GitHub has added pull-request review-stage timing to its enterprise and organization repository-level Copilot usage metrics reports. Announced on September 25, 2026, the addition can help QA leaders distinguish a slow validation handoff from a slow review conversation or a merge queue delay.
What changed in GitHub Copilot usage metrics
The new pull_request_review_times array is included on each repos-1-day report row. For qualifying pull requests merged in a repository on that day, it reports both median and 90th-percentile minutes for three intervals: ready for review to first review, first review to final review, and final review to merge.
- Ready to first review: reveals initial reviewer or test-owner queue time.
- First to final review: highlights rework, validation feedback, and review-loop time.
- Final review to merge: surfaces post-approval gates, merge queues, or release-process friction.
GitHub says the data is attributed to the day a pull request merged. The existing pull_requests fields are unchanged.
Why this matters for QA engineers
A single lead-time number can hide the real bottleneck. If the p90 for ready-to-first-review climbs, a QA team might examine reviewer coverage, test-environment availability, or ownership rules. If first-to-final-review is the outlier, investigate flaky checks, unclear acceptance criteria, or regressions repeatedly discovered late. A high final-review-to-merge value can point to required checks, merge-queue capacity, or release approvals.
Use the new data as an operational signal, not a quality score. Faster review is only helpful when the team still catches meaningful defects and preserves traceable test evidence.
Important reporting limits
- Only pull requests opened by a person and reviewed by at least one different person are timed.
- Copilot code review, other bots, and author reviews are excluded from the timing calculation; a PR with both a human and Copilot review can still qualify.
- The new series has no historical backfill. GitHub excludes PRs made ready for review before September 21, 2026 from this section, so early data will be sparse.
- An empty array means there were no qualifying merged PRs that day; it does not mean a zero-minute review time.
A practical QA rollout check
Before setting an SLA or dashboard alert, first confirm that the Copilot usage metrics policy is enabled and that the intended owner has the required metrics permission. Then collect several weeks of forward-only data, segment repositories by risk and test-suite size, and compare medians with p90 values. Pair a timing spike with CI failure rate, flaky-test rate, and escaped-defect signals before changing a quality gate or reviewer workflow.
GitHub documents daily repository reports for enterprises and organizations, with signed download links and role-based access. That makes this feature most useful when QA, engineering, and platform teams agree on a small shared set of review-health measures.