I have the problem with service module - it sometimes (!) throws an error, when restarting supervisor.
I tried to reproduce this bug in a sandbox project, but wasnt succeed. But in my sandbox supervisor was not handled any processes.
So - my guess, when supervisor supervising some processes - it restarts longer, and because of it Ansible’s service module fails because it has some kind of timeout.
Here is the output with fail: