Examkin

Question

Difficulty: HardOptimizing Resource Costs and Sizing

An enterprise runs a high-throughput, memory-intensive backend application across multiple member accounts in AWS Organizations. The primary production compute environment in the main application account consists of:
- An Auto Scaling Group (ASG) of Amazon EC2 instances launched across 33 Availability Zones using `m6g.2xlarge` instances (88 vCPUs, 3232 GiB RAM). Performance metrics show an average CPU utilization of 25%25\% (peaking at 40%40\%) and an average memory utilization of 75%75\% (peaking at 85%85\%).
- Multiple microservices running on AWS Fargate tasks.
- An API routing layer using AWS Lambda.

All software is compiled specifically for the ARM64 architecture, and the application must maintain high availability across all 33 Availability Zones with zero downtime. The company wants to optimize compute costs. Which of the following is the most cost-effective and architecturally sound strategy?

  1. Modify the EC2 launch template to use `r6g.xlarge` instances (44 vCPUs, 3232 GiB RAM) to match the memory requirements without risking out-of-memory errors, and purchase a Compute Savings Plan in the AWS Organizations management account.Answer
  2. B
    Modify the EC2 launch template to use `r6g.xlarge` instances (44 vCPUs, 3232 GiB RAM) to match the memory requirements. Purchase an EC2 Instance Savings Plan for the `r6g` family in the AWS Organizations management account to cover the EC2 instances, while leaving Fargate and Lambda on On-Demand pricing.
  3. C
    Modify the EC2 launch template to use `m6g.xlarge` instances (44 vCPUs, 1616 GiB RAM) to reduce the compute cost. Use AWS Resource Access Manager (RAM) to share the unused instance capacity and savings benefits with external accounts outside the Organization.
  4. D
    Modify the EC2 launch template to use `r6g.xlarge` instances (44 vCPUs, 3232 GiB RAM). Consolidate outbound traffic by routing all traffic through a single NAT Gateway in one Availability Zone to eliminate redundant NAT Gateway hourly charges, and purchase a Compute Savings Plan in the management account.

Answer

Modify the EC2 launch template to use `r6g.xlarge` instances (44 vCPUs, 3232 GiB RAM) to match the memory requirements without risking out-of-memory errors, and purchase a Compute Savings Plan in the AWS Organizations management account.
The correct answer optimizes compute costs by switching to the memory-optimized `r6g.xlarge` instance type, which matches the resource utilization profile (3.23.2 vCPUs peak, 27.227.2 GiB RAM peak) without causing out-of-memory errors. The Compute Savings Plan is the correct vehicle because it dynamically applies discounts to EC2, Fargate, and Lambda across the entire Organization structure.

Step-by-Step Solution

1
Analyze EC2 resource utilization metrics to identify constraints.
Peak CPU demand is 40%40\% of 88 vCPUs = 3.23.2 vCPUs. Peak memory demand is 85%85\% of 3232 GiB = 27.227.2 GiB.
Determines the minimum vCPU and memory thresholds required to avoid application performance degradation or out-of-memory failures.
2
Evaluate target instance types for right-sizing.
`r6g.xlarge` provides 44 vCPUs (covering the 3.23.2 vCPUs peak) and 3232 GiB RAM (covering the 27.227.2 GiB peak), whereas `m6g.xlarge` only provides 1616 GiB RAM and would crash.
Allows downsizing compute capacity to match actual workload demands while respecting memory constraints.
3
Select the appropriate Savings Plan type for the multi-account heterogeneous workload.
A Compute Savings Plan covers EC2, Fargate, and Lambda across all accounts in the AWS Organization when purchased from the management account.
Ensures that all components of the containerized, serverless, and virtual machine architecture receive discount coverage, unlike EC2 Instance Savings Plans.

Key Concept

Memory-constrained right-sizing and Compute Savings Plans application in multi-account environments
Estimated Time:3m 0s
Rate this question

Topics

Optimizing Resource Costs and Sizing

More questions from this topic