FlakeTriage

FlakeTriage is a web app (with an optional CLI) that turns flaky-test chaos into actionable test case design work. It ingests CI runs from tools like GitHub Actions/Jenkins and correlates failures across time, environments, and code changes to identify which test cases are probabilistically flaky versus genuinely broken. It then clusters failures by signature (stack traces, logs, timing patterns) and proposes concrete design fixes: better waits, deterministic data setup, isolation boundaries, mocking strategies, and assertions that reduce brittleness. The app generates a “flake budget” report per repo and per team, showing the highest-cost tests by wasted minutes and rerun frequency. It also creates Jira/Linear tickets with reproduction steps and suggested refactors, so the work actually lands. This is not magic: it won’t fix bad architecture, but it will reliably surface the worst offenders and guide pragmatic improvements.