Scheduling taking paging, swapping and high CPU IO wait into account

VMs reliability and performance are directly associated with reliability and performance of underlying hardware infrastructure. Hardware health or performance issues such as paging/swapping, high CPU IO or degraded IOPs problems can be collected on Compute level and then exposed in upper layers. With adjusted scheduling logic VMs relying heavily on RAM or permanent storage reliability may benefit from not being instantiated on Compute Nodes which exhibit paging/swapping, high CPI IO wait or degraded IOPs.
Child blueprint for https://blueprints.launchpad.net/nova/+spec/platform-health-aware-scheduling