I’ve been conflicted about how much hardware I need and how to deploy it in a way that will be the most flexible when testing radically different simulations of full IT deployments. For example, Icinga doesn’t run well in Kubernetes but I need to test talking to it from within a Kubernetes application that keeps eyes on PVC and a StorageClass. That means at number of VMs just for the simulated Kubernetes cluster and a few for the things needed in the environment, “external DNS” (for load balancing), and the host systems for Icinga.
I thought I could do this with just all the old hardware I have laying around, but the realization the “over-provisioning” allows virtual machines to share a single CPU has opened my eyes to the possibility of a single, multi-core, lots-of-RAM (which can’t be over-provisioned) machine. Obviously, System76 machines came to mind and are a great example of what I need eventually, but the cost is prohibitive and possibly something that I could get them to sponsor.
In the meantime, I’ll see how far I can push this 8-core MSI server (gaming) tower. Perhaps it will be enough (at least until I can convince System76 to sponsor me and at least cut the price of one of their rigs, give me one for free, or even pay me and give me one, I can only dream).