Distributed Tracing

Follow one request across services using trace IDs and spans.

observe

Detailed Description

Tracing is the best way to understand latency in distributed request paths.

Each service adds spans so teams can see where time and failures occur.

Gateway span
-> UserService span
-> OrderService span
-> DB span
Find slow hop

Pros

Pinpoints bottlenecks across services

Cons

Instrumentation overhead and storage cost

Examples: OpenTelemetry, Jaeger, Zipkin, AWS X-Ray