WarDragon Analytics

February 3, 2026 · View on GitHub

Centralized logging, analysis, and visualization platform for drone detection data from one or more WarDragon kits running DragonSync.

Important: This is designed to run on a separate server or workstation, not on the WarDragon kit itself. While it can technically run on a kit, the resource overhead (TimescaleDB, Grafana) is better suited for a dedicated machine. For lightweight on-kit history, a future integration with the WarDragon ATAK Plugin is planned to provide historical track data directly to TAK users without requiring a full analytics stack.

⚠️ Security Notice: This project is under active development. While every effort is made to review code for security issues, exposing the web UI or API to the public internet without a VPN is not recommended. Note that WarDragon kits must be network-reachable for HTTP polling, which means the DragonSync API would also need protection—you wouldn't want to expose that to the public either. For environments requiring data exchange across untrusted networks, MQTT with TLS encryption (port 8883) may offer the most secure transport option, though this feature is currently untested in production. See SECURITY.md for hardening recommendations.

Overview

WarDragon Analytics aggregates DroneID/Remote ID drone detections, ADS-B aircraft tracks, and FPV signal detections from multiple WarDragon field kits into a single interface. It provides:

  • Real-time map display of all drone and aircraft tracks across all kits
  • Time-series database (TimescaleDB) with 30-day retention and 1-year aggregates
  • Pattern detection to identify surveillance behavior, coordinated swarms, and anomalies
  • Pre-built Grafana dashboards for tactical operations and analysis
  • REST API for integration with other systems
  • CSV export for reporting (KML planned)
  • AI Assistant for natural language queries (optional, requires Ollama)

WarDragon Analytics Web UI

Architecture

Option 1: HTTP Polling (Default)

WarDragon Kits (Field)     Analytics Server (Docker)      User Interfaces
+-----------------+        +----------------------+       +---------------+
| DragonSync API  |  <---  | Collector Service    |  ---> | Web UI :8090  |
| :8088           |  poll  | TimescaleDB          |       | Grafana :3000 |
+-----------------+        +----------------------+       +---------------+

Option 2: MQTT Push (Optional)

WarDragon Kits (Field)     Analytics Server (Docker)      User Interfaces
+-----------------+        +----------------------+       +---------------+
| DragonSync      |  push  | MQTT Broker :1883    |  ---> | Web UI :8090  |
| MQTT Sink       |  --->  | MQTT Ingest Service  |       | Grafana :3000 |
+-----------------+        | TimescaleDB          |       +---------------+
                           +----------------------+

Data collected from each kit:

  • Drones via DJI DroneID (OcuSync) and Remote ID (Bluetooth, Wi-Fi Beacon/NAN)
  • Aircraft via ADS-B
  • FPV signals (5.8GHz analog video) (untested)
  • Kit system health (CPU, memory, disk, temperature, GPS position)

Quick Start

Prerequisites: Docker and Docker Compose (see docs/quickstart.md for Ubuntu installation instructions)

# Clone the repository
git clone https://github.com/alphafox02/WarDragonAnalytics.git
cd WarDragonAnalytics

# Run the quickstart script
./quickstart.sh

The quickstart script will:

  1. Generate secure passwords and create .env
  2. Build and start all Docker containers
  3. Initialize the database schema
  4. Configure Grafana with pre-built dashboards

Access the interfaces:

Configure your kits:

There are two ways to add WarDragon kits:

Option A: Via Web UI (Recommended)

  1. Open http://localhost:8090
  2. Click "Kit Manager" in the sidebar
  3. Enter the kit's API URL (e.g., http://192.168.1.100:8088)
  4. Click "Add Kit" - the collector will start polling immediately

Option B: Via Configuration File

  1. Edit config/kits.yaml:
kits:
  - api_url: "http://192.168.1.100:8088"
    name: "Field Kit Alpha"
    enabled: true
  1. Restart the collector:
docker compose restart collector

Kits added via the Web UI are stored in the database and persist across restarts. Kits in kits.yaml are loaded on collector startup. Both methods can be used together.

Deployment Options

Centralized Server (Recommended)

  • Deploy Analytics on a separate server, workstation, or cloud instance
  • All field kits report to the central server
  • Single pane of glass for all operations
  • Best for multi-kit deployments and operations centers

Per-Kit Local (Testing Only)

  • Can run on the same machine as DragonSync for testing
  • Use http://172.17.0.1:8088 (Linux) or http://host.docker.internal:8088 (Mac/Windows)
  • Not recommended for production due to resource overhead on the kit

Hybrid

  • Central server aggregates from multiple kits
  • Individual operators access via Grafana/Web UI
  • Field units use TAK with DragonSync for real-time situational awareness

Features

Web UI (Port 8090)

  • Interactive Leaflet map with drone, aircraft, and signal markers
  • Drone, pilot, and home location markers (when provided by DroneID/Remote ID)
  • FPV signal detection markers (5.8GHz analog, DJI digital)
  • Track history trails
  • Real-time updates (5-second refresh)
  • Tabbed data table with separate views for Drones, Aircraft, and Signals
    • Sortable columns for each data type
    • Pagination (25/50/100 per page) for large datasets
    • Click any row to zoom to location on map
  • Kit management interface
  • Quick filters for anomalies, repeated contacts, multi-kit detections
  • Light and dark theme support
  • CSV export
  • AI Assistant for natural language queries (optional, requires Ollama)

Enterprise Features (Optional)

  • Authentication: Optional password protection for web UI (JWT-based)
  • Webhook Alerting: Slack, Discord, and generic HTTP webhook notifications
  • Audit Logging: Track admin actions (kit management, exports, logins)
  • Automated Backups: Scheduled database backups with retention policy

All enterprise features are disabled by default and configured via .env. See SECURITY.md for setup.

