Analyzing Distributed Systems using Tracing and Process Mining
Modern information systems produce a tremendous amount of event data. The area of process mining deals with extracting knowledge from this data. Real-life processes can be effectively discovered, analyzed, and optimized with the help of mature process mining techniques. There are various process mining case studies and experience reports from such business areas as healthcare, public, transportation, and education. However, extracting relevant data from modern systems is a complex and cumbersome task that requires in-depth technical and process knowledge. Therefore, this thesis introduces the use of tracing tools to retrieve the digital footprint of the distributed system. On the other hand, process mining techniques have recently received significant attention in the literature to assist in designing complex processes by automatically discovering models that explain the events registered in some log traces provided as input. Following this line of research, the system is utilized, and user interaction is recorded in logs. Each log has a timestamp that corresponds to the occurrence of the operation. These logs are clustered that represent a business transaction. This means a single trace over the model. All the business-related functions of the system are transformed into a log known as an event log.Thus with the help of our examples, we demonstrate that process mining facilitates new forms of software analysis. The user interaction of almost every software system can be mined to improve the software and monitor and measure its real usage.