A company has a stateless application that runs on four Amazon EC2 instances. The application requires four instances at all times to support all traffic. A SysOps administrator must design a highly…

Question

Options

• ADeploy two Auto Scaling groups in two Availability Zones with a minimum capacity of two
• BDeploy an Auto Scaling group across two Availability Zones with a minimum capacity of four
• CDeploy an Auto Scaling group across three Availability Zones with a minimum capacity of four
• DDeploy an Auto Scaling group across three Availability Zones with a minimum capacity of six

• A
  17% (4)
• B
  4% (1)
• C
  70% (16)
• D
  9% (2)

Explanation

Option C is correct because spreading an Auto Scaling group across three AZs with a minimum of four ensures fault tolerance: if one AZ fails, the ASG detects the lost instances and redistributes/relaunches them across the two remaining AZs, always maintaining the required four.

Why the distractors fail:

• A uses two separate ASGs, each with a minimum of two - if one AZ goes down, that group loses its instances and you're left with only two total, breaking the requirement.
• B spans only two AZs with a minimum of four; losing one AZ means losing roughly half the instances, and with only one AZ remaining, you can't guarantee four healthy instances long enough until recovery.
• D works technically, but sets the minimum at six - you'd run six instances during normal operation when only four are required, making it over-provisioned and unnecessarily costly; the question asks for a design that meets requirements, not exceeds them.

Memory tip: Use the "N+1 AZ" rule - if you need N instances and can't lose any, deploy across at least (N/instances_per_AZ)+1 AZs and set the minimum to N, so losing one AZ leaves the ASG room to compensate in the survivors.

No community discussion yet for this question.