How to Configure Zabbix Agent to Monitor CPU Temperature on Linux

Prerequisite: the monitored host must have Zabbix agent installed with the same version as the server.

1. Install lm_sensors and run sensor detection.

yum install lm_sensors -y
sensors-detect   # answer yes to all prompts

2. Create a shell script ( cpu.sh ) that accepts a metric argument ( cpu1 or cpu2 ) and returns the corresponding CPU temperature or 0 if unavailable.

#!/bin/bash
#by 20231214
metric=$1
case $metric in
cpu1)
  cpu1=$(sensors|grep "Package id 0" |cut -c 17-20)
  if [ "$cpu1" == "" ]; then
    echo 0
  else
    echo $cpu1
  fi
  ;;
cpu2)
  cpu2=$(sensors|grep "Package id 1" |cut -c 17-20)
  if [ "$cpu2" == "" ]; then
    echo 0
  else
    echo $cpu2
  fi
  ;;
*)
  echo -e "\e[033mUsage: sh $0 [cpu1|cpu2]\e[0m"
  ;;
esac

3. Add a UserParameter to the Zabbix agent configuration to call the script.

UserParameter=get_temp_cpu[*] /etc/zabbix/cpu.sh

4. Restart the Zabbix agent service to apply the changes.

systemctl restart zabbix-agent

5. On the Zabbix server, create a monitoring item using the key get_temp_cpu[cpu1] (or cpu2 ) and add it to the appropriate host template.