weather-server

The weather-server application is a simple HTTP server that can receive temperature readings from multiple devices, and stores them in a PostgreSQL database.

This allows you to use reporting tools (such as Grafana) to create a dashboard to illustrate a history of temperatures at multiple locations.

Setup instructions

Prerequisites

Setting up this server requires the following:

• Server or VPS running Linux (this documentation assumes Ubuntu 22.04)
• A PostgreSQL database with the following:
  • Empty database named weather
  • User granted privileges on database weather
• Docker

Step by step instructions on setting up these prerequisites can be found in the Appendix section further down.

System requirements

If you intend on running this application on the same machine as a PostgreSQL database and a Grafana instance, these minimum specifications are recommended:

• 2 GB RAM
• 10 GB disk

Installation

Git checkout

First download the code from this repository.

git clone https://github.com/sd-hi/weather-server.git

Go into the created weather-server directory and copy the environment variables file .env.example to .env:

cd weather-server
cp .env.example .env

Now amend the .env file with your preferred text editor e.g.

nano .env

Set the variables based on guidance below

Environment variables

Set the following environment variables:

• DB_HOST - Set this to the IP address of your PostgreSQL database
• DB_USER - PostgreSQL database username (e.g. weatheruser)
• DB_PASS - PostgreSQL user's password
• DB_NAME - The database name in which the measurement will be stored (e.g. weather)
• API_KEY - The API key clients will use as the 'password' (passed on requests as a HTTP header x-api-key)

Save your changes and continue to deployment

Deploy in Docker container

Now it is time to launch the web server by creating a docker container for it

Use the following command to build the weather-server contents into a docker container image

sudo docker build -t weather-server .

Run the docker container:

sudo docker run -d -p 5000:3000 --name weather-server weather-server

Express will be running on port 3000 within the container. In this example

• -p 5000:3000 directs port 3000 within the container to port 5000 on our VPS, so we should submit payloads to port 5000 when setting up clients
• --name weather-server will name our Docker container 'weather-server'

The container is now ready to receive API calls from Raspberry Pi devices running the weather-client

Tip: For debugging, you can inspect the live log of container weather-server with command sudo docker logs -f weather-server e.g.

user1@mymachine:~/weather-server$ sudo docker logs -f weather-server

> weather-server@1.0.0 start
> node index.js

Server running on http://localhost:3000

Test function

Test connectivity with the server using the following curl command (replace 1.2.3.4 with server IP address, and api_key_goes_here with the API key configured in environment variable)

curl --location '1.2.3.4:5000/temperatures' \
--header 'x-api-key: api_key_goes_here' \
--header 'Content-Type: application/json' \
--data '{
    "deviceId": "device1",
    "locationId": "location1",
    "measurements": [
        {
            "dateTime": "2023-08-20T01:42:44.042Z",
            "humidity": 0.79,
            "temperature": 26.62
        },
        {
            "dateTime": "2023-08-20T01:42:46.042Z",
            "humidity": 0.79,
            "temperature": 26.62
        }
    ]
}'

A successful response should be returned

{
  "message": "2 measurements added successfully"
}

Appendix

Installing Docker

Docker is a tool to containerize applications, making them easier to manage. For details on how to install Docker see page https://docs.docker.com/engine/install/ubuntu/

To verify Docker is installed correctly you can run command

sudo docker run hello-world

Which should download a hello-world test container and print a confirmation message.

Installing PostgreSQL

Create a PostgreSQL container

If you already have a PostgreSQL server deployed, this sub-section can be skipped.

Pull image from repository

sudo docker pull postgres

Run the container

sudo docker run --name postgres -e POSTGRES_PASSWORD="rootpassword123" -p 5432:5432 -d postgres

This creates a Docker container named postgres containing the PostgreSQL database with a default user postgres whose password is rootpassword123 (choose a stronger password!). The database can be accessed on port 5432.

Log into PostgreSQL

Connect to the sql prompt of created docker container named postgres, as user postgres:

docker exec -it postgres psql -U postgres

Create a database

Log into the PostgreSQL console, create a database for the weather server application to write its data to.

