How are Lakebase databases and tables governed through Unity Catalog?

Summary

• You govern a Lakebase database in Unity Catalog by registering it as a read-only catalog that mirrors its Postgres schemas, tables, and views, so the operational store lands under the same catalog, permission, lineage, and audit model as your analytical tables.
• Lakebase is generally available and runs standard Postgres (built on Neon, the serverless Postgres engine Databricks acquired in 2025), so this brings transactional data into Unity Catalog without moving it out of Postgres.
• Governance runs on two planes: Unity Catalog controls discovery and query access to the registered catalog through serverless SQL warehouses, while direct Postgres connections keep using native Postgres roles and privileges.
• Synced tables move data between the lakehouse and Lakebase through managed pipelines in snapshot, triggered, or continuous mode, and Unity Catalog records lineage for every sync.
• The payoff is one governance surface: register once, discover in Catalog Explorer, grant access with SQL, and trace end-to-end lineage and user-level audit across operational and analytical data.

How are Lakebase databases and tables governed through Unity Catalog?

You govern a Lakebase database in Unity Catalog by registering it as a read-only catalog that mirrors its Postgres schemas, tables, and views. That single step puts the operational store under the same governance model as your analytical tables: one place for discovery, grants, lineage, and audit. Lakebase is generally available and runs standard Postgres, so registration brings transactional data into Unity Catalog without moving it out of Postgres or copying it into the lakehouse. The design goal is a single governed view of operational and analytical data, not two disconnected permission systems.

What registering a Lakebase database in Unity Catalog does

Registration creates a read-only catalog in Unity Catalog that reflects the live Postgres structure (schemas, tables, and views). You start it from Catalog Explorer, and the database then shows up as a catalog other teams can find and browse. A few properties define how it behaves:

• The catalog is a governed, read-only representation. You still change data by writing to the Postgres database directly through Lakebase, not through Unity Catalog.
• One Lakebase database maps to one Unity Catalog catalog, so you register each database you want to expose.
• Once registered, you get centralized discovery in Catalog Explorer, federated queries that join Lakebase data with lakehouse tables, and access from AI/BI Dashboards and Databricks Apps under one model.

How governance works across the two planes

Governing an operational database next to analytics means being clear about which layer enforces what. With Lakebase and Unity Catalog, analytical access flows through the registered catalog while operational access stays in Postgres.

Governance dimension	How it works with Lakebase and Unity Catalog
Registration and discovery	Register the database from Catalog Explorer; it appears as a read-only catalog that teams can search and browse.
Access control (analytical)	The person who registers the database becomes the catalog owner; grant USE CATALOG and SELECT so others can query it. Metastore admins can manage all registered catalogs.
Access control (operational)	Direct connections to the Postgres endpoint authenticate with native Postgres roles and privileges, independent of Unity Catalog grants.
Query engine	Serverless SQL warehouses can query a registered Lakebase catalog; Pro and Classic SQL warehouses cannot.
Data movement	Synced tables use managed pipelines in snapshot, triggered, or continuous mode to move data between the lakehouse and Lakebase.
Lineage	Unity Catalog automatically tracks lineage for each sync, so you can trace what data moved between Lakebase and the lakehouse.
Audit	Unity Catalog captures user-level audit logs of access across both operational and analytical assets.

How Databricks approaches unified operational and analytical governance

Lakebase is a fully managed, serverless Postgres database built on Neon, the serverless Postgres engine Databricks acquired in 2025, and it is integrated with the lakehouse and Unity Catalog. Registering a Lakebase database brings the operational store into the same governance surface as your Delta tables, so ownership, grants, discovery, lineage, and audit all live in one place rather than in a separate database console.
Data stays connected without hand-built pipelines. Synced tables hydrate a Lakebase database from Unity Catalog tables and keep it current, and a change data feed moves updates back toward the lakehouse, with Unity Catalog recording lineage for each hop. Because Unity Catalog governs both sides, you can define who sees a table once and trace a value from raw lakehouse data through to the operational tables serving apps and agents. For the transactional workload itself, Lakebase keeps standard Postgres roles and privileges, so application access control does not change when you register the database. See the Lakebase general availability announcement for the current product scope.

FAQs

Is Lakebase generally available?

Yes. Lakebase is a generally available, fully managed, serverless Postgres database built on Neon and integrated with the Databricks Platform, including Unity Catalog.

Does registering a Lakebase database in Unity Catalog let me write to it from Databricks?

No. The registered catalog is read-only. You query it through Unity Catalog and serverless SQL warehouses, and you change data by writing to the Postgres database directly through Lakebase.

Do Unity Catalog grants replace Postgres permissions?

No. They work on separate planes. Unity Catalog grants govern query access to the registered catalog, while direct application connections to Postgres continue to use native Postgres roles and privileges.

How does data stay consistent between Lakebase and the lakehouse?

Synced tables run managed pipelines in snapshot, triggered, or continuous mode, and Unity Catalog records lineage for each sync so you can audit what moved and when.

Which SQL warehouses can query a registered Lakebase catalog?

Serverless SQL warehouses. Pro and Classic SQL warehouses do not support querying a registered Lakebase catalog.