Relayer logs should be more meaningful

[nambrot]

The general goal here is to make the relayer logs more useful

That includes, but is not exclusive to:

• Default verbosity should be INFO (might already be)
• Should the "indexed block range" log lines to reduced in verbosity?
• some lower level http logs like incoming body completed� or flushed 806 bytes should be moved to TRACE?
• serialize some objects more concisely (HyperlaneMessage, MetadataBuilder
• remove some traces like this one or this one
• when there is a JsonRPCError, probably don't log it in the inner retries at warn if possible, but when the max number was hit, then do log the errors from at least the last retry
• have dedicated info or warn messages (or maybe debug) when a message could not be estimated, no metadata could be fetched
• Log more data about ISM metadata https://discord.com/channels/935678348330434570/1072862711681142874/1072868391045578812

[nambrot]

another good source for real world relayer logs that seem very hard to read https://discord.com/channels/935678348330434570/1063471868742467644/1072155638135795772

[distributedstatemachine]

another good source for real world relayer logs that seem very hard to read https://discord.com/channels/935678348330434570/1063471868742467644/1072155638135795772

I think these logs are harder than usual to read as i had the pretty var enabled. They are easier to parse without it.