Automatically handle rate limiting

[bmiller59]

In our use of this library, we are receiving 429 / Too many requests errors.

This library should handle rate limiting automatically so that these errors are not encountered, perhaps on an opt-in basis using some config parameters.

If that is not possible, then at a minimum it should expose the X-RateLimit-Limit and X-RateLimit-Remaining headers, and provide a function that will take a job and return a delay to wait in order avoid triggering these errors so that we can implement our own queuing.

Thank you.

Referrence: https://cloudconvert.com/api/v2#errors

[KnorpelSenf]

I think it should already be possible to access the error data via error.response.headers, can you confirm this?

@josiasmontag it may still make sense to handle rate limiting automatically, or document how to obtain the retry-after headers instead.