Apache Druid

Hi, what do you mean by `releasing ingestion jobs`? You can always reduce `taskDuration`

if you are question is can ingestion tasks be paused if there is not data in kafka then the approach would be to use the metrics and check the number of events processed and when that is less than a threshold use the API to submit a new supervisor spec with a lesser number of tasks.

Yes, what Vijay said is what I meant, Thanks

Hi Vijay, what actually idle supervisor configuration would do?
<https://druid.apache.org/docs/latest/development/extensions-core/kafka-supervisor-reference.html#idle-supervisor-configuration>

this should also do what you want. It should be able to scale the tasks up and down based on kafka lag