A method for policing and related equipment
By building an object relationship map in the IT system and using spectral clustering technology, the weight value is calculated based on the correlation characteristics between objects and clustering to process the alarm cluster, the multi-object alarm problem caused by failure in the IT system is solved, and the alarm is improved Effectiveness and fault location efficiency.
Patent Information
- Application Number
- CN202110763673.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2021-07-06
- Publication Date
- 2024-12-06
- Estimated Expiration
- 2041-07-06
AI Technical Summary
In complex IT systems, failures may cause multiple objects to alert, making it difficult for technicians to locate the cause of the failure in a timely manner, reducing the effectiveness of the alarm.
By constructing an object relationship map, spectral clustering technology calculates the weight value of edges based on the correlation characteristics (configuration relationship, call relationship and alarm relationship) between objects, clustering to obtain alarm clusters, thereby reducing the number of alarms and increasing alarms overall effectiveness.
The alarm cluster obtained after clustering processing is more accurate, reducing the number of alarms, and improving the efficiency of technicians to locate the cause of failure and the overall effectiveness of alarms.
Smart Images

Figure CN113486192B_ABST
Abstract
Description
Technical Field
[0001] The present disclosure relates to the field of Internet technology, and in particular to an alarm aggregation method and related equipment. Background Art
[0002] With the annual development of Internet Technology (IT), the IT architecture and IT environment have also undergone profound changes against the backdrop of the gradual popularization of microservices and distributed architecture, the cloudification of Internet infrastructure, and the growing development of Internet business.
[0003] While promoting business development, various emerging Internet technologies have increased the complexity of IT systems, making fault emergency handling face increasing pressure. Fault emergency handling requires timely fault detection capabilities, and the most effective means of fault detection is alarm. However, due to the large number and variety of objects in IT systems and their complex relationships, a fault may cause multiple object alarms. Therefore, compressing or merging alarms can effectively reduce the number of feedback alarms, help technicians locate the cause of the fault in a timely manner, and improve the overall effectiveness of alarms. Summary of the invention
[0004] In view of the above problems, the present disclosure provides an alarm aggregation method and related devices that overcome the above problems or at least partially solve the above problems. The technical solution is as follows:
[0005] An alarm aggregation method, comprising:
[0006] Determining at least one target object that is alerting within a preset time window;
[0007] Determine the target object nodes corresponding to each of the target objects in a preset object relationship graph, the preset object relationship graph is a knowledge graph constructed according to the association features between the objects, the association features include configuration relationships, call relationships, and alarm association relationships, the preset object relationship graph has object nodes and edges connecting the object nodes, each edge corresponds to a preset weight value, wherein the weight value corresponding to each edge is related to the association feature between the two objects connected by the edge;
[0008] According to the weight values of the edges connecting the target object nodes, clustering processing is performed using spectral clustering to obtain at least one alarm cluster.
[0009] Optionally, performing clustering processing by using spectral clustering according to the weight values of the edges connecting the target object nodes to obtain at least one alarm cluster includes:
[0010] Taking each of the target object nodes as a sample set;
[0011] According to the weight values of the edges connecting the target object nodes, an adjacency graph corresponding to the sample set is constructed by using a full connection method to obtain an adjacency matrix and a degree matrix corresponding to the adjacency graph;
[0012] According to the adjacency matrix and the degree matrix, a Laplace matrix is calculated and normalized;
[0013] Performing eigenvalue decomposition on the normalized Laplace matrix to obtain a characteristic matrix;
[0014] Each row vector in the feature matrix is clustered using a k-means clustering algorithm to obtain at least one alarm cluster.
[0015] Optionally, after obtaining at least one alarm cluster, the method further includes:
[0016] Outputting alarm information according to the alarm cluster.
