Shadow GPU - Documentation

Learn how to deploy, manage, and optimize GPU-powered workloads on Shadow, including cluster setup, choosing instance types, job scheduling and load-balancing, performance tuning and profiling, cost optimization strategies, monitoring and alerting, security best practices, and troubleshooting common issues.

Comprehensive guides and tutorials to help you get the most out of Shadow GPU PaaS.

Shadow GPU PaaS Pro offers

Getting Started

Essential tutorials to help you begin.

play_circle

Shadow GPU PaaS

Quickstart, Cluster API setup, Spot instances, load balancing, and Windows GPU instances — all in one place.

dashboard

Skyline Dashboard Overview

Learn how to navigate the Skyline dashboard, access key resources, and manage your user and API settings.

terminal

CLI Guide

Learn how to configure and use the OpenStack CLI to manage your resources.

manage_accounts

Project Management

Best practices for managing your projects.

network_wifi

Networking

A guide to understanding and managing networks using Neutron.

security

Security measures

Learn about the security features and best practices in Keystone.

Advanced Guides

In-depth documentation for more complex workflows.

storage

Storage classes

Overview of available storage classes in Cinder.

image

Managing Images in Shadow OpenStack

Learn how to effectively utilize and manage images with Glance.

public

Regional Endpoints

List of regional endpoints for accessing Horizon endpoints.

info

Understanding Instance Statuses

A comprehensive overview of instance statuses in Nova.

warning

OpenStack Limitations

Important limitations to be aware of when using Shadow GPU PaaS.

Explore Shadow GPU Offers

Discover the range of GPU instances and services available.

Key Resources

Useful links and external documentation.

help

FAQ

Frequently asked questions.

public

OpenStack Public Docs

Official upstream documentation.

code

API Reference

REST API docs for integrations and automation.

Need Help?

Get support from the Platform team.