502 bad gateway error on site

[rudietuesdays]

From @jheretic : "I suspect it might have been a memory issue on the VM, and the OOM killer got it. Memory usage seems stable so far, so we're not likely to get more info unless it happens again. I'm going to suggest modifying the default docker-compose setup so that containers auto-restart on failure, and look into the possibility of adding prometheus metrics if that's easy to do. "