Dealing with fallout from another server failure. We have pretty good redundancy most places, except ironically some of the most stable parts that always felt low on the list to prioritize because they never failed… until they did. Picking up the pieces, improving a few things for later.