Method for analyzing the influence range of environmental faults driven by function inspection

Abstract

This invention relates to the field of fault diagnosis technology and discloses a method for analyzing the impact range of environmental faults driven by functional inspection. The method includes: performing pattern matching between performance parameters in functional status data and preset thresholds to obtain a list of anomaly identifiers for functional components; constructing a dependency graph based on the anomaly identifier list and the dependency graph; performing backpropagation analysis on the dependency direction of functional components based on the anomaly identifier list and the dependency graph to obtain the fault impact range; evaluating the degree of impact on functional components from multiple dimensions based on the impact range to obtain the criticality weights of functional components; aggregating the weights of functional components based on the criticality weights and the impact range to obtain an impact severity index for functional components; and generating an environmental fault handling plan for functional components based on the impact severity index. This invention can improve the efficiency of environmental fault impact range analysis driven by functional inspection.

Description

Technical Field

[0001] This invention relates to the field of fault diagnosis technology, and in particular to a method for analyzing the impact range of environmental faults driven by functional inspection. Background Technology

[0002] In the field of environmental fault impact range analysis driven by functional inspection, existing technologies generally rely on manual parameter comparison and fault diagnosis modes, which makes it difficult to efficiently and accurately identify anomalies in performance parameters in functional status data. These methods often use only a single threshold for static matching and cannot dynamically adjust the judgment criteria in combination with the historical performance trends of functional components, resulting in insufficient accuracy of anomaly identification and easy omissions or misjudgments. This creates hidden dangers for subsequent fault impact range analysis and makes it difficult to quickly locate functional components that are truly abnormal.

[0003] Existing technologies lack systematic topology construction and reverse propagation analysis capabilities when constructing functional component dependencies and analyzing the scope of fault impact. On the one hand, they cannot construct accurate dependency models based on the degree of component association, resulting in a chaotic dependency structure. On the other hand, when analyzing the impact of faults, they can only perform simple forward scope definition, unable to trace back multi-level impact nodes along the dependency direction, making it difficult to fully grasp the scope of fault impact. In addition, existing technologies have a single dimension for assessing the impact of faults, failing to consider multiple factors such as the operational characteristics and contribution of functional components, and unable to scientifically quantify the critical weight of components and the severity of fault impact. Ultimately, this results in fault handling solutions that are insufficiently targeted and effective, making it difficult to meet the fault handling needs in complex environments. Summary of the Invention

[0004] This invention provides a method for analyzing the impact range of environmental faults driven by functional inspection to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides a method for analyzing the impact range of environmental faults driven by functional inspection, comprising: S1. Perform pattern matching between the performance parameters in the functional status data and the preset thresholds to obtain a list of abnormal identifiers for the functional components. S2. Construct a dependency graph of the functional components based on the list of exception identifiers; S3. Based on the anomaly identifier list and the dependency graph, perform reverse propagation analysis on the dependency direction of the functional component to obtain the fault impact range of the functional component; S4. Based on the scope of influence, the degree of influence on the functional component is evaluated from multiple dimensions to obtain the criticality weight of the functional component; S5. Based on the key weights and the scope of influence, perform weight aggregation on the functional components to obtain the severity index of the influence of the functional components; S6. Generate an environmental fault handling plan for the functional component based on the severity index of the impact.

[0006] In a preferred embodiment, the step of performing pattern matching between the performance parameters in the functional status data and preset thresholds to obtain an anomaly identifier list for the functional component includes: The performance parameters in the functional status data are normalized to obtain the standardized performance parameters of the functional components. The standardized performance parameters are compared with the preset threshold using a multi-rule pattern to obtain the dynamic adjustment parameters of the functional components. Based on the historical performance parameters of the functional components, trend analysis is performed on the dynamic adjustment parameters to obtain the abnormal behavior patterns of the functional components. Based on the abnormal behavior pattern, generate a list of abnormal identifiers for the functional components.

[0007] In a preferred embodiment, constructing the dependency graph of the functional components based on the list of exception identifiers includes: The configuration information in the functional components is constructed in a graph form to obtain the dependency topology of the functional components; Based on the dependency topology and the list of exception identifiers, construct the initial dependency graph of the functional components; The initial dependency graph is structurally optimized to obtain the dependency graph of the functional components.

[0008] In a preferred embodiment, constructing the initial dependency graph of the functional components based on the dependency topology and the anomaly identifier list includes: Based on the degree of association of the functional components, extract the component connection matrix of the dependent topology; Based on the list of anomaly identifiers, anomaly enhancement processing is performed on the component connection matrix to obtain the optimized component connection matrix of the functional component; Based on the optimized component connection matrix, construct the weighted directed graph structure of the functional components; The connectivity of the weighted directed graph structure is verified to obtain the initial dependency graph of the functional components.

[0009] In a preferred embodiment, the backpropagation analysis of the dependency direction of the functional component based on the anomaly identifier list and the dependency graph to obtain the fault impact range of the functional component includes: The list of anomaly identifiers is processed to obtain the starting node of the dependency direction in the functional component; Based on the dependency graph and the starting node, a path traversal operation is performed in the opposite direction of the dependency direction to obtain the reverse dependency path of the functional component. Based on the reverse dependency path, a multi-level impact propagation analysis is performed on the functional components to obtain the fault impact nodes of the functional components. The scope of the fault is traced by tracking the affected nodes to obtain the fault impact range of the functional component.

[0010] In a preferred embodiment, the step of performing multi-level impact propagation analysis on the functional component based on the reverse dependency path to obtain the fault impact nodes of the functional component includes: Record the propagation state and influence depth of the functional component to obtain the initial state table of the functional component; Based on the initial state table and the reverse dependency path, the functional component is backtracked to obtain the node to be analyzed in the functional component. Dependency strength is estimated for the nodes to be analyzed to obtain the strongly dependent nodes of the functional components; Based on the strongly dependent nodes, state propagation analysis is performed on the propagation state of the functional components to obtain the fault-affected nodes of the functional components.

[0011] In a preferred embodiment, the assessment of the degree of influence of the functional component based on the scope of influence, to obtain the critical weight of the functional component, includes: Feature extraction is performed on the affected functional components within the scope of influence to obtain the operational feature data of the functional components; The operational feature data is analyzed from multiple dimensions to obtain the dimensional evaluation results of the functional components; Based on the evaluation results of the aforementioned dimensions, a contribution analysis is performed on the environmental operating status of the functional components to obtain the contribution coefficient of the functional components. The key weights of the functional components are obtained by multi-factor fusion of the dimensional evaluation results and the contribution coefficients.

[0012] In a preferred embodiment, the step of weighting the functional components based on the criticality weights and the scope of influence to obtain an impact severity index for the functional components includes: Based on the scope of influence, the dependency chain of the functional component is resolved to obtain the dependency depth of the functional component; Based on the criticality weights and the dependency depth, construct the weight aggregation matrix of the functional components; Based on the weight aggregation matrix, the impact severity index of the functional component is calculated, wherein the calculation formula for the impact severity index is: ; in, This indicates the severity index of the impact. Indicates the first The aforementioned key weights, Indicates the first The aforementioned dependency depth This indicates the total number of affected functional components. This indicates the total number of the functional components. Represents the logarithmic function. It represents the base of the natural logarithm.

[0013] In a preferred embodiment, the step of performing dependency chain resolution on the functional component based on the scope of influence to obtain the dependency depth of the functional component includes: Based on the dependency graph, extract the longest dependency path of the functional component; The longest dependency path is traversed to obtain the path length distribution data of the functional components; The path length distribution data is mapped to a preset depth range to obtain the standardized depth value of the functional component; The topological importance parameter of the functional component is normalized and aggregated with the normalized depth value to obtain the dependency depth of the functional component.

