This python script uses nvme command on Linux to scan NVMe disks on the server and check the discovered disks' health.

1. Install nvme-cli package on your server

Debian/Ubuntu:

$ sudo apt-get install nvme-cli

CentOS/RHEL:

$ yum install nvme-cli

2. Import the zabbix_template.xml into your Zabbix templates dashboard. This template is pre-configured and will add a new one called Template NVMe.

3. Copy the configuration files:

$ mkdir /etc/zabbix/scripts/ && sudo chowd -R zabbix:zabbix /etc/zabbix/scripts/
$ cp zabbix_nvme.conf /etc/zabbix/zabbix_agentd.d/
$ cp nvme_discover.py /etc/zabbix/scripts/
$ cp sudoers_zabbix_nvme /etc/sudoers.d/
$ systemctl restart zabbix-agent

4. Add the template to the hosts you want to monitor.

The pre-defined configuration can be modified directly after importing the template. The default discovery rule is 1 hour and only extracts the percentage_used which shows how healthy the disk is.

The default trigger for the template is above 80.