Announcing Lakebase Change Data Feed (CDF)

Summary

Lakebase Change Data Feed (CDF) is available in Public Preview to reduce the manual effort of moving data from operational databases into downstream systems. The feed is enabled once for all tables in a project, stored and governed in Unity Catalog Managed Tables, and readable by engines, models, and agents without separate extraction pipelines. From one shared feed, teams can build streaming pipelines with SDP, create materialized views with DBSQL, or compute and store embeddings with Agent Bricks, while consumers remain isolated from the primary operational workload. The announcement positions Lakebase as the native Bronze layer in a medallion architecture, complementing Synced Tables and providing governance and lineage across the data lifecycle.

Context

Extracting Change Data Capture from OLTP databases requires teams to configure connectors, monitor replication states, mitigate performance impacts, and track errors through disconnected tools. The source says this creates brittle, ungoverned, and high-friction work for each source and destination, particularly for rapid data branching.

Approach / What changed

Lakebase provides a native Change Data Feed stored and governed in Unity Catalog Managed Tables. Enabling it once applies the feed to all tables in a project, allowing downstream consumers such as SDP, DBSQL, and Agent Bricks to read the same feed while remaining isolated from the primary operational workload.

Takeaways

• Lakebase CDF is in Public Preview and can be enabled in less than a minute for all tables within a project.
• The same feed supports streaming pipelines with SDP, materialized views with DBSQL, and embedding computation and storage with Agent Bricks.
• Lakebase CDF places operational data in the native Bronze layer of the medallion architecture, with governance and lineage through Unity Catalog.

Related reading