0Pricing
gRPC & High Performance APIs · レッスン

OpenTelemetryによるトレーシング

OpenTelemetryを統合して分散トレーシングを実現し、サービス間のエンドツーエンドのリクエストフローを可視化します。

「OpenTelemetryによるトレーシング」はCoddyKit上の無料gRPC & High Performance APIsレッスンです。 これはレッスン2/4です。 下記で完全なレッスンを無料で読むことができます。その後、ブラウザ内の組み込みコードエディタと24時間対応のAIチューターでハンズオン演習できます。 これはgRPC & High Performance APIs学習パスの一部であり、ウェブとCoddyKitアプリ全体で進捗が同期されます。 gRPC & High Performance APIsコースには全4レッスンが含まれています。

このレッスンの一部はまだ翻訳されておらず、英語で表示されています。

Microservices & The Tracing Need

Modern applications often consist of many small, interconnected services, known as microservices. This architecture offers flexibility but can make debugging complex interactions a real challenge.

When a user request travels through several services, pinpointing where a delay or error occurred becomes like finding a needle in a haystack.

Following the Request Path

Distributed tracing helps you visualize the entire journey of a request as it flows through different services in your system.

  • It links all operations related to a single request.
  • It shows the time taken by each part of the request.
  • This helps identify performance bottlenecks and errors across service boundaries.

OpenTelemetry: The Observability Standard

OpenTelemetry (OTel) is an open-source project that provides a standardized way to collect telemetry data from your applications. This includes traces, metrics, and logs.

Instead of vendor-specific SDKs, OTel offers a single set of APIs, SDKs, and tools to instrument your services, making your observability data portable.

Traces Built from Spans

In OpenTelemetry, a trace represents the complete execution path of a request through your system. It's like the entire story of that request.

A trace is composed of one or more spans. Each span represents a single operation or unit of work within that trace, like a step in the story. Spans have a start time, end time, and metadata.

Propagating Trace Context

For spans to form a complete trace across different services, they need to be linked. This is done through context propagation.

When a service calls another, it passes along a "trace context" (often in HTTP/2 headers for gRPC). This context tells the receiving service which trace and parent span its new operations belong to, creating a parent-child relationship between spans.

OTel Tracing Essentials

To implement tracing, you'll work with these core OpenTelemetry components:

  • TracerProvider: Manages Tracer instances and configures span processors and exporters.
  • Tracer: Creates Span objects.
  • SpanProcessor: Processes spans before they are exported (e.g., batching).
  • Exporter: Sends collected spans to a backend (like a console, Jaeger, or OTLP collector).

OTel Setup: TracerProvider & Exporter

Let's set up a basic OpenTelemetry TracerProvider in Go. We'll use a simple stdouttrace exporter to print trace data to the console, so you can see the spans being generated.

This snippet initializes the global TracerProvider and ensures all spans are flushed when the application exits.

Run this code to see the basic setup:

package main

import (
	"context"
	"log"
	"os"
	"time"

	"go.opentelemetry.io/otel"
	"go.opentelemetry.io/otel/attribute"
	"go.opentelemetry.io/otel/exporters/stdout/stdouttrace"
	"go.opentelemetry.io/otel/sdk/resource"
	sdktrace "go.opentelemetry.io/otel/sdk/trace"
	semconv "go.opentelemetry.io/otel/semconv/v1.24.0"
	"go.opentelemetry.io/otel/trace"
)

// initTracerProvider initializes an OpenTelemetry TracerProvider
// with a stdout exporter for demonstration.
func initTracerProvider() *sdktrace.TracerProvider {
	// Create stdout exporter to print traces to the console
	exporter, err := stdouttrace.New(stdouttrace.WithPrettyPrint())
	if err != nil {
		log.Fatalf("failed to create stdout exporter: %v", err)
	}

	// Create a new TracerProvider with the exporter and a service name
	tp := sdktrace.NewTracerProvider(
		sdktrace.WithBatcher(exporter), // Batch spans for efficiency
		sdktrace.WithResource(resource.NewWithAttributes(
			semconv.SchemaURL,
			semconv.ServiceName("my-simple-service"),
			attribute.String("environment", "development"),
		)),
	)
	otel.SetTracerProvider(tp) // Set the global TracerProvider

	return tp
}

func main() {
	tp := initTracerProvider()
	defer func() {
		if err := tp.Shutdown(context.Background()); err != nil {
			log.Fatalf("Error shutting down tracer provider: %v", err)
		}
	}()

	// Get a tracer
	tracer := otel.Tracer("my-app-tracer")

	// Create a root span
	ctx, span := tracer.Start(context.Background(), "main-operation")
	defer span.End()

	log.Println("Hello from main-operation with tracing enabled!")
	time.Sleep(100 * time.Millisecond) // Simulate some work

	// Create a child span
	_, childSpan := tracer.Start(ctx, "sub-operation")
	defer childSpan.End()
	childSpan.SetAttributes(attribute.String("task", "processing"))
	log.Println("Doing some sub-operation...")
	time.Sleep(50 * time.Millisecond) // Simulate more work

	log.Println("Trace data will be printed on shutdown.")
}

Auto-Tracing gRPC Calls

To automatically trace gRPC requests, OpenTelemetry provides interceptors. These are middleware functions that wrap gRPC calls on both the client and server sides.

