Discovery
☀️ Nepxion Discovery is a solution for Spring Cloud with blue green, gray, route, limitation, circuit breaker, degrade, isolation, tracing, dye, failover, active 蓝绿灰度发布、路由、限流、熔断、降级、隔离、追踪、流量染色、故障转移、多活
A Java framework on top of Spring Cloud that adds traffic splitting, blue-green deployments, rate limiting, and circuit breaking to microservice systems, letting teams release new versions to a fraction of users without restarting servers.
Nepxion Discovery is a Java framework built on top of Spring Cloud, aimed at companies that run many small backend services talking to each other (a setup often called microservices). It adds a layer of traffic control on top of what Spring Cloud already provides, letting teams control which version of a service handles which requests, without taking anything offline or restarting servers.
The core use case is called blue-green or gray deployment. When a company wants to release a new version of a service, they do not just swap the old one out. Instead, Discovery lets them send a small fraction of real traffic to the new version first, check that it behaves correctly, and then gradually move more traffic over. If something goes wrong, they can send all traffic back to the old version immediately. This is safer than a full cutover and is how large engineering teams release software with minimal risk to real users.
Beyond traffic splitting, the framework also handles rate limiting (capping how many requests per second a service can receive), circuit breaking (automatically stopping calls to a service that is failing so the failure does not spread), and request tracing (logging exactly which services handled a request as it passed through the system). There are also failover and multi-region features for teams that run their infrastructure across multiple data centers or cloud regions.
The README is written almost entirely in Chinese and is aimed at Java engineers already familiar with Spring Cloud. A companion platform and desktop UI are also mentioned, and there is a commercial version called Polaris with additional features for enterprise buyers. The open-source version covers earlier Spring Cloud releases; newer Spring Cloud versions require the paid tier.
The full README is longer than what was shown.
Where it fits
- Gradually roll out a new service version to 10% of live traffic to test it in production before a full release.
- Instantly route all traffic back to the stable version of a service if the new version shows errors, with no restarts.
- Add rate limiting to a Spring Cloud service to cap how many requests per second it handles during peak load.
- Enable distributed request tracing across a microservice mesh to see which services handled each incoming request.