How to correctly handle OID with dynamic indexes

[xtianfab]

I have OID that returns the CPU usage value per minute, so the index changes every minute. If I perform a discovery I obtain the item correctly, but I obtain items for every minute. Is it possible to keep only the last value or what would be the correct way to handle these OIDs?



Regards

[Hamardaban]

Show some examples of OIDs and values.

[xtianfab]

Hello, Hamardaban, thanks for answering.


I share an example:

This is the OID, and the index changes every minute.

└─$ snmpwalk -v 2c -c public x.x.x.x.x 1.3.6.1.4.1.1397.2.23.13.2.2
iso.3.6.1.4.1.1397.2.23.13.2.2.1.1.1.12679912 = STRING: "2024-02-09 11:52:00"
iso.3.6.1.4.1.1397.2.23.13.2.2.1.2.1.12679912 = Gauge32: 46
iso.3.6.1.4.1.1397.2.23.13.2.2.1.3.1.12679912 = Gauge32: 45
iso.3.6.1.4.1.1397.2.23.13.2.2.1.6.1.12679912 = Gauge32: 10

Next minute:
└─$ snmpwalk -v 2c -c public x.x.x.x 1.3.6.1.4.1.1397.2.23.13.2.2
iso.3.6.1.4.1.1397.2.23.13.2.2.1.1.1.12679914 = STRING: "2024-02-09 11:54:00"
iso.3.6.1.4.1.1397.2.23.13.2.2.1.2.1.12679914 = Gauge32: 46
iso.3.6.1.4.1.1397.2.23.13.2.2.1.3.1.12679914 = Gauge32: 45
iso.3.6.1.4.1.1397.2.23.13.2.2.1.6.1.12679914 = Gauge32: 10

As you can see, the index changes and with each change a new item is generated in zabbix.



Regards

[Hamardaban]

The general idea is this: get a list of IDs and values with the CA key, convert the "SNMPwalk to JSON" preprocessing to json, and the next step is JS preprocessing: replace by regex "iso.3.6.1.4.1.1397.2.23.13.2.2.1.1.1.cxzcvxas a" to "iso.3.6.1.4.1.1397.2.23.13.2.2.1", "iso.3.6.1.4.1.1397.2.23.13.2.2.1.2.1.zxzzcss" to "iso.3.6.1.4.1.1397.2.23.13.2.2.1.2" ,,,,

[xtianfab]

Should this be done in a discovery or should it be a discovery dependent item?

[Hamardaban]

It is possible in both cases.
The goal is to get json data for LLD of the desired content.

"snmp item" -> data correction in item preprocessing -> "dependent lld rule" ->creating item
or
"snmp LLD rule" -> data correction in LLD preprocessing -> creating item

[xtianfab]

The problem is that every time the item is discovered, a new one is created, wouldn't it have to do with the key? I use the snmpindex for the key and since it is a different one, another item is created every minute



I can get the values without problems, but they always get duplicated

Regards

[Hamardaban]

So I suggest you cut out the changing part and leave it unchanged. Then there will be no detection of "new" elements.

[xtianfab]

I think I already achieved it, doing a walk from Item and then preprocessing as you have indicated. I'll be testing and seeing how it works.

Thank you very much for the help