Yesterday a sufficient amount of testers received an access to my game and started playing simultaneously which resulted in all of them constantly receiving this error during most cloudscript requests. This already happened before under the similar conditions.
I do indeed use a firebase realtime database as a third-party backend to store a part of a data and already met this kind of issue which I addressed on the forum. HTTP requests periodically were returning "CloudScriptHTTPRequestError" without any useful information. As a solution I added a retrial mechanism in a form of a method which try/catches all http requests to the firebase and peforms up to 3 re-attempts in a case of an exception. I stopped receiving CloudScriptHTTPRequestError and was quite puzzled by what happened yesterday.
Initially I thought that this happens because of all playfab instances ddos'ing FB database from the same IP, but if http request is a source of the error - then why I still get exceptions with requests positioned in the "try" block? I expected my custom error to be thrown in a case of a failure (and it did, but only a few times). According to the documentation, error 504 may also appear during playfab API requests but then I'd like to know why each ~fifth API call was ending with a failure during 3 hours.
Is it connected somehow to the playfab dynamically scaling resources with an arrival of new players? Game seems to work better right now, even though I don't see significant changes in the server's load. If the FB was a source of problems - I would expect this problem to persist constantly.