# Developer Metrics > Track velocity, review efficiency, and team performance OpenCodeHub includes built-in developer metrics that help teams understand their engineering performance without needing external tools. --- ## Overview ### What Gets Measured - **Cycle Time** — Time from first commit to merge - **Throughput** — PRs merged per week - **Review Time** — Time from PR open to first review - **Review Load** — How many reviews each team member handles - **Code Churn** — Lines added/deleted over time - **Merge Queue Efficiency** — Queue wait times and batch success rates --- ## Dashboard Access metrics at: - **Repository** → Insights → Metrics - **Organization** → Insights (aggregated across repos) - **Personal** → Profile → Contributions ### Key Charts | Chart | Description | |-------|-------------| | Cycle Time Trend | Average days from commit to merge over time | | PR Velocity | PRs opened vs merged per week | | Review Heatmap | Who reviews whom | | Code Frequency | Commits per day/week | | Queue Health | Merge queue wait times and success rates | --- ## Metric Definitions ### Cycle Time ``` Cycle Time = (PR merged_at) - (First commit timestamp) ``` **Benchmarks:** - < 1 day: Excellent - 1-3 days: Good - 3-7 days: Fair - > 7 days: Needs attention ### Review Turnaround ``` Review Time = (First review comment) - (PR opened_at) ``` **Benchmarks:** - < 4 hours: Excellent - 4-24 hours: Good - 1-3 days: Fair - > 3 days: Needs attention ### Review Load Balance Shows whether review work is evenly distributed: ``` Review Load = Number of reviews given per person per week ``` Imbalanced load (>2x difference between team members) indicates: - Knowledge silos - Bottlenecks - Burnout risk --- ## Using Metrics for Improvement ### Identify Bottlenecks High cycle time + high review time = Review bottleneck - Solution: Require faster reviews, add reviewers High cycle time + low review time = CI/test bottleneck - Solution: Optimize pipeline, parallelize tests ### Track Goals Set team goals and track progress: ``` Goal: Reduce average cycle time from 5 days to 2 days Goal: Achieve < 24h review turnaround for 90% of PRs Goal: Keep review load within 20% across team members ``` ### Sprint Retrospectives Use metrics in retrospectives: 1. Review cycle time trends 2. Identify PRs that took unusually long 3. Discuss blockers and process improvements --- ## CLI Access ```bash # View personal metrics och metrics --me # View team metrics och metrics --team # View repository metrics och metrics --repo owner/repo # Export to CSV och metrics --repo owner/repo --format csv --output metrics.csv ``` --- ## API Access ```bash # Get repository metrics curl https://git.yourcompany.com/api/repos/owner/repo/metrics \ -H "Authorization: Bearer $TOKEN" # Get user metrics curl https://git.yourcompany.com/api/users/username/metrics \ -H "Authorization: Bearer $TOKEN" # Get organization metrics curl https://git.yourcompany.com/api/orgs/org-name/metrics \ -H "Authorization: Bearer $TOKEN" ``` --- ## Privacy - Personal metrics are visible to the user and organization admins - Repository metrics are visible to repo members - Raw data is never shared externally - AI models do not train on metrics data --- ## See Also - [Stacked Pull Requests](stacked-prs.md) - [Smart Merge Queue](merge-queue.md) - [Team Workflows](../guides/team-workflows.md)