Software runtime execution trace dataset for runtime anomaly detection

Published: 19 December 2022| Version 1 | DOI: 10.17632/h2zf8gbv42.1
Shiyi Kong


Here is the datasets for software runtime anomaly detection methods. The data in the trainingDatasets folder contains times series collected during software execution. Due to the storage limits, we only upload partiall time series here, other data are provided on request. (email at the author) As for the evaluationDatasets, here are two folder, the norm folder contains the samples with fixed length 1000 collected from normal runs, the anomalies folder contains the samples collected from abnormal runs.


Steps to reproduce

the data is collected by Btrace tools, the software monitored is Cassandra, and the workload is applied using the ycsb tools. all software can be found on GitHub. We use a computer with an Intel i5 processor and 8gb memory.


Beihang University