AWS Lab: Identifying and Implementing Cost Optimization Opportunities

This lab focuses on the practical application of Content Domain 3.5 of the AWS Certified Solutions Architect - Professional (SAP-C02) exam. You will learn to identify underutilized resources and apply optimization strategies including rightsizing, storage tiering, and cost-allocation tagging.

[!WARNING] This lab involves modifying AWS resources. Ensure you follow the Teardown section at the end to avoid unexpected charges.

Prerequisites

An active AWS Account.
AWS CLI installed and configured with AdministratorAccess or equivalent permissions.
Basic familiarity with EC2 and S3 services.
Region: We will use us-east-1 (N. Virginia) for this lab.

Learning Objectives

Implement Resource Tagging: Establish a cost-allocation strategy using tags.
Identify & Rightsize: Simulate the identification of over-provisioned EC2 instances and modify them for cost efficiency.
Automate Storage Tiering: Configure S3 Lifecycle policies to move data to lower-cost storage tiers.
Analyze Trends: Use AWS Cost Explorer to visualize expenditure and usage patterns.

Architecture Overview

This lab demonstrates the flow from resource monitoring to active optimization.

Loading Diagram...

Step-by-Step Instructions

Step 1: Establish Cost Allocation Tags

Tags are the foundation of cost visibility. We will tag an existing or new resource to track its cost center.

CLI Interface:

bash

# Create a dummy EC2 instance for the lab if you don't have one
# Replace <SUBNET_ID> with a valid subnet ID from your VPC
aws ec2 run-instances \
    --image-id ami-0c101f26f147fa7fd \
    --count 1 \
    --instance-type t3.micro \
    --tag-specifications 'ResourceType=instance,Tags=[{Key=CostCenter,Value=Research},{Key=Project,Value=BrainyBee}]' \
    --region us-east-1

▶Console alternative

Open the EC2 Console.
Select Instances > Launch Instances.
Scroll to Advanced Details > Tags.
Add Tag: Key=CostCenter, Value=Research.
Launch the instance.

Step 2: Simulate Rightsizing an EC2 Instance

In a real-world scenario, you would use AWS Compute Optimizer. Here, we will manually "rightsize" an instance from t3.medium (simulated) down to t3.nano to reduce hourly costs.

CLI Interface:

bash

# 1. Stop the instance (replace <INSTANCE_ID> with your actual ID)
aws ec2 stop-instances --instance-ids <INSTANCE_ID>

# 2. Change instance type to a smaller size
aws ec2 modify-instance-attribute \
    --instance-id <INSTANCE_ID> \
    --instance-type "{\"Value\": \"t3.nano\"}"

# 3. Start the instance again
aws ec2 start-instances --instance-ids <INSTANCE_ID>

[!TIP] Moving from x86 (t3) to ARM-based Graviton (t4g) often provides 40% better price/performance.

Step 3: Implement S3 Storage Tiering

To optimize storage costs for infrequently accessed data, we will apply a lifecycle policy.

CLI Interface: Create a file named lifecycle.json:

json

{
    "Rules": [
        {
            "ID": "MoveToIAAfter30Days",
            "Status": "Enabled",
            "Filter": {"Prefix": "logs/"},
            "Transitions": [
                {
                    "Days": 30,
                    "StorageClass": "STANDARD_IA"
                }
            ]
        }
1    ]
}

Apply the policy:

bash

aws s3api put-bucket-lifecycle-configuration \
    --bucket <YOUR_UNIQUE_BUCKET_NAME> \
    --lifecycle-configuration file://lifecycle.json

Checkpoints

Tagging: Run aws ec2 describe-tags --filters "Name=resource-id,Values=<INSTANCE_ID>". Verify the CostCenter tag is present.
Rightsizing: Check the EC2 console; the instance type should now be t3.nano.
S3 Policy: Run aws s3api get-bucket-lifecycle-configuration --bucket <YOUR_BUCKET_NAME>. It should return your JSON configuration.

Analysis: Cost vs. Performance

Understanding the "Sweet Spot" in rightsizing is critical for the SAP-C02 exam. The following graph illustrates the goal of rightsizing: finding the minimum cost that meets the performance threshold.

\begin{tikzpicture}[scale=0.8] % Axes \draw[->] (0,0) -- (6,0) node[right] {Performance (CPU/RAM)}; \draw[->] (0,0) -- (0,5) node[above] {\mbox{Cost ($)}};

code

% Cost Curve
\draw[thick, blue] (0.5,0.5) .. controls (2,1) and (4,2.5) .. (5.5,4.5);
\node[blue] at (5.8,4) {\mbox{Cost}};

% Minimum Performance Requirement
\draw[dashed, red] (3,0) -- (3,5) node[above] {\mbox{Threshold}};

% Optimal Point
\filldraw[black] (3,1.8) circle (2pt);
\draw[<-, thick] (3.1,1.7) -- (4.5,1) node[right] {\mbox{Optimal Rightsize Point}};

\end{tikzpicture}

Troubleshooting

Problem	Potential Cause	Fix
`InsufficientInstanceCapacity`	The `t3.nano` is unavailable in the AZ.	Try a different AZ or use `t3.micro`.
`AccessDenied`	IAM user lacks `ce:*` or `ec2:ModifyInstanceAttribute` permissions.	Attach the `AdministratorAccess` policy for the lab.
S3 Bucket name error	Bucket names must be globally unique.	Add a random suffix like `brainybee-lab-cost-12345`.

Stretch Challenge

Automated Decommissioning: Create an AWS Lambda function that identifies instances with the tag Status: ToBeDecommissioned and stops them automatically at 6 PM daily using Amazon EventBridge.

Cost Estimate

EC2 (t3.nano): ~$0.0052/hour. If run for 30 mins: $0.0026.
S3 Standard Storage: First 50TB is $0.023/GB. For lab amounts: $0.00.
AWS Cost Explorer: Free for the basic interface.
Total Estimated Spend: <$0.01 USD.

Concept Review

Strategy	Best For...	Key Constraint
Spot Instances	Fault-tolerant workloads (e.g., Batch processing)	Can be interrupted with 2-min notice.
Savings Plans	Consistent compute usage over 1-3 years	Commitment to a $/hour spend.
Rightsizing	Over-provisioned legacy migrations	Requires downtime to change instance types.
S3 Glacier	Long-term archival (Compliance logs)	Retrieval times range from minutes to hours.

Clean-Up / Teardown

To avoid ongoing charges, run the following commands immediately:

bash

# 1. Terminate the EC2 instance
aws ec2 terminate-instances --instance-ids <INSTANCE_ID>

# 2. Delete the S3 bucket (empty it first)
aws s3 rb s3://<YOUR_UNIQUE_BUCKET_NAME> --force

# 3. Deactivate any Billing Alarms created during exploration
aws cloudwatch delete-alarms --alarm-names "BudgetAlarm"