|
|
5 months ago | |
|---|---|---|
| .. | ||
| lib | 5 months ago | |
| .gitignore | 5 months ago | |
| .npmignore | 5 months ago | |
| README.md | 5 months ago | |
| app.js | 5 months ago | |
| cdk.context.json | 5 months ago | |
| cdk.json | 5 months ago | |
| package-lock.json | 5 months ago | |
| package.json | 5 months ago | |
| userdata.txt | 5 months ago | |
README.md
CP Docuseal CDK Infrastructure
This directory contains AWS CDK v2 infrastructure code for deploying the CP Docuseal app. At present it only deploys to development and staging; production will follow.
Architecture Overview
This infrastructure is basically nicked wholesale from the Integration Station application, just dialed down a notch as our needs are a bit less than theirs.
- Internal Application Load Balancer (ALB) - Routes traffic to ECS services within the VPC
- Amazon ECS Cluster - Runs containerized applications on EC2 instances (ARM64 t4g.small)
- ECS Service - Manages container deployment and scaling
- ECS Task Definition - Defines container configuration and resource requirements
- CloudWatch Logging - Centralized logging for application monitoring
- Security Groups - Network security controls
- ECR Integration - Uses existing "integration-station" ECR repository
Prerequisites
- AWS CLI configured with appropriate permissions
- Node.js (version 22 or later)
- AWS CDK v2 installed globally:
npm install -g aws-cdk - Existing AWS Infrastructure:
- VPC with public and private subnets
- ECR repository named "cp-docuseal"
- Appropriate IAM permissions for CDK deployment
Setup
-
Install dependencies:
cd cdk_deploy npm install -
Bootstrap CDK (application setup ONLY):
npm run bootstrap -
Update VPC and Subnet IDs: Edit
app.jsand replace the placeholder IDs with your actual VPC and subnet IDs:vpcId: 'vpc-your-actual-vpc-id', privateSubnetIds: ['subnet-your-private-1', 'subnet-your-private-2'], publicSubnetIds: ['subnet-your-public-1', 'subnet-your-public-2']
Environment Configuration
The infrastructure supports three environments with different resource allocations:
Development
- Instances: 1 ECS instance
- CPU: 512 units
- Memory: 1024 MB
Staging
- Instances: 1 ECS instance
- CPU: 512 units
- Memory: 1024 MB
Production
- N/A
Deployment
Deploy to Development
npm run deploy:dev
Deploy to Staging
npm run deploy:staging
Deploy to Production - NOT YET SUPPORTED
npm run deploy:prod
View CloudFormation Template
npm run synth
Compare Changes
npm run diff
Cleanup
Destroy Development Environment
npm run destroy:dev
Destroy Staging Environment
npm run destroy:staging
Destroy Production Environment - NOT YET SUPPORTED
npm run destroy:prod
Important Notes
-
Internal ALB: The Application Load Balancer is configured as internal-only and deployed in private subnets for security.
-
ARM64 Instances: The ECS cluster uses t4g.small ARM64 instances for cost efficiency. Ensure your container images are built for ARM64 architecture.
-
Health Checks: The ALB target group is configured to perform health checks on
/healthendpoint. Make sure your application responds to this endpoint. -
Logging: All ECS tasks automatically log to CloudWatch under
/ecs/cp-docuseal-{environment}log groups. -
Security: Security groups are configured to allow:
- ALB: HTTP (80) and HTTPS (443) traffic
- ECS: Traffic from ALB on port 3000
Troubleshooting
-
VPC Lookup Issues: Ensure the VPC IDs and subnet IDs in
app.jsare correct and exist in your AWS account. -
ECR Repository: Verify that the "cp-docuseal" ECR repository exists and contains the required Docker images.
-
Permissions: Ensure your AWS credentials have sufficient permissions for:
- EC2 (VPC, Security Groups, Launch Templates)
- ECS (Clusters, Services, Tasks)
- ELB (Application Load Balancers, Target Groups)
- CloudWatch (Log Groups)
- IAM (Roles and Policies)
-
Container Health: If services fail to start, check CloudWatch logs for container startup issues.
Customization
You can modify the following aspects:
- Instance Types: Change
ec2.InstanceClass.T4Gandec2.InstanceSize.SMALLin the stack - Container Port: Update port mappings if your application uses a different port
- Resource Limits: Adjust CPU and memory allocations in the environment configurations
- Auto Scaling: Modify
minCapacityandmaxCapacityfor different scaling behaviors
Stack Outputs
After deployment, the stack provides:
- ALB DNS Name: Internal DNS name for the Application Load Balancer
- ECS Cluster Name: Name of the created ECS cluster
- ECS Service Name: Name of the ECS service