DevOps Practitioner · Module 3: Helm and environment configuration · Lab
Return to the Last Working Release After a Failed Upgrade
50 min hands-on · Advanced
A local KinD or k3d cluster with Helm installed and the demo release running a working revision.
Local guide: run the steps below on your own machine in order, then check the validation list.
Objectives
- Upgrade to a broken chart default and verify the failure in traffic
- Roll back to the recorded working revision from history
- Verify restored traffic and leave readable history entries
Step 1
Break it forward
Upgrade the demo release to the provided chart version with a broken default. Verify the failure with real traffic checks, not just pod state, and record the symptom.
Step 2
Read history and roll back
Read the release history, name the last working revision with its values, and roll back to it. Verify traffic restored and quote the history entries before and after.
Step 3
Repeat correctly
Re-run the upgrade with the default corrected through the same value chain, verify traffic, and confirm the history reads like a log: broken revision, rollback, fixed revision.
How to confirm it worked
- Failure verified with traffic evidence, not pod state alone
- Rollback names the working revision from history
- Traffic restored and quoted after rollback
- History reads as a log of break, rollback and fix