Enterprise-grade Stream Reliability Platform for MediaMTX
Features • Screenshots • Quick Start • Architecture • API
MTX Toolkit is an enterprise-grade stream reliability management platform designed for MediaMTX. It provides real-time monitoring, auto-remediation, configuration management, and multi-node fleet management. Supports monitoring thousands of cameras simultaneously with full health checks completed in 10 seconds.
| Feature | Description |
|---|---|
| Live Preview | Grid view with thumbnails, hover-to-play HLS preview, click for fullscreen |
| Dual-layer Health Check | Quick check (API, every 10s) + Deep check (ffprobe, every 5min) |
| Real-time Monitoring | Supports 1000+ streams with millisecond-level status updates |
| Auto Remediation | Smart tiered retry with exponential backoff + jitter |
| Fleet Management | Unified multi-node management across environments (dev/staging/prod) |
| Viewer Management | Real-time viewer sessions, filter by protocol/node, kick viewers |
| Config-as-Code | Terraform-style plan/apply workflow |
| Recording Management | Directory scanning, online playback, search & pagination, auto-cleanup & archiving |
| Event Management | Bulk resolve, cleanup old events, clear resolved alerts |
| i18n | Traditional Chinese / English |
Real-time monitoring of all stream status, health distribution, active alerts, and recent events. Includes event management buttons to resolve all alerts, clear resolved events, or cleanup old events.
Grid view of all streams with auto-generated thumbnails. Hover to play live HLS stream, click for fullscreen player with audio controls.
Unified multi-node management showing stream health status (Healthy/Degraded/Unhealthy) for each node.
Complete stream CRUD operations with status filtering, FPS/bitrate monitoring, manual probe & remediation.
Recording file management with directory scanning, online playback (TS→MP4 transcode), search across all pages, pagination, disk usage monitoring, and auto-cleanup.
Real-time viewer session monitoring across all MediaMTX nodes. Shows client IP, protocol (RTSP/WebRTC/RTMP/SRT), connection duration, data transfer, and allows kicking viewers.
| Status | Color | Description |
|---|---|---|
| Healthy | 🟢 | Stream is normal and playable |
| Degraded | 🟡 | Connecting, on-demand standby, or temporarily unavailable |
| Unhealthy | 🔴 | Path doesn't exist or completely offline |
| Unknown | ⚪ | Not yet checked |
┌─────────────────────────────────────────────────────────────┐
│ Quick Check - Primary Monitoring │
│ (every 10 seconds) │
│ ┌─────────┐ ┌─────────────┐ ┌──────────────────┐ │
│ │ MediaMTX │───▶│ API Query │───▶│ ready: true/false │ │
│ │ API │ │ /v3/paths │ │ Status Update │ │
│ └─────────┘ └─────────────┘ └──────────────────┘ │
│ ⬇ All streams in ~0.2s │
└─────────────────────────────────────────────────────────────┘
┌─────────────────────────────────────────────────────────────┐
│ Deep Check - Detailed Diagnostics │
│ (every 5 minutes) │
│ ┌─────────┐ ┌─────────────┐ ┌──────────────────┐ │
│ │ RTSP │───▶│ ffprobe │───▶│ FPS, Resolution, │ │
│ │ Stream │ │ TCP Mode │ │ Codec, Bitrate │ │
│ └─────────┘ └─────────────┘ └──────────────────┘ │
│ ⬇ Parallel execution │
└─────────────────────────────────────────────────────────────┘
| Stream Count | Quick Check Time |
|---|---|
| 200 | ~0.2s |
| 1,000 | ~1s |
| 5,000 | ~5s |
- Docker & Docker Compose
- Running MediaMTX instance
- 2GB+ RAM
git clone <repo-url> mtx-toolkit
cd mtx-toolkit
docker compose up -d| Service | URL |
|---|---|
| Frontend UI | http://localhost:3001 |
| Backend API | http://localhost:5002 |
Add your MediaMTX node via UI or API:
curl -X POST http://localhost:5002/api/fleet/nodes \
-H "Content-Type: application/json" \
-d '{
"name": "main-mediamtx",
"api_url": "http://your-mediamtx:9998",
"rtsp_url": "rtsp://your-mediamtx:8554",
"environment": "production"
}'curl -X POST http://localhost:5002/api/fleet/sync-all┌────────────────────────────────────────────────────────────────┐
│ MTX Toolkit │
├────────────────────────────────────────────────────────────────┤
