Azure Databricks Workspace Stuck in ApplianceBeingCreated after Terraform Script Interruption

Question

Dear Expert

I'm developing an Azure Databricks environment (following Databricks Secure Reference Architecture, no public access) for a customer using Terraform. Unfortunately, my script aborted due to an issue on my computer. The Databricks Workspace being deployed now stucks in "ApplicanceBeingCreated" state for over 2 hours and can not be deleted following the guidance in various forums and your official docs. Could you please provide me guidance on how to resolve the issue?

Thank you for your support

Answer 1

Hi Elias Rüfenacht,
Thankyou for reaching microsoft Q&A!

This behavior is expected when an Azure Databricks workspace deployment is interrupted. The “ApplianceBeingCreated” state means the managed resource provider has not completed provisioning and is holding a lock on the workspace, which prevents deletion.

To resolve this, follow the steps below:

First, validate whether the deployment is still active. In the Azure Portal, go to Resource Group → Deployments and check if a Databricks-related deployment is still running or failed. If it is in a running state, wait for it to complete or fail before taking further action.

If the deployment is already failed or stuck, remove any locks that may block deletion. Navigate to the resource group and check Locks. If a lock exists (ReadOnly or Delete), remove it.

Next, check the managed resource group (named like databricks-rg-<random>). If it exists, ensure there are no ongoing operations inside it. Do not manually delete resources inside this managed resource group, but confirm its state.

After that, retry deletion using Azure CLI instead of the portal, as it is more reliable for stuck resources:

az resource delete \
  --ids /subscriptions/<subscription-id>/resourceGroups/<rg-name>/providers/Microsoft.Databricks/workspaces/<workspace-name>

If the deletion still fails, re-run your Terraform deployment with the same configuration. Azure Resource Manager will attempt to reconcile the partially created resources and either complete or clean up the deployment. Once the state transitions out of “ApplianceBeingCreated,” deletion will succeed.

Also ensure that: The subnet delegation to Microsoft.Databricks/workspaces is intact There are no NSG or policy restrictions blocking the managed resource creation No concurrent deployments are targeting the same workspace

In most cases, the backend operation times out after a few hours and automatically moves to a failed state, after which deletion works normally.

Please let us know if issue persists.