detect node being hard down after upgrade fails

[mkaczanowski]

TL;DR

With last sommelier upgrade, the node panicked at the migration step:

3:06PM INF cellarfees v1 to v2: Store migration complete module=server
3:06PM INF migrating module crisis from version 1 to version 2 module=server
3:06PM INF migrating module distribution from version 2 to version 3 module=server
3:06PM INF migrating module gov from version 3 to version 4 module=server  
panic: no concrete type registered for type URL /cork.v1.AddManagedCellarIDsProposal against interface *v1beta1.Content  

The node died, but the post check was at:

Post upgrade check: GRPC_RESPONSIVE Waiting for the grpc and cometbft services to be responsive

In the end the post check failed so we knew smth was off. But I would think it would be nice for GRPC response check to also check if the node is down down.

If the node is hard down (docker ps) and there is no restart policy (which there is none) we know the node won't come back online.