│ ┌──────────┐ ┌──────────┐ ┌──────────┐ ┌──────────┐ │
│ │ Frontend │ │ Backend │ │ Celery │ │ Celery │ │
│ │ React │ │ Flask │ │ Worker │ │ Beat │ │
│ │ :3001 │ │ :5002 │ │ │ │ │ │
│ └────┬─────┘ └────┬─────┘ └────┬─────┘ └────┬─────┘ │
│ │ │ │ │ │
│ └─────────────┼─────────────┼─────────────┘ │
│ │ │ │
│ ┌──────┴──────┐ ┌────┴────┐ │
│ │ PostgreSQL │ │ Redis │ │
│ │ :5432 │ │ :6379 │ │
│ └─────────────┘ └─────────┘ │
└────────────────────────────────────────────────────────────────┘
│
▼
┌────────────────────────────────────────────────────────────────┐
│ MediaMTX Nodes │
│ ┌─────────────┐ ┌─────────────┐ ┌─────────────┐ │
│ │ Node 1 │ │ Node 2 │ │ Node N │ │
│ │ Production │ │ Staging │ │ Dev │ │
│ └─────────────┘ └─────────────┘ └─────────────┘ │
└────────────────────────────────────────────────────────────────┘
# Quick check all nodes (milliseconds)
POST /api/health/quick-check
# Quick check single node
POST /api/health/quick-check/{node_id}
# Deep probe stream (ffprobe)
POST /api/health/streams/{stream_id}/probe# List nodes
GET /api/fleet/nodes
# Add node
POST /api/fleet/nodes
# Sync node streams
POST /api/fleet/nodes/{node_id}/sync
# Sync all nodes
POST /api/fleet/sync-all# List streams
GET /api/streams
# Remediate stream
POST /api/streams/{stream_id}/remediate# List recordings (with search & pagination)
GET /api/recordings?search=camera1&page=1&per_page=20
# Scan local recording directory
POST /api/recordings/scan
# Request: { "node_id": 1, "force_rescan": false }
# Stream recording (with transcode for browser playback)
GET /api/recordings/{id}/stream
# Download recording
GET /api/recordings/{id}/download
# Trigger cleanup
POST /api/recordings/retention/cleanup# Resolve all unresolved events
POST /api/dashboard/events/resolve-all
# Clear all resolved events
POST /api/dashboard/events/clear-resolved
# Cleanup old events (default: 7 days)
POST /api/dashboard/events/cleanup
# Request: { "days": 7, "resolved_only": false }# List all viewer sessions (with filters & pagination)
GET /api/sessions?node_id=1&protocol=rtsp&page=1&per_page=50
# Get viewer summary statistics
GET /api/sessions/summary
# Get viewers by node
GET /api/sessions/node/{node_id}
# Get viewers by stream path
GET /api/sessions/path/{stream_path}
# Kick a viewer session
POST /api/sessions/kick
# Request: { "node_id": 1, "session_id": "uuid", "protocol": "rtsp" }# Plan config changes
POST /api/config/plan
# Apply config
POST /api/config/apply
# Rollback config
POST /api/config/rollback/{snapshot_id}| Variable | Default | Description |
|---|---|---|
MEDIAMTX_API_URL |
http://localhost:9998 |
MediaMTX API address |
MEDIAMTX_RTSP_URL |
rtsp://localhost:8554 |
MediaMTX RTSP address |
DATABASE_URL |
postgresql://... |
PostgreSQL connection string |
REDIS_URL |
redis://localhost:6379/0 |
Redis connection string |
Edit docker-compose.yml to modify connection settings:
environment:
- MEDIAMTX_API_URL=http://host.docker.internal:9998
- MEDIAMTX_RTSP_URL=rtsp://host.docker.internal:8554| Service | Port |
|---|---|
| Frontend | 3001 |
| Backend API | 5002 |
| PostgreSQL | 15433 |
| Redis | 6380 |
# Start services
docker compose up -d
# View logs
docker compose logs -f backend
# Rebuild frontend
docker compose build frontend && docker compose up -d frontend
# Rebuild backend
docker compose build backend && docker compose up -d backend celery-worker celery-beat
# Stop services
docker compose down
# Full cleanup (including database)
docker compose down -vVerify the node's RTSP URL is correct:
# Check node settings
curl http://localhost:5002/api/fleet/nodes | jq '.nodes[] | {name, rtsp_url}'
# Update RTSP URL
curl -X PUT http://localhost:5002/api/fleet/nodes/1 \
-H "Content-Type: application/json" \
-d '{"rtsp_url": "rtsp://your-mediamtx:8554"}'Celery tasks are optimized for parallel execution. If issues persist:
# Restart Celery
docker compose restart celery-worker celery-beat# Rebuild and restart frontend
docker compose build frontend && docker compose up -d frontend
# Clear browser cache (Ctrl+Shift+R)MIT License