MQTT Ingest (Optional)

Alternative to HTTP polling - kits push data directly via MQTT:

  • Real-time data: No polling delay, data arrives immediately
  • Auto-registration: Kits appear automatically when they first publish
  • NAT/firewall friendly: Kits initiate outbound connections only
  • Hybrid mode: Use both HTTP polling and MQTT simultaneously

To enable MQTT ingest:

# Enable in .env
MQTT_INGEST_ENABLED=true

# Start with MQTT profile
docker compose --profile mqtt up -d

# Configure DragonSync on kits to publish to this server:1883

See docs/mqtt-ingest.md for detailed setup.

Grafana Dashboards (Port 3000)

Tactical Overview

  • Active drone count and kit status grid
  • Kit health (CPU, memory, disk, temperature)
  • Drone detection timeline
  • Top manufacturers detected
  • Alert summary

Pattern Analysis

  • Repeated drone detections (surveillance indicators)
  • Operator reuse across multiple drones
  • Coordinated activity (potential swarms)
  • Frequency reuse patterns

Multi-Kit Correlation

  • Drones detected by multiple kits (triangulation opportunities)
  • Kit coverage data
  • Detection density heatmap
  • Kit handoff tracking

Anomaly Detection

  • Altitude anomalies (rapid climbs/descents)
  • Speed anomalies
  • Signal strength variations
  • Out-of-pattern behavior

Pattern Detection API

# Drones seen multiple times (surveillance pattern)
curl http://localhost:8090/api/patterns/repeated-drones?hours=24

# Coordinated activity (swarms)
curl http://localhost:8090/api/patterns/coordinated?hours=6

# Operator reuse across drones
curl http://localhost:8090/api/patterns/pilot-reuse?hours=12

# Anomalous behavior
curl http://localhost:8090/api/patterns/anomalies?hours=6

# Multi-kit detections (triangulation)
curl http://localhost:8090/api/patterns/multi-kit?hours=6

Requirements

Server/Host:

  • Docker and Docker Compose
  • 2GB RAM minimum (4GB recommended for multi-kit)
  • 50GB disk (for 30 days of data from 5 kits)

WarDragon Kits:

  • DragonSync with HTTP API enabled (default port 8088)
  • Network connectivity to Analytics host

Documentation

All documentation is located in the docs/ folder.

Getting Started

DocumentDescription
Quick Start GuideGet up and running in 5 minutes
Deployment GuideDetailed deployment for production
ArchitectureSystem design and database schema

User Guides

DocumentDescription
Operator GuideTactical operations workflows
Grafana DashboardsDashboard usage and customization
MQTT Ingest GuidePush data from kits via MQTT
AI Assistant SetupNatural language query interface

Reference

DocumentDescription
API ReferenceREST API documentation
Dashboard QueriesGrafana query reference
Collector ServiceData collection internals
Test Data GeneratorGenerate realistic test data

Operations

DocumentDescription
TroubleshootingCommon issues and solutions
Testing GuideRunning and writing tests
SecuritySecurity hardening and best practices

Common Issues & Quick Fixes

Database schema missing / views not created:

This happens if you ran docker compose up directly without using quickstart.sh, or if the database initialization failed. The init scripts only run on first startup when the database volume is empty.

# Option 1: Full reset (WARNING: -v flag DELETES ALL DATA)
docker compose down -v
./quickstart.sh

# Option 2: Apply schema to existing database (preserves data)
docker exec -i wardragon-timescaledb psql -U wardragon -d wardragon < timescaledb/01-init.sql
docker exec -i wardragon-timescaledb psql -U wardragon -d wardragon < timescaledb/02-pattern-views.sql
docker exec -i wardragon-timescaledb psql -U wardragon -d wardragon < timescaledb/03-extended-fields.sql

Note: Running ./quickstart.sh alone is safe and won't delete data. The -v flag on docker compose down is what removes database volumes.

Containers won't start / port conflicts:

# Check what's using the ports
sudo lsof -i :8090 -i :3000

# Full cleanup and restart
docker compose down -v
docker system prune -f
./quickstart.sh

For detailed troubleshooting, see docs/troubleshooting.md.

Useful Commands

# Start services
docker compose up -d

# Stop services
docker compose down

# View logs
docker compose logs -f collector
docker compose logs -f web

# Restart after config change
docker compose restart collector

# Check service health
./healthcheck.sh

# Database backup
docker exec wardragon-timescaledb pg_dump -U wardragon wardragon > backup.sql

Test Data

Generate realistic test data to explore the dashboards:

python tests/generate_test_data.py --scenario all

Available scenarios: normal, repeated, coordinated, operator, multikit, anomalies, fpv

  • DragonSync - Drone detection and TAK integration for WarDragon kits
  • DragonOS - Linux distribution for SDR and drone detection

License

Apache 2.0

Contributing

  1. Read docs/architecture.md
  2. Open an issue to discuss your idea
  3. Fork, implement, test, submit PR

Support