Macbook Sensor Monitoring

This project utilizes Elixir to monitor CPU and GPU usage on a MacBook. It employs Powermetrics to gather this data. Using a pub-sub approach with RabbitMQ and a fanout exchange, two consumers are set up to aggregate the data into a PostgreSQL database. Another service periodically reads from the database to detect any anomalies in the system's usage.

Usage

To run the project, you need to have Docker and Docker Compose installed, as well as Elixir.

brew install elixir
mix deps.get

Running mix deps.get installs the dependencies listed in mix.exs.

docker-compose up

This command starts RabbitMQ and the PostgreSQL database.

To initialize the database, execute the following commands:

mix ecto.create
mix ecto.migrate

Before starting the server, create a .env file with the following content:

password=YOUR_SUDO_PASSWORD

This is necessary to run the powermetrics command.

The project consists of three components:

• GenSensorData: Generates sensor data from powermetrics and sends it to RabbitMQ.
• ReadSensorData: Reads sensor data from RabbitMQ and stores it in the database.
• ThrottlingDetector: Detects throttling events and stores them in the database. You can start each component by running the following commands in different terminals:

iex -S mix run --no-halt -e "GenSensorsData.start_link"
iex -S mix run --no-halt -e "ReadSensorsData.start_link(\"CPU Power\")"
# or/and
iex -S mix run --no-halt -e "ReadSensorsData.start_link(\"GPU Power\")"

iex -S mix run --no-halt -e "ThrottlingDetector.start_link"

For a "real" deployment, you can add in the application supervisor the children to start the components automatically. In the rabbit_iot/application.ex file, add the following children:

children = [
    Sensors.Repo,
    {GenSensorsData, []},
    Supervisor.child_spec({ReadSensorsData, "CPU Power"}, id: :consumer_1),
    Supervisor.child_spec({ReadSensorsData, "GPU Power"}, id: :consumer_2),
    ThrottlingDetector
]

Then, start the application with:

iex -S mix