AZ-400 Question 37
Single answerYou manage a microservices application hosted on Azure Kubernetes Service (AKS). Your operations team requires real-time insights into CPU usage per container, average request latency, and error rates across all microservices. The manager wants these metrics consolidated into a single dashboard with historical data for capacity planning. Which solution should you implement to gather and visualize these metrics from multiple microservices in one place?
- A
Enable Azure Monitor for containers on the AKS cluster and configure a single Log Analytics workspace. Use Kusto (KQL) queries in Azure Monitor to track CPU usage, request latency, and error rates.
- B
Install the Diagnostics extension on each microservice container to send logs directly to an external PostgreSQL database. Then build a custom dashboard in Power BI connected to the database.
- C
Set up multiple Application Insights resources, each dedicated to one microservice. Create an individual dashboard for each resource and switch between dashboards when needed.
- D
Use only the built-in container metrics (e.g., CPU usage) displayed in the Azure portal and manually gather request latency and error rates from logs stored on each container� local filesystem.
Show answer and explanation
Correct answer: A
Explanation
Azure Monitor and Log Analytics are recommended for centralizing metrics and logs across AKS-based microservices. By configuring Azure Monitor for containers, you can send all data�CPU usage, request latency, and error logs�to a single Log Analytics workspace. Using Kusto (KQL) queries allows you to create unified dashboards for operational insights. For more details, see Microsoft� documentation on AKS monitoring with Azure Monitor (https://docs.microsoft.com/azure/azure-monitor/containers).
- A. Correct.
Option 1 is correct because enabling Azure Monitor for containers on AKS and using a single Log Analytics workspace centralizes CPU, latency, and error data. You can then use Kusto queries (KQL) in Azure Monitor to aggregate and visualize these metrics on one dashboard, aligning with best practices for centralized observability.
- B. Incorrect.
Option 2 is incorrect because installing the Diagnostics extension on each container to push logs to PostgreSQL is typically more complex to manage and not the standard approach for Azure-based solutions. You could build a custom dashboard, but it requires significant custom development and maintenance overhead compared to using Azure Monitor.
- C. Incorrect.
Option 3 is incorrect because using multiple Application Insights resources would fragment your data across separate resources. This makes it harder to manage and visualize metrics in one place, defeating the goal of a consolidated operational view.
- D. Incorrect.
Option 4 is incorrect because relying solely on built-in container metrics in the Azure portal for CPU usage and manually gathering logs for latency and errors does not provide comprehensive, unified visibility. You lose the benefits of a central, query-based approach for historical and real-time analysis.