A sample end‑to‑end service showing how to build, deploy, and operate a Go‑based money‑transfer microservice on Google Cloud:

• gRPC API for processing transfers
• Cloud Pub/Sub for asynchronous audit‑event publishing
• Cloud SQL (PostgreSQL) for ledger persistence
• GKE (Kubernetes) to run the service behind a LoadBalancer
• Terraform for infrastructure as code
• GitHub Actions + Docker for CI/CD

1. gRPC Definition & Codegen
   • transfer.proto with go_package option → Go stubs generated under go-server-demo/transfer/
2. Go Transfer Service (cmd/server)
   • ProcessTransfer handler does currency conversion stub, ACID DB transaction, and publishes audit events to Pub/Sub
3. Pub/Sub Infrastructure (terraform/pubsub.tf)
   • Topic audit-events, DLQ topic audit-events-dlq, subscription audit-events-sub with dead‑letter policy and 7 day retention
4. GKE Deployment & Service (terraform/deployment.tf)
   • Kubernetes Deployment (2 replicas) listening on port 50051 + LoadBalancer Service exposing port 50051
5. Cloud SQL (terraform/cloud_sql.tf)
   • PostgreSQL 17 instance, database ledger, user transfer_user on a supported db-custom-1-3840 tier
6. CI/CD Pipeline (.github/workflows/ci-cd.yml)
   • Regenerates protobuf code, builds & pushes Docker image to GCR, runs Terraform to provision infra and deploy

We set out to build a high‑throughput money‑transfer platform leveraging Go’s concurrency, crypto rails, and event‑driven audit logs. On deck:

1. Massive Parallel Transfers

   • Use Go goroutines and channels to fan out thousands of concurrent transfers per second.
   • Implement worker pools that back‑pressure on Pub/Sub backlog and coordinate idempotent retries.

2. Crypto‑Backend Integration

   • Extend ProcessTransfer to route funds over a blockchain rail:
     • Sign transactions via GCP KMS or HSM
     • Broadcast on‑chain, await confirmations
     • Convert between fiat ↔ crypto using real‑time oracles before ledger commit

3. Advanced Audit & Analytics

   • Build a dedicated subscriber (e.g. in Go or .NET) that consumes audit-events-sub and persists to TimescaleDB or BigQuery.
   • Add dashboards and alerts on transfer volumes, failures, and anomalous patterns.

4. Resiliency & Auto‑Scaling

   • Configure Horizontal Pod Autoscalers based on custom Pub/Sub metrics or gRPC latency.
   • Chaos‑test failure paths (DB timeouts, Pub/Sub nacks) and verify self‑healing.

5. Multi‑Environment & GitOps

   • Split Terraform into dev, staging, and prod workspaces or use Terragrunt
   • Adopt a GitOps tool (Argo CD) so infrastructure changes self‑deploy on PR merge

• Go ≥ 1.24
• protoc compiler
• gcloud SDK
• Terraform ≥ 1.5.0
• Docker

cd go-server-demo
go install google.golang.org/protobuf/cmd/protoc-gen-go@latest
go install google.golang.org/grpc/cmd/protoc-gen-go-grpc@latest
mkdir -p transfer
protoc \
  --go_out=transfer --go_opt=paths=source_relative \
  --go-grpc_out=transfer --go-grpc_opt=paths=source_relative \
  transfer.proto

# (Optional) start Cloud SQL Proxy if using Cloud SQL:
#   cloud-sql-proxy YOUR_INSTANCE_CONN=tcp:5432

export GOOGLE_APPLICATION_CREDENTIALS="../gcp-demo-460104-*.json"
export LEDGER_DSN="postgres://transfer_user:<PASSWORD>@127.0.0.1:5432/ledger?sslmode=disable"

cd go-server-demo
go run cmd/server

cd go-server-demo
go run cmd/client --addr localhost:50051
4. Deploy via CI/CD
Push to main and let GitHub Actions:

Regenerate protobuf

Build & push Docker image

Provision infra with Terraform

Deploy to GKE

5. Inspect the live service

gcloud container clusters get-credentials gcp-demo-cluster \
  --region us-central1 --project YOUR_PROJECT_ID

kubectl get svc go-server-demo -o wide

grpcurl -plaintext <EXTERNAL_IP>:50051 list

docker build -t gcr.io/gcp-demo-460104/go-server-demo:v2 .
docker push gcr.io/gcp-demo-460104/go-server-demo:v2

pull messages command

gcloud pubsub subscriptions tail audit-events-sub \
  --project=gcp-demo-460104 \
  --format="json(message.message.data)"