http_load
HTTP benchmarking tool that runs multiple HTTP fetches in parallel
TLDR
Test with 20 requests/second for 60 seconds
$ http_load -rate 20 -seconds 60 urls.txt
Test with 5 concurrent connections for 60 seconds$ http_load -parallel 5 -seconds 60 urls.txt
Run 1000 requests at 20/second$ http_load -rate 20 -fetches 1000 urls.txt
Run 1000 requests with 5 concurrent connections$ http_load -parallel 5 -fetches 1000 urls.txt
Test single URL directly$ echo "http://example.com/" | http_load -parallel 5 -seconds 30 /dev/stdin
SYNOPSIS
http_load [options] urlfile_
DESCRIPTION
http_load is an HTTP benchmarking tool that runs multiple HTTP fetches in parallel to test the throughput and response time of a web server. It reads URLs from a file and generates load according to specified parameters.
PARAMETERS
-rate N
Generate N requests per second-parallel N
Keep N connections active simultaneously-fetches N
Total number of requests to perform-seconds N
Duration of the test in seconds-timeout N
Timeout for each request in seconds-sip
Source IP addresses (one per line in file)-cipher CIPHER
SSL cipher to use-proxy HOST:PORT
Use HTTP proxy
CAVEATS
Either -rate or -parallel must be specified, not both. Either -fetches or -seconds must be specified to set the test duration. The URL file should contain one URL per line.
HISTORY
http_load was developed by ACME Labs as a simple HTTP load testing tool for benchmarking web server performance.