no monitoring available for multisite radosgw

[Danny Cocks]

Currently, relating a ceph-radosgw cluster from one zone to another zone via the primary/secondary endpoints can create a working multisite radosgw replicated set of environments.

However, there is no monitoring for the state of the replication. Both in terms of the connection and the state of replication.

The ceph-mon package allows for monitoring of this status through the "radosgw-admin sync status" command. The output of the command is similar to:

          realm 13e98d2e-bbf9-4b2f-81b9-0829b3aa4664 (replicated)
      zonegroup 3d328592-2781-48c6-830c-6d14890f056f (mygroup)
           zone b8576ea2-a8dc-42ad-a242-e5021873cd3d (first)
  metadata sync no sync (zone is master)
      data sync source: 2d7911a9-3c77-4b99-be26-fe9bada04db6 (second)
                        init
                        full sync: 128/128 shards
                        full sync: 0 buckets to sync
                        incremental sync: 0/128 shards
                        data is behind on 128 shards
                        behind shards: [0, ...]

It would be nice to have an alert when there is an error in the output of this command or when the command status is failing.

[Danny Cocks (dannycocks)]

I have started an attempt for adding this feature in here:

https://review.opendev.org/c/openstack/charm-ceph-mon/+/901094