[0017] Optionally, the process of determining the weight value corresponding to the edge in the preset object relationship graph includes:
[0018] For an edge between any two object nodes in the preset object relationship graph: using a Gaussian similarity function to calculate an initial weight value of the edge;
[0019] If the two object nodes are in the configuration relationship, the first preset value is subtracted from the initial weight value to obtain the target weight value; if the two object nodes are in the call relationship, the second preset value is added to the initial weight value to obtain the target weight value;
[0020] The target weight value is determined to be the weight value corresponding to the edge.
[0021] Optionally, the process of determining the alarm association relationship includes:
[0022] Obtaining historical alarm records, wherein the historical alarm records include alarm records of at least one object;
[0023] The historical alarm records are analyzed using an association rule analysis algorithm to determine whether the objects have the alarm association relationship.
[0024] Optionally, the process of determining the weight value corresponding to the edge in the preset object relationship graph includes:
[0025] For an edge between any two object nodes in the preset object relationship graph: using a Gaussian similarity function to calculate a first weight value of the edge;
[0026] For the edge between the object nodes corresponding to the two objects having the alarm association relationship in the preset object relationship graph: add the first weight value of the edge to the third preset value to obtain the second weight value; if the two object nodes are in the configuration relationship, subtract the fourth preset value from the second weight value to obtain the target weight value; if the two object nodes are in the call relationship, add the fifth preset value to the second weight value to obtain the target weight value; if the two object nodes do not have the configuration relationship and the call relationship, determine the second weight value as the target weight value;
[0027] The target weight value is determined to be the weight value corresponding to the edge.
[0028] Optionally, the objects include application systems, clusters, middleware, physical devices, virtual devices, databases, storage, and networks in the Internet system.
[0029] An alarm aggregation device comprises: a target object determination unit, a target object node determination unit and an alarm cluster acquisition unit,
[0030] The target object determination unit is used to determine at least one target object that is warning within a preset time window;
[0031] The target object node determination unit is used to respectively determine the target object nodes corresponding to each of the target objects in a preset object relationship graph, wherein the preset object relationship graph is a knowledge graph constructed according to the association features between the objects, wherein the association features include configuration relationships, call relationships, and alarm association relationships, and the preset object relationship graph has object nodes and edges connecting the object nodes, wherein each edge corresponds to a preset weight value, wherein the weight value corresponding to each edge is related to the association feature between the two objects connected by the edge;
[0032] The alarm cluster obtaining unit is used to perform clustering processing by using spectral clustering according to the weight values of the edges connecting the target object nodes, so as to obtain at least one alarm cluster.
[0033] A computer-readable storage medium stores a program, wherein when the program is executed by a processor, the program implements any of the above-mentioned alarm aggregation methods.
[0034] An electronic device, comprising at least one processor, and at least one memory and a bus connected to the processor; wherein the processor and the memory communicate with each other via the bus; and the processor is used to call program instructions in the memory to execute the alarm aggregation method as described in any one of the above items.
[0035] By means of the above technical scheme, the present disclosure provides an alarm aggregation method and related equipment, which determines at least one target object that is alarming within a preset time window; respectively determines the target object nodes corresponding to the target object in the preset object relationship graph, the preset object relationship graph is a knowledge graph constructed according to the association features between the objects, the association features include configuration relationships, call relationships and alarm association relationships, the preset object relationship graph has object nodes and edges connecting the object nodes, each edge corresponds to a preset weight value, and the weight value corresponding to each edge is related to the association feature between the two objects connected by the edge; according to the weight value of the edge connected to each target object node, spectral clustering is used to perform clustering processing to obtain at least one alarm cluster. The present disclosure constructs an object relationship graph and sets the weight value corresponding to the edge between the object nodes, so that the alarm cluster obtained after clustering is more accurate, thereby reducing the number of alarms and improving the overall effectiveness of the alarms.
