Debugging Vertical Pod Autoscaler (VPA)

Follow these steps to diagnose issues with VPA not recommending resources or not updating pods.

1. Check Metrics Server

VPA relies on the Metrics Server to gather historical usage data. If it's down, VPA won't work.

bash
1kubectl get deployment metrics-server -n kube-system
2kubectl top nodes
3kubectl top pods -A

If kubectl top fails or returns "Metrics not available", the Metrics Server is the issue.

2. Check VPA Components

Ensure the VPA recommender, updater, and admission-controller are running.

bash
1kubectl get pods -n kube-system -l app.kubernetes.io/name=vpa

3. Inspect VPA Object

Check the VPA object for Conditions and Recommendations.

bash
1kubectl get vpa -A
2kubectl describe vpa <vpa-name> -n <namespace>

Look for Provided condition being True. If False, check the message field.

4. Check Recommender Logs

If recommendations are missing, check the recommender logs.

bash
1kubectl logs -n kube-system -l app=vpa-recommender

Look for permission errors or "no metrics" warnings.

5. Check Evictions

If recommendations exist but pods aren't updating, check if the VPA Updater is evicting pods.

bash
1kubectl get events -n <namespace> --sort-by='.lastTimestamp'

Look for EvictedByVPA events.