I have prove of cases where errors from same base url are formatted in different ways. And I also have prove of endpoints throwing 200 as status code for error. Your method will work to some extent, but I still stand with this being the best to handle independently for all endpoints.