Point-in-time recovery
Being able to restore a database to an arbitrary point in time, not just the most recent full backup. Requires continuous WAL/binlog archiving plus a base backup. A standard requirement for production DBs.
PostgreSQL: archive_mode = on + pg_basebackup + replay WAL up to recovery_target_time. MySQL: binlog shipping + restore + --stop-datetime. RPO (Recovery Point Objective) is governed by WAL frequency — 5 min is common. RTO depends on base-backup age + WAL volume to replay. Managed databases (RDS, Cloud SQL) make this transparent — a slider for "restore to any second within X days". Open source: pgBackRest, Barman, WAL-G.