From Monolith to Microservices: A Real‑World Online Supermarket Migration Story

Initial Requirements

A few years ago, Xiao Ming and Xiao Pi launched an online supermarket with a basic website for browsing and purchasing products and a separate admin backend for managing users, items, and orders.

Growing Business and Early Problems

Rapid competition forced them to add promotions, mobile apps, and data analysis, leading to duplicated code, tangled interfaces, blurred service boundaries, performance bottlenecks, and painful deployments.

First Refactoring: Shared Services

They abstracted common capabilities into five public services—User, Product, Promotion, Order, and Data‑Analysis—allowing each application to call these services instead of duplicating logic.

Database Splitting and Messaging

To eliminate the shared‑database bottleneck, each service received its own database, and a message‑queue was introduced for real‑time communication.

Monitoring

Each component exposes a uniform /metrics endpoint; Prometheus scrapes these metrics and Grafana visualises them with alerts for memory, latency, error rates, etc.

Link Tracing

Using Zipkin (an implementation of Google’s Dapper), they inject traceId, spanId, parentId, and timestamps into HTTP headers to reconstruct request call trees.

Log Analysis

ELK (Elasticsearch, Logstash, Kibana) is deployed so logs from all services are indexed, searchable, and visualised, enabling fast root‑cause analysis.

Gateway and Service Governance

A central API gateway enforces authentication, provides a unified API catalogue, and can be applied per domain or globally.

Service Registration & Discovery

Services register themselves to a discovery system (e.g., Consul/Eureka) and obtain peer addresses automatically, supporting dynamic scaling.

Circuit Breaker, Degradation & Rate Limiting

When a downstream service fails repeatedly, the circuit breaker opens to return immediate errors; non‑critical features are degraded, and rate limiting protects services from overload.

Testing Strategy

End‑to‑end tests cover core user flows.

Service‑level tests validate each API.

Unit tests ensure code correctness.

Mock servers are used to isolate dependencies during service testing.

Microservice Framework vs. Service Mesh

They built a lightweight framework to handle metrics, tracing, logging, and registration, but faced upgrade coordination challenges. As an alternative, a sidecar‑based Service Mesh (data plane + control plane) provides non‑intrusive traffic management, though it adds memory overhead.

Conclusion

The migration demonstrates that microservices improve modularity and scalability but introduce new operational complexity; continuous monitoring, automated discovery, and disciplined governance are essential for success.