feat: add Netdata monitoring stack

Added Netdata real-time monitoring system at netdata.pivoine.art:
- Real-time infrastructure and container monitoring
- Auto-discovers all Docker containers
- Tracks CPU, memory, disk, network usage per service
- Low overhead monitoring (~1-3% CPU)
- Self-hosted with web dashboard on port 19999

Configuration:
- Created netdata/compose.yaml with full Traefik integration
- Added to main compose.yaml include list
- Added environment variables to arty.yml
- Mounted Docker socket for container metrics
- Mounted system directories for host metrics (/proc, /sys, /var/log)
- Three persistent volumes: config, lib, cache
- Required capabilities: SYS_PTRACE, SYS_ADMIN for system monitoring
- Watchtower enabled for automatic updates

Benefits for infrastructure:
- Monitor 20+ running services in real-time
- Track PostgreSQL, Redis, Traefik performance
- Watch backup processes (Backrest/Restic)
- Monitor Jellyfin transcoding load
- Alert on resource issues before they become critical
- Historical data for capacity planning

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <noreply@anthropic.com>

compose.yaml

@@ -14,6 +14,7 @@ include:
   - jelly/compose.yaml
   - drop/compose.yaml
   - restic/compose.yaml
+  - netdata/compose.yaml
   - umami/compose.yaml
   - sablier/compose.yaml
   - proxy/compose.yaml