How do you handle flaky tests in a CI/CD pipeline, and what metrics do you use to measure test reliability?
