Add support for configuring rclone retries in backup container

Currently, in the helm-chart, the backup container is ran as a sidecar. This creates a scenario where if the S3 endpoint becomes unavailable for whatever reason, the pod frequently restarts and essentially creates a server outage.

This can be mitigated by having the mc-backup able to be configured to indefinitely retry the rclone sync without exiting with a failure, if so the user chooses.

Start in the helm-chart and the mc-backup entry point, then trace how the rclone sync command handles failure. Determine how an opt-in indefinite retry setting should be passed to the backup sidecar and verify that an unavailable S3 endpoint no longer causes the container to exit. Check for existing container or chart tests before making the change.

Written by the indexing model from the issue text.