How to reset an exponential_sleep configuration, after a successfull_threshold again?

[spex66]

Great lib, which saved me definitive from an ugly hack to make "requests" calls retrying on the status_code :)

    def retry_if_result_429(response):
        """Return True if we should retry (in this case when the status_code is 429), False otherwise"""
        # DEBUG to see this in action
        if response.status_code == 429:
            print 'RETRY 429: %s' % response.url
            return True
        else:
            return False

    #
    # functools partial to the help
    # to inject auth parameter in the get/delete/post/put requests
    #
    retry429_decorator = retry(
        retry_on_result=retry_if_result_429, 
        wait_exponential_multiplier=1000, 
        wait_exponential_max=10000
        )

    #1. curry the authentication
    #2. decorate it with the retry logic
    s.get    = retry429_decorator(partial(s.get,    auth=credentials))
    s.put    = retry429_decorator(partial(s.put,    auth=credentials))
    s.post   = retry429_decorator(partial(s.post,   auth=credentials))
    s.delete = retry429_decorator(partial(s.delete, auth=credentials))

But running a huge test, I have the doubt now, that one of my batches had run into the "wait_exponential_max" and that this will not be reset.

Will the wait_function applied to every subsequent call? Or will the delay happen only when it run into a retry again?

Any idea how to achieve a reset? Resetting the "attempt_number" for example?

thx for any hint or clarification, when the delay happens