I don't think it's necessary to split the RequestLogger implementations into 
native/SQL versions, it's enough that there's a new SqlRequestLogLine 
implementation.

If you want to have two different request loggers for SQL and native queries, I 
think would be better to bind the SQL provider to a different configuration 
parameter and inject the desired provider using annotations (maybe like how you 
can get a `@Coordinator` DruidLeaderClient and a separate `@IndexingService` 
instance)

[ Full content available at: 
https://github.com/apache/incubator-druid/pull/6302 ]
This message was relayed via gitbox.apache.org for [email protected]

Reply via email to