Updated recently
Last updated:
Quick Answer: Flaky tests pass and fail inconsistently. Fix them by: (1) identifying root causes (race conditions, timing, external dependencies), (2) implementing automatic retries, (3) quarantining flaky tests, (4) adding network stubs and explicit waits.
Common Causes of Flaky Tests
| Cause | Example | Fix |
|---|---|---|
| Timing issues | Click before element ready | Use explicit waits |
| Network dependency | External API slow/down | Stub network requests |
| Shared state | Tests modify shared data | Isolate test data |
| Parallel execution | Two tests use same resource | Unique data per worker |
| Dynamic content | Timestamps change | Flexible assertions |
Strategy 1: Automatic Retries
// playwright.config.ts
export default defineConfig({
retries: 2,
workers: 4,
});Strategy 2: Quarantine Flaky Tests
- name: Run stable tests
run: npx playwright test --grep-invert="@flaky"
- name: Run flaky tests (non-blocking)
continue-on-error: true
run: npx playwright test --grep="@flaky"Strategy 3: Stub External Dependencies
cy.intercept('GET', 'https://api.external.com/*', {
statusCode: 200,
body: { data: 'mocked' }
});Training Resources
Master Playwright with SkilBrill Playwright Training.
FAQ
Should I delete flaky tests?
No. Quarantine them, fix the root cause, and restore them. Track flaky rates as a team metric.
