MLA-C01 Question 316
Select 3You are managing a machine learning model deployed on an Amazon SageMaker endpoint. The model experiences significant variation in traffic throughout the day, with high demand during business hours and low demand at night. You want to configure the endpoint to automatically scale based on traffic patterns to optimize costs and performance. Which steps should you take to achieve this?
- A
Create an Application Auto Scaling policy targeting the SageMaker endpoint to automatically adjust the number of instances based on request metrics.
- B
Enable Amazon CloudWatch monitoring for the SageMaker endpoint and set up a CloudWatch alarm to trigger scaling actions.
- C
Specify a scaling schedule based on time intervals, such as scaling up during business hours and scaling down at night.
- D
Configure an Elastic Load Balancer (ELB) to handle the dynamic traffic and automatically scale SageMaker endpoints.
- E
Use the SageMaker Inference Recommender to automatically set scaling policies for your endpoint.
Show answer and explanation
Correct answers: A, B, C
Explanation
To meet scalability requirements with SageMaker endpoints, you can use Application Auto Scaling to adjust the number of instances based on demand or time-based schedules. CloudWatch monitoring enables you to track metrics and trigger scaling actions effectively. These steps ensure that your endpoint scales efficiently to handle traffic fluctuations while optimizing costs. However, ELBs and SageMaker Inference Recommender are not directly related to setting up auto-scaling policies for SageMaker endpoints.
- A. Correct.
Correct: Application Auto Scaling integrates with SageMaker to automatically adjust the number of instances based on metrics like InvocationsPerInstance, ensuring scalability and cost efficiency.
- B. Correct.
Correct: CloudWatch monitoring is essential for tracking the endpoint's performance metrics and triggering scaling actions based on traffic demand.
- C. Correct.
Correct: A time-based scaling schedule is suitable for predictable traffic patterns, such as business hours, and helps optimize resource utilization.
- D. Incorrect.
Incorrect: While ELBs are used for distributing traffic, they are not used for auto-scaling SageMaker endpoints. SageMaker endpoints use Application Auto Scaling for this purpose.
- E. Incorrect.
Incorrect: SageMaker Inference Recommender helps optimize endpoint configurations but does not directly create or manage auto-scaling policies.