Alarm correlation analysis method and device, storage medium and electronic equipment

Through the alarm association analysis method, the problem of low alarm processing efficiency in distributed cloud platforms is solved. Through association rule mining and root cause analysis, an alarm association relationship map is built to identify the root cause of the alarm, and rapid response and efficient operation and maintenance are achieved.

CN118916250APending Publication Date: 2024-11-08SHANDONG LANGCHAO YUNTOU INFORMATION TECH CO LTD
View PDF 0 Cites 2 Cited by

Patent Information

Application Number
CN202411159153.0
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-22
Publication Date
2024-11-08

AI Technical Summary

Technical Problem

In distributed cloud platforms, the existing alarm processing methods require independent analysis of a large amount of alarm information, resulting in long response time for operation and maintenance personnel, difficulty in quickly locate abnormal problems, and low operation and maintenance efficiency.

Method used

The alarm correlation analysis method is adopted to construct an alarm correlation map and directed acyclic graph through data standardization, filtering, correlation mining and root cause analysis, to identify the correlation relationship and root causes between alarms, and optimize the alarm processing process.

Benefits of technology

It improves the alarm processing speed, realizes the rapid positioning of abnormal problems, improves operation and maintenance efficiency, and ensures the stable operation of the business system.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118916250A_ABST
    Figure CN118916250A_ABST
Patent Text Reader

Abstract

The invention provides an alarm association analysis method and device, a storage medium and electronic equipment, and belongs to the technical field of security, and the method comprises the steps: carrying out the data standardization and alarm filtering processing of an alarm data set, and obtaining a to-be-analyzed data set; based on an association rule mining algorithm, performing association mining processing on the to-be-analyzed data set to obtain an alarm association analysis result; determining an associated data set according to an alarm association analysis result; establishing an alarm association relationship graph corresponding to the associated data set, and constructing a corresponding alarm directed acyclic graph; based on the alarm directed acyclic graph, carrying out alarm root cause analysis on the associated data set to obtain a root cause analysis result; and taking the alarm association relationship graph, the alarm directed acyclic graph and the root cause analysis result as an association analysis result of the association data set. By applying the method provided by the invention, a large amount of alarm information can be subjected to association analysis, so that operation and maintenance personnel can quickly realize problem positioning and carry out targeted operation and maintenance operation, and the operation and maintenance efficiency can be improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of security technology, and in particular to an alarm correlation analysis method and device, a storage medium and an electronic device. Background Art

[0002] In the IT infrastructure operation and maintenance scenario, alarm processing is one of the common operation and maintenance measures. During the operation of the business system, when the system is abnormal, the system's monitoring unit will send an alarm signal according to the abnormal situation to notify the operation and maintenance personnel to perform operation and maintenance according to the alarm.

[0003] At present, when an alarm signal occurs, the alarm information of each alarm is usually analyzed separately, and based on the analysis result of each alarm, the relevant data of each alarm is sent to the corresponding operation and maintenance personnel, so that the operation and maintenance personnel can respond to each alarm separately.

[0004] With the development of cloud computing technology, various business systems are gradually applying distributed cloud platform architecture. Distributed cloud platforms involve a large number of cloud service nodes. In the actual operation process, each cloud service node usually generates a large number of alarm signals. Based on the existing alarm processing method, it is necessary to analyze large-scale, distributed alarm data sets, perform independent analysis for each alarm, and the operation and maintenance personnel perform operation and maintenance processing for each alarm they receive, which makes the overall alarm processing time-consuming, difficult to quickly locate abnormal problems, and the operation and maintenance efficiency is low, which is not conducive to ensuring the stable operation of the system. Summary of the invention

[0005] In view of this, an embodiment of the present invention provides an alarm correlation analysis method to solve the problem that each alarm is currently analyzed independently and operation and maintenance personnel are required to perform operation and maintenance processing for each alarm separately, which makes alarm processing time-consuming, difficult to quickly locate anomalies, and has low operation and maintenance efficiency.

[0006] The embodiment of the present invention also provides an alarm correlation analysis device to ensure the actual implementation and application of the above method.

[0007] To achieve the above objectives, the embodiments of the present invention provide the following technical solutions:

[0008] An alarm correlation analysis method, comprising:

[0009] Determine an alarm data set to be processed; the alarm data set includes a plurality of alarm information;

[0010] Performing data standardization processing on the alarm data set to obtain a standardized data set; the standardized data set includes each alarm information that has undergone standardization processing;

[0011] Performing alarm filtering processing on the standardized data set, and forming a data set to be analyzed with each alarm information obtained after the processing;

[0012] Based on a preset association rule mining algorithm, an association mining process is performed on the data set to be analyzed to obtain an alarm association analysis result corresponding to the data set to be analyzed;

[0013] Determine a correlation data set according to the alarm correlation analysis result; the correlation data set includes each alarm information having a correlation relationship in the data set to be analyzed;

[0014] Establishing an alarm association relationship graph corresponding to the associated data set, and constructing an alarm directed acyclic graph corresponding to the alarm association relationship graph;

[0015] Based on the alarm directed acyclic graph, performing alarm root cause analysis on the associated data set to obtain a root cause analysis result corresponding to the associated data set;

[0016] The alarm association relationship graph, the alarm directed acyclic graph and the root cause analysis result are used as the association analysis result corresponding to the association data set.

[0017] In the above method, optionally, the step of performing data standardization processing on the alarm data set to obtain a standardized data set includes:

[0018] Performing data cleaning processing on each alarm information in the alarm data set to obtain each alarm information after data cleaning;

[0019] Performing data formatting processing on each of the alarm information that has undergone data cleaning to obtain a data formatting result corresponding to each alarm information in the alarm data set;

[0020] Each of the data formatting results is used as standardized alarm information, and the standardized data set is composed of the various data formatting results.

[0021] In the above method, optionally, performing alarm filtering processing on the standardized data set and forming a data set to be analyzed with each alarm information obtained after the processing includes:

[0022] Preliminarily filtering each alarm information in the standardized data set based on a preset alarm management module to obtain a preliminary screening data set; the preliminary screening data set includes each alarm information that has been preliminarily filtered;

[0023] Based on a preset complex event processing mechanism, the temporal and spatial correlation of the preliminary screening data set is identified to obtain the temporal and spatial correlation relationship of each alarm information in the preliminary screening data set;

[0024] The various alarm information in the primary screening data set is secondary filtered according to the spatiotemporal correlation, and the various alarm information that have undergone secondary filtering are combined into the data set to be analyzed.

