Powering a GPU Node On-Demand with iDRAC and Redfish
A rack server with a GPU is great for self-hosted LLMs and terrible for a power bill. Here's the tiny Redfish-based tooling I use to boot it on demand and shut it down when I'm …
•
1 min read
Zero Open Ports: Publishing Homelab Services with Cloudflare Tunnel
Every public service in my homelab — Grafana, static sites, LLM inference — is reachable over HTTPS without a single forwarded port. Here's the ingress architecture and what I'd do …
•
2 min read
Running AWS ECS Anywhere on a Proxmox Homelab
How I run production SaaS workloads on ECS Anywhere inside a Proxmox VM — getting AWS's control plane, ECR, and deployment tooling while the compute lives in my house.
•
1 min read
Hybrid Cloud Homelab
Self-hosted infrastructure on Proxmox VE running AWS ECS Anywhere clusters on local hardware, blending cloud-native orchestration with on-prem servers. Hosts a full observability …
•
1 min read