| 摘要 |
Monitoring the operational performance of a network-based business service involves defining and detecting significant variances in activities associated with performance of the service. A business service is characterized by corresponding business rhythms, which are derived from patterns of metric values for business activities that are part of business processes corresponding to the business service. Each business rhythm is characterized as a set of statistics about the corresponding metric(s) classified for a period of time or over a group of multiple periods of time, statistically compressed, and persistently stored. For purposes of real-time monitoring of the operational performance of the business service, significant variances in the normal behavior of the business service are automatically detected by comparing real-time metric data with corresponding historical metric data, in view of associated threshold values.
|
