Significance-Ordered Prefix Tree for Root Cause Investigation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for root cause investigation in cloud-based software services are inefficient due to high computational overhead when analyzing large datasets of telemetry point values, leading to reliance on manual guesswork and significant resource consumption.
Innovation Solution
The method involves computing Z-scores for telemetry point values, constructing a significance-ordered prefix tree to identify statistically significant cohorts, and reducing the number of cohorts to evaluate, thereby decreasing computational overhead and increasing efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If existing algorithms are used to identify telemetry point value combinations, then statistical accuracy is improved, but computational overhead increases exponentially
Solution Approach 1:
The patent segments the exponentially growing search space by organizing telemetry point values into a prefix tree structure. This hierarchical segmentation divides the combinatorial problem into manageable levels, where only paths with statistically significant Z-scores are fully explored, dramatically reducing computational overhead while maintaining statistical accuracy.
Solution Approach 2:
The patent performs preliminary statistical filtering by computing Z-scores for individual telemetry point values before exploring combinations. This preliminary action identifies and prunes insignificant values early in the process, preventing exponential exploration of combinatorial spaces that would waste computational resources.
2Ease of operation
If purely algorithmic approaches are used, then objectivity is improved, but time consumption increases
Solution Approach 1:
The prefix tree structure segments the analysis into hierarchical levels, allowing the system to objectively evaluate individual telemetry points first (maintaining algorithmic objectivity) and then progressively explore combinations only where statistically justified, significantly reducing time consumption compared to exhaustive algorithmic approaches.
3Loss of time
If manual guesswork is used, then time consumption is reduced, but reliability decreases
Solution Approach 1:
The system incorporates feedback through Z-score calculations at each level of the prefix tree. This statistical feedback mechanism guides the automated exploration process, ensuring that only telemetry point combinations with statistically significant correlations are pursued, thereby maintaining high investigation reliability while avoiding time-consuming manual guesswork.
Data Source
AI summary
A disclosed method facilitates a computationally efficient way of identifying cohorts of telemetry point values that are statistically significant indicators of a computing event satisfying a condition of interest. The method includes computing Z-scores for the telemetry point values in an event-based dataset; identifying, based on the Z-scores, significant telemetry point values that are statistically significant indicators of the condition of interest; generating a significance-ordered prefix tree based on the significant telemetry point values and the Z-scores; using the significance-ordered prefix tree to identify cohorts to evaluate in combination; computing a cohort Z-score for each of the identified cohorts and identifying, based on the cohort Z-scores, a subset of the cohorts that are statistically significant indicators of the condition of interest.