Azure compute resources that are used to create and manage groups of heterogeneous load-balanced virtual machines.
The VM scale set is in a transient “marked for deletion” / deleting state that blocks any further customer-side updates (HTTP 409 OperationNotAllowed). In this state, dependent resources (Application Gateway backend pool, NSG, subnet, etc.) also cannot be detached because the VMSS cannot be modified.
From the available guidance, there is no supported way to “unblock” or override a VMSS that is already stuck in deletion from the customer side. The only customer-controllable prerequisite is that all VMs belonging to the scale set must be deleted before the scale set itself can be removed:
- When a Virtual Machine Scale Set in Flexible orchestration mode still has associated VMs, attempts to delete the scale set fail with:
OperationNotAllowed. Deletion of Virtual Machine Scale Set isn't allowed as it contains one or more VMs. Please delete or detach the VM(s) before deleting the Virtual Machine Scale Set.
- The documented resolution is to delete all VMs associated with the scale set, then delete the scale set again.
In this scenario, the VMSS is already in a deleting/marked-for-deletion state and is rejecting all update operations, so the remaining steps must be handled by Azure support on the platform side (forcibly completing or clearing the stuck delete operation and then removing the dependent resources). Opening a support ticket with the full error details and resource IDs is required; there is no additional CLI or portal operation documented that can clear a VMSS once it is stuck in this state.
References: