Distributed Tracing: From 100% Error Rate to Root Cause in 60 Seconds

This article demonstrates a step-by-step incident investigation in a microservice architecture, going from detecting a 100% error rate on a recommendations endpoint to identifying the root cause within 60 seconds. Using Uptrace, an OpenTelemetry-native platform, the author walks through service graph analysis, span filtering, timeline reconstruction, and trace-log correlation to pinpoint that the recommendation service cannot reach its downstream dependencies.