Follow one request through several services — X-Ray and Distributed tracing

Your goal

Use a trace to find where a slow request spends its time.

Why this matters

A page calls an API, a worker, and a database. You will connect the request’s pieces, inspect the slow part, and understand why a trace may show only part of the traffic.

Words you will use

A trace follows a request across components. A span or segment describes one part of that work. Instrumentation records timing and context. Sampling selects which requests are traced.

How to proceed

Read each situation and choose an answer. The lesson explains the result. Try again if needed, then continue through all three decisions and finish the tour.

1. Connect the pieces

Each service has separate logs, but the team cannot tell which events belong to the same request. Choose the missing tracing behavior.

What should travel between supported components?

Trace context, with suitable instrumentation.

Instrumentation records work and passes trace context so supported calls can be linked. Configure each relevant component and verify propagation. A service diagram alone does not create a complete request trace.

Related work can be connected into one request view.

Trace context → instrumented calls → linked segments

2. Inspect where time went

A request takes four seconds, and its trace shows most of that time waiting on one database call. Choose the next investigation.

Where should you look first?

The slow call and its dependency context.

Use timing to choose a useful next question, then inspect the dependency’s behavior with relevant metrics and logs. A long segment points to where time was spent; it does not by itself explain the root cause.

The team has a focused place to investigate.

Slow segment → dependency evidence → tested explanation

3. Understand gaps in the view

A teammate cannot find a trace for one reported request. The system samples traffic, and some components may lack instrumentation.

What is the accurate explanation?

Sampling and instrumentation affect what is captured.

Check sampling rules, trace propagation, and component coverage before concluding that nothing happened. Use other evidence such as logs and metrics when needed. Tracing is one view of the system, with deliberate limits.

The investigation accounts for missing or partial traces.

Coverage + sampling → available traces; combine other evidence

Practice, then check your understanding

In the interactive tour, make each decision before seeing the outcome. Wrong answers receive coaching, and replay gives you a fresh attempt without duplicating earned XP.

Read the official AWS documentation for service behavior and production details.