We have a case where we see a lot of long running transactions that are blocking other transactions. The blocking transaction is always the same, but seems to originate from different user sessions. It's a relatively innocent looking select statement. Basically in some cases we've had to kill the transaction because it was running for ~hour.
I know there could be many reasons for the poor performance, but has anyone seen especially this query performing poorly?
Advanced warehouse is used. InventDim table has 3,4 million records.
We are running index maintenance (reorganize if <50% fragmented, otherwise rebuild) twice a week, and a statistics update (Tables and Views, Full scan) every night.
We are using trace flag 1224.