In this meeting, the State Changers are discussing the retry loop pattern as a useful solution when working with external APIs that may not always provide the desired answer. Anthony explains the details of the retry loop, including the validation process and sleeping for a specified amount of time before running it again. The suggestion is made to make the sleep duration dynamic by using a utility function and multiplying it by the index of the loop. This creates a staggered retry approach. They delete unnecessary code and make the necessary adjustments. The meeting participants also discuss the potential action to take if the retry loop is unsuccessful, suggesting considering the next steps and handling the unhappy path scenario. This pattern is deemed useful for those using services with imperfect data reliability, such as open AI or similar platforms. The participants express appreciation for the solution and decide to move on to discussing background tests.
(Source: Office Hours 1/26 )
