All work

Case 06● Live2025 — now

Homelab

A personal cloud on Proxmox: 10 LXC + 2 VM, ~99% uptime

Role
Full cycle: from hardware to alerts
Period
2025 — now
Stack
Proxmox VE · LXC · NPMplus · PostgreSQL · Gitea · k3s · Prometheus · Grafana · Technitium DNS
Grafana: host temperature, load, CPU and memory

Architecture

Cloudflare DNS

Router · port 443 forward

NPMplus reverse proxy · TLS

Proxmox VE 9 · mini-PC · 10 LXC + 2 VM

Apps

  • Gitea
  • n8n
  • Homepage
  • TeamSpeak
  • Frappe LMS · VM
  • k3s · VM

Data

  • PostgreSQL · shared

Observability

  • Prometheus
  • pve-exporter
  • Grafana

Network

  • Technitium DNS
ZFS snapshots & automated backupsTelegram alerts

Outside: DNS on Cloudflare and port 443 forwarded to NPMplus, which terminates TLS and routes traffic to services. Apps live in separate LXC containers, Frappe LMS and a k3s sandbox in QEMU VMs, and a shared PostgreSQL serves four-plus services.

Observability

  • Per-container CPU, memory, disk and network metrics, scraped every 15 seconds
  • Grafana host dashboards: temperatures, load, CPU modes, memory
  • Telegram alerts on anomalies and failures

Screens

3