If you see rare "stuck" requests that either hit the overall request deadline (on frontend instances), or seem to hang "forever" on backend instances, there is a very small chance you are running into this bug we found in App Engine Standard's logging library. The brief summary is if you call logging at import time while other threads are also logging, it is possible that you can deadlock your requests. We were running into this on rare occasions, that recently started happening more often (a few times a day).
We have a workaround that we've been using for the past few days, and it seems to solve the problem. If you run into this, I'd love to hear about it.