Design And Implementation Of Operation And Maintenance Management System Of Microservice Platform

In recent years,with the rapid development of the Internet and cloud computing technologies,enterprise applications have been continuously updated and iterated,becoming more and more complex.The monolithic architecture has been unable to meet the needs of enterprises.The proposal of the microservice architecture broke the traditional software architecture development pattern,and quickly set off a wave of microservices at home and abroad.However,the rapid increase in the number of services and the continuous complexity of technical components have also brought great challenges to the operation and maintenance of the system.Problems such as traditional methods are no longer suitable,and faults are difficult to troubleshoot,and the management of microservices has become extremely complicated,which greatly hinders the process of enterprise microservices.In view of the above problems,this paper designs and implements a operation and maintenance management system of microservice platform.This system not only provides service discovery and service registration capabilities,but also can manage the life cycle of microservices.Besides,it implements microservice governance functions,including multiversion control,fault-tolerant strategies,etc.,and abstracts functions unrelated to business logic to the foundation facility layer to reduce the complexity of the microservice code,the operational complexity of governance,and the need for underlying network functions.In addition,a node agent that monitors and reports the status of cluster nodes is designed and implemented.Based on the Open Tracing framework and Envoy,a distributed call tracing agent with almost no intrusion to microservices is implemented.The data collected by the agent is used as the Jaeger backend,and Jaeger is used for call analysis,which can enhance the monitorability of the microservice system.Finally,an anomaly detection algorithm based on dynamic weighting of node characteristics is proposed,which is more suitable for the microservice scenario.The data collected by the node monitoring agent is used to periodically perform health detection on the nodes in the cluster to ensure the reliability of the microservice system.The main work of this paper is as follows:(1)Demand analysis.According to the current pain points in microservices and situation at home and abroad,the research content of this paper is extracted.The module division of this system is determined,which is mainly composed of five modules,namely,the microservice life cycle management module,the microservice governance module,the monitoring module,the node health detection module and the system management module.The users of the system are defined,and the role permissions are determined.The detailed requirements of each module are analyzed,and the non-functional requirements of the system are described at the same time.(2)Design and implementation.Combined with the requirement analysis of the system,the boundary of the system and the data flow with the outside system are determined.The architecture of the system is designed.The functions of the modules are divided,and the design of the core functions of each module is described using the class diagram and sequence diagram in the modeling language and implementation logic.In addition,the basis and ideas for the improvement of the isolated forest anomaly detection algorithm are expounded.Finally,the attributes of the entities in the system and the relationship between them are analyzed,and the database table is designed accordingly.(3)Test and result analysis.After completing the coding of the system,the system is tested.A test environment was built.Test cases were designed,and the main functions of the system were tested in detail according to the steps of the test cases;non-functional requirements such as system performance and compatibility are also tested;load simulation tools are used to generate abnormal data,and based on this,the improved isolated forest anomaly detection algorithm is compared and verified by experiments,which are in line with the expected results.Experiments show that the system runs stably,which ensures the reliable operation of the microservice system in a distributed environment,and greatly reduces the complexity and learning threshold of service operation and maintenance management.Besides,it provides service governance capabilities,and monitoring capabilities also meet the needs of use.After experimental verification,the performance of the improved anomaly detection algorithm works very well.In conclusion this system provide a powerful tool in the process of microservices.