How a Logistics SaaS Company Scaled to Millions Using Cloud‑Native Microservices

Background

HaiGuanJia, a logistics SaaS provider founded in 2015 and based in Suzhou, grew to serve over one million customers worldwide within eight years. The company’s early investment in AI, big data, and cloud computing allowed it to launch innovative products for ports, shipping lines, and freight forwarders, including visual dashboards, electronic document transmission, and online customs services.

Business Challenges

Rapid user registration caused explosive growth in usage scenarios, creating three major problems:

Slower business iteration and reduced development efficiency due to a traditional monolithic architecture.

Mismatch between business needs and technical architecture, especially the lack of gray‑release capability.

Increasing demand for product innovation and faster iteration under a rapidly changing market.

These issues forced the R&D team to seek a more agile, stable, and scalable solution.

Cloud‑Native Transformation

HaiGuanJia adopted a cloud‑native approach from the start, deploying front‑end web containers, gateways, back‑end microservices, and big‑data components on Kubernetes clusters. Cloud‑native principles—containerization, microservices, continuous delivery—provided loose coupling, distributed resilience, and business‑centric agility.

Service Mesh Selection

After extensive research, the team chose Tencent Cloud Polaris Mesh for its strong control plane, non‑intrusive integration, high stability, rich observability, hybrid‑cloud support, and Kubernetes compatibility. Polaris Mesh provides service management, traffic management, fault tolerance, configuration management, and observability.

Microservice Framework

To unify the technology stack, HaiGuanJia selected Spring Cloud Tencent as the primary framework, integrating its own Spring Cloud + Dubbo base services with Polaris Mesh. This standardization reduced inter‑team communication overhead and multiplied development efficiency.

Environment Isolation Strategy

To avoid conflicts among multiple teams, HaiGuanJia implemented a baseline‑environment and feature‑environment model. The baseline environment provides a stable, complete set of services, while feature environments deploy only the services under test. Traffic routing rules direct requests to the appropriate environment, falling back to the baseline when needed.

Baseline Environment : Full, stable set of services.

Feature Environment : Temporary environment containing only changed services.

Gray Release Implementation

Polaris Mesh supports full‑link gray releases. Requests from a gray domain are routed to gray‑version services via the gateway, which adds the header X-Polaris-Metadata-Transitive-featureenv=v2. If a gray service is unavailable, traffic falls back to the stable version. The approach also extends to databases, caches, and message queues.

Smooth Migration of Legacy Systems

Legacy applications running on VMs were integrated into the Kubernetes‑based mesh using a Java Agent that injects sidecar‑like functionality, allowing both old and new services to be managed uniformly.

Outcomes and Future Direction

The cloud‑native overhaul stabilized the platform, accelerated feature delivery, and reduced operational costs. HaiGuanJia plans to continue deepening its cloud‑native capabilities, collaborating with Tencent Cloud’s native team to further enhance value for customers.