It was originally discussed that we would like to offer the following backup destinations for the backup and restore functionality:

• PVC
• S3

It was decided both solutions were good as long as PVC was offered, as not all users wanted to use S3. For this reason, we have prioritized first the implementation of the backup/restore functionality having PVC as the backup destination.

After having explored in some detail both approaches I would like to discuss the value/usefulness of providing PVC as the backup destination:

• A PVC is created in a Kubernetes Cluster, and it uses a PV (that is manually created by the user or dynamically provisioned by a user configuration) that lives also in the same cluster.
• PVs live in a Kubernetes cluster (and are cluster-level scope)
• A Pod that uses a PVC requires the PVC to exist in the same Kubernetes Namespace of the pod

All of the previous points imply that the backup result is stored in the same cluster, which does not seem ideal to me for Disaster Recovery purposes, as the PVC/PV/Backup would be stored in the same cluster as where the installation was performed.