AZ-400 Question 294
Select 2You manage an Azure-based microservices architecture with multiple APIs and background jobs. The system is instrumented with Azure Application Insights to capture performance and usage metrics. Recently, the DevOps team has been receiving notifications about performance spikes in production. You need to perform a detailed analysis to pinpoint the root cause. Which two approaches should you recommend to effectively analyze the metrics from instrumentation and identify performance bottlenecks?
- A
Use the Performance blade in Application Insights to examine request and dependency call durations.
- B
Create and customize Azure Monitor Workbooks to visualize trends, correlate logs, and drill down into anomalies.
- C
Ignore exception telemetry and focus exclusively on CPU usage metrics from the Azure portal� VM blade.
- D
Schedule your production system to automatically scale out every time a performance spike is detected, ignoring further investigation.
Show answer and explanation
Correct answers: A, B
Explanation
To analyze metrics effectively in a microservices environment instrumented with Application Insights, you should leverage both the Performance blade (for request and dependency insights) and Azure Monitor Workbooks (to correlate performance data with logs). This combined approach follows best practices outlined in Microsoft� Azure Monitor and Application Insights documentation, ensuring that spikes or anomalies are thoroughly investigated to identify their true source.
- A. Correct.
Option 1 is CORRECT because the Performance blade in Application Insights provides insights into response times, dependency call durations, and bottlenecks. It helps you pinpoint which operations are running slowly and why, facilitating root cause analysis.
- B. Correct.
Option 2 is CORRECT because Azure Monitor Workbooks let you visualize metrics, cross-reference them with logs, and perform deeper analysis. By customizing Workbooks, you can correlate performance spikes with exception logs or other telemetry data to identify root causes more effectively.
- C. Incorrect.
Option 3 is INCORRECT because focusing solely on CPU usage and ignoring exceptions neglects other vital performance insights. Exception telemetry often reveals underlying errors or bottleneck issues, so discarding it can lead to incomplete or misleading conclusions.
- D. Incorrect.
Option 4 is INCORRECT because automatically scaling out without thorough investigation only addresses symptoms, not the underlying problem. While scaling can temporarily mitigate performance issues, proper analysis is necessary to discover and resolve the actual root cause.