How do you handle test failures in the deployment pipeline?

test failures
pipeline management
failure handling

Your answer

Answer it first, then compare. Nothing here is saved or scored.

Model answer

Strategy: (1) Blocking failures: critical test failures halt the pipeline, no deployment. (2) Non-blocking warnings: flaky tests or low-priority issues are reported but do not block. (3) Failure categorization: infrastructure failure (retry), flaky test (quarantine and investigate), real regression (block). (4) Fast feedback: run fastest tests first (unit, then integration, then e2e). (5) Failure notifications: Slack/email with failure details and links. (6) Automatic issue creation for persistent failures. Never just re-run and hope, investigate root cause.

What interviewers listen for

Pipeline shape
You design stages by failure cost and runtime budget, not by tool feature lists.
Parallelisation craft
You can shard a real suite and explain the state-isolation choices you made.
Quality gates
You know which checks block merge, which warn, and why those defaults are correct.
Shift-left fluency
You talk about ownership transfer, not just earlier execution.

See all 34 ci/cd questions