etcd full DB backup/restore is not practical for failover, cluster migration etc. because it restores the exact and full etcd state, including data which should not be overwritten such as cluster node IP's etc. Full backup is applicable to restoring etcd in-place from DB corruption.

More fine grained selection of key/value tree data is needed in fail-over and cloning scenarios. Individual selection of ndslabs values, and selected kubernetes keys representing the running state of ndslabs and supportive services.

1. Add key/value backup to cluster-backup daemonset