Question

A machine learning team has a real-time SageMaker endpoint serving a fraud-detection model. They have trained a new model version and want to route only 10% of live traffic to it while keeping 90% on the existing model, so they can monitor CloudWatch metrics before fully committing. They want to accomplish this within a single endpoint without provisioning separate endpoints. Which approach meets these requirements?

Accepted Answer

Deploy both models as two production variants behind the same endpoint and assign initial variant weights of 9 and 1 to split traffic 90/10.

Answer

Create a multi-model endpoint and load both models, letting SageMaker automatically balance traffic between them.

Answer

Deploy the new model to a separate endpoint and use Route 53 weighted routing to send 10% of requests to it.

Answer

Use SageMaker batch transform to score 10% of records with the new model and the rest with the old model.