Hi @StevenWalls, yes for better or worse this is par for the course. Skimming through our log history (not guaranteed to be exhaustive), we had some number of tables -- anywhere between 1 and 26 -- fail with the malformed HTTP response on July 13, July 28, September 20, October 25, December 8, February 21, February 22, May 6, and May 9. I think it is related to a 504 timeout error from Instructure's backend which the DAP Python library does not handle gracefully. It's up to you if you want to catch and handle this error or just ignore it. Our approach is to simply note it in our logs and let the tables that failed get synced on the next job run.