Example below creates a database called weather (this is to be consistent with the environment variable DB_NAME for the weather-server)

CREATE DATABASE weather;

Create a database user weatheruser with password password123 for the weather server to use

CREATE USER weatheruser WITH PASSWORD 'password123';

Grant the user access to everything on the weather database:

GRANT ALL PRIVILEGES ON DATABASE weather TO weatheruser;

Switch to the weather database:

\c weather

Grant the weatheruser user access to the public schema:

GRANT USAGE ON SCHEMA public TO weatheruser;
GRANT CREATE ON SCHEMA public TO weatheruser;
GRANT SELECT ON ALL TABLES IN SCHEMA public TO weatheruser;
ALTER DEFAULT PRIVILEGES IN SCHEMA public GRANT SELECT ON TABLES TO weatheruser;

Using Grafana

Grafana installation

Grafana is a powerful dashboard allowing you to easily view the time series data from the PostgreSQL database For more information see https://grafana.com/docs/grafana/latest/setup-grafana/installation/docker/#use-persistent-storage-recommended Create a persistent volume for Grafana

sudo docker volume create grafana-storage

Verify storage was created correctly

sudo docker volume inspect grafana-storage

Start the Grafana container by running the following command:

sudo docker run -d -p 3000:3000 --name=grafana \
  --volume grafana-storage:/var/lib/grafana \
  grafana/grafana-enterprise

PostgreSQL DB user

Create a DB user grafana with access to the PostgreSQL database. Log into PostgreSQL database SQL prompt

docker exec -it postgres psql -U postgres

Create a grafana user

CREATE USER grafana WITH PASSWORD 'password123';

Grant the user appropriate permissions on the weather database:

GRANT CONNECT ON DATABASE weather TO grafana;

Switch to weather database:

\c weather

Grant the user select access to the public schema:

GRANT USAGE ON SCHEMA public TO grafana;
GRANT SELECT ON ALL TABLES IN SCHEMA public TO grafana;
ALTER DEFAULT PRIVILEGES IN SCHEMA public GRANT SELECT ON TABLES TO grafana;

Grafana setup

Logging in to Grafana

Assuming the Grafana Docker container has been set up on port 3000, navigate to the VPS in web browser (where 1.2.3.4 is address of server)

1.2.3.4:3000

A Grafana login page should be presented The default username and password is admin:admin

Setup PostgreSQL as data source

• Click the menu symbol ≡
• Choose "Data sources"
• On data sources pane, press "Add data source"
• Select PostgreSQL
• Enter the PostgreSQL database credentials for the user created earlier
  • Enter the host as 1.2.3.4:5432 (consistent with VPS IP and PostgreSQL DB port)
  • Database is weather
  • User is grafana (password consistent with creation)
• Click "Save & test" at the bottom You should see a message "Database Connection OK"

Build Grafana dashboard

Tip: Insert some rows into the temperatures table in the PostgreSQL weather database to prove the data points can be seen in the following dashboard. The easiest way to do this is to post some sample payloads to the weather server, see the "Test function" section above

• Click the menu symbol ≡
• Choose "Dashboards"
• Press "New" button
• Choose "New dashboard" from dropdown menu
• Choose "Add visualization"
• Choose PostgreSQL database as the data source

Query pane

In the query pane on the bottom-left:

• Press "+ Query" to add a query
• Choose "Code" mode ("Builder/Code" switch on top-right)
• Enter the following query

SELECT
datetime AS time,
temperature AS "My Office"
FROM temperatures
WHERE $__timeFilter(datetime)
AND locationid = "location1"

The above query will show all data points with a locationid of location1 in the database, the dataset will be labeled "My Office"

Pressing "Run query" button should display the office data Repeat this step, adding a query for for each location you wish to display on the dashboard. Feel free to set an appropriate title and customize to your preference. Press "Apply" button will create the new dashboard. Now you can use this dashboard to view the temperature data submitted to your weather-server!