Support node readiness labels

[liger1978]

The Zalando PostgreSQL Operator depends on node readiness labels to schedule pods on nodes. To trigger a failover of the "leader" role in the PostgreSQL Patroni HA cluster from one pod to another in the event of a node reboot, a label must be updated on the node before it is cordoned and drained. The label must be updated once more after the node is rebooted and uncordened.

Would you accept a PR to optionally update node readiness labels before and after the reboot process?

[evrardjp]

It makes sense to me.

[liger1978]

Please review/merge PR #541 that adds this enhancement. I have built and tested kured from PR #541. It works fine.

[liger1978]

@ckotzbauer @evrardjp