Pull in `source.*` and `destination.*` fieldsets from ECS?

[trask]

It's not clear if these should be pulled in, or if client.* and server.* effectively replace them.

One tangential concern is that currently HTTP messaging semconv has attributes under messaging.source.* and messaging.destination.*, which would conflict if we also adopt the concept from ECS of nested namespaces.

cc @AlexanderWert

[trask]

I suspect that source.* and destination.* will be needed, e.g. #3402 (comment)

there are many scenarios where there are equal players (peers) communicating, e.g. nodes in Paxos or any gossip protocol, actors exchanging messages in systems like Akka, etc. How would one decide who is a client or a server in those scenarios?

[Oberon00]

And note that neither source/destination nor server/client are a conceptual match for host/peer.

So you could make an argument to retain host/peer conventions and add the ECS conventions on top... It's messy.

[AlexanderWert]

I guess this can be closed because #3402 has been merged?

cc @trask