MLS-C01 Question 256
Select 3You are a Machine Learning Engineer responsible for deploying a real-time inference application on Amazon SageMaker. The application uses an endpoint that must be monitored to ensure proper performance and availability. Your team also wants to collect detailed logs for debugging and anomaly detection. Which combination of steps should you take to meet these requirements?
- A
Enable Amazon CloudWatch Logs for the SageMaker endpoint.
- B
Configure Amazon SageMaker Model Monitor to track data drift and monitor endpoint metrics.
- C
Enable AWS CloudTrail to log API calls made to the SageMaker endpoint.
- D
Set up AWS Config to track compliance of SageMaker resources.
- E
Use Amazon EventBridge to trigger automated responses to endpoint anomalies.
Show answer and explanation
Correct answers: A, B, C
Explanation
To log and monitor a SageMaker endpoint, you need to enable CloudWatch Logs to collect detailed logs, use Model Monitor to observe metrics and detect issues like data drift, and configure CloudTrail to log API calls for auditing and anomaly detection. AWS Config and EventBridge are useful in other contexts but are not directly applicable to the requirements of this scenario.
- A. Correct.
Enabling Amazon CloudWatch Logs for the SageMaker endpoint allows you to capture detailed logs for debugging and performance monitoring, meeting the logging requirement.
- B. Correct.
SageMaker Model Monitor helps track data drift and endpoint metrics, providing insight into model performance and supporting monitoring needs.
- C. Correct.
AWS CloudTrail records API calls made to the SageMaker endpoint, which is crucial for auditing and anomaly detection.
- D. Incorrect.
AWS Config tracks compliance of resources but does not directly support logging or monitoring the SageMaker endpoint.
- E. Incorrect.
Amazon EventBridge can trigger automated actions but does not itself provide logging or monitoring capabilities. It would complement monitoring tools but is not a direct solution for this scenario.