Kubernetes on Docker Desktop

You probably already have Kubernetes installed without realising it. Most Docker Desktop setups include a built-in single-node cluster.

This gives us a quick way to test a more production style Bionic install.

1. Switch on Kubernetes in Docker Desktop

Alt text

Docker Desktop with Kubernetes enabled is the easiest option for macOS, Windows, and most Linux desktops.

2. Install Bionic with the Stack CLI

We use Stack which is a Kubernetes Operator for installing applications in a secure and repeatable way. Stack manages the application and generates all required secrets.

Grab the CLI.

curl -fsSL https://stack-cli.com/install.sh | bash

Bootstrap the platform operators into your cluster.

stack init

This command installs CloudNativePG, Keycloak, ingress, the Stack controller, and custom resource definitions that describe your applications.

🔌 Connecting to the cluster...
✅ Connected
🐘 Installing Cloud Native Postgres Operator (CNPG)
⏳ Waiting for Cloud Native Postgres Controller Manager
🛡️ Installing Keycloak Operator
📦 Creating namespace keycloak
⏳ Waiting for Keycloak Operator to be Available
📦 Creating namespace stack-system
📜 Installing StackApp CRD
⏳ Waiting for StackApp CRD
🔐 Setting up roles
🤖 Installing the operator into stack-system
🗄️ Ensuring Keycloak database in namespace keycloak
✅ Keycloak database created.
🛡️ Ensuring Keycloak instance in namespace keycloak

Apply the demo StackApp manifest.

curl -fsSL https://raw.githubusercontent.com/bionic-gpt/bionic-gpt/main/infra-as-code/stack.yaml \
  -o demo.stack.yaml

And then

stack deploy --manifest demo.stack.yaml --profile dev

You should see

🔌 Connecting to the cluster...
✅ Connected
📜 Installing StackApp CRD
⏳ Waiting for StackApp CRD
📦 Creating namespace bionic-gpt
🚀 Applied StackApp `bionic-gpt` in namespace `bionic-gpt`

3. Done

Hopefully if everything went well you should be able to access bionic on http://localhost:30000.