| 主权项 |
1. A method, in a data processing system, for determining threshold conditions for performing operations within a performance monitoring tool, comprising:
collecting, by the data processing system, performance metric data for a plurality of performance metrics from a computing system environment, wherein the performance metric data regards resource utilization and is derived from a plurality of time series raw historical performance metric data and raw current performance metric data obtained from hardware and software components of the computing system environment; initiating, by the data processing system, a performance monitoring rule generation engine in response to a number of errors exceeding a predetermined threshold in the computing system environment; retrieving, by the data processing system, the collected performance metric data in response to the number of errors exceeding the predetermined threshold; performing, by the data processing system, clustering analysis on the retrieved performance metric data to generate a plurality of clusters of performance metric data, wherein the performing the clustering analysis comprises performing, by the data processing system, a parameter reduction operation on the retrieved performance metric data to reduce a number of parameters within the retrieved performance metric data on which the clustering analysis performed, wherein the parameter reduction operation comprises a principle components analysis (PCA); selecting, by the data processing system, a plurality of candidate clusters in the plurality of clusters that are candidate clusters for dynamically generating threshold conditions for performance monitoring, wherein selecting the plurality of candidate clusters further comprises classifying, by the data processing system, each cluster in the plurality of clusters into a classification according to the plurality of cluster definition rules, wherein the plurality of cluster definition rules are fully automatically defined by an automated mechanism automatically defining the plurality of cluster definition rules based on analysis of the retrieved performance metric data; determining, by the data processing system, a plurality of cluster definition rules associated with the selected plurality of candidate clusters, the plurality of cluster definition rules being rules defining metric data values that cause each member of the selected plurality of candidate clusters to be in the respective cluster, wherein the cluster definition rules are not known prior to the clustering analysis being performed and thus, is a result of the clustering analysis operation providing insight into the collected performance metric data; generating, by the data processing system, the threshold conditions for performance monitoring based on the plurality of cluster definition rules that define the members of each of the selected plurality of candidate clusters, wherein the generating of the threshold conditions further comprises searching, by the data processing system, existing threshold conditions for similar threshold conditions to the generated threshold conditions and modifying the similar threshold conditions based on the generated threshold conditions; and applying, by the data processing system, the threshold conditions to future performance metric data to determine whether to perform an operation in response to the future performance metric data meeting or not meeting the at least one threshold conditions, wherein the operation is initiating an operation to reduce processor utilization by offloading processes onto another processor, generating a log report, and output an alert message. |
