Can customers opt out of Redshift's serverless compute engine?

Question

In the recent [Redshift announcement](https://aws.amazon.com/about-aws/whats-new/2020/06/amazon-redshift-now-delivers-better-cold-query-performance/) I read: 
"query compilations are now scaled to a serverless compute engine beyond the compute resources of the leader node of your cluster".

The queries can contain PII information in the text and having this information processed in what looks like a multi-tenant compiling environment might be an issue in some industries.
I have had customers skipping even local logging of query logs to ensure such information can't get exposed.

Marco Ullasci · Accepted Answer

The query parsing, the optimization and the plan creation (the parts that potentially use the PII data in the query text and/or in the statistic) are carried over inside the leader node.
Only the C++ code created at the end of the optimization (that doesn't contain any PII data) is sent to the serverless infrastructure to be compiled.

This means that there is no reason to need an opt-out option.

Can customers opt out of Redshift's serverless compute engine?

Relevant content