Use this for the “validate with metrics” rollout: compare median duration and queue time before and after workflow changes.
OWNER_REPO="allthingslinux/monorepo" # adjust
WORKFLOW="portal-ci.yml"
gh run list --repo "$OWNER_REPO" --workflow "$WORKFLOW" --limit 30 \
--json databaseId,conclusion,createdAt,updatedAt,event \
--jq '.[] | {id: .databaseId, conclusion, createdAt, event}'gh run list --repo "$OWNER_REPO" --workflow "$WORKFLOW" --limit 50 \
--json databaseId,createdAt,updatedAt,conclusion \
--jq '.[] | select(.conclusion=="success") |
((.updatedAt | fromdateiso8601) - (.createdAt | fromdateiso8601)) as $s |
{id: .databaseId, seconds: $s}'RUN_ID=12345678901
gh run view "$RUN_ID" --repo "$OWNER_REPO" --json jobs \
--jq '.jobs[] | {name: .name, status: .status, started: .startedAt, completed: .completedAt}'Remote caching is not enabled by default. To experiment in CI, set TURBO_TOKEN and TURBO_TEAM (or vendor-specific vars) as repository secrets and pass them into jobs that run turbo. Document any adoption in this file when enabled.