K3s Homelab
ActiveK3s cluster on physical hardware. Self-hosted applications, GitOps with Flux, actual networking to deal with. Read the docs, read KubeCraft cover to cover, broke things along the way. Still running.
GitHub →Becoming an architect
Real builds, documented while they happen. Homelab, K3s, GitOps, AI infrastructure. The broken stuff stays in. Not just the wins.
Learn it. Build it. Break it. Document it. Repeat.
Current Build
My goal is to integrate GitOps properly and get this more and more production-similar over time. Physical hardware, real networking, the debugging that only happens when something breaks at 11pm.
Recently started
I recently started my CKAD study, alongside the homelab work and the day-to-day infrastructure learning.
All Builds
K3s cluster on physical hardware. Self-hosted applications, GitOps with Flux, actual networking to deal with. Read the docs, read KubeCraft cover to cover, broke things along the way. Still running.
GitHub →Next up after the base cluster: Prometheus, Grafana, and alerting that survives the usual home-lab chaos. This will only move into active once the current K3s setup is stable enough to deserve it.
Getting Flux properly integrated and the cluster closer to production patterns. Currently studying for the CKAD, which means writing as many K8s manifests by hand as possible instead of relying on generators. That practice feeds directly into this.
How I build
Read the theory first. Actual docs, academic papers, whatever exists on the topic. Then build it. Every time, in that order.
The process gets documented while it's happening, not after it's been cleaned up. If something broke and I had to fix it at midnight, that's in there.
To be honest, the hardest part isn't the technical stuff. Keeping going when things get boring and nothing is working yet. That's the actual job.
Find me
CS student and not sure where any of this fits in your career yet? Start here.