coder/README.md

display_name, description, icon, maintainer_github, verified, tags

display_name	description	icon	maintainer_github	verified	tags
Docker Workspaces with GPU & DooD	Provision Docker-based Coder workspaces with GPU support, Docker-out-of-Docker, Go, and Bun.	../../../site/static/icon/docker.png	coder	true	docker container gpu golang bun sdkman

Feature-Rich Docker Workspaces

Provision powerful Docker containers as Coder workspaces with this advanced template. It features GPU passthrough, Docker-out-of-Docker (DooD), and automatically installs essential development tools on startup.

Prerequisites

Infrastructure

The VM you run Coder on must have a running Docker socket and the coder user must be added to the Docker group:

# Add coder user to Docker group
sudo adduser coder docker

# Restart Coder server
sudo systemctl restart coder

# Test Docker
sudo -u coder docker ps

Features & Architecture

This template provisions the following resources and features:

• Base Image: codercom/enterprise-base:ubuntu
• GPU Support: Passes all host GPUs to the workspace (gpus = "all") for AI/ML and hardware acceleration.
• Docker-out-of-Docker (DooD): Mounts the host's /var/run/docker.sock so you can build and run containers seamlessly from inside your workspace.
• Auto-Provisioned Tools: The workspace startup script automatically handles:
  • System package updates (apt-get update & apt-get upgrade)
  • Docker CLI
  • Nixpacks
  • Bun
  • Golang
  • SDKMAN! & Maven
• Persistent Storage: Uses a persistent host directory mapped to /home/coder to ensure your files survive workspace restarts.
• Rich Telemetry: Displays CPU, RAM, Disk, Load Average, and GPU utilization directly in the Coder UI.
• IDE Support: Native integration with code-server (Browser VS Code) and JetBrains IDEs.

Note

While your /home/coder directory is persistent, the container's root filesystem is ephemeral. The startup script handles reinstalling your system-level tools (Go, Bun, SDKMAN!, etc.) on each start. To speed up boot times, you may want to bake these tools into a custom Docker image.