8000 ci: split heavy bench into one job per benchmark group by adityamukho · Pull Request #117 · project-minigraf/minigraf · GitHub
[go: up one dir, main page]

Skip to content

ci: split heavy bench into one job per benchmark group - #117

Merged
adityamukho merged 2 commits into
mainfrom
fix/bench-heavy-timeouts
Apr 9, 2026
Merged

adityamukho merged 2 commits into
mainfrom
fix/bench-heavy-timeouts

Conversation

@adityamukho
Copy 8000 link
Copy Markdown
Collaborator

Summary

  • 5 of 7 heavy benchmark jobs were hitting the GitHub Actions 6-hour wall-clock limit because each job ran 2–5 Criterion benchmark functions in sequence
  • Rewrote bench.yml to use a matrix strategy (21 jobs, one per benchmark group) matching the approach already used in bench_light.yml
  • timeout-minutes raised from 90 → 360 to match bench_light; fail-fast: false added so a single slow job cannot cancel siblings
  • Regression reporting moved inline (continue-on-error + conditional issue step), eliminating the old report-regression fan-in job
  • Anchored regex filters (e.g. ^insert/ vs ^insert_file/) ensure each job runs exactly the intended benchmarks with no cross-contamination

Test plan

  • Trigger the workflow manually via workflow_dispatch and verify all 21 matrix jobs start and complete without timeout
  • Confirm Bencher receives uploads for each benchmark group
  • Verify a regression in one group does not cancel sibling jobs (fail-fast: false)

Fixes: https://github.com/adityamukho/minigraf/actions/runs/24193906944

🤖 Generated with Claude Code

adityamukho and others added 2 commits April 10, 2026 01:40
Previous jobs combined 2–5 bench functions per runner and hit the
6-hour GitHub Actions wall-clock limit.  Each runner now runs exactly
one Criterion filter, matching the approach already used in
bench_light.yml.  timeout-minutes raised from 90 → 360 (matching
bench_light) and fail-fast disabled so a single slow job cannot cancel
siblings.  Regression reporting moved inline (continue-on-error +
conditional issue step) to eliminate the now-unnecessary
report-regression fan-in job.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
@codecov
codecov Bot commented Apr 9, 2026 •
Copy link
Copy Markdown

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 83.73%. Comparing base (09939a6) to head (3fc1e6d).
⚠️ Report is 1 commits behind head on main.

Additional details and impacted files

Impacted file tree graph

@@           Coverage Diff           @@
##             main     #117   +/-   ##
=======================================
  Coverage   83.73%   83.73%           
=======================================
  Files          26       26           
  Lines        4729     4729           
=======================================
  Hits         3960     3960           
  Misses        769      769           
🚀 New features to boost your workflow:
  • ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

@adityamukho
adityamukho merged commit 3d283c7 into main Apr 9, 2026
17 checks passed
@adityamukho
adityamukho deleted the fix/bench-heavy-timeouts branch April 9, 2026 20:18
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

0