Description of problem:

Cluster with scheduleable masters may generate legitimate but irrelevant HighOverallControlPlaneCPU and/or ExtremelyHighIndividualControlPlaneCPU alerts when user workload consumes a lot of CPU on control plane nodes.
Running scheduleable masters is quite a common design pattern for OpenShift on BareMetal, where node size is too big to waste for control plane only, especially in OpenShift Virtualization use-cases.
Logic behind those alert should consider scheduleable masters use cases. Ideally it should consider only OpenShift platform components and ignore possible user workload running on control plane nodes.