I have just got Airflow working on K8 through the help of this great project, thanks!
I had a question about the embedded DAG's approach. In the event of a deployment or a pod being rescheduled by K8, how would a running DAG survive the process?
Or is the answer quite simply it wouldn't and it would need to retry?
Any other wise words on having a reliable Airflow on K8 are welcome
I have just got Airflow working on K8 through the help of this great project, thanks!
I had a question about the embedded DAG's approach. In the event of a deployment or a pod being rescheduled by K8, how would a running DAG survive the process?
Or is the answer quite simply it wouldn't and it would need to retry?
Any other wise words on having a reliable Airflow on K8 are welcome