- Flask app with /usage/<pv_name> and /metrics endpoints
- 300s cache TTL with threading lock for du command
- ionice/nice for low I/O priority
- Dockerfile with gunicorn
- Helm chart with deployment, service, networkpolicy
- NetworkPolicy allows only gpulive and monitoring namespaces
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Cloud User