As part of #85445, we intend to prioritize queued snapshots, particularly on the receiver side, by using the priority specified by the allocator (or other sending source). The queued snapshots should be separated by sending source, i.e. Raft Snapshot Queue / Replicate Queue / Admin Change Replicas, so that we can ensure fairness between each of the sending sources. The goal of this prioritization is to ensure that receiver stores respect the priority of the snapshots it receives, such that (for example) snapshots of replicas that exist on decommissioning stores can be prioritized higher than a snapshot sent for the purpose of range count rebalancing, which may be sent by a store that no longer has any "decommissioning" replicas to process.

Jira issue: CRDB-18939