Repository navigation
ci(sast): raise the Semgrep job cap to 30 minutes - #25554
Conversation
UX Review (Opus 5.5) — ⏭️ skippedRevision |
First Principles Review (Opus 5.5) — ✅ PASSPremise-level review of First-Principles-Verdict: PASS A one-value cap change that fixes a measured cancellation (#14818). It adds no new surface. What this change shipsInventory (1 items) — 1 justifiedIntent: let a large-diff PR's SAST lane finish and publish a verdict instead of being cancelled by its own time cap. This is a FIX.
[FIRST-PRINCIPLES-REVIEWED] 8692f72 Description read[DESCRIPTION-READ] 49ae163f2decb4f64f51a96c2efc99bc792ce9197997a9a6fec30c35022ea51f That sha256 names the PR description this verdict read. Recompute it with |
Design Review (Opus 5.5) — ✅ PASSDesign-level review of Design-Verdict: PASS Nothing blocks: the cap sat inside the job's measured normal range, and doubling it is a one-line change that a revert undoes. Notes
[DESIGN-REVIEWED] 8692f72 Description read[DESCRIPTION-READ] 73b4db90f7e6a11314fa533f8b749deb70e5782fba6497d9284561749ea5a923 That sha256 names everything this verdict read: the PR description, plus the 1 evidence file(s) the workflow collected from it and handed this review. It covers that evidence because the media strip replaces every attachment URL with the same placeholder, so an edit swapping one attachment for another would leave the description's captured text identical while what this review judged changed. Recompute it with |
Opus 5.5 Review — ✅ no blocking findingsReviewed Review detailsCandidate 1 is a style/comment-convention matter, which the pipeline's deterministic check owns; no observable wrong outcome. Dropped. No findings. [OPUS-REVIEWED] 8692f72 Verdict parsed from the review's SHA-scoped output markers for commit False positive or not applicable? A repository writer can comment: |
GPT 6.1 Review — ✅ no blocking findingsGPT 6.1 completed its review of This comment is updated in place on each push. Review detailsNo findings. False positive or not applicable? A repository writer can comment: |
iamwhatever
left a comment
There was a problem hiding this comment.
APPROVE at 8692f72.
The only change raises the sast job's timeout-minutes from 15 to 30 with a comment explaining why (code-review.yml#L318-L323). No permissions, secrets, triggers, runner selection or steps change, so the scan's scope and trust boundary are untouched. All checks on this head are green or skipped, including SAST itself, and every AI lane passed with no findings.
AI findings judged non-blocking
- Design Review: the comment cites #14818 while the description says the timings came from #13348. This is comment wording only; #14818 is the issue this PR fixes, so the pointer still leads to the history, and the cap value is unaffected. code-review.yml#L318-L322
— crew-pr-reviewer (r134)
Problem / Motivation
Goal: a large-diff PR's
SAST (Semgrep)lane finishes with a verdict instead of being cancelled by its own time cap.The
sastjob hastimeout-minutes: 15. Its baseline scan re-runs the rules on the merge-base for every changed file, so it gets slower as the diff grows. On #13348 (ubuntu-latest) it took 9m10s, 9m54s, then was cancelled twice at 15m03s / 15m04s, then passed at 14m27s.Why it matters
A cancelled lane publishes no verdict, so PR Readiness counts it as blocking and the author has to re-run it. It looks like a flake, but it is a cap the job nearly fills.
Not a goal
Intended? / Repro
Intended? No. The cap is meant to stop a hung scan, not a normal large one.
Repro (current main):
.github/workflows/code-review.yml:318:timeout-minutes: 15on thesastjob.ubuntu-latest, where fix(dev-fleet): harden every git read, and resolve the base branch per repository #13348 measured 9-14 min and two 15-min cancellations.What changed (motivation → approach → change)
Symptom: large-diff SAST runs cancelled at 15 min. Cause: the cap sits inside the job's normal range. Change: double the cap.
.github/workflows/code-review.yml:sasttimeout-minutes: 15→30, with a comment giving the measured range and why it grows with the diff.Security
CI config only. The scan's rules and scope are unchanged; a longer cap can only turn a missing verdict into a real one.
Backwards compatibility
Compatible. Removes nothing.
Tests
None: a workflow timeout value. The lane runs on this PR.
Proof: not provable -- other: timeout only matters on a large-diff run; this PR's own SAST run exercises the edited job
Manual verification
YAML parses (
yaml.safe_load). The SAST lane runs on this PR.Related Issues
Fixes #14818
Sibling sweep
Swept: git grep -n "timeout-minutes" -- .github/workflows/code-review.yml
Sibling: .github/workflows/code-review.yml
sast-- changed hereSibling: other Code Review jobs -- not affected (no measured overrun reported)
Pattern harvest
Not generalizable: one job's cap.
Checklist
feat|fix|docs|style|refactor|perf|test|chore|ci|build|revert: ...)Contribution License Agreement
Contributed under the project's Apache-2.0 license.