Standing up a Talos Linux Cluster on AWS Using Pulumi

NOTE: The code in this repository is currently non-functional due to an issue with all released versions of the Pulumi provider for Talos Linux.

This repository contains a Pulumi program, written in Golang, to automate the process of standing up a Talos Linux cluster on AWS.

Prerequisites

Before using the contents of this repository, you will need to ensure:

You have the Pulumi CLI installed (see here for more information on installing Pulumi).
You have a working AWS CLI installation.
You have a working installation of Golang.
You have manually installed the Pulumi provider for Talos. As of this writing, the Pulumi provider for Talos was still prerelease and needs to be installed manually; see instructions here.

Clone this repository into a directory on your local computer.
Change into the directory where you cloned this repository.
Run pulumi stack init to create a new Pulumi stack.
Use pulumi config set aws:region <region> to set the desired AWS region in which to create the cluster.
Use pulumi config set to set the correct AMI ID for a Talos Linux instance in the desired AWS region. Information on determining the correct AMI ID can be found here in the Talos Linux documentation.
Run pulumi up to run the Pulumi program.

After the Pulumi program finishes running, you can obtain a configuration file for talosctl using this command:

pulumi stack output talosctlCfg --show-secrets > talosconfig

You can then run this command to watch the cluster bootstrap:

talosctl --talosconfig talosconfig health

Once the cluster has finished boostrapping, you can retrieve the Kubeconfig necessary to access the cluster with this command:

talosctl --talosconfig talosconfig kubeconfig

You can then use kubectl to access the cluster as normal, referencing the recently-retrieved Kubeconfig as necessary.