These are the datasets for our ESEC/FSE'22 paper "Actionable and Interpretable Fault Localization for Recurring Failures in Online Service Systems." In each dataset, graph.yml or graphs/*.yml are FDGs, metrics.csv is metrics, and faults.csv is failures (including ground truths).FDG.pkl is a pickle of the FDG object, which contains all the above data. Note that the pickle files are not compatible in different Python and Pandas versions. So if you cannot load the pickles, just ignore and delete them. They are only used to speed up data load. See more at https://github.com/NetManAIOps/DejaVu