AI-103 Certification Practice Question #33

Question

You have a Microsoft Foundry project that contains a prompt agent used by a customer support web app.
The agent is invoked from a Python service that does NOT run in the Foundry portal.
You need to implement end-to-end tracing to capture latency breakdowns and exceptions across agent runs.
Which two components can you use? Each correct answer presents a complete solution.
NOTE: Each correct selection is worth one point.

Accepted Answer

End-to-end tracing of a Python service that invokes a Foundry prompt agent but does NOT run in the Foundry portal is a client-side tracing scenario. MS Learn shows you instrument the code with **OpenTelemetry** (via `opentelemetry-sdk` and the `azure-core-tracing-opentelemetry` plugin) to capture latency breakdowns and exceptions across runs, then export those spans to an **Application Insights** resource so they appear in the Foundry portal Traces view and Azure Monitor. Each of these is independently a valid "component you can use," matching the "two correct answers, each a complete solution" phrasing. A Log Analytics workspace is the underlying store behind App Insights but isn't selected as the tracing instrument; the Azure Monitor Agent collects infrastructure/VM logs rather than gen-AI run traces; and Microsoft Sentinel is a security SIEM, not a latency/exception tracing tool — so all three are eliminated. Confidence is slightly reduced because the captured Suggested-Answer field shows only "C" while the two `✓` marks (B and C) and the multi-select wording indicate both.

More AI-103 practice questions