202510041749 Simplifying The Homelab
I think the homelab is too complicated. There is an unnecessary way of managing various kinds of manifests and I think I am just making my life difficult using Tanka.
One of the big issues is that I’m finding it difficult to actually understand the abstractions I wrote a couple of months back. Perhaps it is memory loss or perhaps it was an unnecessary abstraction.
I think for this new revision of the homelab cluster, I call it V2.
Let’s approach it from this way
- What is the plan to release applications? Do apps go first or infra related software (ie Secrets Operator?)
- I should blog and document more often. I felt that was the reason I made this monorepo and that was one of the goals for this year. Let’s write out a plan and incrementally add the applications and plumbing in.
- Focus on the core cluster essentials. What are the essentials and which is the most important?
- Flux
- Longhorn
- TailScale
- Secrets Operator
- Remove unnecessary abstraction
- Tanka
- Helm
Managing a k8s homelab shouldn’t be made more complicated than it already is