[0025] In the above method, optionally, performing association mining processing on the data set to be analyzed based on a preset association rule mining algorithm to obtain an alarm association analysis result corresponding to the data set to be analyzed includes:

[0026] Determining the alarm data characteristic information corresponding to the data set to be analyzed;

[0027] Determine a target algorithm that matches the characteristic information of the alarm data among a plurality of pre-set distributed association rule mining algorithms;

[0028] The target algorithm is used to perform data mining on the data set to be analyzed, and the data mining result is used as the alarm association analysis result.

[0029] The above method may optionally include performing alarm root cause analysis on the associated data set based on the alarm directed acyclic graph to obtain a root cause analysis result corresponding to the associated data set, including:

[0030] By using a preset path identification strategy, the root node of the alarm directed acyclic graph is identified to obtain the root node in the alarm directed acyclic graph;

[0031] By using a pre-built alarm root source identification model, the root source alarm identification is performed on the associated data set to obtain a score corresponding to each alarm information in the associated data set, wherein the score represents the possibility that the corresponding alarm belongs to the alarm root source;

[0032] The root cause analysis result corresponding to the associated data set is determined according to the root source node in the alarm directed acyclic graph and the score corresponding to each of the alarm information.

[0033] The above method may optionally further include:

[0034] Performing alarm denoising processing on the data set to be analyzed by using a pre-built alarm denoising model, and using the alarm information obtained after the processing as the alarm information to be notified;

[0035] For each of the alarm information to be notified, determining evaluation information corresponding to the alarm information to be notified, the evaluation information including an alarm level and a business impact range corresponding to the alarm information to be notified;

[0036] Determine the notification method and notification personnel corresponding to each of the alarm information to be notified based on the preset notification strategy and the evaluation information corresponding to each of the alarm information to be notified;

[0037] For each of the alarm information to be notified, based on the notification method corresponding to the alarm information to be notified, the alarm notification corresponding to the alarm information to be notified is sent to a corresponding notification person.

[0038] The above method may optionally further include:

[0039] Determining target alarm information in the data set to be analyzed; the target alarm information is alarm information of a known solution;

[0040] Determine, among various preset alarm solutions, a target alarm solution corresponding to the target alarm information;

[0041] A pre-set automated operation and maintenance tool is called to trigger a fault recovery operation corresponding to the target alarm solution to perform an alarm processing response to the target alarm information.

[0042] An alarm correlation analysis device, comprising:

[0043] A first determining unit is used to determine an alarm data set to be processed; the alarm data set includes a plurality of alarm information;

[0044] A standardization unit, used for performing data standardization processing on the alarm data set to obtain a standardized data set; the standardized data set includes each alarm information that has been standardized;

[0045] A filtering unit, used to perform alarm filtering processing on the standardized data set, and form a data set to be analyzed by filtering each alarm information obtained after the processing;

[0046] An association mining unit is used to perform association mining processing on the data set to be analyzed based on a preset association rule mining algorithm to obtain an alarm association analysis result corresponding to the data set to be analyzed;

[0047] A second determining unit is used to determine a correlation data set according to the alarm correlation analysis result; the correlation data set includes each alarm information having a correlation relationship in the data set to be analyzed;

[0048] A graph construction unit, used to establish an alarm association relationship graph corresponding to the associated data set, and to construct an alarm directed acyclic graph corresponding to the alarm association relationship graph;

[0049] A root cause analysis unit, configured to perform an alarm root cause analysis on the associated data set based on the alarm directed acyclic graph, and obtain a root cause analysis result corresponding to the associated data set;

[0050] The third determining unit is used to use the alarm association relationship map, the alarm directed acyclic graph and the root cause analysis result as the association analysis result corresponding to the association data set.

[0051] A storage medium includes stored instructions, wherein when the instructions are executed, the device where the storage medium is located is controlled to execute the alarm correlation analysis method as described above.

[0052] An electronic device comprises a memory and one or more instructions, wherein the one or more instructions are stored in the memory and are configured to be executed by one or more processors as the above-mentioned alarm correlation analysis method.

[0053] An alarm association analysis method provided based on the above-mentioned embodiment of the present invention includes: determining an alarm data set to be processed; performing data standardization processing on the alarm data set to obtain a standardized data set; performing alarm filtering processing on the standardized data set, and forming the various alarm information obtained after the processing into a data set to be analyzed; based on a preset association rule mining algorithm, performing association mining processing on the data set to be analyzed, and obtaining an alarm association analysis result corresponding to the data set to be analyzed; determining an associated data set according to the alarm association analysis result; the associated data set includes various alarm information with an associated relationship in the data set to be analyzed; establishing an alarm association relationship map corresponding to the associated data set, and constructing an alarm directed acyclic graph corresponding to the alarm association relationship map; based on the alarm directed acyclic graph, performing alarm root cause analysis on the associated data set to obtain a root cause analysis result corresponding to the associated data set; and using the alarm association relationship map, the alarm directed acyclic graph and the root cause analysis result as the association analysis result corresponding to the associated data set. By applying the method provided by the embodiment of the present invention, the association mining can be performed on each alarm information to be analyzed through the association rule mining algorithm, and the association analysis results between the alarms can be obtained. The associated data set is determined according to the association analysis results, and the alarm association relationship map and the alarm directed acyclic graph corresponding to the associated data set are established. The root cause analysis of the alarm is performed on the associated data set, and then the association analysis results are obtained. When the operation and maintenance personnel need to process the alarm information in the associated data set, the association analysis results can be fed back to the operation and maintenance personnel, so that the operation and maintenance personnel can perform targeted operations based on the association analysis between the alarms, and quickly complete the operation and maintenance of the alarm from the root, which is conducive to improving the alarm processing speed, realizing the rapid location of abnormal problems, and then improving the operation and maintenance efficiency, and ensuring the operational stability of the business system. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.

[0055] Figure 1 A method flow chart of an alarm correlation analysis method provided by an embodiment of the present invention;

[0056] Figure 2 A schematic diagram of the architecture of an alarm correlation analysis system provided by an embodiment of the present invention;

[0057] Figure 3 A schematic diagram of an alarm correlation analysis process provided by an embodiment of the present invention;

[0058] Figure 4 A schematic diagram of a display interface for alarm correlation analysis results provided by an embodiment of the present invention;

[0059] Figure 5 A schematic diagram of another display interface for alarm correlation analysis results provided by an embodiment of the present invention;

[0060] Figure 6 A schematic diagram of the structure of an alarm correlation analysis device provided by an embodiment of the present invention;

[0061] Figure 7 A schematic diagram of the structure of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0062] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0063] In this application, the terms "comprises", "comprising" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element defined by the sentence "comprising a ..." does not exclude the presence of other identical elements in the process, method, article or device comprising the element.

