Achieving Low‑Cost Full‑Link Gray Release with Alibaba Cloud MSE: Laidian Tech’s Journey

Background and Motivation

Since 2014, Laidian Technology has been a pioneer in the shared‑charging market, operating a nationwide network of charging stations and related services. Rapid growth led to a massive increase in device nodes and business volume, prompting a shift to containerized microservices using Spring Cloud and Dubbo.

Challenges of Deepening Microservice Adoption

The expansion introduced three major concerns: efficiency, stability, and cost. As the number of microservices grew, longer service chains caused performance bottlenecks, and without proper governance, the benefits of microservice architecture could be outweighed by operational complexity.

Benefits of Full Containerization

Simplified deployment and dramatically improved release efficiency.

Elastic scaling and resource‑on‑demand usage, reducing server costs.

Lowered operational overhead through standardized Kubernetes (ACK) management.

These advantages enabled Laidian to adopt continuous integration/continuous delivery (CI/CD) pipelines and achieve high‑availability deployments.

Why Choose Alibaba Cloud MSE

Key factors that attracted Laidian to MSE included:

Non‑intrusive integration via Java Agent bytecode enhancement, supporting all Spring Cloud and Dubbo versions without code changes.

Simple onboarding: install mse-pilot from the Alibaba Cloud Marketplace, enable namespace‑level governance, and restart services.

Comprehensive lifecycle coverage from development to production, providing visualization, observability, and automated gray release capabilities.

Core Capabilities Delivered by MSE

Visualization : Real‑time traffic views and loss‑free routing status.

Cloud‑Native Governance : Tag‑based routing and header‑based traffic segmentation compatible with Kubernetes and Ingress.

Zero‑Loss Deployment : End‑to‑end protection for seamless service upgrades.

Full‑Link Gray Release Scenarios

Laidian identified three practical scenarios:

Scenario 1 – Tag‑Based Traffic Coloring

When a request passes through a tagged node, subsequent calls preferentially select nodes with the same tag, enabling automatic flow coloring across the entire call chain.

Scenario 2 – Header‑Based Routing

Clients add a specific environment header; the ingress layer forwards the request to the corresponding environment gateway, which isolates traffic for that environment.

Scenario 3 – Custom Routing Rules

By propagating a custom header through the call chain, services can apply bespoke routing rules to direct traffic to gray‑release instances.

Scenario 1 satisfies most customer needs, while Scenarios 2 and 3 provide advanced options.

Implementation Details

The architecture places an Nginx gateway at the edge, routing 10 % of traffic to a gray‑release environment and 90 % to the production environment. MSE automatically tags gray traffic, ensuring end‑to‑end isolation. If gray nodes are unavailable, traffic falls back to the production path.

Service Pre‑Warm Capability

MSE offers an agent‑based, non‑intrusive pre‑warm method that gradually ramps up traffic during high‑load periods, preventing cold‑start overload and protecting service availability.

Future Outlook

Beyond gray release and loss‑free upgrades, Laidian anticipates broader governance needs covering release, runtime, fault diagnosis, and recovery. MSE’s extensible platform aims to continuously boost development efficiency and maintain a 99.9 % service availability target.