Kubernetes Simulator
intermediate · ~12 min
The 3am page
You're on call. web is running happily on nginx:1.28… for now. Something is about to change.
When a rollout goes wrong, don't guess. Find which pods are unhappy (kubectl get pods), then ask why: kubectl describe pod NAME ends with the pod's events, and the events usually say exactly what failed.
Two ways out: fix the image and roll forward, or kubectl rollout undo to go back to the last revision that worked.
Objectives
- Find out why the new pods won't start
- Get web healthy: every pod running on an image that exists
- Bonus: Fix it with kubectl rollout undo
Best on a computer with a keyboard. It works on a phone, with shortcuts under the terminal.