport
15476 layer.
Emeric Brun87855892012-10-17 17:39:35 +020015477
Willy Tarreau74ca5042013-06-11 23:12:07 +020015478ssl_c_used : boolean
15479 Returns true if current SSL session uses a client certificate even if current
15480 connection uses SSL session resumption. See also "ssl_fc_has_crt".
Emeric Brun7f56e742012-10-19 18:15:40 +020015481
Willy Tarreau74ca5042013-06-11 23:12:07 +020015482ssl_c_verify : integer
15483 Returns the verify result error ID when the incoming connection was made over
15484 an SSL/TLS transport layer, otherwise zero if no error is encountered. Please
15485 refer to your SSL library's documentation for an exhaustive list of error
15486 codes.
Emeric Brunce5ad802012-10-22 14:11:22 +020015487
Willy Tarreau74ca5042013-06-11 23:12:07 +020015488ssl_c_version : integer
15489 Returns the version of the certificate presented by the client when the
15490 incoming connection was made over an SSL/TLS transport layer.
Emeric Brunce5ad802012-10-22 14:11:22 +020015491
Emeric Brun43e79582014-10-29 19:03:26 +010015492ssl_f_der : binary
15493 Returns the DER formatted certificate presented by the frontend when the
15494 incoming connection was made over an SSL/TLS transport layer. When used for
15495 an ACL, the value(s) to match against can be passed in hexadecimal form.
15496
Willy Tarreau74ca5042013-06-11 23:12:07 +020015497ssl_f_i_dn([<entry>[,<occ>]]) : string
15498 When the incoming connection was made over an SSL/TLS transport layer,
15499 returns the full distinguished name of the issuer of the certificate
15500 presented by the frontend when no <entry> is specified, or the value of the
15501 first given entry found from the beginning of the DN. If a positive/negative
Emeric Brun87855892012-10-17 17:39:35 +020015502 occurrence number is specified as the optional second argument, it returns
Willy Tarreau74ca5042013-06-11 23:12:07 +020015503 the value of the nth given entry value from the beginning/end of the DN.
15504 For instance, "ssl_f_i_dn(OU,2)" the second organization unit, and
15505 "ssl_f_i_dn(CN)" retrieves the common name.
Emeric Brun87855892012-10-17 17:39:35 +020015506
Willy Tarreau74ca5042013-06-11 23:12:07 +020015507ssl_f_key_alg : string
15508 Returns the name of the algorithm used to generate the key of the certificate
15509 presented by the frontend when the incoming connection was made over an
15510 SSL/TLS transport layer.
Emeric Brun7f56e742012-10-19 18:15:40 +020015511
Willy Tarreau74ca5042013-06-11 23:12:07 +020015512ssl_f_notafter : string
15513 Returns the end date presented by the frontend as a formatted string
15514 YYMMDDhhmmss[Z] when the incoming connection was made over an SSL/TLS
15515 transport layer.
Emeric Brun2525b6b2012-10-18 15:59:43 +020015516
Willy Tarreau74ca5042013-06-11 23:12:07 +020015517ssl_f_notbefore : string
15518 Returns the start date presented by the frontend as a formatted string
15519 YYMMDDhhmmss[Z] when the incoming connection was made over an SSL/TLS
15520 transport layer.
Emeric Brun87855892012-10-17 17:39:35 +020015521
Willy Tarreau74ca5042013-06-11 23:12:07 +020015522ssl_f_s_dn([<entry>[,<occ>]]) : string
15523 When the incoming connection was made over an SSL/TLS transport layer,
15524 returns the full distinguished name of the subject of the certificate
15525 presented by the frontend when no <entry> is specified, or the value of the
15526 first given entry found from the beginning of the DN. If a positive/negative
15527 occurrence number is specified as the optional second argument, it returns
15528 the value of the nth given entry value from the beginning/end of the DN.
15529 For instance, "ssl_f_s_dn(OU,2)" the second organization unit, and
15530 "ssl_f_s_dn(CN)" retrieves the common name.
Emeric Brunce5ad802012-10-22 14:11:22 +020015531
Willy Tarreau74ca5042013-06-11 23:12:07 +020015532ssl_f_serial : binary
15533 Returns the serial of the certificate presented by the frontend when the
15534 incoming connection was made over an SSL/TLS transport layer. When used for
15535 an ACL, the value(s) to match against can be passed in hexadecimal form.
Emeric Brun87855892012-10-17 17:39:35 +020015536
Emeric Brun55f4fa82014-04-30 17:11:25 +020015537ssl_f_sha1 : binary
15538 Returns the SHA-1 fingerprint of the certificate presented by the frontend
15539 when the incoming connection was made over an SSL/TLS transport layer. This
15540 can be used to know which certificate was chosen using SNI.
15541
Willy Tarreau74ca5042013-06-11 23:12:07 +020015542ssl_f_sig_alg : string
15543 Returns the name of the algorithm used to sign the certificate presented by
15544 the frontend when the incoming connection was made over an SSL/TLS transport
15545 layer.
Emeric Brun7f56e742012-10-19 18:15:40 +020015546
Willy Tarreau74ca5042013-06-11 23:12:07 +020015547ssl_f_version : integer
15548 Returns the version of the certificate presented by the frontend when the
15549 incoming connection was made over an SSL/TLS transport layer.
15550
15551ssl_fc : boolean
Emeric Brun2525b6b2012-10-18 15:59:43 +020015552 Returns true when the front connection was made via an SSL/TLS transport
15553 layer and is locally deciphered. This means it has matched a socket declared
15554 with a "bind" line having the "ssl" option.
15555
Willy Tarreau74ca5042013-06-11 23:12:07 +020015556 Example :
15557 # This passes "X-Proto: https" to servers when client connects over SSL
15558 listen http-https
15559 bind :80
15560 bind :443 ssl crt /etc/haproxy.pem
15561 http-request add-header X-Proto https if { ssl_fc }
15562
15563ssl_fc_alg_keysize : integer
15564 Returns the symmetric cipher key size supported in bits when the incoming
15565 connection was made over an SSL/TLS transport layer.
15566
15567ssl_fc_alpn : string
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030015568 This extracts the Application Layer Protocol Negotiation field from an
Willy Tarreau74ca5042013-06-11 23:12:07 +020015569 incoming connection made via a TLS transport layer and locally deciphered by
15570 haproxy. The result is a string containing the protocol name advertised by
15571 the client. The SSL library must have been built with support for TLS
15572 extensions enabled (check haproxy -vv). Note that the TLS ALPN extension is
15573 not advertised unless the "alpn" keyword on the "bind" line specifies a
15574 protocol list. Also, nothing forces the client to pick a protocol from this
15575 list, any other one may be requested. The TLS ALPN extension is meant to
15576 replace the TLS NPN extension. See also "ssl_fc_npn".
15577
Willy Tarreau74ca5042013-06-11 23:12:07 +020015578ssl_fc_cipher : string
15579 Returns the name of the used cipher when the incoming connection was made
15580 over an SSL/TLS transport layer.
Willy Tarreauab861d32013-04-02 02:30:41 +020015581
Thierry FOURNIER5bf77322017-02-25 12:45:22 +010015582ssl_fc_cipherlist_bin : binary
15583 Returns the binary form of the client hello cipher list. The maximum returned
15584 value length is according with the value of
Emmanuel Hocdetaaee7502017-03-07 18:34:58 +010015585 "tune.ssl.capture-cipherlist-size".
Thierry FOURNIER5bf77322017-02-25 12:45:22 +010015586
15587ssl_fc_cipherlist_hex : string
15588 Returns the binary form of the client hello cipher list encoded as
15589 hexadecimal. The maximum returned value length is according with the value of
Emmanuel Hocdetaaee7502017-03-07 18:34:58 +010015590 "tune.ssl.capture-cipherlist-size".
Thierry FOURNIER5bf77322017-02-25 12:45:22 +010015591
15592ssl_fc_cipherlist_str : string
15593 Returns the decoded text form of the client hello cipher list. The maximum
15594 number of ciphers returned is according with the value of
15595 "tune.ssl.capture-cipherlist-size". Note that this sample-fetch is only
Davor Ocelice9ed2812017-12-25 17:49:28 +010015596 available with OpenSSL >= 1.0.2. If the function is not enabled, this
Emmanuel Hocdetddcde192017-09-01 17:32:08 +020015597 sample-fetch returns the hash like "ssl_fc_cipherlist_xxh".
Thierry FOURNIER5bf77322017-02-25 12:45:22 +010015598
15599ssl_fc_cipherlist_xxh : integer
15600 Returns a xxh64 of the cipher list. This hash can be return only is the value
15601 "tune.ssl.capture-cipherlist-size" is set greater than 0, however the hash
Emmanuel Hocdetaaee7502017-03-07 18:34:58 +010015602 take in account all the data of the cipher list.
Thierry FOURNIER5bf77322017-02-25 12:45:22 +010015603
Patrick Hemmer65674662019-06-04 08:13:03 -040015604ssl_fc_client_random : binary
15605 Returns the client random of the front connection when the incoming connection
15606 was made over an SSL/TLS transport layer. It is useful to to decrypt traffic
15607 sent using ephemeral ciphers. This requires OpenSSL >= 1.1.0, or BoringSSL.
15608
Willy Tarreau74ca5042013-06-11 23:12:07 +020015609ssl_fc_has_crt : boolean
Emeric Brun2525b6b2012-10-18 15:59:43 +020015610 Returns true if a client certificate is present in an incoming connection over
15611 SSL/TLS transport layer. Useful if 'verify' statement is set to 'optional'.
Emeric Brun9143d372012-12-20 15:44:16 +010015612 Note: on SSL session resumption with Session ID or TLS ticket, client
15613 certificate is not present in the current connection but may be retrieved
15614 from the cache or the ticket. So prefer "ssl_c_used" if you want to check if
15615 current SSL session uses a client certificate.
Emeric Brun2525b6b2012-10-18 15:59:43 +020015616
Olivier Houchardccaa7de2017-10-02 11:51:03 +020015617ssl_fc_has_early : boolean
15618 Returns true if early data were sent, and the handshake didn't happen yet. As
15619 it has security implications, it is useful to be able to refuse those, or
15620 wait until the handshake happened.
15621
Willy Tarreau74ca5042013-06-11 23:12:07 +020015622ssl_fc_has_sni : boolean
15623 This checks for the presence of a Server Name Indication TLS extension (SNI)
Willy Tarreauf7bc57c2012-10-03 00:19:48 +020015624 in an incoming connection was made over an SSL/TLS transport layer. Returns
15625 true when the incoming connection presents a TLS SNI field. This requires
John Roeslerfb2fce12019-07-10 15:45:51 -050015626 that the SSL library is built with support for TLS extensions enabled (check
Willy Tarreauf7bc57c2012-10-03 00:19:48 +020015627 haproxy -vv).
Willy Tarreau7875d092012-09-10 08:20:03 +020015628
Nenad Merdanovic1516fe32016-05-17 03:31:21 +020015629ssl_fc_is_resumed : boolean
Nenad Merdanovic26ea8222015-05-18 02:28:57 +020015630 Returns true if the SSL/TLS session has been resumed through the use of
Jérôme Magnin4a326cb2018-01-15 14:01:17 +010015631 SSL session cache or TLS tickets on an incoming connection over an SSL/TLS
15632 transport layer.
Nenad Merdanovic26ea8222015-05-18 02:28:57 +020015633
Willy Tarreau74ca5042013-06-11 23:12:07 +020015634ssl_fc_npn : string
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030015635 This extracts the Next Protocol Negotiation field from an incoming connection
Willy Tarreau74ca5042013-06-11 23:12:07 +020015636 made via a TLS transport layer and locally deciphered by haproxy. The result
15637 is a string containing the protocol name advertised by the client. The SSL
15638 library must have been built with support for TLS extensions enabled (check
15639 haproxy -vv). Note that the TLS NPN extension is not advertised unless the
15640 "npn" keyword on the "bind" line specifies a protocol list. Also, nothing
15641 forces the client to pick a protocol from this list, any other one may be
15642 requested. Please note that the TLS NPN extension was replaced with ALPN.
Willy Tarreaua33c6542012-10-15 13:19:06 +020015643
Willy Tarreau74ca5042013-06-11 23:12:07 +020015644ssl_fc_protocol : string
15645 Returns the name of the used protocol when the incoming connection was made
15646 over an SSL/TLS transport layer.
Willy Tarreau7875d092012-09-10 08:20:03 +020015647
Emeric Brunb73a9b02014-04-30 18:49:19 +020015648ssl_fc_unique_id : binary
David Sc1ad52e2014-04-08 18:48:47 -040015649 When the incoming connection was made over an SSL/TLS transport layer,
Emeric Brunb73a9b02014-04-30 18:49:19 +020015650 returns the TLS unique ID as defined in RFC5929 section 3. The unique id
15651 can be encoded to base64 using the converter: "ssl_bc_unique_id,base64".
David Sc1ad52e2014-04-08 18:48:47 -040015652
Patrick Hemmer65674662019-06-04 08:13:03 -040015653ssl_fc_server_random : binary
15654 Returns the server random of the front connection when the incoming connection
15655 was made over an SSL/TLS transport layer. It is useful to to decrypt traffic
15656 sent using ephemeral ciphers. This requires OpenSSL >= 1.1.0, or BoringSSL.
15657
Willy Tarreau74ca5042013-06-11 23:12:07 +020015658ssl_fc_session_id : binary
15659 Returns the SSL ID of the front connection when the incoming connection was
15660 made over an SSL/TLS transport layer. It is useful to stick a given client to
15661 a server. It is important to note that some browsers refresh their session ID
15662 every few minutes.
Willy Tarreau7875d092012-09-10 08:20:03 +020015663
Patrick Hemmere0275472018-04-28 19:15:51 -040015664ssl_fc_session_key : binary
15665 Returns the SSL session master key of the front connection when the incoming
15666 connection was made over an SSL/TLS transport layer. It is useful to decrypt
15667 traffic sent using ephemeral ciphers. This requires OpenSSL >= 1.1.0, or
15668 BoringSSL.
15669
15670
Willy Tarreau74ca5042013-06-11 23:12:07 +020015671ssl_fc_sni : string
15672 This extracts the Server Name Indication TLS extension (SNI) field from an
15673 incoming connection made via an SSL/TLS transport layer and locally
15674 deciphered by haproxy. The result (when present) typically is a string
15675 matching the HTTPS host name (253 chars or less). The SSL library must have
15676 been built with support for TLS extensions enabled (check haproxy -vv).
15677
15678 This fetch is different from "req_ssl_sni" above in that it applies to the
15679 connection being deciphered by haproxy and not to SSL contents being blindly
15680 forwarded. See also "ssl_fc_sni_end" and "ssl_fc_sni_reg" below. This
John Roeslerfb2fce12019-07-10 15:45:51 -050015681 requires that the SSL library is built with support for TLS extensions
Cyril Bonté9c1eb1e2012-10-09 22:45:34 +020015682 enabled (check haproxy -vv).
Willy Tarreau62644772008-07-16 18:36:06 +020015683
Willy Tarreau74ca5042013-06-11 23:12:07 +020015684 ACL derivatives :
Willy Tarreau74ca5042013-06-11 23:12:07 +020015685 ssl_fc_sni_end : suffix match
15686 ssl_fc_sni_reg : regex match
Emeric Brun589fcad2012-10-16 14:13:26 +020015687
Willy Tarreau74ca5042013-06-11 23:12:07 +020015688ssl_fc_use_keysize : integer
15689 Returns the symmetric cipher key size used in bits when the incoming
15690 connection was made over an SSL/TLS transport layer.
Willy Tarreaub6fb4202008-07-20 11:18:28 +020015691
Willy Tarreaub6fb4202008-07-20 11:18:28 +020015692
Thierry FOURNIER060762e2014-04-23 13:29:15 +0200156937.3.5. Fetching samples from buffer contents (Layer 6)
Willy Tarreau74ca5042013-06-11 23:12:07 +020015694------------------------------------------------------
Willy Tarreaub6fb4202008-07-20 11:18:28 +020015695
Willy Tarreau74ca5042013-06-11 23:12:07 +020015696Fetching samples from buffer contents is a bit different from the previous
15697sample fetches above because the sampled data are ephemeral. These data can
15698only be used when they're available and will be lost when they're forwarded.
15699For this reason, samples fetched from buffer contents during a request cannot
15700be used in a response for example. Even while the data are being fetched, they
15701can change. Sometimes it is necessary to set some delays or combine multiple
15702sample fetch methods to ensure that the expected data are complete and usable,
15703for example through TCP request content inspection. Please see the "tcp-request
15704content" keyword for more detailed information on the subject.
Willy Tarreau62644772008-07-16 18:36:06 +020015705
Willy Tarreau74ca5042013-06-11 23:12:07 +020015706payload(<offset>,<length>) : binary (deprecated)
Davor Ocelice9ed2812017-12-25 17:49:28 +010015707 This is an alias for "req.payload" when used in the context of a request (e.g.
Willy Tarreau74ca5042013-06-11 23:12:07 +020015708 "stick on", "stick match"), and for "res.payload" when used in the context of
15709 a response such as in "stick store response".
Willy Tarreau0ba27502007-12-24 16:55:16 +010015710
Willy Tarreau74ca5042013-06-11 23:12:07 +020015711payload_lv(<offset1>,<length>[,<offset2>]) : binary (deprecated)
15712 This is an alias for "req.payload_lv" when used in the context of a request
Davor Ocelice9ed2812017-12-25 17:49:28 +010015713 (e.g. "stick on", "stick match"), and for "res.payload_lv" when used in the
Willy Tarreau74ca5042013-06-11 23:12:07 +020015714 context of a response such as in "stick store response".
Willy Tarreau0ba27502007-12-24 16:55:16 +010015715
Thierry FOURNIERd7d88812017-04-19 15:15:14 +020015716req.hdrs : string
15717 Returns the current request headers as string including the last empty line
15718 separating headers from the request body. The last empty line can be used to
15719 detect a truncated header block. This sample fetch is useful for some SPOE
15720 headers analyzers and for advanced logging.
15721
Thierry FOURNIER5617dce2017-04-09 05:38:19 +020015722req.hdrs_bin : binary
15723 Returns the current request headers contained in preparsed binary form. This
15724 is useful for offloading some processing with SPOE. Each string is described
15725 by a length followed by the number of bytes indicated in the length. The
15726 length is represented using the variable integer encoding detailed in the
15727 SPOE documentation. The end of the list is marked by a couple of empty header
15728 names and values (length of 0 for both).
15729
15730 *(<str:header-name><str:header-value>)<empty string><empty string>
15731
15732 int: refer to the SPOE documentation for the encoding
15733 str: <int:length><bytes>
15734
Willy Tarreau74ca5042013-06-11 23:12:07 +020015735req.len : integer
15736req_len : integer (deprecated)
15737 Returns an integer value corresponding to the number of bytes present in the
15738 request buffer. This is mostly used in ACL. It is important to understand
15739 that this test does not return false as long as the buffer is changing. This
15740 means that a check with equality to zero will almost always immediately match
15741 at the beginning of the session, while a test for more data will wait for
15742 that data to come in and return false only when haproxy is certain that no
15743 more data will come in. This test was designed to be used with TCP request
15744 content inspection.
Willy Tarreaua7ad50c2012-04-29 15:39:40 +020015745
Willy Tarreau74ca5042013-06-11 23:12:07 +020015746req.payload(<offset>,<length>) : binary
15747 This extracts a binary block of <length> bytes and starting at byte <offset>
Willy Tarreau00f00842013-08-02 11:07:32 +020015748 in the request buffer. As a special case, if the <length> argument is zero,
15749 the the whole buffer from <offset> to the end is extracted. This can be used
15750 with ACLs in order to check for the presence of some content in a buffer at
15751 any location.
Willy Tarreaua7ad50c2012-04-29 15:39:40 +020015752
Willy Tarreau74ca5042013-06-11 23:12:07 +020015753 ACL alternatives :
15754 payload(<offset>,<length>) : hex binary match
Willy Tarreaua7ad50c2012-04-29 15:39:40 +020015755
Willy Tarreau74ca5042013-06-11 23:12:07 +020015756req.payload_lv(<offset1>,<length>[,<offset2>]) : binary
15757 This extracts a binary block whose size is specified at <offset1> for <length>
15758 bytes, and which starts at <offset2> if specified or just after the length in
15759 the request buffer. The <offset2> parameter also supports relative offsets if
15760 prepended with a '+' or '-' sign.
Willy Tarreaua7ad50c2012-04-29 15:39:40 +020015761
Willy Tarreau74ca5042013-06-11 23:12:07 +020015762 ACL alternatives :
15763 payload_lv(<offset1>,<length>[,<offset2>]) : hex binary match
Willy Tarreaua7ad50c2012-04-29 15:39:40 +020015764
Willy Tarreau74ca5042013-06-11 23:12:07 +020015765 Example : please consult the example from the "stick store-response" keyword.
