bug: health endpoint still reports OK when node has stopped producing blocks.

[tac0turtle]

we should show failure when there is an error with the node.

[chatton]

related to #2676 do we want to just have a single endpoint instead of health + ready?

[auricom]

Logs from sept. 10 about how the erratic behavior

# ev-node
2:40PM INF syncing header and data component=BlockManager height=3016
2:40PM INF syncing header and data component=BlockManager height=3016
2:40PM ERR unrecoverable error in one of the go routines error="failed to sync next block: failed to apply block: failed to execute transactions: failed to get block info: failed to get block at height 3015: not found" component=main
2:40PM INF halting full node and its sub services... component=main

curl localhost:7331/health/live
OK

==> When a fullnode was crashing and hung indefinitely, the health/live endpoint still answered that the process was healthy