Online Performance Problem Detection, Diagnosis, and Visualization with Kieker.

Düllmann, Thomas, Eberlein, Andreas, Endres, Christian, Fetzer, Matthias, Fischer, Markus, Gregorian, Christopher, Képes, Kálmán, Noller, Yannic, Olp, Dominik, Rudolph, Tobias, Scherer, Anton and Scholz, Martin (2014) Online Performance Problem Detection, Diagnosis, and Visualization with Kieker. (Student research project), University of Stuttgart, Institute of Software Technology, Stuttgart, Germany, 149 pp.

[thumbnail of DuellmannEtAl2014KiekerikiProject.pdf]
DuellmannEtAl2014KiekerikiProject.pdf - Published Version

Download (7MB) | Preview


With increasingly large systems Online Performance Monitoring becomes more and more a necessity to find, predict, and recover from failures. The Kieker monitoring tool enables the monitoring and analysis of applications. It allows to gather live data about the systems utilization like RAM-load, Swap-load, CPU-load as well as the latency of executed operations and their qualified name. OPADx provides means to detect anomalous behaviour and RanCorr allows the correlation of anomalies to identifiy the root cause of an anomaly. This project implements the RanCorr approach and extends the OPAD implementation with new forecast algorithms. Also the KiekerWebGUI is extended to visualize the architecture, discovered by RanCorr, and other metrics by using dynamic diagrams. Additionally, an automated test framework is introduced that enables data generation and evaluation of the implemented forecasting and anomaly detection approach.

Document Type: Thesis (Student research project)
Keywords: Kieker, application performance management, application performance monitoring, dynamic analysis, software visualization, performance problem detection, performance problem diagnosis
Research affiliation: Kiel University > Software Engineering
Open Access Journal?: Yes
Projects: Kieker
Date Deposited: 28 Apr 2014 10:51
Last Modified: 28 Apr 2014 10:53

Actions (login required)

View Item View Item