Describe the bug

This is my LogsQL query:

* | collapse_nums | stats count_uniq(_msg) num 

I received an unfriendly error message while executing this LogsQL query:

error from datasource: cannot execute query [_time:[1758009219000000000,1758095619999999999] | collapse_nums | stats count_uniq(_msg) as num | limit 1000]: cannot import state for count_uniq(_msg): unexpected number of imported shards: 2; want 8

To Reproduce

Using cluster mode, vlselect has different CPU cores than vlstorage. Run the query above with a high time range (enough for sharding to happen).

I suspect both statsCountUniqProcessor and statsCountUniqHashProcessor have an incorrect import state. So it could be fixed by:

1. correcting the import state function
2. adding options(concurrency) in vlselect before sending to remotes.