Distributed Tracing with Sleuth and Zipkin [Java Spring Boot Mastery Series – Part 18]

In a microservices architecture, tracing a single request that flows across multiple services is essential for debugging, monitoring, and performance analysis.

Spring Cloud Sleuth and Zipkin help trace and visualize requests across services.

🔍 What is Distributed Tracing?

Distributed tracing allows you to track a request as it moves through multiple microservices. It helps answer:

• Which services were involved?
• How much time did each take?
• Where are the bottlenecks or failures?

🛠️ Key Tools

Tool	Purpose
Sleuth	Automatically adds trace & span IDs to logs
Zipkin	Aggregates traces and provides a UI to analyze them

🧱 Architecture

[Client] → [Service A] → [Service B] → [Service C]
                  \_________ Sleuth sends data to Zipkin _________/

⚙️ Step-by-Step Setup

🔹 1. Add Dependencies (Each Microservice)

In pom.xml:

<dependency>
  <groupId>org.springframework.cloud</groupId>
  <artifactId>spring-cloud-starter-sleuth</artifactId>
</dependency>

<dependency>
  <groupId>org.springframework.cloud</groupId>
  <artifactId>spring-cloud-starter-zipkin</artifactId>
</dependency>

🔁 Make sure your spring-cloud.version matches your Spring Boot version.

🔹 2. Basic Configuration

In application.yml:

spring:
  application:
    name: service-a

  zipkin:
    base-url: http://localhost:9411
    enabled: true

  sleuth:
    sampler:
      probability: 1.0 # 100% of requests will be traced (for development)

🔹 3. Run Zipkin Server

Option 1: Use Docker

docker run -d -p 9411:9411 openzipkin/zipkin

Zipkin UI: http://localhost:9411

Option 2: Run from JAR

Download the Zipkin JAR from https://zipkin.io and run:

java -jar zipkin-server.jar

🔗 How Sleuth Works (Behind the Scenes)

• Trace ID: A unique identifier for the entire request flow.
• Span ID: Represents a single unit of work (e.g., a method or service call).
• Parent Span ID: Refers to the calling span.

Sleuth injects these into logs and headers automatically:

[TraceId=abc123] [SpanId=def456] Calling Service B

HTTP Headers:

Header	Meaning
X-B3-TraceId	ID of the trace
X-B3-SpanId	ID of the current span
X-B3-ParentSpanId	Parent span
X-B3-Sampled	Whether it should be traced

🧪 Test: Simulate a Multi-Service Call

• Service A calls Service B
• Both log their trace and span info

Controller in Service A

@RestController
@RequestMapping("/a")
public class ServiceAController {

    @Autowired
    private RestTemplate restTemplate;

    @GetMapping("/callB")
    public String callB() {
        return restTemplate.getForObject("http://localhost:8082/b/hello", String.class);
    }
}

Controller in Service B

@RestController
@RequestMapping("/b")
public class ServiceBController {

    @GetMapping("/hello")
    public String sayHello() {
        return "Hello from B!";
    }
}

Logs in both services will include the same trace ID.

🔧 RestTemplate + Sleuth (Auto-propagation)

Sleuth automatically injects tracing headers in RestTemplate and WebClient.

Make sure RestTemplate is a Spring Bean:

@Bean
public RestTemplate restTemplate() {
    return new RestTemplate();
}

🔎 View in Zipkin UI

• Visit http://localhost:9411
• Search by service name
• Explore request timelines and call chains

🛡️ Best Practices

Tip	Why It Matters
Use lower sample rate in production	Reduce performance overhead
Always use beans for HTTP clients	Enables auto-injection of trace headers
Combine with ELK/Grafana	For complete observability stack

🧰 Alternatives

Tool	Description
Jaeger	OpenTracing-compatible distributed tracer
OpenTelemetry	Modern standard, supports multiple exporters

📦 Summary

• Sleuth adds trace IDs to logs and headers.
• Zipkin visualizes traces across services.
• Helps debug, monitor, and optimize microservice chains.

➡️ Next Up: Part 19 – Spring Cloud Config & Centralized Configuration Management