From b79cc8f9b5dc83b45db1f82933eaa55d7546351c Mon Sep 17 00:00:00 2001 From: Paul Buetow Date: Fri, 20 Apr 2012 21:48:04 +0200 Subject: initial packaging working --- docs/loadbars.txt | 65 +++++++++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 65 insertions(+) create mode 100644 docs/loadbars.txt (limited to 'docs/loadbars.txt') diff --git a/docs/loadbars.txt b/docs/loadbars.txt new file mode 100644 index 0000000..1cb9764 --- /dev/null +++ b/docs/loadbars.txt @@ -0,0 +1,65 @@ +NAME + httpbench - Simple HTTP benchmarking tool + +SYNOPSIS + httpbench -u -d sec -c concurrent -r rps [-t ms] + [-e expected] + +DESCRIPTION + HTTP Bench is a small program for Linux, UNIX and a like written in C + which can be used to benchmark HTTP servers. For more infos please + consult the project website. + + Arguments + Mandatory + "-u url or urllist.txt" + An URL to benchmark if argument begins with http:// or + https:// or a path to a text file containing a list of all + HTTP-URLs to request during a benchmark. Each line should + start with http:// or https://. + + "-d seconds" + Duration of the benchmark in seconds. E.g. 300 for a five + minutes benchmark. + + "-c concurrent" + Number of cuncurrent threads used by the benchmark. + + "-r rps" Requests per second wanted for the benchmark. + + Optional + "-t timeoutms" + Timeout in milliseconds for each request. This timeout will + actually not interrupt any requests. It's there just for + statistical reasons. E.g. how many requests take longer than + x milliseconds? + + "-e expectedresponsestr" + The expected response of a request. E.g. all requests should + contain HTML in its reponse. All responses not matching are + counted and reported. + + EXAMPLES + "httpbench -u mylist.txt -d 60 -c 10 -r 1000" + Iterates over each url in mylist.txt with 10 cuncurrent + threads for 60 seconds and tries to reach 1000 requests per + second. Each curl failure will be reported at the end of the + test (total count and total %). + + "httpbench -u http://www.example.com -d 300 -c 1000 -r 10000 -t 20 -e + "Hello World"" + Benchmarks the single URL http://www.example.com with 1000 + cuncurrent threads for 5 minutes and tries to reach 10000 + requests per second. Each request which takes longer than 20 + milliseconds will be reported at the end of the test (total + count and total %) and each request which does not contain + "Hello World" in its response string will also be reported + (total count and total %). The curl errors will also be + reported. + +LICENSE + See package description or project website. + +AUTHOR + Paul Buetow - + -- cgit v1.2.3