Complex Event Processing makes it possible to aggregate simple events into complex events with a higher semantic. Complex correlation can be recognized as well as the number of events can be reduced. One problem in CEP is that the workload is not constant, because there are times when more events have to process and times when fewer events have to process. In this work an approach is introduced to handle peaks in the workload by using a Cloud environment. The Cloud makes it possible to assign and reassign resources quickly. A system model is presented that makes it possible to deploy the CEP network in the Cloud environment. Furthermore, the system will be able to reconfigure itself in order to add new resources as well as to delete existing resources and to balance the workload on all existing resources. In addition, a decision criterion is introduced which determine the time, when the allocation and deallocation request has to be sent.