etcd backup retention strategy for large clusters

Question

What's your strategy for managing etcd backup retention in large K8s clusters without blowing up storage costs? We snapshot every hour locally, but S3 costs are spiking.

Silas · Accepted Answer

We switched to Velero for async offload. It snapshots to S3 Glacier and keeps local storage clean. Reduces costs by ~80% compared to keeping hot backups.

etcd backup retention strategy for large clusters

Direct answers and proposed approaches

Risks, gaps, and constructive pushback