While clouds conceptually facilitate very fine-grained resource provisioning, information systems that are able to fully leverage this potential remain an open research problem. This is due to factors such as significant reconfiguration lead-times and non-trivial dependencies between software and hardware resources. In this work we address these factors explicitly and introduce an accurate workload forecasting model, based on Fourier Transformation and stochastic processes, paired with an adaptive provisioning framework. By automatically identifying the key characteristics in the workload process and estimating the residual variation, our model forecasts the workload process in the near future with very high accuracy. Our preliminary experimental evaluation results show great promise. When evaluated empirically on a real Wikipedia trace our resource provisioning framework successfully utilizes the workload forecast module to achieve superior resource utilization efficiency under constant service level objective satisfaction. More generally, this work corroborates the potential of holistic cloud management approaches that fuse domain specific solutions from areas such as workload prediction, autonomic system management, and empirical analysis.