airflow - Storing execution parameters in the Airflow database
问题描述
I am creating DAGs having parameters changing (possibly) every run.
Is there a way to store them in the Airflow metadata db? This would allow having a single db containing full run audit (list of runs with their parameters, statuses, etc).
Currently I use XComs pushing them within the execute method of operators but (except for the fact that's not what XComs are for) it's not easy to query them (as they are serialized by SqlAlchemy to blobs).
解决方案
您可以将它们存储为Airflow 变量- 默认情况下它不会像 XComs 那样存储执行日期,但您应该能够手动添加它。
推荐阅读
- javascript - 相对于 HMD 面的位置变化
- android - 如何通过 hid-usb 驱动旋转安卓触摸屏
- ruby-on-rails - Rails 5 - 检索 MiniMagick 版本的图像(CarrierWave/Google Cloud Storage)
- go - 证明在函数内部模拟函数返回
- java - Java Spring boot - OnceRequestPerFilter 只允许控制器请求映射
- python - 仅针对特定文件夹的图像数据生成器扩充参数 (Keras)
- zabbix - IBM DS5020 SNMP 监控
- sql-server - Authentication="Active Directory Integrated";` 试图以我的计算机而不是使用我的 Active Directory 帐户登录到 SQL Server?
- javascript - 如何从jquery中的动态下拉列表中选择多个值
- java - 如何从线程中获取要在父线程中使用的值?