A company operates a provisioned Amazon Redshift cluster used for ETL that supports critical analytics. A separate Redshift cluster is used by the sales team for BI. The sales team needs to access ETL cluster data weekly and join it with their BI cluster data without disrupting the ETL workloads and while minimizing compute usage on the ETL cluster. Which solution meets these requirements?

Redshift data sharing allows the sales team's BI cluster (consumer) to access data directly from the ETL cluster (producer) without moving or copying data. This minimizes compute usage on the ETL cluster because queries run on the consumer cluster, and it avoids disruption to ETL workloads. Materialized views and database views, even with direct access, would still execute queries on the ETL cluster, consuming its resources and potentially impacting performance. Unloading data to S3 and using Redshift Spectrum would involve data movement, increasing complexity and potentially latency, and Redshift Spectrum is generally slower for complex joins compared to native Redshift queries.