Weird result with SNMP Telegraf plugin

Telegraf
,
1

Hello everyone,

I’ve been using the TIG solution for monitoring for some months.
And I just noticed something weird with the Telegraf SNMP plugin. Especially playing with the OID for the CPU load of my routers.
In the InfluxDB base, I get results as:

> select * from cpu_load where agent_host='10.211.0.5' order by time limit 20
name: cpu_load
time                agent_host host hostname   hrProcessorFrwID hrProcessorLoad network
----                ---------- ---- --------   ---------------- --------------- -------
1720130410000000000 10.211.0.5 TIG  RB_OP_0005 .0.0             1               backbone
1720130422000000000 10.211.0.5 TIG  RB_OP_0005 .0.0             0               backbone
1720130431000000000 10.211.0.5 TIG  RB_OP_0005 .0.0             1               backbone

but if I run the same snmp request (with a walk):

root@TIG:~# snmpwalk -v2c -c supervision 10.211.0.5 HOST-RESOURCES-MIB::hrProcessorTable
HOST-RESOURCES-MIB::hrProcessorFrwID.1 = OID: SNMPv2-SMI::zeroDotZero
[...]
HOST-RESOURCES-MIB::hrProcessorFrwID.9 = OID: SNMPv2-SMI::zeroDotZero
HOST-RESOURCES-MIB::hrProcessorLoad.1 = INTEGER: 1
HOST-RESOURCES-MIB::hrProcessorLoad.2 = INTEGER: 0
HOST-RESOURCES-MIB::hrProcessorLoad.3 = INTEGER: 0
HOST-RESOURCES-MIB::hrProcessorLoad.4 = INTEGER: 0
HOST-RESOURCES-MIB::hrProcessorLoad.5 = INTEGER: 1
HOST-RESOURCES-MIB::hrProcessorLoad.6 = INTEGER: 1
HOST-RESOURCES-MIB::hrProcessorLoad.7 = INTEGER: 1
HOST-RESOURCES-MIB::hrProcessorLoad.8 = INTEGER: 0
HOST-RESOURCES-MIB::hrProcessorLoad.9 = INTEGER: 0

And my Telegraf SNMP conf looks like:

[[inputs.snmp]]
  agents = ["udp://10.211.0.2:161", "udp://10.211.0.5:161"]
  version = 2
  community = "supervision"
  tags = {network = "backbone"}
# More inputs
  [[inputs.snmp.table]]
    name = "cpu_load"
    oid = "HOST-RESOURCES-MIB::hrProcessorTable"
    inherit_tags = ["hostname"]
# More inputs

So I basically don’t understand, where are all the value from my snmpwalk command? Is Telegraf doing an average of the value to build a single one? This is not very bad news since the value are still readable for the CPU load. But this is kinda weird…

2

Since your request (or the table) didn’t include any index, you need to tell telegraf to add it by setting index_as_tag.

3

This make sense, but since I did not configure the index, what is the data I’m getting?

Update:
Looks like I get all the good data in my base, but just without the index…

4

You’re getting 9 identical metrics, so InfluxDB stores/updates to (probably) the last one. There’s definitely no aggregation (like average) happening…

1 Like