The screen shot above illustrates just how easy it is to adjust the size of a Snowflake cluster. Using T-shirt sizing, the operator can choose to resize the cluster from the current Medium size of 4 nodes to a maximum of 128, and this is normally available within seconds.
All queued or subsequently executed queries are automatically started on the larger cluster, and the operation can be scripted and scheduled using SQL to automatically start a larger cluster before running a large and demanding task.
The Challenge Revisited
The charts below illustrate the batch ETL/ELT and business intelligence query profile on the Snowflake platform. Unlike the traditional approach, the user groups can be deployed on entirely independent clusters, each sized as appropriate. This is simply not possible on most database platforms including the cloud based managed solutions provided by AWS Redshift and Oracle or IBM.