Willy Tarreaua7ad50c2012-04-29 15:39:40 +020015766
Willy Tarreau74ca5042013-06-11 23:12:07 +020015767req.proto_http : boolean
15768req_proto_http : boolean (deprecated)
15769 Returns true when data in the request buffer look like HTTP and correctly
15770 parses as such. It is the same parser as the common HTTP request parser which
15771 is used so there should be no surprises. The test does not match until the
15772 request is complete, failed or timed out. This test may be used to report the
15773 protocol in TCP logs, but the biggest use is to block TCP request analysis
15774 until a complete HTTP request is present in the buffer, for example to track
15775 a header.
Willy Tarreaua7ad50c2012-04-29 15:39:40 +020015776
Willy Tarreau74ca5042013-06-11 23:12:07 +020015777 Example:
15778 # track request counts per "base" (concatenation of Host+URL)
15779 tcp-request inspect-delay 10s
15780 tcp-request content reject if !HTTP
Willy Tarreaube4a3ef2013-06-17 15:04:07 +020015781 tcp-request content track-sc0 base table req-rate
Willy Tarreaua7ad50c2012-04-29 15:39:40 +020015782
Willy Tarreau74ca5042013-06-11 23:12:07 +020015783req.rdp_cookie([<name>]) : string
15784rdp_cookie([<name>]) : string (deprecated)
15785 When the request buffer looks like the RDP protocol, extracts the RDP cookie
15786 <name>, or any cookie if unspecified. The parser only checks for the first
15787 cookie, as illustrated in the RDP protocol specification. The cookie name is
15788 case insensitive. Generally the "MSTS" cookie name will be used, as it can
15789 contain the user name of the client connecting to the server if properly
15790 configured on the client. The "MSTSHASH" cookie is often used as well for
15791 session stickiness to servers.
Willy Tarreau04aa6a92012-04-06 18:57:55 +020015792
Willy Tarreau74ca5042013-06-11 23:12:07 +020015793 This differs from "balance rdp-cookie" in that any balancing algorithm may be
15794 used and thus the distribution of clients to backend servers is not linked to
15795 a hash of the RDP cookie. It is envisaged that using a balancing algorithm
15796 such as "balance roundrobin" or "balance leastconn" will lead to a more even
15797 distribution of clients to backend servers than the hash used by "balance
15798 rdp-cookie".
Willy Tarreau04aa6a92012-04-06 18:57:55 +020015799
Willy Tarreau74ca5042013-06-11 23:12:07 +020015800 ACL derivatives :
15801 req_rdp_cookie([<name>]) : exact string match
Willy Tarreau04aa6a92012-04-06 18:57:55 +020015802
Willy Tarreau74ca5042013-06-11 23:12:07 +020015803 Example :
15804 listen tse-farm
15805 bind 0.0.0.0:3389
15806 # wait up to 5s for an RDP cookie in the request
15807 tcp-request inspect-delay 5s
15808 tcp-request content accept if RDP_COOKIE
15809 # apply RDP cookie persistence
15810 persist rdp-cookie
15811 # Persist based on the mstshash cookie
15812 # This is only useful makes sense if
15813 # balance rdp-cookie is not used
15814 stick-table type string size 204800
15815 stick on req.rdp_cookie(mstshash)
15816 server srv1 1.1.1.1:3389
15817 server srv1 1.1.1.2:3389
Willy Tarreau04aa6a92012-04-06 18:57:55 +020015818
Willy Tarreau74ca5042013-06-11 23:12:07 +020015819 See also : "balance rdp-cookie", "persist rdp-cookie", "tcp-request" and the
15820 "req_rdp_cookie" ACL.
Willy Tarreau04aa6a92012-04-06 18:57:55 +020015821
Willy Tarreau74ca5042013-06-11 23:12:07 +020015822req.rdp_cookie_cnt([name]) : integer
15823rdp_cookie_cnt([name]) : integer (deprecated)
15824 Tries to parse the request buffer as RDP protocol, then returns an integer
15825 corresponding to the number of RDP cookies found. If an optional cookie name
15826 is passed, only cookies matching this name are considered. This is mostly
15827 used in ACL.
Willy Tarreau04aa6a92012-04-06 18:57:55 +020015828
Willy Tarreau74ca5042013-06-11 23:12:07 +020015829 ACL derivatives :
15830 req_rdp_cookie_cnt([<name>]) : integer match
Willy Tarreau04aa6a92012-04-06 18:57:55 +020015831
Alex Zorin4afdd132018-12-30 13:56:28 +110015832req.ssl_alpn : string
15833 Returns a string containing the values of the Application-Layer Protocol
15834 Negotiation (ALPN) TLS extension (RFC7301), sent by the client within the SSL
15835 ClientHello message. Note that this only applies to raw contents found in the
15836 request buffer and not to the contents deciphered via an SSL data layer, so
15837 this will not work with "bind" lines having the "ssl" option. This is useful
15838 in ACL to make a routing decision based upon the ALPN preferences of a TLS
Jarno Huuskonene504f812019-01-03 07:56:49 +020015839 client, like in the example below. See also "ssl_fc_alpn".
Alex Zorin4afdd132018-12-30 13:56:28 +110015840
15841 Examples :
15842 # Wait for a client hello for at most 5 seconds
15843 tcp-request inspect-delay 5s
15844 tcp-request content accept if { req_ssl_hello_type 1 }
Jarno Huuskonene504f812019-01-03 07:56:49 +020015845 use_backend bk_acme if { req.ssl_alpn acme-tls/1 }
Alex Zorin4afdd132018-12-30 13:56:28 +110015846 default_backend bk_default
15847
Nenad Merdanovic5fc7d7e2015-07-07 22:00:17 +020015848req.ssl_ec_ext : boolean
15849 Returns a boolean identifying if client sent the Supported Elliptic Curves
15850 Extension as defined in RFC4492, section 5.1. within the SSL ClientHello
Cyril Bonté307ee1e2015-09-28 23:16:06 +020015851 message. This can be used to present ECC compatible clients with EC
15852 certificate and to use RSA for all others, on the same IP address. Note that
15853 this only applies to raw contents found in the request buffer and not to
15854 contents deciphered via an SSL data layer, so this will not work with "bind"
15855 lines having the "ssl" option.
Nenad Merdanovic5fc7d7e2015-07-07 22:00:17 +020015856
Willy Tarreau74ca5042013-06-11 23:12:07 +020015857req.ssl_hello_type : integer
15858req_ssl_hello_type : integer (deprecated)
15859 Returns an integer value containing the type of the SSL hello message found
15860 in the request buffer if the buffer contains data that parse as a complete
15861 SSL (v3 or superior) client hello message. Note that this only applies to raw
15862 contents found in the request buffer and not to contents deciphered via an
15863 SSL data layer, so this will not work with "bind" lines having the "ssl"
15864 option. This is mostly used in ACL to detect presence of an SSL hello message
15865 that is supposed to contain an SSL session ID usable for stickiness.
Willy Tarreau04aa6a92012-04-06 18:57:55 +020015866
Willy Tarreau74ca5042013-06-11 23:12:07 +020015867req.ssl_sni : string
15868req_ssl_sni : string (deprecated)
15869 Returns a string containing the value of the Server Name TLS extension sent
15870 by a client in a TLS stream passing through the request buffer if the buffer
15871 contains data that parse as a complete SSL (v3 or superior) client hello
15872 message. Note that this only applies to raw contents found in the request
15873 buffer and not to contents deciphered via an SSL data layer, so this will not
15874 work with "bind" lines having the "ssl" option. SNI normally contains the
15875 name of the host the client tries to connect to (for recent browsers). SNI is
15876 useful for allowing or denying access to certain hosts when SSL/TLS is used
15877 by the client. This test was designed to be used with TCP request content
15878 inspection. If content switching is needed, it is recommended to first wait
15879 for a complete client hello (type 1), like in the example below. See also
15880 "ssl_fc_sni".
Willy Tarreau04aa6a92012-04-06 18:57:55 +020015881
Willy Tarreau74ca5042013-06-11 23:12:07 +020015882 ACL derivatives :
15883 req_ssl_sni : exact string match
Willy Tarreau04aa6a92012-04-06 18:57:55 +020015884
Willy Tarreau74ca5042013-06-11 23:12:07 +020015885 Examples :
15886 # Wait for a client hello for at most 5 seconds
15887 tcp-request inspect-delay 5s
15888 tcp-request content accept if { req_ssl_hello_type 1 }
15889 use_backend bk_allow if { req_ssl_sni -f allowed_sites }
15890 default_backend bk_sorry_page
Willy Tarreau04aa6a92012-04-06 18:57:55 +020015891
Pradeep Jindalbb2acf52015-09-29 10:12:57 +053015892req.ssl_st_ext : integer
15893 Returns 0 if the client didn't send a SessionTicket TLS Extension (RFC5077)
15894 Returns 1 if the client sent SessionTicket TLS Extension
15895 Returns 2 if the client also sent non-zero length TLS SessionTicket
15896 Note that this only applies to raw contents found in the request buffer and
15897 not to contents deciphered via an SSL data layer, so this will not work with
15898 "bind" lines having the "ssl" option. This can for example be used to detect
15899 whether the client sent a SessionTicket or not and stick it accordingly, if
15900 no SessionTicket then stick on SessionID or don't stick as there's no server
15901 side state is there when SessionTickets are in use.
15902
Willy Tarreau74ca5042013-06-11 23:12:07 +020015903req.ssl_ver : integer
15904req_ssl_ver : integer (deprecated)
15905 Returns an integer value containing the version of the SSL/TLS protocol of a
15906 stream present in the request buffer. Both SSLv2 hello messages and SSLv3
15907 messages are supported. TLSv1 is announced as SSL version 3.1. The value is
15908 composed of the major version multiplied by 65536, added to the minor
15909 version. Note that this only applies to raw contents found in the request
15910 buffer and not to contents deciphered via an SSL data layer, so this will not
15911 work with "bind" lines having the "ssl" option. The ACL version of the test
Davor Ocelice9ed2812017-12-25 17:49:28 +010015912 matches against a decimal notation in the form MAJOR.MINOR (e.g. 3.1). This
Willy Tarreau74ca5042013-06-11 23:12:07 +020015913 fetch is mostly used in ACL.
Willy Tarreaud63335a2010-02-26 12:56:52 +010015914
Willy Tarreau74ca5042013-06-11 23:12:07 +020015915 ACL derivatives :
15916 req_ssl_ver : decimal match
Willy Tarreaud63335a2010-02-26 12:56:52 +010015917
Willy Tarreau47e8eba2013-09-11 23:28:46 +020015918res.len : integer
15919 Returns an integer value corresponding to the number of bytes present in the
15920 response buffer. This is mostly used in ACL. It is important to understand
15921 that this test does not return false as long as the buffer is changing. This
15922 means that a check with equality to zero will almost always immediately match
15923 at the beginning of the session, while a test for more data will wait for
15924 that data to come in and return false only when haproxy is certain that no
15925 more data will come in. This test was designed to be used with TCP response
15926 content inspection.
15927
Willy Tarreau74ca5042013-06-11 23:12:07 +020015928res.payload(<offset>,<length>) : binary
15929 This extracts a binary block of <length> bytes and starting at byte <offset>
Willy Tarreau00f00842013-08-02 11:07:32 +020015930 in the response buffer. As a special case, if the <length> argument is zero,
15931 the the whole buffer from <offset> to the end is extracted. This can be used
15932 with ACLs in order to check for the presence of some content in a buffer at
15933 any location.
Willy Tarreaud63335a2010-02-26 12:56:52 +010015934
Willy Tarreau74ca5042013-06-11 23:12:07 +020015935res.payload_lv(<offset1>,<length>[,<offset2>]) : binary
15936 This extracts a binary block whose size is specified at <offset1> for <length>
15937 bytes, and which starts at <offset2> if specified or just after the length in
15938 the response buffer. The <offset2> parameter also supports relative offsets
15939 if prepended with a '+' or '-' sign.
Willy Tarreaud63335a2010-02-26 12:56:52 +010015940
Willy Tarreau74ca5042013-06-11 23:12:07 +020015941 Example : please consult the example from the "stick store-response" keyword.
Willy Tarreaud63335a2010-02-26 12:56:52 +010015942
Willy Tarreau971f7b62015-09-29 14:06:59 +020015943res.ssl_hello_type : integer
15944rep_ssl_hello_type : integer (deprecated)
15945 Returns an integer value containing the type of the SSL hello message found
15946 in the response buffer if the buffer contains data that parses as a complete
15947 SSL (v3 or superior) hello message. Note that this only applies to raw
15948 contents found in the response buffer and not to contents deciphered via an
15949 SSL data layer, so this will not work with "server" lines having the "ssl"
15950 option. This is mostly used in ACL to detect presence of an SSL hello message
15951 that is supposed to contain an SSL session ID usable for stickiness.
15952
Willy Tarreau74ca5042013-06-11 23:12:07 +020015953wait_end : boolean
15954 This fetch either returns true when the inspection period is over, or does
15955 not fetch. It is only used in ACLs, in conjunction with content analysis to
Davor Ocelice9ed2812017-12-25 17:49:28 +010015956 avoid returning a wrong verdict early. It may also be used to delay some
Willy Tarreau74ca5042013-06-11 23:12:07 +020015957 actions, such as a delayed reject for some special addresses. Since it either
15958 stops the rules evaluation or immediately returns true, it is recommended to
Davor Ocelice9ed2812017-12-25 17:49:28 +010015959 use this acl as the last one in a rule. Please note that the default ACL
Willy Tarreau74ca5042013-06-11 23:12:07 +020015960 "WAIT_END" is always usable without prior declaration. This test was designed
15961 to be used with TCP request content inspection.
Willy Tarreaud63335a2010-02-26 12:56:52 +010015962
Willy Tarreau74ca5042013-06-11 23:12:07 +020015963 Examples :
15964 # delay every incoming request by 2 seconds
15965 tcp-request inspect-delay 2s
15966 tcp-request content accept if WAIT_END
Willy Tarreaud63335a2010-02-26 12:56:52 +010015967
Willy Tarreau74ca5042013-06-11 23:12:07 +020015968 # don't immediately tell bad guys they are rejected
15969 tcp-request inspect-delay 10s
15970 acl goodguys src 10.0.0.0/24
15971 acl badguys src 10.0.1.0/24
15972 tcp-request content accept if goodguys
15973 tcp-request content reject if badguys WAIT_END
15974 tcp-request content reject
15975
15976
Thierry FOURNIER060762e2014-04-23 13:29:15 +0200159777.3.6. Fetching HTTP samples (Layer 7)
Willy Tarreau74ca5042013-06-11 23:12:07 +020015978--------------------------------------
15979
15980It is possible to fetch samples from HTTP contents, requests and responses.
15981This application layer is also called layer 7. It is only possible to fetch the
15982data in this section when a full HTTP request or response has been parsed from
15983its respective request or response buffer. This is always the case with all
15984HTTP specific rules and for sections running with "mode http". When using TCP
15985content inspection, it may be necessary to support an inspection delay in order
15986to let the request or response come in first. These fetches may require a bit
15987more CPU resources than the layer 4 ones, but not much since the request and
15988response are indexed.
15989
15990base : string
15991 This returns the concatenation of the first Host header and the path part of
15992 the request, which starts at the first slash and ends before the question
15993 mark. It can be useful in virtual hosted environments to detect URL abuses as
15994 well as to improve shared caches efficiency. Using this with a limited size
15995 stick table also allows one to collect statistics about most commonly
15996 requested objects by host/path. With ACLs it can allow simple content
15997 switching rules involving the host and the path at the same time, such as
15998 "www.example.com/favicon.ico". See also "path" and "uri".
15999
16000 ACL derivatives :
16001 base : exact string match
16002 base_beg : prefix match
16003 base_dir : subdir match
16004 base_dom : domain match
16005 base_end : suffix match
16006 base_len : length match
16007 base_reg : regex match
16008 base_sub : substring match
16009
16010base32 : integer
16011 This returns a 32-bit hash of the value returned by the "base" fetch method
16012 above. This is useful to track per-URL activity on high traffic sites without
16013 having to store all URLs. Instead a shorter hash is stored, saving a lot of
Willy Tarreau23ec4ca2014-07-15 20:15:37 +020016014 memory. The output type is an unsigned integer. The hash function used is
16015 SDBM with full avalanche on the output. Technically, base32 is exactly equal
16016 to "base,sdbm(1)".
Willy Tarreau74ca5042013-06-11 23:12:07 +020016017
16018base32+src : binary
16019 This returns the concatenation of the base32 fetch above and the src fetch
16020 below. The resulting type is of type binary, with a size of 8 or 20 bytes
16021 depending on the source address family. This can be used to track per-IP,
16022 per-URL counters.
16023
William Lallemand65ad6e12014-01-31 15:08:02 +010016024capture.req.hdr(<idx>) : string
16025 This extracts the content of the header captured by the "capture request
16026 header", idx is the position of the capture keyword in the configuration.
16027 The first entry is an index of 0. See also: "capture request header".
16028
16029capture.req.method : string
16030 This extracts the METHOD of an HTTP request. It can be used in both request
16031 and response. Unlike "method", it can be used in both request and response
16032 because it's allocated.
16033
16034capture.req.uri : string
16035 This extracts the request's URI, which starts at the first slash and ends
16036 before the first space in the request (without the host part). Unlike "path"
16037 and "url", it can be used in both request and response because it's
16038 allocated.
16039
Willy Tarreau3c1b5ec2014-04-24 23:41:57 +020016040capture.req.ver : string
16041 This extracts the request's HTTP version and returns either "HTTP/1.0" or
16042 "HTTP/1.1". Unlike "req.ver", it can be used in both request, response, and
16043 logs because it relies on a persistent flag.
16044
William Lallemand65ad6e12014-01-31 15:08:02 +010016045capture.res.hdr(<idx>) : string
16046 This extracts the content of the header captured by the "capture response
16047 header", idx is the position of the capture keyword in the configuration.
16048 The first entry is an index of 0.
16049 See also: "capture response header"
16050
Willy Tarreau3c1b5ec2014-04-24 23:41:57 +020016051capture.res.ver : string
16052 This extracts the response's HTTP version and returns either "HTTP/1.0" or
16053 "HTTP/1.1". Unlike "res.ver", it can be used in logs because it relies on a
16054 persistent flag.
16055
Willy Tarreaua5910cc2015-05-02 00:46:08 +020016056req.body : binary
16057 This returns the HTTP request's available body as a block of data. It
16058 requires that the request body has been buffered made available using
16059 "option http-buffer-request". In case of chunked-encoded body, currently only
16060 the first chunk is analyzed.
16061
Thierry FOURNIER9826c772015-05-20 15:50:54 +020016062req.body_param([<name>) : string
16063 This fetch assumes that the body of the POST request is url-encoded. The user
16064 can check if the "content-type" contains the value
16065 "application/x-www-form-urlencoded". This extracts the first occurrence of the
16066 parameter <name> in the body, which ends before '&'. The parameter name is
16067 case-sensitive. If no name is given, any parameter will match, and the first
16068 one will be returned. The result is a string corresponding to the value of the
16069 parameter <name> as presented in the request body (no URL decoding is
16070 performed). Note that the ACL version of this fetch iterates over multiple
16071 parameters and will iteratively report all parameters values if no name is
16072 given.
16073
Willy Tarreaua5910cc2015-05-02 00:46:08 +020016074req.body_len : integer
16075 This returns the length of the HTTP request's available body in bytes. It may
16076 be lower than the advertised length if the body is larger than the buffer. It
16077 requires that the request body has been buffered made available using
16078 "option http-buffer-request".
16079
16080req.body_size : integer
16081 This returns the advertised length of the HTTP request's body in bytes. It
16082 will represent the advertised Content-Length header, or the size of the first
16083 chunk in case of chunked encoding. In order to parse the chunks, it requires
16084 that the request body has been buffered made available using
16085 "option http-buffer-request".
16086
Willy Tarreau74ca5042013-06-11 23:12:07 +020016087req.cook([<name>]) : string
16088cook([<name>]) : string (deprecated)
16089 This extracts the last occurrence of the cookie name <name> on a "Cookie"
16090 header line from the request, and returns its value as string. If no name is
16091 specified, the first cookie value is returned. When used with ACLs, all
16092 matching cookies are evaluated. Spaces around the name and the value are
16093 ignored as requested by the Cookie header specification (RFC6265). The cookie
16094 name is case-sensitive. Empty cookies are valid, so an empty cookie may very
16095 well return an empty value if it is present. Use the "found" match to detect
16096 presence. Use the res.cook() variant for response cookies sent by the server.
