Hi Druids I'm ingesting data via kafka supervisor spec with segment granularity as

DAY

and there are 61 segments created by Druid in a Day with around 1000 rows are there in the data source, Since I'm trying to query the data via

Scan

Druid native query but getting this below error, Please help me to get things working.

"{"error":"Resource limit exceeded","errorClass":"org.apache.druid.query.ResourceLimitExceededException","host":"localhost:8083","errorCode":"legacyQueryException","persona":"OPERATOR","category":"INVALID_INPUT","errorMessage":"Time ordering is not supported for a Scan query with 61 segments per time chunk and a row limit of 9,223,372,036,854,775,807. Try reducing your query limit below maxRowsQueuedForOrdering (currently 100,000), or using compaction to reduce the number of segments per time chunk, or raising maxSegmentPartitionsOrderedInMemory (currently 50) above the number of segments you have per time chunk.","context":{"host":"localhost:8083","errorClass":"org.apache.druid.query.ResourceLimitExceededException","legacyErrorCode":"Resource limit exceeded"}}"

There is a limit in the number of segments opened per time interval for some query types. Have you tried compacting your data to reduce the segment count (i.e. fragmentation)? With only 1000 rows I would assume the data can all fit into one segment.

Yes, I tried compacting but it does on next day and not for current day

Are you on a newer release of Druid which has concurrent append/replace? If so, then you can turn that on and do compaction for current day as well. Based on the relatively low volume of records coming in it seems to me that this would work well for you ... https://druid.apache.org/docs/latest/ingestion/concurrent-append-replace

judging by the error, it seems that you are wanting to do an

order by __time

with no

limit

? if you know that there will be only around 1000 rows, you can add

limit 10000

(anything less than

100000

is fine) and you won't run into this error

I had to add "maxSegmentPartitionsOrderedInMemory": 100 in the scan query to make it work

that would work too!