This repo has been consolidated into the
OperationsPAI/aegismonorepo.
- Code now lives under
rcabench-platform/in the monorepo (original directory names preserved).- Pre-migration git history remains viewable here.
- For all new PRs and issues, go to OperationsPAI/aegis.
- This repository is archived as of 2026-04-19.
An experiment framework for Root Cause Analysis (RCA), supporting fast development of RCA algorithms and their evaluation on various datasets.
To add this package to another uv-managed project:
# Install lightweight SDK (for algorithm/sampler development)
uv add rcabench-platform
# Install with platform-internal tools (kubernetes, neo4j, clickhouse, etc.)
uv add "rcabench-platform[internal]"
# Install with analysis/visualization tools (matplotlib, plotly, etc.)
uv add "rcabench-platform[analysis]"
# Install everything (equivalent to the old monolithic install)
uv add "rcabench-platform[all]"The base package includes only the SDK core for developing and evaluating RCA algorithms and trace samplers. Use optional dependency groups to install additional functionality:
[internal]— Platform server clients, cloud storage, dataset converters, SDG builder[analysis]— Visualization, statistical analysis, Streamlit dashboards[all]— All of the above
See Package Restructuring Plan for details.
- User Guide: Complete guide for using rcabench-platform as both a console command and SDK.
- Development Guide: How to set up the development environment and contribute to this project.
- Specifications: Our design details about RCA algorithms and data formats.
- Workflow References: How to use the functionalities of this project.
- Maintenance: Guidelines for maintaining the project and release procedures.
- Package Restructuring Plan: v3 directory structure, v2→v3 migration guide, and optional dependency groups.