I have a table that consists of 56millions rows.
This table is handling high load of UPSERTS every 5 minutes as it's loading streaming data from KAFKA. Approx 200-500k updates every load.
When I run a SELECT with an ORDER BY against one of the timestamp columns, it takes over 5-7 minutes to return a result.
I tried Cluster Key for that column but since there is a high DML operation on that table and high cardinality on the column itself, the clustering was ineffective and costly.
So far, the only think that has significantly reduced query time to about 15 seconds is increasing the warehouse size to an X-Large from a Small.
I am not convinced that the only solution is to increase the warehouse size. Any advice here would be great!