[0036] The above description is only an overview of the technical solution of the present disclosure. In order to more clearly understand the technical means of the present disclosure, it can be implemented according to the contents of the specification. In order to make the above and other purposes, features and advantages of the present disclosure more obvious and easy to understand, the specific implementation methods of the present disclosure are listed below. BRIEF DESCRIPTION OF THE DRAWINGS
[0037] Various other advantages and benefits will become apparent to those of ordinary skill in the art by reading the detailed description of the preferred embodiments below. The accompanying drawings are only for the purpose of illustrating the preferred embodiments and are not to be considered as limiting the present disclosure. Also, the same reference symbols are used throughout the accompanying drawings to represent the same components. In the accompanying drawings:
[0038] Figure 1 A schematic diagram showing a flow chart of an implementation of the alarm aggregation method provided by an embodiment of the present disclosure;
[0039] Figure 2 A schematic diagram of a process for determining an alarm association relationship provided by an embodiment of the present disclosure is shown;
[0040] Figure 3 A schematic diagram showing a process for determining weight values corresponding to edges in a preset object relationship graph provided by an embodiment of the present disclosure is shown;
[0041] Figure 4 A schematic diagram showing another process for determining weight values corresponding to edges in a preset object relationship graph provided by an embodiment of the present disclosure is shown;
[0042] Figure 5 A flow chart showing another implementation of the alarm aggregation method provided in an embodiment of the present disclosure is shown;
[0043] Figure 6A flow chart showing another implementation of the alarm aggregation method provided in an embodiment of the present disclosure is shown;
[0044] Figure 7 A schematic diagram of the structure of an alarm aggregation device provided by an embodiment of the present disclosure is shown;
[0045] Figure 8 A schematic structural diagram of an electronic device provided by an embodiment of the present disclosure is shown. DETAILED DESCRIPTION
[0046] The exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although the exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art.
[0047] like Figure 1 As shown, an implementation of the alarm aggregation method provided by the embodiment of the present disclosure may include:
[0048] S100: Determine at least one target object that is alerting within a preset time window.
[0049] The preset time window may be a pre-set time interval. Optionally, the preset time window may be 5 minutes, 10 minutes, and 30 minutes. Since the alarm is streaming data, it is impossible to determine all alarm objects at once and then cluster them. Therefore, the disclosed embodiment sets a time window and determines the object that is alarming within a time window as the target object, which can improve the efficiency and accuracy of alarm aggregation.
[0050] It is understandable that the objects may include physical objects and logical objects in the Internet system. For example, physical objects including application systems and physical devices and logical objects including clusters. Optionally, the objects in the embodiments of the present disclosure include application systems, clusters, middleware, physical devices, virtual devices, databases, storage, and networks in the Internet system.
[0051] S200. Determine the target object nodes corresponding to each target object in the preset object relationship graph. The preset object relationship graph is a knowledge graph constructed according to the association characteristics between the objects. The association characteristics include configuration relationships, call relationships, and alarm association relationships. The preset object relationship graph has object nodes and edges connecting the object nodes. Each edge corresponds to a preset weight value, wherein the weight value corresponding to each edge is related to the association characteristics between the two objects connected by the edge.
[0052] Optionally, the disclosed embodiment may first construct a first knowledge graph using the configuration relationship between objects, and then, based on the first knowledge graph, construct a second knowledge graph according to the calling relationship between objects, and then, based on the second knowledge graph, construct a third knowledge graph according to the alarm association relationship between objects, and determine the third knowledge graph as the preset object relationship graph.
[0053] The configuration relationship includes the containment relationship, deployment relationship, operation relationship and connection relationship. As a kind of quasi-static information, when the configuration relationship changes, it can be synchronized to the preset object relationship map at a fixed time.
[0054] The calling relationship mainly refers to the calling relationship between application layers. The alarm of the application layer is usually further transmitted along with the calling relationship. The embodiment of the present disclosure can determine the calling relationship between application layers according to the application data between application layers. The application data includes application transaction details data.
[0055] Among them, the alarm association relationship is that when an object is alarming, there will always be other objects that are also alarming, so it can be determined that the object has an alarm association relationship with the other object. The embodiment of the present disclosure can use the association rule analysis algorithm (FP-Growth) to mine the alarm association relationship between objects. The association rule analysis algorithm can mine the implicit frequent patterns between objects, and the frequent patterns can indicate that there is an alarm association relationship between objects.
