Analytics is available from day one. Each metric names its unit and whether higher or lower is better, then shows the current window, previous-window delta, trend buckets, and repository breakdown where evidence exists.

Analytics metrics

• Deployment frequency (deploys/day).
• Failed production deployments.
• Lead time for changes.
• Review turnaround time.
• Stale and aging pull requests.
• Pull requests merged without recorded human approval.

Overview movement

Overview also summarizes PRs merged/opened, shipped rate, deployments total/failed, median time to first review, review comments, unresolved comments, and actionable review threads.

Useful for choosing which repos actually moved, and for Vanta context. Metrics do not replace the weekly digest as the sampling population.

Propose this when

Use metrics (or GitHub merge/deploy volume) to pick 1–2 production-path repos, not to invent extra rules. Extra busy repos are later.

When a metric lacks enough evidence, keep the value unavailable or qualified; do not manufacture a trend from a tiny sample.