Attempting to infer the performance of a database server by looking at resource utilization metrics will fail technically and organizationally. In a shared and dynamic environment, the only way to truly assess the performance of an element of the infrastructure or an application is to measure how long it is actually taking to do its job. This makes response time and its companion metric latency into the two most important metrics for virtualized and cloud based systems.
Transaction rate, an average response time goal, and an allowable upper bound, combined with transaction rate will come to define performance and capacity. These new metrics will get monetized allowing applications owners to shop for price/performance curves across internal and external service providers on a rational and economic basis.