On this page 6 sections
Most apps on Ownkube run on Ownkube Compute and never need to move. Sometimes one does: a customer wants it inside their AWS account, or a contract says where the data lives.
When that happens, the same app moves into your own or a customer’s AWS account with no change to the code, the build or the variables, and its logs, metrics and health checks come with it. This post covers the three steps.
TL;DR
- Connect the AWS account with
okctl aws connect. It creates a scoped IAM role, and Ownkube never receives AWS keys. - Create a cluster in that account from the dashboard or with
okctl clusters create. - Copy the app onto it with
okctl deploy copy, then check its logs, metrics and health checks in the same dashboard. - AWS is the only cloud today. GCP and Azure are coming soon.
What stays the same
The app does not know where it runs. It builds from the same repo or image, reads the same variables, and keeps its health checks and cron schedules. The dashboard, the CLI and your coding agent work the same way against either target.
Observability is the part that usually breaks in a move like this, and here it travels with the app. Logs, CPU and memory metrics, request rates and health check status show up in the same panels, now read from the cluster in your account.
What changes is who runs the compute and who bills it. On Compute you pay from the prepaid wallet. In your own account, AWS bills you for the infrastructure and Ownkube operates the platform on top. See pricing for the current numbers.
Step 1: Connect the AWS account
Connecting runs a CloudFormation stack in the account that creates an IAM role scoped to what the platform needs. Ownkube works through that role and never receives your keys.
okctl aws connectThis opens the stack in the AWS console. If your shell already has AWS credentials, okctl aws connect --deploy runs it end to end from the terminal. You can start the same connection from the dashboard instead. The connect your AWS account guide walks through the screens.
Step 2: Create a cluster
Once the account shows as connected, create a cluster in it from the dashboard. Pick the AWS region and a size, and Ownkube provisions it. When it is active, it appears next to the Compute regions wherever you choose where an app runs.
okctl clusters create --name prod --account <aws-account-id> --region eu-central-1 --type starterokctl clusters status <cluster-id>The clusters docs cover the sizes and what each one suits.
Step 3: Copy the app onto the cluster
Copy the app into an environment on the new cluster. The copy keeps its source, variables, health checks and schedule:
okctl deploy copy api --to-environment <environment-id> --to-cluster <cluster-id>If the app uses a database, create one on the cluster, link it, and restore a pg_dump the same way you would on Compute. Then open the app in the dashboard and confirm the build, logs, metrics and health checks look the way they did before. Keep the Compute copy running until the new one has served real traffic, then delete it.
FAQ
Do I need to rewrite my app to run it in my own AWS account?
No. The same repo or image, variables and settings carry over. What changes is the infrastructure the app runs on and who bills you for it.
Does Ownkube get access to my AWS keys?
No. Connecting creates a scoped IAM role through CloudFormation, and Ownkube works through that role.
Can I use GCP or Azure instead of AWS?
Not yet. AWS is the only cloud today, and GCP and Azure are coming soon.
Do my logs and metrics survive the move?
Yes. They are part of the platform, so they show up in the same dashboard panels after the move.
Where Ownkube fits
Start on Ownkube Compute with no cloud account to set up. If an app later has to run inside your own or a customer’s AWS account, move it there in three steps and keep the logs, metrics and health checks you already rely on. Deploy my first app.