Kubernetes Error Code Evicted: How to Fix It

Medium 30-60 minutes High Severity Verified June 2026
Error Code
Evicted - The node was low on resource: ephemeral-storage
Brand
Kubernetes
Product Type
dev_tool
Severity
High
DIY Difficulty
Medium
Estimated Fix Time
30-60 minutes
The Kubernetes 'Evicted' error occurs when a pod is forcibly removed from a node due to low ephemeral storage resources. This happens when the node runs out of disk space, triggering the kubelet to evict pods to free up storage and maintain cluster stability.
Ad

Tools You'll Need

How to Fix Error Code Evicted - The node was low on resource: ephemeral-storage

  1. Check Pod Eviction Status

    Do not delete evicted pods immediately as they contain important debugging information
  2. Analyze Node Storage Usage

  3. Clean Up Unused Images and Containers

    Ensure no critical workloads are using the images before pruning
  4. Check Container Log Sizes

    Back up important logs before implementing log rotation policies
  5. Increase Ephemeral Storage Requests

  6. Scale Cluster or Add Storage

    Cordoning nodes will affect workload distribution and may impact application availability
  7. Configure Resource Quotas and Limits

  8. Monitor and Set Up Alerts

Parts You May Need

Additional storage volumes
Check Price on Amazon
Additional cluster nodes
Check Price on Amazon
Additional storage volumes
Check Price on Amazon
Additional cluster nodes
Check Price on Amazon
Additional storage volumes
Check Price on Amazon
Additional cluster nodes
Check Price on Amazon
Additional storage volumes
Check Price on Amazon
Additional cluster nodes
Check Price on Amazon
Additional storage volumes
Check Price on Amazon
Additional cluster nodes
Check Price on Amazon
Additional storage volumes
Check Price on Amazon
Additional cluster nodes
Check Price on Amazon
Additional storage volumes
Check Price on Amazon
Additional cluster nodes
Check Price on Amazon
Additional storage volumes
Check Price on Amazon
Additional cluster nodes
Check Price on Amazon
Ad

When to Call a Professional

Call a Kubernetes platform engineer or DevOps specialist if storage issues persist after cleanup, if you need to resize production node storage, or if implementing cluster-wide resource management policies. Professional help is essential for production clusters with strict SLA requirements.

Frequently Asked Questions

Why do Kubernetes pods get evicted?
Pods get evicted when nodes run low on resources like memory, disk space, or ephemeral storage. The kubelet monitors resource usage and evicts pods to maintain node stability and prevent system crashes.
Can I prevent pod evictions in Kubernetes?
Yes, prevent evictions by setting appropriate resource requests and limits, implementing resource quotas, monitoring node resources, and maintaining adequate cluster capacity. Regular cleanup of unused images and logs also helps.
What happens to data when a pod is evicted?
Data in ephemeral storage (emptyDir volumes) is lost when pods are evicted. Data in persistent volumes remains intact. Always use persistent storage for important data that should survive pod restarts.
How do I recover from Kubernetes evicted pods?
Clean up the evicted pod status with 'kubectl delete pod ', address the underlying storage issue, then let the controller (Deployment, StatefulSet) recreate the pod automatically.
What is ephemeral storage in Kubernetes?
Ephemeral storage includes the container's writable layer, logs, and emptyDir volumes. It's temporary storage that's deleted when the pod is removed and counts toward the node's storage capacity limits.