Latest LLM Summary

## Daily summary
• **Proxmox Memory**: High usage on both nodes. `proxmox` at 80.6% and `proxnode` at 80.3% (both >80% threshold). 
• **Docker Health**: 50/50 containers running. 0 unhealthy.
• **Docker Logs**: Anomaly detected in error log volume for `phase-backend` (118 lines), `watchtower` (61 lines), and `personal-startpage` (57 lines). Total error lines (304) are slightly below 7-day avg (331.57).
• **Proxmox VM Status**: 6 VMs running, 6 stopped.
• **Storage**: All volumes safe (<80%). Highest utilization is `nvme1tb` at 73.1% (flat vs baseline).
• **RouterOS**: System stable. CPU 0.0%, Memory 15.5% (0.0% delta vs yesterday). 
• **Network**: Total interface errors (102,931) and link downs (12) remain unchanged from yesterday.

## Actions
- proxmox: investigate memory pressure on nodes: `ssh proxmox 'free -m'`
- docker: inspect `phase-backend` error logs: `docker logs --tail 100 phase-backend`
- docker: inspect `watchtower` logs: `docker logs --tail 100 watchtower`

## FYI (normal but noteworthy)
• RouterOS firewall logs show persistent `dstnat` connection attempts (SYN) on ports 80 and 443 from various external IPs.
• Wireguard: 3/3 peers online; peer4 has most recent handshake (1m 47s).