Spring Cloud Tencent: An Overview of Its Features and Architecture

Spring Cloud and Spring Cloud Alibaba are widely used for Java microservice development, but Tencent also offers its own Spring Cloud implementation, Spring Cloud Tencent, built on the open‑source Polaris (PolarisMesh) platform.

Spring Boot + Spring Cloud remains the dominant Java framework for microservices, with Spring Cloud providing componentization, low‑configuration integration, and a standard SPI for microservice solutions such as Netflix OSS.

Polaris, open‑sourced by Tencent in 2021, delivers a one‑stop microservice suite that combines a service registry, configuration center, and governance capabilities, already serving over 90% of Tencent's internal services and handling millions of instances.

Service Registration & Discovery (Polaris Discovery) implements the Spring Cloud registration API, allowing developers to add a single dependency and instantly use Polaris as the registry. It also supports advanced features like dynamic instance isolation, weight‑based routing, and multi‑registry migration.

The Polaris model consists of Namespace, Service, and Service Instance, where namespaces isolate resources across environments or business units.

Configuration Center (Polaris Config) introduces a three‑level model—Namespace, FileGroup, and File—to manage configuration files. It follows Spring Boot’s native loading mechanism, automatically importing application.yml and profile‑specific files.

Service Routing (Polaris Router) offers metadata routing, rule‑based routing, and proximity routing. Metadata routing uses instance metadata (e.g., environment, data‑center) to direct traffic, while rule routing supports complex scenarios via expressions such as ${http.query.xxx}, ${http.header.xxx}, etc. Proximity routing selects instances based on campus, zone, or region metadata to minimize latency.

Examples of environment variables for metadata routing:

SCT_METADATA_CONTENT_ENV=dev1

SCT_METADATA_CONTENT_TRANSITIVE=ENV

Similarly, proximity routing requires setting:

SCT_METADATA_CAMPUS

SCT_METADATA_ZONE

SCT_METADATA_REGION

Rate Limiting (Polaris Ratelimit) provides both single‑instance and distributed limiting. Single‑instance limiting protects a single service instance, offering fast‑fail or leaky‑bucket queuing. Distributed limiting shares a global quota across all instances, suitable for protecting shared resources like databases or gateway services.

Circuit Breaker (Polaris Circuitbreaker) monitors call results, isolates faulty instances, and performs health‑check probes to recover services automatically.

In summary, Spring Cloud Tencent leverages Polaris to deliver a fully integrated, low‑cost, and stable microservice stack for Java enterprises, aiming to simplify technology selection, reduce operational overhead, and support a wide range of governance scenarios.

Future plans include expanding support for additional Tencent open‑source projects and continuously iterating based on user feedback.