Aquaponic Greenhouse

The goal of this repository is to share what I've created to monitor and automate my aquaponic greenhouse. I tried to use docker almost everywhere to easily replicate/automate the setup and let everyone use it in a simple way.

The setup

I have a raspberry pi 3 in my greenhouse with few sensors connected through I2C on it. It currently monitors :

• pH
• Fishtank temperature
• Humidity
• Greenhouse temperature

Using a bridge to flower power it also monitors the light, the soil moisture, and fertilizer. It uses this project: https://github.com/Parrot-Developers/node-flower-bridge.

On the server side, because we need to store metrics and display them, I use a small Z83-V as a server. It's hosting a grafana and an influxdb.

How to install it

Server (at home)

The packages

The server needs less things than the client in order to run. We only need docker.

Just install docker on your server. And be sure to have docker-compose

The structure of the server

The server has few files located in greenhouse/server:

• docker-compose.yml: contains all the services to launch and the parameters to give to containers
• env.grafana: This file allow to easily pass environment variables to grafana
• env.influxdb: This file allow to easily pass environment variables to influxdb
• nginx.conf: I chose to expose my grafana to public and add a specific url to access it. I use nginx as a reverse proxy here
• telegraf.conf: It allows to expose server metrics

The configuration

If you don't want to expose publicly your grafana, you can remove the nginx block from the docker-compose.yml file.

As you can see, paths are hardcoded, that means you will have to change the /home/seraf/ by your home path.

The influxdb is a bit special about volumes. It uses two volumes :

• /home/seraf/docker-data/influxdb/data is a directory external to this git repository where all the metrics/data are stored
• /home/seraf/nas/backup is a NFS mount where I do my backups

The backup

I'm using this cron to do my backup everyday on my NAS:

0 0 * * *  cd /home/seraf/greenhouse/server/ && /usr/local/bin/docker-compose exec influxdb influxd backup -database grafana /backup/influxdb

Client (in the greenhouse)

The packages

First, you will need to install your raspberry pi. I won't explain here how to do it, it's up to you to choose the distribution you prefer.

The packages needed are :

• python3
• python3-dev
• pigpio
• git

I let you install docker by yourself as it depends of your distribution, but it's easy to do

Ideally, be sure you have pip3 to install the following packages:

• docker-compose
• i2cdev
• pigpio

Then, clone the repository :

git clone https://github.com/Seraf/greenhouse

The structure of the client

I created few directories in the client directory to sort a bit the different components.

• checks directory contains all the checks (written in python to access the GPIO) to get the sensors values.
• crons directory contains script that will be called as crons by your cron scheduler. It is used for example to recalibrate your pH sensor according the temperature changes.
• docker directory contains the docker image source I used to create the containers I'm currently using.
• flower-power directory should contains the credentials file to use if you have flower power devices. Else, ignore this directory.
• telegraf directory contains the configuration I use to send values to the metrics database.
• docker-compose.yml is the docker-compose file to launch all the containers

The configuration

Telegraf

Telegraf is an agent that will trigger the checks, get the values and send them to the metrics database.

The telegraf/telegraf.conf configuration file will need some changes according your own setup.

The main thing to change is the line :

[[outputs.influxdb]]
  urls = ["http://192.168.1.21:8086"] # required

You should obviously use the ip address of your server configured previously.

[[inputs.exec]]
  ## Commands array
  commands = [
    "/checks/htu21df.py",
    "/checks/flower_power.py",
    "/checks/ezoph.py",
    "/checks/ezotemp.py"
  ]

There is also the checks section to modify according your own sensors/checks.

Docker-compose

The docker-compose.yml file is simple when you already understand the concepts of docker.

It will launch 2 services : telegraf (the agent responsible to transmit data), and flower-power which is a bridge to my flower power devices.

If you don't have flower power devices, I recommend to remove all the lines concerned as it's not necessary (and don't forget to remove the check in the telegraf configuration)

Create the crons

Here are the crons I'm currently using:

13,28,43,58 * * * * /usr/bin/python3 /home/pi/greenhouse/client/crons/ph_calibration.py
0 12,15,18 * * *  /usr/bin/curl -X POST "http://192.168.1.21:8086/write?db=grafana&precision=s" --data-binary 'events title="Nourriture",text="Dose niveau 3"'
0 9 * * *  /usr/bin/curl -X POST "http://192.168.1.21:8086/write?db=grafana&precision=s" --data-binary 'events title="Nourriture",text="Dose niveau 4"'

The first one is to calibrate the pH sensor each 15 minutes according to the temperature change.

Why I didn't use */15 ? Because it's doing a read on the I2C interface at the same time than telegraf. Then, one of them is failing. To avoid that, I'm changing the temperature between two checks.

The two other crons are used to create annotations in my metrics database to have an annotation each time my auto-feeder is giving some food to fishes. Then I'm able to track the changes of other parameters according to the quantity of food delivered.

Launch it !

To launch the client, once configured, it's pretty simple, just execute this command :

pi@serre:~/greenhouse/client $ docker-compose up -d