LoRaWAN stack Docker setup

This repository contains a skeleton setting up the LoRa Server, PostgreSQL, Node-RED and Grafana project using docker-compose.

Note: Please use this docker-compose.yml file as a starting point for testing but keep in mind that for production usage it might need modifications.

Directory layout

• docker-compose.yml: the docker-compose file containing the services

• configuration/lora*: directory containing the LoRa Server configuration files, see:

  • https://www.loraserver.io/lora-gateway-bridge/install/config/
  • https://www.loraserver.io/loraserver/install/config/
  • https://www.loraserver.io/lora-app-server/install/config/
  • https://www.loraserver.io/lora-geo-server/install/config/

• configuration/postgresql/initdb/: directory containing PostgreSQL initialization scripts

Persistent storage

Docker volumes are used for persistent storage of data from the different apps and are defined in docker-compose.yml

• postgresql-data: volume containing the PostgreSQL data
• redis-data: volume containing the Redis data
• grafana-data: volume containing the Grafana data
• nodered-data: volume containing the Node-RED data

Configuration

LoRa Server

The LoRa Server components are pre-configured to work with the provided docker-compose.yml file and defaults to the EU868 LoRaWAN band. Please refer to the configuration/loraserver/loraserver.toml configuration file to configure a different band.

Grafana

Add plugins to Grafana in docker-compose.yml under GF_INSTALL_PLUGINS. here's available plugins: https://grafana.com/plugins

PostgreSQL

The DB needs a bit of tuning to increase performance which is automatically done by timescaledb-tune here: configuration/postgresql/initdb/005-postgresql-tuning.sh

Requirements

Before using this docker-compose.yml file, make sure you have Docker installed.

Security

In order to secure the stack following passwords need to be changed before running dokcer-compose up:

• postgres PostgreSQL user here: docker-compose.yml
• iot PostgreSQL user here: configuration/postgresql/initdb/004-init-iot-db.sh
• Grafana readonly PostgreSQL user here: configuration/postgresql/initdb/004-init-iot-db.sh
• Grafana admin user: docker-compose.yml
• Node-RED enable admin auth: data/nodered/settings.js (Currently needs to be done after docker-compose up and entails docker restart nodered)

TODO:

• HTTPS
• Default Node-RED auth

Usage

When configuration and security steps are done start all the LoRaWAN stack components by simply running:

$ docker-compose up

Note: during the startup of services, it is normal to see the following errors:

• ping database error, will retry in 2s: dial tcp 172.20.0.4:5432: connect: connection refused
• ping database error, will retry in 2s: pq: the database system is starting up

After all the components have been initialized and started, you should be able to open http://localhost:8080/ (LoRa Server), http://localhost:3000/ (Grafana) and http://localhost:1880/ (Node-RED) in your browser.

Getting started

First gateway and device on Lora Server

Here's a tutorial for setteing up first gateway and device

Add network-server

When adding the network-server in the LoRa App Server web-interface (see network-servers), you must enter loraserver:8000 as the network-server hostname:IP.

First timeseries table in PostgreSQL

TimescaleDB is used to handle large reading and writing of data.

CREATE TABLE water_level (
	ts timestamptz NOT null,
	pressure float NOT NULL,
	temperature float NOT NULL,
	battery float NOT NULL,
	device_id int4 NULL
);

-- Create the hypertable
SELECT create_hypertable('water_level', 'ts');

First flow in Node-RED

We will create a flow which takes device data from Lora server and writes the data to table in PostgreSQL. In order to read and write data to PostgreSQL, we can make use of the node-red-contrib-postgres nodes, which can be installed under Manage palette in the top left menu. From here we need to:

1. Get data from Lora server.

• Make sure to register a new HTTP integration for the Application on Lora server and point the Uplink data url to http://nodered:1880/<nodered-http-endpoint>

2. Prepare data for table insert in postgres.
3. Write the data to the database.

You can simply refer to postgresql as the host when setting up the connection details in the postgres node.

Here's a bit of flow inspiration, which can easily be imported to yout Node-RED in the top left menu:

[{"id":"a7bcd21.0d4a33","type":"http in","z":"1914bc60.4cc104","name":"","url":"/iot/gps","method":"post","upload":false,"swaggerDoc":"","x":110,"y":60,"wires":[["880cbd3e.d8045","32d5646a.05929c"]]},{"id":"23f16f1d.91b9f","type":"template","z":"1914bc60.4cc104","name":"format query","field":"payload","fieldType":"msg","format":"handlebars","syntax":"mustache","template":"insert into gps(time, latitude, longitude, altitude, temperature, battery) \nvalues (to_timestamp($time/1000.0), $latitude, $longitude, $altitude, $temperature, $battery)","x":490,"y":60,"wires":[["da29f2e3.cec04","590fc4ed.76e5cc"]]},{"id":"880cbd3e.d8045","type":"function","z":"1914bc60.4cc104","name":"setup params","func":"var data = msg.payload.object\n\n\nmsg.queryParameters = msg.queryParameters || {};\nmsg.queryParameters.time = data.timestamp;\nmsg.queryParameters.latitude = data.latitude;\nmsg.queryParameters.longitude = data.longitude;\nmsg.queryParameters.altitude = data.altitude; \nmsg.queryParameters.temperature = data.temperature;    \nmsg.queryParameters.battery = data.battery;\n\n\n\n\nreturn msg;","outputs":1,"noerr":0,"x":300,"y":60,"wires":[["23f16f1d.91b9f"]]},{"id":"da29f2e3.cec04","type":"postgres","z":"1914bc60.4cc104","postgresdb":"17235061.c3afa","name":"iot db","output":false,"outputs":0,"x":650,"y":60,"wires":[]},{"id":"590fc4ed.76e5cc","type":"debug","z":"1914bc60.4cc104","name":"","active":false,"tosidebar":true,"console":false,"tostatus":false,"complete":"payload","x":670,"y":100,"wires":[]},{"id":"32d5646a.05929c","type":"http response","z":"1914bc60.4cc104","name":"","statusCode":"200","headers":{},"x":280,"y":100,"wires":[]},{"id":"17235061.c3afa","type":"postgresdb","z":"","hostname":"postgresql","port":"5432","db":"iot","ssl":false}]

First dashboard in Grafana

1. Add PostgreSQL datasource. Hostname is postgresql and it is reommendeed to use the grafanareader read only user.
2. Create a new dashboard and add a Graph panel.
3. Edit panel and make sure grafanareader has privileges to read from the table.
4. Under Metrics pane it is possible to query the table using a Query Builder or plain SQL. Grafana needs time and metric column in order to generate the gragh.

References

• TimescaleDB
• Node-RED
• Grafana
• Lora Server