While we currently generate a number of traces into a request's tracing span on the sender side of a snapshot operation - for example, upon snapshot delegation, as well as higher up in the replicate queue for rebalance actions, we do not currently have significant tracing instrumentation on the receiver side, when a snapshot is being ingested. Given that we currently have a concurrency limit of 1 on the receiver side (as well as the sender) it would be extremely useful to have better insight into where time is being spent during rebalance and recovery operations.

Jira issue: CRDB-16643