siege - Man Page

siege is a HTTP/FTP load tester and benchmarking utility.

Examples (TL;DR)


  siege [options]
  siege [options] <URL>
  siege [options] -g <URL>
  siege [options] -f urls.txt


siege is a multi-threaded HTTP/FTP load tester and benchmarking  utility. It supports most of the features detailed in RFCs 2616 (HTTP) and 959 (FTP). Properties can be set both from the command line and in a configuration file. When the same property is set in both locations, the command line takes precedence.

The default configuration file is $HOME/.siege/siege.conf If you don't  have a $HOME/.siege directory and a siege.conf and cookies.txt file,  siege will generate a new config directory when it runs. You can  generate your config directory with the following command: siege.config


Option Syntax

siege supports long and short options. Short options look like this:
   -c 25

Long options look like this:

Option Values

-V,  --version

Displays the siege release version and copyright information.

-h,  --help

Prints a help message describing siege's command-line options.

-C,  --config

Prints a detailed summary of all the currently configured options, most of which are sent in $HOME/.siege/siege.conf

-v,  --verbose

This directive puts siege into verbose mode which is actually a  default setting. This command-line option is useful when the config  file is set to 'verbose = false' since it will allow you to override  that.

By default siege's verbose output is displayed in a color-coded  style.
 * HTTP 2xx is coded blue
 * HTTP 3xx is coded cyan
 * HTTP 4xx is coded magenta
 * HTTP 5xx is coded red
 * HTTP cached is coded black

NOTE: You can turn off color in siege.conf like this: 'color = off'

-q,  --quiet

This directive silences siege. It is mostly used for scripting and  is often used in conjunction with -g/--get. You can detect the success  or failure of the run with its exit code.

  siege --quiet -g
  if [ $? -eq 0 ] ; then
    echo "Success"
    echo "Failure"
-g URL, --get=URL

This option allows you to request a URL and watch the header  transaction.  There is a corresponding config file directive that  allows you to set the request method for these requests:  gmethod = HEAD|GET

  $ siege -g ""
  HEAD / HTTP/1.0
  Accept: */*
  User-Agent: Mozilla/5.0 (unknown-x86_64-linux-gnu) Siege/4.0.0-beta5
  Connection: close

  HTTP/1.1 200 OK
  Server: cloudflare-nginx
  Date: Tue, 09 Feb 2016 18:18:41 GMT
  Content-Type: text/html; charset=UTF-8
  Connection: close
  Last-Modified: Wed, 25 Nov 2015 18:46:08 GMT
  Cache-Control: max-age=3, must-revalidate
  Expires: Tue, 09 Feb 2016 18:18:44 GMT
  Vary: Accept-Encoding,Cookie
  CF-RAY: 27219407eeff084a-IAD

NOTE: It's a best practice to quote the URL when it's passed to siege  from the the command-line.

-p URL, --print=URL

This option is similar to -g / --get but it PRINTS the page it received  from the server.

  $ siege -p
  GET / HTTP/1.0
  Accept: */*
  User-Agent: Mozilla/5.0 (unknown-x86_64-linux-gnu) Siege/4.0.3rc1
  Connection: close

  HTTP/1.1 301 Moved Permanently
  Date: Wed, 19 Oct 2016 16:58:13 GMT
  Content-Type: text/html; charset=iso-8859-1
  Server: cloudflare-nginx
  Connection: close

  <title>301 Moved Permanently</title>
  <h1>Moved Permanently</h1>
  <p>The document has moved <a href="">here</a>.</p>
  <address>Apache/2.2.31 (Amazon) Server at Port 80</address>
-c NUM, --concurrent=NUM

This option allows you to set the concurrent number of users. The total  number of users is technically limited to your computer's resources.

You should not configure more users than your web server is configured  to handle. For example, the default apache configuration is capped at  255 threads. If you run siege with -c 1024, then 769 siege users are  left waiting for an apache handler.

For this reason, the default siege configuration is capped at 255 users. You can increase that number inside siege.conf but if you make a mess,  then please don't complain to us.

