For a single Cloud SQL (MySQL second-generation) instance with business-critical transactions, which two features minimize potential data loss in catastrophic failure? (Choose two.)

Automated backups create point-in-time recovery points, allowing you to restore your database to a previous state. This is crucial for recovering from data corruption or accidental deletion. Binary logging (binlog) records all data modifications, enabling point-in-time recovery to a specific transaction, even between backups. This minimizes data loss by allowing you to replay transactions up to the moment of failure. Sharding distributes data across multiple instances, improving scalability but not directly minimizing data loss on a single instance. Read replicas enhance read performance and provide high availability for reads, but they don't prevent data loss on the primary instance. Semisynchronous replication ensures that transactions are committed to at least one replica before being acknowledged to the client, reducing data loss in a primary failure, but it's not a standalone feature for minimizing data loss in catastrophic scenarios like backups and binlogs.