Hi team, we are have a latency issue with a Pinot query for UI, it’s simple doing a select sum of a Int column from a table with 2 billion total records with 6milliom records scanned after filtering by one inverted index and timecolumnname, but it always took more than 2 seconds. Is this normal and can this latency be improved?

Can you give an example query and the index you have?

Numofdocscaned 9m totaldoc:4b, num of segmentqueried:821 what other meta data might be insightful too?

servers

Number of server queried and responded is both 4

For low latency I’d have a range index on datecolumnname, and a star tree defined with id & datecolumnname columns as dimensions, and

SUM__int_field

as the function column.