Lab: Instrument a Python application for Jaeger

What you will learn in this TP

• Auto-instrument an application
• Add spans and attributes
• Search by tag

Introduction #

What you will learn in this section

• Auto-instrument an application
• Add spans and attributes
• Search by tag

Context. Your team is developing a Python API with no tracing at all. You instrument it without rewriting the code, then add business context.

Prerequisites: previous lab completed, Python 3  ·  Estimated time: 45 min

A Python API traced end to end.

Setup #

Guided steps, with their solutions.

1. 1. Write the API

   Why? An API that calls an external service produces a trace with several spans.

   Create app.py with a route that calls httpbin.

2. 2. Auto-instrument

   Why? opentelemetry-instrument adds the Flask and requests instrumentations without touching the code.

   Install the OpenTelemetry distribution and run the instrumented application.

3. 3. Generate and observe

   Why? Each call creates a server trace (Flask) with a client span (requests).

   Call the route about twenty times and open a trace from the stock service.

4. 4. Add a business span

   Why? Automatic spans say nothing about the business logic: a span and attributes fill that gap.

   Add a verifier_stock span with the produit.id attribute around the external call.

5. 5. Search by tag

   Why? Find the trace for a specific product, as the support team would.

   Restart the application, make more calls and search Jaeger for traces with the tag produit.id=7 and a minimum duration of 1 s.

Difficulty level : ●●●○○ (3/5)