Change log message levels according to new system

Signed-off-by: Dr. Jan-Philip Gehrcke <[email protected]>

Author: jgehrcke

cmd/compute-domain-controller/cleanup.go

@@ -123,7 +123,7 @@ func (m *CleanupManager[T]) cleanup(ctx context.Context) {
 			continue
 		}
 
-		klog.Infof("Cleanup: stale %T found for ComputeDomain '%s', running callback", *new(T), uid)
+		klog.V(1).Infof("Cleanup: stale %T found for ComputeDomain '%s', running callback", *new(T), uid)
 		if err := m.callback(ctx, uid); err != nil {
 			klog.Errorf("error running CleanupManager callback: %v", err)
 			continue

cmd/compute-domain-controller/computedomain.go

@@ -232,7 +232,7 @@ func (m *ComputeDomainManager) onAddOrUpdate(ctx context.Context, obj any) error
 		return fmt.Errorf("failed to cast to ComputeDomain")
 	}
 
-	klog.Infof("Processing added or updated ComputeDomain: %s/%s/%s", cd.Namespace, cd.Name, cd.UID)
+	klog.V(2).Infof("Processing added or updated ComputeDomain: %s/%s/%s", cd.Namespace, cd.Name, cd.UID)
 
 	if cd.GetDeletionTimestamp() != nil {
 		if err := m.resourceClaimTemplateManager.Delete(ctx, string(cd.UID)); err != nil {

cmd/compute-domain-controller/daemonset.go

@@ -363,7 +363,7 @@ func (m *DaemonSetManager) onAddOrUpdate(ctx context.Context, obj any) error {
 		return fmt.Errorf("failed to cast to DaemonSet")
 	}
 
-	klog.Infof("Processing added or updated DaemonSet: %s/%s", d.Namespace, d.Name)
+	klog.V(2).Infof("Processing added or updated DaemonSet: %s/%s", d.Namespace, d.Name)
 
 	cd, err := m.getComputeDomain(d.Labels[computeDomainLabelKey])
 	if err != nil {

cmd/compute-domain-kubelet-plugin/device_state.go

@@ -212,7 +212,7 @@ func (s *DeviceState) Unprepare(ctx context.Context, claimRef kubeletplugin.Name
 		// device was never prepared or has already been unprepared (assume that
 		// Prepare+Checkpoint are done transactionally). Note that
 		// claimRef.String() contains namespace, name, UID.
-		klog.Infof("unprepare noop: claim not found in checkpoint data: %v", claimRef.String())
+		klog.V(2).Infof("unprepare noop: claim not found in checkpoint data: %v", claimRef.String())
 		return nil
 	}
 

cmd/compute-domain-kubelet-plugin/driver.go

@@ -199,6 +199,9 @@ func (d *driver) UnprepareResourceClaims(ctx context.Context, claimRefs []kubele
 			if done {
 				results[claim.UID] = err
 				wg.Done()
+				if err != nil {
+					klog.V(0).Infof("Permanent error unpreparing devices for claim %v: %v", claim.UID, err)
+				}
 				return nil
 			}
 			return fmt.Errorf("%w", err)
@@ -248,13 +251,14 @@ func (d *driver) nodePrepareResource(ctx context.Context, claim *resourceapi.Res
 			Err: fmt.Errorf("error preparing devices for claim %s/%s:%s: %w", claim.Namespace, claim.Name, claim.UID, err),
 		}
 		if isPermanentError(err) {
-			klog.V(6).Infof("Permanent error preparing devices for claim %v: %v", claim.UID, err)
+			klog.V(0).Infof("Permanent error preparing devices for claim %v: %v", claim.UID, err)
 			return true, res
 		}
 		return false, res
 	}
 
-	klog.Infof("prepared devices for claim '%s/%s:%s': %v", claim.Namespace, claim.Name, claim.UID, devs)
+	klog.V(1).Infof("prepared devices for claim '%s/%s:%s': %v", claim.Namespace, claim.Name, claim.UID, devs)
+
 	return true, kubeletplugin.PrepareResult{Devices: devs}
 }
 
@@ -269,7 +273,7 @@ func (d *driver) nodeUnprepareResource(ctx context.Context, claimRef kubeletplug
 		return isPermanentError(err), fmt.Errorf("error unpreparing devices for claim '%v': %w", claimRef.String(), err)
 	}
 
-	klog.Infof("unprepared devices for claim '%v'", claimRef.String())
+	klog.V(1).Infof("Unprepared devices for claim '%v'", claimRef.String())
 	return true, nil
 }
 

cmd/compute-domain-kubelet-plugin/health.go

@@ -130,14 +130,14 @@ func (h *healthcheck) Check(ctx context.Context, req *grpc_health_v1.HealthCheck
 		klog.ErrorS(err, "failed to call GetInfo")
 		return status, nil
 	}
-	klog.V(6).Infof("Successfully invoked GetInfo: %v", info)
+	klog.V(7).Infof("Successfully invoked GetInfo: %v", info)
 
 	_, err = h.draClient.NodePrepareResources(ctx, &drapb.NodePrepareResourcesRequest{})
 	if err != nil {
 		klog.ErrorS(err, "failed to call NodePrepareResources")
 		return status, nil
 	}
-	klog.V(6).Info("Successfully invoked NodePrepareResources")
+	klog.V(7).Info("Successfully invoked NodePrepareResources")
 
 	status.Status = grpc_health_v1.HealthCheckResponse_SERVING
 	return status, nil

pkg/workqueue/workqueue.go

@@ -135,7 +135,10 @@ func (q *WorkQueue) processNextWorkItem(ctx context.Context) {
 
 	err := q.reconcile(ctx, workItem)
 	if err != nil {
-		klog.Errorf("Failed to reconcile work item: %v", err)
+		// Most often, this is an expected, retryable error in the context of an
+		// eventually consistent system. Hence, do not log an error level. Rely
+		// on inner business logic to log unexpected errors on error level.
+		klog.V(1).Infof("Reconcile: %v", err)
 		// Only retry if we're still the current operation for this key
 		q.Lock()
 		if q.activeOps[workItem.Key] != nil && q.activeOps[workItem.Key] != workItem {