The Icinga check that links here is checking whether the tftp process is running.
The command line it uses is:
/usr/lib/nagios/plugins/check_procs -c 1:1 -u nobody --ereg-argument-array=\'.*/usr/sbin/atftpd .*\'
It checks if exactly one process (no more and no less) are matching the regular expression in the list of running processes.
What you can do
•ssh to the relevant server and check yourself for processes matching atftp (ps aux | grep atftpd ..)
•Use the exact command line from above that icinga uses
•If it's because there are 0 procs, check the status and try to start the service (systemctl status atftpd; systemctl start atftpd)
•If it's because there are 2 or more processes, stop it, make sure nothing is running anymore, start it again once
•If it still fails to start try to find out why (journalctl -xn ; view /var/log/syslog )
Where are files?
•The config file on the server is in /etc/default/atftpd. See the parameters there.
•The puppet class that sets it up is class install_server::tftp_server.
•The files it serves are in /srv/tftpboot, config files are in puppet, larger files are in volatile/tftpboot.
How bad is it if this is down?
A: Not immediately critical but until it's fixed nobody can install new servers. It should be treated with High but not UBN priority.