Distributed tracing

Distributed tracing follows a single request as it moves through multiple services in a distributed system, recording the time spent in each service and each network hop as a set of connected spans, so engineers can see exactly where in a multi-service call chain time was spent or an error occurred. The complete record of one request's path is called a trace.

Tracing differs from log aggregation and metrics in what question it answers: logs and metrics describe what happened inside one service or one point in time, while a trace reconstructs causality and latency across the entire chain of services a single request touched, which is essential once a request no longer stays within one application. Tools typically propagate a trace ID across service boundaries so spans from different services can be stitched back together.

Tracing matters most in microservice architectures, where a slow or failing request could originate in any of dozens of services, and it is a core data source for application performance monitoring and root-cause investigations tied to a service level indicator breach. A common pitfall is incomplete instrumentation: a trace with gaps where a service was not instrumented hides exactly the segment most likely to be the actual problem.