A better way to monitor diskspace with Zabbix
Zabbix has a log of built-in support to monitor diskspace on servers. But I'm not always interested in triggers that fire on a certain amount or percentage of free space defined in a template. For some filesystems, it may be perfectly normal to be at 90% usage and remain stable there.
What does interest me more is the amount by which disk usage increases in a short timespan, and Zabbix does have the functionality to cater for this. For example:
{Template OS Linux:vfs.fs.size[{#FSNAME},pfree].last(0,3600)}-
{Template OS Linux:vfs.fs.size[{#FSNAME},pfree].last(0)}>10
In the discovery rules for your template will alert you to any increase of more than 10% in one hour.
Happy monitoring :-)
Comments
1 comment postedYou are right, the default way most monitoring systems will alert on file system usage is not useful.
I have made a monitor script that allows each filesystem on all servers to be different and the devs can set the threshold for their application filesystems themself.
Because even sudden spikes can be normal, just thinking about feeds that occur once a day for example (we have those a lot).