Exam AWS Certified Machine Learning - Specialty topic 1 question 186 discussion

BCF; This tutorial doc says so: https://docs.aws.amazon.com/glue/latest/dg/dev-endpoint-tutorial-sage.html

BCF is correct

"C" is definitely wrong SageMaker cannot be created by using the AWS Glue development endpoint. Explanation: This option is not correct. SageMaker notebooks are created using the SageMaker console or APIs, not through AWS Glue development endpoints.

BDF is correct

when creating notebook why do we need Glue development endpoint? it should be D

A. YES - By requirement, the notebook must be in a VPC B. NO - Data is already in S3, we do not need to know it was made with AWS Glue C. NO - Data is already in S3 thanks to AWS Glue, no runtime relationship with SageMaker D. YES - need to create the Notebooks at some point E. NO - no need to decrypt, it is about ACL F. YES - notebooks need to be able to read S3

Creating a SageMaker development endpoint in the data science team's VPC will allow the data science team to access the ETL scripts and the AWS Glue job from within their VPC. Creating an IAM policy and an IAM role for the SageMaker notebooks will allow the data science team to access the ETL scripts and the AWS Glue job with the appropriate permissions. Creating SageMaker notebooks by using the SageMaker console will allow the data science team to easily create and manage the SageMaker notebooks.

In the AWS Glue console, choose Dev endpoints to navigate to the development endpoints list. Select the check box next to the name of a development endpoint that you want to use, and on the Action menu, choose Create SageMaker notebook. Fill out the Create and configure a notebook page as follows: Enter a notebook name. Under Attach to development endpoint, verify the development endpoint. Create or choose an AWS Identity and Access Management (IAM) role.