/vitals-review

Dating a Lab drop to deploys

A Lab drop starts somewhere between the last reference Scan and the first worse Scan. Name the production deploys in that window and the commits they shipped.

Find the deploy workflow

Read .github/workflows/ and pick the workflow that deploys production from the default branch. The repository's AGENTS.md or README.md usually names it. If none deploys, say so and use merge times on the default branch instead.

List the deploys in the window

gh run list --workflow <deploy-workflow.yml> --branch main --status success \
  --created "<last-reference-scan>..<first-worse-scan>" \
  --json headSha,createdAt,updatedAt,url --limit 50

Also take the last successful deploy before the window. Its headSha is the start of the range.

Name the commits

git log --oneline <sha-before-window>..<last-sha-in-window>

Write suspectCommits as <start>..<end>, then the pull request numbers the range merged. One deploy with one squash merge names one pull request. That is the strongest claim this Routine can make.

When the window is empty

No deploy between the two Scans means the code did not change. Say so in the report. Look for an outside cause: a third-party script, a CDN or cache change, or a NuxtSEO scanner change. File the finding only if a repository change can still remove the cost, and write suspectCommits as no deploy in window.