Airflow Scheduler liveness probe crashing (version 2.0)
I managed to fix my restart by setting up the following configs:
[kubernetes]...delete_option_kwargs = {"grace_period_seconds": 10}enable_tcp_keepalive = Truetcp_keep_idle = 30tcp_keep_intvl = 30tcp_keep_cnt = 30
I have another Airflow instance running in AWS - Kubernetes. That one runs fine with any version, I realized the problem is with Azure Kubernetes, the rest api calls to the api server.
Just in case this helps someone else....