A Redshift dashboard repeatedly runs the same expensive aggregate over a large table. The business accepts a defined refresh delay instead of recalculating from the base data for every request. Which feature should the engineer evaluate?

1. Disable refresh and assume all future base-table changes appear in the stored result automatically.
2. A materialized view with a refresh strategy that meets the freshness requirement.
3. A normal view alone, assuming it stores the complete aggregate result permanently.
4. Run ANALYZE after every dashboard request and treat optimizer statistics as cached aggregate results.