nivdror commented on issue #39939: URL: https://github.com/apache/airflow/issues/39939#issuecomment-2140546076
Hi, we won't use datasets. Our architecture is built mainly on three layers - scheduler layer, context layer and compute layer. Each got it's dependencies and definition. So we see airflow as scheduler which doesn't need to what it's running(spark, MapReduce etc') nor where (CDH, EMR, Databricks etc') nor the data artifacts we output. This helps us migrate through different providers of Airflow + plus helps us decouple the dependencies. In summary, we don't put all of our eggs in one basket. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
