---
title: "Study: Sentry PR Friction and ADR Standardization"
description: "A 90-day pull request study of getsentry/sentry showing where review back-and-forth clusters and which ADRs/rules could reduce repeat discussion."
---

This study measures where code review friction concentrates in [`getsentry/sentry`](https://github.com/getsentry/sentry) and identifies which ADRs could reduce repeated discussion cycles.

The goal is not to remove human review. The goal is to move repeated, predictable review debates into explicit decisions and machine-checkable policy.

## Highlights

Across **500 merged PRs**, **500 closed-unmerged PRs**, and **251 open PRs** over a 90-day window:

- **PR size is the strongest friction predictor** -- large PRs hit the high-friction quartile **57%** of the time vs **10%** for tiny PRs.
- **Median time-to-merge is ~5 hours**, but P90 reaches **~71 hours** (a 14x multiplier).
- **Repeated discussion themes** cluster around API semantics, type safety, test evidence, UX flow invariants, and scope splitting.
- **5 proposed ADRs** address the top friction sources, with companion rules and a phased rollout plan.

## Read the full study

The complete study -- methodology, baseline metrics, friction maps, theme analysis, and ADR proposals -- is published at:

**[studies.archgate.dev/studies/sentry-pr-review-friction](https://studies.archgate.dev/studies/sentry-pr-review-friction/)**

Source code and data artifacts are in [`archgate/studies`](https://github.com/archgate/studies/tree/main/studies/sentry-pr-review-friction).