databricks workspace status: Deleting

Hi @Sree Harsha

From the error message and screenshot, the workspace deletion is stuck because of a System Deny Assignment created automatically by Azure Databricks.

This deny assignment is applied on the managed resource group (managed_databrickssreeunity), and it prevents manual deletion of resources inside it.

This usually happens when:

• The Databricks workspace deletion did not complete properly.
• The managed resource group is still locked by the system.
• The backend cleanup process did not finish.

Here’s what you can check:

1. Do not try to delete the managed resource group manually.
2. Check if the main Databricks workspace resource (in rg-adf) still exists.
3. If the workspace shows Deleting for more than a few hours, this is likely a backend issue.

Because this is an Azure for Students subscription, support options are limited. In this case:

• Wait 24–48 hours to see if Azure completes the backend cleanup.
• If it is still stuck after that, this requires Microsoft backend intervention (the deny assignment cannot be removed manually).

This is not a permission issue from your side - it is system-controlled.

If the status remains stuck after 48 hours, let me know and I’ll guide you on the next steps.

Hope this helps.

Hello Sree Harsha,

Thanks for raising this question in Azure Q&A forum.

An Azure Databricks workspace stuck in "Deleting" status indefinitely is a well-known backend issue where the Databricks resource provider's appliance deprovisioning process fails to complete, leaving the workspace in a permanent limbo state that cannot be resolved from the customer side alone.

Why this happens

When you initiate a workspace deletion, Azure Databricks must:

Deprovision the Databricks-managed appliance (control plane VMs)

Remove the system-assigned deny assignment on the managed resource group

Delete all managed resources (storage, networking, etc.)

Remove the workspace ARM resource

If step 1 fails (due to the appliance VM being in an inconsistent state, network timeout, or control plane inconsistency), the workspace stays in "Deleting" indefinitely and none of the subsequent steps proceed — meaning the managed resource group with its deny assignment also cannot be cleaned up.

Step 1 — Check for resource locks (quick win)

Before escalating, verify there are no manual locks preventing deletion:

bash
# Check for locks on workspace

Step 2 — Check and delete failed deployments

Go to Azure Portal → Resource Group → Deployments — if there are any stuck/failed deployments associated with the workspace, delete those first before retrying deletion:​

bash
az deployment group list --resource-group <your-rg> --filter "provisioningState eq 'Failed'" --output table
az deployment group delete --resource-group <your-rg> --name <deployment-name>

Step 3 — Attempt force deletion via CLI

Try force deleting using PowerShell:

powershell
# PowerShell force delete

Wait 30 minutes after each attempt before retrying — the backend may need time to finalize.

Step 4 — Try deleting via REST API with force flag

If portal and CLI are blocked, try the ARM REST API directly:​

bash
# Get access token

Step 5 — Escalate to Microsoft Support (required for persistent cases)

If the workspace has been stuck in "Deleting" for more than 2 hours and all above steps fail, this is a platform-level issue requiring the Azure Databricks resource provider engineering team to manually unblock the deprovisioning. Open a Severity A support ticket with:

text
Title: Azure Databricks workspace stuck in "Deleting" status — unable to complete deletion

Details:
- Workspace Name: <name>
- Resource Group: <rg>
- Subscription ID: <sub-id>
- Region: <region>
- Time stuck in "Deleting" state: <duration>
- Managed Resource Group: databricks-rg-<workspace>-<guid>
- Deny assignments blocking managed RG cleanup: Yes/No

Steps already tried:
- Lock check: None found
- CLI force delete: Failed with <error>
- REST API force delete: Failed with <error>

Request:
Escalate to Databricks resource provider engineering team to:
1. Terminate the stuck appliance deprovisioning process
2. Release the deny assignment on the managed resource group
3. Allow clean deletion of the workspace ARM resource

Microsoft engineering can directly access the Databricks control plane backend and manually terminate the stuck deprovisioning job something that is not possible from the customer side.

If it helps kindly accept the answer.

Best Regards,

Jerald Felix