Intelligent IT Operation System Based On Knowledge Graph

Artificial Intelligence for IT Operations(AIOps)is the product of the rapid development of artificial intelligence and the application of artificial intelligence in the field of operation and maintenance.Its purpose is to use machine learning algorithms to automatically learn operation and maintenance rules and data from operation and maintenance big data.Mode,and add an automatic decision-making brain on the basis of automated operation and maintenance.The current intelligent operation and maintenance can realize the functions of identifying the type of failure,root cause analysis and business flow prediction through algorithms,but lack of inference mechanisms,including logical reasoning and uncertainty reasoning,can not predict the probability and occurrence of other failures based on single point failure The specific business of the failure.In this paper,knowledge graph(KG)technology is introduced to solve the logic reasoning problem in intelligent operation and maintenance and Bayesian network is used in the uncertainty reasoning of knowledge graph.The knowledge graph can be regarded as large-scale semantic networks and displayed in a complex graph structure,providing efficient computing and reasoning capabilities.This article mainly studies the use of OWL(Web Ontology Language)to model IT Operations resource and event classes,and combined with Devops(development and operations)related tools to automatically extract resource class entities and relationships from resource management configuration database to construct resource topology maps.And use big data processing and machine learning to cluster the alarms and logs and identify the operation and maintenance events,and establish the event causality from the time sequence of the events to construct a fault event model.And by marking the conditional probabilities between events,the knowledge graph of events is transformed into a Bayesian network.The Bayesian network circular belief propagation algorithm is used to calculate the edge probability of each event in the failure model.The root cause of the event can be analyzed in conjunction with the event relationship,and the probability of a certain failure in the business can be predicted in conjunction with the business topology map.After the fault occurs,the pre-customized automation script is called according to different fault types to realize the self-healing of the fault,so as to realize the closed loop of intelligent operation and maintenance.This paper proposes an intelligent IT operation system architecture based on knowledge graphs,and constructs an IT operation knowledge graph ontology to achieve automatic identification of entities and relationships.Using Rust language,a prototype of intelligent operation and maintenance system based on knowledge graph was developed,which realized functions such as business topology query,business dependency query,fault root cause analysis and fault self-healing.After testing,it can accurately infer the root cause of the fault based on the incident alarm information and call the automated script to repair the fault.