Alarm fusion system and method based on data center anomaly monitoring

Abstract

The invention provides an alarm fusion method based on data center anomaly monitoring. The alarm fusion method comprises the following steps: S1, alarm collection: collecting alarm information from each data source monitored by a whole operation and maintenance system in a series of time windows; S2, alarm preprocessing: carrying out redundancy elimination operation on the collected alarm information, and extracting key alarm information in the collected alarm information; S3, alarm association analysis: discovering association rules of the preprocessed alarm information according to a certainalgorithm; and S4, alarm integration and output: performing de-duplication and integration operation on the newly arrived alarm information within a certain time according to the association rule ofthe alarm information discovered in the step S3, and outputting the alarm information to a database. According to the invention, redundant information in the alarm information can be eliminated. The alarm information with a causal relationship is fused, so that the workload and the working difficulty of operation and maintenance personnel are reduced. The operation and maintenance cost is reduced.

Description

technical field [0001] The invention relates to the technical field of alarm monitoring for data center operation and maintenance systems, in particular to an alarm fusion method based on data center abnormal monitoring. Background technique [0002] There is such a problem in most of the existing data center operation and maintenance systems: because the monitoring items include the monitoring of each server, the running status of the cluster, and the monitoring of software data running on the server, as well as the complexity of the cluster topology status, resulting in A component that generates an alarm may trigger a chain of chain alarms. That is to say, one fault will lead to another fault in the form of reaction chain, which will generate countless alarms, which will mask the real important fault. Moreover, most of the alarms do not contain information about the real cause of the fault, but a large number of alarms will appear when a fault occurs, some of which are r...

AI Technical Summary

Problems solved by technology

[0002] There is such a problem in most of the existing data center operation and maintenance systems: because the monitoring items include the monitoring of each server, the running status of the cluster, and the monitoring of software data running on the server, as well as the complexity of the cluster topology status, resulting in An alarm generated by a component may trigger a series of chain alarms

For systems with large scale and complex structure, it is difficult to build an accurate network model

[0010] The bottleneck of the method based on artificial neural network is that it is not easy to find training data, and different input parameters have a great influence on the results

Images