kostal2mqtt

publish data from a kostal pv inverter to mqtt broker

This Python scripts grabs content of the html Webinterface of a Kostal PIKO 10.1 Inverter and publishes the data to an MQTT-Broker.

Why MQTT?

There are plenty of solutions available to directly record the data of a inverter with e.g. InfluxDB or a SQL Database.

MQTT was chosen, because it is (from my point of view) the most flexible solution to process the data further on. In my use case, there is a telegraf instance running, that forwards all MQTT data to InfluxDB, thus no additional Interface from "field level" to the monitoring instance. A nice benefit is the possibility to integrate the MQTT stream into Node-Red or any other automation system such as OpenHAB etc.

It was inspired by the following:

https://github.com/gieljnssns/KostalPyko

http://www.steves-internet-guide.com/into-mqtt-python-client/

https://github.com/svijee/kostal-dataexporter (for cotainerization)

extensive refactoring was of the service was based on https://github.com/karrot-dev/fritzinfluxdb

Setup

** CAUTION** From IT_Security Point-of-View, the whole script might be a nightmare! Feel free to enhance and contribute!

Verify access to inverter

As a first step, try to get access with the user credentials to the Inverter frontend with the following Link from your browser: http://<KOSTAL_HOST>/index.fhtml

Add credentials to the config.ini-file

copy the file with the cmd cp config_template.ini config.ini and adjust all the required configuration settings to your config.ini-file

create docker container

building the container docker build -t kostal2mqtt:latest .

docker run kostal2mqtt:latest kostal2mqtt:latest

building the container with docker-compose docker-compose up

• Run python3 kostal2mqtt.py

There's also a Docker Image available on Docker Hub. Note: The Docker-Container runs pretty stable on a amd64 acrchitecture. For arm architectures such as Raspberry Pi, the container is not yet running. in the Dockerfile you might ùncomment FROM arm32v7/python:3.8-slim-buster and give it a try.

Description or the runtime sequence

• get data
• convert to json
• create mqtt message
• publish message

Missing features:

• exception handling implemented only rudimentary, enhancements are on my todo-list
  • connection errors
    • route to PV
    • route to MQTT Broker
  • data inconsistency
  • etc. etc.
• interface to docker logging console

Problems to be solved:

• Docker container environment ==>
• how to make sure that the latest lib is used?
• runtime monitoring shall provide (detailed) information on the issue causing the container to break.

use cases for data consumption

Grafana

By logging the data with this script it's easily possible to create a nice Grafana Dashboard to display some of the interesting data.

NODE-RED

... to be developped

Note

This is just a quick-and-dirty script to grab to content of the of my Kostal Piko 10.1 Inverter. This might be usable on other Inverters aswell.