hard20 pts · first answer countsreliabilityci-cd

Handling flaky tests at scale

Your suite has 2,000 tests and about 3% fail randomly each run, so developers ignore red builds. What is the best strategy?

A.Automatically retry each failed test 5 times and report it as passedB.Track flakiness per test, quarantine flaky tests out of the blocking gate, assign owners to fix root causes, and keep the main gate trustedC.Delete all tests that ever failedD.Stop running tests on pull requests

Breaking changes between services →