AWS Certified Machine Learning – SpecialtyMachine Learning Implementation and OperationsEasy
A financial institution is deploying a new fraud detection model to an Amazon SageMaker real-time endpoint. The model is highly sensitive to latency, and any delay in prediction could lead to significant financial losses. The institution wants to ensure that the model remains available and performs optimally even during unexpected spikes in request traffic. Which SageMaker feature should be implemented to address these requirements?
- ASageMaker Endpoint Auto Scaling with target tracking policies.
- BSageMaker Model Monitor for performance metric tracking.
- CSageMaker Feature Store for low-latency feature retrieval.
- DSageMaker Data Capture for real-time inference data logging.
Show answer & explanationAnswer & explanation
Correct answer: A. SageMaker Endpoint Auto Scaling with target tracking policies.
SageMaker Endpoint Auto Scaling allows the endpoint to dynamically adjust the number of instances based on traffic, ensuring high availability and optimal performance during traffic spikes. Target tracking policies allow for scaling based on specific metrics like CPU utilization or a custom SageMaker-provided metric for invocations.
Why the other options are wrong
- B. SageMaker Model Monitor is for detecting data and model quality issues, not for scaling infrastructure.
- C. SageMaker Feature Store is for managing and serving features, not for scaling inference endpoints.
- D. SageMaker Data Capture logs inference requests and responses, which is useful for monitoring but does not scale the endpoint resources.
SageMaker Endpoint Auto Scaling
A SageMaker feature that automatically adjusts the number of inference instances for a real-time endpoint based on predefined scaling policies.
- Ensures high availability and performance during traffic fluctuations.
- Supports target tracking, step scaling, and scheduled scaling policies.
- Helps optimize costs by scaling down during low traffic.
Memory trick: Scale up and down, when traffic goes 'round!