Ambar AWS Event Sourcing Application Terraform Module

A comprehensive Terraform module for deploying event sourcing applications on AWS using modern cloud-native services and infrastructure as code.

Overview

This Terraform module provides a complete infrastructure foundation for event sourcing applications. It creates and manages all necessary AWS resources, third-party integrations, and networking components required for a production-ready event sourcing architecture.

Infrastructure Components

The module creates the following infrastructure:

• VPC & Networking: Secure VPC with public/private subnets across multiple AZs, security groups, and internet gateway
• Event Store: AWS RDS PostgreSQL database for event stream storage with automated backups
• Projection Store: MongoDB Atlas cluster for read model projections with connection credentials
• Container Services: AWS ECS clusters with ALB/NLB for frontend and backend services
• Container Registry: AWS ECR repositories with GitHub OIDC integration for CI/CD
• Object Storage: S3 buckets for static assets with lifecycle policies and versioning
• Email Service: AWS SES with domain verification, DKIM, and SMTP credentials
• Domain Management: Route53 hosted zone with DNS records for domain setup
• Event Streaming: Ambar Cloud integration for real-time event transmission
• Monitoring: CloudWatch logs, metrics, and SNS alerts for system monitoring

Module Usage

Example 1: Minimal Required Configuration

module "event_sourcing_app" {
  source = "github.com/ambarltd/event-sourcing-infra-aws.git"

  # Required AWS Configuration
  region = "us-east-1"

  # Required MongoDB Atlas Configuration
  mongodbatlas_project_id = "507f1f77bcf86cd799439011"

  # Required Ambar Configuration
  destination_endpoints_to_descriptions = [
    {
      path        = "/projections/users"
      description = "User projection endpoint"
    },
    {
      path        = "/projections/orders"  
      description = "Order projection endpoint"
    }
  ]

  # Required Domain Configuration
  top_level_domain                        = "example.com"
  backend_application_domain_prefix       = "api" # Becomes api.example.com
  from_email                             = "[email protected]"
  hosted_zone_id                         = "Z1D633PJN98FT9"
  hosted_zone_name                       = "example.com"

  # Required GitHub Integration
  github_organization_with_read_write_access     = "myorg"
  frontend_github_repository_with_read_write_access = "frontend-app"
  frontend_github_branch_with_read_write_access     = "main"
  backend_github_repository_with_read_write_access  = "backend-api"
  backend_github_branch_with_read_write_access      = "main"

  # Required Application Images
  frontend_image = "some-git-hash"
  backend_image  = "some-git-hash"

  # Required Application Configuration
  ## Frontend
  frontend_application_port = 8080
  frontend_cpu_capacity     = 256
  frontend_mem_capacity     = 512
  frontend_instance_count   = 1
  ## Backend
  backend_application_port  = 3000
  backend_cpu_capacity      = 512
  backend_mem_capacity      = 1024
  backend_instance_count    = 1

  # Required Monitoring Configuration
  emails_for_alerts = ["[email protected]"]

  # Required Deployment Management
  # Event store should be configured manually or by your application before ambar resources are configured.
  event_store_configured = false
  environment_name       = "production"
}

Example 2: Full Configuration with All Parameters (Defaults Shown)

module "event_sourcing_app" {
  source = "github.com/ambarltd/event-sourcing-infra-aws.git"

  # Required AWS Configuration
  region = "us-east-1"

  # Required MongoDB Atlas Configuration
  mongodbatlas_project_id = "507f1f77bcf86cd799439011"
  mongodbatlas_free_tier  = false  # default: false

  # Required Ambar Configuration
  destination_endpoints_to_descriptions = [
    {
      path        = "/projections/users"
      description = "User projection endpoint"
    },
    {
      path        = "/projections/orders"
      description = "Order projection endpoint"
    },
    {
      path        = "/reactions/notifications"
      description = "Notification reaction endpoint"
    },
    {
      path        = "/reactions/email-triggers"
      description = "Email trigger reaction endpoint"
    }
  ]

  # Required Domain Configuration
  top_level_domain                        = "example.com"
  frontend_application_domain_prefix      = ""
  backend_application_domain_prefix       = "api"
  from_email                             = "notifications"
  hosted_zone_id                         = "Z1D633PJN98FT9"
  hosted_zone_name                       = "example.com"

  # Required GitHub Integration
  github_organization_with_read_write_access           = "myorg"
  frontend_github_repository_with_read_write_access    = "frontend-app"
  frontend_github_branch_with_read_write_access        = "main"  # default: "main"
  backend_github_repository_with_read_write_access     = "backend-api"
  backend_github_branch_with_read_write_access         = "main"  # default: "main"

