Official code of the paper Root Cause Analysis In Microservice Using Neural Granger Causal Discovery .
We propose a novel approach for root cause analysis using neural Granger causal discovery with contrastive learning.
- Enhance the backbone encoder by integrating contextual information from time series
- Leverage a time series forecasting model to conduct neural Granger causal discovery
- In addition, we incorporate Pagerank with a personalization vector to efficiently recommend the top-k root causes
python main.py --root_path ./ --data_path data --root_cause root_cause --trigger_point trigger_point