inter-connected, spanning over a range of computing devices. As software systems are being split into modules and services, coupled with an increasing parallelization, detecting and managing anomalies in such environments is hard. In practice, certain localized areas and subsystems provide strong monitoring support, but cross-system error-correlation, root-cause analysis and prediction are an elusive target. We propose a general approach to what we call Full-spectrum anomaly detection - an architecture that is able to detect local anomalies on data from various sources as well as creating high-level alerts utilizing background knowledge, historical data and forecast models. The methodology can be implemented either completely or partially.