What metrics and SLAs would you define to measure control plane performance, scheduler throughput, and pod launch latency at scale?
