When a service is down, the first five minutes should be boring. A short checklist keeps the investigation grounded.
Check the unit
systemctl status my-service --no-pager -l7/7/26Less than 1 minute
journalctl is one of the fastest ways to understand what a small server is doing.
Recent service logs
journalctl -u my-service --since "30 minutes ago" --no-pager7/6/26Less than 1 minute