[0014] To address the above problems, this invention also provides an environmental fault impact range analysis driven by functional inspection, which includes: The step of generating an environmental fault handling plan for the functional component based on the severity index includes: The severity index of the impact is classified into levels to obtain the fault severity level sequence of the functional component; Based on the scope of influence and the critical weight, a strategy analysis is performed on the fault severity level sequence to obtain the initial fault response plan for the functional component. Resource conflict detection is performed on the initial fault response plan to obtain the environmental fault handling plan for the functional component.

[0015] Compared with the prior art, the present invention has the following beneficial effects: 1. This invention generates an anomaly identification list by standardizing the performance parameters in functional status data and combining multi-rule pattern comparison and trend analysis, which can accurately locate functional component anomalies. Then, based on the anomaly identification list, an optimized dependency graph is constructed, and the nodes and scope of the fault are traced by backpropagation analysis, which greatly improves the accuracy of fault impact range analysis, allowing staff to quickly identify the functional components affected by the fault and reduce processing delays caused by ambiguous range determination.

[0016] 2. This invention determines critical weights by assessing the impact of functional components from multiple dimensions, constructs a weight aggregation matrix by combining dependency depth, and calculates the severity index of impact using a specific formula. Based on this, an environmental fault handling solution is generated. This not only scientifically quantifies the severity of faults but also provides accurate and targeted basis for fault handling, effectively improving the efficiency and quality of environmental fault handling driven by functional inspection and ensuring the overall operational stability of functional components. Attached Figure Description

[0017] Figure 1 A flowchart illustrating the environmental fault impact range analysis method driven by functional inspection provided in an embodiment of the present invention; The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0018] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0019] This application provides a method for analyzing the impact range of environmental faults driven by functional inspection. The executing entity of this method includes, but is not limited to, at least one of the following electronic devices that can be configured to execute the method provided in this application: a server, a terminal, etc. In other words, the method can be executed by software or hardware installed on a terminal device or a server device. The server includes, but is not limited to, a single server, a server cluster, a cloud server, or a cloud server cluster. The server can be an independent server or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDNs), and big data and artificial intelligence platforms.

[0020] Reference Figure 1The diagram shown is a flowchart illustrating a method for analyzing the impact range of environmental faults driven by functional inspection, according to an embodiment of the present invention. In this embodiment, the method includes: S1. Perform pattern matching between the performance parameters in the functional status data and the preset thresholds to obtain a list of abnormal identifiers for the functional components. In this embodiment of the invention, the step of performing pattern matching between the performance parameters in the functional status data and preset thresholds to obtain an anomaly identifier list for the functional component includes: The performance parameters in the functional status data are normalized to obtain the standardized performance parameters of the functional components. The standardized performance parameters are compared with the preset threshold using a multi-rule pattern to obtain the dynamic adjustment parameters of the functional components. Based on the historical performance parameters of the functional components, trend analysis is performed on the dynamic adjustment parameters to obtain the abnormal behavior patterns of the functional components. Based on the abnormal behavior pattern, generate a list of abnormal identifiers for the functional components.

[0021] Specifically, when normalizing the performance parameters in the functional status data, all performance parameters generated by the functional component during normal operation are first collected. These performance parameters include, but are not limited to, response time, resource utilization, and data transmission rate. Then, the value range of each performance parameter is determined by statistically analyzing the maximum and minimum values ​​of the corresponding performance parameters during the historical normal operation of the functional component. Then, the actual value of each performance parameter is subtracted from its minimum value, and the difference is divided by the difference between the maximum and minimum values ​​of the performance parameter. Through this operation, performance parameters of different magnitudes and units are uniformly converted to the range of [0,1], and finally, the standardized performance parameters of the functional component are obtained.

[0022] Furthermore, when comparing standardized performance parameters with preset thresholds using a multi-rule pattern, multiple sets of preset comparison rules are first defined. Each set of rules corresponds to a judgment scenario for an abnormal performance parameter. For example, one set of rules sets a potential anomaly when the standardized response time parameter exceeds 0.8 for five consecutive times, and another set of rules sets a potential anomaly when the standardized resource utilization rate parameter fluctuates by more than 0.5 within 10 minutes. Then, each standardized performance parameter is compared with its corresponding preset threshold according to each set of comparison rules, and the result of meeting the anomaly judgment scenario after each comparison is recorded. Based on these results, the dynamic parameters related to the performance parameter are adjusted so that the adjusted parameters can accurately reflect the degree of deviation between the current performance parameter and the normal state, and finally the dynamic adjustment parameters of the functional components are obtained.

[0023] Furthermore, when performing trend analysis on dynamic adjustment parameters based on the historical performance parameters of functional components, the historical performance parameters of the functional components for the same time period each day over the past 6 months, as well as the dynamic adjustment parameters generated for the corresponding time periods, are first retrieved. These dynamic adjustment parameters are then arranged sequentially according to time to form a dynamic adjustment parameter sequence. The direction and magnitude of change of the dynamic adjustment parameters in this sequence are then observed. For example, it is determined whether the parameters are continuously rising, continuously falling, or fluctuating, as well as the specific increase or decrease in value each time. At the same time, it is combined with whether similar dynamic adjustment parameter change trends have appeared in the historical performance parameters. When it is found that the change trend of the current dynamic adjustment parameter sequence is consistent with the change trend of the dynamic adjustment parameters before the historical failure, the abnormal behavior of the current functional component can be determined, and the abnormal behavior pattern of the functional component can be obtained.

[0024] Furthermore, when generating the list of abnormal identifiers for functional components based on abnormal behavior patterns, the identified abnormal behavior patterns are first categorized to clarify the abnormal type of functional component corresponding to each abnormal behavior pattern. For example, the abnormal behavior pattern of "continuous increase in response time-related dynamic adjustment parameters" is marked as "response delay abnormality," and the abnormal behavior pattern of "significant fluctuation in resource utilization-related dynamic adjustment parameters" is marked as "resource utilization abnormality." Then, the specific identifier information of the functional components exhibiting each abnormal behavior pattern is compiled. This identifier information includes the functional component's number, device location, and functional module. Finally, these abnormal types and their corresponding functional component identifier information are organized into a list according to a unified format, ultimately yielding the list of abnormal identifiers for functional components.

[0025] In summary, normalizing the performance parameters in functional status data to obtain standardized performance parameters can effectively eliminate interference caused by differences in dimensions and ranges of values ​​among different performance parameters. In a diagnostic system that integrates functional inspection and environmental analysis, this process ensures that different types of performance parameters are in a unified data dimension, avoiding the situation where an abnormality in other key parameters is masked by the excessive magnitude of a certain parameter. This provides a consistent data foundation for subsequent accurate comparison with preset thresholds, ensuring the objectivity and accuracy of the subsequent pattern matching process, and aligns with the core idea of ​​systematically optimizing multi-source data.

[0026] In summary, comparing standardized performance parameters with preset thresholds using multiple rule patterns to obtain dynamically adjustable parameters overcomes the limitations of traditional single-threshold judgments. Combining multimodal data fusion and multi-scale feature screening, multi-rule pattern comparison can construct judgment rules from multiple dimensions such as parameter fluctuation frequency, fluctuation amplitude, and duration of continuous exceedance of thresholds. For example, it can simultaneously consider composite conditions such as parameters exceeding the threshold by a certain percentage multiple times consecutively or parameters fluctuating beyond the limit within a specific time period. This more comprehensively captures the deviation characteristics of performance parameters from normal operating conditions, and the generated dynamically adjustable parameters can accurately reflect the actual deviation degree of functional components in the current environment and operating state, providing more targeted intermediate data for subsequent anomaly analysis and avoiding the problems of misjudgment or omission under a single rule.