16097
16098 ACL derivatives :
16099 cook([<name>]) : exact string match
16100 cook_beg([<name>]) : prefix match
16101 cook_dir([<name>]) : subdir match
16102 cook_dom([<name>]) : domain match
16103 cook_end([<name>]) : suffix match
16104 cook_len([<name>]) : length match
16105 cook_reg([<name>]) : regex match
16106 cook_sub([<name>]) : substring match
Willy Tarreaud63335a2010-02-26 12:56:52 +010016107
Willy Tarreau74ca5042013-06-11 23:12:07 +020016108req.cook_cnt([<name>]) : integer
16109cook_cnt([<name>]) : integer (deprecated)
16110 Returns an integer value representing the number of occurrences of the cookie
16111 <name> in the request, or all cookies if <name> is not specified.
Willy Tarreaud63335a2010-02-26 12:56:52 +010016112
Willy Tarreau74ca5042013-06-11 23:12:07 +020016113req.cook_val([<name>]) : integer
16114cook_val([<name>]) : integer (deprecated)
16115 This extracts the last occurrence of the cookie name <name> on a "Cookie"
16116 header line from the request, and converts its value to an integer which is
16117 returned. If no name is specified, the first cookie value is returned. When
16118 used in ACLs, all matching names are iterated over until a value matches.
Willy Tarreau0e698542011-09-16 08:32:32 +020016119
Willy Tarreau74ca5042013-06-11 23:12:07 +020016120cookie([<name>]) : string (deprecated)
16121 This extracts the last occurrence of the cookie name <name> on a "Cookie"
16122 header line from the request, or a "Set-Cookie" header from the response, and
16123 returns its value as a string. A typical use is to get multiple clients
16124 sharing a same profile use the same server. This can be similar to what
Willy Tarreau294d0f02015-08-10 19:40:12 +020016125 "appsession" did with the "request-learn" statement, but with support for
Willy Tarreau74ca5042013-06-11 23:12:07 +020016126 multi-peer synchronization and state keeping across restarts. If no name is
16127 specified, the first cookie value is returned. This fetch should not be used
16128 anymore and should be replaced by req.cook() or res.cook() instead as it
16129 ambiguously uses the direction based on the context where it is used.
Willy Tarreaud63335a2010-02-26 12:56:52 +010016130
Willy Tarreau74ca5042013-06-11 23:12:07 +020016131hdr([<name>[,<occ>]]) : string
16132 This is equivalent to req.hdr() when used on requests, and to res.hdr() when
16133 used on responses. Please refer to these respective fetches for more details.
16134 In case of doubt about the fetch direction, please use the explicit ones.
16135 Note that contrary to the hdr() sample fetch method, the hdr_* ACL keywords
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030016136 unambiguously apply to the request headers.
Willy Tarreaud63335a2010-02-26 12:56:52 +010016137
Willy Tarreau74ca5042013-06-11 23:12:07 +020016138req.fhdr(<name>[,<occ>]) : string
16139 This extracts the last occurrence of header <name> in an HTTP request. When
16140 used from an ACL, all occurrences are iterated over until a match is found.
16141 Optionally, a specific occurrence might be specified as a position number.
16142 Positive values indicate a position from the first occurrence, with 1 being
16143 the first one. Negative values indicate positions relative to the last one,
16144 with -1 being the last one. It differs from req.hdr() in that any commas
16145 present in the value are returned and are not used as delimiters. This is
16146 sometimes useful with headers such as User-Agent.
Willy Tarreaud63335a2010-02-26 12:56:52 +010016147
Willy Tarreau74ca5042013-06-11 23:12:07 +020016148req.fhdr_cnt([<name>]) : integer
16149 Returns an integer value representing the number of occurrences of request
16150 header field name <name>, or the total number of header fields if <name> is
16151 not specified. Contrary to its req.hdr_cnt() cousin, this function returns
16152 the number of full line headers and does not stop on commas.
Willy Tarreaud63335a2010-02-26 12:56:52 +010016153
Willy Tarreau74ca5042013-06-11 23:12:07 +020016154req.hdr([<name>[,<occ>]]) : string
16155 This extracts the last occurrence of header <name> in an HTTP request. When
16156 used from an ACL, all occurrences are iterated over until a match is found.
16157 Optionally, a specific occurrence might be specified as a position number.
16158 Positive values indicate a position from the first occurrence, with 1 being
16159 the first one. Negative values indicate positions relative to the last one,
16160 with -1 being the last one. A typical use is with the X-Forwarded-For header
16161 once converted to IP, associated with an IP stick-table. The function
16162 considers any comma as a delimiter for distinct values. If full-line headers
Lukas Tribus23953682017-04-28 13:24:30 +000016163 are desired instead, use req.fhdr(). Please carefully check RFC7231 to know
Willy Tarreau74ca5042013-06-11 23:12:07 +020016164 how certain headers are supposed to be parsed. Also, some of them are case
Davor Ocelice9ed2812017-12-25 17:49:28 +010016165 insensitive (e.g. Connection).
Willy Tarreaud63335a2010-02-26 12:56:52 +010016166
Willy Tarreau74ca5042013-06-11 23:12:07 +020016167 ACL derivatives :
16168 hdr([<name>[,<occ>]]) : exact string match
16169 hdr_beg([<name>[,<occ>]]) : prefix match
16170 hdr_dir([<name>[,<occ>]]) : subdir match
16171 hdr_dom([<name>[,<occ>]]) : domain match
16172 hdr_end([<name>[,<occ>]]) : suffix match
16173 hdr_len([<name>[,<occ>]]) : length match
16174 hdr_reg([<name>[,<occ>]]) : regex match
16175 hdr_sub([<name>[,<occ>]]) : substring match
16176
16177req.hdr_cnt([<name>]) : integer
16178hdr_cnt([<header>]) : integer (deprecated)
16179 Returns an integer value representing the number of occurrences of request
16180 header field name <name>, or the total number of header field values if
16181 <name> is not specified. It is important to remember that one header line may
16182 count as several headers if it has several values. The function considers any
16183 comma as a delimiter for distinct values. If full-line headers are desired
16184 instead, req.fhdr_cnt() should be used instead. With ACLs, it can be used to
16185 detect presence, absence or abuse of a specific header, as well as to block
16186 request smuggling attacks by rejecting requests which contain more than one
16187 of certain headers. See "req.hdr" for more information on header matching.
16188
16189req.hdr_ip([<name>[,<occ>]]) : ip
16190hdr_ip([<name>[,<occ>]]) : ip (deprecated)
16191 This extracts the last occurrence of header <name> in an HTTP request,
16192 converts it to an IPv4 or IPv6 address and returns this address. When used
16193 with ACLs, all occurrences are checked, and if <name> is omitted, every value
16194 of every header is checked. Optionally, a specific occurrence might be
16195 specified as a position number. Positive values indicate a position from the
Davor Ocelice9ed2812017-12-25 17:49:28 +010016196 first occurrence, with 1 being the first one. Negative values indicate
Willy Tarreau74ca5042013-06-11 23:12:07 +020016197 positions relative to the last one, with -1 being the last one. A typical use
16198 is with the X-Forwarded-For and X-Client-IP headers.
16199
16200req.hdr_val([<name>[,<occ>]]) : integer
16201hdr_val([<name>[,<occ>]]) : integer (deprecated)
16202 This extracts the last occurrence of header <name> in an HTTP request, and
16203 converts it to an integer value. When used with ACLs, all occurrences are
16204 checked, and if <name> is omitted, every value of every header is checked.
16205 Optionally, a specific occurrence might be specified as a position number.
16206 Positive values indicate a position from the first occurrence, with 1 being
16207 the first one. Negative values indicate positions relative to the last one,
16208 with -1 being the last one. A typical use is with the X-Forwarded-For header.
16209
Frédéric Lécailleec891192019-02-26 15:02:35 +010016210
16211
Willy Tarreau74ca5042013-06-11 23:12:07 +020016212http_auth(<userlist>) : boolean
16213 Returns a boolean indicating whether the authentication data received from
16214 the client match a username & password stored in the specified userlist. This
16215 fetch function is not really useful outside of ACLs. Currently only http
16216 basic auth is supported.
16217
Thierry FOURNIER9eec0a62014-01-22 18:38:02 +010016218http_auth_group(<userlist>) : string
16219 Returns a string corresponding to the user name found in the authentication
16220 data received from the client if both the user name and password are valid
16221 according to the specified userlist. The main purpose is to use it in ACLs
16222 where it is then checked whether the user belongs to any group within a list.
Willy Tarreau74ca5042013-06-11 23:12:07 +020016223 This fetch function is not really useful outside of ACLs. Currently only http
16224 basic auth is supported.
16225
16226 ACL derivatives :
Thierry FOURNIER9eec0a62014-01-22 18:38:02 +010016227 http_auth_group(<userlist>) : group ...
16228 Returns true when the user extracted from the request and whose password is
16229 valid according to the specified userlist belongs to at least one of the
16230 groups.
Willy Tarreau74ca5042013-06-11 23:12:07 +020016231
16232http_first_req : boolean
Willy Tarreau7f18e522010-10-22 20:04:13 +020016233 Returns true when the request being processed is the first one of the
16234 connection. This can be used to add or remove headers that may be missing
Willy Tarreau74ca5042013-06-11 23:12:07 +020016235 from some requests when a request is not the first one, or to help grouping
16236 requests in the logs.
Willy Tarreau7f18e522010-10-22 20:04:13 +020016237
Willy Tarreau74ca5042013-06-11 23:12:07 +020016238method : integer + string
16239 Returns an integer value corresponding to the method in the HTTP request. For
16240 example, "GET" equals 1 (check sources to establish the matching). Value 9
16241 means "other method" and may be converted to a string extracted from the
16242 stream. This should not be used directly as a sample, this is only meant to
16243 be used from ACLs, which transparently convert methods from patterns to these
16244 integer + string values. Some predefined ACL already check for most common
16245 methods.
Willy Tarreau6a06a402007-07-15 20:15:28 +020016246
Willy Tarreau74ca5042013-06-11 23:12:07 +020016247 ACL derivatives :
16248 method : case insensitive method match
Willy Tarreau6a06a402007-07-15 20:15:28 +020016249
Willy Tarreau74ca5042013-06-11 23:12:07 +020016250 Example :
16251 # only accept GET and HEAD requests
16252 acl valid_method method GET HEAD
16253 http-request deny if ! valid_method
Willy Tarreau6a06a402007-07-15 20:15:28 +020016254
Willy Tarreau74ca5042013-06-11 23:12:07 +020016255path : string
16256 This extracts the request's URL path, which starts at the first slash and
16257 ends before the question mark (without the host part). A typical use is with
16258 prefetch-capable caches, and with portals which need to aggregate multiple
16259 information from databases and keep them in caches. Note that with outgoing
16260 caches, it would be wiser to use "url" instead. With ACLs, it's typically
Davor Ocelice9ed2812017-12-25 17:49:28 +010016261 used to match exact file names (e.g. "/login.php"), or directory parts using
Willy Tarreau74ca5042013-06-11 23:12:07 +020016262 the derivative forms. See also the "url" and "base" fetch methods.
Willy Tarreau6a06a402007-07-15 20:15:28 +020016263
Willy Tarreau74ca5042013-06-11 23:12:07 +020016264 ACL derivatives :
16265 path : exact string match
16266 path_beg : prefix match
16267 path_dir : subdir match
16268 path_dom : domain match
16269 path_end : suffix match
16270 path_len : length match
16271 path_reg : regex match
16272 path_sub : substring match
Willy Tarreau6a06a402007-07-15 20:15:28 +020016273
Willy Tarreau49ad95c2015-01-19 15:06:26 +010016274query : string
16275 This extracts the request's query string, which starts after the first
16276 question mark. If no question mark is present, this fetch returns nothing. If
16277 a question mark is present but nothing follows, it returns an empty string.
16278 This means it's possible to easily know whether a query string is present
Tim Düsterhus4896c442016-11-29 02:15:19 +010016279 using the "found" matching method. This fetch is the complement of "path"
Willy Tarreau49ad95c2015-01-19 15:06:26 +010016280 which stops before the question mark.
16281
Willy Tarreaueb27ec72015-02-20 13:55:29 +010016282req.hdr_names([<delim>]) : string
16283 This builds a string made from the concatenation of all header names as they
16284 appear in the request when the rule is evaluated. The default delimiter is
16285 the comma (',') but it may be overridden as an optional argument <delim>. In
16286 this case, only the first character of <delim> is considered.
16287
Willy Tarreau74ca5042013-06-11 23:12:07 +020016288req.ver : string
16289req_ver : string (deprecated)
16290 Returns the version string from the HTTP request, for example "1.1". This can
16291 be useful for logs, but is mostly there for ACL. Some predefined ACL already
16292 check for versions 1.0 and 1.1.
Willy Tarreaud63335a2010-02-26 12:56:52 +010016293
Willy Tarreau74ca5042013-06-11 23:12:07 +020016294 ACL derivatives :
16295 req_ver : exact string match
Willy Tarreau0e698542011-09-16 08:32:32 +020016296
Willy Tarreau74ca5042013-06-11 23:12:07 +020016297res.comp : boolean
16298 Returns the boolean "true" value if the response has been compressed by
16299 HAProxy, otherwise returns boolean "false". This may be used to add
16300 information in the logs.
Willy Tarreau6a06a402007-07-15 20:15:28 +020016301
Willy Tarreau74ca5042013-06-11 23:12:07 +020016302res.comp_algo : string
16303 Returns a string containing the name of the algorithm used if the response
16304 was compressed by HAProxy, for example : "deflate". This may be used to add
16305 some information in the logs.
Willy Tarreaud63335a2010-02-26 12:56:52 +010016306
Willy Tarreau74ca5042013-06-11 23:12:07 +020016307res.cook([<name>]) : string
16308scook([<name>]) : string (deprecated)
16309 This extracts the last occurrence of the cookie name <name> on a "Set-Cookie"
16310 header line from the response, and returns its value as string. If no name is
16311 specified, the first cookie value is returned.
Willy Tarreau0ce3aa02012-04-25 18:46:33 +020016312
Willy Tarreau74ca5042013-06-11 23:12:07 +020016313 ACL derivatives :
16314 scook([<name>] : exact string match
Willy Tarreau0ce3aa02012-04-25 18:46:33 +020016315
Willy Tarreau74ca5042013-06-11 23:12:07 +020016316res.cook_cnt([<name>]) : integer
16317scook_cnt([<name>]) : integer (deprecated)
16318 Returns an integer value representing the number of occurrences of the cookie
16319 <name> in the response, or all cookies if <name> is not specified. This is
16320 mostly useful when combined with ACLs to detect suspicious responses.
Willy Tarreaud63335a2010-02-26 12:56:52 +010016321
Willy Tarreau74ca5042013-06-11 23:12:07 +020016322res.cook_val([<name>]) : integer
16323scook_val([<name>]) : integer (deprecated)
16324 This extracts the last occurrence of the cookie name <name> on a "Set-Cookie"
16325 header line from the response, and converts its value to an integer which is
16326 returned. If no name is specified, the first cookie value is returned.
Willy Tarreaud63335a2010-02-26 12:56:52 +010016327
Willy Tarreau74ca5042013-06-11 23:12:07 +020016328res.fhdr([<name>[,<occ>]]) : string
16329 This extracts the last occurrence of header <name> in an HTTP response, or of
16330 the last header if no <name> is specified. Optionally, a specific occurrence
16331 might be specified as a position number. Positive values indicate a position
16332 from the first occurrence, with 1 being the first one. Negative values
16333 indicate positions relative to the last one, with -1 being the last one. It
16334 differs from res.hdr() in that any commas present in the value are returned
16335 and are not used as delimiters. If this is not desired, the res.hdr() fetch
16336 should be used instead. This is sometimes useful with headers such as Date or
16337 Expires.
Willy Tarreau6a06a402007-07-15 20:15:28 +020016338
Willy Tarreau74ca5042013-06-11 23:12:07 +020016339res.fhdr_cnt([<name>]) : integer
16340 Returns an integer value representing the number of occurrences of response
16341 header field name <name>, or the total number of header fields if <name> is
16342 not specified. Contrary to its res.hdr_cnt() cousin, this function returns
16343 the number of full line headers and does not stop on commas. If this is not
16344 desired, the res.hdr_cnt() fetch should be used instead.
Willy Tarreau6a06a402007-07-15 20:15:28 +020016345
Willy Tarreau74ca5042013-06-11 23:12:07 +020016346res.hdr([<name>[,<occ>]]) : string
16347shdr([<name>[,<occ>]]) : string (deprecated)
16348 This extracts the last occurrence of header <name> in an HTTP response, or of
16349 the last header if no <name> is specified. Optionally, a specific occurrence
16350 might be specified as a position number. Positive values indicate a position
16351 from the first occurrence, with 1 being the first one. Negative values
16352 indicate positions relative to the last one, with -1 being the last one. This
16353 can be useful to learn some data into a stick-table. The function considers
16354 any comma as a delimiter for distinct values. If this is not desired, the
16355 res.fhdr() fetch should be used instead.
Willy Tarreau6a06a402007-07-15 20:15:28 +020016356
Willy Tarreau74ca5042013-06-11 23:12:07 +020016357 ACL derivatives :
16358 shdr([<name>[,<occ>]]) : exact string match
16359 shdr_beg([<name>[,<occ>]]) : prefix match
16360 shdr_dir([<name>[,<occ>]]) : subdir match
16361 shdr_dom([<name>[,<occ>]]) : domain match
16362 shdr_end([<name>[,<occ>]]) : suffix match
16363 shdr_len([<name>[,<occ>]]) : length match
16364 shdr_reg([<name>[,<occ>]]) : regex match
16365 shdr_sub([<name>[,<occ>]]) : substring match
16366
16367res.hdr_cnt([<name>]) : integer
16368shdr_cnt([<name>]) : integer (deprecated)
16369 Returns an integer value representing the number of occurrences of response
16370 header field name <name>, or the total number of header fields if <name> is
16371 not specified. The function considers any comma as a delimiter for distinct
16372 values. If this is not desired, the res.fhdr_cnt() fetch should be used
16373 instead.
Willy Tarreau6a06a402007-07-15 20:15:28 +020016374
Willy Tarreau74ca5042013-06-11 23:12:07 +020016375res.hdr_ip([<name>[,<occ>]]) : ip
16376shdr_ip([<name>[,<occ>]]) : ip (deprecated)
16377 This extracts the last occurrence of header <name> in an HTTP response,
16378 convert it to an IPv4 or IPv6 address and returns this address. Optionally, a
16379 specific occurrence might be specified as a position number. Positive values
16380 indicate a position from the first occurrence, with 1 being the first one.
16381 Negative values indicate positions relative to the last one, with -1 being
16382 the last one. This can be useful to learn some data into a stick table.
Willy Tarreau6a06a402007-07-15 20:15:28 +020016383
Willy Tarreaueb27ec72015-02-20 13:55:29 +010016384res.hdr_names([<delim>]) : string
16385 This builds a string made from the concatenation of all header names as they
16386 appear in the response when the rule is evaluated. The default delimiter is
16387 the comma (',') but it may be overridden as an optional argument <delim>. In
16388 this case, only the first character of <delim> is considered.
16389
Willy Tarreau74ca5042013-06-11 23:12:07 +020016390res.hdr_val([<name>[,<occ>]]) : integer
16391shdr_val([<name>[,<occ>]]) : integer (deprecated)
16392 This extracts the last occurrence of header <name> in an HTTP response, and
16393 converts it to an integer value. Optionally, a specific occurrence might be
16394 specified as a position number. Positive values indicate a position from the
16395 first occurrence, with 1 being the first one. Negative values indicate
16396 positions relative to the last one, with -1 being the last one. This can be
16397 useful to learn some data into a stick table.
Alexandre Cassen5eb1a902007-11-29 15:43:32 +010016398
Willy Tarreau74ca5042013-06-11 23:12:07 +020016399res.ver : string
16400resp_ver : string (deprecated)
16401 Returns the version string from the HTTP response, for example "1.1". This
16402 can be useful for logs, but is mostly there for ACL.
Willy Tarreau0e698542011-09-16 08:32:32 +020016403
Willy Tarreau74ca5042013-06-11 23:12:07 +020016404 ACL derivatives :
16405 resp_ver : exact string match
Alexandre Cassen5eb1a902007-11-29 15:43:32 +010016406
Willy Tarreau74ca5042013-06-11 23:12:07 +020016407set-cookie([<name>]) : string (deprecated)
16408 This extracts the last occurrence of the cookie name <name> on a "Set-Cookie"
16409 header line from the response and uses the corresponding value to match. This
Willy Tarreau294d0f02015-08-10 19:40:12 +020016410 can be comparable to what "appsession" did with default options, but with
Willy Tarreau74ca5042013-06-11 23:12:07 +020016411 support for multi-peer synchronization and state keeping across restarts.
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +010016412
Willy Tarreau74ca5042013-06-11 23:12:07 +020016413 This fetch function is deprecated and has been superseded by the "res.cook"
16414 fetch. This keyword will disappear soon.
