Get started with Workspaces

Check your docs version

This version of the Anyscale docs is deprecated. Go to the latest version for up to date information.

Workspace is a fully managed development environment that enables ML practitioners to build distributed Ray applications and advance from research to development to production easily, all within a single environment. It greatly boosts developers' productivity and accelerates the dev-to-prod lifecycles by enabling them to:

Leverage familiar development tools
Program on a scalable cluster like on a laptop
Develop and debug in an environment consistent with production
Collaborate efficiently

Conceptual overview

workspaces-diagram

Workspaces consists of an Anyscale-managed Ray cluster, environment, storage, and development tools (IDEs, Git, etc.)

Ray cluster: Anyscale-managed Ray clusters provide the underlying compute that can horizontally scale up and down.
- Head node: each Ray cluster has exactly one head node. In the context of Workspace, the head node hosts various development tools (see below) and is responsible for distributing workload to worker nodes at runtime.
- Worker nodes: depends on the Compute Config of the Workspace, there could be zero or more worker nodes at a given time that are orchestrated by the head node to run your applications.
Development tools: Workspaces host or integrate with tools like JupyterLab, VS Code, Git, Weights & Biases, etc.
Storage: storage (Network File System (NFS)) persists your application code and allows the entire cluster to access the data.
Environment and dependencies: different environment management options help manage your application dependencies.