My entire Kubernetes cluster runs as VMs on a single beefy Proxmox machine. Yeah, not the most HA setup, but it gets the job done! The cluster consists of three control plane nodes (Odin, Thor, and Frigg) that handle both control plane and worker duties.
Currently running everything on NFS until I can afford to buy some proper NUCs (like the cool kids do!). Once I get those NUCs, the plan is to implement rook-ceph for distributed storage.
For now, I maintain a dedicated 24 TB ZFS server that handles NFS/SMB file sharing, large-scale media storage, and backup operations.
This is a Talos Linux-powered Kubernetes cluster managed with FluxCD for GitOps. Everything you see here is automatically deployed and kept in sync from this Git repository.
Networking: Cilium provides eBPF-based networking and security, with ingress handled by NGINX. External DNS automatically manages Cloudflare records, and Cloudflare Tunnel enables secure external access.
Security: Certificates are automatically provisioned via cert-manager and Let's Encrypt. Secrets are managed through 1Password Connect (via external-secrets) and SOPS for Git-stored secrets.
Storage: OpenEBS provides local persistent volumes. Volsync is installed for backups (configuration in progress).
Observability: Complete monitoring stack with Prometheus, Grafana, and Loki. Kromgo powers the cluster metrics badges at the top of this README.
Media Automation: Complete media management setup with Plex as the streaming platform, automated downloads via the *arr stack (Sonarr, Radarr, Prowlarr, Bazarr), and both torrent and usenet support.
Home Automation: Home Assistant handles all smart home devices and automations, with Zigbee devices connected via Zigbee2MQTT and EMQX as the MQTT broker.
Productivity: Nextcloud for file sync and collaboration, Actual Budget for personal finance tracking, and n8n for workflow automation.
Infrastructure: PostgreSQL clusters managed by CloudNative-PG for application databases.
- Storage: Migrate from NFS to rook-ceph once I get dedicated NUC hardware
- Backups: Properly configure Volsync for automated persistent volume backups
- High Availability: Move to proper bare-metal nodes instead of running everything as VMs on one Proxmox host
- Monitoring: Expand Grafana dashboards and alerting rules
- More Services: Always looking for interesting self-hosted applications to add!
Credits are where credits are due. When I started implementing Talos on my own, a lot of studying went in. Bumping into the cluster-template made life so much easier on many fronts. If you're just like me and like to "FAFO" your way forward, this is a perfect place to start the Talos/Flux journey. You can check out the example setup at onedr0p/cluster-template.
Also make sure to hop in at the Home Operations Discord server for an amazing community of homelabbers!
This repository is licensed under the MIT License - see the LICENSE file for details.
![@renovate[bot]](https://avatars.githubusercontent.com/in/2740?s=64&v=4){kind=small}