DevPod Setup Guide

This guide provides comprehensive setup instructions for using DevPod with various providers, including local Docker, major cloud platforms, and manual setup for services like Rackspace.

Local Docker Setup
DigitalOcean Setup
AWS Setup
GCP Setup
Azure Setup
Fly.io Setup
Rackspace Spot Setup

Local Docker Setup 🐳

The simplest setup option, perfect for getting started. Uses your local Docker installation.

Prerequisites

Docker Desktop (or Docker Engine on Linux) installed and running

Setup Steps

Add the Provider (DevPod usually detects Docker automatically):
```
devpod provider add docker
```

Create a Workspace:

devpod up https://github.com/loft-sh/devpod-example-simple

DevPod will pull the necessary image and start a container on your local machine.

DigitalOcean Setup 💧

Create development workspaces on DigitalOcean Droplets.

Prerequisites

DigitalOcean account
doctl CLI installed and configured
Personal Access Token (PAT) with read/write permissions

Setup Steps

Authenticate with doctl:

doctl auth init --access-token YOUR_DIGITALOCEAN_TOKEN

Add the DigitalOcean Provider:
```
devpod provider add digitalocean
```

Create a Workspace:

devpod up https://github.com/your/repo --provider digitalocean

Customize region and instance size:

devpod up https://github.com/your/repo --provider digitalocean --option region=nyc3

AWS Setup 🌩️

Use Amazon EC2 instances for your workspaces with optional Spot Instance support for cost savings.

Prerequisites

AWS account
AWS CLI installed
AWS credentials configured (aws configure or environment variables)

Setup Steps

Authenticate with AWS CLI:

aws configure
# Follow prompts: Access Key ID, Secret Access Key, region, output format

Add the AWS Provider:
```
devpod provider add aws
```

Create a Workspace (Standard Instance):

devpod up https://github.com/your/repo --provider aws

Create a Workspace (Spot Instance):

devpod up https://github.com/your/repo --provider aws --option spot-instance=true

Set maximum spot price:

devpod up https://github.com/your/repo --provider aws --option spot-instance=true --option spot-price=0.03

GCP Setup ☁️

Provision workspaces on Google Compute Engine (GCE) virtual machines.

Prerequisites

Google Cloud Platform account with project created
gcloud CLI installed
Billing enabled and Compute Engine API enabled for your project

Setup Steps

Authenticate with gcloud:

gcloud auth application-default login
gcloud config set project YOUR_PROJECT_ID

Add the GCP Provider:
```
devpod provider add gcp
```

Create a Workspace:

devpod up https://github.com/your/repo --provider gcp

Customize zone and machine type:

devpod up https://github.com/your/repo --provider gcp --option zone=us-central1-a --option machine-type=e2-medium

Azure Setup 🚀

Use Azure Virtual Machines for your development environments.

Prerequisites

Azure account with active subscription
Azure CLI (az) installed

Setup Steps

Authenticate with Azure CLI:

az login
# Follow browser-based login process

Set active subscription:

az account set --subscription "YOUR_SUBSCRIPTION_NAME_OR_ID"

Add the Azure Provider:
```
devpod provider add azure
```

Create a Workspace:

devpod up https://github.com/your/repo --provider azure

Specify location and VM size:

devpod up https://github.com/your/repo --provider azure --option location=eastus --option vm-size=Standard_B2s

Fly.io Setup ✈️

Use Fly.io Machines for globally distributed development environments with edge deployment capabilities.

Prerequisites

Fly.io account (credit card required for signup)
flyctl CLI installed

Installing flyctl

macOS (Homebrew):

brew install flyctl

Linux:

curl -L https://fly.io/install.sh | sh

Windows (PowerShell):

pwsh -Command "iwr https://fly.io/install.ps1 -useb | iex"

After installation, add flyctl to your PATH if prompted.

Setup Steps

Create a Fly.io Account (if you don't have one):
```
fly auth signup
```

Authenticate with Fly.io:

fly auth login
# This opens a browser for authentication

Verify Authentication:
```
fly auth whoami
```

Add the SSH Provider with Fly.io Machine:

Since Fly.io doesn't have an official DevPod provider, you can use the SSH provider with a Fly.io Machine:

a. Create a Fly.io App for DevPod:

fly apps create devpod-workspace

b. Create a Machine with Docker installed:

fly machine run ubuntu:22.04 \
  --app devpod-workspace \
  --region iad \
  --vm-size shared-cpu-1x \
  --vm-memory 1024

c. SSH into the Machine and install Docker:

fly ssh console --app devpod-workspace
# Inside the machine:
apt-get update && apt-get install -y docker.io

d. Issue SSH credentials:

fly ssh issue --agent

e. Add the SSH Provider:

devpod provider add ssh -o HOST=root@devpod-workspace.fly.dev

Create a Workspace:

devpod up https://github.com/your/repo --provider ssh

Alternative: Using Fly.io with WireGuard

For a more secure connection, set up WireGuard VPN:

Create a WireGuard tunnel:
```
fly wireguard create
```
Import the configuration into your WireGuard client and connect.

