Skip to content
DuraGraph

Cloud Deployment

Deploy DuraGraph on any cloud provider with full control over your infrastructure. DuraGraph is designed to be cloud-agnostic and runs anywhere you can deploy containers.

Deployment Options

Section titled “Deployment Options”

☁️ Any Cloud Provider

Section titled “☁️ Any Cloud Provider”

DuraGraph runs on any cloud that supports Docker containers:

  • AWS - ECS, EKS, EC2
  • Google Cloud - Cloud Run, GKE, Compute Engine
  • Azure - Container Apps, AKS, VMs
  • DigitalOcean - App Platform, Kubernetes
  • Fly.io - Global edge deployment
  • Railway - Simple container hosting
  • Render - Managed infrastructure

🏢 On-Premises

Section titled “🏢 On-Premises”

Deploy in your own data center with:

  • Docker Compose for single-node deployments
  • Kubernetes for production clusters
  • Air-gapped environments supported

🌐 Hybrid

Section titled “🌐 Hybrid”

Mix cloud and on-premises deployments:

  • Manage multiple environments from a single dashboard
  • Federated execution across regions
  • Data residency compliance

Quick Deploy

Section titled “Quick Deploy”

One-Click Deploy Buttons

Section titled “One-Click Deploy Buttons”

Deploy on Fly.io Deploy to Render Deploy on Railway Deploy on DigitalOcean

Docker Compose

Section titled “Docker Compose”
Terminal window
# Clone and deploy
git clone https://github.com/duragraph/duragraph.git
cd duragraph
docker-compose up -d

Kubernetes

Section titled “Kubernetes”
Terminal window
# Using Helm
helm repo add duragraph https://charts.duragraph.ai
helm install duragraph duragraph/duragraph

Cloud-Specific Guides

Section titled “Cloud-Specific Guides”

AWS Deployment

Section titled “AWS Deployment”

ECS with Fargate

Terminal window
# Create ECS cluster
aws ecs create-cluster --cluster-name duragraph


# Deploy using CDK or CloudFormation
cd deploy/aws
cdk deploy

Managed Database Options:

  • Amazon RDS for PostgreSQL
  • Amazon ElastiCache for Redis (optional caching)
  • Amazon MSK or self-managed NATS

Google Cloud Deployment

Section titled “Google Cloud Deployment”

Cloud Run

Terminal window
# Build and push image
gcloud builds submit --tag gcr.io/PROJECT_ID/duragraph


# Deploy to Cloud Run
gcloud run deploy duragraph \
  --image gcr.io/PROJECT_ID/duragraph \
  --platform managed \
  --allow-unauthenticated

Managed Database Options:

  • Cloud SQL for PostgreSQL
  • Memorystore for Redis

Azure Deployment

Section titled “Azure Deployment”

Container Apps

Terminal window
# Create container app
az containerapp create \
  --name duragraph \
  --resource-group myResourceGroup \
  --image duragraph/duragraph:latest \
  --target-port 8080 \
  --ingress external

Managed Database Options:

  • Azure Database for PostgreSQL
  • Azure Cache for Redis

Infrastructure Requirements

Section titled “Infrastructure Requirements”

Minimum Requirements

Section titled “Minimum Requirements”
ComponentRequirement
CPU2 vCPUs
Memory4GB RAM
Storage20GB SSD
DatabasePostgreSQL 14+
Section titled “Recommended Production”
ComponentRequirement
CPU4+ vCPUs
Memory8GB+ RAM
Storage100GB+ SSD
DatabasePostgreSQL 14+ with replicas
NATSJetStream enabled

Environment Variables

Section titled “Environment Variables”
Terminal window
# Required
DATABASE_URL=postgresql://user:pass@host:5432/duragraph
NATS_URL=nats://localhost:4222


# Optional
PORT=8080
LOG_LEVEL=info
AUTH_ENABLED=false

Security Considerations

Section titled “Security Considerations”

Network Security

Section titled “Network Security”
  • Use private subnets for database and NATS
  • Enable TLS for all external connections
  • Configure firewall rules appropriately

Authentication

Section titled “Authentication”

DuraGraph supports multiple authentication methods:

Terminal window
# JWT Authentication
AUTH_ENABLED=true
JWT_SECRET=your-secret-key


# OAuth/OIDC (coming soon)
OAUTH_PROVIDER=google
OAUTH_CLIENT_ID=...
OAUTH_CLIENT_SECRET=...

Encryption

Section titled “Encryption”
  • In Transit: TLS 1.3 for all connections
  • At Rest: Use cloud provider’s encryption (RDS encryption, etc.)

Monitoring

Section titled “Monitoring”

Prometheus Metrics

Section titled “Prometheus Metrics”

DuraGraph exposes Prometheus metrics at /metrics:

prometheus.yml
scrape_configs:
  - job_name: 'duragraph'
    static_configs:
      - targets: ['duragraph:8080']

Health Checks

Section titled “Health Checks”
Terminal window
# Liveness probe
curl http://localhost:8080/health


# Readiness probe
curl http://localhost:8080/health

High Availability

Section titled “High Availability”

Multi-Region Deployment

Section titled “Multi-Region Deployment”
kubernetes/multi-region.yaml
apiVersion: apps/v1
kind: Deployment
metadata:
  name: duragraph
spec:
  replicas: 3
  strategy:
    type: RollingUpdate
  template:
    spec:
      topologySpreadConstraints:
        - maxSkew: 1
          topologyKey: topology.kubernetes.io/zone
          whenUnsatisfiable: DoNotSchedule

Database Replication

Section titled “Database Replication”

Configure PostgreSQL with read replicas for high availability:

Terminal window
# Primary
DATABASE_URL=postgresql://primary:5432/duragraph


# Read replicas (for queries)
DATABASE_READ_URL=postgresql://replica:5432/duragraph

Cost Optimization

Section titled “Cost Optimization”

Right-Sizing

Section titled “Right-Sizing”

Start small and scale based on actual usage:

WorkloadRecommended Instance
Developmentt3.small / e2-small
Production (small)t3.large / e2-standard-2
Production (large)m5.xlarge / n2-standard-4

Spot/Preemptible Instances

Section titled “Spot/Preemptible Instances”

Use spot instances for non-critical workloads:

# EKS spot instances
nodeSelector:
  lifecycle: Ec2Spot

Next Steps

Section titled “Next Steps”