MLS-C01 Question 254
Select 2You are working as a Machine Learning Engineer and are responsible for monitoring an AWS environment where multiple training jobs are executed on Amazon SageMaker. You want to track resource utilization metrics such as CPU, memory, and GPU usage, as well as identify errors in the training jobs. Which combination of services and features should you use to achieve this?
- A
Amazon CloudWatch to monitor resource utilization metrics and set alarms
- B
AWS CloudTrail to track API calls made by SageMaker
- C
Amazon SageMaker Debugger to capture system resource utilization during training jobs
- D
Amazon Athena to query and analyze training logs stored in Amazon S3
- E
Amazon QuickSight to visualize real-time resource utilization metrics
Show answer and explanation
Correct answers: A, C
Explanation
To effectively monitor an AWS environment running SageMaker training jobs, you need Amazon CloudWatch to track resource utilization and set alarms for anomalies. Additionally, Amazon SageMaker Debugger provides detailed insights into system resource usage and errors encountered during training jobs. These two services provide a comprehensive monitoring solution for machine learning workloads.
- A. Correct.
Amazon CloudWatch is the correct service to monitor resource utilization metrics such as CPU, memory, and GPU usage. It also allows you to set alarms for anomalies.
- B. Incorrect.
AWS CloudTrail is used to track API calls and changes in the AWS environment, but it does not monitor resource utilization or provide training job insights.
- C. Correct.
Amazon SageMaker Debugger is designed to capture and analyze system resource utilization and identify issues during training jobs, making it a suitable choice.
- D. Incorrect.
Amazon Athena is used for querying data in Amazon S3 but is not a tool for real-time monitoring or resource utilization tracking.
- E. Incorrect.
Amazon QuickSight is a data visualization service but is not designed to monitor or visualize real-time resource metrics directly.