Fault processing method and device, nonvolatile storage medium and electronic equipment

Abstract

The application discloses a fault processing method and device, a nonvolatile storage medium and an electronic device. The method comprises the following steps: obtaining target fault data of a target network system; analyzing the target fault data by using a preset association relationship model, determining at least one target root cause data of the target fault data and a target association probability of each target root cause data; arranging the at least one target root cause data according to the target association probability, and generating a fault root cause table; and querying a fault removal scheme corresponding to each target root cause data in the fault root cause table in a preset fault processing library, and generating a fault removal table. The application solves the technical problem that the prior art cannot perform fault processing according to the association relationship between data.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of computers, and in particular, to a fault processing method and device, a nonvolatile storage medium and an electronic device. BACKGROUND

[0002] At present, for IP data communication networks, the processing of network faults is mainly based on manual positioning and analysis, emergency processing, or logical analysis between alarm information, or statistical fault analysis through algorithms, and rarely can information correlation modeling and root cause analysis be performed from the system data.

[0003] The current abnormality of equipment hidden dangers is mainly based on manual processing by maintenance personnel, which depends on the professional ability and operation experience of maintenance personnel, and the processing rate is too low, the processing is often incomplete, and the processing flow is difficult to form a closed loop iteration.

[0004] And now for fault analysis and diagnosis, there are the following researches:

[0005] Some focus on associating the title library with the resume library, outputting rule templates and effect analysis, and are applied to the optimization of alarms and the compression of work orders. However, they are not focused on the logical association of alarms themselves, nor are they based on the association of alarms for reasoning and processing of alarms.

[0006] Some focus on the probability relationship between alarms, rather than the correlation between alarms obtained through logical analysis and operation data.

[0007] Some establish an association model through alarm periods and apply the association relationship thus established for processing.

[0008] Some perform vectorization operation on alarm content / words, and then perform operation on the word vector information to train a neural network model, which is an AI learning scheme for alarms. It is not based on expert processing and alarm logical association relationship processing.

[0009] Some perform abnormality detection and root cause analysis on collected information to assist in abnormality modification. It does not involve probability modeling and association model-based processing.

[0010] Some use clustering algorithms to aggregate alarms and use AI algorithms to analyze root causes. It does not involve probability modeling and association model-based processing.

[0011] In view of the problem that the above-mentioned prior art cannot perform fault processing according to the association relationship between data, no effective solution has been proposed so far. SUMMARY

[0012] Embodiments of the present application provide a fault processing method and device, a nonvolatile storage medium and an electronic device to at least solve the technical problem that the prior art cannot process faults according to the association between data.

[0013] According to an aspect of the embodiments of the present application, a fault processing method is provided, characterized in that it comprises: obtaining target fault data of a target network system; analyzing the target fault data by using a preset association relationship model to determine at least one target root cause data of the target fault data and a target association probability of each target root cause data, wherein the preset association relationship model is trained by using multiple sets of data through machine learning, each set of data in the multiple sets of data comprises: preset fault data causing a fault, preset root cause data associated with the preset fault data, and a preset association probability used to represent the association between the preset fault data and the preset root cause data; arranging at least one target root cause data according to the target association probability to generate a fault root cause table; and querying a fault removal scheme corresponding to each target root cause data in the fault root cause table in a preset fault processing library to generate a fault removal table.

[0014] Optionally, before analyzing the target fault data by using the preset association relationship model to determine at least one target root cause data of the target fault data and a target association probability of each target root cause data, the method further comprises: obtaining a first root cause data previously calibrated; obtaining first result data generated by the target network system executing the first root cause data and a first association probability of generating the first result data; taking the first result data as second root cause data, obtaining second result data generated by the target network system executing the second root cause data and a second association probability of generating the second result data; in the case that the second result data is preset fault data, determining that the first root cause data is preset root cause data of the preset fault data; and determining a preset association probability of the preset root cause data according to the first association probability and the second association probability.

[0015] Optionally, before analyzing the target fault data by using the preset association relationship model to determine at least one target root cause data of the target fault data and a target association probability of each target root cause data, the method further comprises: obtaining a preset fault time of preset fault data; taking the preset fault time as a starting point, obtaining at least one upstream data generated by the target network system within a preset time period before the preset fault time; determining the preset root cause data of the preset fault data in at least one upstream data; and determining a preset association probability of the preset root cause data by using the frequency of occurrence of the preset root cause data in at least one upstream data.

[0016] Optionally, before analyzing the target fault data by the preset correlation model to determine at least one target root cause data of the target fault data and a target correlation probability of each target root cause data, the method further comprises: obtaining preset location information of a fault device indicated by the preset fault data; determining at least one preset region data in a preset region space where the preset location information is located, wherein the preset region data is used to represent an operation parameter corresponding to at least one preset region device in the preset region space; determining preset root cause data of the preset fault data in at least one preset region data; and determining a preset correlation probability of the preset root cause data according to a distance value between a preset region device corresponding to the preset root cause data and the fault device.

[0017] Optionally, before analyzing the target fault data by the preset correlation model to determine at least one target root cause data of the target fault data and a target correlation probability of each target root cause data, the method further comprises: obtaining preset parameter data of a fault device indicated by the preset fault data; taking the preset parameter data as preset root cause data of the preset fault data; determining a target parameter threshold value matched with the preset parameter data in a preset threshold value library, wherein a plurality of preset parameter threshold values and a preset correlation probability corresponding to each preset parameter threshold value are pre-set in the preset threshold value library; and determining the preset correlation probability corresponding to the target parameter threshold value.

[0018] Optionally, after querying a fault removal scheme corresponding to each target root cause data in the fault root cause table in the preset fault processing library to generate a fault removal table, the method further comprises: receiving a fault removal report of the target fault data, wherein the target root cause data of the target fault data is carried in the fault removal report; and adjusting the preset correlation model according to the target fault data and the associated target root cause data.

[0019] Optionally, arranging at least one target root cause data according to the target correlation probability to generate a fault root cause table comprises: arranging the at least one target root cause data according to an order from high to low of the target correlation probability to generate the fault root cause table.

[0020] According to another aspect of the embodiments of the present application, there is also provided a fault processing apparatus, characterized by comprising: an acquisition module configured to acquire target fault data of a target network system; an analysis module configured to analyze the target fault data by using a preset correlation relationship model to determine at least one target root cause data of the target fault data and a target correlation probability of each target root cause data, wherein the preset correlation relationship model is trained by using a plurality of sets of data through machine learning, each set of data in the plurality of sets of data comprising: preset fault data causing a fault, preset root cause data associated with the preset fault data, and a preset correlation probability used to represent a correlation relationship between the preset fault data and the preset root cause data; a first generation module configured to arrange the at least one target root cause data according to the target correlation probability to generate a fault root cause table; and a second generation module configured to query a fault removal scheme corresponding to each target root cause data in the fault root cause table in a preset fault processing library to generate a fault removal table.

[0021] According to another aspect of the embodiments of the present application, there is also provided a nonvolatile storage medium, characterized by storing a program, wherein the program, when executed, controls a device in which the nonvolatile storage medium is located to perform the fault processing method.

[0022] According to another aspect of the embodiments of the present application, there is also provided an electronic device, characterized by comprising a memory and a processor, wherein the processor is configured to execute a program stored in the memory, and the program, when executed, performs the fault processing method.

