databricks Databricks Blog ·

How to Implement Secure Per-Viewer Access in Databricks AI/BI Dashboards

blogsecuritydatabricksarchitect
announcement

Databricks details a reference design pattern for implementing secure per-viewer row and column-level access in embedded AI/BI Dashboards. This pattern leverages `__aibi_external_value`, Unity Catalog, and identity provider groups to control data visibility. It allows organizations to serve diverse internal and external audiences from a single dashboard, eliminating the need for multiple dashboard versions. This approach ensures robust data isolation, default-deny access, and protection for direct SQL queries.

Notes (1)
  • Reference Pattern for Secure Per-Viewer Access

    This Databricks reference pattern outlines how to implement granular row-level and column-level security for embedded AI/BI Dashboards. It combines `__aibi_external_value` with Unity Catalog row filters and column masks, and identity provider groups to enforce per-viewer data access and mask sensitive information. This ensures a single dashboard can securely serve both external partners and internal teams with tailored data views.

Read the original announcement →

https://www.databricks.com/blog/beyond-embedding-how-secure-aibi-dashboards-every-viewer

Related releases