Along the code path for operations on time-series collections, we have both the view and the bucket collection namespaces floating around which can be confusing. We should clean up or at least add better documentations to have a clear point where the namespace conversion is done.