[Feature] Add leader-election to the backup-restore sidecar according to the multi-node ETCD proposal.

[amshuman-kr]

Feature (What you would like to be added):

Add leader-election to the backup-restore sidecar according to the multi-node ETCD proposal. Bring the snapshotters (both full and incremental) and the defragmentation control loops/go routines under the control of the leader-election as described here.

Also, introduce the logic to regularly cleanup superfluous ETCD members which do not have a corresponding Pod in the StatefulSet as mentioned in gardener/etcd-druid#189.

Motivation (Why is this needed?):

Pick individually executable pieces of the multi-node proposal.

Approach/Hint to the implement solution (optional):

The preferred leader-election mechanism (because of the implementation and operational simplicity) is based on the leading etcd member. This can be done using the same heart-beat on the corresponding ETCD member's endpoint status as mentioned in #333

[ishan16696]

/assign