[0023] In the embodiments of the present application, target fault data of a target network system is acquired; the target fault data is analyzed by using a preset correlation relationship model to determine at least one target root cause data of the target fault data and a target correlation probability of each target root cause data, wherein the preset correlation relationship model is trained by using a plurality of sets of data through machine learning, each set of data in the plurality of sets of data comprising: preset fault data causing a fault, preset root cause data associated with the preset fault data, and a preset correlation probability used to represent a correlation relationship between the preset fault data and the preset root cause data; the at least one target root cause data is arranged according to the target correlation probability to generate a fault root cause table; and a fault removal scheme corresponding to each target root cause data in the fault root cause table is queried in a preset fault processing library to generate a fault removal table, thereby achieving the technical effect of performing fault processing according to the correlation relationship between data, and further solving the technical problem that the prior art cannot perform fault processing according to the correlation relationship between data. BRIEF DESCRIPTION OF DRAWINGS

[0024] The accompanying drawings, which are included to provide a further understanding of the application and are incorporated in and constitute a part of this application, illustrate embodiments of the application and together with the description serve to explain the application. In the drawings:

[0025] Figure 1 is a flow chart of a fault processing method according to an embodiment of the application;

[0026] Figure 2 is a schematic diagram of a fault analysis system according to an embodiment of the application;

[0027] Figure 3 is a schematic diagram of data associating configuration information according to an embodiment of the application;

[0028] Figure 4 is a schematic diagram of data associating configuration information according to an embodiment of the application;

[0029] Figure 5 is a schematic diagram of a fault processing apparatus according to an embodiment of the application;

[0030] Figure 6 is a structural block diagram of a computer terminal according to an embodiment of the application. DETAILED DESCRIPTION

[0031] In order to make the persons skilled in the art better understand the application scheme, the technical scheme in the embodiments of the application will be described clearly and completely below in conjunction with the drawings in the embodiments of the application. Obviously, the described embodiments are only a part of the embodiments of the application, but not all the embodiments. Based on the embodiments in the application, all the other embodiments obtained by the persons skilled in the art without creative work should belong to the protection scope of the application.

[0032] It should be noted that the terms "first", "second", and the like in the specification and claims of the application and the above-mentioned drawings are used to distinguish similar objects, and do not have to be used to describe a specific order or sequence. It should be understood that the data used in this way can be exchanged under appropriate circumstances, so that the embodiments of the application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion, for example, a process, method, system, product or device including a series of steps or units does not have to be limited to those steps or units clearly listed, but can include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0033] According to an embodiment of the present application, a fault processing method is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from here.

[0034] Figure 1 is a flowchart of a fault processing method according to an embodiment of the present application, as shown in Figure 1 the method comprises the following steps:

[0035] Step S102, obtaining target fault data of a target network system;

[0036] Step S104, analyzing the target fault data by a preset correlation relationship model to determine at least one target root cause data of the target fault data and a target correlation probability of each target root cause data, wherein the preset correlation relationship model is trained by machine learning using a plurality of sets of data, each set of data in the plurality of sets of data comprising: preset fault data causing a fault, preset root cause data associated with the preset fault data, and a preset correlation probability for indicating a correlation between the preset fault data and the preset root cause data;

[0037] Step S106, arranging the at least one target root cause data according to the target correlation probability to generate a fault root cause table;

[0038] Step S108, querying a fault removal scheme corresponding to each target root cause data in the fault root cause table in a preset fault processing library to generate a fault removal table.

[0039] In an embodiment of the present application, target fault data of a target network system is obtained; the target fault data is analyzed by a preset correlation relationship model to determine at least one target root cause data of the target fault data and a target correlation probability of each target root cause data, wherein the preset correlation relationship model is trained by machine learning using a plurality of sets of data, each set of data in the plurality of sets of data comprising: preset fault data causing a fault, preset root cause data associated with the preset fault data, and a preset correlation probability for indicating a correlation between the preset fault data and the preset root cause data; the at least one target root cause data is arranged according to the target correlation probability to generate a fault root cause table; a fault removal scheme corresponding to each target root cause data in the fault root cause table is queried in a preset fault processing library to generate a fault removal table, thereby achieving the technical effect of fault processing according to the correlation between data, and further solving the technical problem that the prior art cannot perform fault processing according to the correlation between data.

[0040] It should be noted that the fault processing method can be executed in a fault analysis system.

[0041] Figure 2 is a schematic diagram of a fault analysis system according to an embodiment of the present application, as shown, the system mainly comprises: data definition module, data association module, expert AI library module; data acquisition module, fault analysis module, fault handling module and result feedback module. Figure 2

[0042] Optionally, the system can be divided into an association layer and a processing layer, wherein the association layer comprises: data definition module, data association module, expert AI library module and result feedback module; the processing layer comprises: data acquisition module, fault analysis module and fault handling module.

[0043] As an optional example, the data definition module is used to define the data that needs to be analyzed by the system, and in the full data space, according to logical and empirical analysis, the data characteristics are labeled and the data is defined. That is, the data that form a correlation relationship are divided into associated source data (such as root cause data) and associated data (such as fault data); wherein the source data (such as root cause data) is the cause or possible cause or impossible cause of the associated data. The same source data (such as root cause data) can have multiple associated data; that is, one cause can lead to multiple results; the same associated data (such as fault data) can have multiple source data (such as root cause data), such as one result, which can be caused by different reasons, or after the result appears, which situations cannot appear.

[0044] Optionally, the data definition module also defines the fault removal operation (such as fault removal scheme) corresponding to each source data (such as root cause data). For example, when the source data (such as root cause data) indicates that the interface is configured as DOWN, the related fault removal operation (such as fault removal scheme) is to configure UP; when the source data (such as root cause data) is that the interface has not been configured with an IP address, the related fault removal operation (such as fault removal scheme) is to obtain the correct IP address configuration. The fault removal operation (such as fault removal scheme) can be an operation flow or scheme composed of multiple operations.

[0045] Optionally, the data definition module also includes: for each associated data (such as fault data), when the corresponding source data (such as root cause data) state or value size needs to be determined, the detection determination operation required.

[0046] For example, when there is a traffic packet loss, the interface traffic congestion is obtained to determine whether the packet loss is caused by traffic congestion.

[0047] ​It should be noted that the definition of data by the data definition module does not need to completely define all the correlations of all data at the beginning. A feasible solution is to first correlate the explicit and direct data, and then continuously iterate and optimize through the expert AI library module and result feedback.

[0048] As an optional example, the expert AI library module is the expert experience and AI data analysis library for system operation and maintenance. Based on system historical data and the latest collected data, the module makes expert (or experience) judgments and conducts AI data analysis to regularly optimize the correlation and correlation measure of the defined associated data (i.e., root cause data and fault data). It analyzes the direct operation suggestions when the corresponding information causes a fault.

[0049] For example, through the expert experience of operation and maintenance experts, the correlation between source data (such as root cause data) and associated data (such as fault data) is increased or decreased, including the relationship between source data (such as root cause data) and associated data (such as fault data) that cannot cause a causal relationship, i.e., the correlation probability of source data (such as root cause data) and associated data (such as fault data) is 0. At the same time, through the AI algorithm (such as based on the frequency of historical data), the correlation measure probability value (such as correlation probability) of the associated source data (such as root cause data) and associated data (such as fault data) is optimized and adjusted.

[0050] Optionally, for the expert AI library module, in the present application, it is not necessary to optimize the data correlation probability in a large amount and very frequently. A feasible solution is to optimize the probability value (i.e., correlation probability) of the associated data (i.e., root cause data and fault data) as necessary and continuously after the accumulation of expert experience; after a certain period of time of system data accumulation, the associated data probability value (i.e., correlation probability) is optimized by a general AI algorithm for reference, so as to reduce the workload of manually adjusting the correlation probability between different data sources.

[0051] As an optional example, the data correlation module is used to correlate data based on data characteristics and mark the correlation degree. The correlation management between the associated data (such as fault data) and the source data (such as root cause data) can be divided into four levels: certain, possible, small probability, and impossible, and the correlation probability is used for data modeling. The modeling process is equivalent to the establishment of a data correlation tree and a correlation forest.

[0052] As an optional example, the data collection module is used to collect the configuration, state, performance, traffic, alarm, operation, and log digital information of the system. These data can be considered as Key-Value to define specific indicators of the related system. At the same time, the module will also save the related data for a long time.

[0053] As an optional example, the fault analysis module is used to analyze the fault information generated by the system based on the data correlation data model (such as the preset correlation model), that is, the module is an application module of the data correlation data model (such as the preset correlation model). The root cause of the system fault is analyzed (such as the most initial hardware or configuration data that causes the fault). The root cause includes possible reasons, small probability reasons, and impossible cases.

[0054] Optionally, the fault analysis module can obtain corresponding operation suggestions according to the root cause analysis of the fault. For example, for the fault of traffic congestion, the possible reason is that the interface member port is down, which causes the interface forwarding bandwidth to decrease; it may be caused by temporary import of other service traffic to the interface, or it may be caused by burst of part of the service traffic in the related service. Then, according to the probability values defined and optimized before, the fault root cause is sorted, and then for each possible reason, the corresponding processing scheme and processing process are associated.

