Skip to content

Benchmark-pr-reporting#10

Open
MichaelDecent wants to merge 2 commits into
tigrbl:masterfrom
MichaelDecent:benchmark-pr-reporting
Open

Benchmark-pr-reporting#10
MichaelDecent wants to merge 2 commits into
tigrbl:masterfrom
MichaelDecent:benchmark-pr-reporting

Conversation

@MichaelDecent
Copy link
Copy Markdown
Contributor

  • Introduced a new workflow to run benchmarks on pull requests.
  • Added a script to format benchmark results into a GitHub PR comment.
  • Updated .gitignore to exclude benchmark artifacts directory.

- Introduced a shuffle option in the performance matrix runner to randomize profile execution order.
- Added a seed parameter for reproducibility of the shuffle.
- Updated PerfRunSummary to include shuffle metadata.
- Implemented tests to verify shuffle behavior and seed consistency.
- Enhanced command-line interface to support new shuffle and seed arguments.
- Introduced a new workflow to run benchmarks on pull requests.
- Added a script to format benchmark results into a GitHub PR comment.
- Updated .gitignore to exclude benchmark artifacts directory.
@MichaelDecent
Copy link
Copy Markdown
Contributor Author

Benchmark Results

Status: ❌ 0/32 passed, 32 failed | Commit: bcc4b6b5
Platform: darwin-arm64-cpython3.14 | Shuffle seed: 4269797450

Failures

Profile Reasons
reload_overhead reload_overhead failed threshold max_rss_kib: observed=49086464.0 threshold=52906.0
logging_on logging_on failed threshold max_rss_kib: observed=49135616.0 threshold=52906.0
logging_off logging_off failed threshold max_rss_kib: observed=49135616.0 threshold=52906.0
http2_multiplex_100 http2_multiplex_100 failed threshold max_rss_kib: observed=49135616.0 threshold=52486.0
ws_http11_permessage_deflate ws_http11_permessage_deflate failed threshold max_rss_kib: observed=49168384.0 threshold=52906.0
ws_http2_permessage_deflate ws_http2_permessage_deflate failed threshold max_rss_kib: observed=49168384.0 threshold=52906.0
http3_loss_jitter http3_loss_jitter failed threshold max_rss_kib: observed=49233920.0 threshold=52486.0
content_coding_under_load content_coding_under_load failed threshold max_rss_kib: observed=49233920.0 threshold=52906.0
ws_http2 ws_http2 failed threshold max_rss_kib: observed=49250304.0 threshold=52906.0
tls_handshake tls_handshake failed threshold max_rss_kib: observed=49250304.0 threshold=52906.0
ws_http11 ws_http11 failed threshold max_rss_kib: observed=49250304.0 threshold=52486.0
http2_multiplex_10 http2_multiplex_10 failed threshold max_rss_kib: observed=49266688.0 threshold=52486.0
ws_http3_permessage_deflate ws_http3_permessage_deflate failed threshold max_rss_kib: observed=49266688.0 threshold=52906.0
alpn_negotiation_cost alpn_negotiation_cost failed threshold max_rss_kib: observed=49283072.0 threshold=52906.0
worker_scaleout worker_scaleout failed threshold max_rss_kib: observed=49283072.0 threshold=52906.0
trailers_under_load trailers_under_load failed threshold max_rss_kib: observed=49315840.0 threshold=52906.0
metrics_on metrics_on failed threshold max_rss_kib: observed=49332224.0 threshold=52906.0
http3_clean_network http3_clean_network failed threshold max_rss_kib: observed=49332224.0 threshold=52486.0
graceful_drain graceful_drain failed threshold max_rss_kib: observed=49348608.0 threshold=52906.0
http2_multiplex_1000 http2_multiplex_1000 failed threshold max_rss_kib: observed=49364992.0 threshold=52486.0
ocsp_strict_mode ocsp_strict_mode failed threshold max_rss_kib: observed=49364992.0 threshold=52906.0
ws_fanout_broadcast ws_fanout_broadcast failed threshold max_rss_kib: observed=49364992.0 threshold=52906.0
http2_tls http2_tls failed threshold max_rss_kib: observed=49364992.0 threshold=52486.0
mtls_handshake mtls_handshake failed threshold max_rss_kib: observed=49381376.0 threshold=52906.0
proxy_headers_on proxy_headers_on failed threshold max_rss_kib: observed=49381376.0 threshold=52906.0
http11_chunked_upload_download http11_chunked_upload_download failed threshold max_rss_kib: observed=49381376.0 threshold=52486.0
metrics_off metrics_off failed threshold max_rss_kib: observed=49397760.0 threshold=52906.0
connect_relay_throughput connect_relay_throughput failed threshold max_rss_kib: observed=49397760.0 threshold=52906.0
ws_http3 ws_http3 failed threshold max_rss_kib: observed=49397760.0 threshold=52906.0
proxy_headers_off proxy_headers_off failed threshold max_rss_kib: observed=49397760.0 threshold=52906.0
http11_baseline http11_baseline failed threshold max_rss_kib: observed=49397760.0 threshold=52486.0
http11_keepalive http11_keepalive failed threshold max_rss_kib: observed=49414144.0 threshold=52486.0

Results

component_regression (9 profiles)
Profile Status Throughput (ops/s) p99 (ms) p99.9 (ms) Error Rate
reload_overhead 5,438.68 0.200 0.201 0.000
logging_on 15,265.24 0.145 0.155 0.000
logging_off 1,014,790.43 0.000 0.000 0.000
worker_scaleout 811,276.46 0.005 0.005 0.000
metrics_on 69,858.75 0.015 0.015 0.000
graceful_drain 983,767.45 0.004 0.004 0.000
proxy_headers_on 130,548.30 0.008 0.008 0.000
metrics_off 1,044,603.90 0.000 0.000 0.000
proxy_headers_off 611,620.93 0.001 0.001 0.000
end_to_end_release (23 profiles)
Profile Status Throughput (ops/s) p99 (ms) p99.9 (ms) Error Rate
http2_multiplex_100 103,283.11 1.059 1.063 0.000
ws_http11_permessage_deflate 64,606.49 0.016 0.016 0.000
ws_http2_permessage_deflate 19,921.62 0.054 0.054 0.000
http3_loss_jitter 35,634.74 0.037 0.038 0.000
content_coding_under_load 70,288.64 0.018 0.018 0.000
ws_http2 23,850.92 0.042 0.042 0.000
tls_handshake 24,144.85 0.047 0.047 0.000
ws_http11 192,351.04 0.005 0.005 0.000
http2_multiplex_10 104,136.51 0.101 0.101 0.000
ws_http3_permessage_deflate 16,255.02 0.068 0.069 0.000
alpn_negotiation_cost 572,333.53 0.001 0.001 0.000
trailers_under_load 158,730.16 0.007 0.007 0.000
http3_clean_network 17,734.86 0.060 0.061 0.000
http2_multiplex_1000 88,796.37 11.447 11.455 0.000
ocsp_strict_mode 16,465.14 0.092 0.094 0.000
ws_fanout_broadcast 1,744,921.60 0.018 0.018 0.000
http2_tls 11,375.31 0.126 0.129 0.000
mtls_handshake 18,337.39 0.057 0.057 0.000
http11_chunked_upload_download 164,665.98 0.006 0.006 0.000
connect_relay_throughput 350,365.03 0.003 0.003 0.000
ws_http3 18,886.06 0.056 0.056 0.000
http11_baseline 175,503.26 0.006 0.006 0.000
http11_keepalive 159,866.98 0.006 0.006 0.000

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

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant