Order Processing Pipeline

The order-api receives POST /orders from the front-end. It immediately returns an acknowledgement + signed URL so the UI can poll status. The heavy work (inventory check, fraud scoring, payment) is off-loaded to order-processor via SQS. The processor writes the final state to DynamoDB and pushes a custom metric OrdersProcessed.

1 Architecture

flowchart TD
    A[Frontend] -->|"POST /orders"| B["order-api (Python/FastAPI)"]
    B -->|"SQS send"| C["AWS SQS – order-queue"]
    C -->|"poll"| D["order-processor (Go)"]
    D -->|"write"| E["DynamoDB – Orders"]
    D -->|"metric"| F[Prometheus]
    C -->|"DLQ"| G["SQS DLQ"]
    B -->|"signed URL"| A

Loading

2 Features

Feature	Description
JWT Login	/login → Bearer <token>
Order API	POST /orders → 202 Accepted + signed poll URL
Async Polling	Client polls DynamoDB via signed URL
Rate Limiting	100/minute per user
Observability	Request ID, structured logs, Prometheus
Health Checks	/health, /ready (API + Processor)
OpenAPI Docs	/docs, /redoc, /openapi.json
Go Processor	Polls SQS, writes to DynamoDB, exports metrics at /metrics
Unit Tests	pytest + go test (100% coverage)

---

3 Quick Start (Local with LocalStack)

3.1 Start LocalStack

cd localstack
docker compose down -v
docker compose up -d
sleep 15   # Wait for boot

3.2 Apply Terraform

cd ../infra
#rm -rf .terraform* terraform.tfstate*  # Optional to Clean slate
terraform init
terraform apply -var-file=dev.tfvars -auto-approve

# verify queue in LocalStack
aws --endpoint-url=http://localhost:4566 sqs list-queues

3.3 Run order API Services

cd order-api
docker build -t order-api .

# Run (with LocalStack env vars)
export SQS_QUEUE_URL=$(cd ../infra && terraform output -raw sqs_queue_url)
export DDB_TABLE=$(cd ../infra && terraform output -raw dynamodb_table_name)
export AWS_ENDPOINT_URL=http://host.docker.internal:4566  # for Docker → LocalStack

docker run -p 8000:8000 \
  -e SQS_QUEUE_URL="$SQS_QUEUE_URL" \
  -e DDB_TABLE="$DDB_TABLE" \
  -e AWS_ENDPOINT_URL="$AWS_ENDPOINT_URL" \
  order-api

3.4 Run order processor

# 1. Build
cd ../order-processor
go mod tidy
docker build --no-cache -t order-processor .

# 2. Export env vars
export SQS_QUEUE_URL=$(cd ../infra && terraform output -raw sqs_queue_url)
export DDB_TABLE=$(cd ../infra && terraform output -raw dynamodb_table)
export AWS_ENDPOINT_URL=http://host.docker.internal:4566

# 3. Run (stops and removes existing container if it exists, then starts new one)
docker rm -f order-processor 2>/dev/null || true; \
docker run -d --name order-processor \
  -e SQS_QUEUE_URL="$SQS_QUEUE_URL" \
  -e DDB_TABLE="$DDB_TABLE" \
  -e AWS_ENDPOINT_URL="$AWS_ENDPOINT_URL" \
  order-processor

docker logs -f order-processor

3.5 API Verification

# 1. Login
TOKEN=$(curl -s -X POST http://localhost:8000/login \
  -H "Content-Type: application/json" \
  -d '{"user_id":"u123","amount":1}' | jq -r .access_token)

# 2. Create order
curl -X POST http://localhost:8000/orders \
  -H "Authorization: Bearer $TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"user_id":"u123","amount":999}'

# 3. Check processor health
curl http://localhost:9090/health
# → {"status":"healthy"}

# 4. Check Prometheus metrics
curl http://localhost:9090/metrics | grep orders_processor
# → orders_processed_total{env="local",status="success"} 1

3.6 API Docs & Monitoring:

• Order API Swagger: http://localhost:8000/docs
• Order API Redoc: http://localhost:8000/redoc
• Order API OpenAPI JSON: http://localhost:8000/openapi.json
• Order API Health: http://localhost:8000/health
• Order Processor Metrics: http://localhost:9090/metrics
• Order Processor Health: http://localhost:9090/health
• Order Processor Readiness: http://localhost:9090/ready

4 Test

4.1 Order API Unit Test

cd order-api
python -m venv venv
source .venv/bin/activate
pytest -v

4.2 Order Processor Unit Test

cd order-processor
go mod tidy
go build ./...
go test ./...

4.3 Smoke Test

# Start infrastructure and services
./scripts/quickstart.sh

# Run smoke tests to verify deployment
./scripts/smoke.sh

5 Follow Ups & Extensions

Feature	Implementation
Kubernetes Deployment	Deploy both order-api-py and order-processor on EKS/GKE with HPA (CPU > 70%) and PodDisruptionBudget. Use multi‑AZ for resilience.
Ingress with TLS	Use NGINX Ingress or Istio Gateway. Auto‑renew certs via Cert‑Manager + Let’s Encrypt or HashiCorp Vault.
HashiCorp Vault Integration	Store JWT_SECRET, AWS credentials, and DB secrets in Vault. Inject via Vault Agent Sidecar or Kubernetes Secrets.
Multi‑Region Active‑Active	Replicate SQS → DynamoDB Global Tables. Route traffic via Route 53 Latency Routing. Use SQS Cross‑Region Replication (via Lambda).
Observability Stack	Scrape /metrics → Prometheus → Grafana. Alert on: • rate(orders_processed_total{status="error"}[5m]) > 0 • http_request_duration_seconds > 1s
GitOps with ArgoCD	Sync k8s/ manifests from Git. Enable ApplicationSet for multi‑cluster. Auto‑sync + prune.
Canary Deployments	Use Argo Rollouts + Istio to shift 5% traffic → validate metrics → full rollout.

---

License

MIT