another kubernetes cluster broke because of a dns issue. how many more layers of yaml do we need to add before this whole thing collapses under its own weight?