Use the internal Fly.io address for SSH:

devpod provider add ssh -o HOST=root@<machine-id>.vm.<app-name>.internal

Available Fly.io Regions

Region ID	Location
ams	Amsterdam, Netherlands
cdg	Paris, France
dfw	Dallas, Texas (US)
ewr	Secaucus, NJ (US)
fra	Frankfurt, Germany
gru	São Paulo, Brazil
iad	Ashburn, Virginia (US)
lax	Los Angeles, California (US)
lhr	London, United Kingdom
nrt	Tokyo, Japan
ord	Chicago, Illinois (US)
sin	Singapore
sjc	San Jose, California (US)
syd	Sydney, Australia
yyz	Toronto, Canada

VM Sizes

Common VM size options for development workspaces:

Size	CPUs	Memory	Use Case
`shared-cpu-1x`	1 shared	256MB	Light development
`shared-cpu-2x`	2 shared	512MB	Standard development
`shared-cpu-4x`	4 shared	1GB	Medium workloads
`performance-1x`	1 dedicated	2GB	Performance-sensitive
`performance-2x`	2 dedicated	4GB	Heavy development

Example with custom sizing:

fly machine run ubuntu:22.04 \
  --app devpod-workspace \
  --region ord \
  --vm-size shared-cpu-2x \
  --vm-memory 2048

Cost Considerations

Fly.io uses pay-as-you-go pricing with billing per second
Stopped machines only incur storage costs (~$0.15/GB/month)
Consider using fly machine stop when not developing to save costs
Machines can be automatically stopped with the --auto-stop flag

Useful Fly.io Commands

# List all machines
fly machines list --app devpod-workspace

# Stop a machine
fly machine stop <machine-id> --app devpod-workspace

# Start a machine
fly machine start <machine-id> --app devpod-workspace

# Delete a machine
fly machine destroy <machine-id> --app devpod-workspace

# View logs
fly logs --app devpod-workspace

# SSH into the machine
fly ssh console --app devpod-workspace

Rackspace Spot Setup 🎯

⚠️ Beta Notice: VM CloudSpaces are currently in beta. The instructions below cover both Kubernetes CloudSpaces (generally available) and VM CloudSpaces (beta).

Use Rackspace Spot for cost-effective development environments with market-based pricing. Rackspace Spot offers cloud infrastructure through an open market auction, delivering fully managed Kubernetes clusters or standalone VMs at significantly reduced prices compared to traditional cloud providers.

Prerequisites

Rackspace Spot account (sign up at spot.rackspace.com)
spotctl CLI installed (for command-line management)
API token from the Rackspace Spot dashboard

Installing spotctl

Download the binary from the releases page:

Linux/macOS:

# Download the appropriate binary for your platform
curl -LO https://github.com/rackspace-spot/spotctl/releases/latest/download/spotctl-linux-amd64
chmod +x spotctl-linux-amd64
sudo mv spotctl-linux-amd64 /usr/local/bin/spotctl

Windows: Download the Windows binary from the releases page and add it to your PATH.

Setup Steps

Create a Rackspace Spot Account:
- Visit spot.rackspace.com
- Sign up using GitHub, Google, or email/password
Get Your API Token:
- Log in to the Rackspace Spot dashboard
- Navigate to API Access in the left sidebar
- Copy your access token

Configure spotctl:

spotctl configure
# Follow the interactive prompts to enter:
# - Organization name
# - Region
# - Refresh token

Option A: Kubernetes CloudSpaces

Use DevPod with Rackspace Spot's fully managed Kubernetes clusters.

Create a Kubernetes CloudSpace

Using spotctl:

spotctl cloudspaces create \
  --name devpod-cluster \
  --region us-central-dfw-1 \
  --org your-org-name \
  --spot-nodepool "serverclass=gp.vs1.medium-dfw,desired=2,bidprice=0.09"

Using Terraform:

terraform {
  required_providers {
    spot = {
      source = "rackerlabs/spot"
    }
  }
}

variable "rackspace_spot_token" {
  description = "Rackspace Spot authentication token"
  type        = string
  sensitive   = true
}

provider "spot" {
  token = var.rackspace_spot_token
}

resource "spot_cloudspace" "devpod" {
  cloudspace_name    = "devpod-cluster"
  region             = "us-central-dfw-1"
  hacontrol_plane    = false
  wait_until_ready   = true
  kubernetes_version = "1.31.1"
  cni                = "calico"
}

resource "spot_spotnodepool" "workers" {
  cloudspace_name      = spot_cloudspace.devpod.cloudspace_name
  server_class         = "gp.vs1.medium-dfw"
  bid_price            = 0.09
  desired_server_count = 2
}

Add the Kubernetes Provider to DevPod

Get your kubeconfig:

spotctl cloudspaces get-config devpod-cluster --file ~/.kube/config-rackspace-spot

Add the Kubernetes provider:
```
devpod provider add kubernetes
```

Configure the provider to use your Rackspace Spot cluster:

export KUBECONFIG=~/.kube/config-rackspace-spot
devpod provider set-options kubernetes --option KUBERNETES_CONTEXT=devpod-cluster

Create a Workspace:

devpod up https://github.com/your/repo --provider kubernetes

Option B: VM CloudSpaces (Beta)

⚠️ Beta: VM CloudSpaces are currently in beta and provide standalone virtual machines with spot pricing.

