Skip to content

Instantly share code, notes, and snippets.

@christiannelson

christiannelson/benchmark.md

Last active August 29, 2015 14:01
Show Gist options
  • Star 0 You must be signed in to star a gist
  • Fork 0 You must be signed in to fork a gist
  • Save christiannelson/c5c8ca51b3fa07780d83 to your computer and use it in GitHub Desktop.
Save christiannelson/c5c8ca51b3fa07780d83 to your computer and use it in GitHub Desktop.
Download ZIP
Comparing "hello world" in node.js under dokku versus bare metal
Raw
benchmark.md

Dokku vs. Bare Metal

Running on an Aleutia R50 Mini-server with an Intel i3 1.7 Ghz, 16 GB and SSD.

Dokku (docker fronted by nginx)

Running 10s test @ http://node-js-sample.kuju0.carbonfive.com/
  2 threads and 10 connections
  Thread Stats   Avg      Stdev     Max   +/- Stdev
    Latency     6.57ms    2.04ms  26.01ms   81.91%
    Req/Sec   798.54    132.61     1.10k    72.46%
  15433 requests in 10.00s, 3.16MB read
Requests/sec:   1543.32
Transfer/sec:    323.96KB

Bare Metal

Running 10s test @ http://kuju0:5000
  2 threads and 10 connections
  Thread Stats   Avg      Stdev     Max   +/- Stdev
    Latency     4.23ms    1.66ms  27.46ms   86.13%
    Req/Sec     1.23k   229.90     1.82k    64.22%
  23564 requests in 10.00s, 4.36MB read
Requests/sec:   2356.40
Transfer/sec:    446.43KB

Notes

I'm impressed that dokku doesn't add more overhead. Also, I suspect much of the overhead is the result of fronting node with ngnix. nginx is fast, but for such a simple app, the tinsy bit of added latency becomes significant. In a real-world app, it would be neglicable compared to servicing the request.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment