Detect and warn about broken replication slots on the Geo primary

Unused replication slots on the gitlab database on the Geo testbed (from the production database restore) caused large amount of log data to build up in pg_xlog. As ~ 10TB of repositories were copied from the primary to the secondary we saw 100GBs of log data on the primary.

Removing the unused replication slots fixed this, causing the postgresql dataset to return to its previous size. We should include information about the number and status of replication slots in the status screen for the Geo primary.