-r NUM, --reps=NUM|once

This option tells each siege user how many times it should run. The value  should generally be a number greater than zero but it may be the keyword 'once'.

If --reps=3 then each siege user will run three times before it exits.  However, if --reps=once, then each user will run through the urls.txt  file exactly one time.

For more information about the urls.txt file, see option -f <file>,  --file=<file>

-t NUMm, --time=NUMm

This option is similar to --reps but instead of specifying the number  of times each user should run, it specifies the amount of time each  should run.

The value format is “NUMm”, where “NUM” is an amount of time and the “m” modifier is either S, M, or H for seconds, minutes and hours. To run  siege for an hour, you could select any one of the following  combinations: -t3600S, -t60M, -t1H.  The modifier is not case sensitive,  but it does require no space between the number and itself.

-d NUM, --delay=NUM

This option instructs siege how long to delay between each page  request.  The value NUM represents the number of seconds between each one. This number can be a decimal value. In fact the default is half a second (--delay=0.5).

The time between delay requests is NOT applied toward the transaction  time. If two 0.1 second transactions have a 2 second delay between them, their average transaction time is 0.1 seconds. It is applied toward the total elapsed time. In this scenario, the elapsed time would be 2.2  seconds.

NOTE: when the parser is enabled (see: -p/--parser), there is no delay between the page and its elements, i.e., style sheets, javascripts, etc.  The delay is only between page requests.

-b,  --benchmark

This directive tells siege to go into benchmark mode. This means there  is no delay between iterations.

-i,  --internet

This option sets siege into what we call internet mode. It makes  requests from the urls.txt file (see: -f <file> / --file=<file>) in  random order.

-f FILE, --file=FILE

This option tells siege to work with a list of urls inside a text  file. The URLs are listed one per line. Unlike URLs that are passed as a command-line argument, the URLs in this file should not be quoted.

siege's urls.txt parser supports comments and variables. Since  siege uses the dollar sign ($) as a prefix for scalar variables, you should escape any variable you want to send to the server:

  https://$(HOST)/siege/jsoner.php POST {"price": "\$10 per mile"}
-R FILE, --rc=FILE

This directive allows you to set an alternative resource file. By  default, the siegerc file is $HOME/.siege/siege.conf With this  directive, you can override the default and use an alternative file.

-L FILE, --log=FILE

The default log file is $prefix/var/log/siege.log. This directive  allows you to specify an alternative file for logging.

-m “string”, --mark=“string”

This option allows you to log a message to the log file before your  stats are written there. It is generally used to identify the  proceeding run. You could, for example, mark the file with your command-line parameters so it's understood what configuration  generated the following data.

-H “header: value”, --header=“Header: value”

This options allows you to set a custom header in the request.  Generally speaking, this header will override an existing header. The Cookie header is a special case.  If you set -H “Cookie: value” then  siege will send that cookie in addition to the other ones.

-A “string”, --agent=“string”

This option allows you to override the default user-agent with a custom  one.

  siege --agent="JoeDog Jr. in da hizzle"

Will set this header:

  User-agent: JoeDog Jr. in da hizzle

Alternatively, you could set the User-agent with the -H/--header option  above.

-T “text”, --content-type=“text”

This is another set header shortcut. You use this option to override  the default Content-type request header.


Turn off the HTML parser. When siege downloads a page, it parses it for additional page elements such as style-sheets, javascript and images. It  will make additional requests for any elements it finds. With this option enabled, siege will stop after it pulls down the main page.


This directive instructs siege not to follow 3xx redirects.

URL Format

siege supports RFC 1738 URL formats but it takes pains to implement commonly used shortcuts for your convenience. In addition to RFC 1738  formats, siege introduces its own URL format to indicate protocol method.

An RFC 1738 URL looks like this:

A siege URL with a method indicator looks like this:
 <scheme>://<username>:<password>@<hostname>:<port>/<path> POST <query>

You can also post the contents of a file using the redirect character like this:
 <scheme>://<username>:<password>@<hostname>:<port>/<path> POST </home/jeff/haha.txt

Here are two examples with the siege method indicator: POST haha=papa&dada=mama PUT </home/jdfulmer/etc/tests/bbc.jpg

NOTE: If you set URLs with method indicators at the command-line, then  you MUST quote the thing or your shell will treat it like three separate arguments. If the URL is in a urls.txt file, then you shouldn't quote it.

