← All updates Databricks Update One-Pager Navy version
databricks Platform · Updated Aug 18, 2025

Lakeflow Spark Declarative Pipelines

Enhanced control over pipeline execution identity.

What it does

This update allows users to set a run-as identity for Spark Declarative Pipelines, enabling the use of service principals instead of user accounts for executing pipeline updates and managing table ownership. This feature enhances security and reliability, especially when the original pipeline owner is no longer available. Users can configure this setting through the pipeline monitoring page or editor, ensuring that the pipeline continues to function seamlessly.

Why it matters

By allowing the use of service principals as run-as identities, organizations can improve the security and stability of their automated workloads. This change mitigates risks associated with user account deactivation and ensures that pipelines remain operational, which is crucial for maintaining data accessibility and governance. It also aligns with best practices for managing automated processes in a cloud environment.

Key capabilities
Service Principal Support
Pipelines can now run under a service principal, enhancing security.
Ownership Management
Easily change the ownership of tables created by the pipeline.
Permission Control
Manage permissions for the run-as user to ensure proper access.
User Deactivation Recovery
Recover pipeline functionality when the original owner is deactivated.
Prerequisites
CAN_MANAGE permissions on the pipeline
CAN_USE role on the service principal if applicable
Workspace access permission for the run-as user
Recommended use cases
Automated workloads requiring stable identities.Recovery of pipelines after user deactivation.Best practices for managing data access in cloud environments.
Limitations · remember
!The run-as change can interrupt an active update.
!Insufficient permissions on the service principal can cause update failures.
Architecture
Lakeflow Pipelines → Service Principal Management → Unity Catalog
Get started
1
Access Pipeline Settings
Open the pipeline monitoring page and select the pipeline.
2
Edit Run-As User
Click on Settings and then Edit next to Run as.
3
Select Identity
Choose either your own account or a service principal.
4
Save Changes
Click Save to apply the new run-as identity.
Key takeaway
The ability to set a service principal as the run-as user for pipelines enhances security and operational reliability.
2025-08-18-set-the-run-as-user-for-spark-declarative-pipelines-on-lakeflow Generated by the Databricks Updates Agent · browse all updates