Krzysztof Piotr Oledzki6b35ce12010-02-01 23:35:44 +010016415
Willy Tarreau74ca5042013-06-11 23:12:07 +020016416status : integer
16417 Returns an integer containing the HTTP status code in the HTTP response, for
16418 example, 302. It is mostly used within ACLs and integer ranges, for example,
16419 to remove any Location header if the response is not a 3xx.
Willy Tarreau25c1ebc2012-04-25 16:21:44 +020016420
Thierry Fournier0e00dca2016-04-07 15:47:40 +020016421unique-id : string
16422 Returns the unique-id attached to the request. The directive
16423 "unique-id-format" must be set. If it is not set, the unique-id sample fetch
16424 fails. Note that the unique-id is usually used with HTTP requests, however this
16425 sample fetch can be used with other protocols. Obviously, if it is used with
16426 other protocols than HTTP, the unique-id-format directive must not contain
16427 HTTP parts. See: unique-id-format and unique-id-header
16428
Willy Tarreau74ca5042013-06-11 23:12:07 +020016429url : string
16430 This extracts the request's URL as presented in the request. A typical use is
16431 with prefetch-capable caches, and with portals which need to aggregate
16432 multiple information from databases and keep them in caches. With ACLs, using
16433 "path" is preferred over using "url", because clients may send a full URL as
16434 is normally done with proxies. The only real use is to match "*" which does
16435 not match in "path", and for which there is already a predefined ACL. See
16436 also "path" and "base".
Willy Tarreau25c1ebc2012-04-25 16:21:44 +020016437
Willy Tarreau74ca5042013-06-11 23:12:07 +020016438 ACL derivatives :
16439 url : exact string match
16440 url_beg : prefix match
16441 url_dir : subdir match
16442 url_dom : domain match
16443 url_end : suffix match
16444 url_len : length match
16445 url_reg : regex match
16446 url_sub : substring match
Willy Tarreau25c1ebc2012-04-25 16:21:44 +020016447
Willy Tarreau74ca5042013-06-11 23:12:07 +020016448url_ip : ip
16449 This extracts the IP address from the request's URL when the host part is
16450 presented as an IP address. Its use is very limited. For instance, a
16451 monitoring system might use this field as an alternative for the source IP in
16452 order to test what path a given source address would follow, or to force an
16453 entry in a table for a given source address. With ACLs it can be used to
16454 restrict access to certain systems through a proxy, for example when combined
16455 with option "http_proxy".
Willy Tarreau25c1ebc2012-04-25 16:21:44 +020016456
Willy Tarreau74ca5042013-06-11 23:12:07 +020016457url_port : integer
16458 This extracts the port part from the request's URL. Note that if the port is
16459 not specified in the request, port 80 is assumed. With ACLs it can be used to
16460 restrict access to certain systems through a proxy, for example when combined
16461 with option "http_proxy".
Willy Tarreau25c1ebc2012-04-25 16:21:44 +020016462
Willy Tarreau1ede1da2015-05-07 16:06:18 +020016463urlp([<name>[,<delim>]]) : string
16464url_param([<name>[,<delim>]]) : string
Willy Tarreau74ca5042013-06-11 23:12:07 +020016465 This extracts the first occurrence of the parameter <name> in the query
16466 string, which begins after either '?' or <delim>, and which ends before '&',
Willy Tarreau1ede1da2015-05-07 16:06:18 +020016467 ';' or <delim>. The parameter name is case-sensitive. If no name is given,
16468 any parameter will match, and the first one will be returned. The result is
16469 a string corresponding to the value of the parameter <name> as presented in
16470 the request (no URL decoding is performed). This can be used for session
Willy Tarreau74ca5042013-06-11 23:12:07 +020016471 stickiness based on a client ID, to extract an application cookie passed as a
16472 URL parameter, or in ACLs to apply some checks. Note that the ACL version of
Willy Tarreau1ede1da2015-05-07 16:06:18 +020016473 this fetch iterates over multiple parameters and will iteratively report all
16474 parameters values if no name is given
Willy Tarreau25c1ebc2012-04-25 16:21:44 +020016475
Willy Tarreau74ca5042013-06-11 23:12:07 +020016476 ACL derivatives :
16477 urlp(<name>[,<delim>]) : exact string match
16478 urlp_beg(<name>[,<delim>]) : prefix match
16479 urlp_dir(<name>[,<delim>]) : subdir match
16480 urlp_dom(<name>[,<delim>]) : domain match
16481 urlp_end(<name>[,<delim>]) : suffix match
16482 urlp_len(<name>[,<delim>]) : length match
16483 urlp_reg(<name>[,<delim>]) : regex match
16484 urlp_sub(<name>[,<delim>]) : substring match
Willy Tarreau25c1ebc2012-04-25 16:21:44 +020016485
Willy Tarreau25c1ebc2012-04-25 16:21:44 +020016486
Willy Tarreau74ca5042013-06-11 23:12:07 +020016487 Example :
16488 # match http://example.com/foo?PHPSESSIONID=some_id
16489 stick on urlp(PHPSESSIONID)
16490 # match http://example.com/foo;JSESSIONID=some_id
16491 stick on urlp(JSESSIONID,;)
Willy Tarreau25c1ebc2012-04-25 16:21:44 +020016492
Jarno Huuskonen676f6222017-03-30 09:19:45 +030016493urlp_val([<name>[,<delim>]]) : integer
Willy Tarreau74ca5042013-06-11 23:12:07 +020016494 See "urlp" above. This one extracts the URL parameter <name> in the request
16495 and converts it to an integer value. This can be used for session stickiness
16496 based on a user ID for example, or with ACLs to match a page number or price.
Willy Tarreaua9fddca2012-07-31 07:51:48 +020016497
Dragan Dosen0070cd52016-06-16 12:19:49 +020016498url32 : integer
16499 This returns a 32-bit hash of the value obtained by concatenating the first
16500 Host header and the whole URL including parameters (not only the path part of
16501 the request, as in the "base32" fetch above). This is useful to track per-URL
16502 activity. A shorter hash is stored, saving a lot of memory. The output type
16503 is an unsigned integer.
16504
16505url32+src : binary
16506 This returns the concatenation of the "url32" fetch and the "src" fetch. The
16507 resulting type is of type binary, with a size of 8 or 20 bytes depending on
16508 the source address family. This can be used to track per-IP, per-URL counters.
16509
Willy Tarreau198a7442008-01-17 12:05:32 +010016510
Willy Tarreau74ca5042013-06-11 23:12:07 +0200165117.4. Pre-defined ACLs
Willy Tarreauc57f0e22009-05-10 13:12:33 +020016512---------------------
Willy Tarreauced27012008-01-17 20:35:34 +010016513
Willy Tarreauc57f0e22009-05-10 13:12:33 +020016514Some predefined ACLs are hard-coded so that they do not have to be declared in
16515every frontend which needs them. They all have their names in upper case in
Patrick Mézard2382ad62010-05-09 10:43:32 +020016516order to avoid confusion. Their equivalence is provided below.
Willy Tarreauced27012008-01-17 20:35:34 +010016517
Willy Tarreauc57f0e22009-05-10 13:12:33 +020016518ACL name Equivalent to Usage
16519---------------+-----------------------------+---------------------------------
Willy Tarreauc57f0e22009-05-10 13:12:33 +020016520FALSE always_false never match
Willy Tarreau2492d5b2009-07-11 00:06:00 +020016521HTTP req_proto_http match if protocol is valid HTTP
Willy Tarreauc57f0e22009-05-10 13:12:33 +020016522HTTP_1.0 req_ver 1.0 match HTTP version 1.0
16523HTTP_1.1 req_ver 1.1 match HTTP version 1.1
Willy Tarreaud63335a2010-02-26 12:56:52 +010016524HTTP_CONTENT hdr_val(content-length) gt 0 match an existing content-length
16525HTTP_URL_ABS url_reg ^[^/:]*:// match absolute URL with scheme
16526HTTP_URL_SLASH url_beg / match URL beginning with "/"
16527HTTP_URL_STAR url * match URL equal to "*"
16528LOCALHOST src 127.0.0.1/8 match connection from local host
Willy Tarreauc57f0e22009-05-10 13:12:33 +020016529METH_CONNECT method CONNECT match HTTP CONNECT method
Daniel Schneller9ff96c72016-04-11 17:45:29 +020016530METH_DELETE method DELETE match HTTP DELETE method
Willy Tarreauc57f0e22009-05-10 13:12:33 +020016531METH_GET method GET HEAD match HTTP GET or HEAD method
16532METH_HEAD method HEAD match HTTP HEAD method
16533METH_OPTIONS method OPTIONS match HTTP OPTIONS method
16534METH_POST method POST match HTTP POST method
Daniel Schneller9ff96c72016-04-11 17:45:29 +020016535METH_PUT method PUT match HTTP PUT method
Willy Tarreauc57f0e22009-05-10 13:12:33 +020016536METH_TRACE method TRACE match HTTP TRACE method
Emeric Brunbede3d02009-06-30 17:54:00 +020016537RDP_COOKIE req_rdp_cookie_cnt gt 0 match presence of an RDP cookie
Willy Tarreauc57f0e22009-05-10 13:12:33 +020016538REQ_CONTENT req_len gt 0 match data in the request buffer
Willy Tarreaud63335a2010-02-26 12:56:52 +010016539TRUE always_true always match
Willy Tarreauc57f0e22009-05-10 13:12:33 +020016540WAIT_END wait_end wait for end of content analysis
16541---------------+-----------------------------+---------------------------------
Willy Tarreauced27012008-01-17 20:35:34 +010016542
Willy Tarreaub937b7e2010-01-12 15:27:54 +010016543
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200165448. Logging
16545----------
Willy Tarreau844e3c52008-01-11 16:28:18 +010016546
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016547One of HAProxy's strong points certainly lies is its precise logs. It probably
16548provides the finest level of information available for such a product, which is
16549very important for troubleshooting complex environments. Standard information
16550provided in logs include client ports, TCP/HTTP state timers, precise session
16551state at termination and precise termination cause, information about decisions
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +010016552to direct traffic to a server, and of course the ability to capture arbitrary
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016553headers.
16554
16555In order to improve administrators reactivity, it offers a great transparency
16556about encountered problems, both internal and external, and it is possible to
16557send logs to different sources at the same time with different level filters :
16558
16559 - global process-level logs (system errors, start/stop, etc..)
16560 - per-instance system and internal errors (lack of resource, bugs, ...)
16561 - per-instance external troubles (servers up/down, max connections)
16562 - per-instance activity (client connections), either at the establishment or
16563 at the termination.
Davor Ocelice9ed2812017-12-25 17:49:28 +010016564 - per-request control of log-level, e.g.
Jim Freeman9e8714b2015-05-26 09:16:34 -060016565 http-request set-log-level silent if sensitive_request
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016566
16567The ability to distribute different levels of logs to different log servers
16568allow several production teams to interact and to fix their problems as soon
16569as possible. For example, the system team might monitor system-wide errors,
16570while the application team might be monitoring the up/down for their servers in
16571real time, and the security team might analyze the activity logs with one hour
16572delay.
16573
16574
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200165758.1. Log levels
16576---------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016577
Simon Hormandf791f52011-05-29 15:01:10 +090016578TCP and HTTP connections can be logged with information such as the date, time,
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016579source IP address, destination address, connection duration, response times,
Simon Hormandf791f52011-05-29 15:01:10 +090016580HTTP request, HTTP return code, number of bytes transmitted, conditions
16581in which the session ended, and even exchanged cookies values. For example
16582track a particular user's problems. All messages may be sent to up to two
16583syslog servers. Check the "log" keyword in section 4.2 for more information
16584about log facilities.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016585
16586
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200165878.2. Log formats
16588----------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016589
William Lallemand48940402012-01-30 16:47:22 +010016590HAProxy supports 5 log formats. Several fields are common between these formats
Simon Hormandf791f52011-05-29 15:01:10 +090016591and will be detailed in the following sections. A few of them may vary
16592slightly with the configuration, due to indicators specific to certain
16593options. The supported formats are as follows :
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016594
16595 - the default format, which is very basic and very rarely used. It only
16596 provides very basic information about the incoming connection at the moment
16597 it is accepted : source IP:port, destination IP:port, and frontend-name.
16598 This mode will eventually disappear so it will not be described to great
16599 extents.
16600
16601 - the TCP format, which is more advanced. This format is enabled when "option
16602 tcplog" is set on the frontend. HAProxy will then usually wait for the
16603 connection to terminate before logging. This format provides much richer
16604 information, such as timers, connection counts, queue size, etc... This
16605 format is recommended for pure TCP proxies.
16606
16607 - the HTTP format, which is the most advanced for HTTP proxying. This format
16608 is enabled when "option httplog" is set on the frontend. It provides the
16609 same information as the TCP format with some HTTP-specific fields such as
16610 the request, the status code, and captures of headers and cookies. This
16611 format is recommended for HTTP proxies.
16612
Emeric Brun3a058f32009-06-30 18:26:00 +020016613 - the CLF HTTP format, which is equivalent to the HTTP format, but with the
16614 fields arranged in the same order as the CLF format. In this mode, all
16615 timers, captures, flags, etc... appear one per field after the end of the
16616 common fields, in the same order they appear in the standard HTTP format.
16617
William Lallemand48940402012-01-30 16:47:22 +010016618 - the custom log format, allows you to make your own log line.
16619
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016620Next sections will go deeper into details for each of these formats. Format
16621specification will be performed on a "field" basis. Unless stated otherwise, a
16622field is a portion of text delimited by any number of spaces. Since syslog
16623servers are susceptible of inserting fields at the beginning of a line, it is
16624always assumed that the first field is the one containing the process name and
16625identifier.
16626
16627Note : Since log lines may be quite long, the log examples in sections below
16628 might be broken into multiple lines. The example log lines will be
16629 prefixed with 3 closing angle brackets ('>>>') and each time a log is
16630 broken into multiple lines, each non-final line will end with a
16631 backslash ('\') and the next line will start indented by two characters.
16632
16633
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200166348.2.1. Default log format
16635-------------------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016636
16637This format is used when no specific option is set. The log is emitted as soon
16638as the connection is accepted. One should note that this currently is the only
16639format which logs the request's destination IP and ports.
16640
16641 Example :
16642 listen www
16643 mode http
16644 log global
16645 server srv1 127.0.0.1:8000
16646
16647 >>> Feb 6 12:12:09 localhost \
16648 haproxy[14385]: Connect from 10.0.1.2:33312 to 10.0.3.31:8012 \
16649 (www/HTTP)
16650
16651 Field Format Extract from the example above
16652 1 process_name '[' pid ']:' haproxy[14385]:
16653 2 'Connect from' Connect from
16654 3 source_ip ':' source_port 10.0.1.2:33312
16655 4 'to' to
16656 5 destination_ip ':' destination_port 10.0.3.31:8012
16657 6 '(' frontend_name '/' mode ')' (www/HTTP)
16658
16659Detailed fields description :
16660 - "source_ip" is the IP address of the client which initiated the connection.
16661 - "source_port" is the TCP port of the client which initiated the connection.
16662 - "destination_ip" is the IP address the client connected to.
16663 - "destination_port" is the TCP port the client connected to.
16664 - "frontend_name" is the name of the frontend (or listener) which received
16665 and processed the connection.
16666 - "mode is the mode the frontend is operating (TCP or HTTP).
16667
Willy Tarreauceb24bc2010-11-09 12:46:41 +010016668In case of a UNIX socket, the source and destination addresses are marked as
16669"unix:" and the ports reflect the internal ID of the socket which accepted the
16670connection (the same ID as reported in the stats).
16671
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016672It is advised not to use this deprecated format for newer installations as it
16673will eventually disappear.
16674
16675
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200166768.2.2. TCP log format
16677---------------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016678
16679The TCP format is used when "option tcplog" is specified in the frontend, and
16680is the recommended format for pure TCP proxies. It provides a lot of precious
16681information for troubleshooting. Since this format includes timers and byte
16682counts, the log is normally emitted at the end of the session. It can be
16683emitted earlier if "option logasap" is specified, which makes sense in most
16684environments with long sessions such as remote terminals. Sessions which match
16685the "monitor" rules are never logged. It is also possible not to emit logs for
16686sessions for which no data were exchanged between the client and the server, by
Willy Tarreauc9bd0cc2009-05-10 11:57:02 +020016687specifying "option dontlognull" in the frontend. Successful connections will
16688not be logged if "option dontlog-normal" is specified in the frontend. A few
16689fields may slightly vary depending on some configuration options, those are
16690marked with a star ('*') after the field name below.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016691
16692 Example :
16693 frontend fnt
16694 mode tcp
16695 option tcplog
16696 log global
16697 default_backend bck
16698
16699 backend bck
16700 server srv1 127.0.0.1:8000
16701
16702 >>> Feb 6 12:12:56 localhost \
16703 haproxy[14387]: 10.0.1.2:33313 [06/Feb/2009:12:12:51.443] fnt \
16704 bck/srv1 0/0/5007 212 -- 0/0/0/0/3 0/0
16705
16706 Field Format Extract from the example above
16707 1 process_name '[' pid ']:' haproxy[14387]:
16708 2 client_ip ':' client_port 10.0.1.2:33313
16709 3 '[' accept_date ']' [06/Feb/2009:12:12:51.443]
16710 4 frontend_name fnt
16711 5 backend_name '/' server_name bck/srv1
16712 6 Tw '/' Tc '/' Tt* 0/0/5007
16713 7 bytes_read* 212
16714 8 termination_state --
16715 9 actconn '/' feconn '/' beconn '/' srv_conn '/' retries* 0/0/0/0/3
16716 10 srv_queue '/' backend_queue 0/0
16717
16718Detailed fields description :
16719 - "client_ip" is the IP address of the client which initiated the TCP
Willy Tarreauceb24bc2010-11-09 12:46:41 +010016720 connection to haproxy. If the connection was accepted on a UNIX socket
16721 instead, the IP address would be replaced with the word "unix". Note that
16722 when the connection is accepted on a socket configured with "accept-proxy"
Bertrand Jacquin93b227d2016-06-04 15:11:10 +010016723 and the PROXY protocol is correctly used, or with a "accept-netscaler-cip"
Davor Ocelice9ed2812017-12-25 17:49:28 +010016724 and the NetScaler Client IP insertion protocol is correctly used, then the
Bertrand Jacquin93b227d2016-06-04 15:11:10 +010016725 logs will reflect the forwarded connection's information.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016726
16727 - "client_port" is the TCP port of the client which initiated the connection.
Willy Tarreauceb24bc2010-11-09 12:46:41 +010016728 If the connection was accepted on a UNIX socket instead, the port would be
16729 replaced with the ID of the accepting socket, which is also reported in the
16730 stats interface.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016731
16732 - "accept_date" is the exact date when the connection was received by haproxy
16733 (which might be very slightly different from the date observed on the
16734 network if there was some queuing in the system's backlog). This is usually
Willy Tarreau590a0512018-09-05 11:56:48 +020016735 the same date which may appear in any upstream firewall's log. When used in
16736 HTTP mode, the accept_date field will be reset to the first moment the
16737 connection is ready to receive a new request (end of previous response for
16738 HTTP/1, immediately after previous request for HTTP/2).
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016739
16740 - "frontend_name" is the name of the frontend (or listener) which received
16741 and processed the connection.
16742
16743 - "backend_name" is the name of the backend (or listener) which was selected
16744 to manage the connection to the server. This will be the same as the
16745 frontend if no switching rule has been applied, which is common for TCP
16746 applications.
16747
16748 - "server_name" is the name of the last server to which the connection was
16749 sent, which might differ from the first one if there were connection errors
16750 and a redispatch occurred. Note that this server belongs to the backend
16751 which processed the request. If the connection was aborted before reaching
16752 a server, "<NOSRV>" is indicated instead of a server name.
16753
16754 - "Tw" is the total time in milliseconds spent waiting in the various queues.
16755 It can be "-1" if the connection was aborted before reaching the queue.
16756 See "Timers" below for more details.
16757
16758 - "Tc" is the total time in milliseconds spent waiting for the connection to
16759 establish to the final server, including retries. It can be "-1" if the
16760 connection was aborted before a connection could be established. See
16761 "Timers" below for more details.
16762
16763 - "Tt" is the total time in milliseconds elapsed between the accept and the
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030016764 last close. It covers all possible processing. There is one exception, if
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016765 "option logasap" was specified, then the time counting stops at the moment
16766 the log is emitted. In this case, a '+' sign is prepended before the value,
16767 indicating that the final one will be larger. See "Timers" below for more
16768 details.
16769
16770 - "bytes_read" is the total number of bytes transmitted from the server to
16771 the client when the log is emitted. If "option logasap" is specified, the
16772 this value will be prefixed with a '+' sign indicating that the final one
16773 may be larger. Please note that this value is a 64-bit counter, so log
16774 analysis tools must be able to handle it without overflowing.
16775
16776 - "termination_state" is the condition the session was in when the session
16777 ended. This indicates the session state, which side caused the end of
16778 session to happen, and for what reason (timeout, error, ...). The normal
16779 flags should be "--", indicating the session was closed by either end with
16780 no data remaining in buffers. See below "Session state at disconnection"
16781 for more details.
