Progressive delivery tools like Flagger promise automated canary analysis and rollback, but will the safety net actually catch you when you fall?

Most teams set up Flagger, run a happy-path canary once, and call it done. The rollback logic sits there untested until a real incident forces it to fire - and by then, it's too late to find out it was misconfigured.

This talks is based on experience bringing a payment system live into production, what we got wrong, and why I would have loved to have a tools like Chaos Mesh instead of just load tests to prove rollbacks worked.

In this session, I'll show you how to combine two CNCF projects - Flagger for progressive delivery and Chaos Mesh for fault injection - to validate that your deployment pipeline actually does what you think it does. You'll see a live demo where we inject latency and failures during a canary rollout, watch Flagger's metric analysis detect the issue, and trigger an automated rollback.

You'll walk away with a practical pattern for testing your delivery pipeline's immune system - not just shipping it and hoping for the best.