[0064] The embodiment of the present invention provides an alarm correlation analysis method, which can be applied to an alarm analysis system of a distributed cloud platform, and the execution subject can be a server of the alarm analysis system. The method flow chart of the method is as follows: Figure 1 As shown, including:

[0065] S101: Determine an alarm data set to be processed; the alarm data set includes a plurality of alarm information;

[0066] In the method provided by the embodiment of the present invention, the alarm events of each cloud service node on the distributed cloud platform are collected, and the alarm information of each alarm event collected within a certain period of time is analyzed and processed in real time.

[0067] When the alarm information processing process is triggered, each alarm information that needs to be processed is used as the alarm data set to be processed. It can be understood that an alarm information represents an alarm event on the platform, which includes relevant information of the alarm event, specifically including the source node of the alarm event, the associated system log, business data and other data.

[0068] S102: performing data standardization processing on the alarm data set to obtain a standardized data set; the standardized data set includes each alarm information that has been standardized;

[0069] In the method provided in the embodiment of the present invention, for the alarm data set, that is, the original alarm information, according to the pre-set data standardization processing strategy, the alarm information in the alarm data set is subjected to data standardization processing, and the alarm information after the standardization processing is used as the standardized data set. The data standardization processing may specifically include operations such as data cleaning and data format normalization processing.

[0070] S103: performing alarm filtering processing on the standardized data set, and forming a data set to be analyzed with each alarm information obtained after the processing;

[0071] In the method provided in an embodiment of the present invention, based on a pre-set alarm filtering method, each alarm information in the standardized data is screened, and the alarm information that does not need to be processed, such as repeated alarms, is filtered out, and the various alarm information obtained after filtering are combined into a data set to be analyzed, that is, the data set to be analyzed includes the various alarm information remaining in the standardized data set after the filtering operation.

[0072] S104: performing association mining processing on the data set to be analyzed based on a preset association rule mining algorithm to obtain an alarm association analysis result corresponding to the data set to be analyzed;

[0073] In the method provided in the embodiment of the present invention, an association rule mining algorithm is pre-set, such as an existing association rule mining algorithm such as an FP-growth algorithm and an Apriori algorithm. Based on the association rule mining algorithm, association mining is performed on each alarm information in the data set to be analyzed to identify frequent patterns and correlations in each alarm information, and obtain an alarm correlation analysis result, that is, an analysis result containing the correlation relationship between each alarm information.

[0074] S105: determining a correlation data set according to the alarm correlation analysis result; the correlation data set includes each alarm information having a correlation relationship in the data set to be analyzed;

[0075] In the method provided by the embodiment of the present invention, based on the correlation relationship between each alarm contained in the alarm correlation analysis result, each alarm information in the data set to be analyzed is classified according to the correlation, and each alarm information with a correlation relationship is combined into a correlation data set.

[0076] It should be noted that, in the specific implementation process, the data set to be analyzed can be divided into multiple related data sets, that is, there are multiple sets of respectively related alarm information. When there are multiple related data sets, each related data set can be subsequently processed separately.

[0077] S106: Establish an alarm association relationship graph corresponding to the associated data set, and construct an alarm directed acyclic graph corresponding to the alarm association relationship graph;

[0078] In the method provided by the embodiment of the present invention, for each alarm information in the associated data set, based on the association relationship between each alarm information, a pre-set graph construction method is used to construct a knowledge graph characterizing the association relationship between each alarm information, and the constructed graph is used as the alarm association relationship graph corresponding to the associated data set. Each graph node in the alarm association relationship graph represents the alarm corresponding to each alarm information, and the connecting edges between each graph node represent the association relationship between each alarm.

[0079] Based on each graph node in the alarm association relationship graph and the connection relationship between each graph node, a corresponding directed acyclic graph is constructed, and the directed acyclic graph is used as the alarm directed acyclic graph corresponding to the associated data set.

[0080] S107: Based on the alarm directed acyclic graph, perform alarm root cause analysis on the associated data set to obtain a root cause analysis result corresponding to the associated data set;

[0081] In the method provided by the embodiment of the present invention, based on the alarm directed acyclic graph corresponding to the associated data set, a pre-set root cause analysis mechanism is applied to perform alarm root cause analysis on the associated data set, that is, analyze which alarm information in the associated data set may be the alarm generated by the most root abnormality, and obtain the root cause analysis result of the associated data set. The root cause analysis result includes indication information for the root alarm information, such as the root alarm identifier, or the probability that each alarm information belongs to the root alarm.

[0082] S108: Using the alarm association relationship graph, the alarm directed acyclic graph, and the root cause analysis result as the association analysis result corresponding to the association data set.

[0083] In the method provided by the embodiment of the present invention, the alarm association relationship graph, the alarm directed acyclic graph and the root cause analysis result corresponding to the associated data set are used as the corresponding association analysis result, and the result can be stored in the database. The association analysis result can be sent to the corresponding operation and maintenance personnel along with the corresponding alarm notification, and can also be fed back to the operation and maintenance personnel when they browse the processing work order of the corresponding alarm information.

[0084] Based on the method provided by the embodiment of the present invention, in the alarm analysis process, an alarm data set to be processed is determined; data standardization processing is performed on the alarm data set to obtain a standardized data set; alarm filtering processing is performed on the standardized data set, and each alarm information obtained after processing is used to form a data set to be analyzed; based on a preset association rule mining algorithm, association mining processing is performed on the data set to be analyzed to obtain an alarm association analysis result corresponding to the data set to be analyzed; according to the alarm association analysis result, an associated data set is determined; the associated data set includes each alarm information with an associated relationship in the data set to be analyzed; an alarm association relationship map corresponding to the associated data set is established, and an alarm directed acyclic graph corresponding to the alarm association relationship map is constructed; based on the alarm directed acyclic graph, an alarm root cause analysis is performed on the associated data set to obtain a root cause analysis result corresponding to the associated data set; the alarm association relationship map, the alarm directed acyclic graph and the root cause analysis result are used as the association analysis result corresponding to the associated data set. By applying the method provided by the embodiment of the present invention, the association mining can be performed on each alarm information to be analyzed through the association rule mining algorithm, and the association analysis results between the alarms can be obtained. The associated data set is determined according to the association analysis results, and the alarm association relationship map and the alarm directed acyclic graph corresponding to the associated data set are established. The root cause analysis of the alarm is performed on the associated data set, and then the association analysis results are obtained. When the operation and maintenance personnel need to process the alarm information in the associated data set, the association analysis results can be fed back to the operation and maintenance personnel, so that the operation and maintenance personnel can perform targeted operations based on the association analysis between the alarms, and quickly complete the operation and maintenance of the alarm from the root, which is conducive to improving the alarm processing speed, realizing the rapid location of abnormal problems, and then improving the operation and maintenance efficiency, and ensuring the operational stability of the business system.

