epleterte/sensehat-exporter

Prometheus metric exporter for the Raspberry Pi SenseHat

Sensehat Prometheus Exporter

A simple Prometheus exporter for exporting values from the Raspberry Pi SenseHat.

Requirements

• Raspberry Pi (3/3+/4)
• Rasperry Pi SenseHat

Installation

git clone https://github.com/epleterte/sensehat-exporter.git
cd sensehat-exporter
pip3 install -r requirements.txt
sudo cp sensehat-exporter.py /usr/local/bin/sensehat-exporter
# start sensehat-exporter!
sensehat-exporter

Run as a service

cp sensehat-exporter.service.example /etc/systemd/system/sensehat-exporter.service
systemctl daemon-reload
systemctl start sensehat-exporter
systemctl enable sensehat-exporter

Usage

usage: simpleexporter.py [-h] [--port [PORT]] [--bind [BIND]]

optional arguments:
  -h, --help     show this help message and exit
  --port [PORT]  The TCP port to listen on (default: 9101)
  --bind [BIND]  The interface/IP to bind to (default: 0.0.0.0)
  --orientation  Enable 'orientation' data (roll, yaw, pitch)

Configuration

Currently there are no other configuration than options IP/port and no configuration file.

Prometheus Configuration

In /etc/prometheus/prometheus.yml, Add a static scrape target under scrape_configs:

scrape_configs:
  - job_name: 'sensehat'
    static_configs:
    - targets: ['<ip>:9101'] 

TODO

• Upload example Grafana dashboard
• Export more sensor data - currently only temperature, humidity and pressure is exported.
• Possibly make additional sensors configurable.

Demo

Here's a screenshot of the data being put to use in a Grafana dashboard:

License

This software is licensed under the ISC license.

sensehat-exporter's Issues

Grafana error using exporter metrics

Using Prometheus 2.33.3 and Grafana 8.3.3. Scraping job for the sensehat-exporter is running and generating metrics. However, when in Grafana to define a Prometheus data source, I set the URL to localhost:9101, click Save and Test and receive the following error:

Error reading Prometheus: bad_response: readObjectStart: expect { or n, but found #, error found in #1 byte of ...|# HELP pyth|..., bigger context ...|# HELP python_gc_objects_collected_total Objects co|...

Apparently the outputted metrics do not follow the current definition for such data. Any help would be appreciated.