1 <?xml version='1.0' encoding='UTF-8' ?>
2 <!DOCTYPE manualpage SYSTEM "../style/manualpage.dtd">
3 <?xml-stylesheet type="text/xsl" href="../style/manual.en.xsl"?>
4 <!-- $LastChangedRevision$ -->
7 Licensed to the Apache Software Foundation (ASF) under one or more
8 contributor license agreements. See the NOTICE file distributed with
9 this work for additional information regarding copyright ownership.
10 The ASF licenses this file to You under the Apache License, Version 2.0
11 (the "License"); you may not use this file except in compliance with
12 the License. You may obtain a copy of the License at
14 http://www.apache.org/licenses/LICENSE-2.0
16 Unless required by applicable law or agreed to in writing, software
17 distributed under the License is distributed on an "AS IS" BASIS,
18 WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
19 See the License for the specific language governing permissions and
20 limitations under the License.
23 <manualpage metafile="ab.xml.meta">
24 <parentdocument href="./">Programs</parentdocument>
26 <title>ab - Apache HTTP server benchmarking tool</title>
29 <p><code>ab</code> is a tool for benchmarking your Apache Hypertext
30 Transfer Protocol (HTTP) server. It is designed to give you an impression
31 of how your current Apache installation performs. This especially shows
32 you how many requests per second your Apache installation is capable of
35 <seealso><program>httpd</program></seealso>
37 <section id="synopsis"><title>Synopsis</title>
38 <p><code><strong>ab</strong>
39 [ -<strong>A</strong> <var>auth-username</var>:<var>password</var> ]
40 [ -<strong>b</strong> <var>windowsize</var> ]
41 [ -<strong>B</strong> <var>local-address</var> ]
42 [ -<strong>c</strong> <var>concurrency</var> ]
43 [ -<strong>C</strong> <var>cookie-name</var>=<var>value</var> ]
44 [ -<strong>d</strong> ]
45 [ -<strong>e</strong> <var>csv-file</var> ]
46 [ -<strong>f</strong> <var>protocol</var> ]
47 [ -<strong>g</strong> <var>gnuplot-file</var> ]
48 [ -<strong>h</strong> ]
49 [ -<strong>H</strong> <var>custom-header</var> ]
50 [ -<strong>i</strong> ]
51 [ -<strong>k</strong> ]
52 [ -<strong>n</strong> <var>requests</var> ]
53 [ -<strong>p</strong> <var>POST-file</var> ]
54 [ -<strong>P</strong> <var>proxy-auth-username</var>:<var>password</var> ]
55 [ -<strong>q</strong> ]
56 [ -<strong>r</strong> ]
57 [ -<strong>s</strong> ]
58 [ -<strong>S</strong> ]
59 [ -<strong>t</strong> <var>timelimit</var> ]
60 [ -<strong>T</strong> <var>content-type</var> ]
61 [ -<strong>u</strong> <var>PUT-file</var> ]
62 [ -<strong>v</strong> <var>verbosity</var>]
63 [ -<strong>V</strong> ]
64 [ -<strong>w</strong> ]
65 [ -<strong>x</strong> <var><table>-attributes</var> ]
66 [ -<strong>X</strong> <var>proxy</var>[:<var>port</var>] ]
67 [ -<strong>y</strong> <var><tr>-attributes</var> ]
68 [ -<strong>z</strong> <var><td>-attributes</var> ]
69 [ -<strong>Z</strong> <var>ciphersuite</var> ]
70 [http[s]://]<var>hostname</var>[:<var>port</var>]/<var>path</var></code></p>
73 <section id="options"><title>Options</title>
75 <dt><code>-A <var>auth-username</var>:<var>password</var></code></dt>
76 <dd>Supply BASIC Authentication credentials to the server. The username and
77 password are separated by a single <code>:</code> and sent on the wire
78 base64 encoded. The string is sent regardless of whether the server needs
79 it (<em>i.e.</em>, has sent an 401 authentication needed).</dd>
81 <dt><code>-b <var>windowsize</var></code></dt>
82 <dd>Size of TCP send/receive buffer, in bytes.</dd>
84 <dt><code>-B <var>local-address</var></code></dt>
85 <dd>Address to bind to when making outgoing connections.</dd>
87 <dt><code>-c <var>concurrency</var></code></dt>
88 <dd>Number of multiple requests to perform at a time. Default is one
89 request at a time.</dd>
