fix: Remove duplicate OTEL observability setup in imc-dispatcher (#8885)

The imc-dispatcher was calling otel.SetupObservabilityOrDie in its
NewController function, but sharedmain.MainWithContext already sets up
observability before calling NewController. This caused Go runtime
metrics to be registered twice, resulting in the error:

  "collected metric 'go_memory_allocations_total' was collected before
   with the same name and label values"

This caused the metrics endpoint (/metrics) to fail with an error,
preventing Prometheus from scraping metrics.

Changes:
- Use otel.GetMeterProvider() and otel.GetTracerProvider() to get the
  global providers already set up by sharedmain
- Remove duplicate pprof server creation (sharedmain handles it)
- Remove pprof config watch (sharedmain handles it)
- Remove provider shutdown calls (sharedmain handles cleanup)
- Clean up unused imports

Author: creydr

pkg/reconciler/inmemorychannel/dispatcher/controller.go

@@ -20,6 +20,7 @@ import (
 	"context"
 	"time"
 
+	"go.opentelemetry.io/otel"
 	"knative.dev/pkg/injection"
 	"knative.dev/pkg/system"
 
@@ -32,16 +33,13 @@ import (
 	"knative.dev/eventing/pkg/channel/multichannelfanout"
 	"knative.dev/eventing/pkg/eventingtls"
 	"knative.dev/eventing/pkg/kncloudevents"
-	"knative.dev/eventing/pkg/observability/otel"
 
 	"knative.dev/pkg/logging"
 
 	"go.uber.org/zap"
 	filteredconfigmapinformer "knative.dev/pkg/client/injection/kube/informers/core/v1/configmap/filtered"
 	"knative.dev/pkg/configmap"
-	configmapinformer "knative.dev/pkg/configmap/informer"
 	"knative.dev/pkg/controller"
-	k8sruntime "knative.dev/pkg/observability/runtime/k8s"
 	pkgreconciler "knative.dev/pkg/reconciler"
 
 	kubeclient "knative.dev/pkg/client/injection/kube/client"
@@ -55,7 +53,6 @@ import (
 	inmemorychannelinformer "knative.dev/eventing/pkg/client/injection/informers/messaging/v1/inmemorychannel"
 	inmemorychannelreconciler "knative.dev/eventing/pkg/client/injection/reconciler/messaging/v1/inmemorychannel"
 	"knative.dev/eventing/pkg/inmemorychannel"
-	o11yconfigmap "knative.dev/eventing/pkg/observability/configmap"
 )
 
 const (
@@ -85,15 +82,12 @@ func NewController(
 ) *controller.Impl {
 	logger := logging.FromContext(ctx)
 
-	pprof := k8sruntime.NewProfilingServer(logger.Named("pprof"))
-
-	mp, tp := otel.SetupObservabilityOrDie(ctx, "inmemorychannel.dispatcher", logger, pprof)
+	// Use the global meter and tracer providers that were set up by sharedmain
+	mp := otel.GetMeterProvider()
+	tp := otel.GetTracerProvider()
 
 	trustBundleConfigMapLister := filteredconfigmapinformer.Get(ctx, eventingtls.TrustBundleLabelSelector).Lister().ConfigMaps(system.Namespace())
 
-	iw := cmw.(*configmapinformer.InformedWatcher)
-	iw.Watch(o11yconfigmap.Name(), pprof.UpdateFromConfigMap)
-
 	var env envConfig
 	if err := envconfig.Process("", &env); err != nil {
 		logger.Panicw("Failed to process env var", zap.Error(err))
@@ -222,16 +216,7 @@ func NewController(
 			logging.FromContext(ctx).Errorw("Failed stopping inMemoryDispatcher.", zap.Error(err))
 		}
 
-		ctx, cancel := context.WithTimeout(ctx, 10*time.Second)
-		defer cancel()
-
-		if err := mp.Shutdown(ctx); err != nil {
-			logger.Errorw("Error flushing metrics", zap.Error(err))
-		}
-
-		if err := tp.Shutdown(ctx); err != nil {
-			logger.Errorw("Error flushing traces", zap.Error(err))
-		}
+		// Note: Meter and tracer provider shutdown is handled by sharedmain
 	}()
 
 	return impl