This message was deleted.

Do you have metrics enabled on the cluster. You should look for bottlenecks in the configuration. This might be a symptom of http thread exhaustion which would tend to queue queries and make them wait. It might be other issues, but reviewing query metrics and resource metrics should shed some light.

You can also look at https://druid.apache.org/docs/latest/operations/basic-cluster-tuning.html to understand more about tuning the cluster. I have a couple of questions, • Could you clarify whether you’re currently running stream ingestion, batch ingestion jobs, or both? • Is there a specific datasource that you’ve observed to be slower than others? • Are you utilizing the MiddleManager or Indexer for your data processing tasks?