Post #2649756

For some longer time we observed pods crashing out of nowhere. Nothing apparent in logs, just pod gone for a minute or two, like once a week. Hard to analyse in a flood of logs. Today we got a customer who was so dedicated he would hammer that website for an hour. Everytime he would open it, pod crashed. He was our "edge case" but this time we got a dedicated one, to finish his task. And finally the logs were obvious... missing return before next() ended up with set headers after response.