databricks Platform · Updated Jul 24, 2026

Dedicated Group Clusters

Enhanced security and efficiency for collaborative workloads.

Public Preview Open docs ↗release note ↗

What it does

Dedicated group clusters allow users to share compute resources securely while running workloads that require specific languages and APIs not supported by standard access mode. When users connect to these clusters, their permissions are automatically adjusted to match the group's permissions, ensuring secure collaboration. This feature requires Unity Catalog and Databricks Runtime 15.4 or higher, and it streamlines the management of shared resources within a group setting.

Why it matters

This update enhances operational efficiency and security for teams working with sensitive data or specialized workloads. By enabling dedicated group clusters by default for compliance-focused workspaces, organizations can ensure better governance and risk management while fostering collaboration among data scientists and analysts. This is particularly beneficial for teams using machine learning and advanced analytics, as it simplifies resource sharing without compromising security.

Key capabilities

Secure Resource Sharing

Users can securely share compute resources while running specialized workloads.

Permission Down-Scoping

User permissions are automatically adjusted to match the group's permissions during cluster sessions.

Support for Advanced Workloads

Supports languages and APIs such as Databricks Runtime for ML and RDD APIs that standard access mode does not.

Audit Logging

Records user and group identities for actions performed on group clusters for compliance and auditing.

Prerequisites

Workspace must be enabled for Unity Catalog.

Must use Databricks Runtime 15.4 or above.

Assigned group must have CAN MANAGE permissions on a workspace folder.

Recommended use cases

Data science teams collaborating on machine learning projects.Analysts sharing compute resources for data processing tasks.Research groups working with sensitive datasets requiring strict access controls.

Limitations · remember

!Jobs created using the API and SDK cannot be assigned group access.

!Lineage system tables do not record identity metadata for workloads on group clusters.

!The %run command uses user permissions instead of group permissions.

!Creating and accessing model serving endpoints is not supported.

!File and folder deletion is not supported in group clusters.

Architecture

Databricks Workspace → Unity Catalog → Dedicated Group Clusters → Compute Resources

Get started

Create Compute Resource

In your Databricks workspace, go to Compute and click Create compute.

Select Access Mode

Expand the Advanced section and select Dedicated from the Access mode dropdown.

Assign Group

In the Single user or group field, select the group you want assigned to this resource.

Configure Settings

Set the desired compute settings and click Create.

Key takeaway

Dedicated group clusters enhance security and collaboration for teams by allowing shared compute resources with tailored permissions.

2026-07-24-ff1e038f Generated by the Databricks Updates Agent · browse all updates