Question

Difficulty: MediumDeployment Strategies and Execution

A company hosts a web application on AWS Elastic Beanstalk. The application runs in a production environment with an Auto Scaling group of 8 Amazon EC2 instances. The company’s deployment policy requires that the application maintains full capacity to handle incoming traffic at all times during updates. To control costs, the company cannot allow a temporary doubling of the environment's compute resources, which rules out immutable deployments. Which deployment policy should the SysOps Administrator select to meet these requirements with the least financial impact?

  1. A
    All at once
  2. B
    Immutable
  3. Rolling with an additional batchAnswer
  4. D
    Rolling

Answer

Rolling with an additional batch
The rolling with an additional batch deployment policy launches a new batch of instances to maintain full capacity while the existing instances are updated in batches. This avoids taking any capacity offline and only increases resources by a single batch size, rather than doubling the entire environment.

Step-by-Step Solution

1
Analyze the capacity requirement during deployment.
The application must maintain 100% capacity (8 instances) during the deployment process.
To prevent performance degradation or downtime under production traffic.
2
Analyze the budget and resource constraints.
The system cannot temporarily double the running compute resources, ruling out immutable or full environment-copy deployments.
To control costs and avoid launching a duplicate Auto Scaling group of 8 instances.
3
Evaluate the Elastic Beanstalk deployment options against both constraints.
Rolling with an additional batch is selected because it launches only one extra batch of instances to offset capacity loss during the rolling update, keeping additional costs to a minimum.
It fulfills the capacity requirement without needing a parallel environment of equal size.

Key Concept

Selecting Elastic Beanstalk deployment strategies to balance application capacity and cost constraints.
Rate this question