wrk

[bertt]

use wrk/wrk instead of ab for benchmark testing?

https://github.com/wg/wrk

sample:

$ wrk -t12 -c400 -d30s http://127.0.0.1:8080/v1.0
Running 30s test @ http://127.0.0.1:8080/v1.0
  12 threads and 400 connections
  Thread Stats   Avg      Stdev     Max   +/- Stdev
    Latency    91.28ms   42.27ms 620.09ms   76.34%
    Req/Sec   358.49     95.67     1.42k    74.44%
  126974 requests in 30.09s, 123.15MB read
  Socket errors: connect 0, read 5, write 0, timeout 0
Requests/sec:   4220.31

pro:

• also used by https://www.techempower.com/benchmarks/
• seems more stable results at first test

con:

• building on linux takes forever (https://github.com/wg/wrk/wiki/Installing-Wrk-on-Linux)

todo:

• figure out how to do posts (something with a lua file) https://github.com/wg/wrk/blob/master/scripts/post.lua

[bertt]

reference: 2107 request/s
other test: 4089 request/s

[bertt]

running wrk in docker:

$ docker run --rm --net=host williamyeh/wrk -t12 -c400 -d30s http://127.0.0.1:8080/v1.0

[bertt]

Posting sample:

$ wrk -t12 -c400 -d30s -s post.lua http://127.0.0.1:8080/v1.0/Locations

Met post.lua file:

wrk.method = "POST"
wrk.body   = '{"name": "Somewhere","description": "my backyard","encodingType": "application/vnd.geo+json","location": {"type": "Point","coordinates": [-117.123,54.123]}}'
wrk.headers["Content-Type"] = "application/json"

Result on my machine: Requests/sec: 1168.43