[0056] Optional, such as Figure 2 As shown, the process of determining the alarm association relationship provided by the embodiment of the present disclosure may include:
[0057] S001. Obtain historical alarm records.
[0058] The historical alarm record includes an alarm record of at least one object. The historical alarm record may be an alarm sequence arranged according to the alarm occurrence time.
[0059] S002. Analyze historical alarm records using an association rule analysis algorithm to determine whether there is an alarm association relationship between objects.
[0060] The disclosed embodiment constructs an object relationship graph through the configuration relationship, call relationship and alarm association relationship between objects, so that the object relationship graph is closer to the association relationship of each object when an alarm occurs in actual situations, so as to make the subsequent aggregation results more accurate.
[0061] It is understandable that an object corresponds to an object node in a preset object relationship graph. The disclosed embodiment can connect the preset object relationship graph to the streaming alarm through an open source stream processing platform, so that after the object alarm, the alarm corresponds to an object node in the preset object relationship graph. Optionally, the open source stream processing platform may include Flink or Kafka.
[0062] In the preset object relationship graph, the weight of an edge depends on the possibility that the two object nodes connected by the edge can be aggregated together when an alarm occurs. The larger the weight, the greater the possibility of aggregation.
[0063] Optional, such as Figure 3 As shown, a process for determining the weight value corresponding to the edge in the preset object relationship graph provided by the embodiment of the present disclosure may include:
[0064] A100. For an edge between any two object nodes in a preset object relationship graph: use a Gaussian similarity function to calculate an initial weight value of the edge.
[0065] Among them, the Gaussian similarity function is also called Gaussian similarity function.
[0066] A200. If the two object nodes are in a configuration relationship, subtract a first preset value from the initial weight value to obtain a target weight value.
[0067] When there is a configuration relationship between two object nodes, the configuration relationship can be regarded as a weak relationship because the alarm generated at the bottom layer may not be transmitted upward. Therefore, the embodiment of the present disclosure can subtract a first preset value from the initial weight value of the edge between two object nodes whose association relationship is a configuration relationship, thereby reducing the possibility of aggregation of the two object nodes.
[0068] A300. If the two object nodes are in a calling relationship, add the initial weight value to the second preset value to obtain the target weight value.
[0069] When there is a calling relationship between two object nodes, the alarm of the called object will usually be passed forward and affect each application on the calling chain. Therefore, the embodiment of the present disclosure can subtract a second preset value from the initial weight value of the edge between two object nodes with a calling relationship, thereby increasing the possibility of aggregation of the two object nodes.
[0070] The second preset value may be the same as or different from the first preset value.
[0071] It is understandable that the first preset value and the second preset value can be set according to requirements.
[0072] A400. Determine the target weight value as the weight value corresponding to the edge.
[0073] The disclosed embodiment reasonably sets the weight value corresponding to the edge between two object nodes of configuration relationship and call relationship, so that the possibility of two object nodes being aggregated is closer to the alarm aggregation in the actual environment, making the subsequent aggregation results more accurate.
[0074] Optional, such as Figure 4 As shown, another process for determining the weight value corresponding to the edge in the preset object relationship graph provided by the embodiment of the present disclosure may include:
[0075] B100. For an edge between any two object nodes in a preset object relationship graph: use a Gaussian similarity function to calculate a first weight value of the edge.
[0076] B200. For an edge between corresponding object nodes of two objects having an alarm association relationship in a preset object relationship graph: add a third preset value to the first weight value of the edge to obtain a second weight value.
[0077] Since the first weight value calculated by the Gaussian similarity function is obtained through deductive reasoning, and two object nodes with an alarm association relationship may not have a configuration relationship and a call relationship, the first weight value of the edge between two object nodes with an implicit frequent pattern can be increased to increase the possibility of aggregation of two objects with an alarm association relationship.