[0055] As an optional example, the fault processing module is used to obtain all root causes (such as root cause data) of the fault to be processed (such as fault data) from the fault analysis module, and the root causes (such as root cause data) have been sorted according to the probability values (such as correlation probability) of the root causes (such as root cause data); then according to the sorted root causes, the suggestions and steps of the root causes for removing the fault are proposed; and the operation steps and schemes are merged and optimized (such as whether repeated operations can be executed only once, whether associated operations have a sequence, whether related operations need to be restored before executing subsequent operations because they are mutually exclusive operations, and this optimization can be handled according to expert experience).

[0056] Optionally, the fault processing module performs corresponding fault processing according to the analysis results of the associated data (such as fault data) and the corresponding operation suggestions / schemes, and the operation and maintenance strategy of the system. For example, the controller can directly control the recovery instruction to be issued to the device, or the maintenance personnel can be guided to operate, information prompt, short message alarm, manufacturer support matrix information, and other ways to perform targeted processing on the fault of the system.

[0057] As an optional example, the result feedback module is used to summarize the real root cause of the fault after the fault handling, that is, among the many root causes (i.e., preset root cause data), which root cause (such as target root cause data) causes the related fault (such as target fault data) is marked. For the root cause that causes the fault, it can be analyzed whether the probability value (such as association probability) associated with the data can be increased accordingly. If other root causes do not cause the fault multiple times, it can be analyzed whether the probability value (such as association probability) associated with the data can be reduced accordingly. And analyze the efficiency of the operation steps, that is, the specific fault removal operation (i.e., fault removal scheme) is performed, and which operations are redundant, and whether these operations can be optimized.

[0058] Optionally, after the result feedback analyzes the fault handling result, the analysis result can be fed back to the data definition module or the data association module.

[0059] Optionally, the result feedback module adjusts and optimizes the association information (i.e., the association probability and the association relationship in the association forest model) and the operation strategy / scheme of the data association forest model through analysis and feedback of the fault handling result, including increasing the association probability value, reducing the association probability, increasing or deleting impossible information (i.e., the probability value of 0 measurement), and also including increasing operations, operation step optimization, and operation merging.

[0060] Optionally, in the result feedback module, the present application does not require that each result is forced to be fed back and the association probability is adjusted. A feasible scheme is to perform necessary iterative adjustment when there is a significant difference between the fault handling result analysis and the system advancing root cause. At this time, the adjustment needs to mainly adjust the association relationship and the association measurement value, which is obtained according to which modeling method, and then adjust the measurement value or the measurement parameter or the calculation function of the specific method, and set the weight value of the combination of various methods.

[0061] Optionally, at the association layer, the definition of the data definition module is applied to establish the basic association relationship model (i.e., the preset association relationship model) of the data, and the data definition and data association are optimized through the expert AI library module. For the system running result, the result feedback module analyzes and outputs the analysis result to support the optimization of the association model information of the data definition and data association.

[0062] Optionally, at the processing layer, the data collection module collects complete system data, which is applied when the system generates a fault and the fault analysis module analyzes. The result obtained by data analysis is applied to the fault handling module.

[0063] Optionally, between the association layer and the processing layer, the expert AI library module can apply the system data collected by the data collection module for analysis and on-demand optimization; the result feedback module can apply the result information of fault processing for analysis and on-demand feedback; and the core module of the present scheme, i.e., the data association module, generates the associated forest data model for maintenance, which is applied to the fault analysis module when network fault processing is performed to analyze the fault root cause and processing steps.

[0064] In the above step S102, the target network system is an IP SDN network system, a communication network system, an IT network system, or a digital system.

[0065] As an optional embodiment, before the target fault data is analyzed by the preset association relationship model to determine at least one target root cause data of the target fault data and a target association probability of each target root cause data, the method further comprises: obtaining a first root cause data labeled in advance; obtaining first result data generated by the target network system executing the first root cause data, and a first association probability of generating the first result data; taking the first result data as a second root cause data, obtaining second result data generated by the target network system executing the second root cause data, and a second association probability of generating the second result data; in the case that the second result data is the preset fault data, determining that the first root cause data is a preset root cause data of the preset fault data; and determining a preset association probability of the preset root cause data according to the first association probability and the second association probability.

[0066] In the above embodiment of the present application, the target fault data and the target root cause data can include multiple sets of association relationships, so that in the case of determining the association relationship between the target fault data and the target root cause data, the target root cause data can be started from, multiple association relationships can be traversed until the target fault data is obtained, and the association relationship between the target fault data and the target root cause data can be determined.

[0067] As an optional example, starting from the fault root cause defined by the data (such as the first root cause data), the association data (such as the first result data) directly associated with the root cause data (such as the first root cause data) of the system fault is defined first, and then the next layer of associated data (the second result data) is associated from this layer of association data (the first result data) as the source data (i.e., the first result data is taken as the second root cause data). In turn, the fault alarm information presented by the system is finally associated, thereby generating the mathematical model of the association data forest based on the system fault root cause (such as the root cause data) and the fault association probability (such as the preset association probability).

[0068] Optionally, in the complete association relationship of the full data domain, it is actually a mesh forest structure starting from root cause data (these points can be regarded as a root cause data surface). Because a data, as a source data, can have multiple associated data, and as associated data, it can have multiple source data with different probability values. When initializing the association probability value, reference can be made to previous operation and maintenance experience values or expert judgment values.

[0069] Optionally, a plurality of association relationships are included between the target fault data and the target root cause data, a corresponding weight can be assigned to the association probability of each association relationship, and then the target association probability between the target fault data and the target root cause data is determined according to the plurality of association probabilities and the corresponding weights between the target fault data and the target root cause data.

[0070] As an optional embodiment, before the target fault data is analyzed by the preset association relationship model to determine at least one target root cause data of the target fault data and the target association probability of each target root cause data, the method further comprises: obtaining a preset fault time of a preset fault data; obtaining at least one upstream data generated by the target network system within a preset time period before the preset fault time, taking the preset fault time as a starting point; determining a preset root cause data of the preset fault data in the at least one upstream data; and determining a preset association probability of the preset root cause data by the frequency of occurrence of the preset root cause data in the at least one upstream data.

[0071] The above embodiments of the application can obtain upstream data generated within a preset time period before the occurrence of the preset fault data according to the time dimension, and then determine the preset root cause data of the preset fault data according to the upstream data generated within the preset time period, thereby realizing the determination of the association relationship between the preset fault data and the preset root cause data and the determination of the preset association probability in the time dimension.

[0072] As an optional embodiment, the association in the time dimension is performed by labeling parameters. The basic method is: in the 1-3 performance, state periods (such as 5 minutes) before the generation of a fault alarm (i.e. a preset fault data), the state transition, performance overrun, traffic overrun, alarm event and other information are regarded as having an association relationship with the fault alarm, and the measurement method is to calculate by the preset gradual function limx→∞f(x)=1 and the frequency parameter.

[0073] For example, in the 3 collection periods before the occurrence of a service packet loss overrun alarm, the interface traffic increase, sub-interface down, configuration change and the like are regarded as having an association with the fault.

[0074] Optionally, the probability measurement value (i.e. the preset association probability) of the association is calculated by the gradual function The calculation is that x is the frequency number of occurrence. In this way, the artificial person only needs to set the previous data time length of the fault association, select the probability calculation calculation function, and set the association degree threshold, such as 25%, so that when a fault occurs, the previous data can be automatically associated and labeled by the system, that is, the information exceeding the association threshold calculated according to the probability calculation tool can be used as the associated data of the fault, and the association probability value is the calculated probability value. Moreover, subsequent artificial persons only need to select a few parameter values and progressive functions, without a large amount of manual processing.

[0075] As an optional embodiment, before the target fault data is analyzed by the preset association relationship model to determine at least one target root cause data of the target fault data and a target association probability of each target root cause data, the method further includes: obtaining preset location information of a fault device indicated by the preset fault data; determining at least one preset area data in a preset area space where the preset location information is located, wherein the preset area data is used to represent running parameters corresponding to at least one preset area device in the preset area space; determining a preset root cause data of the preset fault data in the at least one preset area data; and determining a preset association probability of the preset root cause data according to a distance value between a preset area device corresponding to the preset root cause data and the fault device.

[0076] The above embodiments of the present application can obtain a preset area space of a device indicated by preset fault data according to a spatial dimension, and then determine a preset root cause data of the preset fault data according to running parameters of other preset area devices in the preset area space, so as to realize determination of an association relationship between the preset fault data and the preset root cause data and determination of a preset association probability in the spatial dimension.

