AgentSkillsCN

airflow

通过 VS Code 扩展工具，管理 Apache Airflow 的各项操作，包括列出、运行与调试 DAG，查看日志，以及检查服务器状态。

SKILL.md
--- frontmatter
name: airflow
description: Manages Apache Airflow operations including listing, running, and debugging DAGs, viewing logs, and checking server status using the VS Code extension tools.

Airflow Operations

Use the Airflow VS Code extension tools for all Airflow interactions. Avoid shell commands for Airflow operations.

Tool Quick Reference

ToolPurpose
list_active_dagsList active (unpaused) DAGs
list_paused_dagsList paused DAGs
get_running_dagsShow running or queued DAG runs
pause_dagPause a DAG
unpause_dagUnpause a DAG
trigger_dag_runTrigger a DAG run
get_dag_runsList DAG runs for a DAG
get_dag_historyDaily run history for a DAG
get_failed_runsFailed runs across DAGs
analyse_dag_latest_runFull analysis of latest run (tasks, logs, source)
get_dag_run_detailFull analysis of a specific run
get_dag_source_codeRetrieve DAG source code
go_to_dag_viewOpen DAG view panel
go_to_dag_log_viewOpen log view panel
go_to_connections_viewOpen connections panel
go_to_variables_viewOpen variables panel
go_to_providers_viewOpen providers panel
go_to_configs_viewOpen configs panel
go_to_plugins_viewOpen plugins panel
go_to_server_health_viewOpen server health panel

User Intent Patterns

DAG Operations

  • •"List all DAGs" -> list_active_dags and list_paused_dags
  • •"Show DAG source" -> get_dag_source_code
  • •"Pause DAG X" -> pause_dag
  • •"Resume DAG X" -> unpause_dag

Run Operations

  • •"Trigger DAG X" -> trigger_dag_run
  • •"Show recent runs" -> get_dag_runs
  • •"Why did this fail?" -> analyse_dag_latest_run or get_dag_run_detail

Logs and Troubleshooting

  • •"Show logs for task" -> go_to_dag_log_view or analyse_dag_latest_run
  • •"Analyze latest run" -> analyse_dag_latest_run

System and Config

  • •"Show connections" -> go_to_connections_view
  • •"Show variables" -> go_to_variables_view
  • •"Check server health" -> go_to_server_health_view

Common Workflows

Investigate a Failed Run

  1. •Find failures: get_failed_runs
  2. •Analyze latest run: analyse_dag_latest_run
  3. •If a specific run ID is known: get_dag_run_detail
  4. •Open logs if needed: go_to_dag_log_view

Trigger and Monitor

  1. •Trigger: trigger_dag_run
  2. •Check status: get_dag_runs
  3. •Analyze if failed: analyse_dag_latest_run

Morning Health Check

  1. •Check server health: go_to_server_health_view
  2. •Check running DAGs: get_running_dags
  3. •Review recent failures: get_failed_runs