http_load

Emulate 20 requests based on a given URL list file per second for 60 seconds

$ http_load -rate [20] -seconds [60] [path/to/urls.txt]

$ http_load -parallel [5] -seconds [60] [path/to/urls.txt]

$ http_load -rate [20] -fetches [1000] [path/to/urls.txt]

$ http_load -parallel [5] -fetches [1000] [path/to/urls.txt]

http_load [-parallel N] [-fetches N] [-rate R] [-seconds S] [-timeout T] [-throttle B] [-verbose] [-version] [-help] [urlfile]

-parallel N or -p N
    Number of parallel connections (default: 1)

-fetches N or -f N
    Total fetches to perform (default: infinite)

-rate R or -r R
    Limit to R requests per second

-seconds S or -s S
    Run for S seconds (default: infinite)

-timeout T or -t T
    Timeout per request in seconds (default: 0, no timeout)

-throttle B
    Throttle total bandwidth to B bytes/second

-verbose or -v
    Enable verbose output

-version or -V
    Display version information

-help or -h
    Show usage help

http_load is a lightweight, multi-threaded utility for benchmarking web server performance under concurrent load. It reads a list of URLs from a file or standard input and simulates parallel HTTP requests, allowing precise control over parallelism, request rate, total fetches, runtime, and timeouts.

Designed for stress testing, capacity planning, and tuning, it generates realistic loads using persistent connections and pipelining where possible. Upon completion, it reports detailed statistics including: requests completed, average/median/min/max latency, throughput (requests/second), average bytes/request, total data transferred, and error rates.

URLs in the input file can include repetition prefixes like 10+http://example.com/ to repeat fetches. The tool is single-process but forks worker processes for parallelism, making it efficient on multi-core systems. Written in C, it's portable, dependency-free, and ideal for scripting automated tests. While powerful for basic HTTP loads, it lacks support for advanced features like HTTPS, POST data, or cookies.

Supports HTTP only (no HTTPS). Older tool; limited modern HTTP/2 or auth support. Install via package (http-load) or source. URL file format required; stdin if omitted.

One URL per line. Prefix with N+ for repetition, e.g., 5+https://example.com/path fetches 5 times.

echo 'http://example.com/' | http_load -parallel 10 -fetches 1000 -rate 200 -seconds 30
Generates 10 parallel connections at 200 req/sec for 30s.

Developed by Stu Card at Acme Laboratories around 1996-1998. Open-sourced early, remains popular for its simplicity and effectiveness in Unix environments despite age.

ab(1), siege(1), wrk(1), httperf(1)