Cost Management

First PublishedFeb 16, 2026ByAtif Alam

Cloud costs can grow quickly if unmanaged. AWS provides tools for understanding, forecasting, and controlling costs. The most important practice is tagging — without tags, you can’t attribute costs to teams, projects, or environments.

AWS Billing Dashboard

The billing dashboard (Billing and Cost Management in the console) provides:

• Month-to-date spend — Total and by-service breakdown.
• Forecasted month-end cost — Projection based on current usage.
• Free tier usage — Track how close you are to free tier limits.
• Bills — Detailed monthly invoices.

First thing to do on a new account: Set up a billing alarm so you get notified before costs surprise you.

# Create a billing alarm (CloudWatch, us-east-1 only)
aws cloudwatch put-metric-alarm \
  --alarm-name "MonthlyBillingAlarm" \
  --metric-name EstimatedCharges \
  --namespace AWS/Billing \
  --statistic Maximum \
  --period 21600 \
  --threshold 100 \
  --comparison-operator GreaterThanThreshold \
  --evaluation-periods 1 \
  --alarm-actions arn:aws:sns:us-east-1:123456789012:billing-alerts \
  --dimensions Name=Currency,Value=USD \
  --region us-east-1

Cost Explorer

Cost Explorer is the primary tool for analyzing AWS costs. It visualizes spend over time and lets you filter, group, and forecast.

Key Views

View	What It Shows
Monthly costs by service	Which services cost the most
Daily costs	Spot spending spikes
Cost by account	In multi-account setups (AWS Organizations)
Cost by tag	Costs per team, project, environment
Forecast	Projected spend for the next 3–12 months
Reservation utilization	Are your Reserved Instances being used?

Common Filters

• Service — EC2, RDS, S3, Lambda, NAT Gateway, etc.
• Region — Identify costs by region.
• Tag — Filter by Environment=production, Team=platform, etc.
• Instance type — See cost per EC2 instance type.
• Usage type — Distinguish data transfer, storage, compute hours.

Enabling Cost Explorer

# Enable (must be done from the management account)
aws ce get-cost-and-usage \
  --time-period Start=2026-02-01,End=2026-02-16 \
  --granularity DAILY \
  --metrics BlendedCost \
  --group-by Type=DIMENSION,Key=SERVICE

AWS Budgets

Budgets let you set spending limits and receive alerts when you approach or exceed them.

Budget Types

Type	What It Tracks
Cost budget	Total spend (e.g. “Don’t exceed $5,000/month”)
Usage budget	Resource usage (e.g. “EC2 hours < 10,000”)
Reservation budget	RI/Savings Plan utilization and coverage
Savings Plans budget	Savings Plan utilization

Creating a Budget

aws budgets create-budget --account-id 123456789012 --budget '{
  "BudgetName": "MonthlyTotal",
  "BudgetLimit": {"Amount": "5000", "Unit": "USD"},
  "TimeUnit": "MONTHLY",
  "BudgetType": "COST"
}' --notifications-with-subscribers '[{
  "Notification": {
    "NotificationType": "ACTUAL",
    "ComparisonOperator": "GREATER_THAN",
    "Threshold": 80,
    "ThresholdType": "PERCENTAGE"
  },
  "Subscribers": [{
    "SubscriptionType": "EMAIL",
    "Address": "ops@example.com"
  }]
}]'

This sends an email when actual spend exceeds 80% of the $5,000 budget.

Budget Actions

Budgets can trigger automated actions when thresholds are hit:

Action	What It Does
Apply IAM policy	Deny certain actions (e.g. block launching new EC2 instances)
Apply SCP	Restrict an entire account via AWS Organizations
Notify via SNS	Send to Slack, PagerDuty, custom Lambda

Tagging Strategy

Tags are the foundation of cost management. Without tags, you can’t answer “which team caused this bill?”

Required Tags

Define a minimum set of tags that every resource must have:

Tag Key	Purpose	Example Values
Environment	Which environment	production, staging, development
Team	Owning team	platform, backend, data
Project	Project or product	payments, search, website
CostCenter	Budget category	engineering, marketing, R&D
ManagedBy	How it was created	terraform, manual, cloudformation

Enforcing Tags

Method	How It Works
Tag policies (AWS Organizations)	Define required tags and allowed values across accounts
SCP	Deny resource creation without required tags
Config rules	Detect non-compliant resources after creation
Terraform	Enforce tags via default_tags in the provider block
CI/CD checks	Lint Terraform/CloudFormation for tag compliance before apply

Terraform default tags:

provider "aws" {
  region = "us-east-1"

  default_tags {
    tags = {
      Environment = "production"
      Team        = "platform"
      ManagedBy   = "terraform"
    }
  }
}

Every resource created by this provider automatically gets these tags.

