To enable resource version bumps on restore, we would need to know how far into the future (in terms of revisions) we need to bump. 

We can get this information by requesting endpoint status on each member and using the maximum of all RaftIndex fields as the result. Alternatively by finding the current leader and getting its endpoint status directly.

Even though this is not an expensive operation, this should be polled in a sensible interval, e.g. once every 30s. 

The result should be written as a textfile in the hostPath /var/lib/etcd that is already mounted on all relevant pods. An additional etcd sidecar container should be the most sensible choice to run this code.