Fix the failure to scale-down k8s cluster. We were getting this
failure because as part of scaling, Bay needs to be retrieved for
the respective k8s obj. But if the object passed is a Bay itself
we need not retrieve the Bay. At present even for Bay object the
code proceeds ahead to retrieve bay and cannot find bay_uuid
attribute. This patch fixes this problem.
We might need to backport this fix to stable/liberty as well.
Change-Id: I7ed03af8bdbb8647851751bb550d9529d96b5cb4
Closes-Bug: #1509439