We are going to be a databricks customer and did some PoC tests. Our one test contains dataset in one partitioned table (15colums) is roughly 250M rows, each partition is ~50K-150K rows.

Occasionally we have hundreds (up to one thousand) concurrent users, each querying (sql select) different partition. According to

https://community.databricks.com/s/question/0D58Y00009WgtBiSAJ/data-bricks-sql-is-allowing-10-querie...

https://community.databricks.com/s/question/0D58Y00009ZuIHUSA3/sql-queries-on-databricks

, I can only run 10 concurrent SQL queries per cluster. So if I need to run more queries concurrently, I can increase number of clusters per SQL Warehouse (e.g. if I have 10 clusters per SQL Warehouse I can run 10x10=100 concurrent queries, isn't it?). I think that similar behaviour is possible to accomplish with more 'High concurrent clusters'.

Or are there others possibilities how to accomplish the requirement of 1 thousand of concurrent queries? Thx.

M.