Feature Request: improve observability into the size of the Praefect replication queue

Currently we no have tooling to provide insight into how many jobs are queued in Praefect. This is useful in determining if replication is healthy and for estimating when a cluster will be in sync.

Right now admins can connect to the Praefect DB and run SELECT COUNT(*) FROM replication_queue;, but a CLI tool and prometheus metric would be useful.