Examkin

Soru

Zorluk: ZorOptimizing Resource Costs and Sizing

An enterprise manages a multi-account environment under AWS Organizations with consolidated billing enabled. In one member account, a critical production workload consists of:

* An Amazon ECS on AWS Fargate application running 20 tasks continuously across three Availability Zones (AZs). Each task is configured with 4 vCPU and 16 GB of memory. CloudWatch metrics over the past 90 days show that CPU utilization averages 15%15\%, while memory utilization peaks at 92%92\% during daily traffic spikes.
* A stateful processing layer consisting of 6 self-managed EC2 instances in an Auto Scaling group running on `m6i.2xlarge` instances (88 vCPU, 3232 GiB RAM) across two AZs. CloudWatch metrics show average CPU utilization at 12%12\% and memory utilization at 18%18\%, but the application experiences occasional, unpredictable network throughput bursts.
* Outbound internet traffic for both the ECS tasks and the EC2 instances is routed through a single NAT Gateway located in a public subnet in AZ-A to minimize running costs.

The company wants to optimize its infrastructure costs while maintaining high availability, ensuring performance during peak loads, and maximizing the discount from Savings Plans across the entire AWS Organization.

Which set of actions will achieve the maximum cost optimization while satisfying all availability and performance requirements?

  1. Downsize the Amazon ECS Fargate tasks to 2 vCPU and 16 GB of memory. Downsize the EC2 instances to `m6i.xlarge`. Deploy a NAT Gateway in each of the three Availability Zones and update the route tables. Purchase Compute Savings Plans at the AWS Organizations management account level with sharing enabled.Cevap
  2. B
    Downsize the Amazon ECS Fargate tasks to 2 vCPU and 16 GB of memory. Downsize the EC2 instances to `m6i.xlarge`. Deploy a NAT Gateway in each of the three Availability Zones and update the route tables. Purchase EC2 Instance Savings Plans for the `m6i` family at the AWS Organizations management account level with sharing enabled.
  3. C
    Downsize the Amazon ECS Fargate tasks to 2 vCPU and 16 GB of memory. Downsize the EC2 instances to `m6i.xlarge`. Retain the single NAT Gateway in AZ-A to avoid additional NAT Gateway hourly charges. Purchase Compute Savings Plans at the AWS Organizations management account level with sharing enabled.
  4. D
    Downsize the Amazon ECS Fargate tasks to 2 vCPU and 16 GB of memory. Share the EC2 instances across the other member accounts using AWS Resource Access Manager (RAM) to consolidate idle capacity. Deploy a NAT Gateway in each of the three Availability Zones and update the route tables. Purchase Compute Savings Plans at the AWS Organizations management account level with sharing enabled.

Cevap

Downsize the Amazon ECS Fargate tasks to 2 vCPU and 16 GB of memory, downsize the EC2 instances to `m6i.xlarge`, deploy a NAT Gateway in each of the three Availability Zones, and purchase Compute Savings Plans at the AWS Organizations management account level with sharing enabled.
The correct solution reduces CPU on ECS Fargate from 4 vCPU to 2 vCPU while retaining the 16 GB of memory, which matches the memory-bound workload requirements. It downsizes the EC2 instances to `m6i.xlarge` to cut compute costs while preserving sufficient RAM and network burst capability. To satisfy high availability, it deploys a NAT Gateway in each Availability Zone. Finally, purchasing Compute Savings Plans at the AWS Organizations management account level ensures that both the EC2 instances and ECS Fargate tasks across all member accounts receive the discount.

Adım Adım Çözüm

1
Analyze the Amazon ECS Fargate task metrics to determine right-sizing requirements.
The tasks are CPU-idle (15%15\% of 4 vCPU is 0.60.6 vCPU) but memory-constrained (peaking at 92%92\% of 16 GB is 14.7214.72 GB). The tasks can be downsized to 2 vCPU and 16 GB memory, which is a supported configuration, reducing CPU cost by 50% while fully accommodating memory peaks.
To optimize ECS costs without degrading performance during traffic bursts.
2
Analyze the EC2 instances' metrics and network performance requirements.
The current `m6i.2xlarge` instances (88 vCPU, 3232 GiB RAM) have low CPU and memory utilization (12%12\% CPU, 18%18\% memory). Downsizing to `m6i.xlarge` (44 vCPU, 1616 GiB RAM) leaves ample CPU and memory headroom (1616 GiB is greater than the 5.765.76 GiB needed) and preserves up to 12.5 Gbps network burst capability, keeping costs to a minimum.
To scale down the EC2 instances while retaining capacity for network throughput spikes.
3
Review the outbound network routing and high availability architecture.
A single NAT Gateway in AZ-A is a single point of failure. To maintain high availability across the three AZs, a NAT Gateway must be deployed in each AZ, and the private subnet route tables must be updated to route traffic locally.
To ensure that an outage in a single Availability Zone does not disrupt outbound traffic for resources in other zones.
4
Select the correct Savings Plan type and scope for a multi-account organization.
Compute Savings Plans apply to EC2, Fargate, and Lambda across all accounts under consolidated billing. Purchasing Compute Savings Plans at the management account level maximizes the coverage and utilization of the discount. EC2 Instance Savings Plans would exclude the Fargate workloads.
To optimize savings across both EC2 and ECS Fargate workloads within the AWS Organization.

Anahtar Kavram

Multi-dimensional resource right-sizing (compute vs. memory constraints), high availability network routing design, and the distinction between Compute and Instance Savings Plans in multi-account AWS Organizations.
Tahmini Süre:3m 0s
Bu soruyu puanla

Konular

Optimizing Resource Costs and Sizing

Bu konudan daha fazla soru