Daniel Mateus Pires created AIRFLOW-3420:
--------------------------------------------
Summary: Kubernetes executor sync floods the logs
Key: AIRFLOW-3420
URL: https://issues.apache.org/jira/browse/AIRFLOW-3420
Project: Apache Airflow
Issue Type: Bug
Reporter: Daniel Mateus Pires
Hi there,
I'm using Airflow on Kubernetes with the Kubernetes Executor and I'm getting
flooded by:
[2018-11-29 17:17:23,427] \{kubernetes_executor.py:570} INFO - self.running:
{('xxxx', 'xxxx', datetime.datetime(2018, 11, 29, 17, 6, 32, 935655,
tzinfo=<TimezoneInfo [UTC, GMT, +00:00:00, STD]>))
This logging happens at a very high frequency (I think I'm getting in the order
of 1000s of entries per second coming just from this line)
Could we move this logging to a place that is called less frequently or set the
LEVEL to DEBUG ?
To give you more context, it renders my logs quite hard to browse / or even get
through kubectl, I'm suspecting it is the cause of my scheduler instance pod
being Evicted every few hours for DiskPressure on the imagefs
Those logs come from
[https://github.com/apache/incubator-airflow/blob/master/airflow/contrib/executors/kubernetes_executor.py#L593]
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
