Friction Map
This page breaks down review friction by code area — identifying which domains, file regions, and labels correlate with higher friction.
By Title Scope
Section titled “By Title Scope”Parsed from conventional commit format type(scope): description. Only scopes with ≥5 PRs are shown.
| Scope | PRs | Median Reviews | Median TTM (h) | Changes Requested Rate | Median Churn |
|---|---|---|---|---|---|
| github | 8 | 5.5 | 15.27 | 0.0% | 268.5 |
| seer | 28 | 3.5 | 1.69 | 0.0% | 89.5 |
| attachments | 5 | 3.0 | 13.40 | 0.0% | 113.0 |
| onboarding | 7 | 3.0 | 102.85 | 0.0% | 266.0 |
| workflows | 14 | 2.5 | 18.64 | 0.0% | 152.5 |
| aci | 17 | 2.0 | 18.72 | 0.0% | 76.0 |
| autofix | 29 | 2.0 | 2.57 | 0.0% | 50.0 |
| dashboards | 23 | 2.0 | 1.85 | 0.0% | 55.0 |
| nav | 5 | 2.0 | 12.84 | 0.0% | 52.0 |
| occurrences on eap | 6 | 2.0 | 24.35 | 0.0% | 89.5 |
Observations
Section titled “Observations”githubis the highest-friction scope by median reviews (5.5). The GitHub integration domain involves OAuth flows, pipeline state management, and security considerations — themes we see recurring in the discussion analysis.onboardinghas the longest TTM at 102.85 hours median, despite moderate review counts. This suggests bottlenecks other than review discussion (perhaps waiting for design approval or cross-team coordination).seerandautofixare high-volume but relatively low-friction, suggesting these areas have clearer conventions or dedicated reviewer familiarity.- Zero formal
CHANGES_REQUESTEDacross all scopes reinforces that Sentry’s friction manifests through comment threads, not formal review states.
By File Domain
Section titled “By File Domain”File domain is derived from the first two path segments of changed files. This captures which parts of the codebase generate the most review discussion.
| Domain | PRs | Median Reviews | Median TTM (h) | Changes Requested Rate | Median Churn |
|---|---|---|---|---|---|
| CODEOWNERS | 26 | 4.0 | 13.98 | 0.0% | 164.0 |
| migrations lockfile | 5 | 3.0 | 16.14 | 0.0% | 182.0 |
| static/gsApp | 13 | 3.0 | 19.46 | 0.0% | 145.0 |
| tests/js | 6 | 3.0 | 2.47 | 0.0% | 68.5 |
| tests/sentry | 137 | 3.0 | 4.84 | 0.0% | 89.0 |
| tests/snuba | 10 | 2.5 | 44.93 | 0.0% | 91.0 |
| src/sentry | 245 | 2.0 | 6.62 | 0.0% | 74.0 |
| static/app | 220 | 2.0 | 10.35 | 0.0% | 73.0 |
Observations
Section titled “Observations”- PRs touching CODEOWNERS have the highest median review count (4.0). These are cross-cutting changes that likely require approval from multiple code owners.
static/gsApp(legacy frontend namespace) has both high review counts and high TTM — possibly due to the complexity of maintaining legacy code alongside the newerstatic/appfrontend.tests/snubahas the highest TTM (44.93h) despite moderate review counts, suggesting these PRs involve complex data pipeline changes that take time to validate.src/sentry(245 PRs, 49% of sample) andstatic/app(220 PRs, 44%) are the two dominant file domains, with relatively moderate friction levels.
By Label
Section titled “By Label”Sentry’s label usage is sparse — most PRs carry only scope labels:
| Label | PRs | Median Reviews | Median TTM (h) |
|---|---|---|---|
| Trigger: Revert | 6 | 5.0 | varies |
| Scope: Backend | 274 | 2.0 | varies |
| Scope: Frontend | 239 | 2.0 | varies |
The Trigger: Revert label is associated with the highest review counts. Revert PRs often signal a prior mistake that needs investigation, which drives additional discussion.
The backend/frontend split is roughly even (274 vs 239), but labels alone don’t provide enough granularity for meaningful friction segmentation. Title scope and file domain are more useful dimensions.
Friction Hotspots Summary
Section titled “Friction Hotspots Summary”The domains with consistently elevated friction are:
- GitHub integration (
githubscope) — complex OAuth/security flows - Onboarding (
onboardingscope) — longest TTM, cross-team coordination - Cross-cutting changes (CODEOWNERS-touching PRs) — multi-owner approval
- Legacy frontend (
static/gsApp) — maintaining old patterns alongside new - Snuba-related tests (
tests/snuba) — complex data pipeline validation
These hotspots correlate with the discussion themes identified from comment analysis: security, state management, and API design are the dominant themes in these high-friction domains.