Migrate from OpenCensus library (#1909)

* Update OpenCensus to OpenTelemetry

Change is meant to be transparent to user.

Using OTel SDK for metric composing in: kmetrics, core metrics, resource
group metrics;

Using otlp receiver in otel-agent and otel-collector;

Configured deployment for new ports and component;

Refactor metric composing and recording;

Metric prefix remain the same to minimize breaking change.

Tests updated.

* cleanup

- pass on context in register

* cleanup: restore record.go under kmetrics

* cleanup: restore record.go under resourcegroup metrics

* cleanup: refactor testexporter for readability

* gemini comment: fail fast when record sees metric not initialized

* gemini comment: use context with timeout for exporter graceful shutdown

* fix rebase conflict

* cleanup

- constify conext shutdonw timeout
- restore header
- remove unnecessary metric attribute on exporter setup
- restore test coverage from rebase conflict

* refactor: use lazy initialization for metrics

* copilot: cleanup logs and comments

* refactor testexporter for readability and better error handling

* revert to explicit metric initialization with error check

* cleanup

- restore kmetrics record.go header
- pass on rg manager container name from register

Author: tiffanny29631

cmd/hydration-controller/main.go

@@ -79,20 +79,18 @@ func main() {
 	profiler.Service()
 	ctrl.SetLogger(textlogger.NewLogger(textlogger.NewConfig()))
 
-	// Register the kustomize usage metric views.
-	if err := kmetrics.RegisterKustomizeMetricsViews(); err != nil {
-		klog.Fatalf("Failed to register OpenCensus views: %v", err)
-	}
-
-	// Register the OC Agent exporter
-	oce, err := kmetrics.RegisterOCAgentExporter(reconcilermanager.HydrationController)
+	// Register the OTLP metrics exporter and metrics instruments
+	ctx := context.Background()
+	oce, err := kmetrics.RegisterOTelExporter(ctx, reconcilermanager.HydrationController)
 	if err != nil {
-		klog.Fatalf("Failed to register the OC Agent exporter: %v", err)
+		klog.Fatalf("Failed to register the OTLP metrics exporter: %v", err)
 	}
 
 	defer func() {
-		if err := oce.Stop(); err != nil {
-			klog.Fatalf("Unable to stop the OC Agent exporter: %v", err)
+		shutdownCtx, cancel := context.WithTimeout(context.Background(), kmetrics.ShutdownTimeout)
+		defer cancel()
+		if err := oce.Shutdown(shutdownCtx); err != nil {
+			klog.Fatalf("Unable to stop the OTLP metrics exporter: %v", err)
 		}
 	}()
 

cmd/reconciler-manager/main.go

@@ -182,21 +182,19 @@ func main() {
 	}
 	setupLog.Info("OtelSA controller registration successful")
 
-	// Register the OpenCensus views
-	if err := metrics.RegisterReconcilerManagerMetricsViews(); err != nil {
-		setupLog.Error(err, "failed to register OpenCensus views")
-	}
-
-	// Register the OC Agent exporter
-	oce, err := metrics.RegisterOCAgentExporter(reconcilermanager.ManagerName)
+	// Register the OTLP metrics exporter and metrics instruments
+	ctx := context.Background()
+	oce, err := metrics.RegisterOTelExporter(ctx, reconcilermanager.ManagerName)
 	if err != nil {
-		setupLog.Error(err, "failed to register the OC Agent exporter")
+		setupLog.Error(err, "failed to register the OTLP metrics exporter")
 		os.Exit(1)
 	}
 
 	defer func() {
-		if err := oce.Stop(); err != nil {
-			setupLog.Error(err, "failed to stop the OC Agent exporter")
+		shutdownCtx, cancel := context.WithTimeout(context.Background(), metrics.ShutdownTimeout)
+		defer cancel()
+		if err := oce.Shutdown(shutdownCtx); err != nil {
+			setupLog.Error(err, "failed to stop the OTLP metrics exporter")
 		}
 	}()
 
