Explanation

Backup etcd / dqlite

The Infra Backup Operator does not create backups of the underlying datastore (such as etcd or dqlite). While regularly backing up these databases is a recommended best practice for cluster operators, Velero will still capture all the essential Kubernetes resources required to restore a functional cluster in the event of a disaster. What will be lost in such scenarios are transient data such as events, logs, and other ephemeral Kubernetes metadata that are not critical for cluster recovery.

Non-kubernetes api resources

Velero works just on Kubernetes API level. This means that any kind of customization of kube-apiserver, auditing or other third-parties tools configuration files like Falco won’t be under the protection of the backup using this tool.

Charm lifecycle

This charm does not install workloads and its functionality focuses on leveraging the velero-backups endpoint and continuously monitoring the namespaces configured for backup. Any change in the charm’s configuration or in the Kubernetes cluster is automatically reflected in the set of resources included in the backup. If the configuration specifies a namespace that does not yet exist, no error will be raised. Once that namespace is created, it will automatically be included in the backup process.