A main goal of Enterprise Architecture Management (EAM) is to align business needs with IT capabilities in order to understand and visualise interdependencies between the layers of an Enterprise Architecture (EA) and therefore gain a holistic view.
This thesis aims to provide the required visibility between the business and application layer through a monitoring and analysis approach. The approach enhances traditional process mining with performance indicators obtained from the application layer. With the developed prototype, correlations between user behaviour from the business layer and system performance occurring in the application layer can be detected.
This is achieved by applying process mining discovery techniques on distributed tracing data originating from an instrumented microservice architecture. Through an activity log generation component, low level events of the distributed tracing data are transformed into a high level activity log comprising of two hierarchies: user and system activities.
It is shown, that the developed prototype can be implemented with little effort and provides a cost-efficient bottom-up approach to discover business processes in mi- croservice architectures in near real-time. The approach comes with little instrumenta- tion effort and is agnostic to programming frameworks or architectural styles.
Limitations especially emerge through possible performance losses caused by high sampling rates in the instrumented software component.