kvserver: don't check ConnHealth when releasing proposal quota

The proposal quota release procedure checked node connection health for every
node that appears active after replica activity checks. This is expensive, and
previously caused issues like #84943.

This change removes the ConnHealth check, because other checks, such as
isFollowerActiveSince and paused replicas, provide sufficient protection from
various kinds of overloads.

Release note: None

pkg/kv/kvserver/replica_proposal_quota.go

@@ -163,11 +163,9 @@ func (r *Replica) updateProposalQuotaRaftMuLocked(
 		) {
 			return
 		}
-
-		// Only consider followers that that have "healthy" RPC connections.
-		if err := r.store.cfg.NodeDialer.ConnHealth(rep.NodeID, r.connectionClass.get()); err != nil {
-			return
-		}
+		// At this point, we know that either we communicated with this replica
+		// recently, or we became the leader recently. The latter case is ambiguous
+		// w.r.t. the actual state of that replica, but it is temporary.
 
 		// Note that the Match field has different semantics depending on
 		// the State.
@@ -210,7 +208,7 @@ func (r *Replica) updateProposalQuotaRaftMuLocked(
 		if progress.Match > 0 && progress.Match < minIndex {
 			minIndex = progress.Match
 		}
-		// If this is the most recently added replica and it has caught up, clear
+		// If this is the most recently added replica, and it has caught up, clear
 		// our state that was tracking it. This is unrelated to managing proposal
 		// quota, but this is a convenient place to do so.
 		if rep.ReplicaID == r.mu.lastReplicaAdded && progress.Match >= commitIndex {