[0027] In summary, trend analysis of dynamic adjustment parameters based on historical performance parameters of functional components to identify abnormal behavior patterns fully integrates the technical concept of using historical operating data for trend prediction. This step does not analyze the current dynamic adjustment parameters in isolation, but places them within the time series of historical data. By comparing the historical parameter change patterns of components under the same environmental conditions and similar loads, the changing trends of dynamic adjustment parameters are explored, such as linearly increasing parameter deviations or periodic anomalies. This allows for the identification of potential anomalies that are difficult to detect with short-term single parameter fluctuations. For example, if the dynamic adjustment parameters of a certain component increase month by month, although they have not yet exceeded the severe anomaly threshold, they are already showing signs of impending failure. This analysis method overcomes the limitations of relying solely on real-time data for diagnosis, improves the ability to predict latent anomalies, and is highly consistent with the goal of extending from diagnosis to prediction.

[0028] In summary, generating an anomaly identifier list for functional components based on abnormal behavior patterns transforms abstract data analysis results into concrete and actionable information output. Combined with application scenarios providing decision support for preventative equipment maintenance, the anomaly identifier list clearly marks the location of abnormal components, anomaly types such as performance degradation or intermittent fluctuations, and the severity of anomalies based on the duration and deviation of the abnormal behavior pattern. This avoids the tedious process of manually sifting through massive amounts of data, significantly reducing the time cost of fault location. This output format directly connects to subsequent maintenance operations, ensuring that diagnostic results can be efficiently transformed into actual equipment management actions, further improving the overall efficiency of functional inspection and intelligent diagnosis, aligning with the core value of enhancing diagnostic prediction efficiency and practicality.

[0029] S2. Construct a dependency graph of the functional components based on the list of exception identifiers; In this embodiment of the invention, constructing the dependency graph of the functional components based on the list of exception identifiers includes: The configuration information in the functional components is constructed in a graph form to obtain the dependency topology of the functional components; Based on the dependency topology and the list of exception identifiers, construct the initial dependency graph of the functional components; The initial dependency graph is structurally optimized to obtain the dependency graph of the functional components.

[0030] In this embodiment of the invention, constructing the initial dependency graph of the functional components based on the dependency topology and the anomaly identifier list includes: Based on the degree of association of the functional components, extract the component connection matrix of the dependent topology; Based on the list of anomaly identifiers, anomaly enhancement processing is performed on the component connection matrix to obtain the optimized component connection matrix of the functional component; Based on the optimized component connection matrix, construct the weighted directed graph structure of the functional components; The connectivity of the weighted directed graph structure is verified to obtain the initial dependency graph of the functional components.

[0031] Specifically, when constructing a graph of the configuration information in functional components, the configuration information of all functional components is first comprehensively collected. This configuration information includes the name, function, data interaction object, data transmission path, and other component information that the component must depend on for startup and runtime. Then, each functional component is used as a node in the graph. Based on the dependencies between components recorded in the configuration information, directed line segments are used to connect the nodes with dependencies. The starting point of the directed line segment is the dependent component node, and the ending point is the node that depends on the component. At the same time, the specific type of dependency is marked next to the directed line segment, such as data dependency, service dependency, etc. In this way, all functional components and their dependencies are presented in the form of a graph, and finally the dependency topology of the functional components is obtained.

[0032] Furthermore, when constructing the initial dependency graph of functional components based on the dependency topology and the list of anomaly identifiers, the identifier information of all functional components marked as anomalies is first extracted from the list of anomaly identifiers. Then, the nodes corresponding to these anomaly functional components are found in the constructed dependency topology, and these nodes are marked with a preset special color to clearly distinguish between normal component nodes and anomaly component nodes. Next, other component nodes in the dependency topology that are connected to the anomaly component nodes through directed line segments are checked to confirm whether the dependency direction and dependency type between these nodes and the anomaly component nodes are accurate. If there are cases where the dependency description is ambiguous or incorrect, it is corrected according to the configuration information of the functional components to ensure that the dependency relationships between all component nodes are accurate. Finally, the graph with anomaly nodes marked and the dependency relationships corrected is determined as the initial dependency graph of the functional components.

[0033] Furthermore, when optimizing the initial dependency graph, first check if there are redundant nodes or segments. Redundant nodes refer to isolated nodes that do not have dependencies on any other component nodes in the entire dependency network. Redundant segments refer to segments that repeatedly connect the same pair of component nodes with the same dependency type. Once redundant nodes or segments are found, they are directly deleted from the initial dependency graph. Then, the positional distribution of component nodes in the graph is adjusted, arranging component nodes with direct dependencies as close as possible to avoid the dependency display being chaotic due to overly scattered node distribution. At the same time, the direction of directed segments is adjusted to ensure that there are not too many intersections between segments, improving the readability of the graph. Finally, the optimized graph is checked as a whole to confirm that all abnormal component nodes are clearly marked, dependencies are accurate, and the structure is concise and easy to read, ultimately obtaining the dependency graph of the functional components.

[0034] Specifically, when extracting the component connection matrix of the dependency topology based on the degree of association between functional components, the criteria for judging the degree of association between functional components are first clarified. The degree of association is determined by the frequency of data interaction between components, the amount of data interaction, and the necessity of dependence. For example, if two components interact with each other more than 1,000 times a day and the amount of data interaction reaches more than 1GB, and one component must depend on the other component to start, then the two components are judged to have a high degree of association. Next, all functional components are numbered sequentially to form a component list. A two-dimensional matrix is ​​constructed with the components in the component list as rows and columns, with the rows and columns of the matrix corresponding to different functional components. Then, each element in the matrix is ​​assigned a value according to the degree of association between components. If two components are not related, the value is 0; if the degree of association is low, the value is 1; if the degree of association is moderate, the value is 2; and if the degree of association is high, the value is 3. In this way, the association relationship of components in the dependency topology is transformed into a specific numerical matrix, and finally the component connection matrix of the dependency topology is obtained.

[0035] Furthermore, when performing anomaly enhancement processing on the component connection matrix based on the anomaly identifier list, firstly, all functional components marked as abnormal and their corresponding anomaly types are filtered out from the anomaly identifier list. Different anomaly types reflect the severity of component anomalies; for example, the severity of anomalies of the "core function failure" type is higher than that of anomalies of the "non-core function response delay" type. Next, the rows and columns corresponding to the abnormal functional components are found in the component connection matrix. For the row containing the abnormal functional component, if other components in that row are related to the abnormal component, the matrix element values ​​are adjusted according to the severity of the anomaly. If the anomaly severity is high, the original matrix element values ​​are multiplied by 1.5; if the anomaly severity is medium, the original matrix element values ​​are multiplied by 1.2; if the anomaly severity is low, the original matrix element values ​​are multiplied by 1.1. For the column containing the abnormal functional component, if other components in that column depend on the abnormal component, the matrix element values ​​are also adjusted according to the multiples corresponding to the anomaly severity. After adjusting the matrix elements corresponding to all abnormal components, the optimized component connection matrix of the functional components is obtained.

[0036] Furthermore, when constructing the weighted directed graph structure of functional components based on the optimized component connection matrix, the components corresponding to each row of the optimized component connection matrix are first used as the starting nodes of the directed graph, and the components corresponding to each column are used as the ending nodes of the directed graph. Then, the weights of the directed edges are determined according to the values ​​of the elements in the matrix. The values ​​of the matrix elements are the weights of the directed edges from the corresponding starting nodes to the ending nodes. For example, if the value of the element in the 2nd row and 3rd column of the optimized component connection matrix is ​​3.6, it means that the weight of the directed edge from the component node numbered 2 to the component node numbered 3 is 3.6. In this way, directed edges are constructed for all non-zero elements in the matrix, and the corresponding weights are labeled. All component nodes and the constructed directed edges are integrated together to form a complete graph structure, and finally, the weighted directed graph structure of the functional components is obtained.

