Skip to content

nodeMCU based moisture monitoring for plants 🌱 with a serverless backend


Notifications You must be signed in to change notification settings


Repository files navigation


Greenkeeper badge Build Status

nodeMCU based moisture monitoring for plants 🌱 with a serverless backend

Main Features

  1. Collect sensor readings every 15 minutes
  2. Deep sleep to preserve battery
  3. Submit values to serverless infrastructure using local WiFi network
  4. Automatically fetch weather report for each reading using backend function
  5. View sensor data on web page
  6. Movable navigation panel
  7. Mobile friendly page

Webpage demo


System Overview

system overview


The hardware setup consists of a ESP8266 Microcontroller with a powersupply and two connected sensors.

  • DHT22 connected to D2 digital input pin
  • Soil moisture sensor connected to A0
  • Bridge between D0 and RST to enable deep sleep for optimizing battery life

To program the ESP8266 you need to clear the connection to RST

Fritzing diagram

Power consumption

  • ESP8266 draws ~80mA when active and few µA when in deepsleep
  • Maximum output current per GPIO pin is 12mA
  • Photo resistor has 10kOhm resistor -> 3V3 / 10kOhm = 33mA
  • Moisture sensor consumes 35mA
  • DHT222 max current 2.5mA


The ESP8266 chip is programmed with the esp8266.ino file using the Arduino IDE. Secrets for connecting to local WiFi and endpoints to post data need to be defined in a esp8266.secrets.c file which is specified at esp8266.secrets.template.c.

esp8266 flowchart


Serverless backend

As a backend solution the plantbuddy runs on Google's firebase 🔥 platform. The main functionalities used from firebase are:

  1. Firestore database to store all sensor readings and heartbeats
  2. Firebase functions for providing an HTTP endpoint as well as querying for weather data for each sensor reading

Note: As outbound traffic is required to collect weather data, the project should run in BLAZE plan. Function and database usage should be within the free tier, only outbound traffic will be billed.

Backend code

Before deploying the backend code several variables need to be set so the service can work correctly. An API token for OpenWeatherMap can be obtained from their website. A free account should suffice unless a LOT of plantbuddies are running on one account.

firebase functions:config:set owm.key=<OpenWeatherMap API Token>
firebase functions:config:set owm.city_id=2873776 // according to [OWM City ID List](
firebase functions:config:set header.secret=<some secret>

Local development

For developing the function locally:

# get a copy of the environment config
firebase functions:config:get > functions/.runtimeconfig.json
# start function host locally
firebase serve --only functions


All the Firebase stuff gets automatically deployed via Travis CI. To set it up create a encrypted token and update the Travis CI Config.

# create a firebase CI token
firebase login:ci

# encrypt the token to your project
docker run --rm caktux/travis-cli encrypt "<token from firebase login:ci>" -r anoff/plantbuddy

Useful links