CKAD — Deployment Rollbacks
Why Rollbacks Matter
Updates can go wrong — a bad image, a misconfigured container, or an application bug. Kubernetes keeps a history of Deployment revisions so you can quickly roll back to a known-good state.
Check Rollout History
kubectl rollout history deployment/web
# Expected:
# REVISION CHANGE-CAUSE
# 1 <none>
# 2 kubectl set image deployment/web nginx=nginx:1.27
# View details of a specific revision
kubectl rollout history deployment/web --revision=1
Each time you update a Deployment, Kubernetes creates a new revision. The old ReplicaSet is kept (scaled to 0) so it can be restored.
Roll Back to Previous Version
# Roll back to the previous revision
kubectl rollout undo deployment/web
# Roll back to a specific revision
kubectl rollout undo deployment/web --to-revision=1
# Verify the rollback
kubectl rollout status deployment/web
kubectl describe deployment web | grep Image
What Happens During a Rollback
- Kubernetes finds the ReplicaSet from the target revision
- Scales the old ReplicaSet back up (using rolling update strategy)
- Scales the current ReplicaSet down
- Creates a new revision number for the rollback
A rollback is essentially a rolling update — back to an old version.
Pause and Resume Rollouts
If you need to make multiple changes before triggering a rollout:
# Pause the rollout
kubectl rollout pause deployment/web
# Make multiple changes — none trigger a rollout while paused
kubectl set image deployment/web nginx=nginx:1.27
kubectl set resources deployment/web -c nginx --limits=cpu=500m,memory=512Mi
# Resume — triggers a single rollout with all changes
kubectl rollout resume deployment/web
Why Pause and Resume?
| Without Pause | With Pause |
|---|---|
| Each change triggers a separate rollout | All changes applied in one rollout |
| Multiple Pod restarts | Single Pod restart |
| Intermediate states visible | Clean atomic update |
Rollout Status
# Check if a rollout is in progress
kubectl rollout status deployment/web
# Possible outputs:
# "deployment "web" successfully rolled out" ← done
# "Waiting for deployment rollout to finish..." ← in progress
# Exit code 1 ← failed/timed out
# Check with timeout (useful for scripts and exam)
kubectl rollout status deployment/web --timeout=60s
Revision History Limit
By default, Kubernetes keeps the last 10 ReplicaSets (revisions). You can change this:
spec:
revisionHistoryLimit: 5 # Keep only 5 old ReplicaSets
Setting this to 0 means no rollback capability — old ReplicaSets are deleted immediately.
Rollback Quick Reference
| Task | Command |
|---|---|
| View history | kubectl rollout history deployment/<name> |
| View revision details | kubectl rollout history deployment/<name> --revision=N |
| Undo to previous | kubectl rollout undo deployment/<name> |
| Undo to specific revision | kubectl rollout undo deployment/<name> --to-revision=N |
| Pause rollout | kubectl rollout pause deployment/<name> |
| Resume rollout | kubectl rollout resume deployment/<name> |
| Check status | kubectl rollout status deployment/<name> |
Exam tip: If you push a bad image during the CKAD exam, kubectl rollout undo deployment/<name> is the fastest recovery. Practice the full cycle: update → check → rollback.
Next: Learn operational commands with Managing Deployments →
