We already have the retry logic in place, but to improve the situation on busy servers (e.g. HTTP 503 or 429) we should respect the Retry-After header, if present, when calculating nextSleep.
We still fail if the total time exeeds the configured timeout, but it would be useful to include some info about the server return Retry-After when returning the "retry timeout" error.
We already have the retry logic in place, but to improve the situation on busy servers (e.g. HTTP 503 or 429) we should respect the
Retry-Afterheader, if present, when calculatingnextSleep.We still fail if the total time exeeds the configured timeout, but it would be useful to include some info about the server return
Retry-Afterwhen returning the "retry timeout" error.