Is it possible that the timeouts was due to cloudflare blocking some user agent, maybe the browser integrity check/bad browser detection? When the api is used in a web browser, there was no reported issue with latency/timeouts from users and logging. the latency is not much different than without proxying on my end. but when used in an android and ios app with custom user agent I did see occasional large latency and the some users reported random timeouts. Unfortunately I didn't test updating the app to use chrome user agent and simply turned off proxying for that api server (and other api servers that used by non-browser clients). It was happening last year though, I haven't try it again yet.
Ah sorry, I didn't realize you meant that things were outright timing out, rather than just being slower than when not proxied. That seems like something went wrong. I don't think we have anti-bot measures that would cause API clients to time out. If anything they would get an error back or maybe get served a captcha or JavaScript challenge.