  # Required Application Images
  frontend_image = "a6c58d7df3ea76c5463161eae6c201659e397ece"
  backend_image  = "a6c58d7df3ea76c5463161eae6c201659e397ece"

  # Required Frontend Configuration
  frontend_application_port = 8080
  frontend_cpu_capacity     = 256
  frontend_mem_capacity     = 512
  frontend_instance_count   = 1
  additional_frontend_domains = []  # default: []
  frontend_environment_variables = []  # default: []

  # Required Backend Configuration
  backend_application_port = 3000
  backend_cpu_capacity     = 512
  backend_mem_capacity     = 1024
  backend_instance_count   = 1
  backend_environment_variables = []  # default: []

  # Required Monitoring Configuration
  emails_for_alerts = ["[email protected]", "[email protected]"]

  # Required Deployment Management
  event_store_configured = false  # default: false (set to true after first deployment)
  environment_name       = "production"
}

Event Store Configuration & Ambar Integration

The infrastructure automatically configures a complete PostgreSQL event store with full Ambar integration for real-time event streaming. This provides a production-ready event sourcing foundation with zero manual configuration required.

Automatic Configuration

The event store module automatically creates and configures:

1. Database Schema

• Events Table: event_store with schema matching your Java application exactly:
  • id (BIGSERIAL, primary key) - Serial column for Ambar record ordering
  • event_id (TEXT, UNIQUE) - Unique event identifier
  • event_name (TEXT) - Event type name for filtering
  • aggregate_id (TEXT) - Aggregate identifier for queries
  • aggregate_version (BIGINT) - Version for optimistic concurrency control
  • json_payload (TEXT) - Event data as JSON string
  • json_metadata (TEXT) - Event metadata as JSON string
  • recorded_on (TEXT) - Event timestamp as string
  • causation_id (TEXT) - Causation tracking for event chains
  • correlation_id (TEXT) - Partitioning column for Ambar streaming

2. Performance Indexes & Constraints

• UNIQUE indexes for data integrity:
  • event_store_idx_event_aggregate_id_version on (aggregate_id, aggregate_version)
  • event_store_idx_event_id on (event_id)
• Performance indexes for fast queries:
  • event_store_idx_event_causation_id on (causation_id)
  • event_store_idx_event_correlation_id on (correlation_id) - Critical for Ambar
  • event_store_idx_occurred_on on (recorded_on)
  • event_store_idx_event_name on (event_name)

3. Ambar Streaming Configuration

• Replication User: Dedicated user with REPLICATION privileges for Ambar
• Database Privileges: CONNECT and SELECT privileges on event store
• Logical Replication Publication: ambar_publication configured for streaming
• Replication Slot: ambar_event_store_slot for reliable, ordered event delivery
• Data Source: Automatically configured Ambar data source pointing to your event store

Application Requirements

To work correctly with the pre-configured event store and Ambar streaming, your application must:

1. Follow the Exact Schema

// ✅ Correct: Use the exact column names and data types
const event = {
  event_id: uuidv4(),                    // TEXT (unique)
  event_name: 'UserCreated',             // TEXT 
  aggregate_id: userId,                  // TEXT
  aggregate_version: 1,                  // BIGINT (number)
  json_payload: JSON.stringify(data),    // TEXT (JSON as string)
  json_metadata: JSON.stringify(meta),   // TEXT (JSON as string)
  recorded_on: new Date().toISOString(), // TEXT (ISO string)
  causation_id: causationId,             // TEXT
  correlation_id: correlationId          // TEXT (REQUIRED for Ambar)
};

2. Ensure Correlation ID Consistency

The correlation_id field is critical for Ambar partitioning:

• Must be consistent for related events that should be processed in order
• Should be different for events that can be processed in parallel
• Cannot be null - Ambar uses this for data distribution

3. Use Append-Only Pattern

• Never UPDATE or DELETE events once written (Ambar requirement)
• Only INSERT new events to maintain streaming consistency
• Use aggregate_version for optimistic concurrency control

4. Handle Environment Variables

Your application receives these automatically configured variables:

// Database connection
const eventStore = new EventStore({
  host: process.env.EVENT_STORE_HOST,
  port: process.env.EVENT_STORE_PORT,
  database: process.env.EVENT_STORE_DATABASE_NAME,
  username: process.env.EVENT_STORE_USER,
  password: process.env.EVENT_STORE_PASSWORD
});

// Table names (use these constants)
const EVENTS_TABLE = process.env.EVENT_STORE_EVENTS_TABLE_NAME; // 'event_store'