[0037] Furthermore, when verifying the connectivity of a weighted directed graph structure, first select any component node in the weighted directed graph as the starting verification node. Starting from this node, sequentially visit other component nodes directly connected to it along the direction of the directed edges, recording all visited nodes. Then, using these visited nodes as new starting points, continue visiting their directly connected unrecorded nodes, repeating this process until no new nodes can be visited. Next, check whether all component nodes in the weighted directed graph have been recorded. If there are unrecorded nodes, start from those unrecorded nodes and repeat the above access and recording process. The process continues until all component nodes have been accessed and recorded or it has been confirmed that there are nodes that cannot be connected. If all component nodes can be connected through directed edges, the weighted directed graph structure is deemed to meet the connectivity requirements. If there are nodes that cannot be connected, and these nodes are not marked as abnormal in the anomaly list and belong to non-core functional components, the weighted directed graph structure is retained. If the nodes that cannot be connected are core functional components, the construction process of the component connection matrix is ​​re-examined and optimized, errors are corrected, the weighted directed graph structure is reconstructed, and connectivity verification is performed until a weighted directed graph structure that meets the connectivity requirements is obtained. This structure is the initial dependency graph of the functional components.

[0038] In summary, constructing a dependency topology by graphing the configuration information in functional components can transform the scattered configuration associations between functional components, including hardware interface dependencies, data interaction logic, and control command flow, into structured topological relationships. In the intelligent diagnostic system that integrates functional inspection and environment in this invention, this dependency topology can clearly present the association paths of each functional component, avoiding directional deviations in subsequent anomaly tracing due to ambiguity in the dependencies between components. It provides a precise structural foundation for the construction of the dependency graph, which aligns with the technical logic of this invention to systematically sort out the relationships between multiple components, ensuring that subsequent anomaly analysis can be carried out based on a clear component dependency framework.

[0039] In summary, extracting the component connection matrix based on the degree of correlation between functional components and the dependent topology can transform qualitative correlations in the topology into quantitative data representations. The numerical values ​​of matrix elements intuitively reflect the tightness of the correlation between components, providing a calculable quantitative basis for subsequent anomaly analysis. Anomaly enhancement processing of the component connection matrix based on the anomaly identifier list can specifically strengthen the connection weights between abnormal components and related components. For example, for faulty components marked in the anomaly identifier list, the connection values ​​between them and their upstream and downstream dependent components in the matrix can be increased, highlighting the correlation and radiation range of abnormal components. This makes the optimized component connection matrix more in line with the current anomaly diagnosis needs and avoids the neglect of key dependencies due to the failure to highlight anomaly correlations.

[0040] In summary, constructing a weighted directed graph structure for functional components based on the optimized component connection matrix transforms the quantitative relationships in the matrix into an intuitive graphical representation. The "weighted" characteristic reflects the strength differences in dependencies between components, while the "directed" characteristic clarifies the flow of dependencies. This approach retains the quantitative information of component relationships while providing convenient visualization, facilitating quick location of dependency chains for abnormal components. Connectivity verification of the weighted directed graph structure can identify isolated nodes or broken dependency paths, ensuring that the initial dependency graph fully covers the effective relationships between components. This avoids path interruptions during anomaly tracing due to connectivity issues, ensuring that the initial dependency graph accurately reflects the actual component dependencies and providing a reliable graphical tool for subsequent structural optimization and anomaly diagnosis.

[0041] In summary, structural optimization of the initial dependency graph to obtain the dependency graph of functional components can eliminate redundant associated paths in the initial graph, such as invalid connections without actual data interaction, and correct spurious dependencies caused by data errors. Furthermore, in conjunction with the environmental fusion diagnostic requirements of this invention, the dependency weights can be fine-tuned based on the current environmental parameters, such as increasing the dependency weights of heat dissipation components and core computing components in high-temperature environments. This optimization process improves the accuracy and practicality of the dependency graph, making it more suitable for the diagnostic scenario of "functional inspection and environmental fusion," ensuring that the final dependency graph can efficiently support subsequent diagnostic operations such as anomaly root cause location and fault impact range assessment, further improving the overall efficiency and accuracy of intelligent diagnosis.

[0042] S3. Based on the anomaly identifier list and the dependency graph, perform reverse propagation analysis on the dependency direction of the functional component to obtain the fault impact range of the functional component; In this embodiment of the invention, the step of performing backpropagation analysis on the dependency direction of the functional component based on the anomaly identifier list and the dependency graph to obtain the fault impact range of the functional component includes: The list of anomaly identifiers is processed to obtain the starting node of the dependency direction in the functional component; Based on the dependency graph and the starting node, a path traversal operation is performed in the opposite direction of the dependency direction to obtain the reverse dependency path of the functional component. Based on the reverse dependency path, a multi-level impact propagation analysis is performed on the functional components to obtain the fault impact nodes of the functional components. The scope of the fault is traced by tracking the affected nodes to obtain the fault impact range of the functional component.

[0043] In this embodiment of the invention, the step of performing multi-level impact propagation analysis on the functional component based on the reverse dependency path to obtain the fault impact nodes of the functional component includes: Record the propagation state and influence depth of the functional component to obtain the initial state table of the functional component; Based on the initial state table and the reverse dependency path, the functional component is backtracked to obtain the node to be analyzed in the functional component. Dependency strength is estimated for the nodes to be analyzed to obtain the strongly dependent nodes of the functional components; Based on the strongly dependent nodes, state propagation analysis is performed on the propagation state of the functional components to obtain the fault-affected nodes of the functional components.

[0044] Specifically, when processing the list of exception identifiers to determine the starting node of the dependency direction in the functional components, the information of each abnormal functional component recorded in the list of exception identifiers is first checked one by one, including the component identifier, exception type and exception occurrence time. Then, the abnormal functional component with the earliest exception occurrence time is selected. If there are multiple components with the same exception occurrence time, the dependencies of these components are further checked to determine which component is the direct dependency source of other abnormal components, that is, other abnormal components must depend on this component to run normally. The functional component that has the earliest exception and is the direct dependency source of other abnormal components is determined as the starting node of the dependency direction.

[0045] Furthermore, when performing path traversal operations in the opposite direction of the dependency direction based on the dependency graph and the starting node, the position corresponding to the starting node is first found in the dependency graph, and all the dependency directions pointed out by the node are identified, that is, the connection directions of other components that the node depends on. Then, starting from the starting node, the node moves in the opposite direction to these dependency directions, and visits the component nodes that have reverse dependency relationships with the starting node in turn. The identifier of the node and the access path are recorded for each node visited. This process continues until no new reverse dependency component nodes can be found. After organizing all the recorded access paths, the reverse dependency path of the functional component is obtained.

[0046] Furthermore, when performing multi-level impact propagation analysis on functional components based on reverse dependency paths to obtain fault-affected nodes, the reverse dependency paths are first divided into different levels. The level where the starting node is located is Level 1, the component nodes that have a direct reverse dependency relationship with the starting node are Level 2, the components that depend on the Level 2 component nodes are Level 3, and so on to determine the level of each component node. Then, starting from the starting node of Level 1, the impact of its abnormal state on the Level 2 component nodes is analyzed. It is checked whether the Level 2 components experience fluctuations in operating parameters due to the abnormality of the starting node. Based on the impact of the Level 2 components, the impact on the Level 3 component nodes is analyzed. The impact of the abnormality of each level of component nodes on the level above is analyzed layer by layer. All component nodes that have abnormal operation or have the risk of abnormal operation are marked. These marked nodes are the fault-affected nodes of the functional components.

[0047] Furthermore, when tracing the impact range of a fault-affected node to obtain the fault impact range of a functional component, detailed information about each fault-affected node is first collected, including the functional module, service objects, and data interaction range of the corresponding functional component. Then, the role of the functional module of each fault-affected node in the entire system is analyzed to determine the specific business scenarios served by the module and the user groups involved. Next, the data interaction between the fault-affected node and other non-faulty nodes is tracked to determine whether its anomaly will indirectly affect the normal operation of these non-faulty nodes. Finally, all functional modules, business scenarios, user groups, and related components directly or indirectly affected by the fault-affected node are integrated to form a complete impact range description, which is the fault impact range of the functional component.

