Lab: Node-RED dashboard for sensor data

Aim and what you need

The aim is a live web dashboard of the temperature and humidity your ESP32 publishes. Node-RED represents logic as flows: boxes called nodes joined by wires. It is popular for IoT prototypes because you can connect MQTT, HTTP, databases and dashboards quickly.

• A Raspberry Pi or Linux machine with Docker.
• The secured broker from the MQTT TLS lab, with a dashboard user that can read lab/#, and the lab CA certificate.
• A sensor publishing JSON such as {"t": 21.6, "h": 48} to lab/sensor1/env, or mosquitto_pub to fake it.

Steps

1. Start Node-RED with a named volume so flows survive restarts.
2. Generate a password hash and add an adminAuth block to settings.js in the data volume, then restart.
3. Open http://<pi-address>:1880, log in, and from Manage palette install the dashboard package (Dashboard 2.0 is published as @flowfuse/node-red-dashboard).
4. Drag in an mqtt in node. Add a broker: host broker.lab.example, port 8883, enable TLS with your CA file, and the dashboard username and password. Topic lab/+/env, output: parsed JSON object.
5. Add a change or function node that sets msg.payload to the temperature, then wire it to a gauge and a chart. Repeat for humidity.
6. Deploy, then open the dashboard URL shown in the sidebar.

docker volume create node_red_data
docker run -d --name nodered --restart unless-stopped -p 1880:1880 -v node_red_data:/data nodered/node-red

npx node-red-admin hash-pw

docker exec -it nodered sh
vi /data/settings.js
#   adminAuth: {
#     type: "credentials",
#     users: [{ username: "admin", password: "<bcrypt-hash>", permissions: "*" }]
#   },
exit
docker restart nodered

# function node body
msg.payload = Number(msg.payload.t);
msg.topic = "temperature";
return msg;

# test message
mosquitto_pub -h broker.lab.example -p 8883 --cafile ca.crt -u sensor1 -P '<sensor1-password>' -t lab/sensor1/env -m '{"t":21.6,"h":48}'

How to verify and troubleshoot

The mqtt in node should show a green connected label. Add a debug node to see each message in the sidebar. The gauge should move when you publish a test value, and the chart should build a line over a few minutes. Open the editor in a private browser window: it must ask for a login.

• Node shows disconnected: wrong port, CA file or credentials, or the name does not match the broker certificate.
• Gauge shows nothing: the payload is still an object; check with a debug node and convert to a number.
• Editor still open without login: the adminAuth block is commented out or in the wrong file; check the container logs.
• Lost flows after update: you forgot the volume; always mount /data.

Security notes and clean-up

The Node-RED editor can run JavaScript and shell commands, so an exposed editor is a full compromise of the host. Keep it on the LAN or behind a VPN, use HTTPS through a reverse proxy if anyone else uses it, install only well-maintained nodes and export your flows (menu, Export) to Git, leaving credentials out. To remove the lab: docker rm -f nodered and, if you are sure, docker volume rm node_red_data.

Tip: a screenshot of a working dashboard plus a diagram of the data path makes a strong portfolio entry; see the IoT monitoring project.