VolumeSnapshotClass is now configurable (was hardcoded before) Improved PVC deletion stability with proper wait logic Enhanced test coverage for edge cases Redis StatefulSet backup/restore fully tested end-to-end Code quality improvements: clean linting, better error handling
Why this exists: Velero is great for full-cluster DR, but sometimes you just need fast, snapshot-based backups for StatefulSets without the overhead of external object storage or complex setup. This operator focuses on:
Scheduled snapshots with cron expressions Pre/post backup hooks (e.g., database flush before snapshot) Per-replica retention policies Point-in-time recovery via CRDs 100% declarative, GitOps-friendly
Current limitations: Still alpha (v0.0.3). No cross-cluster restore yet, no webhook validation, container selection for hooks needs work. Velero is still the production choice for serious DR needs. What I'm curious about: Is there actual demand for this kind of focused tool, or does Velero's flexibility already cover these use cases well enough that alternatives don't make sense? Tested successfully on Minikube, Kind, and now Redis workloads. Planning to test on GKE/EKS next. Feedback welcome!