Maximizing AWS Cost Savings

In today's cloud-driven landscape, Amazon Web Services (AWS) has become the go-to platform for businesses worldwide. While AWS offers unparalleled flexibility and scalability for your cloud infrastructure, it's crucial to ensure that you're not overspending on underutilized or idle resources.

One effective strategy to tackle this challenge is by employing custom scripts and tools tailored to your organization's needs. In this blog post, we'll explore three scenarios where custom scripts can make a significant difference in optimizing AWS costs

Scenario 1: Identifying Idle EC2 Instances

The Cost of Idle EC2 Instances

EC2 instances, the virtual machines of AWS, are at the core of many workloads. However, it's common for development and testing instances to run continuously, even when they're needed only during working hours. These idle instances can quickly accumulate costs that could be better allocated elsewhere.

The Solution: Custom Automation

To regain control over your AWS spending, we can employ custom scripts to automate the identification and termination of idle EC2 instances. Here's how it works:

1. Resource Identification

We begin by identifying instances that have been idle for a specified period. For example, we might consider instances that have had no CPU activity or network traffic during non-working hours.

2. Script Development

A custom script is developed using AWS SDKs or APIs to query and filter instances based on our criteria. This script automates the identification process.

3. Termination

Once idle instances are identified, the script initiates their termination. This step is automated, removing the need for manual intervention.

Example Script (Python)

import boto3
import datetime

# Define AWS credentials and region
aws_access_key = 'your_access_key'
aws_secret_key = 'your_secret_key'
aws_region = 'us-east-1'

# Initialize AWS EC2 client
ec2 = boto3.client('ec2', region_name=aws_region,
                   aws_access_key_id=aws_access_key,
                   aws_secret_access_key=aws_secret_key)

# Define criteria for identifying idle instances
current_time = datetime.datetime.now()
idle_threshold_minutes = 60  # Adjust the threshold as needed

# Get a list of all running EC2 instances
instances = ec2.describe_instances(Filters=[{'Name': 'instance-state-name', 'Values': ['running']}])

# Iterate through instances and terminate idle ones
for reservation in instances['Reservations']:
    for instance in reservation['Instances']:
        launch_time = instance['LaunchTime']
        running_time = current_time - launch_time

        if running_time.total_seconds() / 60 > idle_threshold_minutes:
            # Terminate the idle instance
            ec2.terminate_instances(InstanceIds=[instance['InstanceId']])
            print(f"Terminated idle EC2 instance: {instance['InstanceId']}")
        

Scenario 2: Identifying Unattached EBS Volumes

The Cost of Unattached EBS Volumes

Unattached Amazon Elastic Block Store (EBS) volumes can incur charges without providing any value. Identifying and removing these unattached volumes is crucial for cost optimization.

The Solution: Custom Automation

To address this challenge, we can use a custom script to identify and delete unattached EBS volumes:

Example Script (Python)

import boto3

# Define AWS credentials and region
aws_access_key = 'your_access_key'
aws_secret_key = 'your_secret_key'
aws_region = 'us-east-1'

# Initialize AWS EC2 client
ec2 = boto3.client('ec2', region_name=aws_region,
                   aws_access_key_id=aws_access_key,
                   aws_secret_access_key=aws_secret_key)

# Get a list of all EBS volumes
volumes = ec2.describe_volumes(Filters=[{'Name': 'status', 'Values': ['available']}])

# Iterate through volumes and delete unattached ones
for volume in volumes['Volumes']:
    volume_id = volume['VolumeId']
    ec2.delete_volume(VolumeId=volume_id)
    print(f"Deleted unattached EBS volume: {volume_id}")        

Scenario 3: Right-Sizing RDS Instances

The Cost of Unutilized RDS Instances

Relational Database Service (RDS) instances come in various sizes, and choosing the right one is essential for cost efficiency. Unutilized large instances can be downsized to save costs.

The Solution: Custom Automation

To optimize your RDS instances, a custom script can be used to identify instances with low CPU utilization and modify them to a smaller instance type:

Example Script (Python)

import boto3

# Define AWS credentials and region
aws_access_key = 'your_access_key'
aws_secret_key = 'your_secret_key'
aws_region = 'us-east-1'

# Initialize AWS RDS client
rds = boto3.client('rds', region_name=aws_region,
                   aws_access_key_id=aws_access_key,
                   aws_secret_access_key=aws_secret_key)

# Define CPU utilization threshold for identifying unutilized instances
cpu_utilization_threshold = 5  # Adjust the threshold as needed

# Get a list of all RDS instances
instances = rds.describe_db_instances()

# Iterate through instances and modify unutilized ones
for instance in instances['DBInstances']:
    instance_id = instance['DBInstanceIdentifier']
    cpu_utilization = instance['CPUUtilization']

    if cpu_utilization <= cpu_utilization_threshold:
        # Modify the instance to a smaller type (e.g., db.t2.micro)
        rds.modify_db_instance(DBInstanceIdentifier=instance_id,
                               DBInstanceClass='db.t2.micro')
        print(f"Modified unutilized RDS instance: {instance_id}")        

Automating resource identification and optimization through custom scripts is a proactive approach to managing your AWS costs effectively. These scripts are just examples of how you can leverage automation to streamline your AWS costs.

By tailoring these solutions to your organization's unique needs, you can get the most value from your AWS investment while minimizing unnecessary expenses.