An Apache Spark-based analytics platform optimized for Azure.
There is no supported way to bypass or override a system deny assignment created by Azure Databricks while the workspace that owns it still exists. Deny assignments are evaluated after role assignments; if a deny applies at the scope of the VMs or their resource group, it blocks deployments and extension operations even for users with Owner at a higher scope.
Because the deny assignment is at the subscription/resource group scope and explicitly blocks Microsoft.Resources/deployments/* operations, any remediation task or direct extension deployment that relies on ARM deployments to those VMs will fail as long as that Databricks workspace (and its system deny assignment) remains in place.
The only supported remediation is to change the Databricks-side configuration that created the deny assignment (if such an option exists in that specific Databricks feature) or to delete the Databricks workspace, which removes the system deny assignment. If the workspace must remain in use and its deny assignment cannot be reconfigured, then the remediation that requires ARM deployments/extensions at that scope cannot be performed.
Opening a Microsoft support ticket can confirm the specific deny assignment and its impact, but support cannot remove or override a system deny assignment that is owned by an active managed service resource such as Azure Databricks.
References: