Google Professional Cloud Database Engineer Question 154
Single answerGoogle Cloud PlatformA company runs an e-commerce platform on Google Cloud. They use Cloud SQL for the relational database, and their business requires minimal downtime and minimal data loss in case of a disaster. The company sets an RTO of 10 minutes and an RPO of 5 minutes. Which of the following design choices would best meet these requirements?
- A
Enable automated backups and set a backup retention period of 7 days.
- B
Ensure high availability (HA) is enabled for the Cloud SQL instance.
- C
Configure point-in-time recovery (PITR) and enable binary logging on the Cloud SQL instance.
- D
Use a read replica in another region for disaster recovery.
Show answer and explanation
Correct answer: C
Explanation
Meeting a strict RPO of 5 minutes and an RTO of 10 minutes requires a solution that ensures data can be restored to a recent state (RPO) and the database can be operational quickly (RTO). Point-in-time recovery (PITR) with binary logging enables rapid restoration to any point within the retention period, fulfilling these requirements. Other options, while useful for database resilience or performance, do not fully address the defined RTO and RPO.
- A. Incorrect.
Automated backups are important for long-term recovery, but they are insufficient to meet a 5-minute RPO as they typically run on a daily schedule. This option does not meet the RPO or RTO requirements.
- B. Incorrect.
Enabling high availability (HA) increases database resilience within the same region but does not address data loss (RPO) or cross-region disaster recovery needs. This option does not fully meet the requirements.
- C. Correct.
Point-in-time recovery (PITR) with binary logging allows restoration to any point within the specified retention period, which is critical to achieving the 5-minute RPO. This option meets both the RPO and RTO requirements.
- D. Incorrect.
Read replicas improve performance and provide a failover option for certain use cases, but they are not suitable for complete disaster recovery or meeting a 5-minute RPO since replicas can lag behind the primary instance.