Temporal Heterogeneous Graph Pruning for Root Cause Localization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for determining the root cause of faults in a network, such as an Internet application with microservices, are inefficient due to manual annotation and update of causal chains, leading to low efficiency and accuracy in identifying the root cause.
Innovation Solution
A method involving the reconstruction of a temporal heterogeneous graph by deleting normal nodes and edges, followed by determining a root cause in the resulting anomaly subgraph using a probability transfer matrix and random walk probabilities, enhanced by deep neural networks to improve efficiency and accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If manual annotation and manual update of causal chain is used, then the root cause can be determined, but the efficiency of determining the root cause is low
Solution Approach 1:
The system automatically constructs and updates the causal chain by analyzing monitoring data, alarm information, and topology relationships without requiring manual annotation. The root cause localization apparatus autonomously performs data collection, graph construction, and root cause identification, eliminating the need for manual intervention while improving efficiency
Solution Approach 2:
The patent replaces the manual mechanical process of annotating and updating causal chains with an automated computational system that uses graph theory algorithms, probability transfer matrices, and random walk methods to automatically determine root causes from system data
2Loss of time
If the temporal heterogeneous graph includes all nodes and edges, then the graph is complete, but the search space and computation time increase
Solution Approach 1:
The patent extracts and removes normal nodes and their connected edges from the temporal heterogeneous graph, retaining only anomaly nodes and their relationships. This extraction process creates a simplified anomaly subgraph that maintains the essential fault propagation paths while eliminating irrelevant components, thereby reducing search space and computation time
Solution Approach 2:
The patent segments the complete temporal heterogeneous graph into multiple components by separating normal nodes from anomaly nodes. This segmentation creates a focused anomaly subgraph that can be processed independently, reducing the overall computational burden while preserving the critical fault information needed for root cause localization
Data Source
Figure 1~2
Figure 3~4
Figure 5~6
AI summary
This application provides a root cause locating method and apparatus, a device, and a storage medium, and pertains to the field of Internet technologies. The method includes: during root cause locating, obtaining a temporal heterogeneous graph; deleting, based on an attribute of a node in the temporal heterogeneous graph, a normal node and an edge connected to the normal node that are in the temporal heterogeneous graph, and reconstructing a temporal heterogeneous graph obtained through the deletion, to obtain an anomaly subgraph, where a reachable edge exists between a faulty node and another node in the anomaly subgraph; and determining a root cause of the faulty node in the anomaly subgraph. According to the solutions of this application, in the anomaly subgraph obtained through the reconstruction, reachability between the faulty node and another node is retained, and a scale of an original node and noise caused by the normal node are reduced. Therefore, efficiency and accuracy of root cause locating can be improved.