Abnormality Score Calculation for Continuous Metric Monitoring
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
As system size increases, the number of metrics grows, making it difficult for managers to identify continuous abnormalities of high importance within a large number of detected metrics, where both temporary and important metrics are presented.
Innovation Solution
An operation management apparatus that collects and calculates an abnormality score for each metric based on its continuity level, distinguishing between temporary and continuous abnormalities by using a correlation model and abnormality score calculation unit.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the system size increases and more metrics are monitored, then the system monitoring capability is improved, but the number of abnormal metrics increases making it difficult to identify critical continuous abnormalities
Solution Approach 1:
The patent changes the parameter of abnormality assessment by introducing a continuity level parameter that evaluates how long an abnormality persists. This transforms the monitoring approach from simply detecting abnormality presence to assessing abnormality duration, allowing managers to prioritize continuous abnormalities over temporary ones even when many metrics are monitored.
Solution Approach 2:
The patent applies local quality by differentiating between types of abnormalities based on their continuity characteristics. Instead of treating all abnormal metrics uniformly, it assigns different importance levels based on whether the abnormality is continuous or temporary, enabling targeted attention to critical issues while filtering out noise from transient anomalies.
2Reliability
If all detected abnormal metrics are presented to the manager, then comprehensive abnormality detection is achieved, but the manager cannot easily grasp metrics with high importance level continuous abnormalities
Solution Approach 1:
The patent extracts the continuity level information from the set of abnormal metrics and presents it as a separate, prominent parameter. By taking out the continuity characteristic and displaying it separately, the system enables managers to quickly identify which abnormalities are continuous and critical without being overwhelmed by the total number of abnormal metrics.
Solution Approach 2:
The continuity level acts as an intermediary parameter that mediates between the raw abnormality detection data and the manager's decision-making process. It translates the complex set of abnormal metrics into a prioritized view where continuous abnormalities are highlighted, serving as a bridge between comprehensive monitoring and effective anomaly prioritization.
3Area of stationary object
If the number of metrics increases with system size, then monitoring coverage is improved, but the burden on managers to analyze all abnormal metrics increases
Solution Approach 1:
The patent applies partial action by focusing manager attention on only the most critical aspect of abnormal metrics - their continuity level. Instead of requiring managers to analyze all aspects of every abnormal metric, the system provides continuity information that enables quick prioritization, allowing managers to address continuous abnormalities first while temporarily deprioritizing transient issues.
Data Source
AI summary
An operation management to grasp a metric in which a continuous abnormality has occurred in a system, easily, is provided. An operation management apparatus 100 includes a metric collection unit (101) and an abnormality score calculation unit (104). The metric collection unit (101) collects a measured value of each of a plurality of metrics in a system sequentially. The abnormality score calculation unit (104) calculates and outputs, on the basis of a continuity level indicating a degree of continuity of an abnormality of the measurement value for each of the plurality of metrics at each time, an abnormality score for the metric.