Lakebase Postgres Introduces Cost Optimization Features and Sync Modes
- Published
- Sep 30, 2026 — 00:00 UTC
Lakebase Postgres Introduces Cost Optimization Features and Sync Modes
Lakebase now offers a 25% discount on always-on pricing when the scale to zero feature is disabled. This fully managed Postgres database is part of the Databricks Intelligence Platform, which facilitates managed syncs between Lakehouse and Lakebase.
Lakebase provides three sync modes: Snapshot, Triggered, and Continuous. The Snapshot mode, which is recommended when more than 10% of the source data changes, can be up to 10x more efficient compared to other modes. The cost for snapshot storage is set at $0.090/GB-month, while Point-in-time restore (PITR) storage costs $0.200/GB-month. PITR allows for a configurable recovery window of 2 to 30 days, growing with both write activity and the length of the recovery window.
The database supports 10,000 concurrent client connections through a connection pooler and features a default autoscale range of 8 to 16 compute units (CUs). Autoscaling actively adjusts compute resources in response to varying activity levels, enhancing performance and cost efficiency.
Automated snapshots can be scheduled on a daily, weekly, or monthly basis, with the scale to zero feature enabled by default after 24 hours. The introduction of Declarative Automation Bundles (DABs) provides a method for managing Lakebase effectively, while Materialized Views define the data necessary for applications.
These updates follow Databricks' recent enhancements in data integration tools, emphasizing the platform's commitment to optimizing operational costs for users managing large datasets.
By Turing Wire Newsdesk · Sep 30, 2026 · How we work →
Summarised from Databricks Blog's original report by the Turing Wire Newsdesk. Read the original for the full story.
Source: Databricks Blog
