expand Nomad's own Consul health checks and/or tags to include autopilot health

[tgross]

Once https://github.com/hashicorp/nomad/issues/13219 is complete, Nomad will run autopilot metrics on all servers. This could let us have health checks that specifically check that a server has completed restoring its initial snapshot before being marked as healthy for RPC and HTTP clients.

Note that we likely need to separate this out from the service/check we use for Consul autodiscovery, because that's needed to get the raft snapshot in the first place.

[tgross]

@schmichael suggested that this could be a sub-route on https://developer.hashicorp.com/nomad/api-docs/v1.3.x/agent#health