Core concepts

Lakebase is built on a set of features that enable you to develop, test, and scale your database applications efficiently. This section introduces the core concepts that differentiate Lakebase from traditional database systems.

Projects

A Lakebase project is the top-level container for all your database resources. Each project belongs to a Databricks workspace and contains one or more branches, each with its own compute and databases.

Project
└── Branch (e.g., production)
    ├── Compute (read-write)
    └── Database (e.g., databricks_postgres)

Learn more: Projects | Manage projects

Autoscaling

Lakebase automatically adjusts compute resources based on your workload demands. As your application's traffic increases or decreases, compute resources scale up or down within your configured range without manual intervention or downtime.

Key benefits:

  • Automatic adjustment: Resources scale based on actual demand.
  • No downtime: Scaling within your configured range happens without interrupting connections. However, changing the minimum or maximum CU configuration may cause a brief interruption.
  • Cost optimization: Pay only for the resources you actually use.
  • Performance consistency: Maintain responsive performance during traffic spikes.

Learn more: Autoscaling | Configure autoscaling

Scale-to-zero

When your database is idle, Lakebase can automatically scale compute resources down to zero, eliminating costs for unused capacity. When activity resumes, your compute automatically scales back up in seconds.

Key benefits:

  • Zero idle costs: No compute charges when your database is inactive.
  • Instant resume: Databases wake up automatically when accessed.
  • Development-friendly: Ideal for development and staging environments with intermittent usage.
  • Automatic management: No manual intervention required.

Learn more: Scale to zero | Configure scale-to-zero

Database branches

Database branches in Lakebase work similarly to Git branches for your code. You can create instant, isolated branches for development, testing, or experimentation without duplicating data or impacting your production environment.

Key benefits:

  • Instant creation: Branches are created in seconds using copy-on-write technology.
  • Cost-efficient: Only modified data is stored separately, minimizing storage costs.
  • Isolated environments: Test changes safely without affecting your primary database.
  • Easy collaboration: Team members can work on separate branches simultaneously.

Learn more: Database branches | Manage branches

Computes and endpoints

A Lakebase endpoint is the stable connection point your application uses to reach a database. Behind each endpoint, one or more compute instances handle query processing. Your connection string stays the same even as you scale compute resources or add high availability.

A branch typically has one read-write endpoint (your primary compute) and optionally one or more read-only endpoints (read replicas).

Learn more: Computes and endpoints | Manage computes

High availability

High availability pairs a primary read/write compute with one or more secondary compute instances distributed across availability zones. When the primary becomes unavailable, a secondary is automatically promoted and your application continues from the last committed transaction. Your connection string remains unchanged throughout.

Key benefits:

  • Automatic failover: A secondary compute instance is promoted without manual intervention.
  • No data loss: Failover preserves all committed transactions.
  • Stable connection strings: Your application does not need to change its connection configuration after a failover.

Learn more: High availability | Manage high availability

Read replicas

Read replicas are independent read-only computes that perform read operations on the same data as your primary read-write compute. Unlike traditional replicas, Lakebase read replicas don't duplicate data—they read from the same storage layer, enabling instant creation and eliminating additional storage costs.

Key benefits:

  • Horizontal scaling: Distribute read requests across multiple replicas.
  • No data duplication: All replicas read from the same storage.
  • Instant creation: Available in seconds without data copying.
  • Cost-efficient: No additional storage costs, plus autoscaling and scale-to-zero support.

Learn more: Read replicas | Manage read replicas

How they work together

These core concepts work together to create a powerful and flexible database platform:

  1. Projects organize all your database resources and belong to a Databricks workspace.
  2. Autoscaling ensures each branch has the right amount of compute based on demand.
  3. Scale-to-zero reduces costs on idle branches by suspending compute when not in use.
  4. Database branches let you create isolated environments for development and testing without duplicating data.
  5. Computes and endpoints provide stable connection points as you scale or reconfigure your infrastructure.
  6. High availability adds automatic failover across availability zones for production workloads.
  7. Read replicas distribute read traffic across additional compute instances.

This architecture enables you to build and scale database applications with greater flexibility, lower costs, and less operational overhead than traditional database systems.

  • Last updated on