Predictive incident management
AssigneeOracle International Corporation
InventorsPrashanth Kini, Richard Tauriello, Marcos M. Campos, Boriana Milenova, Charlie Sum, Mihaela Dediu, Kevin Timoney
Life of this patent
Estimating the term…
Abstract
An incident predictor system is described herein for predicting impactful incidents in which server computer system operations fail or perform poorly. According to one embodiment of the invention, the incident prediction system trains a generalized linear model (GLM) to predict when a system health indicator will reach a level that represents an incident for the server system.
Claims26
A computer-implemented method for predicting a future period of stress for a computer system, the method comprising:
- collecting a first set of granular monitoring data reflecting activity of the computer system over a first period of time;
- computing a first set of rollup metric values from the first set of granular monitoring data;
- determining a prediction threshold for a particular performance metric of the computer system based at least in part upon at least a portion of the first set of rollup metric values;
- collecting a second set of granular monitoring data reflecting activity of the computer system over a second period of time;
- wherein the second period of time ends after the first period of time ends;
- computing a second set of rollup metric values from the second set of granular monitoring data;
- based at least in part upon the second set of rollup metric values, generating a prediction that predicts a particular value for the particular performance metric at a particular time in the future relative to a time at which the prediction is generated;
- comparing the particular value to the prediction threshold to determine whether an alert should be generated;
- generating the alert in response to determining that the particular value exceeds the prediction threshold;
- wherein the method is performed by one or more computing devices.
A non-transitory computer-readable medium storing instructions which, when executed by one or more processors, cause the one or more processors to perform a method for predicting a future period of stress for a computer system, the method comprising:
- collecting a first set of granular monitoring data reflecting activity of the computer system over a first period of time;
- computing a first set of rollup metric values from the first set of granular monitoring data;
- determining a prediction threshold for a particular performance metric of the computer system based at least in part upon at least a portion of the first set of rollup metric values;
- collecting a second set of granular monitoring data reflecting activity of the computer system over a second period of time;
- wherein the second period of time ends after the first period of time ends;
- computing a second set of rollup metric values from the second set of granular monitoring data;
- based at least in part upon the second set of rollup metric values, generating a prediction that predicts a particular value for the particular performance metric at a particular time in the future relative to a time at which the prediction is generated;
- comparing the particular value to the prediction threshold to determine whether an alert should be generated;
- generating the alert in response to determining that the particular value exceeds the prediction threshold;
- wherein the method is performed by one or more computing devices.
Dependencies are inferred from the claim text.