Script to post CPU load and CPU temp to MQTT

[tubalainen]

Depending on your Frigate setup and how many cameras you are using the machine you are running Frigate on can be tormented with high CPU usage for a long time (if the wind is blowing, lots of movements in the motion areas etc).

Ive created this cron script that posts the CPU temp and CPU usage to MQTT.

EDIT - Or just use Glances instead https://github.com/nicolargo/glances thanks Minglarn!
docker run -d --name glances -e GLANCES_OPT="-w" -v /usr/share/dockers/glances:/glances/conf -v /var/run/docker.sock:/var/run/docker.sock:ro --pid host --network host nicolargo/glances:latest-full

The temperature only works if you are running the OS on directly on your machine, not in a VM.

Pre-reqs
apt install lm-sensors mosqutto-clients coreutils
The script ( i placed it in /root/inform_mqtt_status.sh)
You will need to check that all awk/sed etc are working hence the reporting is different of each machine architecture.

#!/bin/bash
sensors -j | tr -d '[:space:]' > /tmp/cputemp
mosquitto_pub -h 192.168.1.10 -u username -P passwd -t docker05/cputemp -f /tmp/cputemp

# grep 'cpu ' /proc/stat | awk '{usage=($2+$4)*100/($2+$4+$5)} END {print usage ""}' > /tmp/cpuload
awk '{u=$2+$4; t=$2+$4+$5; if (NR==1){u1=u; t1=t;} else print ($2+$4-u1) * 100 / (t-t1) ""; }' <(grep 'cpu ' /proc/stat) <(sleep 1;grep 'cpu ' /proc/stat) > /tmp/cpuload
mosquitto_pub -h 192.168.1.10 -u username -P passwd -t docker05/cpuload -f /tmp/cpuload

Add it to crontab crontab -e (as root, in my case) runs every minute
*/1 * * * * /root/inform_mqtt_status.sh

Sensors in Home Assistant
I use a custom integration called "average", can be found in the HACS. How ever, it doesnt support units_of_measurement so thats why ive got the template sensors.

- platform: mqtt
  state_topic: 'docker05/cputemp'
  name: Docker05 CPU Temp All
  unit_of_measurement: "°C"
  value_template: "{{ value_json['coretemp-isa-0000']['Packageid0']['temp1_input'] }}"
- platform: mqtt
  state_topic: 'docker05/cputemp'
  name: Docker05 CPU Temp Core 0
  unit_of_measurement: "°C"
  value_template: "{{ value_json['coretemp-isa-0000']['Core0']['temp2_input'] }}"
- platform: mqtt
  state_topic: 'docker05/cputemp'
  name: Docker05 CPU Temp Core 1
  unit_of_measurement: "°C"
  value_template: "{{ value_json['coretemp-isa-0000']['Core1']['temp3_input'] }}"
- platform: mqtt
  state_topic: 'docker05/cputemp'
  name: Docker05 CPU Temp Core 2
  unit_of_measurement: "°C"
  value_template: "{{ value_json['coretemp-isa-0000']['Core2']['temp4_input'] }}"
- platform: mqtt
  state_topic: 'docker05/cputemp'
  name: Docker05 CPU Temp Core 3
  unit_of_measurement: "°C"
  value_template: "{{ value_json['coretemp-isa-0000']['Core3']['temp5_input'] }}"

- platform: mqtt
  state_topic: 'docker05/cpuload'
  name: Docker05 CPU Load
  unit_of_measurement: "%"
  value_template: "{{ value_json | round(1) }}"

- platform: average
  name: 'Docker05 average cpu load no unit'
  duration:
    minutes: 15
  entities:
    - sensor.docker05_cpu_load

- platform: average
  name: 'Docker05 average cpu temp no unit'
  duration:
    minutes: 15
  entities:
    - sensor.docker05_cpu_temp_all

- platform: template
  sensors:
    docker05_average_cpu_load:
      friendly_name: 'Docker05 average CPU Load'
      unit_of_measurement: "%"
      value_template: >-
        {{ states.sensor.docker05_average_cpu_load_no_unit.state | round(1) }}
    docker05_average_cpu_temp:
      friendly_name: 'Docker05 average CPU Temp'
      unit_of_measurement: "°C"
      value_template: >-
        {{ states.sensor.docker05_average_cpu_temp_no_unit.state | round(1) }}

[Minglarn]

Nice solution...
However, i ended up with installing glances ..
And after installing the glances integration in HA i've a lot of new sensors watching my ubuntu server...

[tubalainen]

Nice solution...
However, i ended up with installing glances ..
And after installing the glances integration in HA i've a lot of new sensors watching my ubuntu server...

Fräckt!
I wanted to "keep it simple". Do you have any link for "How to"? :)

[tubalainen]

Äh - that was way to easy. :)

Thanks for the tip! Glances the best! <3

docker run -d --name glances -e GLANCES_OPT="-w" -v /usr/share/dockers/glances:/glances/conf -v /var/run/docker.sock:/var/run/docker.sock:ro --pid host --network host nicolargo/glances:latest-full

[Minglarn]

Hahaha... you did that way... It sure works ...
I just followed this guide:
https://www.howtoforge.com/how-to-monitor-system-resources-with-glances-ubuntu-20-04/

😄