[0085] exist Figure 1 On the basis of the method shown, in the method provided by the embodiment of the present invention, the process of performing data standardization processing on the alarm data set to obtain a standardized data set mentioned in step S102 includes:

[0086] Performing data cleaning processing on each alarm information in the alarm data set to obtain each alarm information after data cleaning;

[0087] In the method provided by an embodiment of the present invention, during the data standardization processing of the alarm data set, data cleaning is first performed on the data in each alarm information in the alarm data set. Specifically, an outlier detection algorithm, such as IQR (interquartile range), etc., can be introduced to perform outlier detection and data correction operations on each alarm information.

[0088] Performing data formatting processing on each of the alarm information that has undergone data cleaning to obtain a data formatting result corresponding to each alarm information in the alarm data set;

[0089] Each of the data formatting results is used as standardized alarm information, and the standardized data set is composed of the various data formatting results.

[0090] In the method provided by the embodiment of the present invention, data formatting is performed on each alarm information that has been cleaned according to a predetermined formatting mechanism, and the expressions of fields such as the alarm level, type, and description in each alarm information are counted, and the formatted alarm information is used as the data formatting result. The data formatting result of each alarm information is regarded as the alarm information that has been standardized, thereby obtaining a standardized data set.

[0091] exist Figure 1 On the basis of the method shown, in the method provided by the embodiment of the present invention, the process of performing alarm filtering processing on the standardized data set mentioned in step S103 and forming the data set to be analyzed from the various alarm information obtained after the processing includes:

[0092] Preliminarily filtering each alarm information in the standardized data set based on a preset alarm management module to obtain a preliminary screening data set; the preliminary screening data set includes each alarm information that has been preliminarily filtered;

[0093] In the method provided in an embodiment of the present invention, a pre-deployed alarm management module can be called, such as an existing alarm management module such as Alertmanager, and the alarm filtering rules in the alarm management module are applied to perform preliminary filtering and denoising on each alarm information in the standardized data set, and the initial screening data set is composed of the various alarm information obtained after preliminary filtering.

[0094] Based on a preset complex event processing mechanism, the temporal and spatial correlation of the preliminary screening data set is identified to obtain the temporal and spatial correlation relationship of each alarm information in the preliminary screening data set;

[0095] In the method provided in an embodiment of the present invention, the spatiotemporal correlation between each alarm information in the initial screening data set is identified based on the complex event processing (CEP) mechanism, thereby obtaining the spatiotemporal correlation between each alarm information, such as similar alarms that appear in a short period of time, or alarms that appear in a specific order.

[0096] The various alarm information in the primary screening data set is secondary filtered according to the spatiotemporal correlation, and the various alarm information that have undergone secondary filtering are combined into the data set to be analyzed.

[0097] In the method provided by the embodiment of the present invention, the temporal and spatial correlation between each piece of alarm information is utilized to perform secondary filtering on each piece of alarm information, and each piece of alarm information remaining after filtering is formed into a data set to be analyzed.

[0098] exist Figure 1 On the basis of the method shown, in the method provided by the embodiment of the present invention, the process of performing association mining processing on the data set to be analyzed based on a preset association rule mining algorithm mentioned in step S104 to obtain the alarm association analysis result corresponding to the data set to be analyzed includes:

[0099] Determining the alarm data characteristic information corresponding to the data set to be analyzed;

[0100] In the method provided in the embodiment of the present invention, the alarm data characteristic information characterizing the data characteristics of the data set to be analyzed can be determined based on each alarm information in the data set to be analyzed. For example, characteristic data such as the data volume and number of items of the alarm information in the data set to be analyzed can be counted, and each characteristic data can be used as the alarm data characteristic information.

[0101] Determine a target algorithm that matches the characteristic information of the alarm data among a plurality of pre-set distributed association rule mining algorithms;

[0102] In the method provided in the embodiment of the present invention, a plurality of distributed association rule mining algorithms are preset, such as Apriori, FP-growth and other association rule mining algorithms. Each preset association rule mining algorithm corresponds to processing an alarm data set with corresponding data characteristics, for example, a large amount of data and a large number of items, and the FP-growth algorithm is used accordingly.

[0103] In the method provided by the embodiment of the present invention, the alarm data feature information corresponding to the data set to be analyzed can be matched with the processing object features matched by each association rule mining algorithm, and the association rule mining algorithm that matches the alarm data features of the data set to be analyzed can be found and used as the target algorithm.

[0104] The target algorithm is used to perform data mining on the data set to be analyzed, and the data mining result is used as the alarm association analysis result.

[0105] In the method provided by the embodiment of the present invention, a target algorithm is applied to perform data mining on the data set to be analyzed to identify frequent patterns and correlations between each alarm information in the data set to be analyzed, and the data mining result is used as the alarm correlation analysis result corresponding to the data set to be analyzed.

[0106] exist Figure 1 On the basis of the method shown, in the method provided by the embodiment of the present invention, the process of performing alarm root cause analysis on the associated data set based on the alarm directed acyclic graph mentioned in step S107 to obtain the root cause analysis result corresponding to the associated data set includes:

[0107] By using a preset path identification strategy, the root node of the alarm directed acyclic graph is identified to obtain the root node in the alarm directed acyclic graph;

[0108] In the method provided by the embodiment of the present invention, a path identification strategy is pre-set, such as a graph path identification algorithm such as the shortest path or the maximum weight path. In the association analysis process, the root node in the alarm directed acyclic graph, that is, the node with the most likely fault source, is identified through the pre-set path identification strategy.

