fix: gRPC balancer shutdown panic

The gRPC node discovery hack spawns a task that outlives the gRPC
balancer - once the balancer stops, the task should stop too (and not
panic sending on the closed channel).

router/src/dml_handlers/rpc_write.rs

@@ -42,12 +42,22 @@ where
     tokio::spawn(async move {
         loop {
             for e in &endpoints {
-                tx.send(tower::discover::Change::Insert(
+                // The gRPC balance listener will stop first during shutdown.
+                if tx
+                    .send(tower::discover::Change::Insert(
                         e.uri().to_owned(),
                         e.clone(),
                     ))
                     .await
-                .expect("no grpc balance receiver");
+                    .is_err()
+                {
+                    // The gRPC balancer task has stopped - likely because
+                    // the server has stopped.
+                    //
+                    // Do not leak this task, or panic trying to send on the
+                    // closed channel.
+                    return;
+                }
             }
 
             tokio::time::sleep(Duration::from_secs(5)).await;