Create a VM CloudSpace

Using spotctl:

spotctl cloudspaces create \
  --name devpod-vm \
  --region us-central-dfw-1 \
  --org your-org-name \
  --ondemand-nodepool "serverclass=gp.vs1.large-dfw,desired=1"

Connect with SSH Provider

Get the VM's IP address from the Rackspace Spot dashboard or CLI
Configure SSH access to your VM

Add the SSH Provider:

devpod provider add ssh -o HOST=root@<your-vm-ip>

Create a Workspace:

devpod up https://github.com/your/repo --provider ssh

Available Regions

Region ID	Location
us-central-dfw-1	Dallas, Texas (US)
us-central-dfw-2	Dallas, Texas (US)
us-central-ord-1	Chicago, Illinois (US)
us-east-iad-1	Ashburn, Virginia (US)
eu-west-lon-1	London, United Kingdom
apac-syd-1	Sydney, Australia
apac-hkg-1	Hong Kong

Server Classes

Common server classes for development workspaces:

Server Class	vCPUs	Memory	Use Case
`gp.vs1.small-dfw`	1	1GB	Light development
`gp.vs1.medium-dfw`	2	2GB	Standard development
`gp.vs1.large-dfw`	4	4GB	Medium workloads
`gp.vs1.xlarge-dfw`	8	8GB	Heavy development
`mem.vs1.large-dfw`	4	16GB	Memory-intensive

Note: Replace -dfw suffix with your region code (e.g., -ord, -iad, -lon).

GPU Server Classes

For AI/ML development workloads:

Server Class	GPU	Use Case
`gpu.a30.vs2.xlarge`	NVIDIA A30	ML training/inference
`gpu.h100.vs2.mega-xlarge`	NVIDIA H100	Large-scale AI workloads

Cost Considerations

Market-Based Pricing: Prices are determined by real-time supply and demand
Bid Strategy: Set a maximum bid price; you pay the market rate up to your bid
Preemption: Lower bids may be preempted when demand increases
Savings: Typically 80-95% cheaper than traditional cloud providers
No Commitments: No long-term contracts or reserved instances required

Preemption Handling

Set up Slack notifications for node preemption:

spotctl cloudspaces create \
  --name devpod-cluster \
  --region us-central-dfw-1 \
  --org your-org-name \
  --preemption-webhook "https://hooks.slack.com/services/YOUR/WEBHOOK/URL" \
  --spot-nodepool "serverclass=gp.vs1.medium-dfw,desired=2,bidprice=0.09"

Useful Rackspace Spot Commands

# List all cloudspaces
spotctl cloudspaces list

# Get cloudspace details
spotctl cloudspaces get --name devpod-cluster

# List spot node pools
spotctl nodepools spot list --cloudspace devpod-cluster

# Scale a node pool
spotctl nodepools spot update \
  --name my-nodepool \
  --cloudspace devpod-cluster \
  --desired 3 \
  --bidprice 0.10

# Delete a cloudspace
spotctl cloudspaces delete --name devpod-cluster

# List available server classes
spotctl serverclasses list

# Get kubeconfig for a cluster
spotctl cloudspaces get-config devpod-cluster --file ~/.kube/config

# List organizations
spotctl organizations list

Terraform Resources

The Rackspace Spot Terraform provider is available on the Terraform Registry:

# Initialize Terraform with the Spot provider
terraform init

# Plan your infrastructure
terraform plan -var="rackspace_spot_token=$SPOT_TOKEN"

# Apply changes
terraform apply -var="rackspace_spot_token=$SPOT_TOKEN"

Uh oh!

FilesExpand file tree

devpod_provider_setup_guide.md

Latest commit

History

devpod_provider_setup_guide.md

File metadata and controls

DevPod Setup Guide

Table of Contents

Local Docker Setup 🐳

Prerequisites

Setup Steps

DigitalOcean Setup 💧

Prerequisites

Setup Steps

AWS Setup 🌩️

Prerequisites

Setup Steps

GCP Setup ☁️

Prerequisites

Setup Steps

Azure Setup 🚀

Prerequisites

Setup Steps

Fly.io Setup ✈️

Prerequisites

Installing flyctl

Setup Steps

Alternative: Using Fly.io with WireGuard

Available Fly.io Regions

VM Sizes

Cost Considerations

Useful Fly.io Commands

Rackspace Spot Setup 🎯

Prerequisites

Installing spotctl

Setup Steps

Option A: Kubernetes CloudSpaces

Create a Kubernetes CloudSpace

Add the Kubernetes Provider to DevPod

Option B: VM CloudSpaces (Beta)

Create a VM CloudSpace

Connect with SSH Provider

Available Regions

Server Classes

GPU Server Classes

Cost Considerations

Preemption Handling

Useful Rackspace Spot Commands

Terraform Resources