| 摘要 |
A strategy is described for detecting anomalies in the operation of a data processing environment. The strategy relies on parameter information to detect the anomalies in a detection operation, the parameter information being derived in a training operation. The parameter information is selected such that the detection of anomalies is governed by both a desired degree of sensitivity (determining how inclusive the detection operation is in defining anomalies) and responsiveness (determining how quickly the detection operation reports the anomalies). The detection operation includes specific algorithms for determining undesired trending and spiking in the performance data.
|