[0077] As an optional example, the association in the spatial dimension is performed by using a labeling parameter, and the basic method is as follows: after obtaining a root cause data by processing corresponding fault (such as fault data), the system analyzes the location of the physical space where the root cause data is generated and the location of the physical space where the fault is generated, and then performs the same type of reasoning association.

[0078] For example, a device interface fault is analyzed to be caused by insufficient voltage of a machine room where the device is located, so that an association relationship between the interface of all devices involved in the machine room and the power supply of the machine room is established.

[0079] Optionally, the association degree (that is, the preset association probability) can be described by using a probability function (such as a Lorentz linear function ) related to the spatial position, wherein x is a distance value of the associated data.

[0080] It should be noted that a fault unrelated to spatial information can be regarded as having an association probability of 0.

[0081] Optionally, when a fault is located, all spatial position information data can be automatically associated and labeled by the system, that is, after the system analyzes (extracts) the spatial position information, information exceeding the association threshold calculated according to the probability calculation tool can be used as the associated data of the fault, and the association probability value is the calculated probability value. Moreover, subsequent manual work only needs to select these parameter values and association functions, without a large amount of manual processing.

[0082] As an optional embodiment, before determining at least one target root cause data of the target fault data and a target association probability of each target root cause data by analyzing the target fault data through the preset association relationship model, the method further includes: obtaining preset parameter data of a fault device indicated by the preset fault data; taking the preset parameter data as preset root cause data of the preset fault data; determining a target parameter threshold matching the preset parameter data in a preset threshold library, wherein a plurality of preset parameter thresholds and a preset association probability corresponding to each preset parameter threshold are pre-set in the preset threshold library; and determining the preset association probability corresponding to the target parameter threshold.

[0083] In the above embodiments of the application, the preset parameter data of the fault device can be obtained, and then the preset parameter data is matched with the preset parameter threshold. If the preset parameter data belongs to a first preset parameter threshold, the first preset association probability corresponding to the first preset parameter threshold is determined as the preset association probability of the preset parameter data and the preset fault parameter. If the preset parameter data belongs to a second preset parameter threshold, the second preset association probability corresponding to the second preset parameter threshold is determined as the preset association probability of the preset parameter data and the preset fault parameter.

[0084] As an optional embodiment, the association in the physical dimension is performed through the labeled parameters. After the root cause data is obtained through processing, the system automatically analyzes the physical characteristics of one manufacturer, product, version, and physical transmission characteristics corresponding to the fault, and performs the same reason analysis on the data.

[0085] For example, if the business error code exceeds the limit, the root cause data indicates that the error code of the interface exceeds the limit, all related manufacturers corresponding to the same batch of products, or products using the same interface components, or products of the same version can be labeled, and the association information with the fault can be established. Moreover, the association probability calculation can also refer to the progressive function and the association threshold. In this way, when a fault is located, the system can automatically associate and label according to the physical dimension information of the root cause, that is, information exceeding the association threshold calculated according to the probability calculation tool can be used as the associated data of the fault, and the association probability value is the calculated probability value. Moreover, subsequent manual work only needs to select these parameter values and the progressive function, without a large amount of manual processing.

[0086] Optionally, when a pair of correlation relationship and correlation degree are generated by multiple methods, the final correlation degree can be combined by considering the weight of various methods, and the weight values of various methods are supported to be preset and adjusted.

[0087] As an optional embodiment, after querying the fault removal scheme corresponding to each target root cause data in the fault root cause table in the preset fault processing library and generating the fault removal table, the method further comprises: receiving a fault removal report of the target fault data, wherein the target fault removal report carries the target root cause data of the target fault data; and adjusting the preset correlation relationship model according to the target fault data and the associated target root cause data.

[0088] According to the above embodiments of the present application, the target fault removal report carries the target root cause data of the target fault data, the target correlation relationship between the target root cause data and the target fault data can be verified according to the feedback fault removal report, and the preset correlation relationship model can be adjusted according to the verification result, so that the preset correlation relationship model can more accurately determine the target root cause data associated with the target fault data in the subsequent use process.

[0089] Optionally, by counting multiple fault removal reports, the target correlation probability of the target fault data and the target root cause data can be summarized, and the preset correlation relationship model can be adjusted based on the target fault data and the target root cause data in the multiple fault removal reports, so that the target correlation probability can be accurately evaluated.

[0090] As an optional embodiment, the generating of the fault root cause table according to the target correlation probability of the at least one target root cause data comprises: arranging the at least one target root cause data in descending order of the target correlation probability to generate the fault root cause table.

[0091] According to the above embodiments of the present application, the at least one target root cause data is arranged in descending order of the target correlation probability to generate the fault root cause table, so that the target root cause data with the highest probability can be placed at the front end of the fault root cause table, and the fault indicated by the target fault data can be quickly removed according to the fault root cause table.

[0092] The present application also provides a preferred embodiment, which provides a correlation processing scheme for network fault processing.

[0093] The technical scheme provided by the present application performs correlation determination and logical processing on the configuration, state, flow, performance, alarm and other information of the system in the complete information set of the system on the basis of collecting the full information of the network.

[0094] The technical scheme provided by the application can analyze the root cause of a fault from the perspective of system full data through data correlation, improve fault positioning efficiency by finding the root cause of the fault as soon as possible, analyze the operation of fault processing through the operation suggestion corresponding to the data correlation information and the correlation metric, give the step suggestion of the operation and maintenance operation, thereby improving the efficiency of system operation and maintenance, further improving the automation capability of network operation and maintenance, improving the self-healing capability of the system, reducing the maintenance difficulty and maintenance cost, and achieving good network intelligent operation and maintenance function.

[0095] The technical scheme provided by the application can establish the correlation relationship of data (such as fault data and root cause data) through data definition for system full data, establish the general causal relationship data model (such as a preset correlation relationship model) of the data through the probability metric (i.e., the correlation probability) of the correlation relationship, analyze the direct operation suggestion (such as a fault removal scheme) of the fault through the analysis of the system data by the expert AI module, optimize the probability metric (such as the correlation probability) of the correlation relationship, analyze the direct operation suggestion (such as a fault removal scheme) of the fault, optimize the data definition and data correlation relationship (such as the correlation relationship of the fault data and the root cause data) of the data model (such as the preset correlation relationship model) through result feedback analysis, and optimize the operation strategy at the same time, that is, through the correlation data model (such as the preset correlation relationship model) and the method of system closed-loop feedback processing, the iterative optimization and evolution of the system are formed.

[0096] The technical scheme provided by the application can be applied to IP SDN network operation and maintenance, communication network system operation and maintenance, IT network system operation and maintenance, and digital system operation and maintenance.

[0097] Optionally, in the IP SDN network, in the data definition module, the root cause data of the correlation data is defined as a router hardware interface fault, and the associated data (i.e., the fault data) is selected as the data directly reflected to the operation and maintenance, such as backbone trunk fault, device hardware fault, single-end trunk fault and VPN trunk fault. That is, the associated data should be in a direct causal relationship with the source data, and the correlation alarm relationship in the MIB alarm can be referred to.

[0098] Figure 3 It is a schematic diagram of data correlation hardware information according to an embodiment of the application, as shown in Figure 3 The source data: hardware interface down, interface sending error, such as traffic overrun, can be defined first.

[0099] Optionally, for the root cause data "hardware interface down", the associated data (such as first result data or fault data) such as interface protocol down, interface traffic packet loss, and trunk link down can be associated. When the interface protocol down is used as the source data of the next layer (such as the second root cause data), the associated data (such as the second result data or fault data) such as SRv6 interface failure, VPN interface failure, and backbone trunk failure can be associated.

[0100] Optionally, for the root cause data "interface sending error", the associated data (such as first result data or fault data) such as interface error code and interface traffic packet loss can be associated. When the "interface error code" is used as the source data of the next layer (such as the second root cause data), the fault information (such as the second result data or fault data) such as poor VPN service quality can be associated.

[0101] Figure 4 is a schematic diagram of data association configuration information according to an embodiment of the present application, as shown in Figure 4 the source data is defined first: incomplete configuration and configuration error.

