BlinkyStitt
commented
1 year ago There was a ? on one of the request handlers so that we weren't actually retrying if an error happened on that line. I made an error there continue the loop to try another server.
And I also added another layer of retrying on errors. I might end up removing that one if it just makes error responses slower, but we'll see how it works in practice
we still let too many 500s slip through. investigate more places we can do automated retrying