基于服务网格的微服务故障治理  

Microservice Fault Management Based on Service Mesh

作　　者：赵鑫[1,2] 李铭轩[2] Zhao Xin;Li Mingxuan(Beijing University of Posts and Telecommunications,Beijing 100876,China;China Unicom Research Institute,Beijing 100048,China)

机构地区：[1]北京邮电大学,北京100876 [2]中国联通研究院,北京100048

出　　处：《信息通信技术》2021年第4期50-58,共9页Information and communications Technologies

摘　　要：微服务以其组件之间松耦合、易于扩展、组件内自治性强等特点,成为各大公司分布式服务新的解决方案。但微服务架构复杂,容易发生故障,而且故障与微服务实例之间动态交互相关,难以重现。并且微服务集群稳定性差,易导致整个集群崩溃。为了应对这些挑战,文章提出一套基于服务网格的微服务故障治理方案,选出合适的服务网格软件,使用服务网格提供的分布式追踪、日志收集、故障注入等功能,进行故障治理,并应用在一套微服务软件上。结果证明,这套方案能够及时定位故障位置,减小故障带来的不良影响。Due to the characteristics of loose coupling between components,good expansibility and strong autonomy within components,microservice has become a new solution for distributed service among the big companies.However,the architecture of microservices is complex and fault-prone.And the failures are related to dynamic interactions between microservice instance which are difficult to reproduce.What’s more,the stability of the microservice cluster is poor and a failure is easy to cause the entire cluster to crash.To overcome these challenges,this paper proposes a microservice fault management solution based on service mesh and selects the appropriate service mesh software.The distributed tracing,log collection and fault injection provided by service mesh has been used for fault management and this solution has been applied to a software which is built based on microservice.The result shows that the scheme can locate the fault in time and reduce the adverse effects of the fault.

关 键 词：微服务 服务网格 故障治理 

分 类 号：TP311.5[自动化与计算机技术—计算机软件与理论]