How to fix KubeletHasDiskPressure, pods evicted, node under disk pressure in Kubernetes

KubernetesINTERMEDIATEHIGH

Nodes run out of disk space, triggering eviction of pods. Fix by cleaning up disk space, configuring proper storage limits, and monitoring disk usage.

What this error means

When a node runs low on disk space, the kubelet marks the node with DiskPressure condition and starts evicting pods to reclaim space. Critical pods are evicted last, but best-effort pods go first.

Typical symptoms

Node shows DiskPressure condition
Pods suddenly evicted without user request
kubectl get nodes shows disk pressure warnings
Application pods killed unexpectedly

Common causes

Container logs filling disk
Old images and containers not cleaned up
Large temporary files accumulating
Ephemeral storage limits too low

How to fix "KubeletHasDiskPressure, pods evicted, node under disk pressure"

1Check disk usage

bash

kubectl describe node <node-name> | grep Disk
df -h /var/lib/kubelet

2Clean up old containers and images

bash

sudo docker system prune -a
sudo crictl rmi --prune

3Clean up log files

bash

sudo find /var/log -type f -name "*.log" -exec truncate {} \;

4Set ephemeral storage limits

yaml

resources:
  limits:
    ephemeral-storage: "1Gi"

Advanced notes

Use kubelet flags like --image-gc-high-threshold and --image-gc-low-threshold.

Related errors

Official resources & further reading

📚Kubernetes Official Documentation
📖
Node-pressure Eviction - Kubernetes Docs(Official Docs)