>The high cost arises from needing to provision a bare-bones cluster with redundant management nodes.

?!

How else are you managing your infrastructure besides having redundancy in the control plane? Even if it’s a chef or puppet server, or even sysadmin Dave running scripts on his laptop, you should still have redundancy.

> Moreover, Kubernetes’s slow autoscaling meant I had to over-provision services to ensure availability, paying for unused resources rather than scaling based on demand.

??????

Slow compared to what exactly? Anecdotally, k8s with karpenter is significantly faster to scale than auto scaling groups.