[0078] It is understandable that the third preset value can be set according to demand.
[0079] B300. If the two object nodes are in a configuration relationship, subtract the fourth preset value from the second weight value to obtain a target weight value.
[0080] B400. If the two object nodes are in a calling relationship, add the second weight value to the fifth preset value to obtain a target weight value.
[0081] B500. If there is no configuration relationship or call relationship between the two object nodes, determine the second weight value as the target weight value.
[0082] The fifth preset value may be the same as or different from the fourth preset value.
[0083] It is understandable that the fourth preset value and the fifth preset value can be set according to requirements.
[0084] B600. Determine the target weight value as the weight value corresponding to the edge.
[0085] The embodiment of the present disclosure reasonably sets the weight value corresponding to the edge between the corresponding object nodes of two objects with an alarm association relationship in a preset object relationship graph, so that the possibility of the two object nodes being aggregated is closer to the alarm aggregation in the actual environment, making the subsequent aggregation results more accurate.
[0086] S300 , performing clustering processing by using spectral clustering according to weight values of edges connecting target object nodes to obtain at least one alarm cluster.
[0087] Optionally, the disclosed embodiment may construct an adjacency graph using one of the ε nearest neighbor method, the k nearest neighbor method, and the full connection method. The adjacency graph is also called a similarity graph. A Laplace matrix is obtained from the adjacency graph, and the Laplace matrix is subjected to eigenvalue decomposition. The eigenvalues obtained after the decomposition are processed to construct a cluster, and a cluster is used as an alarm cluster.
[0088] Optional, based on Figure 1 The method shown, such as Figure 5 As shown, in another implementation of the alarm aggregation method provided in the embodiment of the present disclosure, step S300 may include:
[0089] S310: Take each target object node as a sample set.
[0090] S320. According to the weight values of the edges connecting the target object nodes, an adjacency graph corresponding to the sample set is constructed using a full connection method to obtain an adjacency matrix and a degree matrix corresponding to the adjacency graph.
[0091] S330: Calculate a Laplace matrix according to the adjacency matrix and the degree matrix, and perform normalization processing on the Laplace matrix.
[0092] S340, performing eigenvalue decomposition on the normalized Laplace matrix to obtain a characteristic matrix.
[0093] S350 , clustering each row vector in the feature matrix using a k-means clustering algorithm to obtain at least one alarm cluster.
[0094] The number of alarm clusters may be determined by any algorithm including the error square sum method (also called elbow method or elbow method), the average silhouette coefficient method, and the Gap Statistic algorithm.
[0095] Among them, the core idea of the error square sum method is that: as the number of clusters k increases, the sample division will be more refined, and the degree of aggregation of each cluster will gradually increase, so the error square sum SSE will naturally gradually decrease. When k is less than the actual number of clusters, since the increase in k will greatly increase the degree of aggregation of each cluster, the decrease in SSE will be large, and when k reaches the actual number of clusters, the return of the degree of aggregation obtained by increasing k will rapidly decrease, so the decrease in SSE will drop sharply, and then tend to be flat as the k value continues to increase. Therefore, the embodiment of the present disclosure can determine the actual number of clusters as the number of alarm clusters generated.
[0096] The embodiment of the present disclosure adopts the full connection method to construct an adjacency graph, which can reduce the clustering error caused by the uneven density distribution of the target object nodes in the sample set in the preset object relationship graph, so that the alarm cluster obtained after clustering is more accurate.
[0097] The present disclosure provides an alarm aggregation method, which determines at least one target object that is alarming within a preset time window; determines the target object nodes corresponding to the target objects in the preset object relationship graph, the preset object relationship graph is a knowledge graph constructed according to the association features between the objects, the association features include configuration relationships, call relationships and alarm association relationships, the preset object relationship graph has object nodes and edges connecting the object nodes, each edge corresponds to a preset weight value, and the weight value corresponding to each edge is related to the association feature between the two objects connected by the edge; according to the weight value of the edge connected to each target object node, spectral clustering is used to perform clustering processing to obtain at least one alarm cluster. The present disclosure constructs an object relationship graph and sets the weight value corresponding to the edge between the object nodes, so that the alarm cluster obtained after clustering is more accurate, thereby reducing the number of alarms and improving the overall effectiveness of the alarms.
