Once clusters are installed, Hive SyncSets make up the bulk of Hive's etcd traffic as well as underlying storage, particularly in loaded environments. On guidance from the arch team, we should avoid doing any regular large writes to etcd.

Hive's ClusterSync is a perfect candidate for breaking out into some additional storage backend. This resource stores per cluster status for all matching SyncSets. It can easily be reconstructed if lost, and it is not a resource we reconcile on.

Goal is to deploy an alternative storage backend via the Hive Operator, and eliminate the ClusterSync etcd resource.

Once completed this helps Hive clusters scale to our anticipated 1000 per Hive target.