SNMP Indexed Tables - filtering and aggregation

[amish1975]

This is duplicate of:

https://www.zabbix.com/forum/showthread.php?t=5726

The great value of important SNMP data is available only as indexed tables.
There's need not just to summarize table data (as was shown above), but also to filter it out (based on certain criteria (RegExp?) and apply aggregate functions on the resultset (avg, count, etc).

The functionality is really HIGHLY DEMANDED, as now the only workaround is to call external scripts, do all the work (snmpwalk, parse the output) externally.

Please consider to extend the functionality of index search - something, like:

UCD-DISKIO-MIB::diskIOWrites["index","UCD-DISKIO-MIB::diskIODevice","dm-*","sum"]
- summarize the "diskIOWrites" values when the "diskIODevice" values match "dm-*" mask (SAN volumes, presented externally);

or:

HOST-RESOURCES-MIB::hrProcessorLoad["index","HOST-RESOURCES-MIB::hrProcessorFrwID","*","avg"]
- average load of all the server's processors

or:

HOST-RESOURCES-MIB::hrSWRunName["index","HOST-RESOURCES-MIB::hrSWRunName","*oracle*","cnt"]
- the number of spawned processes by oracle user

Hope this idea will be inflammatory!
Thanks!

[amish1975]

Will somebody respond?

Unless we have this feature (or reasonable workaround), we'll continue using our custom SNMP monitoring scripts based on RRDTool. The necessity to describe hundreds of counters manually seems terrifying!

[amish1975]

Will somebody from development respond please?

[Pada]

I'm also very interested in something like this!

I recently created a thread, because I was unable to make use of the SNMP Dynamic Indexing of Zabbix due to it not supporting partial matching it seems: https://www.zabbix.com/forum/showthread.php?t=42444

If I have time I can also try to look at the C code. I've already done a few frontend patches of my own.