The yo-yo ride of the retweet API gave me this idea. It depends on
proper versioning of the API by Twitter.
Twitter creates an API call that returns the current working API
version. We query that method and use that version of the API for our
If something goes down, Twitter simply pushes out the version number
of an older API version, which is still working correctly. Our systems
will then automatically fall back to using that older version, until
Twitter again pushes out the new version number when it's back online.