Skip to content
CloudBudgie logo

EC2 rightsizing

Size EC2 instances to the work they actually do.

CloudBudgie identifies oversized and underutilized EC2 instances, estimates your potential savings, and helps you make the right changes—directly in AWS or through infrastructure-as-code pull requests.

See an example

No credit card required

One smaller instance. An estimated $140/month saved.

An oversized general-purpose instance, the evidence behind the recommendation, and what the change is estimated to save.

  • Evidence: p95 CPU stayed at 31% of 8 vCPUs. Memory is shown when memory metrics are available for the instance.
  • Headroom: On half the vCPUs, the same load would put p95 CPU near 62%: less headroom than before, but still room for peaks.
  • Confidence: High confidence means recent usage strongly supports the change. It is not a promise of unchanged performance, so watch utilization after it ships.

Illustrative example, not customer data. Savings use us-east-1 Linux On-Demand rates. Actual savings vary with your pricing, Savings Plans or Reserved Instance commitments, usage, and when the change is deployed.

EC2 instance

reports-worker-2

m5.2xlarge → m5.xlarge

High confidence
CPU utilization
9% avg / 31% p95
Memory utilization
22% avg / 38% p95
vCPU
8 → 4 (-50%)
Memory
32GB → 16GB (-50%)

Estimated savings

$140/month

Find wasted EC2 spending

CloudBudgie looks for three kinds of EC2 waste and shows the estimated monthly savings for each.

  • Oversized instances

    Utilization stays well below what the instance type provides. Move to a smaller type and stop paying for unused capacity.

    Apply NowIaC pull request

  • Low-utilization instances

    CPU and network activity stay low for days at a time. Stop the instance, and start it again whenever you need it.

    Apply Now

  • Stopped instances

    Stopped instances still pay for attached EBS volumes and Elastic IPs. Terminate the ones you no longer need. Termination is permanent.

    Apply NowIaC pull request

From recommendation to fix

Choose how each finding becomes a change.

  1. Recommendation only

    Review the finding and its estimated savings. Nothing in AWS changes.

  2. Apply Now

    CloudBudgie makes the change directly in AWS after you confirm it.

  3. Apply with AI

    BudgieBot edits your Terraform or CloudFormation and opens a pull request for your team to review and deploy.

Available actions vary by optimization. Stopping a low-utilization instance is applied directly, not through a pull request, because removing an instance from infrastructure code terminates it.

How IaC pull requests work

Rightsize with confidence

Savings without guesswork.

CloudBudgie shows utilization, available performance headroom, and confidence levels so you can evaluate each recommendation before making a change. Some changes may require a maintenance window.

Recommendations reflect recent usage, so consider peaks it may not include, such as month-end jobs, before you downsize.

Find out how much you’re overspending on EC2.

Connect your AWS account for free, discover rightsizing opportunities, and choose which savings to put into action.

See pricing

No credit card required