AWS Certified DevOps Engineer – Professional — Question 1

A DevOps Engineer administers an application that manages video files for a video production company. The application runs on Amazon EC2 instances behind an ELB Application Load Balancer. The instances run in an Auto Scaling group across multiple Availability Zones. Data is stored in an Amazon RDS PostgreSQL
Multi-AZ DB instance, and the video files are stored in an Amazon S3 bucket. On a typical day, 50 GB of new video are added to the S3 bucket. The Engineer must implement a multi-region disaster recovery plan with the least data loss and the lowest recovery times. The current application infrastructure is already described using AWS CloudFormation.
Which deployment option should the Engineer choose to meet the uptime and recovery objectives for the system?

Answer options

• A. Launch the application from the CloudFormation template in the second region, which sets the capacity of the Auto Scaling group to 1. Create an Amazon RDS read replica in the second region. In the second region, enable cross-region replication between the original S3 bucket and a new S3 bucket. To fail over, promote the read replica as master. Update the CloudFormation stack and increase the capacity of the Auto Scaling group.
• B. Launch the application from the CloudFormation template in the second region, which sets the capacity of the Auto Scaling group to 1. Create a scheduled task to take daily Amazon RDS cross-region snapshots to the second region. In the second region, enable cross-region replication between the original S3 bucket and Amazon Glacier. In a disaster, launch a new application stack in the second region and restore the database from the most recent snapshot.
• C. Launch the application from the CloudFormation template in the second region, which sets the capacity of the Auto Scaling group to 1. Use Amazon CloudWatch Events to schedule a nightly task to take a snapshot of the database, copy the snapshot to the second region, and replace the DB instance in the second region from the snapshot. In the second region, enable cross-region replication between the original S3 bucket and a new S3 bucket. To fail over, increase the capacity of the Auto Scaling group.
• D. Use Amazon CloudWatch Events to schedule a nightly task to take a snapshot of the database and copy the snapshot to the second region. Create an AWS Lambda function that copies each object to a new S3 bucket in the second region in response to S3 event notifications. In the second region, launch the application from the CloudFormation template and restore the database from the most recent snapshot.

Correct answer: A

Explanation

Option A is correct as it ensures minimal data loss and allows for quick recovery by creating a read replica and enabling cross-region S3 replication, which is efficient for failover. Options B and C involve snapshots and Glacier, which can lead to longer recovery times due to the need to restore from snapshots or Glacier. Option D also relies on snapshots and adds complexity with Lambda, making it less optimal than A.