[0109] By using a pre-built alarm root source identification model, the root source alarm identification is performed on the associated data set to obtain a score corresponding to each alarm information in the associated data set, wherein the score represents the possibility that the corresponding alarm belongs to the alarm root source;

[0110] In the method provided in the embodiment of the present invention, an alarm root source identification model is pre-constructed based on a machine learning algorithm, for example, based on a machine learning algorithm such as a random forest or a gradient boosting algorithm, and the alarm root source identification model is constructed by means of ensemble learning. Specifically, historical alarm data can be used as training samples to perform model training to obtain the alarm root source identification model. The input data of the alarm root source identification model is the alarm information of each alarm to be identified, and its output data is a score for each alarm. The score output by the alarm root source identification model represents the predicted probability that the corresponding alarm belongs to the alarm root source. The higher the score, the greater the probability of belonging to the alarm root source.

[0111] In the method provided by the embodiment of the present invention, during the root cause analysis process, each alarm information in the associated data set can be used as input data and input into the alarm root source identification model, and the score for each alarm information is obtained from the output layer of the alarm root source identification model.

[0112] The root cause analysis result corresponding to the associated data set is determined according to the root source node in the alarm directed acyclic graph and the score corresponding to each of the alarm information.

[0113] In the method provided by the embodiment of the present invention, the root source node determined in the alarm directed acyclic graph using path identification and the score corresponding to each alarm information predicted by the alarm root source identification model are combined to comprehensively evaluate the root source alarms in each alarm information to obtain the root cause analysis result. For example, the identification result of the root source node can be converted into a score, and the identification result of the root source node and the score output by the alarm root source identification model can be weighted by using a weighted method to obtain the final score, and the root cause analysis result is determined based on the final score.

[0114] exist Figure 1 On the basis of the method shown, the method provided in the embodiment of the present invention further includes:

[0115] Performing alarm denoising processing on the data set to be analyzed by using a pre-built alarm denoising model, and using the alarm information obtained after the processing as the alarm information to be notified;

[0116] In the method provided in the embodiment of the present invention, an alarm denoising model is pre-constructed based on a machine learning algorithm to suppress alarm information, for example, suppressing alarm information that can be solved automatically and non-critical alarm information, so as to screen out these alarms during the alarm notification process to reduce the burden of operation and maintenance. Specifically, historical alarm data and its corresponding solutions (such as notifying manual or automatic responses) can be used as training samples to perform model training to obtain an alarm denoising model. The input data of the alarm denoising model can be each alarm information to be denoised, and the output result can be an alarm list, that is, the identifiers of each alarm obtained after alarm suppression of each alarm information.

[0117] In the method provided by the embodiment of the present invention, each alarm information in the data set to be analyzed can be used as input data and input into the alarm denoising model. After the alarms of each alarm information are suppressed by the alarm denoising model, an alarm list output by the alarm denoising model is obtained, that is, a list of alarms that need to be notified, that is, the alarm information obtained after the alarm denoising processing, and these alarm information are used as the alarm information to be notified.

[0118] For each of the alarm information to be notified, determining evaluation information corresponding to the alarm information to be notified, the evaluation information including an alarm level and a business impact range corresponding to the alarm information to be notified;

[0119] In the method provided by the embodiment of the present invention, the corresponding alarm level and service impact range and other information can be determined according to the data in the alarm information to be notified, and the above information can be used as the evaluation information corresponding to the alarm information to be notified.

[0120] Determine the notification method and notification personnel corresponding to each of the alarm information to be notified based on the preset notification strategy and the evaluation information corresponding to each of the alarm information to be notified;

[0121] In the method provided in the embodiment of the present invention, the notification strategy of the alarm is pre-set, which specifically includes the alarm notification method of the alarm information of various situations and the range of personnel to be notified. For the alarm information to be notified, the notification method and notification personnel matching it can be found through its corresponding evaluation information.

[0122] For each of the alarm information to be notified, based on the notification method corresponding to the alarm information to be notified, the alarm notification corresponding to the alarm information to be notified is sent to a corresponding notification person.

[0123] In the method provided in the embodiment of the present invention, based on the notification method corresponding to the alarm information to be notified, an alarm notification is sent to the corresponding notification personnel to remind the relevant personnel of the alarm event. The notification method can be through channels such as email, text message or instant messaging tools. Specifically, the push of alarm notification can be realized by integrating the alarm notification system.

[0124] Based on the method provided in the embodiment of the present invention, the alarm can be denoised through the alarm noise reduction model, and non-critical alarms or alarm notifications of known solutions can be suppressed to reduce the information receiving burden of the notification recipient and avoid information overload.

[0125] exist Figure 1 On the basis of the method shown, the method provided in the embodiment of the present invention further includes:

[0126] Determining target alarm information in the data set to be analyzed; the target alarm information is alarm information of a known solution;

[0127] In the method provided by the embodiment of the present invention, various alarm solutions can be preset to determine the alarm information of known solutions in the data set to be analyzed, and the alarm information of the known solutions can be used as the target alarm information.

[0128] It should be noted that in the specific implementation process, there may be no alarm information of a known solution in the data set to be analyzed, or there may be alarm information of multiple known solutions. If there is no alarm information of a known solution, there is no need to respond to the subsequent automatic alarm processing. If there is alarm information of multiple known solutions, the alarm information of each known solution can be used as the target information and processed separately.

[0129] Determine, among various preset alarm solutions, a target alarm solution corresponding to the target alarm information;

[0130] In the method provided by the embodiment of the present invention, the alarm solution matching the target alarm information can be found among the pre-set alarm solutions by matching information, that is, the alarm solution used to process the alarm event of the type corresponding to the target alarm information, and the alarm solution can be used as the target alarm solution.

[0131] A pre-set automated operation and maintenance tool is called to trigger a fault recovery operation corresponding to the target alarm solution to perform an alarm processing response to the target alarm information.

[0132] In the method provided in the embodiment of the present invention, an automated operation and maintenance tool, such as Ansible, is pre-integrated. The automated operation and maintenance tool is called to trigger an automated script or workflow to execute a fault recovery operation corresponding to the target alarm solution to process the alarm event corresponding to the target alarm information.

[0133] Based on the method provided in the embodiment of the present invention, automatic operation and maintenance can be implemented for alarm information of known solutions, which can reduce the burden of manual alarm processing and improve the efficiency of alarm operation and maintenance.

