Trend analysis method and device for monitoring indexes in monitoring system

Abstract

The invention discloses a trend analysis method and a trend analysis device for monitoring indexes in a monitoring system. The method comprises the following steps of: setting predictive analysis parameters of the indexes in advance; performing predictive analysis and periodical analysis by utilizing an acquired index value sequence respectively; in the predictive analysis, obtaining prediction results by comprehensively adopting a linear regression algorithm and an exponential regression algorithm, and externally outputting control signals according to the results of comparison between the prediction results and a preset threshold value; and in the periodical analysis, obtaining a periodical regression function according to acquired index values, and then drawing a periodical curve by using the periodical regression function. The device comprises an index value acquisition module, a predictive analysis module, a prediction alarming module, a periodical regression analysis module and a periodical curve drawing module. In the method and the device, the predictive analysis is performed by comprehensively adopting the linear regression algorithm and the exponential regression algorithm so as to achieve high accuracy; and an index variation periodical analysis function is added, so the index variation periodical curve can be obtained.

Description

technical field [0001] The invention relates to the technical field of monitoring such as network monitoring, network management, and IT operation and maintenance, and in particular to a trend analysis method and device for monitoring indicators in a monitoring system. Background technique [0002] With the continuous improvement of the information level of modern society, the scale of enterprise information technology infrastructure construction continues to expand, and network monitoring, network management, and IT operation and maintenance systems are widely used. This type of system monitors network devices, hosts, operating systems, database systems, and various application systems, collects and monitors units based on monitoring indicators, and generates alarm signals according to preset thresholds. At present, most of the monitoring systems adopt the method of "alarm after the event", that is, the current value of the indicator is compared with the threshold value set...

AI Technical Summary

Problems solved by technology

However, the system has the following disadvantages: (1) It uses a trend definition library, in which each algorithm needs to be used independently, and multiple algorithms cannot be combined with each other to make up for their own shortcomings, so the accuracy is not high; (2) Without periodic analysis, the periodic curve of indicator changes cannot be obtained

Images