fix: ignored conflicts and watch updates (#1472)

- Reset the retry backoff when syncing is skipped due to
  prior successful sync and there are no conflicts or watch updates.
  This ensures the retry event will still fire to check for new
  conflicts and watch update requests from the remediator.
- Add logging with more details for debugging, but at level 3, to
  avoid spamming customer logs.

pkg/parse/event_handler.go

@@ -124,7 +124,9 @@ func (s *EventHandler) Handle(event events.Event) events.Result {
 		} else if opts.needToUpdateWatch() {
 			trigger = triggerWatchUpdate
 		} else {
-			// No RunFunc call
+			// Skip RunFunc and reset the backoff to keep checking for conflicts & watch updates.
+			klog.V(3).Info("Sync retry skipped; resetting retry backoff")
+			eventResult.ResetRetryBackoff = true
 			break
 		}
 
@@ -139,7 +141,11 @@ func (s *EventHandler) Handle(event events.Event) events.Result {
 	}
 
 	// If the run succeeded or source changed, reset the retry backoff.
-	if runResult.Success || runResult.SourceChanged {
+	if runResult.Success {
+		klog.V(3).Info("Sync attempt succeeded; resetting retry backoff")
+		eventResult.ResetRetryBackoff = true
+	} else if runResult.SourceChanged {
+		klog.V(3).Info("Source change detected; resetting retry backoff")
 		eventResult.ResetRetryBackoff = true
 	}
 	return eventResult

pkg/parse/events/publisher.go

@@ -161,7 +161,7 @@ func (s *RetrySyncPublisher) Publish(subscriber Subscriber) Result {
 
 	retryDuration := s.currentBackoff.Step()
 	retries := s.retryLimit - s.currentBackoff.Steps
-	klog.Infof("a retry is triggered (retries: %v/%v)", retries, s.retryLimit)
+	klog.V(3).Infof("Sending retry event (step: %v/%v)", retries, s.retryLimit)
 
 	result := subscriber.Handle(Event{Type: s.EventType})