Apache Druid

you should first narrow it down to how much time, an individual sub-task is taking. if that's high and volume processed is low, then I will suggest optimizing that task runtime

you can look at the sub tasks for your main task by filtering on the task group id

I have only one task per compaction interval

actually trying to edit the value for `maxNumConcurrentSubTasks` causes the task to fail without any meaningful error in the logs

if you're having issues with classic native batch ingest or compaction, I'd suggest upgrading and trying out the SQL-based approach. It has been totally rewritten to be more reliable

a relevant example: <https://druid.apache.org/docs/latest/multi-stage-query/examples.html#replace-for-replacing-a-specific-time-segment>

this example is basically equivalent to a compaction: it does a `REPLACE` from a datasource into itself for a specific time range

<@U0344FW86DD> Thank you
but upgrade is not an option right now