[0048] Specifically, when recording the propagation status and influence depth of functional components to obtain the initial state table, the criteria for classifying propagation status are first clarified. Propagation status includes three types: normal, potentially abnormal, and abnormal. The normal status means that all operating parameters of the functional component are within the preset normal range. The potentially abnormal status means that some operating parameters of the functional component are close to the preset threshold but have not exceeded it. The abnormal status means that some operating parameters of the functional component exceed the preset threshold. At the same time, the definition of influence depth is determined. Influence depth refers to the hierarchical distance between the functional component and the starting node in the reverse dependency path. The influence depth of the starting node is 0, the influence depth of functional components directly connected to the starting node is 1, and so on. Then, each functional component is examined one by one to determine its current propagation status and the corresponding influence depth. The identifier, propagation status, and influence depth of each functional component are recorded in a table in sequence. Each row of the table corresponds to the relevant information of a functional component, and finally, the initial state table of functional components is formed.

[0049] Furthermore, when performing path backtracking on functional components based on the initial state table and the reverse dependency path to obtain the nodes to be analyzed, firstly, functional components whose propagation state is abnormal or potentially abnormal are selected from the initial state table, and the positions of these components in the reverse dependency path are determined. Then, starting from these components, backtracking is performed along the reverse dependency path in the opposite direction to the starting node. All functional components passed through are recorded during the process. At the same time, referring to the initial state table, functional components whose propagation state is clearly normal and whose influence depth differs from that of the abnormal components by more than 3 levels are excluded. The remaining recorded functional components are sorted out, and these components are the nodes to be analyzed for the functional components.

[0050] Furthermore, when estimating the dependency strength of the nodes to be analyzed to obtain strongly dependent nodes, the evaluation dimensions of dependency strength are first determined, including the data interaction frequency, criticality of data interaction, and dependency duration of the node on the upstream component in the reverse dependency path between the node to be analyzed and the upstream component. The data interaction frequency is measured by the number of interactions per day, the criticality of data interaction is determined based on whether the interaction data is core business data, and the dependency duration is measured by the continuous running time of the node depending on the upstream component. Then, for each node to be analyzed, it is evaluated from these three dimensions. If the number of data interactions between the node to be analyzed and the upstream component exceeds 500 times per day, the interaction data is core business data, and the continuous running time of the node depending on the upstream component exceeds 30 days, then the node to be analyzed is judged to have a high dependency strength with the upstream component. Nodes with such high dependency strength are marked as strongly dependent nodes of the functional component.

[0051] Furthermore, when performing state propagation analysis on the propagation status of functional components based on strongly dependent nodes to obtain fault-affected nodes, the current propagation status of each strongly dependent node is first analyzed. If the propagation status of a strongly dependent node is abnormal, the downstream functional components of that strongly dependent node are examined. Components located after that node in the reverse dependency path are checked to determine whether their operating parameters have changed due to the abnormality of the strongly dependent node. If the operating parameters of a downstream component change beyond the normal fluctuation range, its propagation status is updated to potentially abnormal or already abnormal. If the propagation status of a strongly dependent node is potentially abnormal, its downstream components are also examined. If the operating parameters of a downstream component are close to a preset threshold, its propagation status is updated to potentially abnormal. During the state propagation process, components whose propagation status changes or are already in an abnormal or potentially abnormal state are continuously recorded. These components are the fault-affected nodes of the functional components.

[0052] In summary, by processing the list of anomaly identifiers to determine the starting node, the source of fault propagation can be accurately located, avoiding deviations in the direction of subsequent reverse propagation analysis due to incorrect determination of the starting node, and laying a correct starting point for the analysis of the scope of fault impact.

[0053] In summary, traversing the execution path in the opposite direction of the dependency yields the reverse dependency path, which can clearly show the reverse path from which the fault may spread from the source, providing a clear path basis for subsequent multi-level impact analysis and reducing the blindness of the analysis.

[0054] In summary, multi-level impact propagation analysis based on reverse dependency paths can identify fault-affected nodes, systematically analyze the fault propagation process, comprehensively identify affected components, and avoid missing key fault-affected nodes.

[0055] In summary, tracing the scope of the impact of a fault by tracking the affected nodes can clarify the boundaries of the fault's impact from multiple dimensions, including functionality, business, and users, providing detailed information for developing targeted fault handling solutions.

[0056] In summary, by recording the propagation status and depth of influence to form an initial state table, we can comprehensively and clearly grasp the basic state information of each functional component, providing accurate initial data support for subsequent path backtracking and state transmission analysis, and avoiding analysis bias due to missing initial information.

[0057] In summary, by combining the initial state table and reverse dependency path to backtrack and filter the nodes to be analyzed, we can accurately focus on components that may be affected by the fault, exclude components that are clearly unaffected, reduce the workload of subsequent analysis, and improve analysis efficiency.

[0058] In summary, estimating the dependency strength of nodes under analysis from multiple dimensions and filtering out strongly dependent nodes can accurately identify components that play a key role in fault propagation, making subsequent state propagation analysis more targeted and avoiding wasting analysis resources on non-critical components.

[0059] In summary, state propagation analysis based on strongly dependent nodes can identify the nodes affected by a fault, systematically analyze the fault propagation process between components, comprehensively and accurately identify all components affected by the fault, and provide a reliable basis for subsequently determining the scope of the fault's impact.

[0060] S4. Based on the scope of influence, the degree of influence on the functional component is evaluated from multiple dimensions to obtain the criticality weight of the functional component; In this embodiment of the invention, the step of evaluating the degree of influence of the functional component based on the scope of influence, and obtaining the critical weight of the functional component, includes: Feature extraction is performed on the affected functional components within the scope of influence to obtain the operational feature data of the functional components; The operational feature data is analyzed from multiple dimensions to obtain the dimensional evaluation results of the functional components; Based on the evaluation results of the aforementioned dimensions, a contribution analysis is performed on the environmental operating status of the functional components to obtain the contribution coefficient of the functional components. The key weights of the functional components are obtained by multi-factor fusion of the dimensional evaluation results and the contribution coefficients.

[0061] Specifically, when extracting features from the affected functional components within the scope of impact, the specific scope of the affected functional components is first defined, that is, all functional components with operational abnormalities or potential abnormalities within the determined scope of the fault impact. Then, for each affected functional component, its operational data before and after the fault occurs is collected, including but not limited to the hourly response time, the resource utilization rate per minute, the data processing volume per 5 minutes, the frequency of error codes after the fault occurs, and the success rate of interaction with other components. These collected specific data that can reflect the operational status of the components are organized and summarized to finally obtain the operational feature data of the functional components.

[0062] Furthermore, when performing multi-dimensional analysis of operational feature data, the dimensions to be analyzed are first determined, including response efficiency, resource consumption, data processing, error frequency, and interaction stability. For the response efficiency dimension, the changes in response time in the operational feature data are analyzed to determine whether the component's response speed slows down under the influence of a fault. For the resource consumption dimension, resource utilization data is examined to determine whether the component is experiencing excessive or insufficient resource usage due to a fault. For the data processing dimension, the data processing volume data is used to assess whether the component's data processing capability has decreased. For the error frequency dimension, the frequency of error codes is statistically analyzed to understand the frequency of component fault errors. For the interaction stability dimension, interaction success rate data is used to determine whether the component's interaction with other components is stable. The analysis results for each dimension are graded, such as "Excellent," "Good," "Average," and "Poor." The grading results of each dimension are integrated to obtain the dimensional evaluation results of the functional component.

[0063] Furthermore, when analyzing the contribution of functional components to the environmental operation status based on the dimensional evaluation results, the overall operating environment in which the functional component resides and the core operating objectives of that environment are first clarified. For example, the core objective of a data processing system is to efficiently complete data calculation and transmission. Then, based on the dimensional evaluation results, the degree to which each functional component supports the core operating objectives in each dimension is determined. If a component is "excellent" in the response efficiency dimension and "good" in the data processing dimension, and its data processing capability directly affects the calculation progress of the entire system, it indicates that the component has a strong supporting role in the environmental operation status. If a component performs "poor" in multiple dimensions, and its operation status has little impact on the core objectives, then its supporting role is weak. Based on this difference in the degree of support, a specific value is assigned to each functional component to quantify its contribution. This value is the contribution coefficient of the functional component.