[0098] Optional, based on Figure 1 The method shown, such as Figure 6 As shown, another implementation of the alarm aggregation method provided by the embodiment of the present disclosure, after step S300, the method further includes:
[0099] S400: Output alarm information according to the alarm cluster.
[0100] The embodiment of the present disclosure may output corresponding alarm information to each alarm cluster respectively. Optionally, the alarm information may include the fault location and the fault cause.
[0101] The embodiment of the present disclosure can promptly notify a technician of an object failure by outputting alarm information, so that the technician can promptly eliminate the failure based on the alarm information.
[0102] It should be understood that although the present disclosure depicts operations in a particular order, this should not be construed as requiring that the operations be performed in the particular order shown or in a sequential order.Multitasking and parallel processing may be advantageous under certain circumstances.
[0103] It should be understood that the various steps described in the method embodiments of the present disclosure may be performed in different orders and / or in parallel. In addition, the method embodiments may include additional steps and / or omit the steps shown. The scope of the present disclosure is not limited in this respect.
[0104] Corresponding to the above method embodiment, the present disclosure embodiment also provides an alarm aggregation device, whose structure is as follows: Figure 7 As shown, it may include: a target object determining unit 100 , a target object node determining unit 200 and an alarm cluster obtaining unit 300 .
[0105] The target object determination unit 100 is used to determine at least one target object that is warning within a preset time window.
[0106] The target object node determination unit 200 is used to respectively determine the target object nodes corresponding to each target object in the preset object relationship graph. The preset object relationship graph is a knowledge graph constructed based on the association characteristics between each object. The association characteristics include configuration relationships, call relationships, and alarm association relationships. The preset object relationship graph has object nodes and edges connecting each object node. Each edge corresponds to a preset weight value, wherein the weight value corresponding to each edge is related to the association characteristics between the two objects connected by the edge.
[0107] The alarm cluster obtaining unit 300 is used to perform clustering processing by using spectral clustering according to the weight values of the edges connecting the target object nodes, so as to obtain at least one alarm cluster.
[0108] Optionally, the alarm cluster acquisition unit 300 can be specifically used to take each target object node as a sample set; based on the weight value of the edge connecting each target object node, use the full connection method to construct an adjacency graph corresponding to the sample set, and obtain the adjacency matrix and degree matrix corresponding to the adjacency graph; based on the adjacency matrix and the degree matrix, calculate the Laplace matrix and normalize the Laplace matrix; perform eigenvalue decomposition on the normalized Laplace matrix to obtain a feature matrix; cluster each row vector in the feature matrix using the k-means clustering algorithm to obtain at least one alarm cluster.
[0109] Optionally, the alarm aggregation device may further include an alarm information output unit.
[0110] The alarm information output unit is used to output alarm information according to the alarm cluster.
[0111] Optionally, the alarm aggregation device may further include: a first weight value determining unit.
[0112] The first weight value determination unit is used to calculate the initial weight value of the edge between any two object nodes in the preset object relationship graph using a Gaussian similarity function; if the two object nodes are in a configuration relationship, the initial weight value is subtracted from the first preset value to obtain a target weight value; if the two object nodes are in a call relationship, the initial weight value is added to the second preset value to obtain the target weight value; and the target weight value is determined to be the weight value corresponding to the edge.
[0113] Optionally, the alarm aggregation device may further include: an alarm association relationship determination unit.
[0114] The alarm association relationship determination unit is used to obtain historical alarm records, wherein the historical alarm records include alarm records of at least one object; and analyze the historical alarm records using an association rule analysis algorithm to determine whether there is an alarm association relationship between the objects.
[0115] Optionally, the alarm aggregation device may further include: a second weight value determining unit.
