How git4example/genai-fsx-netapp-ontap-workshop-on-eks-auto works
This repository is a hands-on workshop for building a self-hosted GenAI and agentic-AI system on AWS, not an application runtime. Its user journey moves from provisioning an EKS/FSx ONTAP environment through model serving, chatbot and agent deployment, then monitoring and optional storage/failover exercises. The graph retains those documented capabilities and their representative workshop assets. Source inspection was limited to the README and the agent implementation; links between unsampled manifests and lesson steps are therefore left out where their wiring cannot be verified.
Main entry points: litellm-deployment.yaml LLM gateway · agent.py Strands agent service · mistral-ontap.yaml vLLM model serving · index.en.md Setup lessons
Components 28
Workshop Journey
- Setup lessons
content/020_setup/index.en.md - EKS and storage
content/100_module1_eks_fsxontap/index.en.md - GenAI deployment
content/200_module2_genai/index.en.md - Observability lessons
content/300_module3_observability/index.en.md - Agent workflow
content/400_module4_agentic_fsxn/index.en.md - Storage exercises
content/600_module6_dynamic_prov/index.en.md - Failover exercises
content/700_module7_maz_failover/index.en.md
AWS Platform
- Workshop deployment
static/scripts/quick-deploy-on-demand.sh - Workshop stack
static/GenAIFSXWorkshopOnEKS.yaml - EKS infrastructure
static/terraform/main.tf - Trident storage setup
static/eks/FSxONTAP/trident-backend-config.yaml - Model preload job
static/eks/FSxONTAP/model-loading-job.yaml
GenAI Stack
- vLLM model serving
static/eks/genai/mistral-ontap.yaml - LLM gateway
static/eks/genai/litellm-deployment.yaml - Gateway configuration
static/eks/genai/litellm-config.yaml - Open WebUI
static/eks/genai/open-webui-helm/values.yaml - Monitoring stack
static/eks/genai/observability/kube-prom-stack.yaml
Agents And Operations
- Strands agent service
static/eks/agentic-agents/agent-app/agent.py - Agent deployments
static/eks/agentic-agents/finance-agent-deployment.yaml - Shared agent data
static/eks/agentic-agents/agent-shared-pvc.yaml - Prepare agent data
content/400_module4_agentic_fsxn/410_PrepareData.md - Model inspection
content/500_module5_inspect_data/index.en.md - FSx snapshots
content/500_module5_inspect_data/520_FSxNSnapshots.md - Dynamic provisioning
content/600_module6_dynamic_prov/610_dynamic_prov.md - MAZ failover
content/700_module7_maz_failover/720_TriggerFailover.md
Other components
- Workshop learner
- AWS services
- FSx ONTAP
Connections 26
- Workshop learner → Setup lessons: follows
- Workshop learner → EKS and storage: follows
- Workshop learner → GenAI deployment: follows
- Workshop learner → Observability lessons: follows
- Workshop learner → Agent workflow: follows
- Setup lessons → Workshop deployment: runs
- Workshop deployment → Workshop stack: deploys
- Workshop stack → AWS services: provisions
- EKS infrastructure → AWS services: provisions
- Trident storage setup → FSx ONTAP: configures storage
- Model preload job → FSx ONTAP: loads model
- vLLM model serving → FSx ONTAP: reads model
- LLM gateway → vLLM model serving: routes inference
- LLM gateway → Gateway configuration: uses config
- Open WebUI → LLM gateway: sends requests
- Monitoring stack → vLLM model serving: monitors
- Agent workflow → Prepare agent data: prepares data
- Agent deployments → Strands agent service: runs
- Strands agent service → LLM gateway: requests inference
- Strands agent service → Shared agent data: reads documents
- Workshop learner → Model inspection: explores
- Workshop learner → FSx snapshots: explores
- Workshop learner → Storage exercises: explores
- Workshop learner → Dynamic provisioning: explores
- Workshop learner → Failover exercises: explores
- Workshop learner → MAZ failover: triggers