Just a short post to provide some context around some upcoming posts.

Around a year ago, I set up a bare-metal(ish) kubernetes cluster pretty much from scratch using the kubeadm method. Installing and managing services on it has gone through a fairly predictable pattern for an educational cluster:

  1. Started out with applying bare manifests
  2. Quickly moved on to installing services with helm charts
  3. Let it all sit happily for a while until charts get out-dated, and realise that remembering what charts are sitting around where is a PITA

And so I’m currently in the process of migrating all the helm charts (and those few initial bare manifests) to be managed with Argo CD

While I’m at it, I’m also taking the opportunity to fix some mistakes learnings made along the way (such as cleaning up everything into their appropriate namespaces). I’ll also be moving to a better way of handling storage across the cluster from the current unmentionable jank.