Nexoryx_GPU_Cluster

Production-ready GPU-enabled Kubernetes AI infrastructure platform with NVIDIA GPU Operator, Ollama, CUDA workloads, autoscaling, monitoring, and distributed AI compute support.

Features

• NVIDIA GPU Operator
• GPU-enabled Kubernetes workloads
• CUDA runtime support
• Ollama GPU inference
• GPU autoscaling
• Prometheus monitoring
• Grafana dashboards
• Node Feature Discovery
• NVIDIA Device Plugin
• Persistent storage
• Ingress support
• Production-ready manifests

Stack

• Kubernetes
• NVIDIA GPU Operator
• CUDA
• Ollama
• Docker
• Prometheus
• Grafana
• NGINX Ingress
• NVIDIA Container Toolkit

Deployment

kubectl apply -f kubernetes/

Namespace

nexoryx-gpu

Components

• GPU Operator
• GPU Device Plugin
• Ollama GPU Pods
• Prometheus
• Grafana
• Ingress
• Autoscaling

Requirements

• NVIDIA GPU Nodes
• NVIDIA Drivers
• Kubernetes Cluster
• Helm installed
• NVIDIA Container Toolkit

Notes

Update domains, storage classes, and GPU resource limits before production deployment.

Project Roadmap

• Kubernetes Helm charts
• GitOps support
• CI/CD improvements
• Monitoring dashboards
• Multi-cloud support
• Security hardening

GitHub Actions

This repository includes:

• Shell validation
• Markdown linting
• Terraform validation (where applicable)

Example Deployments

See:

• examples/
• docs/

Related Nexoryx Projects

This repository is part of the Nexoryx infrastructure ecosystem.