用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
直接命令不会经过审查 Prompt；运行前请先检查来源。
npx skills add https://github.com/vamseeachanta/workspace-hub --skill airflow-common-issues命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地？可下载 SkillsMP 当前能够提供的文件。
基于 SOC 职业分类
正在显示 SKILL.md
| name | airflow-common-issues |
| description | Sub-skill of airflow: Common Issues (+1). |
| version | 1.0.0 |
| category | operations |
| type | reference |
| scripts_exempt | true |
Issue: DAG not appearing in UI
# Check for import errors
airflow dags list-import-errors
# Validate DAG file
python dags/my_dag.py
# Check scheduler logs
docker logs airflow-scheduler-1 | grep -i error
Issue: Tasks stuck in queued state
# Check worker status
airflow celery status
# Verify executor configuration
airflow config get-value core executor
# Check for resource constraints
kubectl top pods -n airflow
Issue: XCom size limits
# Use external storage for large data
def store_large_result(**context):
# Store in S3 instead of XCom
s3_hook.load_string(large_data, key='results/data.json', bucket='my-bucket')
return 's3://my-bucket/results/data.json' # Return reference only
Issue: Scheduler performance
# Tune scheduler settings
AIRFLOW__SCHEDULER__PARSING_PROCESSES: 4
AIRFLOW__SCHEDULER__MIN_FILE_PROCESS_INTERVAL: 30
AIRFLOW__SCHEDULER__DAG_DIR_LIST_INTERVAL: 60
# Add detailed logging
import logging
logger = logging.getLogger(__name__)
def my_task(**context):
logger.info(f"Starting task with context: {context}")
# ... task logic
logger.debug(f"Intermediate result: {result}")
# Test specific task
airflow tasks test my_dag my_task 2026-01-15
# Clear task state for re-run
airflow tasks clear my_dag -t my_task -s 2026-01-15 -e 2026-01-15
# Trigger DAG run
airflow dags trigger my_dag --conf '{"key": "value"}'