Microsoft Certified: Azure Developer Associate (AZ-204)Monitor, troubleshoot, and optimize Azure solutionsHard
A company uses Azure API Management (APIM) to expose its backend APIs. They are experiencing intermittent performance issues and higher-than-expected latency for certain API calls. They need to identify the specific policies or backend service calls within APIM that contribute most to the latency. Which APIM feature or Azure Monitor integration should they leverage to diagnose this issue effectively?
- AAPIM Developer Portal's testing console
- BAzure Application Insights integration with APIM
- CAzure Network Watcher's Connection Troubleshoot
- DAPIM built-in metrics in Azure Monitor
Show answer & explanationAnswer & explanation
Correct answer: B. Azure Application Insights integration with APIM
Azure Application Insights, when integrated with APIM, provides end-to-end transaction tracing, including detailed metrics and durations for each policy execution step and backend call within an API request. This allows for pinpointing performance bottlenecks precisely.
Why the other options are wrong
- A. The Developer Portal's testing console is for functional testing and basic response viewing, not for in-depth performance diagnostics of policy execution.
- C. Azure Network Watcher focuses on network connectivity and routing, not on the internal processing times of an API Management gateway or its policies.
- D. APIM built-in metrics provide aggregate data like total requests, average latency, and errors, but not granular, per-request tracing of individual policy execution times.
APIM Integration with Application Insights
Connecting Azure API Management to Azure Application Insights to enable detailed monitoring, distributed tracing, and performance analysis of API requests flowing through APIM.
- Provides end-to-end transaction visibility.
- Traces individual policy execution times.
- Helps identify performance bottlenecks within the APIM pipeline and backend.
Memory trick: App Insights traces every step, showing where the API time is lost.