Skip to content
Get packs

Managed Airflow DAG Troubleshooting

Diagnoses failed Airflow DAG runs and task instances on Google Cloud's managed Airflow using logs and code.

Source Download skill
Repository
google/skills
Stars
21,782
Licence
Apache-2.0
Files
2 files

What it does

Airflow is a tool for scheduling data workflows, and a DAG is one workflow made of tasks. When a workflow run fails on Google Cloud's Managed Service for Apache Airflow (formerly Cloud Composer), this skill tells the agent how to investigate: it fetches the task's error logs, checks each task's state, downloads the DAG's source code, and compares the code's settings against the error messages. The agent then reports the specific failed task, the exact log text as evidence, and a concrete fix.

Great for

  • Data engineers whose Airflow DAG run or task instance has failed on Cloud Composer or Managed Airflow
  • Platform teams who need a repeatable, evidence-based method for reading Airflow failure logs
  • Developers who want the agent to match error codes such as 403 or 404 to the DAG's table IDs or paths

Skip it if

  • Overall performance tuning or sizing recommendations for an Airflow environment, which the skill explicitly excludes; using it for that gives off-target results
  • Self-hosted Airflow or other clouds, since the commands target Google Cloud's gcloud composer and logging tools
  • Users without gcloud access to the project: the skill cannot read logs or files without credentials, and it may not be able to fully diagnose the failure

Try asking

  • “My Composer DAG run 'daily_sales_load' failed overnight. Find the failed task and explain why using the logs.”
  • “Task 'load_to_bq' in my Managed Airflow DAG keeps failing with a 404. Check the dataset and table IDs in the DAG code against the error.”
  • “Why did task instances in run 2025-01-10 of my orders DAG time out with AirflowTaskTimeout?”
Step-by-step guide: how to use Managed Airflow DAG TroubleshootingInstall, first run, examples, tips and FAQs

Install in 3 steps

  1. 1

    Download the skill

    The ZIP contains the 2 original files, the Apache-2.0 licence and an attribution note.

    managed-airflow-dag-troubleshooting.zip
  2. 2

    Add it to your agent

    Claude Code (all projects)

    mkdir -p ~/.claude/skills && unzip managed-airflow-dag-troubleshooting.zip -d ~/.claude/skills/

    Claude Code / Agent SDK (this project only)

    mkdir -p .claude/skills && unzip managed-airflow-dag-troubleshooting.zip -d .claude/skills/

    Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.

  3. 3

    Ask for the task

    No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.

Get better results

  • Give the agent the environment name, location, DAG ID and run ID so it can run the state and log commands directly.
  • Review every gcloud command the agent proposes before approval, since it will read logs and copy files from your project's storage bucket.
  • Ask for the verbatim log line and code line behind each conclusion, as the skill is designed to show evidence rather than general advice.

What's inside

  • references/attribution.md
  • SKILL.md

Redistributed unmodified from google/skills under the Apache-2.0 licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.