Thank you, I'm aware of the operator pattern. I'm not sure what this does for me beyond the things I've just said. There are already Ceph, NFS, and other persis

Thank you, I'm aware of the operator pattern. I'm not sure what this does for me beyond the things I've just said. There are already Ceph, NFS, and other persistent volume provisioners that integrate by providing storage classes that can satisfy persistent volume claims. Implementing CRDs and operators for things other than persistent volumes makes sense, too. But does Rook provide a uniform interface for those to do backups/disaster recovery? Or is it just a collection of heterogeneous storage-y things? Because it seems like the latter reading the docs, and I think that's okay, but the way the page sells it seems inaccurate: > Rook turns storage software into self-managing, self-scaling, and self-healing storage services. It does this by automating deployment, bootstrapping, configuration, provisioning, scaling, upgrading, migration, disaster recovery, monitoring, and resource management. Which resources have all of those things automated? It looks like Ceph is the only one that has any documentation on disaster recovery at all. For that matter, those docs are solely about getting a cluster in a bad state to restart after quorum is lost. Nothing in the documentation describes how to backup, ship backups, and restore backups.