Examkin

Question

Difficulty: Very hardHigh Availability and Disaster Recovery Design

A healthcare provider is designing a multi-region disaster recovery (DR) architecture for a critical patient telemetry application. The application's web and ingestion layers run on AWS Fargate behind Application Load Balancers (ALBs) in both the primary region (`us-east-1`) and the recovery region (`us-west-2`). The ingestion tasks in both regions must query a shared metadata store and resolve the internal microservice address `api.telemetry.local` using a Route 53 Private Hosted Zone (PHZ) created in the primary VPC. The application must also send outbound alert notifications to external clinical endpoints via the internet. The architecture must support a Recovery Time Objective (RTO) of 1010 minutes and a Recovery Point Objective (RPO) of 11 minute. The recovery region must run as a cost-efficient warm standby, keeping running resources to a minimum until a failover is declared. The solution must also remain highly available within each region, ensuring no single Availability Zone (AZ) failure can disrupt telemetry ingestion or outbound alerting. Which configuration meets these requirements while minimizing cost?

  1. A
    Deploy Amazon Aurora Global Database with the primary cluster in us-east-1 and a secondary cluster in us-west-2, running a single db.r6g.large reader instance in the secondary cluster. Configure Route 53 Active-Passive Failover routing policy with health checks to route public traffic to the primary ALB, failing over to the secondary ALB if health checks fail. Set up a VPC Peering connection between the primary and secondary VPCs to route DNS queries to the Route 53 Private Hosted Zone (PHZ) associated with the primary VPC. In each VPC, deploy a NAT Gateway in each of the two utilized Availability Zones.
  2. B
    Deploy Amazon Aurora Global Database with the primary cluster in us-east-1 and a secondary cluster in us-west-2, running a single db.r6g.large reader instance in the secondary cluster. Associate the api.telemetry.local Private Hosted Zone with both the primary and secondary VPCs. In each VPC, deploy a single NAT Gateway in one Availability Zone to handle outbound internet traffic for Fargate tasks across all private subnets. Configure Route 53 Active-Passive Failover routing policy with health checks to route public traffic to the primary ALB, failing over to the secondary ALB if health checks fail.
  3. Deploy Amazon Aurora Global Database with the primary cluster in us-east-1 and a secondary cluster in us-west-2, running a single db.r6g.large reader instance in the secondary cluster. Associate the api.telemetry.local Private Hosted Zone with both the primary and secondary VPCs. In each VPC, deploy a NAT Gateway in each of the two utilized Availability Zones. Configure Route 53 Active-Passive Failover routing policy with health checks to route public traffic to the primary ALB, failing over to the secondary ALB if health checks fail.Answer
  4. D
    Deploy a single-region Amazon RDS for MySQL database in us-east-1. Configure AWS Backup to copy automated database snapshots to us-west-2 every hour. Associate the api.telemetry.local Private Hosted Zone with both the primary and secondary VPCs. Deploy a NAT Gateway in each of the two utilized Availability Zones in both VPCs. Configure Route 53 Active-Passive Failover routing policy with health checks to route public traffic to the primary ALB, and use an AWS Lambda function to restore the database from the latest copied snapshot during a regional failover.

Answer

The configuration using Amazon Aurora Global Database, associating the Private Hosted Zone with both VPCs, deploying NAT Gateways across multiple Availability Zones in each region, and routing traffic using Route 53 Active-Passive Failover with health checks.
The correct solution leverages Amazon Aurora Global Database to achieve cross-region sub-second data replication and quick managed failovers, satisfying the 1-minute RPO and 10-minute RTO. It ensures local name resolution for the microservice endpoint by associating the Route 53 Private Hosted Zone with both VPCs. Furthermore, it protects outbound alerting availability by deploying multi-AZ NAT Gateways in each region.

Step-by-Step Solution

1
Select a database disaster recovery pattern that supports the RPO of 1 minute and RTO of 10 minutes.
Identify Amazon Aurora Global Database as the correct choice since its storage-based physical replication is asynchronous but achieves sub-second lag, and managed regional failover can be executed within a few minutes.
Traditional backup and restore or simple cross-region read replicas may require manual promotion processes or snapshot restoration that violate the aggressive RPO/RTO bounds.
2
Ensure name resolution for the internal microservice endpoint works in both the primary and recovery regions.
Associate the Route 53 Private Hosted Zone created in the primary VPC with the secondary VPC in the recovery region.
VPC peering does not automatically share Private Hosted Zone name resolution; the hosted zone must be explicitly associated with each VPC that needs to resolve the private DNS names.
3
Design highly available internet egress routing for outbound telemetry alerts in each region.
Deploy a NAT Gateway in each Availability Zone in both regions.
Using a single NAT Gateway per VPC creates a single point of failure. If the Availability Zone containing the NAT Gateway experiences an outage, all resources in other Availability Zones lose outbound internet access.
4
Establish global routing that automatically detects failures and redirects traffic.
Configure Route 53 Active-Passive Failover routing using health checks linked to the Application Load Balancer in each region.
Active-Passive failover routes all traffic to the primary region under normal operations and dynamically switches traffic to the recovery region when the primary endpoint health check fails.

Key Concept

Designing multi-region disaster recovery for low RTO/RPO, ensuring cross-VPC DNS resolution, and maintaining multi-AZ redundant egress paths.
Rate this question

Topics

High Availability and Disaster Recovery Design

More questions from this topic