Dynamic Metric Threshold Adjustment via User Sentiment Feedback
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Legacy techniques for setting system metric thresholds in distributed computing systems fail to accurately correlate with user-perceived performance, leading to ineffective or costly adjustments, as they rely on default values that do not account for dynamic user environments and may trigger unnecessary alerts or miss performance issues.
Innovation Solution
The implementation of techniques that correlate system metrics with user-specified performance indicators using learning models to generate a mapping function, allowing for dynamic adjustment of thresholds based on user feedback, thereby improving the accuracy of performance monitoring and reducing unnecessary actions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If default threshold values are used for system metrics, then the system can be easily configured and deployed, but the thresholds do not accurately reflect user-perceived performance
Solution Approach 1:
The system implements feedback loops where user performance feedback is continuously collected and used to adjust metric thresholds. The threshold adjustment agent monitors user feedback about performance sentiment and dynamically modifies thresholds to better reflect actual user-perceived performance, resolving the contradiction between easy configuration and accurate measurement.
Solution Approach 2:
The patent transforms static default thresholds into dynamic, adaptive thresholds that automatically adjust based on user feedback and system conditions. The thresholds evolve over time to match user-perceived performance characteristics, maintaining both ease of initial configuration and ongoing measurement accuracy.
2Device complexity
If static default thresholds are used, then the system configuration is simple, but the thresholds may trigger unnecessary alerts or miss performance issues
Solution Approach 1:
The system implements self-service automation where the threshold adjustment agent autonomously monitors user feedback and adjusts thresholds without requiring manual intervention. This maintains configuration simplicity while significantly improving monitoring reliability by eliminating false alerts and detecting actual performance issues.
Solution Approach 2:
User feedback about performance sentiment is continuously fed back to the threshold adjustment agent, which uses this information to refine thresholds. This feedback mechanism ensures that thresholds remain reliable indicators of user-perceived performance while keeping the system configuration simple.
3Measurement precision
If user feedback collection is implemented to improve threshold accuracy, then performance monitoring becomes more accurate, but system complexity increases
Solution Approach 1:
The patent introduces a threshold adjustment agent as an intermediary component that mediates between user feedback and threshold configuration. This intermediary simplifies the overall system architecture by centralizing the feedback processing logic and providing a clear interface between user sentiment and threshold adjustments, thereby managing complexity while improving accuracy.
4Speed
If actions are taken based on threshold breaches without statistical validation, then rapid responses can be made, but the actions may not correlate with actual performance improvement
Solution Approach 1:
The system performs preliminary statistical validation and correlation analysis before taking corrective actions. The threshold adjustment agent analyzes historical data to establish statistically significant correlations between threshold breaches and user-perceived performance, ensuring that actions are based on validated relationships rather than isolated events, thus improving reliability while maintaining response speed.
Data Source
AI summary
Systems for autonomous management of hyperconverged distributed computing and storage systems. A method embodiment commences upon receiving a set of system measurements that correspond to system metrics associated with the computing system. A user interface is presented to users to capture a set of user sentiment indications. Over a period of time, a time series of system measurements and a time series of user sentiment indications are captured and used to form a learning model that comprises dynamically-changing user sentiment correlations between the system measurements and the user sentiment. At some moment in time, a system metric threshold breach event occurs. The learning model is consulted to determine a tracking value between the set of user sentiment indications and the system metric pertaining to the system metric threshold. Based on the tracking value, the respective system metric threshold is adjusted to more closely track with the historical user sentiment indications.