[0134] In order to better illustrate the method provided by the embodiment of the present invention, based on the methods provided by the above embodiments, the embodiment of the present invention provides another alarm correlation analysis method. The method provided by the embodiment of the present invention can be deployed based on the alarm correlation analysis system. The method is applied to realize the alarm analysis scenario of the distributed cloud platform. The architecture of the alarm correlation analysis system provided by the embodiment of the present invention can be as follows: Figure 2 As shown, it specifically includes: data acquisition module, data processing module, association rule mining module, intelligent analysis module, integrated machine learning model, graph database module, notification and response module. The processing process of each functional module is mainly as follows:

[0135] Data acquisition module;

[0136] Use Kafka message queues to collect real-time alarm information from different cloud services, applications, and infrastructure to ensure low-latency data transmission. Add a message confirmation mechanism to the Kafka message queue configuration to ensure that the message is delivered at least once, and implement deduplication logic at the data receiving end to avoid duplicate data processing.

[0137] Data processing module;

[0138] On the Apache Spark cluster, the original alarm data (i.e., each alarm information) is cleaned and normalized, and an outlier detection algorithm, such as the IQR (interquartile range) method, is introduced to remove or correct outliers during the data cleaning stage to improve data quality. The cleaned alarm data is then standardized for subsequent analysis.

[0139] Association rule mining module;

[0140] Use the distributed association rule algorithm based on Apriori or FP-growth to mine the association of alarm data. Specifically, decide which association rule algorithm to use based on the characteristics of the alarm data (such as the amount of data and the number of items). If the amount of data is large and the number of items is large, the FP-growth algorithm is used to improve efficiency. Through the association rule algorithm, the frequent patterns and associations between alarms are discovered, and the association relationship between alarms is obtained. This step is performed on the Hadoop or Spark framework to effectively process massive data sets.

[0141] Intelligent analysis module;

[0142] Through pre-built integrated machine learning models, such as time series analysis, isolation forests, or deep learning networks, pattern recognition and anomaly detection are performed on alarm sequences with correlations (sequences of temporal relationships, sequences of causal relationships), abnormal patterns of alarms and potential precursors to failures are identified, so as to obtain the root causes of alarms in the alarm sequence. During the model training process, model training sets, validation sets, and test sets are set up, and cross-validation and other methods are used to evaluate model performance, such as accuracy, recall, F1 score, and other indicators to ensure model effectiveness and generalization capabilities.

[0143] Ensemble machine learning models;

[0144] Such as time series analysis, isolation forest or deep learning network, which is used to perform pattern recognition and anomaly detection on alarm sequences, and identify abnormal patterns of alarms and potential precursors to failures.

[0145] Graph database module;

[0146] Build an alarm relationship graph, use Neo4j or JanusGraph to store and query the dependencies between alarms, and implement root cause analysis through graph algorithms such as the shortest path algorithm. Perform partition management and index optimization on the alarm relationship graph, for example, pre-partition according to alarm type or service category, and use the built-in index mechanism of the graph database to speed up the query process.

[0147] Notification and response module;

[0148] Integrate the alarm notification system with automated operation and maintenance tools, such as PagerDuty and Ansible, to automatically trigger alarm notifications and preset fault response strategies based on correlation analysis results. Implement alarm priority sorting logic to ensure that critical alarms are notified first. At the same time, introduce a notification deduplication mechanism to avoid sending the same alarm to the same recipient repeatedly in a short period of time.

[0149] like Figure 3As shown, the overall implementation process of the alarm correlation analysis provided by the embodiment of the present invention mainly includes:

[0150] Distributed log collection and storage;

[0151] Use tools such as Fluentd or Logstash to deploy log agents on each cloud service node to collect application logs, system logs, and various alarm events in real time. The collected data is aggregated to a central log storage cluster such as Elasticsearch or Apache Hadoop HDFS through message queue systems such as Kafka or Flume to ensure high availability and scalability of data.

[0152] Alarm standardization and processing;

[0153] Through the designed alarm standardization model, the collected raw alarm information is formatted and the alarm level, type, description and other fields are unified to facilitate subsequent analysis. Alertmanager or other alarm management systems are used to perform preliminary filtering and denoising of alarm data, and to perform preliminary classification and aggregation of alarms according to predefined rules.

[0154] Relevance analysis engine;

[0155] Real-time stream processing and analysis: Use Apache Flink or Spark Streaming to build real-time stream processing pipelines to perform real-time analysis on the continuously flowing log and alarm data streams.

[0156] Complex event processing (CEP) technology is applied to identify the temporal and spatial correlations between alarms, such as similar alarms that appear in a short period of time or alarm sequences that appear in a specific order.

[0157] Distributed association rule mining;

[0158] The association rule mining based on FP-growth algorithm is introduced to run on the distributed computing framework (such as Hadoop or Spark) for the historical alarm data set to discover the implicit association rules between alarms.

[0159] Use graph databases (such as Neo4j) to build relationship graphs between alarms, visualize the dependencies and causal relationships between alarms, and help maintenance personnel understand the system status. Figure 4 and Figure 5 shown.

[0160] Root cause analysis and alarm noise reduction;

[0161] Root cause location based on graph analysis technology: Use alarm log correlation analysis technology based on the above relationship graph to build a directed acyclic graph (DAG) between alarms, determine the shortest path or the maximum weight path, and identify the most likely fault source in the DAG.

[0162] Combine machine learning models (such as random forests or gradient boosted trees) to score each alarm in the relationship graph and further perform precise root cause analysis.

[0163] Intelligent alarm noise reduction;

[0164] By learning the data of historical alarms and solutions, the alarm noise reduction model is trained to automatically suppress non-critical alarms or repeated alarms with known solutions, thus reducing the burden of operation and maintenance. In other words, through the alarm noise reduction model, critical alarms or alarms that need to be manually resolved can be screened out from various alarm information.

[0165] Real-time notification and response;

[0166] Combine Prometheus and Alertmanager to achieve intelligent routing of alarms, and send notifications to relevant personnel or teams via email, SMS or instant messaging tools based on the alarm level, business impact scope and pre-set notification strategy.

[0167] Integrate automated operation and maintenance tools (such as Ansible and Terraform) to trigger automated scripts or workflows to perform fault recovery operations for alarms with known solutions.

[0168] In general, the method provided by the embodiment of the present invention:

[0169] Through the data acquisition modules distributed in each node of the network, the alarm data is captured and summarized in real time;

[0170] Preprocess the collected alarm data, including standardization and denoising, to prepare for subsequent analysis; use stream processing technology to perform online analysis on real-time data streams to identify alarm patterns and anomalies;

[0171] Perform association rule mining based on historical data to discover the correlation and causal chain between alarms;

[0172] Apply machine learning models for intelligent analysis and root cause location;

[0173] The analysis results are displayed through visualization tools to assist operation and maintenance decision-making.

