Databricks workspace stuck in Deleting state for 3+ days, cannot delete managed resource group

Question

My Databricks workspace "dbw-medico-prd-eus2-001" in resource group "rg-medico-prd-eus2-001" has been stuck in Deleting state for over 3 days. The managed resource group "databricks-rg-dbw-medico-prd-eus2-001-pgv256sq2fkfa" cannot be deleted due to system deny assignments created by Databricks. I cannot modify any resources inside the managed RG. How can I force delete both resource groups? Subscription ID: 67bbee80-5ecf-4835-a4b7-0d9b62cfc19d

Answer 1

Hi Julian David Agredo Cortes,
Thanks for sharing the details. Since your Azure Databricks workspace has been in a Deleting state for over 3 days, this indicates the deletion process is stuck on the backend rather than an issue with your configuration.

Azure Databricks creates a managed resource group (MRG) with system deny assignments to protect its underlying infrastructure. These protections prevent any manual deletion or modification of resources inside the managed resource group. Under normal conditions, once the workspace deletion completes successfully, the managed resource group and its associated locks are automatically removed.

In your case, the backend cleanup has not completed, which is why the workspace remains in Deleting and the managed resource group cannot be deleted.

It’s important to clarify that there is no supported customer‑side mechanism to bypass these system deny assignments or manually force‑delete the managed resource group. CLI or PowerShell commands, including standard --force-deletion options, will not clear a stuck backend‑inconsistent state (such as ApplianceBeingDeleted or long‑lived Deleting).

This situation requires the backend deletion process to be completed or reset so that the deny assignments are released and the managed resource group can be cleaned up automatically.

In summary, this is a backend issue rather than something that can be resolved through user actions, and the behavior you’re seeing is expected when the deletion workflow becomes stuck. The recommended next step is to open an Azure Support ticket referencing:

and ask the Azure / Databricks backend team to investigate the stuck deletion and clear the resource state so cleanup can complete.

Answer 2

Use force deletion on the Databricks workspace and, if needed, on the resource group to break the stuck Deleting state and remove the managed resource group.

1. Force delete the Databricks workspace (recommended first)

If the workspace is still present as a resource (even if its provisioning state is Deleting), use either Azure PowerShell or Azure CLI with the force deletion flag. This also removes the managed resource group and its Unity Catalog data.

• Azure PowerShell:

Remove-AzDatabricksWorkspace \
  -Name "dbw-medico-prd-eus2-001" \
  -ResourceGroupName "rg-medico-prd-eus2-001" \
  -ForceDeletion

• Azure CLI:

az databricks workspace delete \
  --name dbw-medico-prd-eus2-001 \
  --resource-group rg-medico-prd-eus2-001 \
  --force-deletion \
  --yes

This instructs Azure to:

• Delete the Databricks workspace.
• Force delete the workspace catalog and managed resource group, including DBFS storage and access connector.

Warning: Force deletion is irreversible and permanently removes all data in the workspace catalog, including managed tables and volumes. Ensure any required data is backed up before running these commands.

2. If the parent resource group must be removed and the workspace is blocking it

If the workspace resource is still blocking deletion of rg-medico-prd-eus2-001, delete the resource group with force deletion types so that the Databricks workspace is forcibly removed as part of the group deletion:

• Azure CLI:

az group delete \
  --name rg-medico-prd-eus2-001 \
  --force-deletion-types Microsoft.Databricks/workspaces

• Azure PowerShell:

Remove-AzResourceGroup \
  -Name "rg-medico-prd-eus2-001" \
  -ForceDeletionType Microsoft.Databricks/workspaces \
  -Force

This ensures the Databricks workspace is force-deleted even if it is stuck, which in turn allows the managed resource group to be cleaned up.

3. Behavior with Unity Catalog–enabled workspaces

If Unity Catalog is enabled by default on the workspace and force deletion is not used, the default workspace catalog and its storage (in the managed resource group) are retained even after workspace deletion. To fully remove the managed resource group and catalog data, the force deletion options above must be used.

---

References:

• Delete a workspace
• az databricks workspace delete
• Remove-AzDatabricksWorkspace
• Databricks.WorkspacesDelete Method
• Deploy a workspace with PowerShell