Skip to content

ci(performance): restore the 1.10 cold limit of the fresh ratio - #1287

Merged
kvlonge merged 1 commit into
mainfrom
ci/restore-fresh-ratio-cold-limit
Nov 1, 2026
Merged

kvlonge merged 1 commit into
mainfrom
ci/restore-fresh-ratio-cold-limit

Conversation

@kvlonge

@kvlonge kvlonge commented Nov 1, 2026

Copy link
Copy Markdown
Contributor

Why

#1283 gave the same-runner fresh ratio a temporary cold allowance (1.25 wall, 1.41 CPU) because its base, main, did not run the default SQL Rules yet. Main now runs them, so both sides pay the same cost and the allowance would only hide regressions.

Changes

  • Fresh ratio matrix entry back to 1.10 for cold, warm and edit.
  • Remove the --cold-max-ratio and --cold-max-cpu-ratio options, which existed only for the allowance. verify.yml and compare.py now match their state before feat(rules): run SQL Rules through the SQLClean engine #1283.
  • COMPILE_BENCHMARKS.md records the measured Rules cost instead of the allowance.

Verification

  • fensu check: 0 faults. Ratio script unit tests pass (32). verify.yml parses.
  • The fresh ratio job on this PR compares against main with default Rules on both sides; it must pass at 1.10.

Main now runs the default SQL Rules, so both sides of the same-runner fresh
comparison pay their cost. Remove the temporary 1.25 wall and 1.41 CPU cold
allowance and the cold-only ratio options that existed for it.
@kvlonge
kvlonge enabled auto-merge (squash) November 1, 2026 16:07
@kvlonge
kvlonge merged commit 45f102e into main Nov 1, 2026
44 checks passed
@kvlonge
kvlonge deleted the ci/restore-fresh-ratio-cold-limit branch November 1, 2026 16:45
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