16782
16783 - "actconn" is the total number of concurrent connections on the process when
Jamie Gloudonaaa21002012-08-25 00:18:33 -040016784 the session was logged. It is useful to detect when some per-process system
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016785 limits have been reached. For instance, if actconn is close to 512 when
16786 multiple connection errors occur, chances are high that the system limits
16787 the process to use a maximum of 1024 file descriptors and that all of them
Willy Tarreauc57f0e22009-05-10 13:12:33 +020016788 are used. See section 3 "Global parameters" to find how to tune the system.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016789
16790 - "feconn" is the total number of concurrent connections on the frontend when
16791 the session was logged. It is useful to estimate the amount of resource
16792 required to sustain high loads, and to detect when the frontend's "maxconn"
16793 has been reached. Most often when this value increases by huge jumps, it is
16794 because there is congestion on the backend servers, but sometimes it can be
16795 caused by a denial of service attack.
16796
16797 - "beconn" is the total number of concurrent connections handled by the
16798 backend when the session was logged. It includes the total number of
16799 concurrent connections active on servers as well as the number of
16800 connections pending in queues. It is useful to estimate the amount of
16801 additional servers needed to support high loads for a given application.
16802 Most often when this value increases by huge jumps, it is because there is
16803 congestion on the backend servers, but sometimes it can be caused by a
16804 denial of service attack.
16805
16806 - "srv_conn" is the total number of concurrent connections still active on
16807 the server when the session was logged. It can never exceed the server's
16808 configured "maxconn" parameter. If this value is very often close or equal
16809 to the server's "maxconn", it means that traffic regulation is involved a
16810 lot, meaning that either the server's maxconn value is too low, or that
16811 there aren't enough servers to process the load with an optimal response
16812 time. When only one of the server's "srv_conn" is high, it usually means
16813 that this server has some trouble causing the connections to take longer to
16814 be processed than on other servers.
16815
16816 - "retries" is the number of connection retries experienced by this session
16817 when trying to connect to the server. It must normally be zero, unless a
16818 server is being stopped at the same moment the connection was attempted.
16819 Frequent retries generally indicate either a network problem between
16820 haproxy and the server, or a misconfigured system backlog on the server
16821 preventing new connections from being queued. This field may optionally be
16822 prefixed with a '+' sign, indicating that the session has experienced a
16823 redispatch after the maximal retry count has been reached on the initial
16824 server. In this case, the server name appearing in the log is the one the
16825 connection was redispatched to, and not the first one, though both may
16826 sometimes be the same in case of hashing for instance. So as a general rule
16827 of thumb, when a '+' is present in front of the retry count, this count
16828 should not be attributed to the logged server.
16829
16830 - "srv_queue" is the total number of requests which were processed before
16831 this one in the server queue. It is zero when the request has not gone
16832 through the server queue. It makes it possible to estimate the approximate
16833 server's response time by dividing the time spent in queue by the number of
16834 requests in the queue. It is worth noting that if a session experiences a
16835 redispatch and passes through two server queues, their positions will be
Davor Ocelice9ed2812017-12-25 17:49:28 +010016836 cumulative. A request should not pass through both the server queue and the
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016837 backend queue unless a redispatch occurs.
16838
16839 - "backend_queue" is the total number of requests which were processed before
16840 this one in the backend's global queue. It is zero when the request has not
16841 gone through the global queue. It makes it possible to estimate the average
16842 queue length, which easily translates into a number of missing servers when
16843 divided by a server's "maxconn" parameter. It is worth noting that if a
16844 session experiences a redispatch, it may pass twice in the backend's queue,
Davor Ocelice9ed2812017-12-25 17:49:28 +010016845 and then both positions will be cumulative. A request should not pass
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016846 through both the server queue and the backend queue unless a redispatch
16847 occurs.
16848
16849
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200168508.2.3. HTTP log format
16851----------------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016852
16853The HTTP format is the most complete and the best suited for HTTP proxies. It
16854is enabled by when "option httplog" is specified in the frontend. It provides
16855the same level of information as the TCP format with additional features which
16856are specific to the HTTP protocol. Just like the TCP format, the log is usually
16857emitted at the end of the session, unless "option logasap" is specified, which
16858generally only makes sense for download sites. A session which matches the
16859"monitor" rules will never logged. It is also possible not to log sessions for
16860which no data were sent by the client by specifying "option dontlognull" in the
Willy Tarreauc9bd0cc2009-05-10 11:57:02 +020016861frontend. Successful connections will not be logged if "option dontlog-normal"
16862is specified in the frontend.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016863
16864Most fields are shared with the TCP log, some being different. A few fields may
16865slightly vary depending on some configuration options. Those ones are marked
16866with a star ('*') after the field name below.
16867
16868 Example :
16869 frontend http-in
16870 mode http
16871 option httplog
16872 log global
16873 default_backend bck
16874
16875 backend static
16876 server srv1 127.0.0.1:8000
16877
16878 >>> Feb 6 12:14:14 localhost \
16879 haproxy[14389]: 10.0.1.2:33317 [06/Feb/2009:12:14:14.655] http-in \
16880 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 +010016881 {} "GET /index.html HTTP/1.1"
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016882
16883 Field Format Extract from the example above
16884 1 process_name '[' pid ']:' haproxy[14389]:
16885 2 client_ip ':' client_port 10.0.1.2:33317
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020016886 3 '[' request_date ']' [06/Feb/2009:12:14:14.655]
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016887 4 frontend_name http-in
16888 5 backend_name '/' server_name static/srv1
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020016889 6 TR '/' Tw '/' Tc '/' Tr '/' Ta* 10/0/30/69/109
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016890 7 status_code 200
16891 8 bytes_read* 2750
16892 9 captured_request_cookie -
16893 10 captured_response_cookie -
16894 11 termination_state ----
16895 12 actconn '/' feconn '/' beconn '/' srv_conn '/' retries* 1/1/1/1/0
16896 13 srv_queue '/' backend_queue 0/0
16897 14 '{' captured_request_headers* '}' {haproxy.1wt.eu}
16898 15 '{' captured_response_headers* '}' {}
16899 16 '"' http_request '"' "GET /index.html HTTP/1.1"
Willy Tarreaud72758d2010-01-12 10:42:19 +010016900
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016901Detailed fields description :
16902 - "client_ip" is the IP address of the client which initiated the TCP
Willy Tarreauceb24bc2010-11-09 12:46:41 +010016903 connection to haproxy. If the connection was accepted on a UNIX socket
16904 instead, the IP address would be replaced with the word "unix". Note that
16905 when the connection is accepted on a socket configured with "accept-proxy"
Bertrand Jacquin93b227d2016-06-04 15:11:10 +010016906 and the PROXY protocol is correctly used, or with a "accept-netscaler-cip"
Davor Ocelice9ed2812017-12-25 17:49:28 +010016907 and the NetScaler Client IP insertion protocol is correctly used, then the
Bertrand Jacquin93b227d2016-06-04 15:11:10 +010016908 logs will reflect the forwarded connection's information.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016909
16910 - "client_port" is the TCP port of the client which initiated the connection.
Willy Tarreauceb24bc2010-11-09 12:46:41 +010016911 If the connection was accepted on a UNIX socket instead, the port would be
16912 replaced with the ID of the accepting socket, which is also reported in the
16913 stats interface.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016914
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020016915 - "request_date" is the exact date when the first byte of the HTTP request
16916 was received by haproxy (log field %tr).
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016917
16918 - "frontend_name" is the name of the frontend (or listener) which received
16919 and processed the connection.
16920
16921 - "backend_name" is the name of the backend (or listener) which was selected
16922 to manage the connection to the server. This will be the same as the
16923 frontend if no switching rule has been applied.
16924
16925 - "server_name" is the name of the last server to which the connection was
16926 sent, which might differ from the first one if there were connection errors
16927 and a redispatch occurred. Note that this server belongs to the backend
16928 which processed the request. If the request was aborted before reaching a
16929 server, "<NOSRV>" is indicated instead of a server name. If the request was
16930 intercepted by the stats subsystem, "<STATS>" is indicated instead.
16931
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020016932 - "TR" is the total time in milliseconds spent waiting for a full HTTP
16933 request from the client (not counting body) after the first byte was
16934 received. It can be "-1" if the connection was aborted before a complete
John Roeslerfb2fce12019-07-10 15:45:51 -050016935 request could be received or a bad request was received. It should
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020016936 always be very small because a request generally fits in one single packet.
16937 Large times here generally indicate network issues between the client and
Willy Tarreau590a0512018-09-05 11:56:48 +020016938 haproxy or requests being typed by hand. See section 8.4 "Timing Events"
16939 for more details.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016940
16941 - "Tw" is the total time in milliseconds spent waiting in the various queues.
16942 It can be "-1" if the connection was aborted before reaching the queue.
Willy Tarreau590a0512018-09-05 11:56:48 +020016943 See section 8.4 "Timing Events" for more details.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016944
16945 - "Tc" is the total time in milliseconds spent waiting for the connection to
16946 establish to the final server, including retries. It can be "-1" if the
Willy Tarreau590a0512018-09-05 11:56:48 +020016947 request was aborted before a connection could be established. See section
16948 8.4 "Timing Events" for more details.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016949
16950 - "Tr" is the total time in milliseconds spent waiting for the server to send
16951 a full HTTP response, not counting data. It can be "-1" if the request was
16952 aborted before a complete response could be received. It generally matches
16953 the server's processing time for the request, though it may be altered by
16954 the amount of data sent by the client to the server. Large times here on
Willy Tarreau590a0512018-09-05 11:56:48 +020016955 "GET" requests generally indicate an overloaded server. See section 8.4
16956 "Timing Events" for more details.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016957
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020016958 - "Ta" is the time the request remained active in haproxy, which is the total
16959 time in milliseconds elapsed between the first byte of the request was
16960 received and the last byte of response was sent. It covers all possible
16961 processing except the handshake (see Th) and idle time (see Ti). There is
16962 one exception, if "option logasap" was specified, then the time counting
16963 stops at the moment the log is emitted. In this case, a '+' sign is
16964 prepended before the value, indicating that the final one will be larger.
Willy Tarreau590a0512018-09-05 11:56:48 +020016965 See section 8.4 "Timing Events" for more details.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016966
16967 - "status_code" is the HTTP status code returned to the client. This status
16968 is generally set by the server, but it might also be set by haproxy when
16969 the server cannot be reached or when its response is blocked by haproxy.
16970
16971 - "bytes_read" is the total number of bytes transmitted to the client when
16972 the log is emitted. This does include HTTP headers. If "option logasap" is
John Roeslerfb2fce12019-07-10 15:45:51 -050016973 specified, this value will be prefixed with a '+' sign indicating that
Willy Tarreaucc6c8912009-02-22 10:53:55 +010016974 the final one may be larger. Please note that this value is a 64-bit
16975 counter, so log analysis tools must be able to handle it without
16976 overflowing.
16977
16978 - "captured_request_cookie" is an optional "name=value" entry indicating that
16979 the client had this cookie in the request. The cookie name and its maximum
16980 length are defined by the "capture cookie" statement in the frontend
16981 configuration. The field is a single dash ('-') when the option is not
16982 set. Only one cookie may be captured, it is generally used to track session
16983 ID exchanges between a client and a server to detect session crossing
16984 between clients due to application bugs. For more details, please consult
16985 the section "Capturing HTTP headers and cookies" below.
16986
16987 - "captured_response_cookie" is an optional "name=value" entry indicating
16988 that the server has returned a cookie with its response. The cookie name
16989 and its maximum length are defined by the "capture cookie" statement in the
16990 frontend configuration. The field is a single dash ('-') when the option is
16991 not set. Only one cookie may be captured, it is generally used to track
16992 session ID exchanges between a client and a server to detect session
16993 crossing between clients due to application bugs. For more details, please
16994 consult the section "Capturing HTTP headers and cookies" below.
16995
16996 - "termination_state" is the condition the session was in when the session
16997 ended. This indicates the session state, which side caused the end of
16998 session to happen, for what reason (timeout, error, ...), just like in TCP
16999 logs, and information about persistence operations on cookies in the last
17000 two characters. The normal flags should begin with "--", indicating the
17001 session was closed by either end with no data remaining in buffers. See
17002 below "Session state at disconnection" for more details.
17003
17004 - "actconn" is the total number of concurrent connections on the process when
Jamie Gloudonaaa21002012-08-25 00:18:33 -040017005 the session was logged. It is useful to detect when some per-process system
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017006 limits have been reached. For instance, if actconn is close to 512 or 1024
17007 when multiple connection errors occur, chances are high that the system
17008 limits the process to use a maximum of 1024 file descriptors and that all
Willy Tarreauc57f0e22009-05-10 13:12:33 +020017009 of them are used. See section 3 "Global parameters" to find how to tune the
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017010 system.
17011
17012 - "feconn" is the total number of concurrent connections on the frontend when
17013 the session was logged. It is useful to estimate the amount of resource
17014 required to sustain high loads, and to detect when the frontend's "maxconn"
17015 has been reached. Most often when this value increases by huge jumps, it is
17016 because there is congestion on the backend servers, but sometimes it can be
17017 caused by a denial of service attack.
17018
17019 - "beconn" is the total number of concurrent connections handled by the
17020 backend when the session was logged. It includes the total number of
17021 concurrent connections active on servers as well as the number of
17022 connections pending in queues. It is useful to estimate the amount of
17023 additional servers needed to support high loads for a given application.
17024 Most often when this value increases by huge jumps, it is because there is
17025 congestion on the backend servers, but sometimes it can be caused by a
17026 denial of service attack.
17027
17028 - "srv_conn" is the total number of concurrent connections still active on
17029 the server when the session was logged. It can never exceed the server's
17030 configured "maxconn" parameter. If this value is very often close or equal
17031 to the server's "maxconn", it means that traffic regulation is involved a
17032 lot, meaning that either the server's maxconn value is too low, or that
17033 there aren't enough servers to process the load with an optimal response
17034 time. When only one of the server's "srv_conn" is high, it usually means
17035 that this server has some trouble causing the requests to take longer to be
17036 processed than on other servers.
17037
17038 - "retries" is the number of connection retries experienced by this session
17039 when trying to connect to the server. It must normally be zero, unless a
17040 server is being stopped at the same moment the connection was attempted.
17041 Frequent retries generally indicate either a network problem between
17042 haproxy and the server, or a misconfigured system backlog on the server
17043 preventing new connections from being queued. This field may optionally be
17044 prefixed with a '+' sign, indicating that the session has experienced a
17045 redispatch after the maximal retry count has been reached on the initial
17046 server. In this case, the server name appearing in the log is the one the
17047 connection was redispatched to, and not the first one, though both may
17048 sometimes be the same in case of hashing for instance. So as a general rule
17049 of thumb, when a '+' is present in front of the retry count, this count
17050 should not be attributed to the logged server.
17051
17052 - "srv_queue" is the total number of requests which were processed before
17053 this one in the server queue. It is zero when the request has not gone
17054 through the server queue. It makes it possible to estimate the approximate
17055 server's response time by dividing the time spent in queue by the number of
17056 requests in the queue. It is worth noting that if a session experiences a
17057 redispatch and passes through two server queues, their positions will be
Davor Ocelice9ed2812017-12-25 17:49:28 +010017058 cumulative. A request should not pass through both the server queue and the
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017059 backend queue unless a redispatch occurs.
17060
17061 - "backend_queue" is the total number of requests which were processed before
17062 this one in the backend's global queue. It is zero when the request has not
17063 gone through the global queue. It makes it possible to estimate the average
17064 queue length, which easily translates into a number of missing servers when
17065 divided by a server's "maxconn" parameter. It is worth noting that if a
17066 session experiences a redispatch, it may pass twice in the backend's queue,
Davor Ocelice9ed2812017-12-25 17:49:28 +010017067 and then both positions will be cumulative. A request should not pass
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017068 through both the server queue and the backend queue unless a redispatch
17069 occurs.
17070
17071 - "captured_request_headers" is a list of headers captured in the request due
17072 to the presence of the "capture request header" statement in the frontend.
17073 Multiple headers can be captured, they will be delimited by a vertical bar
17074 ('|'). When no capture is enabled, the braces do not appear, causing a
17075 shift of remaining fields. It is important to note that this field may
17076 contain spaces, and that using it requires a smarter log parser than when
17077 it's not used. Please consult the section "Capturing HTTP headers and
17078 cookies" below for more details.
17079
17080 - "captured_response_headers" is a list of headers captured in the response
17081 due to the presence of the "capture response header" statement in the
17082 frontend. Multiple headers can be captured, they will be delimited by a
17083 vertical bar ('|'). When no capture is enabled, the braces do not appear,
17084 causing a shift of remaining fields. It is important to note that this
17085 field may contain spaces, and that using it requires a smarter log parser
17086 than when it's not used. Please consult the section "Capturing HTTP headers
17087 and cookies" below for more details.
17088
17089 - "http_request" is the complete HTTP request line, including the method,
17090 request and HTTP version string. Non-printable characters are encoded (see
17091 below the section "Non-printable characters"). This is always the last
17092 field, and it is always delimited by quotes and is the only one which can
17093 contain quotes. If new fields are added to the log format, they will be
17094 added before this field. This field might be truncated if the request is
17095 huge and does not fit in the standard syslog buffer (1024 characters). This
17096 is the reason why this field must always remain the last one.
17097
17098
Cyril Bontédc4d9032012-04-08 21:57:39 +0200170998.2.4. Custom log format
17100------------------------
William Lallemand48940402012-01-30 16:47:22 +010017101
Willy Tarreau2beef582012-12-20 17:22:52 +010017102The directive log-format allows you to customize the logs in http mode and tcp
William Lallemandbddd4fd2012-02-27 11:23:10 +010017103mode. It takes a string as argument.
William Lallemand48940402012-01-30 16:47:22 +010017104
Davor Ocelice9ed2812017-12-25 17:49:28 +010017105HAProxy understands some log format variables. % precedes log format variables.
William Lallemand48940402012-01-30 16:47:22 +010017106Variables can take arguments using braces ('{}'), and multiple arguments are
17107separated by commas within the braces. Flags may be added or removed by
17108prefixing them with a '+' or '-' sign.
17109
17110Special variable "%o" may be used to propagate its flags to all other
17111variables on the same format string. This is particularly handy with quoted
Dragan Dosen835b9212016-02-12 13:23:03 +010017112("Q") and escaped ("E") string formats.
William Lallemand48940402012-01-30 16:47:22 +010017113
Willy Tarreauc8368452012-12-21 00:09:23 +010017114If a variable is named between square brackets ('[' .. ']') then it is used
Willy Tarreaube722a22014-06-13 16:31:59 +020017115as a sample expression rule (see section 7.3). This it useful to add some
Willy Tarreauc8368452012-12-21 00:09:23 +010017116less common information such as the client's SSL certificate's DN, or to log
17117the key that would be used to store an entry into a stick table.
17118
William Lallemand48940402012-01-30 16:47:22 +010017119Note: spaces must be escaped. A space character is considered as a separator.
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030017120In order to emit a verbatim '%', it must be preceded by another '%' resulting
Willy Tarreau06d97f92013-12-02 17:45:48 +010017121in '%%'. HAProxy will automatically merge consecutive separators.
William Lallemand48940402012-01-30 16:47:22 +010017122
Dragan Dosen835b9212016-02-12 13:23:03 +010017123Note: when using the RFC5424 syslog message format, the characters '"',
17124'\' and ']' inside PARAM-VALUE should be escaped with '\' as prefix (see
17125https://tools.ietf.org/html/rfc5424#section-6.3.3 for more details). In
17126such cases, the use of the flag "E" should be considered.
