Amazon Web Services (RKE) Deployment

Deploy TrustGraph on Amazon Web Services using RKE2 Kubernetes and AWS Bedrock integration.

Overview

TrustGraph provides a complete AWS deployment solution using Pulumi (Infrastructure as Code) that automatically provisions and configures an RKE2 Kubernetes cluster with AWS Bedrock LLM integration for a production-ready TrustGraph deployment.

What You Get

The AWS deployment includes:

  • RKE2 Kubernetes cluster with 1 server and multiple agent nodes
  • IAM roles and policies with Bedrock access permissions
  • EBS CSI driver for persistent storage provisioning
  • Complete TrustGraph stack deployed and configured
  • AWS Bedrock integration with Claude 3.5 Haiku (configurable)
  • Mistral Nemo Instruct endpoint configured
  • Secrets management for secure credential handling
  • Monitoring and observability with Grafana
  • Web workbench for document processing and Graph RAG

Deployment Method

The deployment uses Pulumi, an Infrastructure as Code tool that:

  • Has an open-source license
  • Uses general-purpose programming languages (TypeScript/JavaScript)
  • Provides testable infrastructure code
  • Offers retryable deployments
  • Supports local or S3 state management

Architecture

Kubernetes Platform: RKE2 (Rancher Kubernetes Engine 2) Node Configuration: t3a.xlarge instances (configurable) Default Setup: 3 nodes, Amazon Linux 2023 Network: Custom VPC with configurable CIDR blocks Storage: EBS volumes with automatic provisioning LLM Service: AWS Bedrock with multiple model options

Quick Process Overview

  1. Install Pulumi and dependencies
  2. Configure AWS credentials via .aws/credentials
  3. Set up Bedrock model access in AWS console
  4. Customize configuration in Pulumi.STACKNAME.yaml
  5. Deploy with pulumi up
  6. Wait for initialization (~30 seconds after pods are running)
  7. Access services via port-forwarding

Configuration Options

Customizable settings include:

  • Environment: dev, staging, production
  • Region: us-west-2 (default, configurable)
  • VPC CIDR: 172.38.0.0/16
  • Node type: t3a.xlarge (configurable)
  • Node count: 3 (configurable)
  • AMI: Amazon Linux 2023

Access Points

Once deployed, you’ll have access to:

  • TrustGraph API: Port 8088
  • Web Workbench: Port 8888 (document processing, Graph RAG)
  • Grafana Monitoring: Port 3000

Bedrock Integration

The deployment includes AWS Bedrock integration with:

  • Default Model: Claude 3.5 Haiku
  • Alternative Models: Mistral Nemo Instruct and others
  • Model Access: Requires enabling specific models in Bedrock console
  • Secure Access: IAM roles and policies automatically configured

Complete Documentation

For detailed step-by-step instructions, configuration options, and troubleshooting, visit:

TrustGraph AWS RKE Deployment Guide

The repository contains:

  • Complete Pulumi deployment code
  • RKE2 cluster configuration
  • AWS Bedrock integration setup
  • Detailed setup instructions
  • Troubleshooting guides
  • Customization options

Important Notes

Storage Cleanup: The EBS CSI driver doesn’t automatically clean up volumes on teardown. Manual cleanup may be required in the EC2 console.

Initialization Time: Allow ~30 seconds after all pods are running for application initialization to complete.

Bedrock Access: Ensure you have enabled the required models in your AWS Bedrock console before deployment.

Next Steps

After deployment, you can:

  • Load documents through the web workbench
  • Test Graph RAG queries with Bedrock models
  • Monitor processing through Grafana
  • Scale the cluster by adjusting node count
  • Integrate with other AWS services