Causal Graph Fault Localization in Micro-Service Architectures
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The complexity of dependency relationships among micro-services in cloud-native applications makes fault localization difficult, especially with limited observability and unknown topology, leading to inefficiencies in identifying and addressing application errors.
Innovation Solution
A system and method that employs active causal learning to infer causal graphs among micro-services, using selective error injection in a staging environment to generate a learned causal graph, which is then leveraged in a production environment to dynamically localize application errors through an ancestral matrix and error log data analysis.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If micro-service architecture is used to enable scalability and modularity, then system adaptability and deployability are improved, but the complexity of dependency relationships increases making fault localization more difficult
Solution Approach 1:
The patent introduces a causal graph as an intermediary data structure that models dependency relationships among micro-services. This causal graph serves as a mediator between the complex micro-service architecture and the fault localization process, enabling systematic analysis of error propagation without requiring direct inspection of all dependency relationships.
Solution Approach 2:
The patent replaces traditional mechanical/manual fault localization methods with an AI-based causal inference system. Machine learning models automatically infer causal relationships from operational data, substituting manual analysis and traditional debugging approaches with automated intelligent systems that can handle the complexity of micro-service dependencies.
2Ease of manufacture
If traditional fault localization methods are used without causal models, then implementation simplicity is maintained, but fault localization accuracy and efficiency deteriorate in complex micro-service environments
Solution Approach 1:
The patent performs preliminary action by constructing causal graphs and training AI models in advance during the staging environment phase. This preliminary modeling of normal and erroneous operational patterns enables rapid and accurate fault localization during production without requiring complex real-time analysis, thus maintaining implementation simplicity while improving accuracy.
Solution Approach 2:
The patent transitions from traditional single-dimension fault detection to multi-dimensional causal analysis by incorporating temporal sequences, error propagation paths, and contextual operational data. This dimensional expansion enables more accurate fault localization by considering multiple factors simultaneously rather than relying on simple error matching.
3Measurement precision
If extensive observational data is collected for fault analysis, then fault localization accuracy may be improved, but system overhead and processing complexity increase
Solution Approach 1:
The patent extracts only the most relevant features and error patterns from operational data for causal graph construction and fault analysis. By selectively extracting critical information rather than processing all available data, the system achieves accurate fault localization while minimizing data volume and processing overhead.
Solution Approach 2:
The patent changes parameters by transforming raw operational data into standardized causal representations with specific features and attributes. This parameter transformation enables efficient processing by converting diverse data formats into a unified causal model structure that can be quickly analyzed for fault detection.
Data Source
AI summary
A computer-implemented method is provided that includes learning causal relationships between two or more application micro-services, and applying the learned causal relationships to dynamically localize an application fault. First micro-service error log data corresponding to selectively injected errors is collected. A learned causal graph is generated based on the collected first micro-service error log data. Second micro-service error log data corresponding to a detected application and an ancestral matrix is built using the learned causal graph and the second micro-service error log data. The ancestral matrix is leveraged to identify the source of the error, and the micro-service associated with the identified error source is also subject to identification. A computer system and a computer program product are also provided.