For monitoring, controlling, and improving software development projects, project and quality managers need tool support to analyze engineering processes within development environments. Unfortunately, technical and semantic gaps between the engineering tools and related data models make it hard to observe and analyze the implemented tool-based engineering processes. In this paper we build on a service-oriented platform for technically and semantically integrating heterogeneous engineering tools and propose an approach to monitor, analyze, and improve tool-based engineering processes. We empirically evaluate the approach using the"continuous integration and test"process and discuss strengths and limitations. Major result was that the approach enabled comparing expected and real-life engineering processes with respect to process structure, performance of individual process steps, and risk of bottlenecks.