17127
William Lallemand48940402012-01-30 16:47:22 +010017128Flags are :
17129 * Q: quote a string
Jamie Gloudonaaa21002012-08-25 00:18:33 -040017130 * X: hexadecimal representation (IPs, Ports, %Ts, %rt, %pid)
Dragan Dosen835b9212016-02-12 13:23:03 +010017131 * E: escape characters '"', '\' and ']' in a string with '\' as prefix
17132 (intended purpose is for the RFC5424 structured-data log formats)
William Lallemand48940402012-01-30 16:47:22 +010017133
17134 Example:
17135
17136 log-format %T\ %t\ Some\ Text
17137 log-format %{+Q}o\ %t\ %s\ %{-Q}r
17138
Dragan Dosen835b9212016-02-12 13:23:03 +010017139 log-format-sd %{+Q,+E}o\ [exampleSDID@1234\ header=%[capture.req.hdr(0)]]
17140
William Lallemand48940402012-01-30 16:47:22 +010017141At the moment, the default HTTP format is defined this way :
17142
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020017143 log-format "%ci:%cp [%tr] %ft %b/%s %TR/%Tw/%Tc/%Tr/%Ta %ST %B %CC \
17144 %CS %tsc %ac/%fc/%bc/%sc/%rc %sq/%bq %hr %hs %{+Q}r"
William Lallemand48940402012-01-30 16:47:22 +010017145
William Lallemandbddd4fd2012-02-27 11:23:10 +010017146the default CLF format is defined this way :
William Lallemand48940402012-01-30 16:47:22 +010017147
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020017148 log-format "%{+Q}o %{-Q}ci - - [%trg] %r %ST %B \"\" \"\" %cp \
17149 %ms %ft %b %s %TR %Tw %Tc %Tr %Ta %tsc %ac %fc \
17150 %bc %sc %rc %sq %bq %CC %CS %hrl %hsl"
William Lallemand48940402012-01-30 16:47:22 +010017151
William Lallemandbddd4fd2012-02-27 11:23:10 +010017152and the default TCP format is defined this way :
17153
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020017154 log-format "%ci:%cp [%t] %ft %b/%s %Tw/%Tc/%Tt %B %ts \
17155 %ac/%fc/%bc/%sc/%rc %sq/%bq"
William Lallemandbddd4fd2012-02-27 11:23:10 +010017156
William Lallemand48940402012-01-30 16:47:22 +010017157Please refer to the table below for currently defined variables :
17158
William Lallemandbddd4fd2012-02-27 11:23:10 +010017159 +---+------+-----------------------------------------------+-------------+
Willy Tarreauffc3fcd2012-10-12 20:17:54 +020017160 | R | var | field name (8.2.2 and 8.2.3 for description) | type |
William Lallemandbddd4fd2012-02-27 11:23:10 +010017161 +---+------+-----------------------------------------------+-------------+
17162 | | %o | special variable, apply flags on all next var | |
17163 +---+------+-----------------------------------------------+-------------+
Willy Tarreau2beef582012-12-20 17:22:52 +010017164 | | %B | bytes_read (from server to client) | numeric |
17165 | H | %CC | captured_request_cookie | string |
17166 | H | %CS | captured_response_cookie | string |
William Lallemand5f232402012-04-05 18:02:55 +020017167 | | %H | hostname | string |
Andrew Hayworth0ebc55f2015-04-27 21:37:03 +000017168 | H | %HM | HTTP method (ex: POST) | string |
17169 | H | %HP | HTTP request URI without query string (path) | string |
Andrew Hayworthe63ac872015-07-31 16:14:16 +000017170 | H | %HQ | HTTP request URI query string (ex: ?bar=baz) | string |
Andrew Hayworth0ebc55f2015-04-27 21:37:03 +000017171 | H | %HU | HTTP request URI (ex: /foo?bar=baz) | string |
17172 | H | %HV | HTTP version (ex: HTTP/1.0) | string |
William Lallemanda73203e2012-03-12 12:48:57 +010017173 | | %ID | unique-id | string |
Willy Tarreau4bf99632014-06-13 12:21:40 +020017174 | | %ST | status_code | numeric |
William Lallemand5f232402012-04-05 18:02:55 +020017175 | | %T | gmt_date_time | date |
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020017176 | | %Ta | Active time of the request (from TR to end) | numeric |
William Lallemandbddd4fd2012-02-27 11:23:10 +010017177 | | %Tc | Tc | numeric |
Willy Tarreau27b639d2016-05-17 17:55:27 +020017178 | | %Td | Td = Tt - (Tq + Tw + Tc + Tr) | numeric |
Yuxans Yao4e25b012012-10-19 10:36:09 +080017179 | | %Tl | local_date_time | date |
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020017180 | | %Th | connection handshake time (SSL, PROXY proto) | numeric |
17181 | H | %Ti | idle time before the HTTP request | numeric |
17182 | H | %Tq | Th + Ti + TR | numeric |
17183 | H | %TR | time to receive the full request from 1st byte| numeric |
17184 | H | %Tr | Tr (response time) | numeric |
William Lallemand5f232402012-04-05 18:02:55 +020017185 | | %Ts | timestamp | numeric |
William Lallemandbddd4fd2012-02-27 11:23:10 +010017186 | | %Tt | Tt | numeric |
17187 | | %Tw | Tw | numeric |
Willy Tarreau2beef582012-12-20 17:22:52 +010017188 | | %U | bytes_uploaded (from client to server) | numeric |
William Lallemandbddd4fd2012-02-27 11:23:10 +010017189 | | %ac | actconn | numeric |
17190 | | %b | backend_name | string |
Willy Tarreau2beef582012-12-20 17:22:52 +010017191 | | %bc | beconn (backend concurrent connections) | numeric |
17192 | | %bi | backend_source_ip (connecting address) | IP |
17193 | | %bp | backend_source_port (connecting address) | numeric |
William Lallemandbddd4fd2012-02-27 11:23:10 +010017194 | | %bq | backend_queue | numeric |
Willy Tarreau2beef582012-12-20 17:22:52 +010017195 | | %ci | client_ip (accepted address) | IP |
17196 | | %cp | client_port (accepted address) | numeric |
William Lallemandbddd4fd2012-02-27 11:23:10 +010017197 | | %f | frontend_name | string |
Willy Tarreau2beef582012-12-20 17:22:52 +010017198 | | %fc | feconn (frontend concurrent connections) | numeric |
17199 | | %fi | frontend_ip (accepting address) | IP |
17200 | | %fp | frontend_port (accepting address) | numeric |
Willy Tarreau773d65f2012-10-12 14:56:11 +020017201 | | %ft | frontend_name_transport ('~' suffix for SSL) | string |
Willy Tarreau7346acb2014-08-28 15:03:15 +020017202 | | %lc | frontend_log_counter | numeric |
Willy Tarreaud9ed3d22014-06-13 12:23:06 +020017203 | | %hr | captured_request_headers default style | string |
17204 | | %hrl | captured_request_headers CLF style | string list |
17205 | | %hs | captured_response_headers default style | string |
17206 | | %hsl | captured_response_headers CLF style | string list |
Willy Tarreau812c88e2015-08-09 10:56:35 +020017207 | | %ms | accept date milliseconds (left-padded with 0) | numeric |
William Lallemand5f232402012-04-05 18:02:55 +020017208 | | %pid | PID | numeric |
Willy Tarreauffc3fcd2012-10-12 20:17:54 +020017209 | H | %r | http_request | string |
William Lallemandbddd4fd2012-02-27 11:23:10 +010017210 | | %rc | retries | numeric |
Willy Tarreau1f0da242014-01-25 11:01:50 +010017211 | | %rt | request_counter (HTTP req or TCP session) | numeric |
William Lallemandbddd4fd2012-02-27 11:23:10 +010017212 | | %s | server_name | string |
Willy Tarreau2beef582012-12-20 17:22:52 +010017213 | | %sc | srv_conn (server concurrent connections) | numeric |
17214 | | %si | server_IP (target address) | IP |
17215 | | %sp | server_port (target address) | numeric |
William Lallemandbddd4fd2012-02-27 11:23:10 +010017216 | | %sq | srv_queue | numeric |
Willy Tarreauffc3fcd2012-10-12 20:17:54 +020017217 | S | %sslc| ssl_ciphers (ex: AES-SHA) | string |
17218 | S | %sslv| ssl_version (ex: TLSv1) | string |
Willy Tarreau2beef582012-12-20 17:22:52 +010017219 | | %t | date_time (with millisecond resolution) | date |
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020017220 | H | %tr | date_time of HTTP request | date |
17221 | H | %trg | gmt_date_time of start of HTTP request | date |
Jens Bissinger15c64ff2018-08-23 14:11:27 +020017222 | H | %trl | local_date_time of start of HTTP request | date |
William Lallemandbddd4fd2012-02-27 11:23:10 +010017223 | | %ts | termination_state | string |
Willy Tarreauffc3fcd2012-10-12 20:17:54 +020017224 | H | %tsc | termination_state with cookie status | string |
William Lallemandbddd4fd2012-02-27 11:23:10 +010017225 +---+------+-----------------------------------------------+-------------+
William Lallemand48940402012-01-30 16:47:22 +010017226
Willy Tarreauffc3fcd2012-10-12 20:17:54 +020017227 R = Restrictions : H = mode http only ; S = SSL only
William Lallemand48940402012-01-30 16:47:22 +010017228
Willy Tarreau5f51e1a2012-12-03 18:40:10 +010017229
172308.2.5. Error log format
17231-----------------------
17232
17233When an incoming connection fails due to an SSL handshake or an invalid PROXY
17234protocol header, haproxy will log the event using a shorter, fixed line format.
17235By default, logs are emitted at the LOG_INFO level, unless the option
17236"log-separate-errors" is set in the backend, in which case the LOG_ERR level
Davor Ocelice9ed2812017-12-25 17:49:28 +010017237will be used. Connections on which no data are exchanged (e.g. probes) are not
Willy Tarreau5f51e1a2012-12-03 18:40:10 +010017238logged if the "dontlognull" option is set.
17239
17240The format looks like this :
17241
17242 >>> Dec 3 18:27:14 localhost \
17243 haproxy[6103]: 127.0.0.1:56059 [03/Dec/2012:17:35:10.380] frt/f1: \
17244 Connection error during SSL handshake
17245
17246 Field Format Extract from the example above
17247 1 process_name '[' pid ']:' haproxy[6103]:
17248 2 client_ip ':' client_port 127.0.0.1:56059
17249 3 '[' accept_date ']' [03/Dec/2012:17:35:10.380]
17250 4 frontend_name "/" bind_name ":" frt/f1:
17251 5 message Connection error during SSL handshake
17252
17253These fields just provide minimal information to help debugging connection
17254failures.
17255
17256
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200172578.3. Advanced logging options
17258-----------------------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017259
17260Some advanced logging options are often looked for but are not easy to find out
17261just by looking at the various options. Here is an entry point for the few
17262options which can enable better logging. Please refer to the keywords reference
17263for more information about their usage.
17264
17265
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200172668.3.1. Disabling logging of external tests
17267------------------------------------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017268
17269It is quite common to have some monitoring tools perform health checks on
17270haproxy. Sometimes it will be a layer 3 load-balancer such as LVS or any
17271commercial load-balancer, and sometimes it will simply be a more complete
17272monitoring system such as Nagios. When the tests are very frequent, users often
17273ask how to disable logging for those checks. There are three possibilities :
17274
17275 - if connections come from everywhere and are just TCP probes, it is often
17276 desired to simply disable logging of connections without data exchange, by
17277 setting "option dontlognull" in the frontend. It also disables logging of
17278 port scans, which may or may not be desired.
17279
17280 - if the connection come from a known source network, use "monitor-net" to
17281 declare this network as monitoring only. Any host in this network will then
17282 only be able to perform health checks, and their requests will not be
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030017283 logged. This is generally appropriate to designate a list of equipment
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017284 such as other load-balancers.
17285
17286 - if the tests are performed on a known URI, use "monitor-uri" to declare
17287 this URI as dedicated to monitoring. Any host sending this request will
17288 only get the result of a health-check, and the request will not be logged.
17289
17290
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200172918.3.2. Logging before waiting for the session to terminate
17292----------------------------------------------------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017293
17294The problem with logging at end of connection is that you have no clue about
17295what is happening during very long sessions, such as remote terminal sessions
17296or large file downloads. This problem can be worked around by specifying
Davor Ocelice9ed2812017-12-25 17:49:28 +010017297"option logasap" in the frontend. HAProxy will then log as soon as possible,
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017298just before data transfer begins. This means that in case of TCP, it will still
17299log the connection status to the server, and in case of HTTP, it will log just
17300after processing the server headers. In this case, the number of bytes reported
17301is the number of header bytes sent to the client. In order to avoid confusion
17302with normal logs, the total time field and the number of bytes are prefixed
17303with a '+' sign which means that real numbers are certainly larger.
17304
17305
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200173068.3.3. Raising log level upon errors
17307------------------------------------
Willy Tarreauc9bd0cc2009-05-10 11:57:02 +020017308
17309Sometimes it is more convenient to separate normal traffic from errors logs,
17310for instance in order to ease error monitoring from log files. When the option
17311"log-separate-errors" is used, connections which experience errors, timeouts,
17312retries, redispatches or HTTP status codes 5xx will see their syslog level
17313raised from "info" to "err". This will help a syslog daemon store the log in
17314a separate file. It is very important to keep the errors in the normal traffic
17315file too, so that log ordering is not altered. You should also be careful if
17316you already have configured your syslog daemon to store all logs higher than
17317"notice" in an "admin" file, because the "err" level is higher than "notice".
17318
17319
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200173208.3.4. Disabling logging of successful connections
17321--------------------------------------------------
Willy Tarreauc9bd0cc2009-05-10 11:57:02 +020017322
17323Although this may sound strange at first, some large sites have to deal with
17324multiple thousands of logs per second and are experiencing difficulties keeping
17325them intact for a long time or detecting errors within them. If the option
17326"dontlog-normal" is set on the frontend, all normal connections will not be
17327logged. In this regard, a normal connection is defined as one without any
17328error, timeout, retry nor redispatch. In HTTP, the status code is checked too,
17329and a response with a status 5xx is not considered normal and will be logged
17330too. Of course, doing is is really discouraged as it will remove most of the
17331useful information from the logs. Do this only if you have no other
17332alternative.
17333
17334
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200173358.4. Timing events
17336------------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017337
17338Timers provide a great help in troubleshooting network problems. All values are
17339reported in milliseconds (ms). These timers should be used in conjunction with
17340the session termination flags. In TCP mode with "option tcplog" set on the
17341frontend, 3 control points are reported under the form "Tw/Tc/Tt", and in HTTP
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020017342mode, 5 control points are reported under the form "TR/Tw/Tc/Tr/Ta". In
17343addition, three other measures are provided, "Th", "Ti", and "Tq".
17344
Guillaume de Lafondf27cddc2016-12-23 17:32:43 +010017345Timings events in HTTP mode:
17346
17347 first request 2nd request
17348 |<-------------------------------->|<-------------- ...
17349 t tr t tr ...
17350 ---|----|----|----|----|----|----|----|----|--
17351 : Th Ti TR Tw Tc Tr Td : Ti ...
17352 :<---- Tq ---->: :
17353 :<-------------- Tt -------------->:
17354 :<--------- Ta --------->:
17355
17356Timings events in TCP mode:
17357
17358 TCP session
17359 |<----------------->|
17360 t t
17361 ---|----|----|----|----|---
17362 | Th Tw Tc Td |
17363 |<------ Tt ------->|
17364
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020017365 - Th: total time to accept tcp connection and execute handshakes for low level
Davor Ocelice9ed2812017-12-25 17:49:28 +010017366 protocols. Currently, these protocols are proxy-protocol and SSL. This may
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020017367 only happen once during the whole connection's lifetime. A large time here
17368 may indicate that the client only pre-established the connection without
17369 speaking, that it is experiencing network issues preventing it from
Davor Ocelice9ed2812017-12-25 17:49:28 +010017370 completing a handshake in a reasonable time (e.g. MTU issues), or that an
Willy Tarreau590a0512018-09-05 11:56:48 +020017371 SSL handshake was very expensive to compute. Please note that this time is
17372 reported only before the first request, so it is safe to average it over
17373 all request to calculate the amortized value. The second and subsequent
17374 request will always report zero here.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017375
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020017376 - Ti: is the idle time before the HTTP request (HTTP mode only). This timer
17377 counts between the end of the handshakes and the first byte of the HTTP
17378 request. When dealing with a second request in keep-alive mode, it starts
Willy Tarreau590a0512018-09-05 11:56:48 +020017379 to count after the end of the transmission the previous response. When a
17380 multiplexed protocol such as HTTP/2 is used, it starts to count immediately
17381 after the previous request. Some browsers pre-establish connections to a
17382 server in order to reduce the latency of a future request, and keep them
17383 pending until they need it. This delay will be reported as the idle time. A
17384 value of -1 indicates that nothing was received on the connection.
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020017385
17386 - TR: total time to get the client request (HTTP mode only). It's the time
17387 elapsed between the first bytes received and the moment the proxy received
17388 the empty line marking the end of the HTTP headers. The value "-1"
17389 indicates that the end of headers has never been seen. This happens when
17390 the client closes prematurely or times out. This time is usually very short
17391 since most requests fit in a single packet. A large time may indicate a
17392 request typed by hand during a test.
17393
17394 - Tq: total time to get the client request from the accept date or since the
17395 emission of the last byte of the previous response (HTTP mode only). It's
Davor Ocelice9ed2812017-12-25 17:49:28 +010017396 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 +020017397 returns -1 as well. This timer used to be very useful before the arrival of
17398 HTTP keep-alive and browsers' pre-connect feature. It's recommended to drop
17399 it in favor of TR nowadays, as the idle time adds a lot of noise to the
17400 reports.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017401
17402 - Tw: total time spent in the queues waiting for a connection slot. It
17403 accounts for backend queue as well as the server queues, and depends on the
17404 queue size, and the time needed for the server to complete previous
17405 requests. The value "-1" means that the request was killed before reaching
17406 the queue, which is generally what happens with invalid or denied requests.
17407
17408 - Tc: total time to establish the TCP connection to the server. It's the time
17409 elapsed between the moment the proxy sent the connection request, and the
17410 moment it was acknowledged by the server, or between the TCP SYN packet and
17411 the matching SYN/ACK packet in return. The value "-1" means that the
17412 connection never established.
17413
17414 - Tr: server response time (HTTP mode only). It's the time elapsed between
17415 the moment the TCP connection was established to the server and the moment
17416 the server sent its complete response headers. It purely shows its request
17417 processing time, without the network overhead due to the data transmission.
17418 It is worth noting that when the client has data to send to the server, for
17419 instance during a POST request, the time already runs, and this can distort
17420 apparent response time. For this reason, it's generally wise not to trust
17421 too much this field for POST requests initiated from clients behind an
17422 untrusted network. A value of "-1" here means that the last the response
17423 header (empty line) was never seen, most likely because the server timeout
17424 stroke before the server managed to process the request.
17425
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020017426 - Ta: total active time for the HTTP request, between the moment the proxy
17427 received the first byte of the request header and the emission of the last
17428 byte of the response body. The exception is when the "logasap" option is
17429 specified. In this case, it only equals (TR+Tw+Tc+Tr), and is prefixed with
17430 a '+' sign. From this field, we can deduce "Td", the data transmission time,
17431 by subtracting other timers when valid :
17432
17433 Td = Ta - (TR + Tw + Tc + Tr)
17434
17435 Timers with "-1" values have to be excluded from this equation. Note that
17436 "Ta" can never be negative.
17437
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017438 - Tt: total session duration time, between the moment the proxy accepted it
17439 and the moment both ends were closed. The exception is when the "logasap"
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020017440 option is specified. In this case, it only equals (Th+Ti+TR+Tw+Tc+Tr), and
17441 is prefixed with a '+' sign. From this field, we can deduce "Td", the data
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030017442 transmission time, by subtracting other timers when valid :
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017443
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020017444 Td = Tt - (Th + Ti + TR + Tw + Tc + Tr)
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017445
17446 Timers with "-1" values have to be excluded from this equation. In TCP
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020017447 mode, "Ti", "Tq" and "Tr" have to be excluded too. Note that "Tt" can never
17448 be negative and that for HTTP, Tt is simply equal to (Th+Ti+Ta).
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017449
17450These timers provide precious indications on trouble causes. Since the TCP
17451protocol defines retransmit delays of 3, 6, 12... seconds, we know for sure
17452that timers close to multiples of 3s are nearly always related to lost packets
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020017453due to network problems (wires, negotiation, congestion). Moreover, if "Ta" or
17454"Tt" is close to a timeout value specified in the configuration, it often means
17455that a session has been aborted on timeout.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017456
17457Most common cases :
17458
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020017459 - If "Th" or "Ti" are close to 3000, a packet has probably been lost between
17460 the client and the proxy. This is very rare on local networks but might
17461 happen when clients are on far remote networks and send large requests. It
17462 may happen that values larger than usual appear here without any network
17463 cause. Sometimes, during an attack or just after a resource starvation has
17464 ended, haproxy may accept thousands of connections in a few milliseconds.
17465 The time spent accepting these connections will inevitably slightly delay
17466 processing of other connections, and it can happen that request times in the
17467 order of a few tens of milliseconds are measured after a few thousands of
17468 new connections have been accepted at once. Using one of the keep-alive
17469 modes may display larger idle times since "Ti" measures the time spent
Patrick Mezard105faca2010-06-12 17:02:46 +020017470 waiting for additional requests.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017471
17472 - If "Tc" is close to 3000, a packet has probably been lost between the
17473 server and the proxy during the server connection phase. This value should
17474 always be very low, such as 1 ms on local networks and less than a few tens
17475 of ms on remote networks.
