Repo Health Metrics

Key Findings

• Fewer maintainers: Across all dotnet repos, open issues and PRs per maintainer are rising. runtime lost ~28% of active maintainers since 2023 peak, yet workload hasn't shrunk proportionally.
• Community contributors concentrating: runtime active contributors (community + maintainer) down ~25% from 2022 peak, but PR volume held steady — remaining contributors are shouldering more work. The funnel is narrowing.
• First-timer retention halved: runtime's 12-month return rate dropped from ~40% (2020) to 22% (2024). This isn't just a bigger pool — total first-timers fell from 234 to 137/yr while returners collapsed from 117 to 34.
• Community merge latency gap: maui community PRs wait 8× longer than maintainer PRs. runtime and roslyn show a 2× gap. This potentially drives the retention drop.
• maui Syncfusion dependency: Syncfusion engineers account for a large share of maui community PRs since mid-2024. Community health metrics are heavily dependent on this single partnership.
• Issue turnaround holding: Despite fewer maintainers, runtime's community issue turnaround (~50% closed within 30 days) has held steady since 2020. But the open backlog % is creeping up (14% → 30%).
• Copilot adoption accelerating: runtime is the most aggressive adopter of Copilot SWE Agent PRs, with rapid month-over-month growth across all four dotnet repos.
• How dotnet compares: vscode has 2× the maintainer pool. rust merges ~160 PRs/week via bors with 1,000+ contributors. dotnet repos are healthy but less automated and more maintainer-constrained than peers.

Potential Actions

• Copilot Code Review impact (NEW): See the charts below for current data on Copilot review coverage, comment density, churn reduction, and time-to-first-feedback. Early signs suggest AI review reduces human rework; the key metric to watch is whether human review effort decreases over time as Copilot review improves.
• Reduce community PR merge latency: maui's 8× gap and runtime's 2× gap potentially drive the retention drop. Faster first review could help?
• Nurture top contributors: Top 10% of community authors (~110 people in runtime) produce ~76% of community PRs. Identify and invest in these contributors — losing even a few has outsized impact.
• Diversify maui community: High dependence on Syncfusion for community PRs since mid-2024. Broaden the contributor base to reduce single-partner dependency.
• Explore automation: rust's near-parity in merge latency and higher retention suggest automated reviewer assignment and merge tooling could help — worth investigating?