[0174] The method provided in the embodiment of the present invention can dynamically adjust the alarm threshold and reduce the false alarm rate based on the historical alarm frequency and business load, geographic location, time period and other data through the alarm noise reduction model. Use Spark Streaming or Flink to process the real-time alarm stream to ensure the timeliness of alarm correlation analysis. Continuously learn new alarm patterns through machine learning models to optimize the accuracy of association rules and root cause analysis. It is also possible to consider introducing expert knowledge and rules to form a hybrid intelligent system to improve the interpretation and generalization capabilities of the model. Density-based DBSCAN or hierarchical clustering methods are used to select appropriate algorithms based on alarm characteristics and business scenarios. Identify alarm groups, eliminate redundant alarms, and focus on real problem points. Visualize the alarm association relationship, allow batch closure, and improve practicality. Provide an interactive alarm association analysis diagram. Rapid fault location and response in distributed cloud platform scenarios can be achieved to ensure efficient and stable cloud services.

[0175] The method provided in the embodiment of the present invention combines tools and technologies such as Fluentd, Spark Streaming, FP-growth, and Neo4j to achieve full-link optimization from data collection to fault response, significantly improving the problem diagnosis efficiency and problem location speed, and can be widely used in intelligent manufacturing, network monitoring, cloud computing and other fields. Through accurate correlation analysis and root cause location, it is helpful to reduce false alarms and missed alarms, significantly improve the operation and maintenance efficiency of the cloud service platform, shorten the mean time to resolve (MTTR), and improve service quality. The operation and maintenance team can quickly identify the root cause of the problem from massive alarms, reduce manual intervention, achieve more efficient alarm management and fault response, ensure the high availability of cloud services, and improve user experience.

[0176] and Figure 1 Corresponding to the alarm correlation analysis method shown in FIG. 1 , the embodiment of the present invention further provides an alarm correlation analysis device for Figure 1 The specific implementation of the method shown in is shown in the structural diagram Figure 6 As shown, including:

[0177] A first determining unit 201 is used to determine an alarm data set to be processed; the alarm data set includes a plurality of alarm information;

[0178] The standardization unit 202 is used to perform data standardization processing on the alarm data set to obtain a standardized data set; the standardized data set includes each alarm information that has been standardized;

[0179] A filtering unit 203 is used to perform alarm filtering processing on the standardized data set, and form each alarm information obtained after the processing into a data set to be analyzed;

[0180] The association mining unit 204 is used to perform association mining processing on the data set to be analyzed based on a preset association rule mining algorithm to obtain an alarm association analysis result corresponding to the data set to be analyzed;

[0181] The second determining unit 205 is used to determine a correlation data set according to the alarm correlation analysis result; the correlation data set includes each alarm information having a correlation relationship in the data set to be analyzed;

[0182] A graph construction unit 206, configured to establish an alarm association relationship graph corresponding to the associated data set, and to construct an alarm directed acyclic graph corresponding to the alarm association relationship graph;

[0183] A root cause analysis unit 207, configured to perform an alarm root cause analysis on the associated data set based on the alarm directed acyclic graph, and obtain a root cause analysis result corresponding to the associated data set;

[0184] The third determining unit 208 is configured to use the alarm association relationship graph, the alarm directed acyclic graph and the root cause analysis result as the association analysis result corresponding to the association data set.

[0185] By using the device provided by the embodiment of the present invention, the association mining of each alarm information to be analyzed can be performed through the association rule mining algorithm to obtain the association analysis results between the alarms, determine the associated data set according to the association analysis results, establish the alarm association relationship map and the alarm directed acyclic graph corresponding to the associated data set, and perform the root cause analysis of the alarm on the associated data set, and then obtain the association analysis results. When the operation and maintenance personnel need to process the alarm information in the associated data set, the association analysis results can be fed back to the operation and maintenance personnel, so that the operation and maintenance personnel can perform targeted operations based on the association analysis between the alarms, and quickly complete the operation and maintenance of the alarm from the root, which is conducive to improving the alarm processing speed, realizing the rapid location of abnormal problems, and then improving the operation and maintenance efficiency, and ensuring the operational stability of the business system.

[0186] exist Figure 6 Based on the device shown, the device provided by the embodiment of the present invention can be further expanded into multiple units. The function of each unit can refer to the description of each embodiment provided in the above text for the alarm correlation analysis method, and no further examples will be given here.

[0187] An embodiment of the present invention further provides a storage medium, which includes stored instructions, wherein when the instructions are executed, the device where the storage medium is located is controlled to execute the above-mentioned alarm correlation analysis method.

[0188] The embodiment of the present invention further provides an electronic device, the structural diagram of which is shown in FIG. Figure 7As shown, it specifically includes a memory 301 and one or more instructions 302, wherein the one or more instructions 302 are stored in the memory 301 and are configured to be executed by one or more processors 303 to perform the following operations:

[0189] Determine an alarm data set to be processed; the alarm data set includes a plurality of alarm information;

[0190] Performing data standardization processing on the alarm data set to obtain a standardized data set; the standardized data set includes each alarm information that has undergone standardization processing;

[0191] Performing alarm filtering processing on the standardized data set, and forming a data set to be analyzed with each alarm information obtained after the processing;

[0192] Based on a preset association rule mining algorithm, an association mining process is performed on the data set to be analyzed to obtain an alarm association analysis result corresponding to the data set to be analyzed;

[0193] Determine a correlation data set according to the alarm correlation analysis result; the correlation data set includes each alarm information having a correlation relationship in the data set to be analyzed;

[0194] Establishing an alarm association relationship graph corresponding to the associated data set, and constructing an alarm directed acyclic graph corresponding to the alarm association relationship graph;

[0195] Based on the alarm directed acyclic graph, performing alarm root cause analysis on the associated data set to obtain a root cause analysis result corresponding to the associated data set;

[0196] The alarm association relationship graph, the alarm directed acyclic graph and the root cause analysis result are used as the association analysis result corresponding to the association data set.

[0197] Each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can refer to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the system or system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can refer to the partial description of the method embodiment. The system and system embodiments described above are merely schematic, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without creative work.

