AWS Cost Optimization
OhChimp cuts your AWS bill: idle EC2, gp2 volumes, idle NAT gateways, weak Reserved Instance coverage. Read-only access, savings verified on the bill.
What OhChimp optimizes for AWS
Some of the cost levers OhChimp checks for AWS. Each one becomes a reviewable plan you approve before anything changes.
Reserved Instances and Savings Plans
OhChimp pulls your EC2 Reserved Instance recommendations from Cost Explorer across 1-year and 3-year terms with every payment option, plus Compute Savings Plan recommendations for both terms. It also reads your existing active commitments. The plan shows where committing to steady usage cuts the on-demand premium, with the estimated monthly saving and break-even months spelled out.
Idle and over-provisioned EC2
OhChimp reads Cost Explorer rightsizing recommendations to find instances running near-empty or sized larger than their real CPU use. Each one becomes a resize plan with the target instance type, or a termination plan for instances flagged idle, alongside the reported CPU utilization.
EBS volumes and gp2 to gp3
Unattached volumes keep billing after their instance is gone, and gp2 lists about 20% more than gp3 for the same baseline performance. OhChimp finds both. It writes a deletion plan for unattached volumes, and an online aws ec2 modify-volume migration for attached gp2 volumes. Volumes at or above 1000 GiB are skipped, since their gp2 IOPS baseline already beats the gp3 default.
Cold S3 buckets still on Standard
S3 Standard-IA holds objects at the same millisecond access for about 46% less at rest ($0.0125 against Standard's $0.023 per GB-month for the first 50 TB), but only a lifecycle Transition rule moves data already sitting in a bucket. OhChimp reads each bucket's size and BytesDownloaded from CloudWatch, then writes that rule only where the reads are measured and cold, the mean object clears the 128 KB size below which S3 refuses to transition anything, and the one-time transition request fee pays itself back inside three months. A bucket with request metrics switched off is skipped, never assumed cold.
Idle NAT gateways and unused network resources
A NAT gateway bills around $33 a month whether or not anything routes through it. OhChimp sums CloudWatch BytesOutToDestination over the window and flags gateways under a 1 GB per 30 days idle rate. It also flags unattached Elastic IPs and load balancers with no healthy targets. The Elastic IP survives deletion, so a flagged gateway can be recreated at the same address.
Over-provisioned ECS and Fargate
OhChimp lists your ECS clusters and services, estimates each service's monthly Fargate or EC2 cost from its task definition, and traces the autoscaling chain to its CloudWatch alarm or SQS trigger. It flags services running while their metric source sits at zero, services whose minimum capacity is set well above what is actually running, and services with no autoscaling policy at all.
Bedrock model spend and idle provisioned throughput
OhChimp reads per-model Bedrock token usage from the CloudWatch AWS/Bedrock namespace and prices it by input and output tokens. It surfaces frontier-tier models carrying material spend, where routing some traffic to a standard-tier model is usually the largest lever. It also flags provisioned throughput commitments with zero invocations in the window, since provisioned capacity bills per model unit regardless of traffic.
AWS cost optimization FAQ
How does OhChimp connect to my AWS account?
With an IAM access key, plus an optional session token. OhChimp verifies the connection using a read-only STS GetCallerIdentity call. The cost tools can also assume a role you provide, with an optional external ID. From there it reads Cost Explorer recommendations and EC2, EBS, CloudWatch, ECS, and Bedrock metadata to find waste. It never stores your application data, secrets, or workload contents.
Does OhChimp need write access or see my data?
No. OhChimp reads billing and resource metadata only and never stores your application data, secrets, or workload contents. When you apply a fix, the plan runs the exact commands you reviewed, like aws ec2 modify-volume for a gp2 to gp3 migration, and you are the one who clicks apply.
What AWS cost problems does OhChimp actually find?
Weak Reserved Instance and Savings Plan coverage, idle and over-provisioned EC2, unattached EBS volumes, gp2 volumes that should be gp3, cold S3 buckets still billing at the Standard rate, idle NAT gateways, unused Elastic IPs, idle load balancers, over-provisioned ECS and Fargate services, and Bedrock spend including idle provisioned throughput. It can also scan your Terraform for AWS resources and a connected EKS cluster through its Kubernetes cost scan.
Who applies the changes, and can they be rolled back?
You do. Each fix is a reviewable plan with a confidence score, a risk level, and rollback steps, like snapshotting a volume before reverting its type. Nothing changes until you click apply, and OhChimp applies the matching code and infrastructure changes together.
How do I know the savings are real, and how is it priced?
OhChimp checks your actual AWS bill after you apply a plan. Savings are marked VERIFIED only after 7 or more days, a drop of at least 10%, and 3 consecutive positive checks. Otherwise the plan stays flagged "not implemented." Pricing is flat monthly with no cut of your savings, plus a year-one ROI guarantee: a full refund of subscription fees if it does not pay for itself in your first 12 months on a paid plan.
Where the Reserved Instances and Savings Plans lever breaks even: AWS Savings Plans: 73% utilization is break-even.
The gp2-to-gp3 lever, start to finish: EBS gp2 to gp3: cut storage costs about 20%.
The three billing rules that decide whether the Standard-IA lever pays: S3 Standard-IA lifecycle rules: cut 46% without losing money.
The idle NAT gateway lever, start to finish: Find idle NAT gateways before they bill $33 a month.
The Bedrock idle provisioned-throughput lever, start to finish: Bedrock provisioned throughput bills when idle.
Related integrations
Teams running AWS usually run these too. OhChimp finds the waste in each and proves it on the bill.
Google Cloud
Idle VMs, disk tiers, committed-use discounts
Azure
Underused VMs, disk tiers, reservation coverage
DigitalOcean
Oversized Droplets, Spaces storage, idle nodes