Comment 2 for bug 1431291

As per the move to the conductor, it was also agreed on not leaving the scheduler know the list of instance_uuids
Consequently, the recent Scheduler RPC API version bump (4.0) removed that field from the request_spec which was passed in because it was no longer needed.

If we want to backtrack which instance was failing with the logs, we then only have 2 possibilities :
 #1 request_spec still continues to provide a full information of the first instance being requested, including its uuid - so we can show this up but it would only be accurate for the first one
 #2 that instance information still includes an EC2-compatible field called reservation_id which is set even for OpenStack API and describe the request by itself. We can show it up too.