DAG-Sentinel

• 

  Architecture diagram

• 

  Login

• 

  Dashboard

• 

  adk web app

• 

  adk run app

• 

  agents-cli run

• 

  adk web

In modern enterprise environments, data pipelines (such as Apache Airflow / Google Cloud Composer) form the backbone of business intelligence and analytics. However, when a DAG (Directed Acyclic Graph) fails, it halts downstream operations. Resolving these failures is currently a painful, manual process:

1. An on-call engineer gets paged.
2. They manually dig through massive, verbose Airflow task logs.
3. They inspect raw storage files in Google Cloud Storage (GCS) or verify schemas in BigQuery (BQ) to diagnose the issue.
4. They find that the root cause was a simple typo, a transient connection timeout, or a missing upstream table partition.
5. They manually clear/retry the task.

Built With

• adk
• agents
• agents-cli
• cloudrun
• gcp
• gemini
• gemini-enterprise
• terraform