expose more fine-grained information about what faults happen

[ralexstokes]

the current faults API only exposes an unsigned integer as a fault count.

the relay monitor should expose more information (which it will necessarily have) to any interested parties

this could be a V2 of the faults endpoint or even a new API which accepts the fault type and the relay pubkey(s) and returns details information about the fault

for example, it could be nice to know when a given fault happened; if something was 3000 epochs ago, it is less actionable than if it was 3 epochs ago