A retailer uses Amazon Aurora PostgreSQL for live transactions and Amazon Redshift as a data warehouse. A daily ETL updates Redshift from PostgreSQL. To reduce Redshift costs, the engineer must keep only the most recent 15 months of data in Redshift, archive historical data, and still run analytics that combine live PostgreSQL data, current Redshift data, and archived historical data. Which steps will satisfy these requirements? (Choose two.)

The correct options address the need to query live PostgreSQL data, manage Redshift data retention, and access archived historical data. Configuring Amazon Redshift Federated Query to query live transactional data in PostgreSQL directly allows analytics to combine current Redshift data with real-time operational data without ETL. Scheduling a monthly job to UNLOAD data older than 15 months to Amazon S3, deleting it from Redshift, and configuring Redshift Spectrum to access this historical data in S3 satisfies the cost reduction and archiving requirements. Redshift Spectrum can query data directly in S3, enabling combined analytics with current Redshift data and live PostgreSQL data (via Federated Query). Incorrect options: Redshift Spectrum cannot directly query live data in a PostgreSQL database; it queries data in S3 or other data lakes. S3 Glacier Flexible Retrieval is for archival storage and is not directly queryable by Redshift Spectrum without prior restoration to S3 Standard, making it unsuitable for immediate analytical access. Materialized views in Redshift combine data within Redshift or from external sources accessible by Redshift, but they don't directly query live PostgreSQL data or data in Glacier Flexible Retrieval.