Text decision tree
Adapted from Chapter 125 of Decision Trees: Frameworks for Technology Choices. Sources reviewed 2026-11-07.
START: Separate platform enforcement from enterprise ownership and stewardship.
Q1. Databricks is a production platform?
Yes -> Use Unity Catalog for its scope. Permissions, lineage and data / AI assets. [Q2]
No -> Otherwise keep the enterprise or cloud catalog primary and govern any Databricks consumption explicitly. [STOP]
Q2. A wider enterprise catalog exists?
Yes -> Keep enterprise stewardship there. Map it to platform enforcement. [Q3]
No -> Otherwise begin with clearly assigned ownership in Unity Catalog and define the wider governance roadmap. [Q3]
Q3. Other engines need governed access?
Yes -> Validate supported access paths. Iceberg REST support; check feature limits. [Q4]
No -> Otherwise keep the implementation focused on the Databricks scope while documenting the boundary. [Q4]
Q4. Legacy permission debt remains?
Yes -> Migrate in controlled waves. Test identities, grants and workloads. [END]
No -> Otherwise move to ongoing ownership, lineage, access review and audit checks. [END]
END: Define owners, access rules, lineage and audit evidence. Migrate legacy permissions in controlled waves.
AI-generated chapter illustration
Conceptual illustration for Choose the governance boundary
The illustration depicts a broad governance control plane. Its platform and enterprise panels are a visual metaphor; they do not establish that the same policy is enforced in every external engine. The graph explicitly requires checking each access path.
The illustration was generated in BookWriter. Its visual arrangement is interpretive; use the explicit nodes and routes in the graph and text tree for the decision.