I can't think of a solution right now, haven't looked through the codebase. However I suggest error handling should be worked on so if some step fails, the instance creation is aborted or at the least, a slightly more descriptive error message than two status codes that are vague (400...