91 <dt><code>-C <var>cookie-name</var>=<var>value</var></code></dt>
92 <dd>Add a <code>Cookie:</code> line to the request. The argument is
93 typically in the form of a <code><var>name</var>=<var>value</var></code>
94 pair. This field is repeatable.</dd>
96 <dt><code>-d</code></dt>
97 <dd>Do not display the "percentage served within XX [ms] table". (legacy
100 <dt><code>-e <var>csv-file</var></code></dt>
101 <dd>Write a Comma separated value (CSV) file which contains for each
102 percentage (from 1% to 100%) the time (in milliseconds) it took to serve
103 that percentage of the requests. This is usually more useful than the
104 'gnuplot' file; as the results are already 'binned'.</dd>
106 <dt><code>-f <var>protocol</var></code></dt>
107 <dd>Specify SSL/TLS protocol (SSL2, SSL3, TLS1, or ALL).</dd>
109 <dt><code>-g <var>gnuplot-file</var></code></dt>
110 <dd>Write all measured values out as a 'gnuplot' or TSV (Tab separate
111 values) file. This file can easily be imported into packages like Gnuplot,
112 IDL, Mathematica, Igor or even Excel. The labels are on the first line of
115 <dt><code>-h</code></dt>
116 <dd>Display usage information.</dd>
118 <dt><code>-H <var>custom-header</var></code></dt>
119 <dd>Append extra headers to the request. The argument is typically in
120 the form of a valid header line, containing a colon-separated field-value
121 pair (<em>i.e.</em>, <code>"Accept-Encoding: zip/zop;8bit"</code>).</dd>
123 <dt><code>-i</code></dt>
124 <dd>Do <code>HEAD</code> requests instead of <code>GET</code>.</dd>
126 <dt><code>-k</code></dt>
127 <dd>Enable the HTTP KeepAlive feature, <em>i.e.</em>, perform multiple
128 requests within one HTTP session. Default is no KeepAlive.</dd>
130 <dt><code>-n <var>requests</var></code></dt>
131 <dd>Number of requests to perform for the benchmarking session. The default
132 is to just perform a single request which usually leads to
133 non-representative benchmarking results.</dd>
135 <dt><code>-p <var>POST-file</var></code></dt>
136 <dd>File containing data to POST. Remember to also set <code>-T</code>.</dd>
138 <dt><code>-P <var>proxy-auth-username</var>:<var>password</var></code></dt>
139 <dd>Supply BASIC Authentication credentials to a proxy en-route. The
140 username and password are separated by a single <code>:</code> and sent on
141 the wire base64 encoded. The string is sent regardless of whether the
142 proxy needs it (<em>i.e.</em>, has sent an 407 proxy authentication
145 <dt><code>-q</code></dt>
146 <dd>When processing more than 150 requests, <code>ab</code> outputs a
147 progress count on <code>stderr</code> every 10% or 100 requests or so. The
148 <code>-q</code> flag will suppress these messages.</dd>
150 <dt><code>-r</code></dt>
151 <dd>Don't exit on socket receive errors.</dd>
153 <dt><code>-s</code></dt>
154 <dd>When compiled in (<code>ab -h</code> will show you) use the SSL
155 protected <code>https</code> rather than the <code>http</code> protocol.
156 This feature is experimental and <em>very</em> rudimentary. You probably
157 do not want to use it.</dd>
159 <dt><code>-S</code></dt>
160 <dd>Do not display the median and standard deviation values, nor display
161 the warning/error messages when the average and median are more than
162 one or two times the standard deviation apart. And default to the
163 min/avg/max values. (legacy support).</dd>
165 <dt><code>-t <var>timelimit</var></code></dt>
166 <dd>Maximum number of seconds to spend for benchmarking. This implies a
167 <code>-n 50000</code> internally. Use this to benchmark the server within a
168 fixed total amount of time. Per default there is no timelimit.</dd>
170 <dt><code>-T <var>content-type</var></code></dt>
171 <dd>Content-type header to use for POST/PUT data, eg.
172 <code>application/x-www-form-urlencoded</code>.
