Debug false POE SNMP values result in excessive usage percentage

Owned by Robin Gierse
Last updated: Dec 10, 2024 by Matthew Hierholzer

This article provides details about errors with SNMP values that result in increased usage percentages.

APPLICABLE TO ALL CHECKMK VERSIONS

Table of Contents

Problem

A POE-enabled network interface shows an excessive usage percentage.
e.g. POE usage (6000W/1800W): : 333% (warn/crit at 90.0%/95.0%)

Reason

Some hardware vendors violate the SNMP RFC and output milliwatts instead of watts for pethMainPseConsumptionPower.

When looking at an SNMP walk of the affected device, it will become evident that the value in pethMainPseConsumptionPower (1.3.6.1.2.1.105.1.3.1.1.4.1) is not output in watts:

OMD[mysite]:~$ grep .1.3.6.1.2.1.105.1.3.1.1. ~/var/check_mk/snmpwalks/my_device
.1.3.6.1.2.1.105.1.3.1.1.2.1 1800
.1.3.6.1.2.1.105.1.3.1.1.3.1 1
.1.3.6.1.2.1.105.1.3.1.1.4.1 6000


Checkmk then uses these values to calculate the percentages:

6000/1800 → 333%


Those are then wrong, of course, because if your input data is faulty, the result of your calculation will also be faulty.

Solution

As the bug lies within the SNMP implementation of the affected device, the hardware vendor must fix this bug in the firmware.
Please open a ticket with the vendor and ask them to fix the bug.