Manage high availability

info

Lakebase Autoscaling is available in the following regions: us-east-1, us-east-2, us-west-2, ca-central-1, sa-east-1, eu-central-1, eu-west-1, eu-west-2, ap-south-1, ap-southeast-1, ap-southeast-2.

Lakebase Autoscaling is the latest version of Lakebase with autoscaling compute, scale-to-zero, branching, and instant restore. For feature comparison with Lakebase Provisioned, see choosing between versions.

This guide covers enabling and managing high availability for your Lakebase endpoints. For background on how high availability works and how secondary compute instances differ from standalone read replicas, see High availability.

Enable high availability

note

If your endpoint has scale to zero enabled, disable it before enabling high availability. To disable it in the UI, set Scale to zero to Off in the edit compute drawer.

After creating a project, click the primary compute link from the project dashboard to open the edit compute drawer.

Set Compute type to High availability, then choose a Configuration under High availability:

• 2 (1 primary, 1 secondary),
• 3 (1 primary, 2 secondaries),
• or 4 (1 primary, 3 secondaries) total compute instances.

Lakebase provisions secondary compute instances in different availability zones. Once all compute instances are active, the endpoint has automatic failover.

Configure read-only access to secondary compute instances

Allow access to read-only compute instances controls whether secondary compute instances serve read traffic via the -ro connection string.

1. On the Computes tab, click Edit on the primary compute.
2. Under High availability, check or uncheck Allow access to read-only compute instances.
3. Click Save.

warning

With only one secondary compute instance and read access enabled, all read traffic on the -ro connection string is interrupted during a failover until a replacement is added. For resilient read access, configure two or more secondary compute instances with read access enabled.

Change the number of secondary compute instances

1. On the Computes tab, click Edit on the primary compute.
2. Under High availability, choose a new Configuration from the dropdown (2, 3, or 4 total compute instances).
3. Click Save.

note

To disable high availability, set Compute type back to Single compute. This removes all secondary compute instances and your endpoint returns to a single-compute configuration.

View high availability status and roles

The Computes tab shows every compute instance in your high availability configuration with its current role, status, and access level.

Column	Values
Role	Primary, Secondary
Status	Starting, Active
Access	Read/Write (primary), Read-only (secondary compute instance with access enabled), Disabled (secondary compute instance without read access)

The primary compute header also shows the endpoint ID, autoscaling range, and secondary count (e.g. 8 ↔ 16 CU · 3 secondaries).

Obtain connection strings

Click Connect on the primary compute to open the connection details dialog. The Compute dropdown lists both connection options for your high availability endpoint.

Compute option	Connection string	Use for
Primary (name) ● Active	{endpoint-id}.database.{region}.databricks.com	All writes and read/write connections
Secondary (name) ● Active RO	{endpoint-id}-ro.database.{region}.databricks.com	Read offload to secondary compute instances

The -ro connection string is only available when Allow access to read-only compute instances is enabled.

For complete connection string reference, see Connection strings.

Next steps

• High availability — concepts, failover behavior, and best practices
• Read replicas — standalone read replicas for additional read capacity without high availability
• Connection strings