Extending Zabbix Monitoring with Custom Scripts and Handling Stale NFS Handles

Zabbix monitoring supports custom scripts and self‑made templates to extend its capabilities, allowing you to monitor business‑specific metrics by writing scripts in shell, Python, or other languages.

Below is an example of a debug‑run shell script that checks disk usage and evaluates a threshold:

[root@mha ~]# sh -x watch_disk.sh
++ df -k
++ grep centos-root
++ awk '{print $4}'
df: ‘/NFS’: Stale file handle
+ Disk=320455292
+ [ 320455292 -le 83886080 ]
+ echo 1
1

The script captures the output of df -k, filters the relevant line, extracts the free space column, and compares it against a predefined limit. In this run the NFS mount returned a “Stale file handle” error, causing the script to treat the free space as a large value.

To resolve the stale NFS handle, three approaches are suggested:

Kill the processes using the mount point: fuser -k /NFS Search for processes holding the directory:

ps -ef | awk '{print $2}' | grep -v PID | xargs /usr/proc/bin/pwdx

Force unmount the NFS share: umount -f /NFS The article concludes with a list of recommended reading links covering topics such as Kubernetes deployment, Ansible roles, and other DevOps practices.