Observability Pipelines at Kubecon

The complexity meme doesn't really make much sense.

k8s is a the centre of a lot of transformational stories, where companies that couldn't do things like CI/CD have been helped making the leap, because it pretty much forces you to do IaC and makes it hard to screw up.

It's harder than clicking on the GUI, starting a VM, ssh'ing to a box and running a process, but is it harder than doing the same with terraform on AWS?

The problem is more that it is actually simple. And that simplicity allowed the creation of more and more things, all building on the simple frameworks, that work somewhat the same way. In a world where rarely do we enjoy fitting in one or two machines, and where there's a lot of moving bits to handle, that commonality, that Operating System for Data Center, is a godsend.

And "simple" is not the same as "easy". K8s is a classic example of complexity arising from simple system, as there are pretty much two major algorithm ideas in the whole thing (blackboard multi agent system and OODA loop controllers).