The environment is provided in the conda format (environment.yaml
).
Please run conda env create -f environment.yaml
to create the required environment.
The root cause analysis algorithm is in ours_full_pipeline.py
.
We provide a test case data in sample_data/anomaly_host_metrics.csv
.
Run python ours_full_pipeline.py
to get the test output on this case.
These visualization and analyisis code is in visualization-analysis.ipynb
.
Due to commercial secrecy requirement, the host anomaly data is not provided. We only provide the sample data of 1 anomaly case.
The cross-evaluation experiment code and the baselines are concluded in exp-analysis.ipynb
.
For simplicity, we do not provide the intermediate results and datasets. This code is only for the demonstration purpose.