Issue

Backup jobs to cloud storage fail with an error indicating that the Pod exceeded its ephemeral local storage limit.

Environment

• Neo4j self-managed deployment on Kubernetes

• Neo4j Helm chart 

Symptoms

Backup jobs fail, and describe pod logs show: Pod ephemeral local storage usage exceeds the total limit of containers 40Gi

Cause

The database backup process requires temporary local disk space before uploading data to cloud storage.
Even though the data is streamed to S3/Blob, the backup operation still generates temporary files that must fit within the Pod’s ephemeral storage allocation. If the database backup size exceeds the pod’s configured limits, Kubernetes terminates the pod due to storage pressure.

Resolution

1 - Configure a tempVolume in the Helm chart

Use the tempVolume parameter to provide additional scratch space for the backup job.

This volume is used as temporary storage for backup files before they are uploaded to the cloud provider. This allows backup jobs to use dedicated storage instead of relying solely on pod ephemeral space.

Example (Helm values):

  tempVolume:
    persistentVolumeClaim:
      claimName: neo4j-backup-pvc

2 - Increase the Pod ephemeral storage limit

If you prefer to avoid attaching a temp volume, increase the pod ephemeral storage request and limit so the backup job can temporarily hold the backup data.

For more details, check: Backup storage options