trunk-recorder-dashboard

A small Node.js web server that subscribes to events from the tr-plugin-mqtt trunk-recorder plugin and serves a live status dashboard in your browser.

Why not the old status server? The original trunk-recorder-status-server (and its companion tr-plugin-websocket-server) relied on Boost.Beast WebSockets built into trunk-recorder. Starting with trunk-recorder v4.7+/v5, changes to the Boost dependencies broke that WebSocket interface (issue #4, issue #1). This dashboard replaces that approach entirely — it uses the standalone MQTT status plugin (which uses the Paho MQTT library instead of Boost) and requires no WebSocket support in trunk-recorder itself.

Features

• Live decode rates per system
• Total calls and captured audio time per system
• Active calls table (talkgroup, tag, unit, length, flags)
• Recorders table, live graph split by SDR device, missed-calls table, and conspicuous no-recorder alerts
• Systems cards (sys_num, type, sysid, wacn, nac, rfss, site)
• Sources cards (driver, device, gain, ranges) from the retained config message
• Audio playback of completed calls (when mqtt_audio: true)
• Recent unit activity log
• Recently ended calls
• Multi-instance support via instance_id
• MQTT + WebSocket connection indicators

Quick start

git clone https://github.com/theficus/trunk-recorder-mqtt-status-server.git
cd trunk-recorder-mqtt-status-server
npm install
./setup.sh

The script prompts for your broker URL, credentials, topic prefixes, and HTTP port, installs dependencies, writes a .env, and starts the server.

Non-interactive:

MQTT_URL=mqtt://broker.local:1883 \
MQTT_USERNAME=user MQTT_PASSWORD=pass \
MQTT_TOPIC=trunk-recorder \
MQTT_UNIT_TOPIC=trunk-recorder/units \
./setup.sh --no-prompt

Install only (no auto-start):

./setup.sh --no-start

After install:

./run.sh                  # loads .env and starts the server
npm start                 # same thing, without .env auto-load

Then open http://localhost:3080.

Demo mode with mock data is available at http://localhost:3080/demo or by adding ?demo=1 to the dashboard URL. It renders synthetic values only and does not use live MQTT data.

Configuration

setup.sh writes a .env (chmod 600) that run.sh sources at start.

Variable	Default	Purpose
PORT	3080	HTTP port
MQTT_URL	mqtt://localhost:1883	Broker URL (mqtt://, mqtts://, ws://, wss://)
MQTT_USERNAME	–	Optional
MQTT_PASSWORD	–	Optional
MQTT_TOPIC	trunk-recorder	Match topic in your plugin config
MQTT_UNIT_TOPIC	trunk-recorder/units	Match unit_topic
MQTT_MESSAGE_TOPIC	trunk-recorder/messages	Match message_topic (if set)
MQTT_AUDIO_ENABLED	false	Enable subscription to <topic>/audio and on-disk caching
RECORDINGS_DIR	./recordings	Where audio files are written (served at /recordings/)
AUDIO_RETENTION	100	Max recent audio entries to keep (older files are pruned)

Audio playback

Set mqtt_audio: true in the trunk-recorder plugin config and MQTT_AUDIO_ENABLED=true in the dashboard. The dashboard subscribes to <MQTT_TOPIC>/audio, decodes audio_wav_base64 / audio_m4a_base64 to files under RECORDINGS_DIR, and serves them with an <audio> player in the "Recent Audio" section.

⚠️ Audio messages are large (a 30s WAV is ~700 KB base64). Make sure your broker's message_size_limit is high enough. Mosquitto: set message_size_limit 10485760 (or larger) in mosquitto.conf.

trunk-recorder plugin config

Add this to your trunk-recorder config.json (broker URL uses tcp://, the dashboard uses mqtt://; both go to the same broker):

{
  "plugins": [
    {
      "name": "MQTT Status",
      "library": "libmqtt_status_plugin.so",
      "broker": "tcp://your-broker:1883",
      "topic": "trunk-recorder",
      "unit_topic": "trunk-recorder/units",
      "username": "",
      "password": "",
      "console_logs": true,
      "mqtt_audio": false,
      "mqtt_qos": 0
    }
  ]
}

Topics handled

Status (MQTT_TOPIC/...): rates, config, systems, system, calls_active, recorders, recorder, call_start, call_end, plugin_status, console.

Unit (MQTT_UNIT_TOPIC/...): call, end, on, off, ackresp, join, data, ans_req, location.

Trunking (MQTT_MESSAGE_TOPIC/...): messages.

See the official example messages for full payload schemas.

How it works

1. server.js connects to MQTT and subscribes to the three topic prefixes.
2. Each message is parsed and merged into an in-memory per-instance_id state.
3. State is broadcast to browsers over Socket.IO.
4. public/app.js re-renders on every update.

No database, no audio handling — just status.

Requirements

• Node.js 18 or newer
• An MQTT broker (e.g. Mosquitto) reachable from wherever you run the dashboard
• A trunk-recorder instance with libmqtt_status_plugin.so enabled

License

MIT — see LICENSE.