As mentioned above, siege goes to great lengths to allow commonly  used shortcuts that you're used to from most browser implementations.  It treats many parts of the 1738 URL as optional. In this example, the  parts in brackets are optional:
 [scheme://] [:port] [/path/file]

When siege receives a host name it builds the URL with default  assumptions. becomes


From the section called Option Syntax above we learn that siege  can take a URL as an argument. siege -c -r2 will  request the JoeDog index page twice.  But what if you want to hit large portions of the site? siege will allow you to fill a file  with URLs so that it can run through list.

The format for the file is one URL per line: POST homer=simpson&marge=doestoo

The file also supports UNIX-style commenting:
 # Comment looks like this POST homer=simpson&marge=doestoo

It supports shell-style variable declaration and references. This is  convenient if you want to run the same test on two different tiers or two different schemes:

  $(SCHEME)://$(HOST)/haha/ POST homer=simpson&marge=doestoo

You can tell siege about this file with the -f/--file option:
 siege -c1 -r50 -f /home/jeff/urls.txt

Performance Statistics

When its run is complete, siege will gather performance data from all its clients and summarize them after the run. (You can also choose to log these numbers). The command-line output is modeled after Lincoln  Stein's script:

  Transactions:                   2000 hits
  Availability:                 100.00 %
  Elapsed time:                  58.57 secs
  Data transferred:               5.75 MB
  Response time:                  0.25 secs
  Transaction rate:              34.15 trans/sec
  Throughput:                     0.10 MB/sec
  Concurrency:                    8.45
  Successful transactions:        2000
  Failed transactions:               0
  Longest transaction:            4.62
  Shortest transaction:           0.00

      This number represents the total number of HTTP requests. In this
      example, we ran 25 simulated users [-c25] and each ran ten times
      [-r10]. Twenty-five times ten equals 250 so why is the transaction 
      total 2000? That's because siege counts every request. This run 
      included a META redirect, a 301 redirect and the page it requested 
      contained several elements that were also downloaded.

      This is the percentage of socket connections successfully handled 
      by the server. It is the result of socket failures (including 
      timeouts) divided by the sum of all connection attempts. This 
      number does not include 400 and 500 level server errors which are
      recorded in "Failed transactions" described below.  

  Elapsed time
      The duration of the entire siege test. This is measured from the 
      time the user invokes siege until the last simulated user 
      completes its transactions. Shown above, the test took 58.57 
      seconds to complete.

  Data transferred
      The sum of data transferred to every siege simulated user. It 
      includes the header information as well as content. Because it
      includes header information, the number reported by siege will 
      be larger then the number reported by the server. In internet 
      mode, which hits random URLs in a configuration file, this 
      number is expected to vary from run to run.

  Response time
      The average time it took to respond to each simulated user's requests.

  Transaction rate
      The average number of transactions the server was able to handle
      per second, in a nutshell: it is the count of all transactions 
      divided by elapsed time.

      The average number of bytes transferred every second from the 
      server to all the simulated users.

      This is the average number of simultaneous connections. The metric
      is calculated like this: the sum of all transaction times divided
      by elapsed time (how long siege ran) 

  Successful transactions
      The number of times the server responded with a return code < 400.

  Failed transactions
      The number of times the socket transactions failed which includes 
      socket timeouts.

  Longest transaction
      The greatest amount of time that any single transaction took, out 
      of all transactions.

  Shortest transaction
      The smallest amount of time that any single transaction took, out
      of all transactions.


Jeffrey Fulmer, et al. <> is the primary author of siege. Numerous people  throughout the globe also contributed to this program. Their  contributions are noted in the source code ChangeLog


The most recent released version of siege is available by HTTP  download:

See Also

siege.config(1) bombardment(1) siege2csv(1)

Referenced By


2023-01-05 JoeDog Siege Load Tester