Advanced parsing of SNMP Values / Monitoring partition usage

[scoopex]

In our setup a snmpv2 item provides a string which needs to be analysed in a non trivial way.

We run hundreds of linux-servers and we want to monitor the partition usage
of these servers. An already existing net-snmp "extend" returns the following value.

"79#/:32#/boot:100#/log-local:7#/var:30#/export"

All of the numbers describe the percentage of filesystem-usage of a certain partition.
If one of these values excesses 97% a trigger should fire an alert.
To analyse the percentage-values of the snmp-oid, we need some intelligence at the zabbix server. Is there a possibility to add programmatic intelligence at the zabbix server to derive a status for a trigger?

Where can i find detailed information about the usage an possibilities regexes in zabbix.
(The Documentation in the manual is not very useful: http://www.zabbix.com/documentation/...config/regexps)

Is there a better way to monitor automatically all mounted partitions which are not pseudo-filesystems (like /proc, /dev, /sys, ...)

Regards
Marc

[scoopex]

It´s easy - but very intuitive :-)

Documentation should be enhanced : http://www.zabbix.com/documentation/...config/regexps

1.) Define a regular expression in
     Administration => General => Dropdown at the right side => New Regular Expression
2.) Add a expression : 
     Expression: "(100|9[789])#.+?:"
     Expression Type: "RESULT is true"
3.) Enter test string: "97#/:94#/boot:100#/log-local:7#/var:30#/export:"
4.) Hit "Test"
5.) Define a Name: "DiskUsageRegex97"
6.) Save
7.) Create a Item which monitors the value to parse
     i.e. SNMP OID: NET-SNMP-EXTEND-MIB::nsExtendOutputFull."df/mountpoint/usage"
8.) Define a trigger with the following expression:
     "{Linux Standard Monitoring:DiskUsage.regexp(@DiskUsageRegex97)}>0"