Real-Time Event Streaming

Once configured, Ambar automatically streams events to your application endpoints:

1. Automatic Event Detection

• New events in the event_store table are automatically detected
• Ordered delivery based on the id column (serial)
• Partitioned delivery based on correlation_id for parallel processing

2. Destination Endpoints

Events are streamed to your configured endpoints:

// Your application should handle these HTTP POST requests
app.post('/projections/users', (req, res) => {
  const { events } = req.body; // Array of new events
  // Update your read models/projections
});

app.post('/reactions/notifications', (req, res) => {
  const { events } = req.body; // Array of new events  
  // Trigger side effects (emails, notifications, etc.)
});

Deployment Process

1. Deploy Infrastructure: Event store and Ambar are automatically configured
2. Deploy Application: Use provided environment variables to connect
3. Start Writing Events: Events are immediately streamed to your endpoints
4. Monitor Streaming: Use CloudWatch and Ambar monitoring for observability

Best Practices

Event Design

• Include correlation_id in every event for proper partitioning
• Use meaningful event_name values for filtering and monitoring
• Store rich data in json_payload for projection building
• Include causation_id for tracing event chains

Error Handling

• Handle duplicate events (Ambar provides at-least-once delivery)
• Implement idempotency in your event handlers
• Use exponential backoff for transient failures

Monitoring

• Monitor replication lag via CloudWatch metrics
• Track event processing in your application endpoints
• Set up alerts for streaming failures

PostgreSQL Configuration

• Version: PostgreSQL 15.10 (Aurora) with logical replication enabled
• Parameters: All Ambar-required parameters automatically configured
• Security: TLS encryption enforced, dedicated replication user
• Backup: Automated backups and point-in-time recovery enabled

The event store is production-ready and requires no manual configuration. Your application can immediately start writing events and receiving real-time streams through Ambar.

For advanced configuration and troubleshooting, refer to the Ambar Documentation.

Requirements

Name	Version
terraform	>= 1.0.0
aws	5.90.0*
random	>= 3.1.0
mongodbatlas	>= 1.4.0
ambar	>= 1.0.11
postgresql	>= 1.15.0

*AWS Version currently pinned due to issues with ap-southeast-5 and other regions.

Inputs

Name	Description	Type	Default	Required
region	AWS region to deploy resources	string	n/a	yes
mongodbatlas_project_id	MongoDB Atlas Project Identifier	string	n/a	yes
mongodbatlas_free_tier	If the projection store should use the M0 or M10 cluster size	bool	false	no
destination_endpoints_to_descriptions	List of destinations objects describing endpoint path and a description	list(object({path=string, description=string}))	n/a	yes
github_organization_with_read_write_access	The github organization name	string	n/a	yes
frontend_github_repository_with_read_write_access	The name of the github repo which contains the source for your frontend application	string	n/a	yes
frontend_github_branch_with_read_write_access	The name of the deployable github branch for your frontend application	string	n/a	yes
frontend_image	Frontend container image	string	n/a	yes
frontend_application_port	Frontend application port	number	n/a	yes
frontend_cpu_capacity	Frontend CPU capacity	number	n/a	yes
frontend_mem_capacity	Frontend memory capacity	number	n/a	yes
frontend_instance_count	Frontend instance count	number	n/a	yes
additional_frontend_domains	Additional frontend domains	list(string)	[]	no
top_level_domain	Domain name for frontend hosting	string	n/a	yes
frontend_application_domain_prefix	A prefix (if any) to apply to the domain for hosting the frontend portion of the application	string	""	no
frontend_environment_variables	Additional environment variables to configure for the service, beside base Ambar configs	list(object({name=string, value=string}))	[]	no
hosted_zone_id	ID of the hosted zone for the domain	string	n/a	yes
hosted_zone_name	Name of the hosted zone for the domain	string	n/a	yes
backend_github_repository_with_read_write_access	The name of the github repo which contains the source for your backend application	string	n/a	yes
backend_github_branch_with_read_write_access	The name of the deployable github branch for your backend application	string	n/a	yes
backend_image	Backend container image	string	n/a	yes
backend_application_domain_prefix	A prefix (if any) to apply to the domain for hosting the backend portion of the application	string	n/a	yes
backend_application_port	Backend application port	number	n/a	yes
backend_cpu_capacity	Backend CPU capacity	number	n/a	yes
backend_mem_capacity	Backend memory capacity	number	n/a	yes
backend_instance_count	Backend instance count	number	n/a	yes
from_email	Identity to send emails from the backend as (E.G. 'noreply' for [email protected])	string	n/a	yes
backend_environment_variables	Additional environment variables to configure for the service, beside base Ambar configs	list(object({name=string, value=string}))	[]	no
emails_for_alerts	List of email addresses for alerts	list(string)	n/a	yes
event_store_configured	If the application has been deployed at least once and successfully connected to and configured the event store for ambar use	bool	n/a	yes
environment_name	Resource name prefix for easy identification and allowing multiple template deployments to one AWS account	string	n/a	yes

