TL;DR
This is zolty.systems – a technical blog about building and running a production-grade Kubernetes homelab. Expect deep dives into k3s, Proxmox, networking, monitoring, CI/CD, and all the things that break along the way.
Why Another Homelab Blog?
Because every homelab is different, and every failure teaches something new.
I’ve been running a multi-node k3s cluster on repurposed mini PCs for a while now. What started as a “let me just run a few containers” experiment has evolved into a full production-grade platform with:
- High-availability Kubernetes control plane
- Distributed block storage
- Automated TLS certificate management
- Self-hosted CI/CD runners
- Full observability stack (metrics, logs, alerts)
- Automated remediation pipelines
Along the way, I’ve hit every possible failure mode: networking issues that took days to debug, storage deadlocks that required rethinking deployment strategies, DNS configurations that silently broke half the cluster, and LACP bonds that bricked a NAS.
This blog exists to document those experiences – not just the polished end result, but the messy debugging sessions and the “oh, that’s why” moments.
What to Expect
Infrastructure Deep Dives
Step-by-step guides for building and configuring homelab infrastructure. Not the “just run this script” kind – the “here’s why each setting matters” kind.
Kubernetes Patterns and Anti-Patterns
Real-world lessons from running k3s in production. What works, what doesn’t, and what will silently break at 3 AM.
Monitoring and Observability
How to know when things break before your users do (or in my case, before I notice the dashboard is down).
Post-Mortems
Honest write-ups of what went wrong and how it got fixed. The best learning comes from failure.
The Tech Stack
This blog itself is a product of the homelab:
- Hugo for static site generation
- AWS CloudFront + S3 for global CDN delivery
- GitHub Actions running on self-hosted k3s runners for CI/CD
- AWS Bedrock for AI-assisted content drafting
- Terraform for infrastructure management
Every piece of infrastructure is codified, automated, and version-controlled. No ClickOps allowed.
Stay Tuned
Subscribe via RSS or check back regularly. There’s a lot to cover.