Work out flaky test rate instantly with clear inputs, formula shown and shareable results.
A flaky test is one that fails and then passes on retry without any code change. The damaging metric is the share of all failures that are flaky: once it passes roughly half, developers stop reading failures and start reflexively retrying, which means genuine regressions reach production. The wasted compute is real but secondary to that loss of trust.
Flakiness
flaky rate = flaky failures / runs; false failure share = flaky failures / total failures; wasted time = flaky failures x rerun duration
Quarantine immediately to restore signal, then fix on a deadline. Leaving them in the blocking suite trains the team to ignore red builds.
Timing assumptions and shared mutable state — fixed sleeps instead of waiting for a condition, test order dependence, and unreset databases or caches between tests.