/RPiDeviceInfo

Basic Linux service to monitor and publish basic Raspberry Pi device telemetry via MQTT.

Primary LanguagePython

RPiDeviceInfo

Basic Linux service to monitor and publish basic Raspberry Pi device telemetry via MQTT.

How It Works

The Python script runs as a systemd service and utilizes the psutil Python library to collect device data from the Raspberry Pi including CPU temperature, CPU load, and memory utilization. Also, it utilizes built-in functionality to determine the hostname and uptime for the Raspberry Pi. Telemetry data is published to an MQTT server utilizing the Eclipse Paho MQTT Python library.

Please note: The script does assume an MQTT server is listening at the specified address and port in the function mqtt_publish_single. Eclipse Mosquitto runs really well on Raspberry Pi if you are in need of setting up your own.

Installation

The repository includes an installation script that assumes several steps have been completed:

  1. The pi user has created the directories necessary to support the path /var/pyapps/device
    1. mkdir /var/pyapps
    2. mkdir /var/pyapps/device
  2. The pi user has cloned the repository to the folder /var/pyapps/device
    1. git clone https://github.com/dad2cl3/RPiDeviceInfo.git .
  3. The Raspberry Pi has python3 and pip3 pre-installed and the superuser has installed virtualenv.
    1. sudo pip3 install virtualenv --upgrade
  4. The pi user has created a virtual environment titled "venv" at the path /var/pyapps/device/venv
    1. cd /var/pyapps/device
    2. virtualenv venv
  5. The pi user has activated the virtual environment created in step 4.
    1. cd /var/pyapps/device
    2. source venv/bin/activate
  6. The pi user has installed the necessary dependencies utilizing pip
    1. pip3 install -r requirements.txt --upgrade
  7. The pi user has made the Korn shell script executable after cloning
    1. chmod +x device-info-install.sh

Once those steps are complete, it's as simple as executing the Korn shell script:

(venv) pi@PiDesktop:/var/pyapps/device $ ./device-info-install.sh

The Korn shell script does the following:

  1. Creates symlink /etc/systemd/system/device-info.service to the local service file /var/pyapps/device/device-info.service
  2. Reloads the systemd daemon in order to recognize the new service file
  3. Enables the new service within systemd
  4. Starts the new service within systemd

Testing

The easiest way to test the service is to install an MQTT client of your choice. The Raspberry Pi package mosquitto-clients can be installed as follows: pi@PiDesktop:~ $ sudo apt install mosquitto-clients -y

Should you choose to utilize the mosquitto-clients package on a Raspberry Pi, the command to listen to messages from the service is as follows:

pi@PiDesktop:~ $ mosquitto_sub -h localhost -t RPiDeviceInfo/# -q 1

where the -h switch is followed by the address of your MQTT server and the -t switch is followed by the hostname of the Raspberry Pi where the service is running.

The messages published by the service are in JSON format:

{
    "timestamp": "2021-03-01 16-57-53 -0500",
    "hostname": "PiDesktop",
    "address": "192.168.0.80",
    "uptime": {
      "days": 18,
      "hours": 3,
      "minutes": 49,
      "seconds": 5.59
    },
    "temperature": 5.5,
    "cpu_load": 7.6,
    "memory_load": 40.3,
    "disk_usage": 43.9
}

Removal

The Korn shell script, device-info-uninstall.sh, will remove the service and code from your Raspberry Pi. It assumes the following:

  1. The pi user has made the Korn shell script executable after cloning 
    1. 1. chmod +x device-info-uninstall.sh

Once those steps are complete, it's as simple as executing the Korn shell script:

(venv) pi@PiDesktop:/var/pyapps/device $ ./device-info-uninstall.sh

Node-RED and Blynk

Telemetry data can be processed any number of ways. This repository utilizes Node-RED running locally on the Raspberry Pi itself. The Node-RED import file node-red.json will allow someone to easily import the necessary nodes assuming the following:

  1. Node-RED is running locally on the Raspberry Pi where telemetry is being collected.
  2. The library node-red-contrib-blynk-ws has been added to the palette within Node-RED.
  3. The user has setup a project in Blynk, generated the application key, and added that application key to the Node-RED configuration file prior to import.

The Node-RED import should yield a flow that resembles the following screenshot: Node-RED flow example

Blynk has a multitude of options that users can leverage to build applications to their liking. This repository utilizes a Device Selector, four gauges, and three Value Displays with the following PIN settings:

  1. V0 - CPU Temperature
  2. V1 - CPU Utilization
  3. V2 - Memory Utilization
  4. V3 - Disk Usage
  5. V4 - Last Reading
  6. V5 - IP Address
  7. V7 - Gateway IP Address
  8. V8 - CPU Throttle Status
  9. V9 - Remote Command Terminal
  10. V10 - Uptime
  11. V11 - Styled Button (Clear remote terminal)
  12. V15 - Styled Button (Clear CPU throttle status)
  13. Device Selector
  14. Notifications

Blynk Device Selector

The Blynk application supports many Raspberry Pi devices in a single application. The Device Selector handles all of the devices by assigning a unique key to each device used in the application. As a result, the Node-RED flow will require a unique Blynk connection for each device to be viewed in the Blynk application. The gauges and value displays will change as the device is changed in the Device Selector. Open Blynk Device Selector Each device is added to the Blynk Device Selector in order to generate a unique Blynk key for each device. Open Blynk Device Selector configuration Each Blynk widget also needs to be configured to use the Blynk Device Selector. Open Blynk Gauge configuration to use Device Selector

The screenshot of the Blynk application on an iPhone 11 Pro XS Max: Blynk application