MLS-C01 Question 261
Select 3Your organization has implemented an Amazon SageMaker model endpoint for real-time predictions. To comply with governance policies, you need to monitor and log API calls to the endpoint, track unusual activity, and generate automated alerts when a high number of requests are made to the endpoint within a short period. Which combination of AWS services should you use to achieve this?
- A
Enable AWS CloudTrail to log API calls to the SageMaker endpoint
- B
Use Amazon CloudWatch Alarms to monitor the request count metric for the SageMaker endpoint
- C
Enable Amazon SageMaker Model Monitor to track API requests to the SageMaker endpoint
- D
Set up Amazon CloudWatch Logs to store and analyze logs generated by CloudTrail
- E
Use AWS Config to monitor configuration changes to the SageMaker endpoint
Show answer and explanation
Correct answers: A, B, D
Explanation
To monitor and log API calls, AWS CloudTrail is essential for recording API activities on the SageMaker endpoint. For detecting unusual activity such as high request counts, Amazon CloudWatch Alarms can monitor relevant metrics and send notifications. Additionally, CloudTrail logs can be analyzed in depth using Amazon CloudWatch Logs. Together, these services provide a comprehensive solution for tracking and alerting on unusual activity on a SageMaker endpoint.
- A. Correct.
Correct. AWS CloudTrail is used to log API calls for AWS services, including SageMaker endpoints. This ensures governance and compliance by tracking who accessed the service and what actions were performed.
- B. Correct.
Correct. Amazon CloudWatch Alarms can monitor SageMaker endpoint metrics, such as the number of requests, and trigger alerts when thresholds are breached.
- C. Incorrect.
Incorrect. While SageMaker Model Monitor is used for monitoring data quality and model drift, it does not track API requests or unusual activity on the endpoint.
- D. Correct.
Correct. Amazon CloudWatch Logs can store logs generated by CloudTrail, enabling detailed analysis of API call patterns and identifying anomalies.
- E. Incorrect.
Incorrect. AWS Config is used for tracking configuration changes and ensuring compliance with configuration policies, but it does not monitor API activity or request metrics for SageMaker endpoints.