Docs

Network requirements

Network requirements are the connections a customer's environment makes for a deployment that your Private Alien Instance manages. A customer's network team can use this page to decide what to allow. Every connection uses HTTPS on port 443, and nothing connects inbound to the customer's network.

To your Private Alien Instance

FromWhy
Workloads of every deployment, unless telemetry is turned offSend logs
Kubernetes pull and Single machine pull deploymentsGet releases and commands, and report status
The setup CLI, when every platform in the release uses the same Private Alien Instance as its defaultRegister the deployment

The address is the Private Alien Instance's endpoint: its default endpoint, or the custom domain you chose. Give the customer that hostname, and have them allow it by name, since its addresses can change.

Push deployments are updated through the cloud's own APIs with the management identity, so the Private Alien Instance doesn't connect into the customer's network.

To the customer's own cloud

Workloads call the cloud services of the customer's own account, such as the registry they pull images from and the secret store they read values from. A network with no internet route needs a NAT gateway or the cloud's private endpoints for those services.

During initial setup

FromTo
The admin's browserThe deployment portal, at Alien's address or your domain
The admin's browser or machineThe cloud provider's console and APIs
The admin's machinepkg.alien.dev, or pkg.<your domain>, to download setup packages
The setup CLI, in other casesAlien's API at api.alien.dev

On AWS, CloudFormation reports the new stack to Alien through a custom resource. AWS makes that call, so it doesn't pass through the customer's network.

Still goes to Alien

Machines in a Machines deployment connect to Alien's machine service, not to the Private Alien Instance. Check each feature you use before you tell a customer that nothing leaves their network for Alien.

Next

On this page