Automated Root Cause Analysis for IT Systems
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current incident management procedures in large IT environments are inefficient, requiring significant labor and resources to identify the root cause of technical issues, often resulting in high costs and temporary fixes that fail to address the underlying problems.
Innovation Solution
A computer-implemented method using machine learning and natural language processing to collect and analyze data from multiple systems, building a classification model that performs root cause analysis and predictive analysis to identify and prevent technical issues, providing suggested corrective and optimization actions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional incident management procedures are used to identify root causes, then technical issues can be resolved, but significant labor hours and resources are spent parsing messages, errors, and logs
Solution Approach 1:
The patent replaces manual mechanical analysis of logs and error messages with an automated computer-implemented system that uses machine learning models and natural language processing to automatically parse, analyze, and identify root causes from system data, significantly reducing both time and human resources required
Solution Approach 2:
The system enables self-service root cause analysis by automatically collecting data from multiple systems, training classification models on historical data, and independently identifying root causes without requiring extensive human intervention or expert analysis
2Measurement precision
If manual analysis of logs and error messages is performed to identify root causes, then technical issues can be diagnosed, but the process is expensive and involves significant numbers of people at different support levels
Solution Approach 1:
The patent creates a universal automated system that handles multiple functions including data collection from various sources, machine learning model training, natural language processing of logs, root cause identification, and predictive analysis, replacing the need for multiple specialized human roles across different support levels
Solution Approach 2:
The system introduces an intermediary automated analysis layer between raw system data and human decision-makers, using machine learning models and NLP to process and interpret logs, errors, and messages, thereby reducing the complexity of direct human analysis while maintaining or improving diagnostic accuracy
3Speed
If temporary fixes are applied to resolve technical issues quickly, then service restoration is achieved, but the underlying root causes remain unaddressed
Solution Approach 1:
The patent performs preliminary root cause analysis using machine learning models and predictive analysis before issues fully manifest or recur, identifying underlying causes in advance and enabling proactive corrective actions that address root causes rather than just applying temporary fixes after problems occur
Data Source
AI summary
A process collects information about multiple systems of a computing environment from environmental tools web sources. The process builds and trains a classification model that extracts relevant information that classifies technical issues of the computing environment. The process performs root cause and predictive analyses, and identifies root causes of experienced technical issues and predicted technical issues. Based on identifying the root causes of the experienced technical issues, or identifying predicted issues, the process provides suggested corrective actions or optimization actions. Effectiveness of the suggested corrective actions and suggested optimization actions is tracked, and results of the tracking are fed into a training process that further trains the classification model.