AZ-400 Question 302
Select 2You maintain a microservices-based application hosted on Azure App Services, with each microservice configured to use Application Insights. Users report occasional high latency, and you suspect one microservice is causing a bottleneck. You want to perform distributed tracing to identify the root cause. Which two actions should you perform to effectively diagnose latency across all microservices using Application Insights?
- A
Use the Application Map feature in the Azure portal to identify the microservice with elevated latency.
- B
Configure correlation headers (e.g., Request-Id and Request-Context) for incoming and outgoing calls in each microservice.
- C
Disable automatic dependency tracking to reduce noise and overhead in telemetry data.
- D
Create separate Application Insights resources for each microservice to isolate logs and minimize correlation complexity.
- E
Enable only trace-level logs for each microservice and rely on Log Analytics queries to piece together request flows.
Show answer and explanation
Correct answers: A, B
Explanation
To troubleshoot latency in a distributed microservices environment, you must use Application Insights features that correlate requests and dependencies across services. The most critical practices include enabling the Application Map for a real-time overview of service-to-service relationships and ensuring correlation headers are preserved in requests. By doing so, requests across microservices can be traced end-to-end, highlighting bottlenecks. Refer to Microsoft� official documentation on distributed tracing with Application Insights for further guidance on correlation headers and Application Map usage.
- A. Correct.
Correct. The Application Map in Application Insights provides an end-to-end view of how requests flow between your microservices. It highlights which resources have higher latency or errors, making it a key tool for diagnosing performance bottlenecks.
- B. Correct.
Correct. Ensuring correlation headers (Request-Id, Request-Context, and correlation-conversation-id where relevant) are preserved allows Application Insights to associate requests and dependencies across services. This practice enables true distributed tracing and end-to-end visibility.
- C. Incorrect.
Incorrect. Disabling automatic dependency tracking would prevent Application Insights from collecting critical data about service-to-service calls. You need dependency details to understand how latency flows through each microservice.
- D. Incorrect.
Incorrect. While separate resources can be preferred in some scenarios, splitting telemetry for each microservice can make correlation more difficult. Using a shared Application Insights resource (or at least ensuring consistent correlation across resources) is often recommended for simpler distributed tracing.
- E. Incorrect.
Incorrect. Trace-level logs are useful, but enabling them alone without proper correlation headers or the Application Map would make it difficult to see an end-to-end transaction. You need a combination of correlation and the Application Map for effective distributed tracing.