Gryvia Documentation

Documentation for Gryvia, a Kubernetes-native GPU platform. Gryvia is alpha software (gryvia.io/v1alpha1): pages say what is verified, what is experimental and what is only a CRD or design so far. Start with the status section of the README.

Documentation Structure

Getting Started

• Quick Start Guide - kind demo without GPUs, Helm install, k3s + GPU bootstrap, remote k3s

User Guides

• Submitting Jobs - How to submit and manage AI workloads
• GPU as a Service - Tenants, GPU SKU catalog, usage metering and estimate invoices

Administrator Guides

• Cluster Setup - Setting up a Gryvia cluster
• GPU Nodes - Drivers, CUDA and the device plugin through the bundled NVIDIA GPU Operator
• Authentication and TLS - API key, sessions, OIDC roles and tenants, certificates
• Deployment Guide - Complete deployment instructions
• Operations - Upgrade, uninstall, backup, troubleshooting
• Operational Playbooks - Day-2 operations

Developer Guides

• API Reference - Every API gateway route and who may call it
• CRD Reference - The 49 custom resources and which ones have a controller
• CLI Guide - Command-line interface usage
• Storage & Network Operators - Operator internals

Feature Guides

• ML Workflows - Hyperparameter tuning, DAG pipelines, model registry, inference serving, workspaces
• Network Intelligence - eBPF-powered observability, security, and traffic analysis
• Scheduling - What the scheduler does today and the design for gang scheduling, fair-share queues and elastic training

Reference

• FAQ - Frequently asked questions
• Roadmap - Feature roadmap
• Advanced Features - Advanced platform features
• Integrations - Third-party integrations
• Budget Management - Budget design (see its status note)
• Flight Recorder - Node-local, job-attributed eBPF timeline (preview)
• GPU validation checklist - What to run on real GPU hardware

Examples

• Job Examples - Sample job configurations
• Workflow Examples - Argo Workflow templates
• Integration Examples - JupyterHub, VSCode, Ray
• Templates - Pre-configured job templates
• Complete Setup - Full production example

Quick Links

For End Users

• New to Gryvia? Start with the Quick Start Guide
• Submit your first job: See Job Submission Guide

For Administrators

• Install Gryvia: Start with the Quick Start, then the Deployment Guide
• Configure cluster: See Cluster Setup

For Developers

• REST API: Browse API Reference
• CLI commands: See CLI Guide

Support

• Issues: https://github.com/zyvorai/gryvia/issues
• Discussions: https://github.com/zyvorai/gryvia/discussions

Contributing

We welcome contributions! See Contributing Guide for details.

License

Apache License 2.0 - see LICENSE for details.