I wonder if anyone is rate-limiting APIs by CPU time consumed on the server end
This thought brought to you by pondering how more efficient API designs (like GraphQL etc.) are also more difficult to rate-limit than less efficient API designs (REST etc.)
