Introduction

This cookbook shows how to use Chicory AI agents to generate a Root Cause Analysis report for a failed DAG

Tools & Integrations

  • BigQuery MCP – SQL execution and runtime insights

  • REST API – Deployment mechanism for Chicory agents

  • Github MCP – Github code repository for supporting code/documentation

  • DBT MCP – DBT connection for pipeline runs

Problem: Developers frequently receive PagerDuty/incident alerts for pipeline failures. Diagnosing root causes requires manually connecting information across multiple systems—logs, data lineage tools, orchestration platforms, and data quality monitors, which is time-consuming and delays incident resolution.

Quick Start

  1. Build an Airflow Pipeline Dag

  2. Set Pager Duty Alerts for incident notifications

  3. Build the agent and deploy it

PreviousAutomated Root Cause AnalysisNextWorkflow

Last updated