Abnormal load detection method for cloud calculation on-line business

Abstract

The invention relates to an abnormal load detection method for cloud calculation on-line business, and belongs to the technical field of cloud calculation application. According to the method, information data of various load items of all hosts of a certain on-line business is collected and born by employing a fixed cycle sampling method; the information data of each load item of the current on-line business is processed to time sequences with fixed time intervals, and the time sequences of all load item data of the current business are obtained; discrete wavelet transformation of each time sequence of the on-line business is performed, statistics and analysis of each coefficient vector of an obtained coefficient matrix are conducted, and the probability of the abnormal load is calculated; the obtained probability and a confidence interval given by a confidence function are compared, and whether the abnormal load exists is determined; and a bearing server which is abnormal in the current on-line business is searched by employing the K-means clustering algorithm. Compared with the conventional method, according to the abnormal load detection method, higher accuracy can be obtained, and the adaptability is better.

Description

technical field [0001] The invention belongs to the technical field of cloud computing applications, and in particular relates to a method for identifying abnormal loads and abnormal operating conditions of services by utilizing historical load data of online services. Background technique [0002] With the development of cloud computing technology, more and more users choose to deploy or migrate services to platforms based on cloud architecture. Using cloud computing technology, resources such as computing, storage, and network allocated to specific businesses can be increased or decreased as needed, thereby maximizing resource utilization and reducing business operating costs. Online business accounts for a large proportion of all businesses deployed on cloud platforms. Since online businesses often directly provide service interfaces for users, the load of online businesses is more likely to be affected by the number of user visits. Monitoring the business load is the b...

AI Technical Summary

Problems solved by technology

Since the characteristics of the business are constantly changing throughout the life cycle of the business, these "fingerprints" also need to be constantly adjusted and corrected, making it difficult to achieve complete automation

[0009] To sum up, the existing methods of using load data to judge the abnormal conditions of online services cannot obtain high accuracy, and there is a problem of high false alarm rate

At the same time, these methods rely to a large extent on the experience of operation and maintenance personnel, and cannot achieve complete automatic monitoring

Images