Tasks

Loading tasks...

This challenge was created by a community author. Community content is reviewed by the iximiuz Labs team on a best effort basis.

Challenge, Easy, on Kubernetes

Resolve OOMKilled Errors by Reading VPA Memory Recommendations

by Omkar Shelke

Scenario

A Deployment web-app is running in the production namespace with 2 replicas. The web-container consumes more memory than the configured limit, causing repeated OOMKilled events — the container is being terminated by the kernel because it tried to use more memory than its limit allows.

A Vertical Pod Autoscaler (VPA) named web-app-vpa is already configured for this Deployment with updateMode: Off. In this mode, the VPA continuously observes the container's actual resource usage and calculates a recommendation — but it never applies that recommendation automatically. It is your job to read the recommendation and apply it manually.

Task

Check the VPA recommendation for the web-container
Update the Deployment web-app to set the memory request and memory limit to match the VPA's recommended target value
Ensure the Deployment rollout completes successfully and the replica count remains 2

Do not modify CPU values — requests.cpu and limits.cpu must remain 50m.

Wait a few minutes for the VPA to collect metrics before checking recommendations — if you check too early, the recommendation may not be populated yet.

Hint 1 — Confirm the OOMKilled Pods

Start by confirming the container is actually being OOMKilled. Check the Pod status and recent events:

kubectl get pods -n production
kubectl describe pod -n production -l app=web-app

Look for OOMKilled in the Last State section and a high Restart Count — this confirms the container is exceeding its memory limit.

Documentation

Resource Management for Pods and Containers

Hint 2 — Read the VPA Recommendation

The VPA does not change anything automatically when updateMode: Off — it only calculates recommendations. View them with:

kubectl describe vpa web-app-vpa -n production

Look for the Recommendation section near the bottom of the output. It will show several values per container — Target, Lower Bound, Upper Bound, and Uncapped Target. Use the Target memory value — this is the recommended value to apply.

You can also extract just the target memory value directly:

kubectl get vpa web-app-vpa -n production \
  -o jsonpath='{.status.recommendation.containerRecommendations[0].target.memory}'

Documentation

Vertical Pod Autoscaler

Hint 3 — Apply the Recommendation to the Deployment

Update only the memory request and memory limit — leave CPU untouched. You can do this with kubectl set resources in one command, or by editing the Deployment directly:

kubectl set resources deployment web-app -n production \
  --containers=web-container \
  --requests=memory=<target-value> \
  --limits=memory=<target-value>

Or edit the Deployment manually:

kubectl edit deployment web-app -n production

After saving, watch the rollout complete and confirm the Pods are stable:

kubectl rollout status deployment web-app -n production
kubectl get pods -n production

Documentation

Test Cases

Start challenge