A release pipeline should automatically revert to the last known good version when post-deployment health checks fail. Which design achieves this?
Show answer and explanation
Answer: C
Post-deployment gates evaluate signals such as health endpoints or monitoring alerts after the release is live. When a gate fails, a rollback task can swap back to the previous slot or redeploy the prior artifact without waiting for a person. This closes the loop automatically, so a bad release is reverted quickly. Skipping checks lets failures through, leaving the decision to a later manual step is slow, and removing checks hides problems rather than handling them.
Why the other options are wrong
- A. Deploying without any health validation means a broken release stays live until users report it, which defeats automated recovery.
- B. Paging an operator is manual recovery and extends the outage while someone is found, so it does not meet the requirement for automation.
- D. Removing health checks makes the pipeline look green while the application may be failing, so it hides the problem rather than fixing it.