A data engineer must enable data scientists to access ETL scripts in AWS Glue directly from Amazon SageMaker notebooks inside a VPC. The data scientists should be able to run the Glue job and then trigger SageMaker training. Which combination of steps should the engineer take? (Choose three.)

To allow SageMaker notebooks within a VPC to access and run AWS Glue ETL scripts, an AWS Glue development endpoint must be created within that same VPC. This endpoint provides a network path for the SageMaker notebooks to interact with Glue. The SageMaker notebooks themselves need to be configured to connect to this Glue development endpoint, enabling data scientists to directly execute Glue jobs from their notebook environment. Finally, an IAM policy and an associated IAM role are crucial for granting the SageMaker notebooks the necessary permissions to interact with AWS Glue, including starting and monitoring jobs, ensuring secure and authorized access. Creating a SageMaker development endpoint is incorrect; SageMaker notebooks are the interface. Creating standalone SageMaker notebooks without connecting them to the Glue development endpoint would not achieve the desired integration. Attaching a decryption policy is irrelevant to this scenario.