Experiment setup with a single service jpetstore to evaluate user behavior.
compare-behavior.shcompare and compile results of different behaviors with te evaluate behavior toolconfig.templatetemplate of the experiment configurationexecute-all-fuzzy.shrun all observation tasks with randomized behaviors (executes execute-observation.sh)execute-analysis-ssp.sh(outdated) analysis for the SSP paper 2017execute-analysis.shrun application and analysis, uses docker-composeexecute-kieker-analysis.shuse observed data and analyze it with Kiekerexecute-observation.shup to date JPetStore experiment execution, generates user dataexecute-session-reconstruction.shtool to fix collected data and add missing session datareconstructor.configexample reconstruction configuration
All scripts share a common configuration file config.
DATA_DIRdirectory for the monitoring data. A running collector will store all events in the specified directory or a sub-directory thereof.COLLECTORrefers to the executable script of an collector service of the https://github.com/research-iobserve/iobserve-analysis projectWORKLOAD_RUNNERrefers to our selenium based workload driver for the JPetStore which can be found in https://github.com/research-iobserve/selenium-workloadsPHANTOM_JSphantomJs driver for selenium
execute-observation.sh takes two parameters. The first is a workload specification file and the second an experiment id. If no workload is specified, the script runs without workload driver and the JPetStore can be run interactively. In case a workload is specified, it is also necessary to specify the experiment id. The script assumes the use of phantomJs. However, the workload runner works better with the Chrome driver. If you want to switch, please adjust the script.