Cost Allocation Tags

After defining tags, activate them as cost allocation tags in the Billing console. This makes them available in Cost Explorer and billing reports.

# Activate cost allocation tags
aws ce create-cost-category-definition --name "Team" \
  --rules '[{"Value":"platform","Rule":{"Tags":{"Key":"Team","Values":["platform"]}}}]' \
  --rule-version 1

Common Cost Traps

Trap	Why It’s Expensive	How to Fix
Unattached EBS volumes	Pay for volumes not attached to any instance	Audit and delete unused volumes
Old snapshots	EBS snapshots accumulate over time	Lifecycle policies, retention limits
NAT Gateway data processing	$0.045/GB adds up with high traffic	Use VPC endpoints for S3/DynamoDB (free)
Idle EC2 instances	Running 24/7 but only used during business hours	Auto-stop schedules, right-size
Oversized instances	Running m5.4xlarge when m5.large is enough	Right-size with CloudWatch metrics
Unused Elastic IPs	Charged when not attached to a running instance	Release unused EIPs
Data transfer between AZs	$0.01/GB per direction	Minimize cross-AZ traffic where possible
CloudWatch Logs retention	Default is “never expire” — logs grow forever	Set retention periods (7, 30, 90 days)
Forgotten dev/test resources	Sandbox environments left running	Tagging + auto-shutdown schedules

Cost Optimization Strategies

Right-Sizing

Match instance types to actual usage:

CloudWatch CPU/Memory metrics → consistently below 20%?
  → Downsize the instance type
  → Or switch to a smaller class

AWS Compute Optimizer analyzes your usage and recommends right-size changes.

Reserved Instances and Savings Plans

Option	Commitment	Flexibility	Discount
Reserved Instances	1 or 3 years, specific instance type	Low (locked to instance family/region)	Up to 72%
Compute Savings Plans	1 or 3 years, dollar amount	High (any instance type, region, OS)	Up to 66%
EC2 Instance Savings Plans	1 or 3 years, specific instance family	Medium (locked to family + region)	Up to 72%

Rule of thumb: If an instance has been running steadily for 3+ months, it’s a candidate for a Savings Plan.

Spot Instances

Use Spot for fault-tolerant workloads at up to 90% discount:

• CI/CD build runners
• Data processing / batch jobs
• Dev/test environments
• Kubernetes worker nodes (with mixed instance groups)

Other Optimizations

Strategy	Savings
S3 Intelligent-Tiering	Auto-moves objects to cheaper storage classes
VPC endpoints for S3/DynamoDB	Eliminates NAT Gateway data charges
Lambda right-sizing	More memory = faster = same or lower cost
Auto Scaling to zero	Scale dev/test to 0 at night and weekends
Graviton instances	ARM-based (t4g, m6g) — ~20% cheaper than x86

AWS Organizations and Multi-Account Strategy

For larger teams, use AWS Organizations to manage multiple accounts:

Management Account (billing)
  ├── Production OU
  │     ├── prod-app (account)
  │     └── prod-data (account)
  ├── Development OU
  │     ├── dev-app (account)
  │     └── sandbox (account)
  └── Security OU
        └── audit (account)

Benefits:

• Consolidated billing — One bill, volume discounts apply across all accounts.
• SCPs — Restrict what accounts/OUs can do (e.g. block expensive instance types in dev).
• Blast radius — A compromised dev account can’t affect production.

Key Takeaways

• Set up billing alerts and budgets immediately — don’t wait for a surprise bill.
• Tag everything with at least Environment, Team, and Project. Activate tags as cost allocation tags.
• Use Cost Explorer to understand where money goes. Filter by service, tag, and region.
• Right-size instances based on actual CloudWatch metrics. Use Compute Optimizer for recommendations.
• Commit to Savings Plans for steady workloads (1- or 3-year). Use Spot for fault-tolerant batch work.
• Eliminate waste: delete unused EBS volumes, set log retention, auto-stop dev environments, use VPC endpoints.
• Use AWS Organizations for multi-account billing consolidation and SCPs for guardrails.