I believe it's due to limitation in Redash's legacy architecture (which was acquired by Databricks). Hence Databricks is doing great efforts in increasing this threshold (was originally 20K), so with patience it will get to close to 1 Million. In the meantime, it's recommended to run aggregation/reduction operations on your data.

@Amine El Helou​  when this limit will be lifted? my users won't switch from SSMS to dbx sql unless they can see all results set in UI ( which sometimes is more than 10mm rows)