Design Of Operation And Maintenance Platform Based On Microservice And Research On Anomaly Detection

Due to the historical structure,the operation and maintenance platform in the traditional telecommunication system is coupled with the business module,which makes the access and compatibility of the new module cumbersome,low development efficiency,and increases the difficulty of additional operation and maintenance work.At the same time,the operation and maintenance platform needs to pay attention to the log information of the service,and get abnormal data through filtering for alarm.The traditional operation and maintenance platform uses static threshold to filter data,resulting in a large number of invalid alarms in high load machines,which affects the operation and maintenance efficiency.In this paper,a new architecture pattern and different anomaly detection algorithms are used to solve the above two problems.Firstly,according to the software engineering development process,this thesis re develops the simple operation and maintenance platform system based on the core concept of service governance in the micro service architecture.Go language is used to independently develop the gateway component of non-invasive management service module,and onion structure middleware is used to realize the service governance functions such as service discovery,load balancing,heartbeat detection,permission control and flow control in the microservice architecture.A complete system under the microservice architecture is successfully built.The operation and maintenance platform has the ability to control common protocol service nodes,and designs tenant objects to make more fine-grained statistics,control and management of service traffic and permissions.Access to a new service on this operation and maintenance platform can achieve almost non-invasive at the same time of service monitoring,authority,traffic and load management,greatly improving the scalability and ease of use of the operation and maintenance system architecture level.Secondly,according to the analysis results of real telecom system server logs,the optimization of log anomaly detection is realized through three steps: eigenvalue definition,cycle mining and dynamic threshold.Using statistical eigenvalues to solve the standard of log rule measurement,combined with the idea of pheromone in ant colony algorithm,to solve the problem of high complexity and fuzzy boundary conditions encountered in periodic mining recursion.Finally,the dynamic threshold of each target interval is determined,which can filter the log and leave more effective alarm data.The operation and maintenance platform uses the anomaly detection algorithm,which can filter useless alarms in the case of high load,and provide more sensitive alarm monitoring in the case of low load,so as to improve the operation and maintenance efficiency.