A database selection must satisfy a 5-minute recovery point objective and a 30-minute recovery time objective, not just an attractive benchmark. Which approach best fits?

1. Compare supported recovery mechanisms and test restores against both objectives.
2. Separate consistency-sensitive reads from replica-eligible reads.
3. Select and validate a topology whose documented failure behavior meets the outage scenario.
4. Plan and monitor the supported index-build process under representative workload.