A company has an application that runs on five Amazon EC2 instances. The instances run behind an Application Load Balancer (ALB). The application processes messages from an Amazon Simple Queue…

Question

Options

• ACreate an EC2 Auto Scaling group. Add the existing EC2 instances to the Auto Scaling group.
• BDouble the capacity of the SQS queue. Add another five EC2 instances to handle the increase in
• CDouble the capacity of the SQS queue. Double the memory of the EC2 instances to handle the
• DCreate an EC2 Auto Scaling group. Add the existing EC2 instances to the Auto Scaling group.

• A
  13% (7)
• B
  4% (2)
• C
  6% (3)
• D
  77% (41)

Explanation

Option D is correct because Auto Scaling dynamically adds EC2 instances during peak load and removes them when demand drops - you only pay for capacity when it's actually needed, making it the most cost-effective approach. The key differentiator between A and D (both appear truncated here) is that D almost certainly includes configuring a scaling policy tied to the SQS ApproximateNumberOfMessagesVisible metric, which directly reflects queue backlog and is the appropriate trigger for this workload.

Options B and C are wrong for two reasons: SQS is a managed, fully elastic service with no "capacity" you need to double - it handles virtually unlimited messages automatically - and permanently doubling EC2 instances or memory means paying for idle capacity 24/7, which is wasteful during off-peak hours.

Memory tip: When you see SQS + EC2 + variable load, always think Auto Scaling on queue depth. The formula is: fluctuating demand = Auto Scaling (pay as you go), never pre-provision double capacity (pay always). If SQS is the workload driver, the scaling metric should be SQS-based, not CPU.

No community discussion yet for this question.