OrderFlow — Event-Driven Microservices Platform on AWS

Took an empty folder and turned it into a live AWS microservices platform. No templates. No shortcuts.

---

What Is This?

OrderFlow is a production-grade, event-driven microservices platform deployed entirely on AWS. It was built from scratch to demonstrate real cloud engineering skills — not concepts, not tutorials, but live running infrastructure provisioned with zero manual console interaction.

Every resource in this project — VPC, subnets, ECS cluster, ALB, ECR, SQS queues, SNS topic, EventBridge rules, CloudWatch alarms, X-Ray tracing — was created and managed through modular Terraform IaC.

---

Architecture

Event Flow

Client Request
      │
      ▼
Application Load Balancer (port 80)
      │
      ▼
ECS Fargate Cluster
      │
      ├── order-service     (Flask · port 5000 · /orders/*)
      ├── inventory-service (Flask · port 5001 · /inventory/*)
      └── notification-service (Flask · port 5002 · /notify/*)
      │
      ▼
order.created event fired
      │
      ▼
Amazon EventBridge
      │
      ├──► SQS inventory-queue   (direct target)
      │
      └──► Amazon SNS orders-topic
                  │
                  └──► SQS notification-queue (fan-out)

---

Tech Stack

Layer	Technology
Compute	Amazon ECS Fargate
Networking	AWS VPC, ALB, public/private subnets, IGW
Containers	Docker, Amazon ECR
Messaging	Amazon EventBridge, SQS, SNS
CI/CD	GitHub Actions (OIDC — zero hardcoded keys)
Observability	CloudWatch Dashboard, Metric Alarms, AWS X-Ray
IaC	Terraform (modular)
Language	Python 3.11, Flask

---

Microservices

Service	Port	Route	Responsibility
order-service	5000	/orders/*	Accepts orders, fires EventBridge events
inventory-service	5001	/inventory/*	Consumes SQS inventory queue
notification-service	5002	/notify/*	Consumes SQS notification queue via SNS

Health Check

curl http://orderflow-alb-1358729812.us-east-1.elb.amazonaws.com/health
# {"service":"order-service","status":"healthy"}

---

Infrastructure

Terraform Module Structure

terraform/
├── environments/
│   └── primary/               # us-east-1 deployment
│       ├── main.tf            # Module orchestration
│       ├── variables.tf
│       └── terraform.tfvars
└── modules/
    ├── networking/            # VPC, subnets, route tables, SGs, IGW
    ├── ecs/                   # Fargate cluster, task defs, services, ECR, ALB
    ├── messaging/             # SQS, SNS, EventBridge rules and targets
    └── observability/         # CloudWatch dashboard, alarms, X-Ray group

Deploy from Scratch

git clone https://github.com/aboodi679/orderflow-aws.git
cd orderflow-aws/terraform/environments/primary
terraform init
terraform apply

Requires: AWS CLI configured, Terraform >= 1.6.0

Tear Down (cost saving)

terraform destroy

---

CI/CD Pipeline

Each microservice has its own GitHub Actions workflow triggered on path-specific pushes.

Push to services/order-service/**
          │
          ▼
GitHub Actions (ubuntu-latest)
          │
          ├── Configure AWS via OIDC (no stored credentials)
          ├── Login to Amazon ECR
          ├── Docker build + tag (${{ github.sha }} + latest)
          ├── Push to ECR
          └── Force ECS rolling redeploy

OIDC means zero AWS access keys stored in GitHub Secrets — the pipeline assumes an IAM role directly via web identity token. Enterprise-grade security from day one.

---

Observability

CloudWatch Dashboard

Live metrics tracked:

• ECS CPU utilization (all 3 services)
• ECS Memory utilization (all 3 services)
• ALB request count
• ALB target response time
• SQS messages sent/received (inventory queue)
• SQS messages sent/received (notification queue)

Alarms

Alarm	Threshold	Action
ECS CPU high (×3 services)	> 80% for 2 periods	Alert
SQS inventory queue depth	> 100 messages	Alert
SQS notification queue depth	> 100 messages	Alert

X-Ray

Distributed tracing configured with custom sampling rule (5% fixed rate) and service group filter for the orderflow namespace.

---

Screenshots

ECS Services — All Running

CloudWatch Alarms — All OK

EventBridge Rule — Enabled

SQS Queues — Live with Messages

GitHub Actions CI/CD — Green

---

Event-Driven Test

Manually fire an order.created event to verify the full messaging pipeline:

aws events put-events --entries file://test-event.json --region us-east-1

[
  {
    "Source": "orderflow.order-service",
    "DetailType": "order.created",
    "Detail": "{\"order_id\":\"ORD-001\",\"item\":\"laptop\",\"qty\":1}",
    "EventBusName": "default"
  }
]

Verify message arrived in SQS:

aws sqs receive-message \
  --queue-url https://sqs.us-east-1.amazonaws.com/026243800492/orderflow-inventory-queue \
  --region us-east-1

---

Cost

Estimated monthly cost with all resources running: ~$22–25/month

Service	Est. Cost
ECS Fargate (3 services, minimal config)	~$8–10
ALB	~$8
DynamoDB / SQS / SNS / EventBridge	~$1–2
CloudWatch + X-Ray	~$2–3
ECR + S3	~$1

Tip: Run terraform destroy when not actively using to eliminate compute costs.

---

Author

Muhammad Abdullah Fresh Software Engineering graduate | Cloud & DevOps Engineer

---

Related Projects

• AI Hitman — Cloud Game Backend — Flask REST API on AWS Lambda, DynamoDB, S3, API Gateway, Terraform
• AWS 3-Tier Web Architecture — ALB, EC2, RDS MySQL, modular Terraform, multi-AZ