How to Monitor CPU, Disk and Load on Linux with NRPE and Nagios

Updated: 28 September 2026 · Applies to: NRPE 4.1 with Nagios plugins 2.4 and 2.5 on AlmaLinux and Rocky Linux 9 and 10 (EPEL), Ubuntu 24.04 and 26.04, Debian 13

NRPE (Nagios Remote Plugin Executor) lets a Nagios server run checks on another Linux server: load, disk space, swap, processes and anything else a plugin can measure. The monitored server runs the small NRPE service on TCP port 5666, and the Nagios server asks it for results with check_nrpe. We tested the setup below with NRPE 4.1.2 on AlmaLinux 9.

1. Install NRPE on the monitored server

dnf install epel-release
dnf install nrpe nagios-plugins-load nagios-plugins-disk nagios-plugins-swap nagios-plugins-procs nagios-plugins-users   # AlmaLinux, Rocky Linux
apt install nagios-nrpe-server monitoring-plugins-basic                                                                   # Ubuntu, Debian

2. Allow your Nagios server

In /etc/nagios/nrpe.cfg, add the Nagios server's address to allowed_hosts:

allowed_hosts=127.0.0.1,::1,203.0.113.50

Only listed addresses may ask for results.

3. Define the checks

Put your own commands in a separate file, /etc/nrpe.d/checks.cfg (on Ubuntu and Debian, /etc/nagios/nrpe.d/checks.cfg, with plugins in /usr/lib/nagios/plugins):

command[check_root_disk]=/usr/lib64/nagios/plugins/check_disk -w 20% -c 10% -p /
command[check_load]=/usr/lib64/nagios/plugins/check_load -r -w 2,1.5,1 -c 4,3,2
command[check_swap]=/usr/lib64/nagios/plugins/check_swap -w 20% -c 10%
command[check_zombies]=/usr/lib64/nagios/plugins/check_procs -w 5 -c 10 -s Z

-w and -c are the warning and critical limits. check_load -r divides the load by the number of CPU cores, so the same limits suit small and large servers. The standard plugins have no check for RAM use; a small third-party script is common for that, while swap use (above) shows memory pressure.

4. Start NRPE and open the port

systemctl enable --now nrpe                    # AlmaLinux, Rocky Linux
systemctl restart nagios-nrpe-server           # Ubuntu, Debian
firewall-cmd --permanent --add-rich-rule='rule family="ipv4" source address="203.0.113.50" port port="5666" protocol="tcp" accept'
firewall-cmd --reload

5. Test from the Nagios server

/usr/lib64/nagios/plugins/check_nrpe -H 203.0.113.10                 # prints the NRPE version
/usr/lib64/nagios/plugins/check_nrpe -H 203.0.113.10 -c check_root_disk

A result such as DISK OK - free space: / ... means everything works. Install check_nrpe there with dnf install nagios-plugins-nrpe or apt install nagios-nrpe-plugin.

6. Add the service in Nagios

define service {
    use                   generic-service
    host_name             web1
    service_description   Root disk
    check_command         check_nrpe!check_root_disk
}

Common problems

• "CHECK_NRPE: Error - Could not connect ... Connection refused": NRPE does not run, or a firewall blocks port 5666.
• "CHECK_NRPE: (ssl_err != 5) Could not complete SSL handshake" or the connection closes at once: the Nagios server is not in allowed_hosts. Restart NRPE after changing it.
• "NRPE: Command 'x' not defined": a typo in the command name, or the file in nrpe.d was not loaded; check include_dir in nrpe.cfg.

Official documentation: NRPE on GitHub and check_disk (Monitoring Plugins).