I’d like to use query cancellation api (http://druid.io/docs/0.8.1/querying/querying.html) when query computing time is too long for users to wait and reduce burden of druid cluster.
However, a query id in the response header field ‘X-Druid-Query-Id’ can only be found when full response has came over, which is not useful for my case.
Is there any idea or point I have missed for this case? Please let me know how to use query cancellation api efficiently.
(In my case, I use Java client to query to druid cluster.)