Support precious quota enforcement with added latency

The [quotas](https://cloud.google.com/endpoints/docs/openapi/quotas-configure) limit is per minute, so if we get status code `429 Too Many Request` and we wait a minute we should be fine but after we wait a minute the next request has status code `429`, it happens even if we wait more than twice like `150` seconds. I can reproduce the issue with a simple example in [this repo](https://github.com/jyeros/esp-quota-limit). In the github action logs, there is one waiting `61` seconds with the issue and other waiting `150` with the same issue.

Reproduce the quota behavior using the linked esp-quota-limit repository and inspect the GitHub Action runs showing waits of 61 and 150 seconds. Compare the proxy's handling with the Cloud Endpoints quota documentation; done means requests no longer receive 429 after the documented quota window has elapsed.

Written by the indexing model from the issue text.