I wisely(?) decieded that something I wanted to do was have Kubernetes deployed in the cloud, so I picked an archived project using a tool I hadn't used much (Terraform) that deployed already obsolete versions of everything. All in all, a strong start. Actually, its not all as bad as it sounds. My main motivation for using this setup is that it's inexpensive and resilient. Most k8s clusters aim to be $1 a day, while this paticular bit tries to be $3-$6 a month, which sounds better for experimenting. So I download and try to run stuff, and it has a bunch of dependencies, so I get those all sorted out. I create a new IAM role for this, and then go through a bunch of pain to deploy the cluster all the way one time. Well, kind of at least, there is still an error but I can run "kubectl get no" and it brings back some k8s output. I call this a success and run a destroy, which brings all the resources down. Having done it once, I get a bit cocky and run the scripts again, but this time I want to run something on the cluster. I deploy, and then struggle a bit to get things working. While I've done some Kubernetes work, I'm still not a kubectl master so I need to figure out some of the commands. I eventually get a container to run, but realise I'm not sure how to map yet from public IPs onto the cluster's internal private IPs -- this will be nodePort or Ingress, but that will come. I call this another good session, and spin stuff down again. I then dig into the scripts and Terraform configs a bit. I have to understand this a bit, to modernize the infrastructure. The TF files seem straightforward. The scripts feel simple, but there will be a bit of learning to do in modifying them. My goals are modest -- move to a recent or supported version of k8s, change the volume type to gp3 or standard from gp2, and run the infra for a couple days to get an idea on costs, with some containers I'd like to run persistently. As stretch goals, I'd like to try and bring the TF to a recent version and maybe even deploy my containers automatically. Let's see what can be done.