  • Server interceptor: Starts a new span for incoming requests, linking it to the trace context from the client.
  • Client interceptor: Injects the current trace context into outgoing requests, allowing propagation to the server.

Here's how you'd apply them to your gRPC server and client:

package main

import (
	"context"
	"log"
	"net"
	"time"

	"google.golang.org/grpc"
	"google.golang.org/grpc/codes"
	"google.golang.org/grpc/status"

	// OpenTelemetry imports
	"go.opentelemetry.io/otel"
	"go.opentelemetry.io/otel/attribute"
	"go.opentelemetry.io/otel/exporters/stdout/stdouttrace"
	"go.opentelemetry.io/otel/sdk/resource"
	sdktrace "go.opentelemetry.io/otel/sdk/trace"
	semconv "go.opentelemetry.io/otel/semconv/v1.24.0"

	// gRPC OpenTelemetry instrumentation
	"go.opentelemetry.io/contrib/instrumentation/google.golang.org/grpc/otelgrpc"

	// For this example, we'll define a simple service in-line
	pb "google.golang.org/grpc/examples/helloworld/helloworld" // Using a standard example proto
)

// initTracerProvider initializes an OpenTelemetry TracerProvider
// with a stdout exporter for demonstration.
func initTracerProvider() *sdktrace.TracerProvider {
	exporter, err := stdouttrace.New(stdouttrace.WithPrettyPrint())
	if err != nil {
		log.Fatalf("failed to create stdout exporter: %v", err)
	}

	tp := sdktrace.NewTracerProvider(
		sdktrace.WithBatcher(exporter),
		sdktrace.WithResource(resource.NewWithAttributes(
			semconv.SchemaURL,
			semconv.ServiceName("grpc-otel-demo"),
			attribute.String("environment", "development"),
		)),
	)
	otel.SetTracerProvider(tp)
	return tp
}

// helloServer is a simple gRPC server for demonstration
type helloServer struct {
	pb.UnimplementedGreeterServer
}

func (s *helloServer) SayHello(ctx context.Context, in *pb.HelloRequest) (*pb.HelloReply, error) {
	log.Printf("Received: %v", in.GetName())
	// Simulate some work
	time.Sleep(50 * time.Millisecond)
	return &pb.HelloReply{Message: "Hello " + in.GetName()}, nil
}

func main() {
	// 1. Initialize OpenTelemetry TracerProvider
	tp := initTracerProvider()
	defer func() {
		if err := tp.Shutdown(context.Background()); err != nil {
			log.Fatalf("Error shutting down tracer provider: %v", err)
		}
	}()

	// 2. Start gRPC Server with OpenTelemetry Interceptor
	lis, err := net.Listen("tcp", ":50051")
	if err != nil {
		log.Fatalf("failed to listen: %v", err)
	}
	s := grpc.NewServer(
		grpc.UnaryInterceptor(otelgrpc.UnaryServerInterceptor()), // Apply OTel server interceptor
	)
	pb.RegisterGreeterServer(s, &helloServer{})
	log.Println("Server listening on :50051")
	go func() {
		if err := s.Serve(lis); err != nil {
			log.Fatalf("failed to serve: %v", err)
		}
	}()
	defer s.Stop()

	// Allow server to start
	time.Sleep(1 * time.Second)

	// 3. Create gRPC Client with OpenTelemetry Interceptor
	conn, err := grpc.Dial(
		"localhost:50051",
		grpc.WithInsecure(), // For simplicity, use insecure
		grpc.WithBlock(),
		grpc.WithUnaryInterceptor(otelgrpc.UnaryClientInterceptor()), // Apply OTel client interceptor
	)
	if err != nil {
		log.Fatalf("did not connect: %v", err)
	}
	defer conn.Close()
	c := pb.NewGreeterClient(conn)

	// 4. Make a gRPC call
	log.Println("Making gRPC call...")
	ctx, cancel := context.WithTimeout(context.Background(), time.Second)
	defer cancel()

	r, err := c.SayHello(ctx, &pb.HelloRequest{Name: "CoddyKit User"})
	if err != nil {
		log.Fatalf("could not greet: %v", err)
	}
	log.Printf("Greeting: %s", r.GetMessage())

	// Give some time for spans to be processed and exported
	time.Sleep(1 * time.Second)
	log.Println("Trace data should now be visible in console.")
}

Reading the Trace Story

After running the previous example, you'll see a detailed JSON output in your console. This output represents the trace data.

  • Look for "TraceID": This unique ID links all spans belonging to the same request.
  • Each entry is a "Span": It has a "SpanID", "ParentSpanID" (if it's not the root), name, start/end times, and attributes.
  • The hierarchy of spans (parent-child) shows the flow of execution.

This "story" helps you understand how long each step took and which service was responsible.

Enriching Spans with Attributes

While auto-instrumentation provides basic spans, you often need to add more specific information to your traces. Span attributes (key-value pairs) allow you to do this.

