Modify the OnDemandWorkflows blueprint for creating workflow environments in a shared VPC - Amazon SageMaker Unified Studio

Modify the OnDemandWorkflows blueprint for creating workflow environments in a shared VPC

In order to support creating workflow environments in a shared VPC setup, where the VPC is in one AWS account and the project and the Amazon Managed Workflows for Apache Airflow (Amazon MWAA) environment are in another AWS account, the domain administrator must complete the following procedure to modify the endpointManagement parameter of the OnDemand Workflows blueprint.

  1. Navigate to the Amazon SageMaker management console at https://console.aws.amazon.com/datazone and use the region selector in the top navigation bar to choose the appropriate AWS Region.

  2. Choose View domains and choose the domain’s name from the list. The name is a hyperlink.

  3. On the domain's details page, navigate to the Project profiles tab.

  4. In the Project profiles tab, choose a project profile, for example, All capabilities. The name of the project profile is a hyperlink.

  5. On the project profile details page, choose OnDemand Workflows blueprint.

  6. In the OnDemand Workflows details page, choose Edit.

  7. In the Blueprint parameters section, choose endpointManagement and then choose Edit.

  8. In the Edit blueprint parameter pop up window, choose Customer in the Value drop-down.

    This value defines whether the VPC endpoints configured for the environment are created and managed by the customer or by Amazon MWAA. If Value is set to SERVICE, Amazon MWAA creates and manages the required VPC endpoints in your VPC. If Value is set to CUSTOMER, you must create and manage the VPC endpoints for your VPC. If you choose to create an environment in a shared VPC, you must set this value to CUSTOMER.

The domain users can then create workflow environments and the domain administrators then can follow the steps and procedures described here to automate deployment of Amazon Amazon MWAA environments using customer-managed endpoints in a VPC.