[0102] Optionally, for the fault of the configuration error root cause, the data association method of the system can also be used for processing. For example, for the root cause data configuration error, the configuration protocol error, configuration parameter error, configuration version mismatch, and configuration atomic capability conversion error can be associated. The configuration parameter error can be associated with the service not being available, and the forwarding behavior of the SR Policy not meeting the expectation. The Policy behavior not meeting the expectation can be associated with the intuitively manifested fault information such as Policy bandwidth error, Policy latency error, and Policy master and backup path risk differentiation error.

[0103] Optionally, in the data association module, the association degree of all associated relationships is expressed based on probability. That is, the association degree (such as the association probability) between the source data (such as the root cause data) and the associated data (such as the fault data) is divided into four levels: certain, possible (a percentage can be considered to be set), small probability, and impossible. When initializing the association probability value, the previous operation and maintenance experience value can be referred to. The operation and maintenance experience in the system of the present application is derived from the expert AI library module.

[0104] Optionally, for example, the physical hardware interface failure will inevitably lead to the interface protocol down. For example, the failure of the member port can cause the traffic congestion of the aggregation port. For example, the error code of the interface can cause the detection protocol report down with a small probability. For example, reading the device temperature cannot cause the device power failure.

[0105] Optionally, the definition of the association degree (such as the association probability) is divided into positive definition and reverse definition.

[0106] Optionally, for the positive definition of the correlation degree (such as the correlation probability), the prior probability can be used, such as the source data (or root cause data) A, the correlation data (or fault data) B and the correlation data (or fault data) C, which can be directly defined according to the probabilities of the occurrence of the correlation data (or fault data) B and C under the condition of the occurrence of the source data (or root cause data), the correlation probability P(B|A) of A and B, and the correlation probability P(C|A) of A and C.

[0107] Optionally, for the initial value definition of the probability value of the source data (or root cause data) and the correlated data (or fault data) of the possibility relationship, examples are as follows: when 5 member ports are aggregated into an interface, the probability of causing congestion of the aggregated port traffic is 1 / 5=20% when each member port fails.

[0108] Optionally, for the reverse definition of the correlation degree, the posterior probability can be used, and the Bayes formula can be used: such as the source data space Ω and the correlation data B, then: Among them, for any event B, as long as P(B)>0, there is:

[0109] Optionally, the correlation forest data model (i.e. the preset correlation relationship model) of the system is constructed through the definition of the correlation relationship between the source data (or root cause data) and the correlation data (or fault data), and the positive and reverse probability-based correlation degree measurement of the correlation relationship.

[0110] Optionally, for the data of each node, through the expert AI library module, the related direct operation that needs to be performed when the data of this node is wrong can be defined.

[0111] Optionally, when the user needs to handle a fault, that is, in the fault analysis module, the fault needs to be handled for the fault analysis of the root cause, the fault information is taken as a correlation data (or fault data) A, and based on the probability of the correlation degree measurement value (i.e. the correlation probability) of the correlation data (or fault data), all possible related source data (or root cause data) nodes are analyzed. The analysis method is: according to the source data (or root cause data) value / state collected by the system, whether the corresponding source data (or root cause data) is the cause of the fault is sorted. If the state or value of the source data (or root cause data) is not clear, the state and value of the source data (or root cause data) can be obtained or confirmed according to the operation of the correlated data (or fault data) corresponding to the source data (or root cause data). If the source data (or root cause data) of the fault data is also in the fault state, or the value is out of limit, the corresponding source data (or root cause data) is taken as the next correlated data (or fault data) B, and the fault state and fault value of the source data (or root cause data) corresponding to the correlated data (or fault data) B are continuously analyzed. Until all possible root cause data are analyzed.

[0112] Optionally, after the root cause is analyzed, the root cause data of these error states is sorted according to the probability value of the correlation degree (i.e. correlation probability), and the operation strategy and suggestion information related to the sorted source data (or root cause data) node corresponding operation suggestion (i.e. fault resolution scheme) are analyzed. That is, the possible causes of the fault are given, and the fault resolution scheme is sorted according to the probability (i.e. correlation probability).

[0113] Optionally, the fault resolution scheme can use the user's habit of fault maintenance to list the related operations that can be handled, and in the fault handling module, the corresponding suggestions are executed step by step to execute the related operation scheme; if one scheme does not work, the next scheme can be executed.

[0114] Optionally, the source data (or root cause data) of fault (or fault data) A is B, C, D, and the source data (or root cause data) of C is E, F. When judging the state of B, C, and D, B is faulty, C is unknown, and D is good. Then, B and C are sorted by probability, and when the direction probability (i.e. correlation probability) of C to A is greater, the system executes the data collection operation of C, and obtains the state of C through the data collection module. When the state of C is wrong, the fault state E and F of C is determined. When E state is faulty and F state is normal, the fault resolution scheme can be given based on the operation suggestion of E and B, and the related processing operation of E fault is preferentially recommended according to the reverse probability degree value.

[0115] Optionally, in each operation process of fault handling, the fault judgment can also be based on the impossible relationship (probability value is 0) in the correlation degree to optimize the fault root cause analysis. For example, through the connectivity check interface, all source data corresponding to the associated data of this step of processing are determined, such as fault C, associated source data (or fault data) A and source data (or root cause data) B. When C is processed, the interface connectivity is checked, and it is found that the interface is up, and the data with normal interface connectivity is used as the source data (or root cause data). The correlation between the data A is 0, that is, the possibility of the correlation between the data A and the source data (or root cause data) with normal interface connectivity is 0. Then, the fault C only needs to be traced back to the source data (or root cause data) B for related analysis and processing, and the source data (or root cause data) A does not need to be analyzed and processed.

[0116] Optionally, for the expert AI library module, the construction and regular update of the expert experience library can be carried out by sorting network maintenance cases, maintaining problem ticket information, system characteristic logical analysis, and expert maintenance cases. Then, according to the expert experience library, the (forward) correlation probability value (such as correlation probability) of the associated data (or fault data) is regularly optimized and adjusted. At the same time, the (reverse) correlation probability value (such as correlation probability) of the associated data (such as fault data and root cause data) can be optimized (adjusted) by AI analysis of system data, such as calculating the correlation probability value (such as correlation probability) of the associated data (such as fault data and root cause data) by the frequency of simultaneous (same period) occurrence, and referring to the probability value calculated by AI.

[0117] Optionally, after each successful fault handling, the system root cause analysis is carried out in the result feedback module, the necessity of all operation steps and the efficiency of the operation process are sorted out, the correlation probability value (such as correlation probability) is increased for the necessary ones, the correlation probability value (such as correlation probability) is decreased for the unnecessary ones, and the redundant operation is optimized and integrated, so as to optimize the data definition and the correlation forest data model (such as the preset correlation relationship model) of data correlation, and the efficiency of fault handling operation.

[0118] Optionally, in the closed loop system as described above, based on the data model of data correlation (such as the preset correlation relationship model) and operation optimization, the system manually handles the fault according to the operation suggestion based on the initially defined correlation forest model (such as the preset correlation relationship model) at the beginning, optimizes the correlation forest model (such as the preset correlation relationship model) according to the handling result, and optimizes the operation scheme (such as the fault removal scheme), and evolves into that the system automatically executes part of the confirmed and low-risk operation according to the optimized correlation forest model (such as the preset correlation relationship model) and operation scheme, so as to realize part of the automatic detection, automatic judgment and automatic fault recovery. When this system is continuously iterated in maintenance, finally, the automatic fault positioning of the system can be realized, the automatic detection and recovery of fault handling are realized, and the self-healing evolution of IP SDN network operation and maintenance is realized.

[0119] Optionally, for the related fault alarm information, when the system impact analysis is carried out, such as network simulation, network planning, network service, network risk prediction, etc., the fault information is taken as a fault data, the root cause data corresponding to the fault data is analyzed based on the probability correlation value (such as correlation probability), and all the root cause data involved can be analyzed in turn, so as to obtain the possible influence of the fault, so as to support the analysis of the influence of the fault on the system, support the analysis of the optimization scheme of the system, and support the iterative evolution of the network system.

[0120] The technical scheme provided by the application can improve system operation and maintenance efficiency, reduce system operation and maintenance cost, improve system self-healing capability, and support system evolution.