[0064] Furthermore, when performing multi-factor fusion of dimensional assessment results and contribution coefficients, the dimensional assessment results are first converted into corresponding numerical values, such as "excellent" corresponding to 5 points, "good" corresponding to 4 points, "average" corresponding to 3 points, and "poor" corresponding to 2 points. The average score of each dimension assessment result for each functional component is calculated. Then, the average score is multiplied by the contribution coefficient to obtain the preliminary fusion result. Then, the recovery difficulty of the component under the influence of the fault is combined. If the recovery difficulty is high, the weight is increased by 10% on the basis of the preliminary fusion result. If the recovery difficulty is low, the weight is reduced by 5%. In this way, the numerical values ​​corresponding to the dimensional assessment results, the contribution coefficient, and the recovery difficulty factor are integrated to finally obtain the key weight of the functional component.

[0065] In summary, by extracting features from the affected functional components within the scope of influence to obtain operational feature data, we can deeply capture the core operational information of functional components in the process of interacting with environmental factors. This includes key content such as performance fluctuation data, response latency data, and collaborative operation adaptation data of components under different environmental parameters. These data directly reflect the actual operating status of functional components under actual working conditions, providing accurate and comprehensive data support for subsequent multi-dimensional evaluation work. This effectively avoids the problem of evaluation benchmark deviation caused by incomplete or one-sided data collection, ensuring that the entire critical weight calculation process is reliable and accurate from the source.

[0066] In summary, multi-dimensional analysis of operational characteristic data and generation of dimensional evaluation results overcome the limitations of single-dimensional analysis in traditional evaluation methods. It comprehensively deconstructs and evaluates the operational status of components after being affected from multiple perspectives, including operational stability, performance compliance, environmental adaptability, and compatibility with other related components. For example, the operational stability dimension can analyze the frequency and amplitude of parameter fluctuations under environmental disturbances, while the environmental adaptability dimension can assess the component's adjustment response capability under different environmental compensation data. Through multi-dimensional cross-validation and analysis, the evaluation results can more comprehensively and objectively present the degree of impact on functional components, reducing evaluation errors that may arise from single-perspective analysis and laying a precise evaluation foundation for subsequent contribution analysis.

[0067] In summary, analyzing the contribution of functional components to the environmental operation status based on dimensional assessment results and obtaining contribution coefficients can clearly define the importance levels of different functional components in the overall equipment environmental operation system. By analyzing the role of components in maintaining the overall environmental adaptability of the equipment, ensuring the normal operation of other related components, and improving the overall operating efficiency of the equipment, it is possible to distinguish between core supporting components and auxiliary components that play a core supporting role in the equipment environmental operation status. This avoids treating all affected components equally during the assessment process, ensuring that the contribution coefficients can accurately reflect the value weight of components in the overall operation of the equipment. This provides differentiated and important parameters for subsequent multi-factor fusion, ensuring that the calculation of key weights can highlight the influence of core components.

[0068] In summary, by fusing dimensional assessment results with contribution coefficients through multiple factors to obtain the critical weights of functional components, a combination of the degree of impact on components and their own importance is achieved. This fusion process is not a simple data aggregation, but rather a deep coupling of the current impact status of components reflected in the dimensional assessment results with the core value of components in the overall operation of equipment as reflected in the contribution coefficients, through the establishment of a multi-factor correlation model. This ensures that the final critical weights can accurately quantify the degree of impact of components on environmental factors and fully reflect the importance of components to the overall operation of equipment. This provides a scientific and accurate decision-making basis for subsequent intelligent equipment diagnosis processes, including fault priority determination, optimized allocation of maintenance resources, and focused diagnostic efforts. It further enhances the practicality and reliability of the functional inspection and environmental fusion intelligent diagnosis and prediction method, effectively supporting the efficient implementation of preventive maintenance work.

[0069] S5. Based on the key weights and the scope of influence, perform weight aggregation on the functional components to obtain the severity index of the influence of the functional components; In this embodiment of the invention, the step of weighting the functional components based on the critical weights and the scope of influence to obtain an impact severity index for the functional components includes: Based on the scope of influence, the dependency chain of the functional component is resolved to obtain the dependency depth of the functional component; Based on the criticality weights and the dependency depth, construct the weight aggregation matrix of the functional components; Based on the weight aggregation matrix, the impact severity index of the functional component is calculated, wherein the calculation formula for the impact severity index is: ; in, This indicates the severity index of the impact. Indicates the first The aforementioned key weights, Indicates the first The aforementioned dependency depth This indicates the total number of affected functional components. This indicates the total number of the functional components. Represents the logarithmic function. It represents the base of the natural logarithm.

[0070] In this embodiment of the invention, the step of performing dependency chain resolution on the functional component based on the scope of influence to obtain the dependency depth of the functional component includes: Based on the dependency graph, extract the longest dependency path of the functional component; The longest dependency path is traversed to obtain the path length distribution data of the functional components; The path length distribution data is mapped to a preset depth range to obtain the standardized depth value of the functional component; The topological importance parameter of the functional component is normalized and aggregated with the normalized depth value to obtain the dependency depth of the functional component.

[0071] Specifically, when performing dependency chain analysis on functional components based on the scope of impact to obtain dependency depth, firstly, a detailed list of all functional components within the scope of impact is defined. Then, components with direct or indirect dependencies on the fault initiation node are selected from this list, constructing a complete dependency chain extending from the initiation node to each affected component. Next, the hierarchical structure of each dependency chain is determined: the level of the initiation node is recorded as level 0, the level of components directly dependent on the initiation node is recorded as level 1, the level of components dependent on level 1 components is recorded as level 2, and so on, determining the level of each component in the dependency chain. Then, for each functional component, all its dependency chains are found, and the level value corresponding to the component in each dependency chain is recorded. The maximum value among these level values ​​is taken as the initial depth value of the component. Finally, combined with the importance parameter of the component in the entire system topology, if the component is a key support component of the core business module, the initial depth value is increased by 1; if it is a non-core auxiliary component, the initial depth value remains unchanged, ultimately obtaining the dependency depth of the functional component.

[0072] Furthermore, when constructing the weight aggregation matrix of functional components based on key weights and dependency depth, all functional components within the scope of influence are first uniformly numbered to form a component sequence. The component number is used as the row and column index of the matrix to construct a two-dimensional matrix framework. Then, in the matrix, the key weight of each component is filled into the cell with the same row and column index, i.e., the cell at the diagonal position. For example, the key weight of component number 3 is filled into the cell at the 3rd row and 3rd column. Then, in the non-diagonal cell, the dependency association value between the corresponding row component and the corresponding column component is filled in. If the row component is a direct dependency of the column component, the association value is the product of the row component's key weight and the column component's dependency depth. If it is an indirect dependency, the association value is 0.5 times the product of the row component's key weight and the column component's dependency depth. If there is no dependency relationship, the association value is 0. The values ​​of all cells are filled in accordance with this rule, and the weight aggregation matrix of functional components is finally obtained.

[0073] Furthermore, when calculating the impact severity index of functional components based on the weight aggregation matrix, the values ​​of all non-zero cells in the weight aggregation matrix are first extracted. These values ​​include the criticality weight of the component itself and the dependency relationships between components. Then, the total number of affected functional components within the scope of influence is counted. and the total number of all functional components in the system. Next, process the values ​​in each non-zero cell by multiplying each value by 1 and adding the natural logarithm of the component dependency depth plus 1. Sum all the processed values ​​to get the numerator. Then calculate the sum of squares of the criticality weights of all components and take the square root of the sum as the denominator. Finally, calculate 1 minus the natural constant. of The value of the power is obtained by dividing the sum of the numerator by the denominator and then multiplying by the calculated result. Through this process, the severity index of the impact of the functional component is finally obtained.

