Dev MaxxingAdvancedReliability and Observability90 sec

Trace

A trace follows a single request as it moves through multiple services in a system, showing how long each step along the way took.

Mark as completed

Prerequisite: Metric

What matters

In a distributed system made of many services, a trace is often the only way to see exactly which specific step in a request's journey caused a slowdown or failure.

Example

A trace for one slow checkout request might show it spent 10ms in the API, 480ms waiting on a payment service, and 5ms writing to the database.

Common misconception

A trace is not the same as a log; a trace specifically follows one request's full path across multiple services, rather than recording isolated events separately.

Recall question

What can a trace reveal about a slow request that separate logs from each service might not?

Keywords and related concepts

tracedistributed tracingrequest pathlog