[0198] Professionals may further appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in the above description according to function. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians may use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0199] The above description of the disclosed embodiments enables one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for analyzing alarm association, characterized in that: include: Determine an alarm data set to be processed; the alarm data set includes a plurality of alarm information; Performing data standardization processing on the alarm data set to obtain a standardized data set; The standardized data set includes each warning information that has been standardized; Performing alarm filtering processing on the standardized data set, and forming a data set to be analyzed with each alarm information obtained after the processing; Based on a preset association rule mining algorithm, an association mining process is performed on the data set to be analyzed to obtain an alarm association analysis result corresponding to the data set to be analyzed; Determine a correlation data set according to the alarm correlation analysis result; the correlation data set includes each alarm information having a correlation relationship in the data set to be analyzed; Establishing an alarm association relationship graph corresponding to the associated data set, and constructing an alarm directed acyclic graph corresponding to the alarm association relationship graph; Based on the alarm directed acyclic graph, performing alarm root cause analysis on the associated data set to obtain a root cause analysis result corresponding to the associated data set; The alarm association relationship graph, the alarm directed acyclic graph and the root cause analysis result are used as the association analysis result corresponding to the association data set.

2. The alarm correlation analysis method according to claim 1, characterized in that: The step of performing data standardization processing on the alarm data set to obtain a standardized data set includes: Performing data cleaning processing on each alarm information in the alarm data set to obtain each alarm information after data cleaning; Performing data formatting processing on each of the alarm information that has undergone data cleaning to obtain a data formatting result corresponding to each alarm information in the alarm data set; Each of the data formatting results is used as standardized alarm information, and the standardized data set is composed of the various data formatting results.

3. The alarm correlation analysis method according to claim 1, characterized in that: The step of performing alarm filtering processing on the standardized data set and forming a data set to be analyzed with each alarm information obtained after the processing includes: Preliminarily filtering each alarm information in the standardized data set based on a preset alarm management module to obtain a preliminary screening data set; the preliminary screening data set includes each alarm information that has been preliminarily filtered; Based on a preset complex event processing mechanism, the temporal and spatial correlation of the preliminary screening data set is identified to obtain the temporal and spatial correlation relationship of each alarm information in the preliminary screening data set; The various alarm information in the primary screening data set is secondary filtered according to the spatiotemporal correlation, and the various alarm information that have undergone secondary filtering are combined into the data set to be analyzed.

4. The alarm correlation analysis method according to claim 1, characterized in that: The method of performing association mining processing on the data set to be analyzed based on a preset association rule mining algorithm to obtain an alarm association analysis result corresponding to the data set to be analyzed includes: Determining the alarm data characteristic information corresponding to the data set to be analyzed; Determine a target algorithm that matches the characteristic information of the alarm data among a plurality of pre-set distributed association rule mining algorithms; The target algorithm is used to perform data mining on the data set to be analyzed, and the data mining result is used as the alarm association analysis result.

5. The alarm correlation analysis method according to claim 1, characterized in that: The step of performing alarm root cause analysis on the associated data set based on the alarm directed acyclic graph to obtain a root cause analysis result corresponding to the associated data set includes: By using a preset path identification strategy, the root node of the alarm directed acyclic graph is identified to obtain the root node in the alarm directed acyclic graph; By using a pre-built alarm root source identification model, the root source alarm identification is performed on the associated data set to obtain a score corresponding to each alarm information in the associated data set, wherein the score represents the possibility that the corresponding alarm belongs to the alarm root source; The root cause analysis result corresponding to the associated data set is determined according to the root source node in the alarm directed acyclic graph and the score corresponding to each of the alarm information.

6. The alarm correlation analysis method according to claim 1, characterized in that: Also includes: Performing alarm denoising processing on the data set to be analyzed by using a pre-built alarm denoising model, and using the alarm information obtained after the processing as the alarm information to be notified; For each of the alarm information to be notified, determining evaluation information corresponding to the alarm information to be notified, the evaluation information including an alarm level and a business impact range corresponding to the alarm information to be notified; Determine the notification method and notification personnel corresponding to each of the alarm information to be notified based on the preset notification strategy and the evaluation information corresponding to each of the alarm information to be notified; For each of the alarm information to be notified, based on the notification method corresponding to the alarm information to be notified, the alarm notification corresponding to the alarm information to be notified is sent to a corresponding notification person.

7. The alarm correlation analysis method according to claim 1, characterized in that: Also includes: Determining target warning information in the data set to be analyzed; The target warning information is warning information of a known solution; Determine, among various preset alarm solutions, a target alarm solution corresponding to the target alarm information; A pre-set automated operation and maintenance tool is called to trigger a fault recovery operation corresponding to the target alarm solution to perform an alarm processing response to the target alarm information.

8. An alarm correlation analysis device, characterized in that: include: A first determining unit is used to determine an alarm data set to be processed; the alarm data set includes a plurality of alarm information; A standardization unit, used for performing data standardization processing on the alarm data set to obtain a standardized data set; The standardized data set includes each warning information that has been standardized; A filtering unit, used to perform alarm filtering processing on the standardized data set, and form a data set to be analyzed by filtering each alarm information obtained after the processing; An association mining unit, configured to perform association mining processing on the data set to be analyzed based on a preset association rule mining algorithm, and obtain an alarm association analysis result corresponding to the data set to be analyzed; A second determining unit is used to determine a correlation data set according to the alarm correlation analysis result; the correlation data set includes each alarm information having a correlation relationship in the data set to be analyzed; A graph construction unit, used to establish an alarm association relationship graph corresponding to the associated data set, and to construct an alarm directed acyclic graph corresponding to the alarm association relationship graph; A root cause analysis unit, configured to perform an alarm root cause analysis on the associated data set based on the alarm directed acyclic graph, and obtain a root cause analysis result corresponding to the associated data set; The third determining unit is used to use the alarm association relationship map, the alarm directed acyclic graph and the root cause analysis result as the association analysis result corresponding to the association data set.

9. A storage medium, characterized in that: The storage medium includes stored instructions, wherein when the instructions are executed, the device where the storage medium is located is controlled to execute the alarm correlation analysis method according to any one of claims 1 to 7.

10. An electronic device, characterized in that: The system comprises a memory and one or more instructions, wherein the one or more instructions are stored in the memory and configured to be executed by one or more processors to perform the alarm correlation analysis method according to any one of claims 1 to 7.

Citation Information

Cited By

  • Alarm data processing method, program product, electronic equipment and storage medium

    CN120223413A

  • Alarm root cause positioning and repairing method, device and equipment of cloud platform

    CN120750739A