Skip to main content
Provision once with OpenTofu, configure with Ansible, run forever. The infrastructure layer is OpenTofu-managed; the terraform-* repo names are legacy, and the engine inside is tofu. Every module is opinionated about deployment shape. It uses LXC for production homelab workloads, Docker on a dedicated VM only when vendor-locked, and AWS for disaster recovery and managed services.

How the pieces fit together

Every guest follows the same three-stage pipeline. OpenTofu provisions it, Ansible configures it, and then the app stack deploys on top. CI/CD gates every change before it reaches provisioning. The AWS module map exists mainly as a failover target and for managed services the homelab doesn’t run itself.

The Proxmox stack

OpenTofu builds VMs and LXCs (coral). Ansible takes the inventory and configures hosts (green), then deploys the app stack on top.

AWS module map

Repos in this section

tofu-proxmox

VMs and LXC containers on the Proxmox cluster.

tofu-unifi

UniFi networks, VLANs, port profiles, firewall rules — network-as-code.

ansible-proxmox

Host config — ZFS, networking, users, hardening.

ansible-proxmox-apps

App deploy — HAProxy, Cribl Edge, Cribl Stream.

tofu-aws (private)

AWS DR footprint for Splunk failover. Cold infra, ready to go warm. Private infrastructure, no public source.

CI/CD

GitHub Actions policy and remote plan/apply pattern.

orbstack-kubernetes

Local Kubernetes cluster on macOS — monitoring stack + AI gateway.

tofu-splunk-aws

Cost-optimized Splunk deployment on AWS.

Cross-cutting topics

Kubernetes overview

OrbStack as the local control plane; what runs on K8s vs LXC vs Docker.

LXC vs Docker

The four-question decision tree for every new workload.

Self-hosted Netflix

Media library on a dedicated VLAN.

What runs where

Deployment philosophy

LXC is the default for production homelab services. Native packages where possible. Docker runs only when a vendor ships Docker-only images and there’s no native path. It also runs only on a dedicated docker-host VM, so high-volume network traffic never crosses Docker’s virtualized network stack.
For configuration of provisioned hosts, see Configuration.