rpc: add trace events for network send/receive

Latency investigations sometimes point towards slowness in the network, but this can't be conclusively shown in traces, and we often have to resort to Go runtime traces which are hard to obtain.

It might be helpful to add trace events for when a gRPC request is put on the network and received by the remote peer. This should be done as close to the network/OS as possible, i.e. it should ignore gRPC network buffers/queues. It might also be useful to log the estimated node clock skew, if we can do so reliably, to account for this when comparing timestamps across peers.

For writes, this would involve adding tracing through Raft as well #104035.

Jira issue: CRDB-32858