17476
Willy Tarreau55165fe2009-05-10 12:02:55 +020017477 - If "Tr" is nearly always lower than 3000 except some rare values which seem
17478 to be the average majored by 3000, there are probably some packets lost
17479 between the proxy and the server.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017480
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020017481 - If "Ta" is large even for small byte counts, it generally is because
17482 neither the client nor the server decides to close the connection while
17483 haproxy is running in tunnel mode and both have agreed on a keep-alive
17484 connection mode. In order to solve this issue, it will be needed to specify
17485 one of the HTTP options to manipulate keep-alive or close options on either
17486 the frontend or the backend. Having the smallest possible 'Ta' or 'Tt' is
17487 important when connection regulation is used with the "maxconn" option on
17488 the servers, since no new connection will be sent to the server until
17489 another one is released.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017490
17491Other noticeable HTTP log cases ('xx' means any value to be ignored) :
17492
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020017493 TR/Tw/Tc/Tr/+Ta The "option logasap" is present on the frontend and the log
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017494 was emitted before the data phase. All the timers are valid
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020017495 except "Ta" which is shorter than reality.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017496
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020017497 -1/xx/xx/xx/Ta The client was not able to send a complete request in time
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017498 or it aborted too early. Check the session termination flags
17499 then "timeout http-request" and "timeout client" settings.
17500
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020017501 TR/-1/xx/xx/Ta It was not possible to process the request, maybe because
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017502 servers were out of order, because the request was invalid
17503 or forbidden by ACL rules. Check the session termination
17504 flags.
17505
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020017506 TR/Tw/-1/xx/Ta The connection could not establish on the server. Either it
17507 actively refused it or it timed out after Ta-(TR+Tw) ms.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017508 Check the session termination flags, then check the
17509 "timeout connect" setting. Note that the tarpit action might
17510 return similar-looking patterns, with "Tw" equal to the time
17511 the client connection was maintained open.
17512
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020017513 TR/Tw/Tc/-1/Ta The server has accepted the connection but did not return
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030017514 a complete response in time, or it closed its connection
Thierry FOURNIER / OZON.IO4cac3592016-07-28 17:19:45 +020017515 unexpectedly after Ta-(TR+Tw+Tc) ms. Check the session
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017516 termination flags, then check the "timeout server" setting.
17517
17518
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200175198.5. Session state at disconnection
17520-----------------------------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017521
17522TCP and HTTP logs provide a session termination indicator in the
17523"termination_state" field, just before the number of active connections. It is
175242-characters long in TCP mode, and is extended to 4 characters in HTTP mode,
17525each of which has a special meaning :
17526
17527 - On the first character, a code reporting the first event which caused the
17528 session to terminate :
17529
17530 C : the TCP session was unexpectedly aborted by the client.
17531
17532 S : the TCP session was unexpectedly aborted by the server, or the
17533 server explicitly refused it.
17534
17535 P : the session was prematurely aborted by the proxy, because of a
17536 connection limit enforcement, because a DENY filter was matched,
17537 because of a security check which detected and blocked a dangerous
17538 error in server response which might have caused information leak
Davor Ocelice9ed2812017-12-25 17:49:28 +010017539 (e.g. cacheable cookie).
Willy Tarreau570f2212013-06-10 16:42:09 +020017540
17541 L : the session was locally processed by haproxy and was not passed to
17542 a server. This is what happens for stats and redirects.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017543
17544 R : a resource on the proxy has been exhausted (memory, sockets, source
17545 ports, ...). Usually, this appears during the connection phase, and
17546 system logs should contain a copy of the precise error. If this
17547 happens, it must be considered as a very serious anomaly which
17548 should be fixed as soon as possible by any means.
17549
17550 I : an internal error was identified by the proxy during a self-check.
17551 This should NEVER happen, and you are encouraged to report any log
17552 containing this, because this would almost certainly be a bug. It
17553 would be wise to preventively restart the process after such an
17554 event too, in case it would be caused by memory corruption.
17555
Simon Horman752dc4a2011-06-21 14:34:59 +090017556 D : the session was killed by haproxy because the server was detected
17557 as down and was configured to kill all connections when going down.
17558
Justin Karnegeseb2c24a2012-05-24 15:28:52 -070017559 U : the session was killed by haproxy on this backup server because an
17560 active server was detected as up and was configured to kill all
17561 backup connections when going up.
17562
Willy Tarreaua2a64e92011-09-07 23:01:56 +020017563 K : the session was actively killed by an admin operating on haproxy.
17564
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017565 c : the client-side timeout expired while waiting for the client to
17566 send or receive data.
17567
17568 s : the server-side timeout expired while waiting for the server to
17569 send or receive data.
17570
17571 - : normal session completion, both the client and the server closed
17572 with nothing left in the buffers.
17573
17574 - on the second character, the TCP or HTTP session state when it was closed :
17575
Willy Tarreauf7b30a92010-12-06 22:59:17 +010017576 R : the proxy was waiting for a complete, valid REQUEST from the client
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017577 (HTTP mode only). Nothing was sent to any server.
17578
17579 Q : the proxy was waiting in the QUEUE for a connection slot. This can
17580 only happen when servers have a 'maxconn' parameter set. It can
17581 also happen in the global queue after a redispatch consecutive to
17582 a failed attempt to connect to a dying server. If no redispatch is
17583 reported, then no connection attempt was made to any server.
17584
17585 C : the proxy was waiting for the CONNECTION to establish on the
17586 server. The server might at most have noticed a connection attempt.
17587
17588 H : the proxy was waiting for complete, valid response HEADERS from the
17589 server (HTTP only).
17590
17591 D : the session was in the DATA phase.
17592
17593 L : the proxy was still transmitting LAST data to the client while the
17594 server had already finished. This one is very rare as it can only
17595 happen when the client dies while receiving the last packets.
17596
17597 T : the request was tarpitted. It has been held open with the client
17598 during the whole "timeout tarpit" duration or until the client
17599 closed, both of which will be reported in the "Tw" timer.
17600
17601 - : normal session completion after end of data transfer.
17602
17603 - the third character tells whether the persistence cookie was provided by
17604 the client (only in HTTP mode) :
17605
17606 N : the client provided NO cookie. This is usually the case for new
17607 visitors, so counting the number of occurrences of this flag in the
17608 logs generally indicate a valid trend for the site frequentation.
17609
17610 I : the client provided an INVALID cookie matching no known server.
17611 This might be caused by a recent configuration change, mixed
Cyril Bontéa8e7bbc2010-04-25 22:29:29 +020017612 cookies between HTTP/HTTPS sites, persistence conditionally
17613 ignored, or an attack.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017614
17615 D : the client provided a cookie designating a server which was DOWN,
17616 so either "option persist" was used and the client was sent to
17617 this server, or it was not set and the client was redispatched to
17618 another server.
17619
Willy Tarreau996a92c2010-10-13 19:30:47 +020017620 V : the client provided a VALID cookie, and was sent to the associated
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017621 server.
17622
Willy Tarreau996a92c2010-10-13 19:30:47 +020017623 E : the client provided a valid cookie, but with a last date which was
17624 older than what is allowed by the "maxidle" cookie parameter, so
17625 the cookie is consider EXPIRED and is ignored. The request will be
17626 redispatched just as if there was no cookie.
17627
17628 O : the client provided a valid cookie, but with a first date which was
17629 older than what is allowed by the "maxlife" cookie parameter, so
17630 the cookie is consider too OLD and is ignored. The request will be
17631 redispatched just as if there was no cookie.
17632
Willy Tarreauc89ccb62012-04-05 21:18:22 +020017633 U : a cookie was present but was not used to select the server because
17634 some other server selection mechanism was used instead (typically a
17635 "use-server" rule).
17636
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017637 - : does not apply (no cookie set in configuration).
17638
17639 - the last character reports what operations were performed on the persistence
17640 cookie returned by the server (only in HTTP mode) :
17641
17642 N : NO cookie was provided by the server, and none was inserted either.
17643
17644 I : no cookie was provided by the server, and the proxy INSERTED one.
17645 Note that in "cookie insert" mode, if the server provides a cookie,
17646 it will still be overwritten and reported as "I" here.
17647
Willy Tarreau996a92c2010-10-13 19:30:47 +020017648 U : the proxy UPDATED the last date in the cookie that was presented by
17649 the client. This can only happen in insert mode with "maxidle". It
Jarno Huuskonen0e82b922014-04-12 18:22:19 +030017650 happens every time there is activity at a different date than the
Willy Tarreau996a92c2010-10-13 19:30:47 +020017651 date indicated in the cookie. If any other change happens, such as
17652 a redispatch, then the cookie will be marked as inserted instead.
17653
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017654 P : a cookie was PROVIDED by the server and transmitted as-is.
17655
17656 R : the cookie provided by the server was REWRITTEN by the proxy, which
17657 happens in "cookie rewrite" or "cookie prefix" modes.
17658
17659 D : the cookie provided by the server was DELETED by the proxy.
17660
17661 - : does not apply (no cookie set in configuration).
17662
Willy Tarreau996a92c2010-10-13 19:30:47 +020017663The combination of the two first flags gives a lot of information about what
17664was happening when the session terminated, and why it did terminate. It can be
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017665helpful to detect server saturation, network troubles, local system resource
17666starvation, attacks, etc...
17667
17668The most common termination flags combinations are indicated below. They are
17669alphabetically sorted, with the lowercase set just after the upper case for
17670easier finding and understanding.
17671
17672 Flags Reason
17673
17674 -- Normal termination.
17675
17676 CC The client aborted before the connection could be established to the
17677 server. This can happen when haproxy tries to connect to a recently
17678 dead (or unchecked) server, and the client aborts while haproxy is
17679 waiting for the server to respond or for "timeout connect" to expire.
17680
17681 CD The client unexpectedly aborted during data transfer. This can be
17682 caused by a browser crash, by an intermediate equipment between the
17683 client and haproxy which decided to actively break the connection,
17684 by network routing issues between the client and haproxy, or by a
17685 keep-alive session between the server and the client terminated first
17686 by the client.
Willy Tarreaud72758d2010-01-12 10:42:19 +010017687
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017688 cD The client did not send nor acknowledge any data for as long as the
17689 "timeout client" delay. This is often caused by network failures on
Cyril Bontédc4d9032012-04-08 21:57:39 +020017690 the client side, or the client simply leaving the net uncleanly.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017691
17692 CH The client aborted while waiting for the server to start responding.
17693 It might be the server taking too long to respond or the client
17694 clicking the 'Stop' button too fast.
17695
17696 cH The "timeout client" stroke while waiting for client data during a
17697 POST request. This is sometimes caused by too large TCP MSS values
17698 for PPPoE networks which cannot transport full-sized packets. It can
17699 also happen when client timeout is smaller than server timeout and
17700 the server takes too long to respond.
17701
17702 CQ The client aborted while its session was queued, waiting for a server
17703 with enough empty slots to accept it. It might be that either all the
17704 servers were saturated or that the assigned server was taking too
17705 long a time to respond.
17706
17707 CR The client aborted before sending a full HTTP request. Most likely
17708 the request was typed by hand using a telnet client, and aborted
17709 too early. The HTTP status code is likely a 400 here. Sometimes this
17710 might also be caused by an IDS killing the connection between haproxy
Willy Tarreau0f228a02015-05-01 15:37:53 +020017711 and the client. "option http-ignore-probes" can be used to ignore
17712 connections without any data transfer.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017713
17714 cR The "timeout http-request" stroke before the client sent a full HTTP
17715 request. This is sometimes caused by too large TCP MSS values on the
17716 client side for PPPoE networks which cannot transport full-sized
17717 packets, or by clients sending requests by hand and not typing fast
17718 enough, or forgetting to enter the empty line at the end of the
Willy Tarreau2705a612014-05-23 17:38:34 +020017719 request. The HTTP status code is likely a 408 here. Note: recently,
Willy Tarreau0f228a02015-05-01 15:37:53 +020017720 some browsers started to implement a "pre-connect" feature consisting
17721 in speculatively connecting to some recently visited web sites just
17722 in case the user would like to visit them. This results in many
17723 connections being established to web sites, which end up in 408
17724 Request Timeout if the timeout strikes first, or 400 Bad Request when
17725 the browser decides to close them first. These ones pollute the log
17726 and feed the error counters. Some versions of some browsers have even
17727 been reported to display the error code. It is possible to work
Davor Ocelice9ed2812017-12-25 17:49:28 +010017728 around the undesirable effects of this behavior by adding "option
Willy Tarreau0f228a02015-05-01 15:37:53 +020017729 http-ignore-probes" in the frontend, resulting in connections with
17730 zero data transfer to be totally ignored. This will definitely hide
17731 the errors of people experiencing connectivity issues though.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017732
17733 CT The client aborted while its session was tarpitted. It is important to
17734 check if this happens on valid requests, in order to be sure that no
Willy Tarreau55165fe2009-05-10 12:02:55 +020017735 wrong tarpit rules have been written. If a lot of them happen, it
17736 might make sense to lower the "timeout tarpit" value to something
17737 closer to the average reported "Tw" timer, in order not to consume
17738 resources for just a few attackers.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017739
Willy Tarreau570f2212013-06-10 16:42:09 +020017740 LR The request was intercepted and locally handled by haproxy. Generally
17741 it means that this was a redirect or a stats request.
17742
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +010017743 SC The server or an equipment between it and haproxy explicitly refused
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017744 the TCP connection (the proxy received a TCP RST or an ICMP message
17745 in return). Under some circumstances, it can also be the network
Davor Ocelice9ed2812017-12-25 17:49:28 +010017746 stack telling the proxy that the server is unreachable (e.g. no route,
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017747 or no ARP response on local network). When this happens in HTTP mode,
17748 the status code is likely a 502 or 503 here.
17749
17750 sC The "timeout connect" stroke before a connection to the server could
17751 complete. When this happens in HTTP mode, the status code is likely a
17752 503 or 504 here.
17753
17754 SD The connection to the server died with an error during the data
17755 transfer. This usually means that haproxy has received an RST from
17756 the server or an ICMP message from an intermediate equipment while
17757 exchanging data with the server. This can be caused by a server crash
17758 or by a network issue on an intermediate equipment.
17759
17760 sD The server did not send nor acknowledge any data for as long as the
17761 "timeout server" setting during the data phase. This is often caused
Davor Ocelice9ed2812017-12-25 17:49:28 +010017762 by too short timeouts on L4 equipment before the server (firewalls,
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017763 load-balancers, ...), as well as keep-alive sessions maintained
17764 between the client and the server expiring first on haproxy.
17765
17766 SH The server aborted before sending its full HTTP response headers, or
17767 it crashed while processing the request. Since a server aborting at
17768 this moment is very rare, it would be wise to inspect its logs to
17769 control whether it crashed and why. The logged request may indicate a
17770 small set of faulty requests, demonstrating bugs in the application.
17771 Sometimes this might also be caused by an IDS killing the connection
17772 between haproxy and the server.
17773
17774 sH The "timeout server" stroke before the server could return its
17775 response headers. This is the most common anomaly, indicating too
17776 long transactions, probably caused by server or database saturation.
17777 The immediate workaround consists in increasing the "timeout server"
17778 setting, but it is important to keep in mind that the user experience
17779 will suffer from these long response times. The only long term
17780 solution is to fix the application.
17781
17782 sQ The session spent too much time in queue and has been expired. See
17783 the "timeout queue" and "timeout connect" settings to find out how to
17784 fix this if it happens too often. If it often happens massively in
17785 short periods, it may indicate general problems on the affected
17786 servers due to I/O or database congestion, or saturation caused by
17787 external attacks.
17788
17789 PC The proxy refused to establish a connection to the server because the
17790 process' socket limit has been reached while attempting to connect.
Cyril Bontédc4d9032012-04-08 21:57:39 +020017791 The global "maxconn" parameter may be increased in the configuration
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017792 so that it does not happen anymore. This status is very rare and
17793 might happen when the global "ulimit-n" parameter is forced by hand.
17794
Willy Tarreaued2fd2d2010-12-29 11:23:27 +010017795 PD The proxy blocked an incorrectly formatted chunked encoded message in
17796 a request or a response, after the server has emitted its headers. In
17797 most cases, this will indicate an invalid message from the server to
Davor Ocelice9ed2812017-12-25 17:49:28 +010017798 the client. HAProxy supports chunk sizes of up to 2GB - 1 (2147483647
Willy Tarreauf3a3e132013-08-31 08:16:26 +020017799 bytes). Any larger size will be considered as an error.
Willy Tarreaued2fd2d2010-12-29 11:23:27 +010017800
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017801 PH The proxy blocked the server's response, because it was invalid,
17802 incomplete, dangerous (cache control), or matched a security filter.
17803 In any case, an HTTP 502 error is sent to the client. One possible
17804 cause for this error is an invalid syntax in an HTTP header name
Willy Tarreaued2fd2d2010-12-29 11:23:27 +010017805 containing unauthorized characters. It is also possible but quite
17806 rare, that the proxy blocked a chunked-encoding request from the
17807 client due to an invalid syntax, before the server responded. In this
17808 case, an HTTP 400 error is sent to the client and reported in the
17809 logs.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017810
17811 PR The proxy blocked the client's HTTP request, either because of an
17812 invalid HTTP syntax, in which case it returned an HTTP 400 error to
17813 the client, or because a deny filter matched, in which case it
17814 returned an HTTP 403 error.
17815
17816 PT The proxy blocked the client's request and has tarpitted its
17817 connection before returning it a 500 server error. Nothing was sent
17818 to the server. The connection was maintained open for as long as
17819 reported by the "Tw" timer field.
17820
17821 RC A local resource has been exhausted (memory, sockets, source ports)
17822 preventing the connection to the server from establishing. The error
17823 logs will tell precisely what was missing. This is very rare and can
17824 only be solved by proper system tuning.
17825
Willy Tarreau996a92c2010-10-13 19:30:47 +020017826The combination of the two last flags gives a lot of information about how
17827persistence was handled by the client, the server and by haproxy. This is very
17828important to troubleshoot disconnections, when users complain they have to
17829re-authenticate. The commonly encountered flags are :
17830
17831 -- Persistence cookie is not enabled.
17832
17833 NN No cookie was provided by the client, none was inserted in the
17834 response. For instance, this can be in insert mode with "postonly"
17835 set on a GET request.
17836
17837 II A cookie designating an invalid server was provided by the client,
17838 a valid one was inserted in the response. This typically happens when
Jamie Gloudonaaa21002012-08-25 00:18:33 -040017839 a "server" entry is removed from the configuration, since its cookie
Willy Tarreau996a92c2010-10-13 19:30:47 +020017840 value can be presented by a client when no other server knows it.
17841
17842 NI No cookie was provided by the client, one was inserted in the
17843 response. This typically happens for first requests from every user
17844 in "insert" mode, which makes it an easy way to count real users.
17845
17846 VN A cookie was provided by the client, none was inserted in the
17847 response. This happens for most responses for which the client has
17848 already got a cookie.
17849
17850 VU A cookie was provided by the client, with a last visit date which is
17851 not completely up-to-date, so an updated cookie was provided in
17852 response. This can also happen if there was no date at all, or if
17853 there was a date but the "maxidle" parameter was not set, so that the
17854 cookie can be switched to unlimited time.
17855
17856 EI A cookie was provided by the client, with a last visit date which is
17857 too old for the "maxidle" parameter, so the cookie was ignored and a
17858 new cookie was inserted in the response.
17859
17860 OI A cookie was provided by the client, with a first visit date which is
17861 too old for the "maxlife" parameter, so the cookie was ignored and a
17862 new cookie was inserted in the response.
17863
17864 DI The server designated by the cookie was down, a new server was
17865 selected and a new cookie was emitted in the response.
17866
17867 VI The server designated by the cookie was not marked dead but could not
17868 be reached. A redispatch happened and selected another one, which was
17869 then advertised in the response.
17870
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017871
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200178728.6. Non-printable characters
17873-----------------------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017874
17875In order not to cause trouble to log analysis tools or terminals during log
17876consulting, non-printable characters are not sent as-is into log files, but are
17877converted to the two-digits hexadecimal representation of their ASCII code,
17878prefixed by the character '#'. The only characters that can be logged without
17879being escaped are comprised between 32 and 126 (inclusive). Obviously, the
17880escape character '#' itself is also encoded to avoid any ambiguity ("#23"). It
17881is the same for the character '"' which becomes "#22", as well as '{', '|' and
17882'}' when logging headers.
17883
17884Note that the space character (' ') is not encoded in headers, which can cause
17885issues for tools relying on space count to locate fields. A typical header
17886containing spaces is "User-Agent".
17887
17888Last, it has been observed that some syslog daemons such as syslog-ng escape
17889the quote ('"') with a backslash ('\'). The reverse operation can safely be
17890performed since no quote may appear anywhere else in the logs.