[0121] The technical scheme provided by the application generates a correlation forest mathematical model (i.e., a preset correlation relationship model) of system data according to probability-based correlation (possibility) root cause analysis of fault data, and has the following advantages in system analysis and fault processing using the mathematical model compared with the prior art:

[0122] 1. The root cause of the problem is more clearly located through data correlation, and the correlation forest data model (i.e., the preset correlation relationship model) can establish correlation relationships based on necessity and then iteratively optimize them in a day-by-day and month-by-month manner as needed.

[0123] 2. The data correlation relationships are iteratively optimized as needed in the system operation life cycle through result feedback and expert analysis.

[0124] 3. The correlation relationship of the system is more clear and comprehensive through the correlation forest data model (i.e., the preset correlation relationship model), and the system operation and maintenance capability can be improved by applying the corresponding processing operation information in the corresponding model node.

[0125] 4. The correlation relationship of system data is comprehensively obtained through a multi-dimensional method, and the correlation forest data model (i.e., the preset correlation relationship model) is created, which can improve the automation level of data correlation and make the obtained correlation forest data model (i.e., the preset correlation relationship model) more comprehensive and real in reflecting the system.

[0126] 5. The correlation forest model (i.e., the preset correlation relationship model) based on probability can be used as a strategy and approach for system fault risk analysis.

[0127] According to an embodiment of the application, a fault processing device is also provided, and it should be noted that the fault processing device can be used to execute the fault processing method in the embodiment of the application, and the fault processing method in the embodiment of the application can be executed in the fault processing device.

[0128] Figure 5 is a schematic diagram of a fault processing device according to an embodiment of the application, as Figure 5As shown, the apparatus can comprise: an acquisition module 52 configured to acquire target fault data of a target network system; an analysis module 54 configured to analyze the target fault data by using a preset correlation relationship model to determine at least one target root cause data of the target fault data and a target correlation probability of each target root cause data, wherein the preset correlation relationship model is trained by using multiple sets of data through machine learning, and each set of data in the multiple sets of data comprises: preset fault data causing a fault, preset root cause data associated with the preset fault data, and a preset correlation probability used to represent a correlation relationship between the preset fault data and the preset root cause data; a first generation module 56 configured to arrange the at least one target root cause data according to the target correlation probability to generate a fault root cause table; and a second generation module 58 configured to query a fault removal scheme corresponding to each target root cause data in the fault root cause table in a preset fault handling library to generate a fault removal table.

[0129] It should be noted that the acquisition module 52 in this embodiment can be configured to perform step S102 in the embodiments of the present application, the analysis module 54 in this embodiment can be configured to perform step S104 in the embodiments of the present application, the first generation module 56 in this embodiment can be configured to perform step S106 in the embodiments of the present application, and the second generation module 58 in this embodiment can be configured to perform step S108 in the embodiments of the present application. The above modules have the same examples and application scenarios as the corresponding steps, but are not limited to the contents disclosed in the above embodiments.

[0130] In the embodiments of the present application, target fault data of a target network system is acquired; the target fault data is analyzed by using a preset correlation relationship model to determine at least one target root cause data of the target fault data and a target correlation probability of each target root cause data, wherein the preset correlation relationship model is trained by using multiple sets of data through machine learning, and each set of data in the multiple sets of data comprises: preset fault data causing a fault, preset root cause data associated with the preset fault data, and a preset correlation probability used to represent a correlation relationship between the preset fault data and the preset root cause data; the at least one target root cause data is arranged according to the target correlation probability to generate a fault root cause table; and a fault removal scheme corresponding to each target root cause data in the fault root cause table is queried in a preset fault handling library to generate a fault removal table, thereby achieving the technical effect of fault handling according to the correlation relationship between data, and further solving the technical problem that the prior art cannot perform fault handling according to the correlation relationship between data.

[0131] As an optional embodiment, the device further comprises: a first acquisition submodule, configured to acquire first root cause data calibrated in advance before analyzing the target fault data by using the preset correlation model to determine at least one target root cause data of the target fault data and a target correlation probability of each target root cause data; a second acquisition submodule, configured to acquire first result data generated by the target network system executing the first root cause data and a first correlation probability of generating the first result data; a third acquisition submodule, configured to acquire second result data generated by the target network system executing second root cause data by taking the first result data as the second root cause data and a second correlation probability of generating the second result data; a first determination submodule, configured to determine that the first root cause data is preset root cause data of preset fault data in a case where the second result data is the preset fault data; and a second determination submodule, configured to determine a preset correlation probability of the preset root cause data according to the first correlation probability and the second correlation probability.

[0132] As an optional embodiment, the device further comprises: a fourth acquisition submodule, configured to acquire a preset fault time of the preset fault data before analyzing the target fault data by using the preset correlation model to determine at least one target root cause data of the target fault data and a target correlation probability of each target root cause data; a fifth acquisition submodule, configured to acquire at least one upstream data generated by the target network system in a preset time period before the preset fault time, taking the preset fault time as a starting point; a third determination submodule, configured to determine preset root cause data of the preset fault data in the at least one upstream data; and a fourth determination submodule, configured to determine a preset correlation probability of the preset root cause data according to a frequency of the preset root cause data appearing in the at least one upstream data.

[0133] As an optional embodiment, the device further comprises: a sixth acquisition submodule, configured to acquire preset location information of a fault device indicated by the preset fault data before analyzing the target fault data by using the preset correlation model to determine at least one target root cause data of the target fault data and a target correlation probability of each target root cause data; a fifth determination submodule, configured to determine at least one preset region data in a preset region space where the preset location information is located, wherein the preset region data is used to represent running parameters corresponding to at least one preset region device in the preset region space; a sixth determination submodule, configured to determine preset root cause data of the preset fault data in the at least one preset region data; and a seventh determination submodule, configured to determine a preset correlation probability of the preset root cause data according to a distance value between a preset region device corresponding to the preset root cause data and the fault device.

[0134] As an optional embodiment, the device further comprises a seventh acquisition sub-module configured to acquire preset parameter data of the fault equipment indicated by the preset fault data before analyzing the target fault data by using the preset correlation model to determine at least one target root cause data of the target fault data and a target correlation probability of each target root cause data; an eighth determination sub-module configured to determine the preset parameter data as preset root cause data of the preset fault data; a ninth determination sub-module configured to determine a target parameter threshold matching the preset parameter data in a preset threshold library, wherein the preset threshold library is previously provided with a plurality of preset parameter thresholds and a preset correlation probability corresponding to each preset parameter threshold; and a tenth determination sub-module configured to determine the preset correlation probability corresponding to the target parameter threshold.

[0135] As an optional embodiment, the device further comprises a receiving sub-module configured to receive a fault removal report of the target fault data after querying a fault removal scheme corresponding to each target root cause data in the fault root cause table in the preset fault processing library and generating the fault removal table, wherein the target root cause data of the target fault data is carried in the fault removal report; and an adjusting sub-module configured to adjust the preset correlation model according to the target fault data and the associated target root cause data.

[0136] As an optional embodiment, the first generation module comprises a generation sub-module configured to arrange the at least one target root cause data in a sequence from high to low according to the target correlation probability and generate the fault root cause table.

[0137] Embodiments of the present application can provide a computer terminal, which can be any one of computer terminal devices in a computer terminal group. Alternatively, in the present embodiment, the computer terminal can be replaced by a mobile terminal or other terminal device.

[0138] Alternatively, in the present embodiment, the computer terminal can be located in at least one network device of a plurality of network devices of a computer network.

[0139] In the embodiment, the computer terminal can execute program codes of the following steps in the fault processing method: obtaining target fault data of a target network system; analyzing the target fault data through a preset correlation relationship model to determine at least one target root cause data of the target fault data and a target correlation probability of each target root cause data, wherein the preset correlation relationship model is trained by machine learning using a plurality of sets of data, each set of data in the plurality of sets of data includes: preset fault data causing a fault, preset root cause data associated with the preset fault data, and a preset correlation probability used to represent a correlation between the preset fault data and the preset root cause data; arranging the at least one target root cause data according to the target correlation probability to generate a fault root cause table; and querying a fault removal scheme corresponding to each target root cause data in the fault root cause table in a preset fault processing library to generate a fault removal table.

[0140] Optionally, Figure 6 is a structural block diagram of a computer terminal according to an embodiment of the present application. As shown in the figure, the computer terminal 60 can include one or more (only one is shown in the figure) processors 62 and a memory 64.

