`/v3/app` delete does not wait until service binding are unbound

Issue

User tries to call DELETE v3/app/<guid> for an app with a service binding where the service broker does not reply synchronously.
This can result in a failed v3 app deletion job although the service binding is deleted eventually.
App delete job:

{
   "guid":"<guid>",
   "created_at":"2023-06-28T18:46:13Z",
   "updated_at":"2023-06-28T18:46:16Z",
   "operation":"app.delete",
   "state":"FAILED",
   "errors":[
      {
         "detail":"An operation for the service binding between app myapp and service instance slow-broker-service is in progress.",
         "title":"CF-UnprocessableEntity",
         "code":10008
      }
   ],
   "warnings":[
      
   ],
   "links":{
      "self":{
         "href":"https://api.cf.bommel/v3/jobs/<guid>"
      },
      "app":{
         "href":"https://api.cf.bommel/v3/apps/<guid>"
      }
   }
}

Context

The binding deletion is implemented here:
https://github.com/cloudfoundry/cloud_controller_ng/blob/main/app/actions/mixins/bindings_delete.rb#L19-L24

Which is called during app delete:
https://github.com/cloudfoundry/cloud_controller_ng/blob/main/app/actions/app_delete.rb#L101-L104

Possible Expected Results

• Current behavior
• App deletion job should wait
• App deletion job should NOT wait for binding deletion BUT not fail

-> Needs to be discussed

Current result

App deletion job fails although the binding is deleted eventually

Start by reading app/actions/mixins/bindings_delete.rb and app/actions/app_delete.rb at the referenced call sites to trace how asynchronous service-binding deletion affects the app deletion job. Resolve whether deletion should wait for unbinding or avoid failing while it is in progress, then verify that the chosen behavior no longer produces the reported failed job.