Auto-compaction clarification

When I submit a compaction task via druid console ( Load Data/ Other / Submit Task)

The task is correctly submitted and performed (I can change segment granularity for example).
I understand this action as being run once.

example of submitted json :

  "type": "compact",
  "dataSource": "a-datasource",
  "tuningConfig" : {
    "type" : "index_parallel",
    "maxRowsPerSegment" : 5000000,
    "maxRowsInMemory" : 1000000
  "ioConfig" : {
    "type" : "compact",
    "inputSpec": {
        "type" : "interval",
        "interval" : "2018-01-01T00:00:00.000/2020-12-31T23:59:00.000"
  "segmentGranularity" : "all",
  "context" : {
  "priority" : 100,
  "forceTimeChunkLock": true

In documentation, I find mentions of auto-compaction, which seem to be a periodic compaction task run on segments of a datasource.

Please clarify the difference between compaction and auto-compaction, and the way autocompaction can be activated via API (until now, the only way I found was to activate it via druid console, in datasources menu, Compaction column

Things I've tried

curl -H “Content-Type: application/json” -X GET http://IP:8081/druid/coordinator/v1/compaction/status

is it the status of auto-compaction tasks ? or all compaction tasks ?


Relates to Apache Druid 0.20.1

I think you might find an answer in an earlier post -

1 Like