17891
17892
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200178938.7. Capturing HTTP cookies
17894---------------------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017895
17896Cookie capture simplifies the tracking a complete user session. This can be
17897achieved using the "capture cookie" statement in the frontend. Please refer to
Willy Tarreauc57f0e22009-05-10 13:12:33 +020017898section 4.2 for more details. Only one cookie can be captured, and the same
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017899cookie will simultaneously be checked in the request ("Cookie:" header) and in
17900the response ("Set-Cookie:" header). The respective values will be reported in
17901the HTTP logs at the "captured_request_cookie" and "captured_response_cookie"
Willy Tarreauc57f0e22009-05-10 13:12:33 +020017902locations (see section 8.2.3 about HTTP log format). When either cookie is
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017903not seen, a dash ('-') replaces the value. This way, it's easy to detect when a
17904user switches to a new session for example, because the server will reassign it
17905a new cookie. It is also possible to detect if a server unexpectedly sets a
17906wrong cookie to a client, leading to session crossing.
17907
17908 Examples :
17909 # capture the first cookie whose name starts with "ASPSESSION"
17910 capture cookie ASPSESSION len 32
17911
17912 # capture the first cookie whose name is exactly "vgnvisitor"
17913 capture cookie vgnvisitor= len 32
17914
17915
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200179168.8. Capturing HTTP headers
17917---------------------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017918
17919Header captures are useful to track unique request identifiers set by an upper
17920proxy, virtual host names, user-agents, POST content-length, referrers, etc. In
17921the response, one can search for information about the response length, how the
17922server asked the cache to behave, or an object location during a redirection.
17923
17924Header captures are performed using the "capture request header" and "capture
17925response header" statements in the frontend. Please consult their definition in
Willy Tarreauc57f0e22009-05-10 13:12:33 +020017926section 4.2 for more details.
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017927
17928It is possible to include both request headers and response headers at the same
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +010017929time. Non-existent headers are logged as empty strings, and if one header
17930appears more than once, only its last occurrence will be logged. Request headers
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017931are grouped within braces '{' and '}' in the same order as they were declared,
17932and delimited with a vertical bar '|' without any space. Response headers
17933follow the same representation, but are displayed after a space following the
17934request headers block. These blocks are displayed just before the HTTP request
17935in the logs.
17936
Willy Tarreaud9ed3d22014-06-13 12:23:06 +020017937As a special case, it is possible to specify an HTTP header capture in a TCP
17938frontend. The purpose is to enable logging of headers which will be parsed in
17939an HTTP backend if the request is then switched to this HTTP backend.
17940
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017941 Example :
17942 # This instance chains to the outgoing proxy
17943 listen proxy-out
17944 mode http
17945 option httplog
17946 option logasap
17947 log global
17948 server cache1 192.168.1.1:3128
17949
17950 # log the name of the virtual server
17951 capture request header Host len 20
17952
17953 # log the amount of data uploaded during a POST
17954 capture request header Content-Length len 10
17955
17956 # log the beginning of the referrer
17957 capture request header Referer len 20
17958
17959 # server name (useful for outgoing proxies only)
17960 capture response header Server len 20
17961
17962 # logging the content-length is useful with "option logasap"
17963 capture response header Content-Length len 10
17964
Davor Ocelice9ed2812017-12-25 17:49:28 +010017965 # log the expected cache behavior on the response
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017966 capture response header Cache-Control len 8
17967
17968 # the Via header will report the next proxy's name
17969 capture response header Via len 20
17970
17971 # log the URL location during a redirection
17972 capture response header Location len 20
17973
17974 >>> Aug 9 20:26:09 localhost \
17975 haproxy[2022]: 127.0.0.1:34014 [09/Aug/2004:20:26:09] proxy-out \
17976 proxy-out/cache1 0/0/0/162/+162 200 +350 - - ---- 0/0/0/0/0 0/0 \
17977 {fr.adserver.yahoo.co||http://fr.f416.mail.} {|864|private||} \
17978 "GET http://fr.adserver.yahoo.com/"
17979
17980 >>> Aug 9 20:30:46 localhost \
17981 haproxy[2022]: 127.0.0.1:34020 [09/Aug/2004:20:30:46] proxy-out \
17982 proxy-out/cache1 0/0/0/182/+182 200 +279 - - ---- 0/0/0/0/0 0/0 \
17983 {w.ods.org||} {Formilux/0.1.8|3495|||} \
Willy Tarreaud72758d2010-01-12 10:42:19 +010017984 "GET http://trafic.1wt.eu/ HTTP/1.1"
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017985
17986 >>> Aug 9 20:30:46 localhost \
17987 haproxy[2022]: 127.0.0.1:34028 [09/Aug/2004:20:30:46] proxy-out \
17988 proxy-out/cache1 0/0/2/126/+128 301 +223 - - ---- 0/0/0/0/0 0/0 \
17989 {www.sytadin.equipement.gouv.fr||http://trafic.1wt.eu/} \
17990 {Apache|230|||http://www.sytadin.} \
Willy Tarreaud72758d2010-01-12 10:42:19 +010017991 "GET http://www.sytadin.equipement.gouv.fr/ HTTP/1.1"
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017992
17993
Willy Tarreauc57f0e22009-05-10 13:12:33 +0200179948.9. Examples of logs
17995---------------------
Willy Tarreaucc6c8912009-02-22 10:53:55 +010017996
17997These are real-world examples of logs accompanied with an explanation. Some of
17998them have been made up by hand. The syslog part has been removed for better
17999reading. Their sole purpose is to explain how to decipher them.
18000
18001 >>> haproxy[674]: 127.0.0.1:33318 [15/Oct/2003:08:31:57.130] px-http \
18002 px-http/srv1 6559/0/7/147/6723 200 243 - - ---- 5/3/3/1/0 0/0 \
18003 "HEAD / HTTP/1.0"
18004
18005 => long request (6.5s) entered by hand through 'telnet'. The server replied
18006 in 147 ms, and the session ended normally ('----')
18007
18008 >>> haproxy[674]: 127.0.0.1:33319 [15/Oct/2003:08:31:57.149] px-http \
18009 px-http/srv1 6559/1230/7/147/6870 200 243 - - ---- 324/239/239/99/0 \
18010 0/9 "HEAD / HTTP/1.0"
18011
18012 => Idem, but the request was queued in the global queue behind 9 other
18013 requests, and waited there for 1230 ms.
18014
18015 >>> haproxy[674]: 127.0.0.1:33320 [15/Oct/2003:08:32:17.654] px-http \
18016 px-http/srv1 9/0/7/14/+30 200 +243 - - ---- 3/3/3/1/0 0/0 \
18017 "GET /image.iso HTTP/1.0"
18018
18019 => request for a long data transfer. The "logasap" option was specified, so
Krzysztof Piotr Oledzkif8645332009-12-13 21:55:50 +010018020 the log was produced just before transferring data. The server replied in
Willy Tarreaucc6c8912009-02-22 10:53:55 +010018021 14 ms, 243 bytes of headers were sent to the client, and total time from
18022 accept to first data byte is 30 ms.
18023
18024 >>> haproxy[674]: 127.0.0.1:33320 [15/Oct/2003:08:32:17.925] px-http \
18025 px-http/srv1 9/0/7/14/30 502 243 - - PH-- 3/2/2/0/0 0/0 \
18026 "GET /cgi-bin/bug.cgi? HTTP/1.0"
18027
18028 => the proxy blocked a server response either because of an "rspdeny" or
18029 "rspideny" filter, or because the response was improperly formatted and
Willy Tarreau3c92c5f2011-08-28 09:45:47 +020018030 not HTTP-compliant, or because it blocked sensitive information which
Willy Tarreaucc6c8912009-02-22 10:53:55 +010018031 risked being cached. In this case, the response is replaced with a "502
18032 bad gateway". The flags ("PH--") tell us that it was haproxy who decided
18033 to return the 502 and not the server.
18034
18035 >>> haproxy[18113]: 127.0.0.1:34548 [15/Oct/2003:15:18:55.798] px-http \
Willy Tarreaud72758d2010-01-12 10:42:19 +010018036 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 +010018037
18038 => the client never completed its request and aborted itself ("C---") after
18039 8.5s, while the proxy was waiting for the request headers ("-R--").
18040 Nothing was sent to any server.
18041
18042 >>> haproxy[18113]: 127.0.0.1:34549 [15/Oct/2003:15:19:06.103] px-http \
18043 px-http/<NOSRV> -1/-1/-1/-1/50001 408 0 - - cR-- 2/2/2/0/0 0/0 ""
18044
18045 => The client never completed its request, which was aborted by the
18046 time-out ("c---") after 50s, while the proxy was waiting for the request
Davor Ocelice9ed2812017-12-25 17:49:28 +010018047 headers ("-R--"). Nothing was sent to any server, but the proxy could
Willy Tarreaucc6c8912009-02-22 10:53:55 +010018048 send a 408 return code to the client.
18049
18050 >>> haproxy[18989]: 127.0.0.1:34550 [15/Oct/2003:15:24:28.312] px-tcp \
18051 px-tcp/srv1 0/0/5007 0 cD 0/0/0/0/0 0/0
18052
18053 => This log was produced with "option tcplog". The client timed out after
18054 5 seconds ("c----").
18055
18056 >>> haproxy[18989]: 10.0.0.1:34552 [15/Oct/2003:15:26:31.462] px-http \
18057 px-http/srv1 3183/-1/-1/-1/11215 503 0 - - SC-- 205/202/202/115/3 \
Willy Tarreaud72758d2010-01-12 10:42:19 +010018058 0/0 "HEAD / HTTP/1.0"
Willy Tarreaucc6c8912009-02-22 10:53:55 +010018059
18060 => The request took 3s to complete (probably a network problem), and the
Willy Tarreauc57f0e22009-05-10 13:12:33 +020018061 connection to the server failed ('SC--') after 4 attempts of 2 seconds
Willy Tarreaucc6c8912009-02-22 10:53:55 +010018062 (config says 'retries 3'), and no redispatch (otherwise we would have
18063 seen "/+3"). Status code 503 was returned to the client. There were 115
18064 connections on this server, 202 connections on this proxy, and 205 on
18065 the global process. It is possible that the server refused the
18066 connection because of too many already established.
Willy Tarreau844e3c52008-01-11 16:28:18 +010018067
Willy Tarreau52b2d222011-09-07 23:48:48 +020018068
Christopher Fauletc3fe5332016-04-07 15:30:10 +0200180699. Supported filters
18070--------------------
18071
18072Here are listed officially supported filters with the list of parameters they
18073accept. Depending on compile options, some of these filters might be
18074unavailable. The list of available filters is reported in haproxy -vv.
18075
18076See also : "filter"
18077
180789.1. Trace
18079----------
18080
Christopher Faulet31bfe1f2016-12-09 17:42:38 +010018081filter trace [name <name>] [random-parsing] [random-forwarding] [hexdump]
Christopher Fauletc3fe5332016-04-07 15:30:10 +020018082
18083 Arguments:
18084 <name> is an arbitrary name that will be reported in
18085 messages. If no name is provided, "TRACE" is used.
18086
18087 <random-parsing> enables the random parsing of data exchanged between
18088 the client and the server. By default, this filter
18089 parses all available data. With this parameter, it
18090 only parses a random amount of the available data.
18091
Davor Ocelice9ed2812017-12-25 17:49:28 +010018092 <random-forwarding> enables the random forwarding of parsed data. By
Christopher Fauletc3fe5332016-04-07 15:30:10 +020018093 default, this filter forwards all previously parsed
18094 data. With this parameter, it only forwards a random
18095 amount of the parsed data.
18096
Davor Ocelice9ed2812017-12-25 17:49:28 +010018097 <hexdump> dumps all forwarded data to the server and the client.
Christopher Faulet31bfe1f2016-12-09 17:42:38 +010018098
Christopher Fauletc3fe5332016-04-07 15:30:10 +020018099This filter can be used as a base to develop new filters. It defines all
18100callbacks and print a message on the standard error stream (stderr) with useful
18101information for all of them. It may be useful to debug the activity of other
18102filters or, quite simply, HAProxy's activity.
18103
18104Using <random-parsing> and/or <random-forwarding> parameters is a good way to
18105tests the behavior of a filter that parses data exchanged between a client and
18106a server by adding some latencies in the processing.
18107
18108
181099.2. HTTP compression
18110---------------------
18111
18112filter compression
18113
18114The HTTP compression has been moved in a filter in HAProxy 1.7. "compression"
18115keyword must still be used to enable and configure the HTTP compression. And
Christopher Faulet27d93c32018-12-15 22:32:02 +010018116when no other filter is used, it is enough. When used with the cache enabled,
18117it is also enough. In this case, the compression is always done after the
18118response is stored in the cache. But it is mandatory to explicitly use a filter
18119line to enable the HTTP compression when at least one filter other than the
18120cache is used for the same listener/frontend/backend. This is important to know
18121the filters evaluation order.
Christopher Fauletc3fe5332016-04-07 15:30:10 +020018122
Christopher Faulet27d93c32018-12-15 22:32:02 +010018123See also : "compression" and section 9.4 about the cache filter.
Christopher Fauletc3fe5332016-04-07 15:30:10 +020018124
18125
Christopher Fauletf7e4e7e2016-10-27 22:29:49 +0200181269.3. Stream Processing Offload Engine (SPOE)
18127--------------------------------------------
18128
18129filter spoe [engine <name>] config <file>
18130
18131 Arguments :
18132
18133 <name> is the engine name that will be used to find the right scope in
18134 the configuration file. If not provided, all the file will be
18135 parsed.
18136
18137 <file> is the path of the engine configuration file. This file can
18138 contain configuration of several engines. In this case, each
18139 part must be placed in its own scope.
18140
18141The Stream Processing Offload Engine (SPOE) is a filter communicating with
18142external components. It allows the offload of some specifics processing on the
Davor Ocelice9ed2812017-12-25 17:49:28 +010018143streams in tiered applications. These external components and information
Christopher Fauletf7e4e7e2016-10-27 22:29:49 +020018144exchanged with them are configured in dedicated files, for the main part. It
18145also requires dedicated backends, defined in HAProxy configuration.
18146
18147SPOE communicates with external components using an in-house binary protocol,
18148the Stream Processing Offload Protocol (SPOP).
18149
Tim Düsterhus4896c442016-11-29 02:15:19 +010018150For all information about the SPOE configuration and the SPOP specification, see
Christopher Fauletf7e4e7e2016-10-27 22:29:49 +020018151"doc/SPOE.txt".
18152
18153Important note:
18154 The SPOE filter is highly experimental for now and was not heavily
18155 tested. It is really not production ready. So use it carefully.
18156
Christopher Faulet99a17a22018-12-11 09:18:27 +0100181579.4. Cache
18158----------
18159
18160filter cache <name>
18161
18162 Arguments :
18163
18164 <name> is name of the cache section this filter will use.
18165
18166The cache uses a filter to store cacheable responses. The HTTP rules
18167"cache-store" and "cache-use" must be used to define how and when to use a
John Roeslerfb2fce12019-07-10 15:45:51 -050018168cache. By default the corresponding filter is implicitly defined. And when no
Christopher Faulet27d93c32018-12-15 22:32:02 +010018169other filters than cache or compression are used, it is enough. In such case,
18170the compression filter is always evaluated after the cache filter. But it is
18171mandatory to explicitly use a filter line to use a cache when at least one
18172filter other than the compression is used for the same
18173listener/frontend/backend. This is important to know the filters evaluation
18174order.
Christopher Faulet99a17a22018-12-11 09:18:27 +010018175
Christopher Faulet27d93c32018-12-15 22:32:02 +010018176See also : section 9.2 about the compression filter and section 10 about cache.
Christopher Faulet99a17a22018-12-11 09:18:27 +010018177
William Lallemand86d0df02017-11-24 21:36:45 +01001817810. Cache
18179---------
18180
18181HAProxy provides a cache, which was designed to perform cache on small objects
18182(favicon, css...). This is a minimalist low-maintenance cache which runs in
18183RAM.
18184
18185The cache is based on a memory which is shared between processes and threads,
Cyril Bonté7b888f12017-11-26 22:24:31 +010018186this memory is split in blocks of 1k.
William Lallemand86d0df02017-11-24 21:36:45 +010018187
18188If an object is not used anymore, it can be deleted to store a new object
18189independently of its expiration date. The oldest objects are deleted first
18190when we try to allocate a new one.
18191
Cyril Bonté7b888f12017-11-26 22:24:31 +010018192The cache uses a hash of the host header and the URI as the key.
William Lallemand86d0df02017-11-24 21:36:45 +010018193
18194It's possible to view the status of a cache using the Unix socket command
18195"show cache" consult section 9.3 "Unix Socket commands" of Management Guide
18196for more details.
18197
18198When an object is delivered from the cache, the server name in the log is
18199replaced by "<CACHE>".
18200
Cyril Bonté7b888f12017-11-26 22:24:31 +01001820110.1. Limitation
18202----------------
William Lallemand86d0df02017-11-24 21:36:45 +010018203
18204The cache won't store and won't deliver objects in these cases:
18205
18206- If the response is not a 200
18207- If the response contains a Vary header
Frédéric Lécaille5f8bea62018-10-23 10:09:19 +020018208- If the Content-Length + the headers size is greater than "max-object-size"
William Lallemand86d0df02017-11-24 21:36:45 +010018209- If the response is not cacheable
18210
18211- If the request is not a GET
18212- If the HTTP version of the request is smaller than 1.1
William Lallemand8a16fe02018-05-22 11:04:33 +020018213- If the request contains an Authorization header
William Lallemand86d0df02017-11-24 21:36:45 +010018214
Christopher Faulet99a17a22018-12-11 09:18:27 +010018215Caution!: For HAProxy version prior to 1.9, due to the limitation of the
18216filters, it is not recommended to use the cache with other filters. Using them
18217can cause undefined behavior if they modify the response (compression for
18218example). For HAProxy 1.9 and greater, it is safe, for HTX proxies only (see
18219"option http-use-htx" for details).
William Lallemand86d0df02017-11-24 21:36:45 +010018220
Cyril Bonté7b888f12017-11-26 22:24:31 +01001822110.2. Setup
18222-----------
William Lallemand86d0df02017-11-24 21:36:45 +010018223
18224To setup a cache, you must define a cache section and use it in a proxy with
18225the corresponding http-request and response actions.
18226
Cyril Bonté7b888f12017-11-26 22:24:31 +01001822710.2.1. Cache section
18228---------------------
William Lallemand86d0df02017-11-24 21:36:45 +010018229
18230cache <name>
18231 Declare a cache section, allocate a shared cache memory named <name>, the
18232 size of cache is mandatory.
18233
18234total-max-size <megabytes>
Davor Ocelice9ed2812017-12-25 17:49:28 +010018235 Define the size in RAM of the cache in megabytes. This size is split in
Frédéric Lécaillee3c83d82018-10-25 10:46:40 +020018236 blocks of 1kB which are used by the cache entries. Its maximum value is 4095.
William Lallemand86d0df02017-11-24 21:36:45 +010018237
Frédéric Lécaille5f8bea62018-10-23 10:09:19 +020018238max-object-size <bytes>
Frédéric Lécaillee3c83d82018-10-25 10:46:40 +020018239 Define the maximum size of the objects to be cached. Must not be greater than
18240 an half of "total-max-size". If not set, it equals to a 256th of the cache size.
18241 All objects with sizes larger than "max-object-size" will not be cached.
Frédéric Lécaille5f8bea62018-10-23 10:09:19 +020018242
William Lallemand86d0df02017-11-24 21:36:45 +010018243max-age <seconds>
18244 Define the maximum expiration duration. The expiration is set has the lowest
18245 value between the s-maxage or max-age (in this order) directive in the
18246 Cache-Control response header and this value. The default value is 60
18247 seconds, which means that you can't cache an object more than 60 seconds by
18248 default.
18249
Cyril Bonté7b888f12017-11-26 22:24:31 +01001825010.2.2. Proxy section
18251---------------------
William Lallemand86d0df02017-11-24 21:36:45 +010018252
Jarno Huuskonen251a6b72019-01-04 14:05:02 +020018253http-request cache-use <name> [ { if | unless } <condition> ]
William Lallemand86d0df02017-11-24 21:36:45 +010018254 Try to deliver a cached object from the cache <name>. This directive is also
18255 mandatory to store the cache as it calculates the cache hash. If you want to
18256 use a condition for both storage and delivering that's a good idea to put it
18257 after this one.
18258
Jarno Huuskonen251a6b72019-01-04 14:05:02 +020018259http-response cache-store <name> [ { if | unless } <condition> ]
William Lallemand86d0df02017-11-24 21:36:45 +010018260 Store an http-response within the cache. The storage of the response headers
18261 is done at this step, which means you can use others http-response actions
18262 to modify headers before or after the storage of the response. This action
18263 is responsible for the setup of the cache storage filter.
18264
18265
18266Example:
18267
18268 backend bck1
18269 mode http
18270
18271 http-request cache-use foobar
18272 http-response cache-store foobar
18273 server srv1 127.0.0.1:80
18274
18275 cache foobar
18276 total-max-size 4
18277 max-age 240
18278
Willy Tarreau0ba27502007-12-24 16:55:16 +010018279/*
18280 * Local variables:
18281 * fill-column: 79
18282 * End:
18283 */