Or, you can get the output from journalctl or syslog of the previous (non-working) boot, too. That should suffice.