Launching soon

Launching soon · Amazon ECS in your own AWS account

Production on AWS, without a DevOps team

podpilot deploys your Docker images to ECS in your AWS account. It sets up the network, the load balancer and the cluster, and gives every environment its versions and a one-click rollback. You don't need to know IAM, VPC or ECS to run it well.

See how it works

We are getting ready to open it to the first teams.

How it works

From an empty AWS account to production in three steps

  1. 1

    Connect your AWS account

    Run the AWS CLI commands podpilot gives you: they create a role it assumes, with an external ID and a permissions boundary. Nothing is copied out of your account.

  2. 2

    Create a cluster

    Pick a region and a size. podpilot provisions the VPC, load balancer, ECS cluster, instances, image registry, security groups and IAM roles, with sensible defaults.

  3. 3

    Deploy

    Choose an image tag in the app, or deploy every new image on its own: your CI calls us after the build, or we pick up the push to your registry. Each deployment is a version you can promote to the next environment or roll back to.

Deploy from your CI

any CI

One request after the build: the tag you just pushed is the version.

curl -X POST https://app.podpilot.com/api/hooks/deploy \
  -H "Authorization: Bearer $OPS_DEPLOY_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"version": "'"$IMAGE_TAG"'"}'

Ready-made steps for GitHub Actions, GitLab, Bitbucket, CircleCI, Jenkins and more.

Deploy on every push to your registry

no deploy step

Your CI keeps building and pushing images as it does now. We see each new tag in your registry and deploy it, with nothing to add to the pipeline. Each environment takes the tags you choose: everything to dev, release-* to prod.

  • GitHub Container Registry Install our GitHub App. Nothing else to set up.
  • Amazon ECR One CloudFormation stack, generated for you.

Features

The defaults a platform team would set up, ready on day one

Environments

dev, staging and prod for every application, each with its own settings, scaling and secrets.

Versions and rollback

Every deployment is recorded with who made it. Roll back to any earlier version in one click.

Auto-deploy from CI

A token per environment: your pipeline sends the new version, podpilot deploys it in order.

Secrets you keep

Values go to AWS Secrets Manager in your account and reach the container from there. podpilot keeps only the reference.

Least-privilege IAM

Each environment gets its own pod role, generated for you. No shared admin credentials in containers.

Network zones

Keep environments apart, and decide per service who may reach it, also across zones. Network ACLs follow the rules.

Service discovery

Services find each other by name inside the cluster, with the host and port given to every container.

Scheduled jobs

Run a job on a schedule next to your services, with the same image, secrets and network.

Roles and access

Invite the whole team. Decide who can only look, who can deploy and who can change an application.

Bring your own cloud

Your account, your resources, your bill

podpilot is the control plane; everything it builds lives in your AWS account, tagged and visible in your console. You pay AWS directly for what you run, with your discounts and credits.

If you stop using podpilot, nothing is taken down: your services keep serving traffic, and you remove our role when you are ready.

  • Limited access. A role with an external ID and a permissions boundary, which you can revoke at any time.
  • No secret values with us. They are written to your Secrets Manager and never stored by podpilot.
  • Plain AWS underneath. CloudFormation stacks, ECS services and load balancers you can inspect yourself.
  • No lock-in. Cancel and keep everything that runs, exactly as it is.

Where it runs

AWS first. Kubernetes on any cloud next.

We start with Amazon ECS. Next, we are bringing the same workflow to your Kubernetes clusters on Amazon EKS, Google GKE and Azure AKS.

What you set up with us is described once, not for one cloud, so your clusters will join the place where your team already works.

  1. Amazon ECS

    At launch

    In your AWS account, with the network, load balancer and cluster set up for you.

  2. Kubernetes

    Next

    On the cluster you already have, wherever it runs.

    • Amazon EKS
    • Google GKE
    • Azure AKS

  • One workflow

    Environments, versions, promotion and one-click rollback, the same on a cluster as on ECS.

  • Same access rules

    Network zones and per-service access become NetworkPolicies, every environment closed by default.

  • Secrets in your cluster

    Values go to native Kubernetes Secrets in your cluster. They never reach us.

  • Nothing to redo

    Your team, roles and deployment history stay where they are. New environments sit next to the ones you have.

FAQ

Questions

When can I start?

Soon. We are getting podpilot ready for the first teams, and this page will say when you can sign up.

Do I need to know AWS?

No. podpilot chooses the network layout, security groups and IAM for you. When you want to look under the hood, everything is plain AWS in your own console.

What do I need to start?

An AWS account and an application packaged as a Docker image. podpilot sets up the rest, including a registry for your images.

Who pays for the AWS resources?

You do, directly to AWS, as for anything else in your account. We never bill you for AWS resources.

What happens if I stop using it?

Your services keep running in your account. podpilot stops deploying for you, and shows which stacks it created and which role to remove. You can still undeploy what you no longer need.

Can podpilot read my secrets?

Secret values are written to AWS Secrets Manager in your account and read by your containers from there. podpilot keeps only the reference, never the value.

What about Kubernetes?

It is what we are building now: your Kubernetes clusters on EKS, GKE and AKS, with the same environments, versions and access rules as on ECS. See what carries over.

Production on your own AWS, soon

We are getting ready to open podpilot to the first teams. Your first cluster will take minutes. Want to be one of them, or have a question? Write to us.

contact@podpilot.com