AI-Powered Waste Detection Finds $4,200 in Unused AWS Resources

The $4,200 Mistake

An engineer discovered $4,200 in monthly AWS costs stemming from a forgotten staging environment that had zero traffic and users, despite remaining active for six months. This highlights a common issue: resources left running unnecessarily.

Why This Matters

AWS Cost Explorer provides visibility into spending, but lacks proactive identification of waste. The ideal scenario involves automated detection and remediation of unused resources; however, the reality is significant financial loss due to orphaned instances, idle databases, and over-provisioned services. Unaddressed cloud waste can easily reach tens of thousands of dollars per year for even modestly sized deployments.

Key Insights

• $4,200/month: Cost of the discovered AWS waste in the author’s account.
• 70% of waste: Percentage attributed to orphaned or idle resources, according to the author’s analysis.
• OpenAI GPT-4: Used for generating cost-saving recommendations within the CloudWise tool.

Working Example

# Example of identifying orphaned EBS volumes using boto3
import boto3

ec2 = boto3.client('ec2')

response = ec2.describe_volumes()

for volume in response['Volumes']:
    if not volume.get('Attachments'):
        print(f"Orphaned EBS Volume: {volume['VolumeId']} - Cost: (Estimate based on volume type and size)")

Practical Applications

• CloudWise (Startup): Automates AWS cost optimization by identifying and recommending removal of wasted resources.
• Pitfall: Relying solely on manual cost reviews leads to consistent overspending due to the difficulty of tracking all running resources and their utilization.

References:

• https://dev.to/cloudwiseteam/i-built-an-ai-powered-aws-waste-detector-and-found-4200-in-my-own-account-2cc4