My major concern is that this would hide serious problems with the connection to the DB (say the DB crashed and isn't running anymore, or the tables are corrupt) - this would then end up creating lots of logs but just keep trying endlessly.
Maybe it's worth only allowing _n_ failures in a 60s window, or do you think I am worrying needlessly? [ Full content available at: https://github.com/apache/incubator-airflow/pull/3650 ] This message was relayed via gitbox.apache.org for [email protected]
