Today, scalable servers are key elements of most information systems. Their operational requirements, like good response time, low power consumption or high availability, are more and more strict. To fulfill these requirements continuously a proper management policy must be used. But the complexity of these servers prevents the application of manual policies and the own servers must manage themselves as much as possible. Therefore, there is a growing need of self-managing techniques and systems for this type of complex servers. This paper presents a self-managing technique for balancing response time and power consumption that can be integrated and configured in any server cluster very easily.