Details
-
Improvement
-
Status: Closed
-
Major
-
Resolution: Fixed
-
None
-
None
Description
Our current log config template is very generic and invites users to always set the root logger to DEBUG if they want to get more details. Since Flink depends on libraries like Akka and it's common to run Flink with other systems like Hadoop or Kafka, this also increases the log levels for those systems.
I would propose to split the default logger configuration up and provide a separate debugging template. Furthermore, some noisy loggers might turned off by default, for example HadoopFileSystem.