Can you give a little more detail?

Because typically how databricks works is to deploy temporary VMs in a resource group. When their job is done they are removed.

The cluster definition in databricks however remains.

So what exactly do you mean by 'compute resources'? The VMs or the cluster definition?