Domain-assets linking
Before running the domain-assets linking script, crawl your AWS SageMaker Unified Studio assets to catalog them into Atlan.
When AWS SageMaker Unified Studio assets are cataloged into Atlan, the domain information associated with Projects, Published Assets, and Subscribed Assets is captured as metadata. However, this domain information isn't automatically linked to the corresponding Data Domains in Atlan's governance hierarchy.
The domain-assets linking script bridges this gap by programmatically establishing the relationship between cataloged AWS SageMaker Unified Studio assets and their respective Data Domains.
Why is this needed?
The AWS SageMaker Unified Studio connector extracts domain metadata from your AWS environment and stores it as custom attributes on the cataloged assets. While this metadata is valuable, it exists as plain text and isn't connected to Atlan's Data Domain governance structure.
By running the domain-assets linking script, your AWS SageMaker Unified Studio assets are properly governed under the correct Data Domains in Atlan.
What the script does
The script automates the following:
- Discovers all Data Domains defined in your Atlan instance
- Identifies AWS SageMaker Unified Studio Project assets that contain domain information in their custom attributes
- Links each Project asset to its corresponding Atlan Data Domain based on the domain name
- Propagates the domain linkage to all Published Assets and Subscribed Assets associated with each Project
Benefits
- Unified domain-based governance across your data catalog
- Streamlined data discovery within domain boundaries
- Consistent governance hierarchy between AWS SMUS and Atlan
Get started
For detailed instructions on prerequisites, environment setup, and running the script, see the AWS SageMaker Unified Studio - Atlan Domain Linking Script documentation on GitHub.
See also
- Set up AWS SageMaker Unified Studio: Configure AWS authentication for SMUS integration
- Crawl AWS SageMaker Unified Studio: Configure and run the crawler to extract metadata from SMUS