tylerjl/benchmarking-proxies

Benchmarking wishlist

Opened this issue · 2 comments

tylerjl commented

I've had lots of feedback since publishing my findings so I'm collecting all of those suggestions here for later reference:

  • Benchmarks with logging enabled
  • Varying quantities of response headers
  • Dramatically boost the number of nginx connection workers
  • Reduce potential noise in results through use of bare-metal systems under test
centminmod commented

What about

  • HTTP gzip compressed tests instead of non-compressed?
  • Also testing their HTTPS handling i.e. HTTP/2 for various SSL configs i.e. RSA 2048bit vs ECDSA 256bit SSL certificate/ciphers?
  • ❤️3
git001 commented

Thank you for your excellent Benchmark.
May I add HAProxy to the whitelist as HAProxy is by default and in his "DNA" a reverse proxy 😄.

Note: I'm a long term HAProxy contributor and nginx user so maybe I'm biased 🤔

  • 👀1