基于改进PageRank的调用链异常节点定位研究

Research on Location of Abnormal Nodes in Call Chain Based on Improved PageRank

微服务、云原生是当前信息系统发展的主流方向,给信息系统带来高可用的同时也让IT系统变得前所未有的复杂,这对IT运维工作带来了巨大的挑战。机器学习正是当前时代下应对复杂系统和海量信息的可选措施。文章将讨论基于PageRank的算法在接口服务调用链上定位异常节点,并且经过测试,可在错综复杂的调用关系上实现快速准确的异常定位。

Microservices and cloud native are the mainstream development directions of current information systems. While bringing high availability to information systems, they also make IT systems more complex than ever before, which bring huge challenges to IT operation and maintenance. Machine learning is an optional measure to deal with complex systems and massive amounts of information in the current era. This paper will discuss the algorithm based on PageRank to locate the exception node on the interface service call chain,and after testing, it can realize fast and accurate exception location on the complex call relationship.