Intelligent Alert Configuration System for Network Monitoring
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Configuring alerts for complex computing systems is challenging due to the need for expertise in both domain knowledge and data science, leading to sub-optimal performance and false positives, especially in managing thousands of dynamic performance metrics.
Innovation Solution
An automated system using machine learning and rule-based engines to analyze data from sensors, decoupling domain knowledge from data science expertise, and providing a scalable network backend for real-time monitoring and alert generation with high confidence levels.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If automated alert configuration is implemented, then productivity and ease of operation improve, but reliability may worsen due to potential false positives from automated systems lacking domain expertise
Solution Approach 1:
The patent introduces an intermediary system that bridges domain expertise and data science capabilities. This intermediary automatically translates domain knowledge requirements into data science algorithms, enabling automated alert configuration while maintaining high accuracy by preserving the essence of expert judgment without requiring direct human expert involvement in each configuration task.
Solution Approach 2:
The system performs preliminary action by pre-configuring alert rules based on historical data analysis and domain knowledge encapsulation before actual monitoring begins. This allows the automated system to start with pre-validated rules that have been optimized using domain expertise, thereby maintaining reliability while achieving automation.
2Reliability
If manual configuration by domain experts is used, then reliability improves through deep technology understanding, but productivity deteriorates due to lack of data science expertise and time-consuming processes
Solution Approach 1:
The system enables self-service by allowing domain experts to configure alerts using simple, intuitive interfaces that automatically handle the complex data science transformations. The system self-adapts by learning from configured rules and automatically optimizing alert parameters, eliminating the need for domain experts to acquire data science skills while maintaining configuration accuracy.
Solution Approach 2:
The patent replaces the mechanical system of manual expert configuration with an automated intelligent system that uses machine learning algorithms to perform the same function. This substitution maintains the quality of expert-level configuration while dramatically improving speed and scalability by removing the bottleneck of human expert availability.
3Ease of operation
If traditional rule-based alerting is used, then ease of operation improves with simple threshold settings, but measurement precision deteriorates due to inability to handle complex dynamic performance metrics
Solution Approach 1:
The system applies dynamics by transforming static threshold-based alerts into dynamic, adaptive alerting mechanisms. The alert rules automatically adjust to changing system conditions and learn from historical patterns, enabling the system to maintain simplicity of operation while achieving high measurement precision through continuous adaptation to dynamic performance metrics.
Solution Approach 2:
The patent utilizes parameter changes by automatically adjusting alert thresholds and parameters based on learned patterns from historical data. This allows the system to maintain easy operation with minimal user configuration while achieving high precision through automatic parameter optimization that adapts to changing system behavior and performance characteristics.
Data Source
AI summary
A method and system of creating an alert for a monitored network system. Key performance indicators (KPI's) of a plurality of components of a monitored network system are displayed on a user interface. A selection of one or more components of the plurality of components related to a malfunction is received. A present status and/or a pattern of performance of the one or more selected components is extracted. A preliminary alert is created based on the at least one of (i) the present status and (ii) the pattern of performance of the one or more selected components. Historical data related to the one or more selected components is retrieved. The preliminary alert is trained based on at least some of the retrieved historical scenarios. The preliminary alert is promoted to a primary alert upon training the preliminary to a confidence level that is above a predetermined threshold.