Re: [druid-user] QueryInterruptedException occurs when querying for large number of rows in druid cluster

Running multiple parallel queries will get you faster results for the same resources. When a select query without order by executes the broker streams the results. When you execute as a single query you can make use of only one http thread on the broker to stream out the results. With multiple queries you can make use of multiple http threads on the broker. This will increase the throughput. For aggregation queries…the historicals use one merge thread per query typically. If you have multiple queries then you get multiple merge threads. So basically it boils down to parallelizing the work. Single large queries tend to create bottlenecks which will be avoided using multiple concurrent queries.