Pull request hygiene metrics help your team maintain a healthy code review process. CompassHQ automatically analyzes PR activity from your connected repositories and surfaces patterns that can slow down delivery or introduce risk.
Key Metrics Tracked
PR Size
CompassHQ measures the total number of lines changed (additions plus deletions) in each pull request and assigns it to one of four size buckets:
| Bucket | Lines Changed |
|---|---|
| Small | 0—100 |
| Medium | 101—400 |
| Large | 401—1,000 |
| XLarge | 1,000+ |
These thresholds reflect a well-documented pattern in code review research: reviewer effectiveness drops as the size of a change grows. Reviewers can comfortably process roughly 200—400 lines in a single session before fatigue sets in and defect detection rates decline sharply. Beyond 1,000 lines, reviews tend to become superficial — reviewers skim rather than read, and subtle bugs slip through.
The 100-line “small” cutoff marks the range where reviews are consistently thorough. The 400-line boundary represents the practical upper limit of a focused review session. Above 1,000 lines, the review is large enough that splitting it into smaller changes will almost always produce better outcomes.
The overall size assessment shown on the dashboard is based on the median PR size across all merged PRs in the selected period, so a handful of outliers will not skew the rating.
Time to First Review
This metric captures how long a pull request waits before receiving its first review. Long wait times indicate bottlenecks in your review process and directly impact lead time for changes. CompassHQ tracks this from the moment the PR is opened (or marked ready for review) to the first substantive review.
Review Coverage
Review coverage measures the percentage of merged PRs that received at least one approved review before merging. A low coverage rate may signal that code is being merged without adequate oversight, which can increase your change failure rate.
PR Churn
Churn measures how much code changes after a review — the rework cycle. A churn event is triggered when the PR author pushes changes exceeding a dynamic threshold (10% of the PR size, with a 20-line floor) after a reviewer submits feedback. Multiple rework rounds on the same PR score exponentially higher. See PR Churn for the full methodology.
Stale PR Detection
CompassHQ flags pull requests that have been open without activity beyond a configurable threshold (default: 7 days). Stale PRs often indicate blocked work, unclear ownership, or scope creep. Addressing them keeps your backlog clean and your cycle time predictable.
Viewing PR Hygiene Data
- Navigate to Metrics > PR Hygiene from the sidebar.
- Select a team or repository to filter the results.
- Review the summary cards at the top for average PR size, median review time, and coverage percentage.
- Scroll down to see individual PR details, sorted by age or size.
Improving Your PR Practices
Use the trend charts to identify whether metrics are moving in the right direction over time. If review times are climbing, consider adjusting team review assignments or introducing review-time SLAs. If PR sizes are growing, encourage developers to break work into smaller, reviewable units.