Tracing Fundamentals

Overview

Distributed tracing tracks a request across service boundaries, revealing latency contributors and failure points.

Trace Model

• Trace: full request journey
• Span: timed operation within a trace
• Context propagation: trace and span identifiers across calls

sequenceDiagram
    participant Client
    participant API
    participant Payments
    participant DB
    Client->>API: Request
    API->>Payments: Charge
    Payments->>DB: Write transaction
    DB-->>Payments: OK
    Payments-->>API: OK
    API-->>Client: Response

Practical Example

traceparent: 00-4bf92f3577b34da6a3ce929d0e0e4736-00f067aa0ba902b7-01

Production Example

During latency incidents, traces quickly identify whether delays originate in application code, database calls, or external APIs.

Common Pitfalls

• Missing context propagation between services
• Sampling too aggressively and losing critical traces
• Capturing traces without linking to logs and metrics

Key Takeaways

• Tracing provides request-level visibility in distributed systems.
• Propagation and sampling strategy determine trace usefulness.
• Correlating signals across telemetry types accelerates diagnosis.