Ah, that’s annoying. I’d assumed you were launching docker containers with systemd.
A quick way to check if any of them have this problem is to use docker-compose kill -s SIGTERM on each. If any of them are hanging that should take longer than usual.
Alas, not option. Writing a unit to SIGKILL a specific container (whichever ones end up hanging) is probably the best option in this case. I’m not sure if docker-compose has some strange interaction with how containers are killed on shutdown, so maybe my theory is wrong.