@@ -205,9 +203,11 @@ func main() {
 	setupLog.Info("starting manager")
 	if err := mgr.Start(ctrl.SetupSignalHandler()); err != nil {
 		setupLog.Error(err, "problem running manager")
-		// os.Exit(1) does not run deferred functions so explicitly stopping the OC Agent exporter.
-		if err := oce.Stop(); err != nil {
-			setupLog.Error(err, "failed to stop the OC Agent exporter")
+		// os.Exit(1) does not run deferred functions so explicitly stopping the OTLP metrics exporter.
+		shutdownCtx, cancel := context.WithTimeout(context.Background(), metrics.ShutdownTimeout)
+		defer cancel()
+		if err := oce.Shutdown(shutdownCtx); err != nil {
+			setupLog.Error(err, "failed to stop the OTLP metrics exporter")
 		}
 		os.Exit(1)
 	}

cmd/reconciler/main.go

@@ -15,6 +15,7 @@
 package main
 
 import (
+	"context"
 	"flag"
 	"fmt"
 	"os"
@@ -140,20 +141,18 @@ func main() {
 		status.EnablePanicOnMisuse()
 	}
 
-	// Register the OpenCensus views
-	if err := ocmetrics.RegisterReconcilerMetricsViews(); err != nil {
-		klog.Fatalf("Failed to register OpenCensus views: %v", err)
-	}
-
-	// Register the OC Agent exporter
-	oce, err := ocmetrics.RegisterOCAgentExporter(reconcilermanager.Reconciler)
+	// Register the OTLP metrics exporter and metrics instruments
+	ctx := context.Background()
+	oce, err := ocmetrics.RegisterOTelExporter(ctx, reconcilermanager.Reconciler)
 	if err != nil {
-		klog.Fatalf("Failed to register the OC Agent exporter: %v", err)
+		klog.Fatalf("Failed to register the OTLP metrics exporter: %v", err)
 	}
 
 	defer func() {
-		if err := oce.Stop(); err != nil {
-			klog.Fatalf("Unable to stop the OC Agent exporter: %v", err)
+		shutdownCtx, cancel := context.WithTimeout(context.Background(), ocmetrics.ShutdownTimeout)
+		defer cancel()
+		if err := oce.Shutdown(shutdownCtx); err != nil {
+			klog.Fatalf("Unable to stop the OTLP metrics exporter: %v", err)
 		}
 	}()
 

docs/custom-metric-filter.md

@@ -171,7 +171,7 @@ To **turn off the custom metrics** in Cloud Monitoring, remove this section:
 ```yaml
 # Remove this entire section
 metrics/cloudmonitoring:
-  receivers: [opencensus]
+  receivers: [otlp]
   processors: [batch, filter/cloudmonitoring, metricstransform/cloudmonitoring, resourcedetection]
   exporters: [googlecloud]
 ```
@@ -183,7 +183,7 @@ To **turn off the report to Cloud Monarch**, remove this section:
 ```yaml
 # Remove this entire section
 metrics/kubernetes:
-  receivers: [opencensus]
+  receivers: [otlp]
   processors: [batch, filter/kubernetes, metricstransform/kubernetes, resourcedetection]
   exporters: [googlecloud/kubernetes]
 ```
@@ -235,7 +235,12 @@ metadata:
 data:
   otel-collector-config.yaml: |
     receivers:
-      opencensus:
+      otlp:
+        protocols:
+          grpc:
+            endpoint: 0.0.0.0:4317
+          http:
+            endpoint: 0.0.0.0:4318
     exporters:
       debug:
         verbosity: detailed
@@ -245,7 +250,7 @@ data:
       batch:
       pipelines:
         metrics:
-          receivers: [opencensus]
+          receivers: [otlp]
           processors: [batch]
           exporters: [debug]
 ```