[Feature request] Cluster stats

[NeQuissimus]

Consider adding support for reporting statistics such as the following:

• Number of active pods
• Number of entities
• Number of shards
• Number of register/unregister events
• Results of health checks
• Start and stop (and by extension runtime) times of rebalances

Having the above in something like StatsD or Prometheus would enable visualizing cluster size, status and health very easily.

[ghostdogpr]

Currently we expose:

• ShardManager#getAssignments gives you shard assignments, we could expose getPods as well
• ShardManager#getShardingEvents gives you a stream of assignments/unassignments events. We could add register and unregister there.
• Sharding#getPods gives you all pods with their addresses

Note that most of these metrics would be on the shard manager side (rebalances, register/unregister events, etc), but some would be on the pods side (mostly entities: the shard manager doesn't know about them).

[NeQuissimus]

ZIO Metrics should provide an easy interface to send these values to various backends such as DataDog or Prometheus.

[ghostdogpr]

ZIO Metrics integration done in #116