Outputs

Application URLs

Name	Description
frontend_url	URL of the frontend application
backend_url	URL of the backend API

Container Registry

Name	Description
frontend_ecr_repository_url	ECR repository URL for frontend container images
frontend_github_assumable_role_read_write	GitHub assumable role for frontend ECR access
backend_ecr_repository_url	ECR repository URL for backend container images
backend_github_assumable_role_read_write	GitHub assumable role for backend ECR access

Email Service

Name	Description
ses_identity	SES domain identity ARN for email service

Environment Variables

The module automatically configures environment variables for both frontend and backend applications. These variables are injected into the ECS task definitions and are available to your applications at runtime.

Frontend Application Environment Variables

Variable Name	Description	Example Value
API_ADDRESS	Backend API endpoint for frontend to connect to	http://backend-nlb-xxx.amazonaws.com
PRODUCTION	Production environment flag	TRUE
SERVER_PORT	Port the frontend server should listen on	8080
SERVER_HOSTNAME	Server hostname binding (IPv6 wildcard)	::
DOMAIN	Comma-separated list of all frontend domains	app.example.com,www.example.com
LOAD_BALANCER	DNS name of the Application Load Balancer	frontend-alb-xxx.amazonaws.com

Backend Application Environment Variables

Event Store Configuration (PostgreSQL RDS)

Variable Name	Description
EVENT_STORE_HOST	RDS PostgreSQL instance endpoint
EVENT_STORE_PORT	Database connection port (typically 5432)
EVENT_STORE_DATABASE_NAME	PostgreSQL database name (postgres)
EVENT_STORE_USER	Database authentication username
EVENT_STORE_PASSWORD	Database authentication password
EVENT_STORE_EVENTS_TABLE_NAME	Name of the events table (event_store)
EVENT_STORE_IDEMPOTENT_REACTION_TABLE_NAME	Name of the idempotent reactions table (event_store_idempotent_reaction)
EVENT_STORE_CREATE_REPLICATION_USER_WITH_USERNAME	Username for database replication
EVENT_STORE_CREATE_REPLICATION_USER_WITH_PASSWORD	Password for database replication
EVENT_STORE_CREATE_REPLICATION_PUBLICATION	Name of the replication publication (replication_publication)

MongoDB Projection Store Configuration

Variable Name	Description
MONGODB_PROJECTION_HOST	MongoDB Atlas cluster host
MONGODB_PROJECTION_PORT	MongoDB connection port (typically 27017)
MONGODB_PROJECTION_AUTHENTICATION_DATABASE	MongoDB authentication database (admin)
MONGODB_PROJECTION_DATABASE_NAME	MongoDB database for projections (projections)
MONGODB_PROJECTION_DATABASE_USERNAME	MongoDB authentication username
MONGODB_PROJECTION_DATABASE_PASSWORD	MongoDB authentication password

SMTP Configuration (SES)

Variable Name	Description
SMTP_HOST	SES SMTP endpoint
SMTP_PORT	SMTP connection port
SMTP_USERNAME	SES SMTP authentication username
SMTP_PASSWORD	SES SMTP authentication password
SMTP_FROM_EMAIL_FOR_ADMINISTRATORS	From email address for administrative emails ([email protected])

Ambar Configuration

Variable Name	Description
AMBAR_HTTP_USERNAME	HTTP authentication username for Ambar service (8-character random string)
AMBAR_HTTP_PASSWORD	HTTP authentication password for Ambar service (16-character random password)

S3 Configuration

Variable Name	Description
S3_ENDPOINT_URL	S3 service endpoint URL
S3_ACCESS_KEY	S3 authentication access key
S3_SECRET_KEY	S3 authentication secret key
S3_BUCKET_NAME	Name of the S3 bucket for blob storage
S3_REGION	AWS region for S3 operations

Other Configuration

Variable Name	Description
FRONTEND_DOMAIN	Domain name of the frontend application

Security Notes

• Sensitive values (passwords, database credentials, API keys) are automatically generated and managed by Terraform
• The Ambar HTTP credentials are randomly generated during deployment for security
• Database and service credentials are sourced from the respective AWS services (RDS, SES, etc.)