[szibis]

Summary

The scheduled daily performance run (perf-daily.yaml) measured main every day: a 4-hour A/B that then updated or opened the daily report pull request, even when nothing it measures had changed since the previous day. A scheduled run now skips, with no build, no run, and no new or updated pull request, when nothing it measures has changed on main since the commit the last full daily measurement used. While a daily pull request is open, its unmerged days count. A manual dispatch always runs.

Changes

• bench/ab/history.py last-commit prints the commit that the newest full measurement of every current set used.

  • Only the sets in bench/ab/shapes.json count, so a renamed or removed set's old history file cannot disable the skip.
  • A line counts only if it was a full daily measurement: the daily run count (7, now one constant shared by daily.py) and every range of the set up to 24h. A cut-down manual run therefore does not count as a measurement.
  • If any current set has no full line, it prints nothing and the run goes ahead.

• perf-daily.yaml: after the existing carry-history step, a scheduled run compares that commit with HEAD over what a day measures:

  • cmd/, internal/, pkg/, go.mod, go.sum;
  • bench/ab/ without its results, history, tests and README;
  • the stack's compose file, the Loki config and the log generator;
  • the workflow itself.

  Go test files (**/*_test.go) are excluded. If nothing changed, the build, run, report and pull-request steps are skipped, and the job summary says why. A history read error falls back to running. Go setup moves after the check (still 1.27.3, cache: false), so a skipped day does not install Go.

• bench/ab/README.md, and a CHANGELOG entry under CI.

Proofs

• Gate logic against real commits:
  • the 2026-10-21 run's commit 5c6dfed3 compared with a main that has only report and metadata changes since: skip;
  • compared with main after a proxy change: run;
  • no history, or a commit not in the repository: run;
  • a test-only Go commit: skip (git diff --quiet returns 0 with the exclude, 1 without);
  • a production change: run.
• Unit tests: python3 -m unittest discover -s bench/ab/tests passes 100 tests. New cases cover agreeing sets, a stale set file, too few runs, too short a range, a set never fully measured, and lines without a commit.
• Gates: scripts/ci/tests: 77 pass. actionlint (pinned and local), zizmor 1.30.1 (pinned, no findings), and the conformance and changelog gates.

Not applicable: no proxy code changes, so e2e vs Loki, the Grafana range matrix, visual proof, A/B, benchstat and the load test do not apply. The workflow's own pull-request validation job runs on this PR.

[github-actions]

⚪ Performance A/B: no change beyond noise

8 shape×range — 8 ⚪ same · results vs Loki: 7 same, 0 new difference(s), 1 pre-existing

Every shape is within noise of the base build (±30% and 100 ms).

8 unchanged shape×range (within noise)

	shape	range	base p50	PR p50	change	Loki p50	PR ÷ Loki	result vs Loki
⚪	drilldown service landing	1h	33 ms	30 ms	⚪ -9%	26 ms	1.15×	✅ same
⚪	instant ungrouped rate	instant	10 ms	7 ms	⚪ -30%	78 ms	0.09×	✅ same
⚪	json rate by level (no filter, no drop)	1h	400	400	⚪ +0%	400	—	✅ same
⚪	json volume no filters	1h	194 ms	201 ms	⚪ +4%	52 ms	3.87×	✅ same
⚪	logs plain	1h	45 ms	48 ms	⚪ +7%	59 ms	0.81×	✅ same
⚪	logs | json	1h	78 ms	81 ms	⚪ +4%	77 ms	1.05×	✅ same
⚪	stream-label count by app	1h	26 ms	27 ms	⚪ +4%	40 ms	0.67×	✅ same
⚪	topk rate by app	1h	44 ms	44 ms	⚪ +0%	97 ms	0.45×	➖ differs (as base)

Cold (first-run) timings — what a user opening a panel sees

shape	range	base cold	PR cold	Loki cold	PR ÷ Loki cold
drilldown service landing	1h	31 ms	29 ms	897 ms	0.03×
instant ungrouped rate	instant	7 ms	8 ms	80 ms	0.10×
json rate by level (no filter, no drop)	1h	400	400	400	—
json volume no filters	1h	187 ms	188 ms	1.13 s	0.17×
logs plain	1h	90 ms	45 ms	135 ms	0.33×
logs | json	1h	126 ms	142 ms	63 ms	2.25×
stream-label count by app	1h	28 ms	26 ms	769 ms	0.03×
topk rate by app	1h	44 ms	51 ms	840 ms	0.06×

Why these shapes

• control smoke → 8: control/drilldown service landing, control/instant ungrouped rate, control/json rate by level (no filter, no drop), control/json volume no filters, control/logs plain, control/logs | json, control/stream-label count by app, control/topk rate by app

_{base e069c371db → PR 889bcdd5bf · 4 runs (first is cold), warm p50 · ranges 1h · noise ±30% and 100 ms · VictoriaLogs restarts 0→0 · 2.2 min. Same runner, one fresh stack, targets interleaved per request. How to read and reproduce: bench/ab/README.md.}

[github-actions]

PR Quality Report

Compared against base branch main.

Coverage and tests

Signal	Base	PR	Delta
Test count	11763	11826	63
Coverage	88.3%	88.3%	0.0% (stable)

Compatibility

Track	Base	PR	Delta
Loki API	100.0%	11/11 (100.0%)	0.0% (stable)
Logs Drilldown	100.0%	17/17 (100.0%)	0.0% (stable)
VictoriaLogs	100.0%	11/11 (100.0%)	0.0% (stable)

Performance smoke

Performance smoke was skipped for this PR because no perf-sensitive paths changed.

State

• Coverage, compatibility, and sampled performance are reported here from the same PR workflow.
• This is a delta report, not a release gate by itself. Required checks still decide merge safety.
• Performance smoke was intentionally skipped because no perf-sensitive paths changed in this PR.
• Delta states use the same noise guards as the quality gate (percent + absolute + low-baseline checks), so report labels match merge-gate behavior.

[github-actions]

✅ Visual smoke: passed

Base e069c371db vs PR 6012e2bcd0, Grafana through the proxy builds and Loki on one fresh stack; run.
Core set: 5 capture(s) at 1h. Detailed set: 0 entries (nothing visual touched). 5 capture(s), 15m/1h/6h, 24h and 7d stay out of PR runs.

page	range	set	base = PR (data)	vs Loki (PR)	pixel diff	result
dd-fields	1h	core	✅ identical	identical (4 explained, 1 history-dependent)	0.11%	✅
dd-landing	1h	core	✅ identical	identical (3 explained)	0.14%	✅
dd-service-logs	1h	core	✅ identical	identical (3 explained, 1 history-dependent)	0.11%	✅
explore-logs-plain	1h	core	✅ identical	identical (1 explained)	0.07%	✅
explore-metric-plain	1h	core	✅ identical	identical (1 explained)	0.04%	✅

Gate: a base-vs-PR data difference passes when the PR is closer to Loki, and fails when the PR diverges from Loki or Loki cannot decide it (at a range Loki does not hold, a difference passes when the same page proved it closer to Loki at a shorter range; otherwise label the pull request visual-change-expected to accept an intended change). Also failing: a panel empty on the PR but not on the base, an error banner, panel error, error answer or new "No data" panel on the PR, a PR side that never settled, a difference that did not reproduce on the recapture. Pixel differences only warn.

Core set montages (5): base | PR | Loki

dd-fields-1h

dd-landing-1h

dd-service-logs-1h

explore-logs-plain-1h

explore-metric-plain-1h

All montages of this run: pr-visuals/pr-796.