Abnormity prediction method and device in full-link monitoring system

Abstract

The invention discloses an abnormity prediction method in a full-link monitoring system, which comprises the following steps of: when current monitoring data of a target system are received, carryingout dimensionality reduction and cleaning operation on the current monitoring data to obtain target monitoring data; and transmitting the target monitoring data to an operation and maintenance management and control model trained by adopting a preset training method for prediction to obtain an abnormal prediction result of the target system in the next time interval, and displaying the abnormal prediction result to operation and maintenance personnel. According to the prediction method, the abnormal state of the target system in the next time interval can be predicted according to the currentmonitoring data, and the abnormal state of the target system in the next time interval is analyzed from the perspective of pre-event analysis.

Description

technical field [0001] The invention relates to the technical field of maintenance, in particular to an abnormality prediction method and device in a full-link monitoring system. Background technique [0002] System operation and maintenance is now a very complicated task. In addition, the management of application software such as systems and databases has increased the requirements for the level of operation and maintenance personnel. The rise in IT system operation and maintenance costs is obvious, and intelligent operation and maintenance Currently in the exploratory stage, traditional operation and maintenance personnel have the following two ways to implement the performance and link call monitoring of business applications: [0003] (1) Analyze the abnormal log by printing the log; [0004] (2) Passively analyze error stacks and Dump files. When anomalies have occurred and bottlenecks have occurred, the system should be expanded. In severe cases, the application syst...

AI Technical Summary

Problems solved by technology

[0007] In view of this, the present invention provides an abnormality prediction method and device in a full-link monitoring system to solve the problem that the operation and maintenance of the system in the prior art belongs to post-event analysis. With the increase of IT basic platforms, the complexity of the system and The diversity is constantly changing, and there is a lack of prediction of the future trend of the system, which makes it impossible to analyze the positioning problem from the perspective of prior analysis

Images