ravisharm opened a new issue, #17169:
URL: https://github.com/apache/druid/issues/17169

   Currently, at Broker, while Druid does emit ‘query/time’ metric that 
captures overall time taken to complete the query, it doesn’t emit any Broker 
metric that captures time taken to plan the query before the query gets sent to 
the Data nodes. This is essentially the time spent to determine segments to 
query, grouping segments by server etc. When `query/time` shows a high value, 
OnCall needs to understand the reason for the high latency. The only relevant 
metric is `query/node/time`. If it shows normal values, the natural question 
becomes how much time did broker take to process the query. Broker spends some 
CPU cycles before fanning out the query (the subject of the proposed metric) 
and later spends some CPU cycles to do the final aggregation. Using the 
proposed metric we can pinpoint if fanout took longer and further investigate 
that part of the query lifecycle if needed.
   


-- 
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