# Cost optimization

> Wholesale AWS pricing, spot placement, free Starter clusters, and transparent Production platform fees. Know what you'll pay before you ship.

Ownkube is built around a simple premise: **you pay AWS directly for compute, with a transparent platform fee on Production and no Ownkube fee on Starter.** No markup, no opaque bundles, no surprise usage tiers.

## The cost model in one paragraph

AWS bills you for the underlying compute ([EC2](https://aws.amazon.com/ec2/pricing/), storage, bandwidth) at their standard rates. Any credits you have ([AWS Activate](https://aws.amazon.com/activate/), [committed-use discounts](https://aws.amazon.com/savingsplans/)) apply directly. Ownkube adds a transparent platform fee on top: **\$5 / vCPU + \$1 / GB RAM** on the Production shape ([EKS](https://aws.amazon.com/eks/pricing/)), metered per second on requested resources. Starter clusters (K3s on one AWS instance) are free, with no Ownkube platform fee at all.

See [Pricing](/docs/configuration/pricing) for the full breakdown.

## Where the savings come from

### No compute markup

Traditional PaaS marks up AWS compute 3 to 5 times. Ownkube doesn't. You pay wholesale for the instance and a transparent platform fee on Production only.

### Starter is free

A Starter cluster (K3s on one AWS instance) runs unlimited apps, workers, and databases with **no Ownkube platform fee**. You only pay AWS for the EC2 instance.

### Cost agent (roadmap)

Will right-size over-provisioned workloads, auto-sleep idle environments, and flag spend anomalies. Aimed at the recurring FinOps work a DevOps engineer would otherwise do by hand.

### Scaling agent without waste

Replica-level and node-level scaling means you only pay for capacity you're using. The Scaling agent owns this. See [Autoscaling](/docs/features/autoscaling).

### Burn AWS credits efficiently

Every dollar of [AWS Activate](https://aws.amazon.com/activate/) or other startup credits applies directly to your compute bill at wholesale rates. Credits expire in 12 to 24 months, redeem them against real EC2, S3, and bandwidth instead of letting them lapse.

### Skip the $200K DevOps hire

A US senior DevOps engineer runs $180K to $250K loaded. The Cost and Scaling agents cover the ops work day to day, so a 5- to 20-person team can stay on engineers building product.

## A worked example

A small app (one web service + one worker + one Postgres database) on a Starter cluster (K3s on one AWS instance):

| Component | Monthly cost |
|---|---|
| Ownkube platform fee | \$0 (free) |
| EC2 `t3.medium` (2 vCPU / 4 GiB) | ~\$30 |
| EBS storage (30 GB gp3) | ~\$3 |
| Outbound bandwidth (modest) | ~\$2 |
| **Total** | **~\$35** |

Adding another web service or worker to the same node: **\$0 extra**. That's the wedge over per-service PaaS pricing.

## On Production at scale

For a production workload running 8 vCPU / 16 GiB across multiple services on Production (EKS):

| Component | Monthly cost |
|---|---|
| EKS control plane | ~\$73 |
| Worker nodes (mixed on-demand + spot) | ~\$150-200 |
| Ownkube platform fee: 8 vCPU x \$5 + 16 GB x \$1 | \$56 |
| **Total** | **~\$280-330** |

The platform fee ramps linearly with requested resources, billed per second from the first second.

## On the Cost agent's roadmap

In-dashboard cost visibility and the Cost agent are on the roadmap. Planned capabilities:

- **Auto-sleep schedules** for non-production environments during off-hours. Example output: "3 staging envs idle since 8pm. Auto-slept. ~$127/mo saved this week."
- **Right-sizing recommendations** to flag deployments with chronically low utilization. Example output: "api-worker over-provisioned: 2GB allocated, 340MB peak. Right-size to save ~$18/mo."
- **Cost anomaly alerts** when spend jumps outside its normal band. Example output: "EKS data-transfer-out 4x higher than 7-day baseline. Investigate before month-end."

## Limits and constraints

- **Actual AWS costs** come from AWS at their standard rates. Your AWS bill is the canonical source of truth for compute cost.
- **Reserved instances and Savings Plans** you purchase on AWS apply directly; Ownkube doesn't need to know about them
- **Compute markup** is zero. This is the wedge and we're not changing it.

- [Pricing](/docs/configuration/pricing)
- [Clusters](/docs/features/clusters)

---

**Don't see a feature you need?** Email [support@ownkube.io](mailto:support@ownkube.io?subject=Feature%20request). Ownkube is shaped by the teams using it and we ship what our users ask for.