fixing-flaky-tests

Tests passing in isolation but failing concurrently usually indicate flaky tests caused by shared state, cross-test pollution, or race conditions. This Skill diagnoses and fixes these concurrency issues by applying isolation patterns and synchronized waits.

To fix flaky tests caused by shared database state, this Skill applies per-test transactions and resource isolation patterns. It mitigates resource conflicts for ports, files, or global resources across concurrent workers to ensure deterministic execution.

The best way to handle race conditions in asynchronous test suites is using condition-based-waiting for reliable operations. This Skill implements synchronized waits to resolve race conditions and ensure deterministic test execution across concurrent processes.

Yes, you can use these test isolation patterns for both Python and JavaScript test suites. The Skill diagnoses flaky tests across languages, handling shared databases, global state, and file locks across different programming environments.

To reproduce and validate flaky test fixes, run the test suite multiple times to trigger failures, then apply per-worker resource management. Validate stability by rerunning the suite with parallelism enabled to confirm the fixes resolved the flakiness.

Shared state isolation patterns require identifying specific cross-test pollution sources like global variables or file locks. If resource conflicts are highly dynamic or rely on external system states, deterministic execution may be harder to achieve without per-test resource management.