[0074] Specifically, when extracting the longest dependency path of a functional component from the dependency graph, first, the nodes corresponding to each functional component and the directed edges representing the dependency relationships between nodes are identified in the dependency graph. All possible dependency paths that can extend from the fault initiation node to other affected component nodes are determined. Then, the number of directed edges contained in each dependency path is analyzed one by one. The more directed edges there are, the more component levels involved in the path and the longer the path length. By comparing the number of directed edges of all dependency paths, the path with the most directed edges is selected, which is the longest dependency path of the functional component.

[0075] Furthermore, when traversing the longest dependency path to obtain the path length distribution data of the functional components, starting from the starting node of the longest dependency path, each component node on the path is visited sequentially according to the dependency direction indicated by the directed edge. During the visit, the position number of each component node in the path is recorded. The position number of the starting node is recorded as 1, and the position number of each subsequent node is incremented by 1. At the same time, the number of component nodes corresponding to each position number is counted. If multiple component nodes are at the same position number in the longest dependency path, the position number and the corresponding number of nodes are recorded. All position numbers, corresponding nodes and node counts are organized into a data table, which is the path length distribution data of the functional components.

[0076] Furthermore, when mapping path length distribution data to a preset depth range to obtain the standardized depth value of a functional component, the preset depth range and the standardized value corresponding to each range are first determined. For example, the preset depth range is divided into [1-3], [4-6], and [7-9], with corresponding standardized values ​​of 0.3, 0.6, and 0.9, respectively. Then, the position number of each component node in the path length distribution data is checked to determine the preset depth range to which the position number belongs. The standardized value corresponding to the range is then assigned to the component node. If the position number of a component node is 2, which belongs to the [1-3] range, then its standardized value is 0.3. Standardized values ​​are assigned to all component nodes in this way, and these values ​​are the standardized depth values ​​of the functional component.

[0077] Furthermore, when normalizing and aggregating the topology importance parameters and standardized depth values ​​of functional components to obtain the dependency depth of functional components, the value range of the topology importance parameters is first determined. This range is determined by statistically analyzing the number of connections of all functional components in the dependency graph and whether they are core business nodes. For example, the value range of the topology importance parameters is [1-5]. The topology importance parameter of each component is divided by 5 to obtain the normalized topology importance value. At the same time, it is ensured that the standardized depth value is in the range of [0-1]. If the standardized depth value exceeds this range, it is adjusted. For example, values ​​greater than 1 are set to 1, and values ​​less than 0 are set to 0. Then, the normalized topology importance value is added to the standardized depth value, and then divided by 2. The result is the dependency depth of the functional component.

[0078] Specifically, the critical weights are derived from the critical weight data of the corresponding functional components in the weight aggregation matrix, the dependency depths are derived from the dependency depth data of the corresponding functional components in the weight aggregation matrix, the total number of affected functional components is obtained by counting the number of affected functional components in the weight aggregation matrix, and the total number of functional components is obtained by counting the total number of all functional components.

[0079] Furthermore, this calculation aims to obtain an impact severity index for functional components. By combining the criticality weight and dependency depth of the affected functional components, and considering the proportion of the number of affected functional components to the total number of functional components, the impact severity of the functional components is comprehensively quantified.

[0080] Furthermore, the severity index increases when the criticality weight of the affected functional components increases; the severity index increases when the dependency depth of the affected functional components increases; the severity index increases when the total number of affected functional components increases; and the severity index decreases when the total number of functional components increases, provided the total number of affected functional components remains unchanged.

[0081] In summary, determining dependency depth by combining dependency chain resolution with system topology importance can accurately reflect the position and importance of components in the fault propagation path, avoiding the bias in dependency depth judgment caused by relying solely on hierarchical values, and providing reasonable depth parameters for subsequent weight aggregation.

[0082] In summary, constructing a weight aggregation matrix based on criticality weights and dependency depth can integrate the importance of components and their dependency location information in matrix form, clearly presenting the correlation weight relationships between components and providing a structured data foundation for calculating the severity index.

[0083] In summary, calculating the severity index based on the weighted aggregation matrix can comprehensively consider the criticality of components, the depth of dependence, and the proportion of affected components. This allows the calculation results to fully reflect the degree of impact of the fault on the components, providing accurate quantitative basis for the formulation of subsequent fault handling solutions.

[0084] In summary, extracting the longest dependency path from the dependency graph allows us to focus on the critical path of fault propagation, ensuring that subsequent in-depth analysis is based on the most representative path and avoiding deviations in dependency depth calculation due to improper path selection.

[0085] In summary, traversing the longest dependency path to obtain path length distribution data allows for a clear understanding of the location distribution of each component in the critical path, providing detailed locational basis for subsequent mapping of standardized depth values ​​and ensuring the accuracy of the standardization process.

[0086] In summary, mapping path length distribution data to a preset depth range to obtain standardized depth values ​​can uniformly quantify the location of paths of different lengths, eliminate the problem of inconsistent depth measurement standards caused by differences in path length, and lay the foundation for subsequent aggregation operations.

[0087] In summary, normalizing and aggregating the topology importance parameters and standardized depth values ​​can comprehensively consider the importance of components in the system topology and the depth of their path location. This allows the final dependency depth to more fully reflect the actual impact of components in fault propagation, providing reliable parameters for subsequent calculation of impact severity indicators.

[0088] S6. Generate an environmental fault handling plan for the functional component based on the severity index of the impact.

[0089] In this embodiment of the invention, generating an environmental fault handling scheme for the functional component based on the severity index includes: The severity index of the impact is classified into levels to obtain the fault severity level sequence of the functional component; Based on the scope of influence and the critical weight, a strategy analysis is performed on the fault severity level sequence to obtain the initial fault response plan for the functional component. Resource conflict detection is performed on the initial fault response plan to obtain the environmental fault handling plan for the functional component.

[0090] Specifically, when classifying the severity index, the specific criteria for classification are first determined. These criteria are based on the correspondence between the severity index and the actual harm of the fault in historical fault handling data. For example, the severity index is divided into four levels: a value between 0 and 0.3 is considered "minor," between 0.3 and 0.6 is "moderate," between 0.6 and 0.8 is "relatively severe," and between 0.8 and 1.0 is "severe." Next, the severity index of each functional component is obtained one by one. By comparing it with the preset classification criteria, the fault severity level corresponding to each functional component is determined. Then, the fault severity levels of each component are arranged in order according to their numbering to form an ordered list. This list is the fault severity level sequence of the functional components.

[0091] Furthermore, when conducting strategy analysis on the fault severity level sequence based on the scope of impact and criticality weight to obtain the initial fault response plan, the impact scope of the functional components corresponding to each level is first examined in conjunction with the fault severity level sequence. If a component belongs to the "severe level" and its impact scope covers the core business modules, and its criticality weight is high, it indicates that the failure of this component has a significant impact on the system and needs to be prioritized. The response strategy is to immediately activate the backup component to replace the faulty component and arrange for technical personnel to arrive on-site within 1 hour to investigate the cause of the failure. If the component belongs to the "minor level" and its impact scope is limited to non-core auxiliary modules with a low criticality weight, the response strategy is to record the fault information and repair the fault during a system idle period. In this way, for each component with a fault severity level, corresponding handling strategies are formulated based on the size of its impact scope and the level of its criticality weight. After integrating all strategies, the initial fault response plan for the functional component is obtained.

[0092] Furthermore, when performing resource conflict detection on the initial fault response plan to obtain an environmental fault handling plan, first clarify the types and quantities of resources available for fault handling in the system, including the number of backup components, the number of technical personnel, the types and quantities of maintenance tools, and the time window required to handle the fault. Then, check the resources required for each handling strategy in the initial fault response plan one by one. If the handling strategies of two "severity level" components both require the use of the same unique backup server, or both require the same core technical personnel to handle them in the same time period, a resource demand conflict occurs. At this time, the processing order needs to be adjusted to prioritize the use of resources for the component with a larger impact and higher critical weight, postpone the processing time of the other component to a resource idle period, or allocate other alternative resources. After resolving all resource conflict issues, the adjusted fault response plan is checked again to confirm that there are no resource conflicts and that all handling strategies are executable, and finally, the environmental fault handling plan for the functional components is obtained.

