OpenTelemetry

Introduction

OpenTelemetry is the vendor-neutral standard for traces, metrics and logs. Spring Boot 3.x auto-instruments HTTP, JDBC and Kafka — export to Jaeger, Grafana Tempo or Datadog with zero code changes.

Informative example

Manual span + Spring Boot config:

@Service
public class OrderService {
    private final Tracer tracer;

    public Order create(OrderRequest req) {
        Span span = tracer.spanBuilder("createOrder").startSpan();
        try (var scope = span.makeCurrent()) {
            span.setAttribute("order.sku", req.sku());
            return repo.save(new Order(req));
        } finally { span.end(); }
    }
}

# application.yml
management.tracing.sampling.probability=1.0
management.otlp.tracing.endpoint=http://localhost:4318/v1/traces

Best practices

• Propagate trace context across HTTP and Kafka headers.
• Sample in production (1-10%) — 100% sampling is expensive.
• Correlate traceId with log MDC for unified debugging.

Purpose of this lesson

Master OpenTelemetry so you can apply it confidently in production Java code, technical interviews, and code reviews.

Step-by-step explanation

1. Understand the core idea behind OpenTelemetry.
2. Walk through the runnable example and tweak it in the playground.
3. Apply the pattern in a small Spring Boot or CLI exercise of your own.
4. Re-read the common mistakes and interview Q&A to lock the concept in.

Interactive workflow diagram

Debugging tips

• Read the full stack trace — Java's exception messages name the offending class and line.
• Reproduce in the smallest possible main() method before fixing in the real app.
• Use IntelliJ's debugger breakpoints and 'Evaluate Expression' rather than scattering System.out.

Optimization strategies

• Profile before optimizing — JFR (Java Flight Recorder) and async-profiler reveal real hotspots.
• Prefer immutable data and stream pipelines over hand-rolled loops when readability matters.
• Reach for the right JDK collection (ArrayList vs LinkedList vs ArrayDeque) before writing custom data structures.

Enterprise example

Grafana's LGTM stack (Loki, Grafana, Tempo, Mimir) ingests OpenTelemetry natively — one SDK, full observability.

Interview questions & answers

Summary

In this lesson you learned OpenTelemetry — the concept, syntax, a runnable example, and the production pitfalls to avoid. Apply it in the playground before moving on.