I'd say 3 or more distinct technology stacks. You can run 200-machine clusters comfortably with lighter tooling if they're all Ruby or all JVM or what have you (and I have), but once your deployment process requires remembering the foibles of 4 different Python build tools it becomes easier to shove it all in Kubernetes.

Kubernetes starts to make sense when you don't know the number of machines you need (and therefore desire auto-scaling based on system load).