Erich Blume 251873906a Disable SPA mode, remove index files, relax wiki-link constraints

Fix the Facebook crawler spider trap by disabling Quartz SPA mode and
removing the nginx fallback to index.html. Non-existent URLs now return
404.html instead of the root SPA shell, preventing infinite recursive
crawling.

Remove hand-curated category index files (tutorials.md, reference.md,
how-to.md, explanation.md) — Quartz auto-generates folder pages. Drop
docs-check-index and docs-check-filenames hooks. Update docs-check-links
to allow path-based wiki-links and only error on true ambiguity. Remove
robots.txt exclusions since they're no longer needed.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

2026-03-09 11:57:35 -07:00

4 KiB

Raw Blame History

title

modified

Replicating BlumeOps

Audiences: Replicator

This tutorial provides a roadmap for building your own homelab GitOps environment inspired by BluemeOps. It links to detailed component tutorials for each major piece.

What You'll Build

By following this guide, you'll have:

A secure mesh network connecting your devices
A Kubernetes cluster for running containerized services
GitOps-driven deployments via ArgoCD
Observability with metrics, logs, and dashboards
Backup and disaster recovery capabilities

Hardware Requirements

BluemeOps runs on modest hardware. At minimum:

Component	BlumeOps Uses	Minimum Alternative
Server	Mac Mini M1	Any machine with sufficient RAM (16GB recommended)
NAS	Synology DS920+	USB drive or second machine
Workstation	MacBook Air M4	Whatever you use daily

You can start with a single machine and add storage later.

The Journey

Phase 1: Networking Foundation

Before deploying services, establish secure connectivity.

tailscale-setup

Create a tailnet and connect your devices
Configure ACLs for service access
Set up MagicDNS for convenient naming

This replaces: traditional VPNs, port forwarding, dynamic DNS

Phase 2: Core Services

Bootstrap the essential services that everything else depends on.

core-services

Set up forgejo for git hosting and CI/CD
Optionally set up zot container registry
Configure SSH access and deploy keys

Forgejo is central to GitOps - it's where your infrastructure definitions live and where CI/CD workflows run.

Phase 3: Kubernetes Cluster

A cluster for running containerized workloads.

kubernetes-bootstrap

Install minikube (or k3s, kind, etc.)
Configure persistent storage
Expose the API securely via Tailscale

BlumeOps uses minikube for simplicity, but the patterns apply to any distribution.

Phase 4: GitOps with ArgoCD

Declarative, git-driven deployments.

argocd-config

Install ArgoCD in your cluster
Connect to your git repository
Deploy your first application
Set up the app-of-apps pattern

This is the heart of GitOps - changes in git automatically sync to your cluster.

Phase 5: Observability Stack

Know what's happening in your infrastructure.

observability-stack

Deploy Prometheus for metrics
Deploy Loki for logs
Deploy Grafana for dashboards
Configure Alloy for collection

Without observability, you're flying blind.

Phase 6: Your First Services

With the foundation in place, deploy actual workloads. BluemeOps runs:

miniflux - RSS reader
jellyfin - Media server
immich - Photo management
navidrome - Music streaming
docs - Documentation site (Quartz)

Pick what matters to you. Each service follows similar patterns:

Create Kubernetes manifests
Create ArgoCD Application
Configure ingress routing
Sync and verify

Phase 7: Backups and Resilience

Protect your data.

Set up borgmatic for backup automation
Configure NAS as backup target
Test restore procedures
Document disaster recovery

Alternative Approaches

BluemeOps makes specific choices that may not suit everyone:

BlumeOps Choice	Alternative
macOS server	Linux server (more common)
Minikube	k3s, kind, or managed K8s
Tailscale	WireGuard, Nebula
ArgoCD	Flux, manual kubectl
Ansible	NixOS, Docker Compose

The principles (GitOps, IaC, observability) matter more than specific tools.

Getting Started

Begin with tailscale-setup - networking is the foundation everything else builds on.

Reference - See BlumeOps' specific configurations
contributing - Help improve BlumeOps instead

4 KiB Raw Blame History