Skip to content

Latest commit

 

History

History
102 lines (74 loc) · 6.04 KB

README.md

File metadata and controls

102 lines (74 loc) · 6.04 KB
Raw

self_hosted_backend

Project for a docker-compose managed self hosted backend for PFC (openag-device-software based) machines

Project structure

Componenets

There are several components contained in the docker-compose setup

eclipse-mosquitto

The Eclipse Mosquitto project is used as an MQTT broker. Devices will communicate with the broker. A separate python service parses and process the messages from the PFCs. Commands out to the PFCs are also sent via the MQTT broker.

InfluxDB

InfluxDB is a time-series database that holds all the telemetry sent from the PFCs (plus other time based data).

Chronograf

Chronograf is a web based interface to Influx for administration as well as visualization/exploration of the data

mqtt-service

A 'local' mode branch of the OpenAg Mqtt Service is used to parse the data coming in from the PFCs and inserts the data in the InfluxDB database.

upload-file-service

The upload file service is a tiny alpine docker based image running a Flask web application to handle uploading, and downloading of images via http. This is the only non-mqtt based communication from a PFC. Note that this service is completly contained in this project (i.e. not a seperate Github project), but it doesn't have to be run inside a docker-compose setup as it is a very simple Flask application.

DataAPI service

A light weight Flask Data API has been created under the upload-file-service for getting a list of machines seen, their latest readings, and the ability to download data over a span of time in CSV format. Please Note: There is zero authentication or security on this API.

Control UI

A simple ReactJS front end has been included in the upload-file-service docker image to access the information available from the DataAPI. If you point a browser to your host machine's port 80, it will load up. Please Note: There is zero authentication or security on this UI or the underlaying API.

Running

First Start - Server side

  1. Install Perquisites if you have't yet.
  2. docker-compose up will build and start the docker instances. (First time may take a few moments)
  3. Create an openag_local database in InfluxDB via Chronograf (See Below)
  4. Setup dashboards / document using Chronograf with PFC data
  5. Once everything is running, and you have PFCs reporting data, you should be able to point a browser at your server (http://server-ip/) and you'll be presented with a simple UI to see the state of your PFCs and download data.

PFC setup

In order to use this backend with PFCs, you need to run a modified version of the openag-device-software code.

Switch to the local_mqtt branch of the openag-device-software git project.

First you'll want to fetch the remote code if you've already cloned it: From inside the openag-device-software directory run the command: git fetch Then you'll want to checkout the local_mqtt branch: From inside the openag-device-software directory run the command: git branch local_mqtt Now you're ready to configure the software to point to your local backend.

Setup local_mqtt.json for PubSub

Rather than sending MQTT messages up to the Google Cloud, PFCs will want to hit our local mosquitto instance. You'll need to find the IP address for the machine running the self_hosted_backend docker-compose setup. Once you have that, you'll want to change the openag-device-software/device/iot/setups/local_mqtt.json file. The contents of that file should have the IP and port (default of 1883) for the MQTT Broker.

{
 "broker": "192.168.1.27",
 "ports": [1883]
}

Setup the save image URL

Now that the MQTT messages are going to right place, you'll want to change the URL that the PFC will post images to. You'll find that in openag-device-software/app/settings.py, and look for IMAGE_UPLOAD_URL.

In the basic configuration of just running the docker compose file as is, this will be an unencrypted HTTP post, rather than using HTTPS. So, you'll want to set it to something like:

IMAGE_UPLOAD_URL = 'http://192.168.1.27/saveImage'

At this point, you're ready to run the software on your PFCs. If you're running balena, you can push the firmware up through it's command line interface, or if you're deployed the code directly onto your RaspberryPi computer, you'll want to restart the device software on the machine. Just make sure you've already configured your backend by creating the openag_local database inside of InfluxDB.

Creating a DB with Chronograf

Once you've got the docker-compose setup running. You can go to Chronograph on you server's prot 8888. Initial Chronograf

Once you click Get Started, you'll be presented with a screen asking for connection information. Set Connection URL to http://influxdb:8086. You can name the connection whatever you want, in the image it is OpenAg Local DB. Since we are not using Telegraf, you can leave that field blank.

Connection Information

After adding the connection, you can skip adding a Dashboard or the Kapacitor connection, and just click View All Connections. You'll now want to click InfluxDB Admin on the sidebar.

InfluxDB Admin

On the admin screen, click + Create Database, and enter openag_local as the name. Add Database

You are now set to start sending data from the PFCs. Once you have data in your DB, you can use the Explore side tab to start looking at your data with graphs and queries. (See Getting Started with Chronograf)