AI-102 Question 61
Select 3You are managing a Cognitive Services resource in Azure that supports a production application. You need to monitor the performance and usage of this resource to ensure it meets business requirements. Which of the following steps should you take to monitor the AI resource effectively?
- A
Enable diagnostic logs for the Cognitive Services resource and configure them to export to Azure Monitor Logs.
- B
Use the Azure AI Health Dashboard to monitor service outages and issues related to the Cognitive Services resource.
- C
Configure alerts in Azure Monitor to notify you of high latency or request failures.
- D
Manually review the API usage logs by downloading them from the Cognitive Services portal monthly.
- E
Integrate Application Insights to collect telemetry data for client-side usage of the Cognitive Services API.
Show answer and explanation
Correct answers: A, B, C
Explanation
Monitoring an Azure AI resource requires a combination of tools and configurations to ensure continuous and effective tracking of performance and usage. Enabling diagnostic logs, leveraging the Azure AI Health Dashboard, and configuring alerts in Azure Monitor are key practices that provide real-time insights and proactive monitoring. Manually reviewing logs is not scalable, and Application Insights is more suited for client-side monitoring than server-side metrics.
- A. Correct.
Enabling diagnostic logs and exporting them to Azure Monitor Logs helps track performance metrics and identify potential issues. This is an effective monitoring strategy.
- B. Correct.
The Azure AI Health Dashboard provides real-time updates on service outages or issues, making it essential for proactive monitoring.
- C. Correct.
Configuring alerts in Azure Monitor allows you to respond quickly to performance issues such as high latency or request failures.
- D. Incorrect.
Manually reviewing API usage logs monthly is not an effective or scalable monitoring practice. Automated solutions are more suitable for production environments.
- E. Incorrect.
While Application Insights is useful for monitoring client-side telemetry, it is not the best tool for monitoring server-side performance metrics of the Cognitive Services resource.