173 Default: <code>text/plain</code>.</dd>
175 <dt><code>-u <var>PUT-file</var></code></dt>
176 <dd>File containing data to PUT. Remember to also set <code>-T</code>.</dd>
178 <dt><code>-v <var>verbosity</var></code></dt>
179 <dd>Set verbosity level - <code>4</code> and above prints information on
180 headers, <code>3</code> and above prints response codes (404, 200, etc.),
181 <code>2</code> and above prints warnings and info.</dd>
183 <dt><code>-V</code></dt>
184 <dd>Display version number and exit.</dd>
186 <dt><code>-w</code></dt>
187 <dd>Print out results in HTML tables. Default table is two columns wide,
188 with a white background.</dd>
190 <dt><code>-x <var><table>-attributes</var></code></dt>
191 <dd>String to use as attributes for <code><table></code>. Attributes
192 are inserted <code><table <var>here</var> ></code>.</dd>
194 <dt><code>-X <var>proxy</var>[:<var>port</var>]</code></dt>
195 <dd>Use a proxy server for the requests.</dd>
197 <dt><code>-y <var><tr>-attributes</var></code></dt>
198 <dd>String to use as attributes for <code><tr></code>.</dd>
200 <dt><code>-z <var><td>-attributes</var></code></dt>
201 <dd>String to use as attributes for <code><td></code>.</dd>
203 <dt><code>-Z <var>ciphersuite</var></code></dt>
204 <dd>Specify SSL/TLS cipher suite (See openssl ciphers)</dd>
208 <section id="output"><title>Output</title>
209 <p>The following list describes the values returned by <code>ab</code>:
213 <dt>Server Software</dt>
214 <dd>The value, if any, returned in the <var>server</var> HTTP header
215 of the first successful response. This includes all characters in the
216 header from beginning to the point a character with decimal value of 32
217 (most notably: a space or CR/LF) is detected.</dd>
219 <dt>Server Hostname</dt>
220 <dd>The DNS or IP address given on the command line</dd>
223 <dd>The port to which ab is connecting. If no port is given on the
224 command line, this will default to 80 for http and 443 for
227 <dt>SSL/TLS Protocol</dt>
228 <dd>The protocol parameters negotiated between the client and server.
229 This will only be printed if SSL is used.</dd>
231 <dt>Document Path</dt>
232 <dd>The request URI parsed from the command line string.</dd>
234 <dt>Document Length</dt>
235 <dd>This is the size in bytes of the first successfully returned document.
236 If the document length changes during testing, the response is
237 considered an error.</dd>
239 <dt>Concurrency Level</dt>
240 <dd>The number of concurrent clients used during the test</dd>
242 <dt>Time taken for tests</dt>
243 <dd>This is the time taken from the moment the first socket connection
244 is created to the moment the last response is received</dd>
246 <dt>Complete requests</dt>
247 <dd>The number of successful responses received</dd>
249 <dt>Failed requests</dt>
250 <dd>The number of requests that were considered a failure. If the
251 number is greater than zero, another line will be printed showing the
252 number of requests that failed due to connecting, reading, incorrect
253 content length, or exceptions.</dd>
255 <dt>Write errors</dt>
256 <dd>The number of errors that failed during write (broken pipe).</dd>
258 <dt>Non-2xx responses</dt>
259 <dd>The number of responses that were not in the 200 series of response
260 codes. If all responses were 200, this field is not printed.</dd>
262 <dt>Keep-Alive requests</dt>
263 <dd>The number of connections that resulted in Keep-Alive requests</dd>
265 <dt>Total body sent</dt>
266 <dd>If configured to send data as part of the test, this is the total
267 number of bytes sent during the tests. This field is omitted if the test
268 did not include a body to send.</dd>
270 <dt>Total transferred</dt>
271 <dd>The total number of bytes received from the server. This number
272 is essentially the number of bytes sent over the wire.</dd>
274 <dt>HTML transferred</dt>
275 <dd>The total number of document bytes received from the server. This
276 number excludes bytes received in HTTP headers</dd>
278 <dt>Requests per second</dt>
279 <dd>This is the number of requests per second. This value is the result
280 of dividing the number of requests by the total time taken</dd>
282 <dt>Time per request</dt>
283 <dd>The average time spent per request. The first value is calculated
284 with the formula <code>concurrency * timetaken * 1000 / done</code>
285 while the second value is calculated with the formula
286 <code>timetaken * 1000 / done</code></dd>
288 <dt>Transfer rate</dt>
289 <dd>The rate of transfer as calculated by the formula
290 <code>totalread / 1024 / timetaken</code></dd>
294 <section id="bugs"><title>Bugs</title>
295 <p>There are various statically declared buffers of fixed length. Combined
296 with the lazy parsing of the command line arguments, the response headers
297 from the server and other external inputs, this might bite you.</p>
299 <p>It does not implement HTTP/1.x fully; only accepts some 'expected' forms
300 of responses. The rather heavy use of <code>strstr(3)</code> shows up top
301 in profile, which might indicate a performance problem; <em>i.e.</em>, you
302 would measure the <code>ab</code> performance rather than the server's.</p>