Shell Monitoring Script

I have a service (Jupyter) that crashes now and then and the Data scientist need to reboot their machines or login via SSH and restart it. I was thinking in adding a service (shell) that keeps monitoring the service status and after X amount of continuos failures restart it. Is it a good pattern or seems hacky? most of the time service crashes because OOM. (Shell script may run systemctl and check for service status)