The following is a primer for setting up your Raspberry PI to use SNMP to monitor the CPU and the GPU temperature in Cacti. The instructions below cover both Raspian and Centos 7. I am assuming you are familiar with SNMP.

Raspian

Create a file named /usr/local/sbin/cputemp

#!/bin/bash
cpu=$(</sys/class/thermal/thermal_zone0/temp)
echo print $cpu/1000 | /usr/bin/perl
echo

Create a file named  /usr/local/sbin/gputemp

#!/bin/bash
gpu=$(/usr/bin/vcgencmd measure_temp)
echo $gpu | /usr/bin/awk -F '=' {'print $2'} | /usr/bin/awk -F "'" {'print $1'}

Set the permissions on the above files

sudo chmod 775 /usr/local/sbin/cputemp
sudo chmod 775 /usr/local/sbin/gputemp

Install snmp by:

sudo apt-get update
sudo apt-get install snmpd

edit the file /etc/snmp/snmpd.conf file and add the following lines to the end

extend cputemp /usr/local/sbin/cputemp
extend gputemp /usr/local/sbin/gputemp
view systemonly included .1.3.6.1.4.1.8072.1.3.2

If you want to be able to monitor the network interfaces via snmp/cacti, add the following line

view systemonly included .1.3.6.1.2.1.2

If you want to be able to monitor the storage usage and memory via snmp/cacti add the following line

view systemonly included .1.3.6.1.2.1.25.2.3.1

find the line near the top of the file that reads

agentAddress udp:127.0.0.1:161

and add a “#” in front of it. Then find the line (two lines down) that reads

#agentAddress udp:161,udp6:[::1]:16

and remove the “#” in front of it

save your changes and execute

usermod -a -G video Debian-snmp 

/etc/init.d/snmpd restart

Centos 7

Easy Way

su -
rpm -Uvh https://yum.seei.biz/altarch/7/os/armhfp/Packages/PiSNMPTemperature-1.0.0-1.noarch.rpm

Follow the instructions provided then –

Edit /etc/snmp/snmpd.conf and set the community string – then

systemctl restart snmpd

Jump down to “Testing the snmp using net-snmp-utils” below

Manual Method

Create a file named /usr/local/sbin/cputemp

#!/bin/bash
cpu=$(</sys/class/thermal/thermal_zone0/temp)
echo print $cpu/1000 | /bin/perl
echo

Create a file named  /usr/local/sbin/gputemp

#!/bin/bash
gpu=$(/usr/bin/vcgencmd measure_temp)
echo $gpu | /bin/awk -F '=' {'print $2'} | /bin/awk -F "'" {'print $1'}

Execute:

yum -y install net-snmp

edit the file /etc/snmp/snmpd.conf file and add the following two lines to the end

extend cputemp /usr/local/sbin/cputemp
extend gputemp /usr/local/sbin/gputemp

Scroll up the snmpd.conf file and find the lines that start with

view systemvew included.....

and add the following line:

view systemview included .1.3.6.1.4.1.8072.1.3.2

save your changes and exit your editor

execute

systemctl enable snmpd

edit ipchains to permit inbound udp traffic to port 161

firewall-cmd --zone=public --add-port=161/udp --permanent
firewall-cmd --reload

Important: Restart the pi now

Testing the snmp using net-snmp-utils

replace “public” with your community string

[user@mymachine ~]$ snmpwalk -v 1 -c public my.pi.ip.address nsExtendOutLine
NET-SNMP-EXTEND-MIB::nsExtendOutLine."cputemp".1 = STRING: 49.388
NET-SNMP-EXTEND-MIB::nsExtendOutLine."gputemp".1 = STRING: 49.9
[user@mymachine ~]$

Depending on your snmp configuration you may get data from other points in the snmp besides the above two

Cacti Graph Template

Download this zip file, unzip it and import it into cacti, then you can quickly graph the CPU and GPU temperatures