Amazon Web Services

Stop paying for AWS resources you don't use

CostSave runs 90+ checks across EC2, RDS, S3, Lambda, EKS, ECS, DynamoDB, CloudWatch, and more. Connect once with a read-only CloudFormation role — no agents, no CLI.

One-click CloudFormation setup·90+ checks·Results in ~5 min·Free forever
Scan my AWS account free
90+
Cost checks
18
AWS services covered
<5 min
Scan time
$0
Cost, forever
What we scan

Every AWS check CostSave runs

Real-time scans across your live AWS account. Every finding includes the exact monthly dollar amount — sorted by impact.

AWS
AWS — 90+ cost checks across all major services
Connect via a one-click read-only CloudFormation role. CostSave scans your live billing data and every active region automatically.
EC2
EC2 & EBS
  • Idle instances — max CPU <10% over 30 days
  • Unattached EBS volumes (billed even when unused)
  • io1/io2 volumes that can migrate to gp3 (same IOPS, lower cost)
  • Unused Elastic IPs not associated with any resource
  • Old EBS snapshots older than 90 days
  • Multiple NAT Gateways — consolidation opportunity
  • Compute Optimizer right-sizing for EC2 & EBS
RDS
RDS Databases
  • Idle RDS instances — 0 connections in 30 days
  • Small instances (t3/t4) with Multi-AZ in dev environments
  • Old manual RDS snapshots older than 30 days
  • Compute Optimizer right-sizing for RDS
S3
S3 Storage
  • Versioning enabled but no lifecycle rule
  • Buckets with no lifecycle policy and non-zero storage
  • Missed tiering to S3-IA or Glacier
ELB
Load Balancers
  • Idle ALBs/NLBs — no targets or 0 requests in 30 days
  • Low-traffic load balancers — under 1,000 requests/month
Lambda
Lambda
  • Unused functions — 0 invocations in 30 days
  • High-memory (≥1 GB) with <100 invocations/month
  • Compute Optimizer memory right-sizing
ECS
ECS
  • Idle clusters — 0 running tasks, 0 active services
  • Services with desired count = 0 (scaled down, forgotten)
  • Unhealthy services — running < desired count
EKS
EKS
  • Empty clusters — 0 worker nodes (~$73/mo control plane)
  • Node groups with desired size = 0
  • On-Demand node groups (Spot saves 60–70%)
  • Oversized node groups — maxSize ≥ 3× desiredSize
  • Clusters running unsupported Kubernetes version
SQS
SQS Queues
  • Idle queues — 0 messages in 7 days
  • Short-polling queues (increases API calls & cost)
  • Active queues with no Dead Letter Queue
  • High empty-receive ratio — ≥80% return nothing
DynamoDB
DynamoDB
  • Idle tables — 0 reads and 0 writes in 30 days
  • Over-provisioned — using <20% of provisioned capacity
  • On-demand tables with 0 traffic
CloudFront
CloudFront & VPC
  • Unused distributions — 0 requests in 30 days
  • Distributions not enforcing HTTPS
  • Unused VPC Endpoints (no attached route tables)
  • Idle VPN Connections & unused ENIs
CloudWatch
CloudWatch & EFS
  • Log groups with no retention policy
  • Large log groups — >1 GB stored
  • Idle EFS file systems — 0 connections in 14 days
  • EFS without lifecycle policy
ECR
ECR (Container Registry)
  • Idle repos — no image pushes in 90+ days
  • Repos without a lifecycle policy
  • Repos with untagged images (orphaned build artifacts)
  • Images older than 6 months
KMS
KMS & Secrets Manager
  • Unused KMS keys — no crypto ops in 30 days
  • KMS keys without rotation enabled
  • Stale secrets — not accessed in 90 days
  • Secrets without automatic rotation
API Gateway
API Gateway & Route 53
  • Idle REST APIs — 0 requests in 30 days
  • REST APIs with no stages deployed
  • Empty hosted zones (only SOA + NS records)
  • Unused Route 53 health checks
WAF
WAF & GuardDuty
  • Idle WebACLs — 0 requests in 30 days
  • Unattached WebACLs (not associated with any resource)
  • GuardDuty detectors with 0 findings in 30 days
CloudTrail
CloudTrail & ACM
  • Stopped/inactive CloudTrail trails
  • Trails without log file validation
  • Expired or soon-expiring SSL certificates
  • Unused certificates not attached to any resource
CodePipeline
CodeBuild, Pipeline & Amplify
  • Idle CodeBuild projects — 0 builds in 30 days
  • Inactive pipelines — no executions in 30 days
  • V1 pipelines (legacy type, not V2)
  • Idle Amplify apps — no deployments in 90 days
Billing
Billing & Savings Plans
  • Savings Plans coverage & utilization %
  • AWS-recommended Compute SP commitment
  • Cost anomaly detection — last 90 days
  • Billing trend spike — ≥20% MoM increase
  • 6-month spend trend and forecast
Read-only access — nothing ever written

Find out what your AWS account
is wasting — in under 5 minutes

Connect with a one-click CloudFormation role. No credit card, no agents, free forever.

Start free AWS scan →