Introducing Governance Hub: Intelligent, account-level governance over your Databricks estate
Databricks introduces Governance Hub, a centralized account-level tool for monitoring data health, AI usage, and costs across multiple cloud platforms, now available in beta.
Databricks has launched Governance Hub, a new account-level governance tool designed to address challenges faced by platform and governance teams managing large Databricks estates. The tool provides a centralized view of data health, AI usage, and spending across AWS, Azure, and GCP, replacing scattered insights from system tables and third-party tools. It aims to simplify governance for teams responsible for hundreds of workspaces and multiple regions, offering prioritized recommendations and drill-down capabilities without workspace constraints.
The Data page in Governance Hub offers a high-level overview of assets, including the percentage of tables tagged, owned, and classified, with detailed drill-downs to identify missing metadata. Users can manage governed tags, create policies, and edit values directly within the interface. The tool replaces manual processes for tracking data quality and ownership, enabling teams to quickly identify undocumented or unhealthy metastores and close compliance gaps.
Access Insights provides a principal-centric view to determine what users, groups, or service principals can access across the entire account, including direct grants and inherited permissions. This feature simplifies complex access audits, allowing teams to debug access issues, review vendor group permissions, or identify assets owned by departing employees in seconds. The tool consolidates data from multiple sources into a single interface, reducing the need for manual queries and cross-referencing.
The AI page in Governance Hub tracks token consumption, model activity, per-user spend, and guardrail coverage for models managed through Unity AI Gateway, including both Databricks-hosted and externally-hosted models. The Cost page displays 30-day spend, month-to-date trends, and daily averages, with a focus on tagged spend to highlight costs invisible to chargeback processes. Integration with Genie allows users to ask natural-language questions about governance data and receive actionable recommendations without writing SQL.