What is a flaky test?

A flaky test is one that passes and fails on the same code without a deterministic cause - timing races, shared test data, network blips, or environment drift are common triggers. Flakiness erodes trust: teams re-run until green, ignore failures, or disable suites, which hides real bugs. UI automation is especially prone because of async rendering, animations, and third-party scripts. Fixes usually involve stable selectors, explicit waits instead of fixed sleeps, isolated data, and idempotent setup/teardown. Track flake rate like any quality metric and quarantine or fix offenders quickly. One ignored flaky test often becomes ten; treat instability as a defect, not noise.