fix: remove pending actors metric, fix actor error tracker, engine runner error print

Author: MasterPtato

engine/docker/dev-host/grafana/dashboards/guard.json

@@ -1324,7 +1324,7 @@
   },
   "timepicker": {},
   "timezone": "browser",
-  "title": "Rivet Guard",
+  "title": "Guard",
   "uid": "cen785ige8fswd",
   "version": 3
 }

engine/docker/dev-multidc-multinode/core/grafana/dashboards/guard.json

@@ -254,6 +254,7 @@ impl Init {
 	}
 }
 
+#[tracing::instrument(skip_all)]
 pub async fn handle_init(
 	ctx: &StandaloneCtx,
 	conn: &Conn,

engine/docker/dev-multidc/core/grafana/dashboards/guard.json

@@ -1058,11 +1058,11 @@ async fn handle_stopped(
 
 				// Set Crashed failure reason for actual crashes.
 				// Runner failure reasons are already set at the start of handle_stopped.
-				if let StoppedVariant::Normal { code, message } = &variant {
-					ensure!(
-						*code != protocol::mk2::StopCode::Ok,
-						"expected non-Ok stop code in crash handler, got Ok"
-					);
+				if let StoppedVariant::Normal {
+					code: protocol::mk2::StopCode::Error,
+					message,
+				} = &variant
+				{
 					ctx.v(3)
 						.activity(runtime::SetFailureReasonInput {
 							failure_reason: FailureReason::Crashed {

engine/docker/dev-multinode/grafana/dashboards/guard.json

@@ -971,7 +971,7 @@ pub(crate) async fn allocate_pending_actors(
 	let runner_eligible_threshold = ctx.config().pegboard().runner_eligible_threshold();
 
 	// NOTE: This txn should closely resemble the one found in the allocate_actor activity of the actor wf
-	let (allocations, pending_actor_count) = ctx
+	let allocations = ctx
 		.udb()?
 		.run(|tx| async move {
 			let start = Instant::now();
@@ -993,7 +993,6 @@ pub(crate) async fn allocate_pending_actors(
 				// the one we choose
 				Snapshot,
 			);
-			let mut pending_actor_count = 0;
 			let ping_threshold_ts = util::timestamp::now() - runner_eligible_threshold;
 
 			'queue_loop: loop {
@@ -1006,8 +1005,6 @@ pub(crate) async fn allocate_pending_actors(
 					break;
 				};
 
-				pending_actor_count += 1;
-
 				let (queue_key, generation) =
 					tx.read_entry::<keys::ns::PendingActorByRunnerNameSelectorKey>(&queue_entry)?;
 
@@ -1123,20 +1120,15 @@ pub(crate) async fn allocate_pending_actors(
 						},
 					});
 
-					pending_actor_count -= 1;
 					continue 'queue_loop;
 				}
 			}
 
-			Ok((allocations, pending_actor_count))
+			Ok(allocations)
 		})
 		.custom_instrument(tracing::info_span!("runner_allocate_pending_actors_tx"))
 		.await?;
 
-	metrics::ACTOR_PENDING_ALLOCATION
-		.with_label_values(&[&input.namespace_id.to_string(), &input.name.to_string()])
-		.set(pending_actor_count as i64);
-
 	Ok(AllocatePendingActorsOutput { allocations })
 }
 

engine/docker/dev/grafana/dashboards/guard.json

@@ -649,7 +649,7 @@ pub(crate) async fn allocate_pending_actors(
 	let runner_eligible_threshold = ctx.config().pegboard().runner_eligible_threshold();
 
 	// NOTE: This txn should closely resemble the one found in the allocate_actor activity of the actor wf
-	let (allocations, pending_actor_count) = ctx
+	let allocations = ctx
 		.udb()?
 		.run(|tx| async move {
 			let start = Instant::now();
@@ -671,7 +671,6 @@ pub(crate) async fn allocate_pending_actors(
 				// the one we choose
 				Snapshot,
 			);
-			let mut pending_actor_count = 0;
 			let ping_threshold_ts = util::timestamp::now() - runner_eligible_threshold;
 
 			'queue_loop: loop {
@@ -684,8 +683,6 @@ pub(crate) async fn allocate_pending_actors(
 					break;
 				};
 
-				pending_actor_count += 1;
-
 				let (queue_key, generation) =
 					tx.read_entry::<keys::ns::PendingActorByRunnerNameSelectorKey>(&queue_entry)?;
 
@@ -801,20 +798,15 @@ pub(crate) async fn allocate_pending_actors(
 						},
 					});
 
-					pending_actor_count -= 1;
 					continue 'queue_loop;
 				}
 			}
 
-			Ok((allocations, pending_actor_count))
+			Ok(allocations)
 		})
 		.custom_instrument(tracing::info_span!("runner_allocate_pending_actors_tx"))
 		.await?;
 
-	metrics::ACTOR_PENDING_ALLOCATION
-		.with_label_values(&[&input.namespace_id.to_string(), &input.name.to_string()])
-		.set(pending_actor_count as i64);
-
 	Ok(AllocatePendingActorsOutput { allocations })
 }
 
@@ -856,12 +848,6 @@ async fn drain_older_versions(
 	ctx: &ActivityCtx,
 	input: &DrainOlderVersionsInput,
 ) -> Result<crate::ops::runner::drain::Output> {
-	tracing::info!(
-		namespace_id = %input.namespace_id,
-		name = %input.name,
-		version = input.version,
-		"drain_older_versions activity called"
-	);
 	ctx.op(crate::ops::runner::drain::Input {
 		namespace_id: input.namespace_id,
 		name: input.name.clone(),