Deployment

Docker Compose

The recommended deployment method runs all services as Docker containers.

Full Stack

cd redirect
docker compose up -d

This starts:

• Infrastructure: MongoDB 7, ClickHouse, Redis 7, MinIO, Fluvio (SC + SPU), RabbitMQ, gglsbl-rest (Safe Browsing)
• Init containers: MinIO bucket setup, Fluvio topic creation, ClickHouse migrations
• Application services: click-router, click-router-api, click-tracker, click-aggregator, click-aggregator-api, route-verifier, route-icon-worker, domain-verifier, domains

Management API

cd api
docker compose up -d

Starts the .NET API and its PostgreSQL database.

UI

cd ui
docker compose up -d

Starts the dashboard and landing page. This compose file includes the redirect compose file via the include directive.

Service Dependencies

click-router         → mongo, redis, clickhouse, fluvio (topics), rabbitmq, domains
click-router-api     → mongo, redis, rabbitmq
click-tracker        → redis, fluvio (topics)
click-aggregator     → clickhouse, fluvio (topics)
click-aggregator-api → clickhouse (after migrations)
route-verifier       → mongo, rabbitmq, gglsbl-rest
route-icon-worker    → rabbitmq, minio
domain-verifier      → mongo, rabbitmq
domains              → (standalone)
management-api       → postgresql, elasticsearch, click-router-api, click-aggregator-api
dashboard            → management-api

Environment Configuration

Rust Services

Configuration is loaded from TOML files in each service’s config/ directory. The APP_RUN_MODE environment variable selects the profile:

• development — local defaults with debug logging
• production — production-optimized settings

Docker containers default to APP_RUN_MODE=production.

Management API

Configured via appsettings.json and environment variables:

Docker Images

Multi-stage builds minimize image sizes:

• Rust services: Built from the workspace Dockerfile, producing a Debian slim image with a non-root user
• .NET API: Three-stage build (base, build/publish, runtime) targeting mcr.microsoft.com/dotnet/aspnet:9.0
• UI apps: Node-based builds outputting static assets

Health Checks

All services expose health endpoints. Docker Compose healthchecks monitor service status automatically:

make ps    # Show service status and health
make logs  # Tail service logs

Networking

All services share the shortas-net Docker bridge network. Inter-service communication uses container hostnames (e.g., mongo, clickhouse, click-router-api).

AWS Deployment

For production deployments, Shortas can be deployed to AWS using Terraform. This replaces local services with AWS managed equivalents for scalability and reliability.

Service Mapping

Prerequisites

• AWS CLI configured with appropriate credentials
• Terraform >= 1.5.0
• Docker (for building images)
• Domain name with Route53 hosted zone (optional)

Quick Start

# Initialize Terraform
cd infra/aws/terraform/environments/dev
terraform init

# Configure variables
cp terraform.tfvars.example terraform.tfvars
# Edit terraform.tfvars with your settings

# Deploy infrastructure
terraform apply

# Build and push Docker images
cd ../scripts
./build-push-images.sh dev

# Deploy services to ECS
./deploy-services.sh dev

Environments

Two environments are provided:

Authentication with AWS Cognito

AWS deployments use Cognito instead of Keycloak for authentication:

• User Pool with email-based sign-up/sign-in
• Dashboard app client (public SPA with PKCE)
• API app client (confidential with client credentials)
• Configurable MFA and password policy
• Hosted UI for sign-in/sign-up

Configure the dashboard with Cognito endpoints:

REACT_APP_COGNITO_ISSUER_URL=https://cognito-idp.us-east-1.amazonaws.com/us-east-1_xxxxxxxx
REACT_APP_COGNITO_CLIENT_ID=xxxxxxxxxxxxxxxxxxxxxxxxxx

Configure the .NET API with appsettings.Aws.json:

{
  "Authentication": {
    "Provider": "Cognito",
    "Cognito": {
      "UserPoolId": "${COGNITO_USER_POOL_ID}",
      "Authority": "${COGNITO_ISSUER_URL}"
    }
  }
}

Monitoring

CloudWatch provides:

• Log aggregation for all ECS services
• Custom metrics dashboards
• Alarms for error rates, latency, resource usage
• SNS notifications for critical alerts

Terraform Outputs

After deployment, retrieve important values:

terraform output

# Outputs include:
# - cognito_user_pool_id
# - cognito_hosted_ui_url
# - alb_dns_name
# - ecr_repository_urls
# - dynamodb_tables

See infra/aws/terraform/README.md for detailed documentation.