Easy: the perception of infinite overhead. The cloud itself (e.g. EC2) gives you capacity-on-demand, but the glue to make all the nodes work together is missing

Easy: the perception of infinite overhead. The cloud itself (e.g. EC2) gives you capacity-on-demand, but the glue to make all the nodes work together is missing. And its a really hard problem, in general, because its distributed systems. K8s fills that demand, or seeks to. The alternative is to roll-your-own, which is possible but expensive, error-prone, and difficult to hire for. (My last company discovered this to their detriment after sinking a LOT of time into building out a really complex Salt+Vagrant+AWS solution, and then decided to migrate to k8s).