Retry and timeouts
Use cases Achieve high availability (HA) from the standpoint of Users or Proxies assuming that: the failure of some attempt is temporary or localized to some of the workers workers are redundant and interchangeable there are enough time and workers There are several types or levels of retrying: A general control flow of retries: Polling for a status change without knowing when it’ll happen Sleep interval can be linear or exponential, based on your modeling of probability distribution of when the change will happen. ...