Testbed stats

[mniedermaier]

Summary

• Integrated host network statistics directly into landing container
• Added comprehensive system monitoring page with real-time stats and historical data
• Landing container now runs with host network mode on port 80

Changes

Container Architecture:

• Removed host-stats container from both production and dev compose files
• Updated landing container to use network_mode: host instead of bridge networking
• Landing now accessible directly on port 80 (no port mapping needed)
• Removed build step for host-stats from build.sh

Landing Application (software/landing/):

• Integrated psutil to collect host network stats directly (filtering Docker interfaces)
• Added background threads for collecting system and Docker container statistics
• Changed Flask port from 8081 to 80
• Added new API endpoints:
  • /api/stats - Current system statistics (CPU, memory, swap, disk, network, Docker containers)
  • /api/stats/history - Historical data for the past hour
• Added new /stats route with comprehensive monitoring dashboard

System Stats Page (software/landing/templates/stats.html):

• Real-time system metrics: CPU, Memory, Swap, Disk, Network I/O
• System uptime display
• Historical charts (1 hour) for CPU, Memory/Swap, Disk, and Network traffic
• Service monitoring for OpenPLC, FUXA, and Virtual Hardware IO
• Docker container statistics with CPU, memory, and network usage per container
• Auto-refresh every 1 second for current stats, charts update every 10 seconds

Bug Fixes:

• Fixed network traffic historical data spike on startup by initializing counters with current values
• Memory chart now includes swap usage as a second line

Other Changes:

• Fixed hardware abstraction in hwio-virtual to use NiceGUI properly with async updates
• Updated landing page navigation to include link to stats page
• Moved CTF reset button position for better UX

Test Plan

• Verify landing page accessible on http://localhost after compose up
• Check stats page displays real-time system metrics
• Confirm historical charts populate correctly without initial spike
• Validate Docker container stats show CybICS containers only
• Test service monitoring for OpenPLC, FUXA, and VHIO

[mniedermaier]

Looks good