🔍

Distributed Tracing: Spans, Traces & OpenTelemetry

Medium

Free preview · 2 of 5 steps

Tracing a request across 10 services

In microservices, a single user request might touch 10-20 services. When it's slow, which service is the bottleneck? Logs from each service are isolated — you can't correlate them without a shared identifier.

Distributed tracing propagates a trace ID through every service call as an HTTP header. Each service records its processing time as a span. A tracing backend assembles all spans into a waterfall timeline showing exactly where time was spent.

Packets in motion — click to inspect:

1 / 2

Space to play/pause · ← → to step