Team .,, in our of the prod cluster where we are seeing latency issue , it takes 4 to 5s to get the result. other datasource works fine and doesnt have any latency. any suggestion please

select tenantId, systemId, TIMESTAMP_TO_MILLIS(__time) as "timestamp",  sum(iops_pref_pct) as iops_pref_pct  from (select DISTINCT(__time),* from "xyzdatasource" where  systemId='aaajjjjccccc' and __time >= MILLIS_TO_TIMESTAMP(1742252400000) and  __time <= MILLIS_TO_TIMESTAMP(1742338800000)) group by __time, tenantId,systemId order by __time asc

Hi Sivakumar, Your subquery doesn't make sense to me. First, the function syntax

distinct(colName)

doesn't look like valid SQL to me, at least not Druid SQL ... but the parser appears to be accepting it and treating it like

select distinct colName

... and then you have added

, *

, which should be invoking a group by on every column in your table ... which can be very expensive, if not overruning the max subquery buffer. What are you trying to achieve with that subquery? Thanks. John