Latest LLM Summary

## Daily summary
• Proxmox Memory: Both nodes at critical threshold (proxmox: 80.5%, proxnode: 80.0%), stable vs 7-day avg (80.26%).
• Proxmox Compute: 6 VMs and 2 containers running; 6 VMs currently stopped.
• Proxmox Storage: Max utilization 73.1% (nvme1tb), well below 80% threshold.
• RouterOS Health: CPU (0.0%) and Memory (15.43%) are stable and within 7-day baselines.
• RouterOS Network: `wireguard1` reporting significant TX errors (90,764) and RX errors (647).
• RouterOS Connectivity: WireGuard `peer5` failed handshake after 20 attempts at 06:58.
• Docker Status: 48/48 containers running (100% availability).
• Docker Logs: Total error lines (336) are below the 7-day average (363.86).
• Docker Noise: `phase-backend` is the primary log contributor (119 error lines).

## Actions
- routeros: Investigate `wireguard1` interface errors and peer5 connectivity: `/log print where topics~wireguard`
- docker: Review `phase-backend` error logs: `docker logs phase-backend | tail -n 100`
- proxmox: Evaluate memory overhead on `proxnode` and `proxmox` to reduce usage below 80%: `qm list` / `pct list`

## FYI (normal but noteworthy)
• RouterOS logs show consistent TCP SYN activity on ether1 (ports 80/443) from various external IPs.
• `AgentApiRead` is performing regular REST API polling every ~10 minutes.