Overview of serverless workspaces

Note

Currently, serverless workspaces are only available for accounts created with Databricks express setup.

A serverless workspace is a Databricks workspace deployment in your account that comes pre-configured with serverless compute and default storage to provide a completely serverless experience.

Compute in serverless workspaces

All workloads in serverless workspaces are powered by serverless compute. For information about the types of serverless compute available on Databricks, see Connect to serverless compute.

If you need custom compute resources, you must create a new workspace in your account console that supports these resources. Select Create workspace, then under Advanced options select Custom compute and storage. See Create a workspace using the AWS Quickstart (Recommended).

Storage in serverless workspaces

Serverless workspaces come with default storage. Default storage provides you with a fully-managed storage location that you can use in your workspace. For more information about your workspace’s default storage, see Default storage in serverless workspaces.

Using data in your cloud storage

You can also create a connection between a cloud storage account and your serverless workspace. Among other things, this allows you to ingest raw data into your workspace, create and read managed tables in secure cloud storage, create external tables, and read and write unstructured data.

For instructions, see Connect to cloud object storage and services using Unity Catalog.

Create a serverless workspace

If you want to create an additional serverless workspace, you can do so from your Databricks account console. New workspaces are serverless by default.

  1. In the Workspace tab of the account console, click Create workspace.

  2. Give your workspace a name and select a region.

  3. Click Create.

Serverless workspace limitations

Because serverless workspaces use serverless compute, they share the same limitations. See Serverless compute limitations for a full list of limitations.

Default storage on serverless workspaces includes the following limitations:

  • Data assets in default storage can be shared using Delta Sharing within the same Databricks account, but not to other Databricks accounts or open Delta Sharing clients.

  • External readers and writers cannot access default storage.

  • Default storage is not compatible with advanced networking configurations or customer-managed keys.