We provide users the ability to register/sign-up with our service using Twitter. We use a third party python API wrapper (Tweepy) to register users and do the OAuth dance for us. After the OAuth steps are complete, and the user is redirected to our site, we use said third party wrapper to retrieve basic information about the user to populate a final registration form. This process invokes /users/show.json.
Recently, new members have complained that the twitter redirect takes an awfully long time. Investigating this issue has highlighted that the HTTP request to /users/show.json alone takes 28 seconds for non-authenticated users.
I also found out that one of the developers on my team was using the “live” twitter app consumer key/secret pair in their local development environment, and was frequently getting rate-limited when creating new features.
I understand the rate limit component (Code: 88, 15 minute window etc…), but my question is: Does the fact that the app has been rate-limited several times impact/penalized the performance time in which non-rate-limited requests execute? I.e are the requests that we make to the twitter API slowing down because we are being throttled due to repeated rate-limiting?
Thanks for your help!