Dynamic threshold analysis method for operation index of IT system

Abstract

The invention relates to a dynamic threshold analysis method for an operation index of an IT (Information Technology) system. The method comprises the following steps of: 1, inquiring historical change conditions of the operation index; 2, calculating a normal value interval R of the index under the current time granularity on the basis of a normal distribution principle, wherein the upper limit of the normal value interval is marked as an RU, and the lower limit of the normal value interval is marked as an RD; 3, connecting the RUs of all time granularities to be used as a dynamic threshold upper base line, and connecting the RDs to be used as a dynamic threshold lower base line; and 4, judging newly collected operation indexes under the current time granularity, wherein if the newly collected operation index is greater than the RU, the condition is the too-high-value abnormity, and if the newly collected operation index is lower than the RD, the condition is the too-low-value abnormity. The dynamic threshold analysis method has the beneficial effects that defects of a static threshold are overcome; and the historical operation rule of the operation index is intelligently studied, the normal value range of the operation index in different time points is calculated, and the dynamic threshold base lines are generated to be used for the monitoring and the analysis of the operation index. The dynamic threshold analysis method has the beneficial effects that mass manual operation is not needed; the calculation quantity is small; and the analysis is accurate, so that the operation and maintenance labor cost can be greatly reduced, and the monitoring coverage range and the accuracy are improved.

Description

technical field [0001] The invention relates to the field of IT operation and maintenance, and mainly relates to a dynamic threshold analysis method for an IT system operation index. Background technique [0002] IT system operation status monitoring is an indispensable part of the development of modern informatization construction. It provides operation and maintenance personnel with 7x24 hours of IT system operation status information, real-time monitoring of fault points and timely and accurate early warning, which is very important for reducing IT System operation and maintenance costs, to achieve its long-term stable operation is of great help. [0003] The operating status of an IT system is often reflected in its various software and hardware operating indicators, including the underlying network traffic to the high-level business transaction volume. After these operating indicators are collected, further analysis is required to determine the operating health status ...

AI Technical Summary

Problems solved by technology

[0004] Traditional operating index analysis techniques often use artificially established static thresholds for monitoring and analysis, which has human experience errors

And a large-scale IT system can contain as many as tens of thousands of various software and hardware indicators, so it is even more impossible to use static rules to artificially configure reasonable static rules for these indicators

Images