Flaky tests erode trust in CI and nobody tracks which ones
“We just hit re-run until it's green. Nobody knows if a test failed because the code is broken or because it's Tuesday — the suite cried wolf so often that real failures get ignored.”— Hacker News ↗
Flake-detection SaaS (BuildPulse, Launchable, CI-vendor analytics) is enterprise-priced and wants deep CI integration; small teams get nothing between 'ignore it' and '$500/mo'. A self-hosted board that ingests plain JUnit XML from any CI and ranks tests by flake rate is the missing middle.
Small teams with a growing E2E/integration suite
Build a flaky-test tracker in Next.js. CI POSTs each run's JUnit XML to an ingest endpoint (repo, branch, commit, run id); store per-test pass/fail history in SQLite. Flake score = tests that both failed and passed on the same commit (retries) or flip state with no code change. Dashboard ranks the flakiest tests and exports a quarantine list. First 3 steps: (1) JUnit XML parser + authenticated ingest route, (2) schema for runs/results keyed by test name with history, (3) the flake-score algorithm + ranked dashboard with per-test failure timelines.