You can record details like user IDs, request parameters, or specific business logic outcomes directly on a span. This makes debugging and analysis much more effective.

Use span.SetAttributes() to add custom attributes.

package main

import (
	"context"
	"log"
	"time"

	"go.opentelemetry.io/otel"
	"go.opentelemetry.io/otel/attribute"
	"go.opentelemetry.io/otel/exporters/stdout/stdouttrace"
	"go.opentelemetry.io/otel/sdk/resource"
	sdktrace "go.opentelemetry.io/otel/sdk/trace"
	semconv "go.opentelemetry.io/otel/semconv/v1.24.0"
	"go.opentelemetry.io/otel/trace"
)

// initTracerProvider initializes an OpenTelemetry TracerProvider
func initTracerProvider() *sdktrace.TracerProvider {
	exporter, err := stdouttrace.New(stdouttrace.WithPrettyPrint())
	if err != nil {
		log.Fatalf("failed to create stdout exporter: %v", err)
	}
	tp := sdktrace.NewTracerProvider(
		sdktrace.WithBatcher(exporter),
		sdktrace.WithResource(resource.NewWithAttributes(
			semconv.SchemaURL,
			semconv.ServiceName("custom-attributes-demo"),
		)),
	)
	otel.SetTracerProvider(tp)
	return tp
}

func main() {
	tp := initTracerProvider()
	defer func() {
		if err := tp.Shutdown(context.Background()); err != nil {
			log.Fatalf("Error shutting down tracer provider: %v", err)
		}
	}()

	tracer := otel.Tracer("my-custom-tracer")

	// Start a root span
	ctx, span := tracer.Start(context.Background(), "process-order")
	defer span.End()

	// Add custom attributes to the root span
	span.SetAttributes(
		attribute.String("user.id", "user123"),
		attribute.Int("order.id", 45678),
		attribute.Bool("is_premium_user", true),
	)
	log.Println("Processing order with custom attributes...")
	time.Sleep(100 * time.Millisecond)

	// Create a child span
	_, childSpan := tracer.Start(ctx, "database-query")
	defer childSpan.End()
	childSpan.SetAttributes(
		attribute.String("db.system", "postgres"),
		attribute.String("db.statement", "SELECT * FROM orders WHERE id=45678"),
	)
	log.Println("Executing database query...")
	time.Sleep(50 * time.Millisecond)

	log.Println("Check the console output for custom attributes in the spans!")
	time.Sleep(500 * time.Millisecond) // Ensure spans are exported
}

Tracing Concepts Check

You've learned about distributed tracing and how OpenTelemetry helps. Let's test your understanding.

Tracing with OpenTelemetry Recap

In this lesson, you learned how distributed tracing helps navigate the complexities of microservices by showing the full journey of a request.

We explored OpenTelemetry as the standard for collecting traces, understanding core concepts like traces, spans, and context propagation. You saw how to set up an OTel TracerProvider and how gRPC interceptors make instrumenting your services straightforward. Finally, we covered enriching your spans with custom attributes for deeper insights.

With OpenTelemetry, you gain crucial visibility into your gRPC application's behavior!

よくある質問

「OpenTelemetryによるトレーシング」レッスンは無料ですか？

はい。「OpenTelemetryによるトレーシング」の完全なテキストはこのウェブで無料で読めます。インタラクティブに演習し（組み込みコードエディタと24時間対応のAIチューター）、gRPC & High Performance APIsコースの残りをアンロックするには、CoddyKit PROにアップグレードしてください。 gRPC & High Performance APIsコースには全4レッスンが含まれています。

「OpenTelemetryによるトレーシング」で何を学びますか？

OpenTelemetryを統合して分散トレーシングを実現し、サービス間のエンドツーエンドのリクエストフローを可視化します。 ブラウザで直接実行するハンズオンコードでgRPC & High Performance APIsを演習し、24時間対応のAIチューターがレッスンを進める中での質問に答えます。

gRPC & High Performance APIsを始めるのに経験は必要ですか？

事前経験は必要ありません。CoddyKitのgRPC & High Performance APIsは初級者から上級者向けに構成されているため、ここから始めるか最初から始めて、自分のペースで進むことができます。 これはレッスン2/4です。

「OpenTelemetryによるトレーシング」レッスンにはどのくらい時間がかかりますか？

ほとんどのCoddyKitレッスンは約5～10分かかります。各レッスンはコンパクトでインタラクティブなので、着実に進歩し、ウェブとアプリ全体で正確に前回の場所から再開できます。

このgRPC & High Performance APIsレッスンでコードを書いて実行できますか？

はい。すべてのgRPC & High Performance APIsレッスンに組み込みコードエディタが含まれているため、ブラウザでリアルコードを書いて実行し、即座のAIフィードバックを取得できます。ローカル設定は不要です。

このコースのすべてのレッスン

  1. gRPCインタラクションのロギング
  2. OpenTelemetryによるトレーシング
  3. gRPCメトリクスの監視
  4. ヘルスチェックとReadiness Probe
← gRPC & High Performance APIsに戻る