AcmeRetail uses the SageMaker Model Registry to manage production-ready models. The ML team wants to configure automated deployment to a staging endpoint whenever a new model package version is approved in the registry. Which design will reliably trigger an automated deployment when a model package version is set to Approved?

The correct option leverages Amazon EventBridge to reliably detect the specific state change of a SageMaker model package. When a model package version is approved, EventBridge can trigger a downstream process, such as an AWS Lambda function or a Step Functions state machine. This process then executes the necessary SageMaker API calls (CreateModel, CreateEndpointConfig, and UpdateEndpoint/CreateEndpoint) to deploy the approved model to a staging endpoint. The other options are incorrect because: The Model Registry does not have an "automatic deployment checkbox" that pushes approved models to endpoints. S3 PutObject events are for objects being placed in S3, not for model package approval status changes, and SageMaker endpoints do not automatically update from SNS topic notifications in this manner. CloudWatch Events is the older name for EventBridge, but CodeDeploy is for deploying application code, not for directly managing SageMaker model deployments based on model package approvals.