[0141] The memory can be used to store software programs and modules, such as program instructions / modules corresponding to the fault processing method and device in the embodiments of the present application. The processor executes various functions and data processing by running the software programs and modules stored in the memory, i.e., implements the above-mentioned fault processing method. The memory can include a high-speed random access memory, and can also include a non-volatile memory, such as one or more magnetic storage devices, flash memories, or other non-volatile solid-state memories. In some examples, the memory can further include a memory remotely arranged with respect to the processor, which can be connected to the terminal 60 through a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and a combination thereof.

[0142] The processor can call information and application programs stored in the memory through the transmission device to execute the following steps: obtaining target fault data of a target network system; analyzing the target fault data through a preset correlation relationship model to determine at least one target root cause data of the target fault data and a target correlation probability of each target root cause data, wherein the preset correlation relationship model is trained by machine learning using a plurality of sets of data, each set of data in the plurality of sets of data includes: preset fault data causing a fault, preset root cause data associated with the preset fault data, and a preset correlation probability used to represent a correlation between the preset fault data and the preset root cause data; arranging the at least one target root cause data according to the target correlation probability to generate a fault root cause table; and querying a fault removal scheme corresponding to each target root cause data in the fault root cause table in a preset fault processing library to generate a fault removal table.

[0143] Optionally, the processor can further execute program codes of the following steps: before analyzing the target fault data by the preset correlation relationship model to determine at least one target root cause data of the target fault data and a target correlation probability of each target root cause data, obtaining a first root cause data calibrated in advance; obtaining first result data generated by the target network system executing the first root cause data and a first correlation probability of generating the first result data; taking the first result data as a second root cause data, obtaining second result data generated by the target network system executing the second root cause data and a second correlation probability of generating the second result data; in the case that the second result data is the preset fault data, determining that the first root cause data is a preset root cause data of the preset fault data; and determining a preset correlation probability of the preset root cause data according to the first correlation probability and the second correlation probability.

[0144] Optionally, the processor can further execute program codes of the following steps: before analyzing the target fault data by the preset correlation relationship model to determine at least one target root cause data of the target fault data and a target correlation probability of each target root cause data, obtaining a preset fault time of the preset fault data; taking the preset fault time as a starting point, obtaining at least one upstream data generated by the target network system within a preset time period before the preset fault time; determining a preset root cause data of the preset fault data in the at least one upstream data; and determining a preset correlation probability of the preset root cause data according to a frequency number of the preset root cause data appearing in the at least one upstream data.

[0145] Optionally, the processor can further execute program codes of the following steps: before analyzing the target fault data by the preset correlation relationship model to determine at least one target root cause data of the target fault data and a target correlation probability of each target root cause data, obtaining preset location information of a fault device indicated by the preset fault data; determining at least one preset region data in a preset region space where the preset location information is located, wherein the preset region data is used to represent running parameters corresponding to at least one preset region device in the preset region space; determining a preset root cause data of the preset fault data in the at least one preset region data; and determining a preset correlation probability of the preset root cause data according to a distance value between a preset region device corresponding to the preset root cause data and the fault device.

[0146] Optionally, the processor can further execute program codes of the following steps: obtaining preset parameter data of the fault device indicated by the preset fault data before analyzing the target fault data by using the preset correlation model to determine at least one target root cause data of the target fault data and a target correlation probability of each target root cause data; taking the preset parameter data as preset root cause data of the preset fault data; determining a target parameter threshold matching the preset parameter data in a preset threshold library, wherein the preset threshold library is previously provided with a plurality of preset parameter thresholds and a preset correlation probability corresponding to each preset parameter threshold; and determining the preset correlation probability corresponding to the target parameter threshold.

[0147] Optionally, the processor can further execute program codes of the following steps: after querying the fault removal scheme corresponding to each target root cause data in the fault root cause table in the preset fault processing library and generating the fault removal table, receiving a fault removal report of the target fault data, wherein the target root cause data of the target fault data is carried in the fault removal report; and adjusting the preset correlation model according to the target fault data and the associated target root cause data.

[0148] Optionally, the processor can further execute program codes of the following steps: arranging the at least one target root cause data in a sequence from high to low according to the target correlation probability to generate the fault root cause table.

[0149] By using the embodiment of the present application, a fault processing scheme is provided. The target fault data of the target network system is obtained; the target fault data is analyzed by using the preset correlation model to determine at least one target root cause data of the target fault data and a target correlation probability of each target root cause data, wherein the preset correlation model is trained by using a plurality of sets of data through machine learning, each set of data in the plurality of sets of data includes preset fault data causing a fault, preset root cause data associated with the preset fault data, and a preset correlation probability used to represent the correlation between the preset fault data and the preset root cause data; the at least one target root cause data is arranged according to the target correlation probability to generate a fault root cause table; and the fault removal scheme corresponding to each target root cause data in the fault root cause table is queried in the preset fault processing library to generate a fault removal table, thereby achieving the technical effect of fault processing according to the correlation between data, and further solving the technical problem that the prior art cannot perform fault processing according to the correlation between data.

[0150] Those skilled in the art can understand that the structure shown in the figure is only schematic, and the computer terminal can also be a smart phone (such as an Android phone, an iOS phone, etc.), a tablet computer, a palm computer, a Mobile Internet Device (MID), a PAD, and the like. Figure 6It does not limit the structure of the electronic device. For example, the computer terminal 60 can further include more or less components (such as a network interface, a display device, etc.) or have a different configuration from that shown in FIG. 6. Figure 6 Figure 6

[0151] Those skilled in the art can understand that all or part of the steps in the above-mentioned various methods of the embodiments can be instructed by programs to the hardware related to the terminal device, and the programs can be stored in a computer readable storage medium, which can include a flash disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc.

[0152] The embodiments of the present application also provide a non-volatile storage medium. Optionally, in the present embodiment, the above-mentioned storage medium can be used to save the program code executed by the fault processing method provided by the above-mentioned embodiments.

[0153] Optionally, in the present embodiment, the above-mentioned non-volatile storage medium can be located in any one of the computer terminals in the computer terminal group in the computer network, or in any one of the mobile terminals in the mobile terminal group.

[0154] Optionally, in the present embodiment, the non-volatile storage medium is configured to store program code for performing the following steps: obtaining target fault data of a target network system; analyzing the target fault data by a preset association relationship model to determine at least one target root cause data of the target fault data and a target association probability of each target root cause data, wherein the preset association relationship model is trained by machine learning using a plurality of sets of data, each set of data in the plurality of sets of data includes preset fault data causing a fault, preset root cause data associated with the preset fault data, and a preset association probability for representing the association relationship between the preset fault data and the preset root cause data; arranging at least one target root cause data according to the target association probability to generate a fault root cause table; and querying a fault removal scheme corresponding to each target root cause data in the fault root cause table in a preset fault processing library to generate a fault removal table.

[0155] ​​Optionally, in the embodiment, the non-volatile storage medium is configured to store program code for performing the following steps: obtaining a first root cause data calibrated in advance before analyzing the target fault data by using the preset correlation model, determining at least one target root cause data of the target fault data and a target correlation probability of each target root cause data, and obtaining a first result data generated by the target network system executing the first root cause data and a first correlation probability of generating the first result data; obtaining a second result data generated by the target network system executing the first result data as a second root cause data and a second correlation probability of generating the second result data; determining the first root cause data as a preset root cause data of the preset fault data in a case that the second result data is the preset fault data; and determining a preset correlation probability of the preset root cause data according to the first correlation probability and the second correlation probability.

[0156] Optionally, in the embodiment, the non-volatile storage medium is configured to store program code for performing the following steps: obtaining a preset fault time of the preset fault data before analyzing the target fault data by using the preset correlation model, determining at least one upstream data generated by the target network system in a preset time period before the preset fault time from the preset fault time, determining a preset root cause data of the preset fault data in at least one of the upstream data, and determining a preset correlation probability of the preset root cause data according to a frequency number of the preset root cause data appearing in at least one of the upstream data.

[0157] Optionally, in the embodiment, the non-volatile storage medium is configured to store program code for performing the following steps: obtaining preset location information of a fault device indicated by the preset fault data before analyzing the target fault data by using the preset correlation model, determining at least one preset region data in a preset region space of the preset location information, wherein the preset region data is used to represent an operating parameter corresponding to at least one preset region device in the preset region space, determining a preset root cause data of the preset fault data in at least one of the preset region data, and determining a preset correlation probability of the preset root cause data according to a distance value between a preset region device corresponding to the preset root cause data and the fault device.

