abhishekrb19 commented on PR #16247:
URL: https://github.com/apache/druid/pull/16247#issuecomment-2044137684

   @kfaraz - thanks for the swift review!
   
   Re:
   > After this change, the default druid.coordinator.kill.period effectively 
changes from 1 day to 30 mins (default value of indexing period). So 
potentially, there could be a kill task submitted for every datasource every 30 
mins. I wonder if there can be clusters for which this is harmful / wasteful in 
terms of task resources.
   
   I was thinking of changing the default `killTaskSlotRatio` from 1 to 0.1, so 
the number of kill tasks spun will be bound, preventing starvation of other 
tasks. This will also be in alignment with compaction task slot ratio.
   
   Upon reviewing the reasoning in the code 
[here](https://github.com/apache/druid/blob/master/server/src/main/java/org/apache/druid/server/coordinator/CoordinatorDynamicConfig.java#L443)
 for the current default values of kill task slots, it appears that the 
unbounded values were chosen to preserve old behavior, which I'd argue was not 
ideal to start off with. Given that we're changing the kill period's default in 
this patch, we might as well take an opportunity to adjust adjacent kill 
configurations to set them to safe values for a better out-of-the-box 
experience, and highlight the changes in the release notes. What do you think?
   
   


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]


---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to