[0093] In summary, by classifying the severity of impact indicators to obtain a fault severity level sequence, the severity of different component failures can be clearly distinguished, providing a clear basis for subsequent development of differentiated handling strategies and avoiding resource waste or critical fault delays caused by adopting a uniform handling method for all faulty components.

[0094] In summary, combining the scope of impact and criticality weights to conduct strategy analysis on the rank sequence to obtain an initial plan can ensure that the handling strategy matches the actual impact of the fault and the importance of the components, allowing high-impact and high-importance faults to receive priority attention and handling, and improving the pertinence and effectiveness of fault handling.

[0095] In summary, detecting and adjusting resource conflicts in the initial solution to arrive at the final solution can proactively avoid resource contention during the troubleshooting process, ensure the smooth execution of all troubleshooting strategies, prevent the solution from failing due to insufficient resources, and ensure that troubleshooting work is carried out in an orderly and efficient manner.

[0096] In the several embodiments provided by this invention, it should be understood that the disclosed method can be implemented in other ways.

[0097] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0098] The embodiments of this application can acquire and process relevant data based on artificial intelligence technology. Artificial intelligence is the theory, method, and technology that uses digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to obtain optimal results.

[0099] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method of analyzing the impact range of an environmental failure caused by a functional patrol drive, characterized by, The method includes: S1. Perform pattern matching between the performance parameters in the functional status data and the preset thresholds to obtain a list of abnormal identifiers for the functional components. S2. Construct a dependency graph of the functional components based on the list of exception identifiers; S3. Based on the anomaly identifier list and the dependency graph, perform reverse propagation analysis on the dependency direction of the functional component to obtain the fault impact range of the functional component; S4. Based on the scope of influence, the degree of influence on the functional component is evaluated from multiple dimensions to obtain the criticality weight of the functional component; S5. Based on the key weights and the scope of influence, perform weight aggregation on the functional components to obtain the severity index of the influence of the functional components; S6. Generate an environmental fault handling plan for the functional component based on the severity index of the impact.

2. The method of claim 1, wherein the functional walk-through driven environmental fault impact analysis method is characterized by, The step of performing pattern matching between performance parameters in the functional status data and preset thresholds to obtain an anomaly identifier list for the functional components includes: The performance parameters in the functional status data are normalized to obtain the standardized performance parameters of the functional components. The standardized performance parameters are compared with the preset threshold using a multi-rule pattern to obtain the dynamic adjustment parameters of the functional components. Based on the historical performance parameters of the functional components, trend analysis is performed on the dynamic adjustment parameters to obtain the abnormal behavior patterns of the functional components. Based on the abnormal behavior pattern, generate a list of abnormal identifiers for the functional components.

3. The method of claim 1, wherein the functional walk-through driven environmental fault impact analysis method is characterized by, The step of constructing the dependency graph of the functional components based on the list of exception identifiers includes: The configuration information in the functional components is constructed in a graph form to obtain the dependency topology of the functional components; Based on the dependency topology and the list of exception identifiers, construct the initial dependency graph of the functional components; The initial dependency graph is structurally optimized to obtain the dependency graph of the functional components.

4. The method for analyzing the impact range of environmental faults driven by functional inspection as described in claim 3, characterized in that, The step of constructing the initial dependency graph of the functional components based on the dependency topology and the anomaly identifier list includes: Based on the degree of association of the functional components, extract the component connection matrix of the dependent topology; Based on the list of anomaly identifiers, anomaly enhancement processing is performed on the component connection matrix to obtain the optimized component connection matrix of the functional component; Based on the optimized component connection matrix, construct the weighted directed graph structure of the functional components; The connectivity of the weighted directed graph structure is verified to obtain the initial dependency graph of the functional components.

5. The method for analyzing the impact range of environmental faults driven by functional inspection as described in claim 1, characterized in that, Based on the anomaly identifier list and the dependency graph, a backpropagation analysis is performed on the dependency direction of the functional component to obtain the fault impact range of the functional component, including: The list of anomaly identifiers is processed to obtain the starting node of the dependency direction in the functional component; Based on the dependency graph and the starting node, a path traversal operation is performed in the opposite direction of the dependency direction to obtain the reverse dependency path of the functional component. Based on the reverse dependency path, a multi-level impact propagation analysis is performed on the functional components to obtain the fault impact nodes of the functional components. The scope of the fault is traced by tracking the affected nodes to obtain the fault impact range of the functional component.

6. The method for analyzing the impact range of environmental faults driven by functional inspection as described in claim 5, characterized in that, The step of performing multi-level impact propagation analysis on the functional components based on the reverse dependency path to obtain the fault impact nodes of the functional components includes: Record the propagation state and influence depth of the functional component to obtain the initial state table of the functional component; Based on the initial state table and the reverse dependency path, the functional component is backtracked to obtain the node to be analyzed in the functional component. Dependency strength is estimated for the nodes to be analyzed to obtain the strongly dependent nodes of the functional components; Based on the strongly dependent nodes, state propagation analysis is performed on the propagation state of the functional components to obtain the fault-affected nodes of the functional components.

7. The method for analyzing the impact range of environmental faults driven by functional inspection as described in claim 1, characterized in that, Based on the scope of influence, the degree of impact on the functional component is evaluated from multiple dimensions to obtain the critical weight of the functional component, including: Feature extraction is performed on the affected functional components within the scope of influence to obtain the operational feature data of the functional components; The operational feature data is analyzed from multiple dimensions to obtain the dimensional evaluation results of the functional components; Based on the evaluation results of the aforementioned dimensions, a contribution analysis is performed on the environmental operating status of the functional components to obtain the contribution coefficient of the functional components. The key weights of the functional components are obtained by multi-factor fusion of the dimensional evaluation results and the contribution coefficients.

8. The method for analyzing the impact range of environmental faults driven by functional inspection as described in claim 1, characterized in that, The step of aggregating the weights of the functional components based on the critical weights and the scope of influence to obtain the severity index of the influence of the functional components includes: Based on the scope of influence, the dependency chain of the functional component is resolved to obtain the dependency depth of the functional component; Based on the criticality weights and the dependency depth, construct the weight aggregation matrix of the functional components; Based on the weight aggregation matrix, the impact severity index of the functional component is calculated, wherein the calculation formula for the impact severity index is: ； in, This indicates the severity index of the impact. Indicates the first The aforementioned key weights, Indicates the first The aforementioned dependency depth This indicates the total number of affected functional components. This indicates the total number of the functional components. Represents the logarithmic function. It represents the base of the natural logarithm.

9. The method for analyzing the impact range of environmental faults driven by functional inspection as described in claim 8, characterized in that, The process of resolving the dependency chain of the functional component based on the scope of influence to obtain the dependency depth of the functional component includes: Based on the dependency graph, extract the longest dependency path of the functional component; The longest dependency path is traversed to obtain the path length distribution data of the functional components; The path length distribution data is mapped to a preset depth range to obtain the standardized depth value of the functional component; The topological importance parameter of the functional component is normalized and aggregated with the normalized depth value to obtain the dependency depth of the functional component.

10. The method for analyzing the impact range of environmental faults driven by functional inspection as described in claim 1, characterized in that, The step of generating an environmental fault handling plan for the functional component based on the severity index includes: The severity index of the impact is classified into levels to obtain the fault severity level sequence of the functional component; Based on the scope of influence and the critical weight, a strategy analysis is performed on the fault severity level sequence to obtain the initial fault response plan for the functional component. Resource conflict detection is performed on the initial fault response plan to obtain the environmental fault handling plan for the functional component.

Images