[0158] Optionally, in the embodiment, the nonvolatile storage medium is configured to store program code for performing the following steps: obtaining preset parameter data of a fault device indicated by preset fault data before analyzing the target fault data by the preset correlation relationship model, determining at least one target root cause data of the target fault data and a target correlation probability of each target root cause data; taking the preset parameter data as preset root cause data of the preset fault data; determining a target parameter threshold matching the preset parameter data in a preset threshold library, wherein the preset threshold library is previously provided with a plurality of preset parameter thresholds and a preset correlation probability corresponding to each preset parameter threshold; and determining the preset correlation probability corresponding to the target parameter threshold.

[0159] Optionally, in the embodiment, the nonvolatile storage medium is configured to store program code for performing the following steps: after querying a fault removal scheme corresponding to each target root cause data in the fault root cause table in a preset fault processing library and generating a fault removal table, receiving a fault removal report of the target fault data, wherein the target root cause data of the target fault data is carried in the fault removal report; and adjusting the preset correlation relationship model according to the target fault data and the associated target root cause data.

[0160] Optionally, in the embodiment, the nonvolatile storage medium is configured to store program code for performing the following steps: arranging the at least one target root cause data in a sequence from high to low according to the target correlation probability, and generating the fault root cause table.

[0161] The above-mentioned serial numbers of the embodiments of the application are only for description, and do not represent the advantages and disadvantages of the embodiments.

[0162] In the above-mentioned embodiments of the application, the description of each embodiment has its own focus, and the parts not described in detail in a certain embodiment can be referred to the related description of other embodiments.

[0163] In several embodiments provided in the present application, it should be understood that the disclosed technical contents can be implemented by other ways. Among them, the above-mentioned device embodiments are only schematic, for example, the division of the units can be a logical function division, and actual implementation can have another division way, for example, a plurality of units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the displayed or discussed each other can be through some interface, indirect coupling or communication connection between units or modules, which can be electrical or other forms.

[0164] The units described as separate components may or may not be physically separate, and the components displayed as units may or may not be physical units, that is, may be located in one place, or may be distributed to multiple units. Part or all of the units can be selected according to actual needs to achieve the purpose of the embodiment scheme.

[0165] In addition, each functional unit in each embodiment of the present application can be integrated in one processing unit, or each unit can exist physically, or two or more units can be integrated in one unit. The integrated unit can be realized in the form of hardware or in the form of a software functional unit.

[0166] The integrated unit, if realized in the form of a software functional unit and sold or used as an independent product, can be stored in a computer readable storage medium. Based on such understanding, the technical scheme of the present application or the part of the present application which contributes to the prior art or the whole or part of the technical scheme can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes a plurality of instructions for causing a computer device (which can be a personal computer, a server or a network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The foregoing storage medium includes a U disk, a read-only memory (ROM), a random access memory (RAM), a mobile hard disk, a magnetic disk or an optical disk, and various program code storage media.

[0167] The above is only the preferred embodiment of the present application, and it should be pointed out that for ordinary skilled in the art, without departing from the principles of the present application, a number of improvements and refinements can be made, and these improvements and refinements should be considered as the protection scope of the present application.

Claims

1. A failure handling method characterized by, The method comprises: obtaining target fault data of a target network system; analyzing the target fault data by a preset correlation relationship model to determine at least one target root cause data of the target fault data and a target correlation probability between the target fault data and each target root cause data, wherein the preset correlation relationship model is trained by machine learning using a plurality of sets of data, each set of data in the plurality of sets of data comprising: preset fault data causing a fault, preset root cause data associated with the preset fault data, and a preset correlation probability representing the correlation between the preset fault data and the preset root cause data; arranging at least one target root cause data according to the target correlation probability to generate a fault root cause table; querying a fault removal scheme corresponding to each target root cause data in the fault root cause table in a preset fault handling library to generate a fault removal table; Before analyzing the target fault data by the preset correlation relationship model to determine at least one target root cause data of the target fault data and a target correlation probability between the target fault data and each target root cause data, the method further comprises: obtaining a first root cause data labeled in advance; obtaining first result data generated by the target network system executing the first root cause data, and a first correlation probability of generating the first result data; taking the first result data as second root cause data, obtaining second result data generated by the target network system executing the second root cause data, and a second correlation probability of generating the second result data; in the case that the second result data is preset fault data, determining that the first root cause data is preset root cause data of the preset fault data; determining a preset correlation probability of the preset root cause data according to the first correlation probability and the second correlation probability; or, obtaining a preset fault time of the preset fault data; obtaining at least one upstream data generated by the target network system within a preset time period before the preset fault time; determining the preset root cause data of the preset fault data in at least one upstream data; determining the preset correlation probability of the preset root cause data by the frequency of the preset root cause data appearing in at least one upstream data; or, obtaining preset location information of a fault device indicated by the preset fault data; determining at least one preset area data in a preset area space where the preset location information is located, wherein the preset area data is used to represent the running parameters of at least one preset area device in the preset area space; determining the preset root cause data of the preset fault data in at least one preset area data; determining the preset correlation probability of the preset root cause data according to the distance value between the preset area device corresponding to the preset root cause data and the fault device.

2. The method of claim 1, wherein, After querying the fault removal scheme corresponding to each target root cause data in the fault root cause table in the preset fault handling library to generate the fault removal table, the method further comprises: receive a fault resolution report carrying the target fault data and target root cause data of the target fault data; adjust the preset correlation model according to the target fault data and the target root cause data associated with the target fault data.

3. The method according to any one of claims 1-2, characterized in that, arrange at least one target root cause data according to the target correlation probability to generate a fault root cause table, including: arranging at least one target root cause data in descending order of the target correlation probability to generate the fault root cause table.

4. A failure handling apparatus characterized by comprising: including: an acquisition module configured to acquire target fault data of a target network system; an analysis module configured to analyze the target fault data by using a preset correlation model to determine at least one target root cause data of the target fault data and a target correlation probability between the target fault data and each target root cause data, wherein the preset correlation model is trained by using multiple sets of data through machine learning, and each set of data in the multiple sets of data includes preset fault data causing a fault, preset root cause data associated with the preset fault data, and a preset correlation probability used to represent a correlation between the preset fault data and the preset root cause data; a first generation module configured to arrange at least one target root cause data according to the target correlation probability to generate a fault root cause table; a second generation module configured to query a fault resolution scheme corresponding to each target root cause data in the fault root cause table in a preset fault resolution library to generate a fault resolution table; before analyzing the target fault data by using the preset correlation model to determine at least one target root cause data of the target fault data and a target correlation probability between the target fault data and each target root cause data, the analysis module is further configured to: acquire first root cause data calibrated in advance; acquire first result data generated by the target network system executing the first root cause data and a first correlation probability of generating the first result data; acquire second result data generated by the target network system executing second root cause data taking the first result data as the second root cause data and a second correlation probability of generating the second result data; in a case where the second result data is preset fault data, determine that the first root cause data is preset root cause data of the preset fault data; and determine a preset correlation probability of the preset root cause data according to the first correlation probability and the second correlation probability; or acquire a preset fault time of preset fault data; acquire at least one upstream data generated by the target network system within a preset time period before the preset fault time, taking the preset fault time as a starting point; determine preset root cause data of the preset fault data in the at least one upstream data; and determine a preset correlation probability of the preset root cause data by counting a frequency of occurrence of the preset root cause data in the at least one upstream data; or Obtaining preset location information of a fault device indicated by preset fault data; determining at least one preset area data in a preset area space where the preset location information is located, wherein the preset area data is used to represent running parameters corresponding to at least one preset area device in the preset area space; determining preset root cause data of the preset fault data in at least one preset area data; and determining a preset correlation probability of the preset root cause data according to a distance value between a preset area device corresponding to the preset root cause data and the fault device.

5. A non-volatile storage medium, characterized by, The non-volatile storage medium stores a program, wherein the program controls a device in which the non-volatile storage medium is located to execute the fault processing method in any one of claims 1 to 3 when the program is running.

6. An electronic device, comprising: Comprise: A memory and a processor, the processor is used to run the program stored in the memory, wherein the program executes the fault processing method in any one of claims 1 to 3 when the program is running.

Images