MLA-C01 Question 409
Select 3You are a Machine Learning Engineer designing a system to monitor and automate model retraining on AWS. You want to use AWS CloudTrail to log and trigger retraining activities whenever a SageMaker model endpoint is updated. Which combination of steps should you take to accomplish this?
- A
Enable AWS CloudTrail logging to monitor SageMaker API calls for model updates.
- B
Configure a CloudWatch Events rule to trigger a Lambda function based on CloudTrail logs for SageMaker model updates.
- C
Set up an EventBridge rule to directly react to SageMaker endpoint update events.
- D
Modify the SageMaker endpoint using the AWS Management Console to automatically trigger retraining.
- E
Use the AWS SDK to query CloudTrail logs periodically and check for SageMaker endpoint update events.
Show answer and explanation
Correct answers: A, B, C
Explanation
To automate model retraining using AWS services, CloudTrail must log SageMaker API activities, such as endpoint updates. These logs can be monitored through EventBridge (formerly CloudWatch Events), which can directly react to specific SageMaker events or trigger workflows like invoking a Lambda function. This combination ensures a real-time and automated retraining pipeline.
- A. Correct.
Correct. Enabling AWS CloudTrail logging is necessary to monitor API activity, including SageMaker model updates.
- B. Correct.
Correct. CloudWatch Events (now part of EventBridge) can be used to trigger workflows, such as invoking a Lambda function, based on patterns found in CloudTrail logs.
- C. Correct.
Correct. EventBridge can directly react to SageMaker events, including endpoint updates, streamlining the retraining process.
- D. Incorrect.
Incorrect. Modifying the SageMaker endpoint using the Management Console does not inherently trigger retraining activities; it is a manual process.
- E. Incorrect.
Incorrect. While the AWS SDK can query CloudTrail logs, this approach is inefficient and not recommended for real-time or automated workflows.