Which Elastic Beanstalk deployment policy minimizes impact and maximizes availability?

The correct answer is B. Use an immutable deployment policy to deploy to Amazon EC2 instances.

An immutable deployment in AWS Elastic Beanstalk launches a completely new set of instances with the updated application version in a separate Auto Scaling group. Once the new instances pass health checks, the load balancer is updated to route traffic to them, and the old instances are terminated. This approach ensures zero downtime during the deployment process, which directly satisfies the requirement to maximize availability. Additionally, because the old instances continue to serve traffic until the new ones are fully validated, there is minimal performance impact on the running application.

Let's examine why the other options are incorrect:

• Option A (Rolling deployment): This policy updates instances in batches, meaning some instances run the old version while others run the new version simultaneously. This can lead to downtime or degraded performance during the transition, failing to maximize availability.

• Option C (All-at-once deployment): This policy updates all instances simultaneously, which results in complete downtime during the deployment. This is the opposite of maximizing availability.

• Option D (Canary deployment): While canary deployments are excellent for testing new features on a limited scale and reducing risk, they do not minimize the overall deployment duration. The gradual rollout process takes longer, which conflicts with the requirement to deploy new features as quickly as possible. As noted by community members, canary deployments prioritize risk mitigation over speed and availability.