Databricks preview releases
Databricks regularly releases previews to allow you to evaluate and provide feedback on features before they’re generally available (GA). Previews come in various degrees of maturity, each of which is defined in this article. The following tables list Databricks support options associated with each release type.
For information on accessing preview releases, see Manage Databricks Previews.
Platform releases
Databricks platform features can be released as any of the following types:
Release type | Who can use? | For production? | Interface stable | SLA | Support | Notes |
|---|---|---|---|---|---|---|
Invite only | No | No | No | Engineering team | Available by invite-only and usually for a small set of customers. Previews are OFF by default, and admins can enable (or disable) the preview in workspaces. There is often no accompanying Databricks documentation. | |
Most customers | No | No | No | Engineering team | Available to most customers. The default Beta state can be enabled or disabled. If you do not see a Beta feature in your workspace and want to enable it, contact your Databricks account team. Admins can opt out of Beta by disabling it in workspaces. These features are documented in Databricks documentation and are intended to advance to either Public Preview or GA. A notification appears in the Databricks UI if GA is approaching soon for a feature. | |
Everyone | Yes | Yes | Yes | Support team | Available to all customers. The default preview state can be enabled or disabled. The preview state set during the Private Preview phase does not change for Public Preview. Admins can opt out of the preview by disabling it in workspaces. These features are documented in the Databricks documentation and are also stable and intended to advance to GA. A notification appears in the Databricks UI if GA is approaching soon for a feature. | |
General Availability (GA) | Everyone | Yes | Yes | Yes | Support team | Generally available to all customers. The feature is fully supported and production-ready. |
For information about the platform feature support lifecycle, see Platform feature lifecycle.
Databricks Runtime releases
Databricks Runtime releases can be any of the following types. Previews of Databricks Runtime are always labeled Beta.
Release type | Who can use? | Maturity | For production? | Interface stable | SLA | Support |
|---|---|---|---|---|---|---|
Beta | Everyone | Low-Med | No | No | No | Engineering team |
GA | Everyone | High | Yes | Yes | Yes | Support team |
Features included in a Databricks Runtime can be released at any of the following preview types:
Preview type | Who can use? | Maturity | For production? | Interface stable | SLA | Support | Notes |
|---|---|---|---|---|---|---|---|
Private Preview | Invite only | Low-Med | No | No | No | Engineering team | Features are usually not documented in the public Databricks documentation. The API might change. |
Experimental | Everyone | Low-Med | No | No | No | Engineering team | Features are documented in the public Databricks documentation. The API might change. |
Public Preview | Everyone | Med-High | Yes | Yes | Yes | Support team | Features are documented in the public Databricks documentation. The API will not change, either during the Public Preview or when the feature becomes GA. Most public previews are open to all customers. Some, however, require that you request access. |
For more information about Databricks Runtime releases, including support lifecycle and long-term-support (LTS), see Databricks Runtime support lifecycles.