Examkin

Soru

Zorluk: ZorMulti-AZ, Multi-Region Architectures and Disaster Recovery (DR)

A company operates a food delivery dispatch platform in the us-east-1 Region. The architecture utilizes Amazon ECS on AWS Fargate for the microservices and an Amazon RDS for PostgreSQL Multi-AZ DB instance for the transactional database. The company needs to implement a disaster recovery (DR) strategy in the us-west-2 Region. The DR solution must achieve a Recovery Time Objective (RTO) of 30 minutes, a Recovery Point Objective (RPO) of 5 minutes, and minimize ongoing running costs in the secondary Region. Which strategy should a solutions architect recommend to meet these requirements?

  1. Configure an RDS for PostgreSQL cross-region read replica in the us-west-2 Region. Deploy an Application Load Balancer (ALB) in us-west-2 and configure an ECS service with a desired task count of zero. Set up Amazon Route 53 with failover routing and health checks. During a disaster, promote the read replica to a standalone DB instance, update the ECS service desired count to match production levels, and let Route 53 redirect traffic to the secondary ALB.Cevap
  2. B
    Configure an RDS for PostgreSQL cross-region read replica in the us-west-2 Region. Deploy an ALB and configure an ECS service in us-west-2. Set up Route 53 with active-active latency routing to distribute traffic. If the primary Region experiences an outage, rely on the RDS cross-region read replica to automatically promote itself and begin accepting write traffic from the secondary Region's ECS tasks.
  3. C
    Configure AWS Backup to take hourly snapshots of the primary RDS database and copy them to the us-west-2 Region. Deploy an ALB in us-west-2 and run a scaled-down ECS service with a desired task count of one. During a failover event, restore the database from the copied snapshot and scale up the ECS tasks.
  4. D
    Configure an RDS for PostgreSQL cross-region read replica in the us-west-2 Region. Deploy an ALB and ECS tasks in us-west-2 running at a desired task count of zero. Configure Route 53 with latency-based routing to direct traffic, assuming Route 53 will automatically divert all users to the secondary Region's ALB during a primary Region outage.

Cevap

Configure an RDS for PostgreSQL cross-region read replica, set the secondary ECS task count to zero, use Route 53 failover routing with health checks, and promote the replica and scale up tasks during a failover event.
The correct strategy uses a Pilot Light approach. A cross-region RDS read replica replicates data asynchronously, keeping the RPO to a few seconds, which satisfies the 5-minute requirement. Running the secondary ECS service with a desired task count of zero minimizes running costs. Promoting the database to primary status and scaling up the ECS service tasks to production levels can be done in a few minutes, satisfying the 30-minute RTO.

Adım Adım Çözüm

1
Select database replication strategy based on the RPO of 5 minutes.
Determine that an RDS cross-region read replica is necessary to continuously replicate data, ensuring data loss remains under 5 minutes.
Hourly backups or snapshots copied to another region would violate the 5-minute RPO limit.
2
Select the compute layer deployment strategy to minimize standby costs while meeting the 30-minute RTO.
Deploy an ALB and configure the secondary Region's ECS service with a desired task count of 0.
An active but scaled-down fleet increases running costs needlessly. Launching and scaling ECS Fargate tasks can be completed within minutes during a failover, well within the 30-minute limit.
3
Configure the DNS routing and failover automation.
Configure Route 53 failover routing with health checks pointing to the primary ALB, and define the manual or scripted steps to promote the RDS read replica and scale up the ECS tasks.
Active-active latency routing without health checks fails to steer traffic away from the degraded primary region.

Anahtar Kavram

Disaster Recovery (DR) Pilot Light Pattern on AWS
Bu soruyu puanla

Konular

Multi-AZ, Multi-Region Architectures and Disaster Recovery (DR)

Bu konudan daha fazla soru