Automated Root Cause Analysis
This cookbook demonstrates how to integrate Chicory AI with your data stack in order to perform Root Cause Analysis. The agent is triggered on an Airflow DAG pipeline failure.
What You’ll Build
An Airflow Dag that:
Runs Data Quality Validation checks across the data source and transformed data via DBT
Connects to Pager Duty to send incident alerts for pipeline failures
Connects to Chicory Agent to perform Root Cause Analysis
An agent that:
That scans the metadata across Bigquery, DBT and Airflow via MCP
Scans through supporting documentation and code repositories
Performs detailed upstream and downstream analysis
Generates Root Cause Analysis and suggests fix's by raising a Github Issue.
Git Action:
That updates Pager Duty Incident once the Git Issue is reviewed, acknowledged and closed.
Contents
Introduction – Introduction to setup and tools
Workflow - In detail workflow
Agent Creation – Creating a Chicory Agent
Sample Catalog – Sample analysis output
Troubleshooting – Common issues & fixes
Last updated