What is your systematic approach when a test passes locally but fails in CI?

This is the most frustrating debugging scenario I encounter regularly. Test passes 100% locally, fails 30–80% in CI. My current process: 1. Check timing — add `--timeout` logging to see if CI is just slow 2. Check environment — compare env vars, browser version, OS 3. Check test isolation — does the test depend on previous test state? 4. Add verbose logging and re-run CI 5. Give up and add a retry (bad) Is there a better systematic approach? Especially for Playwright tests where the failure is often a subtle timing difference between local Chrome and CI Chromium.