Question

A team has a SageMaker Pipeline in a development account that trains a model, evaluates it, and registers approved versions to the SageMaker Model Registry. The company mandates that all production endpoints run in a separate production account, and that the deployment to production must be triggered automatically whenever a model version is approved in the registry. The team wants to minimize custom code and avoid manual copying of model artifacts between accounts. Which approach best meets these requirements?

Accepted Answer

Configure an EventBridge rule in the dev account for Model Registry approval-status-change events, forward the event to an EventBridge bus in the production account, and trigger a deployment pipeline (CodePipeline/CloudFormation) that reads the shared S3 model artifacts and creates the production endpoint.

Answer

Add a deployment step to the training pipeline that uses boto3 to assume a role in the production account and directly call CreateEndpoint from within the pipeline's processing step.

Answer

Export the approved model artifacts to a local file, have an engineer manually upload them to the production account's S3 bucket, and run a CloudFormation template to create the endpoint.

Answer

Configure the production account to poll the dev Model Registry every hour with a scheduled Lambda function, and deploy any newly approved version it finds.