[0116] The second weight value determination unit is used to calculate the first weight value of the edge between any two object nodes in the preset object relationship graph using a Gaussian similarity function; for the edge between the object nodes corresponding to two objects with an alarm association relationship in the preset object relationship graph: add the first weight value of the edge to the third preset value to obtain a second weight value; if the two object nodes are in a configuration relationship, subtract the fourth preset value from the second weight value to obtain a target weight value; if the two object nodes are in a calling relationship, add the fifth preset value to the second weight value to obtain the target weight value; if the two object nodes do not have a configuration relationship or a calling relationship, determine the second weight value as the target weight value; and determine the target weight value as the weight value corresponding to the edge.
[0117] Optionally, the objects in the embodiments of the present disclosure include application systems, clusters, middleware, physical devices, virtual devices, databases, storage, and networks in an Internet system.
[0118] The present disclosure provides an alarm aggregation device, which determines at least one target object that is alarming within a preset time window; determines the target object nodes corresponding to the target objects in the preset object relationship graph, the preset object relationship graph is a knowledge graph constructed according to the association features between the objects, the association features include configuration relationships, call relationships and alarm association relationships, the preset object relationship graph has object nodes and edges connecting the object nodes, each edge corresponds to a preset weight value, and the weight value corresponding to each edge is related to the association feature between the two objects connected by the edge; according to the weight value of the edge connected to each target object node, spectral clustering is used to perform clustering processing to obtain at least one alarm cluster. The present disclosure constructs an object relationship graph and sets the weight value corresponding to the edge between the object nodes, so that the alarm cluster obtained after clustering is more accurate, thereby reducing the number of alarms and improving the overall effectiveness of the alarms.
[0119] Regarding the device in the above embodiment, the specific manner in which each module performs operations has been described in detail in the embodiment of the method, and will not be elaborated here.
[0120] The alarm aggregation device includes a processor and a memory. The target object determination unit 100, the target object node determination unit 200 and the alarm cluster acquisition unit 300 are all stored in the memory as program units, and the processor executes the program units stored in the memory to implement corresponding functions.
[0121] The processor contains a kernel, which calls the corresponding program unit from the memory. One or more kernels can be set. By adjusting the kernel parameters to build an object relationship graph and setting the weight values corresponding to the edges between object nodes, the alarm clusters obtained after clustering are more accurate, thereby reducing the number of alarms and improving the overall effectiveness of alarms.
[0122] An embodiment of the present disclosure provides a computer-readable storage medium on which a program is stored. When the program is executed by a processor, the alarm aggregation method is implemented.
[0123] An embodiment of the present disclosure provides a processor, which is used to run a program, wherein the alarm aggregation method is executed when the program is running.
[0124] like Figure 8 As shown, the embodiment of the present disclosure provides an electronic device 400, which includes at least one processor 401, at least one memory 402 connected to the processor 401, and a bus 403; wherein the processor 401 and the memory 402 communicate with each other through the bus 403; the processor 401 is used to call the program instructions in the memory 402 to execute the above-mentioned alarm aggregation method. The electronic device in this article can be a server, a PC, a PAD, a mobile phone, etc.
[0125] The present disclosure also provides a computer program product, which, when executed on an electronic device, is suitable for executing the program of initializing the steps of the alarm aggregation method.
[0126] The present disclosure is described with reference to flowcharts and / or block diagrams of methods, apparatuses, electronic devices (systems), and computer program products according to embodiments of the present disclosure. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable device to generate a machine, so that the instructions executed by the processor of the computer or other programmable device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.
[0127] In a typical configuration, an electronic device includes one or more processors (CPU), a memory, and a bus. The electronic device may also include an input / output interface, a network interface, and the like.
[0128] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM, and the memory includes at least one memory chip. The memory is an example of a computer-readable medium.
[0129] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.
[0130] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. It should also be noted that the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including 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, commodity or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.
[0131] It will be appreciated by those skilled in the art that the embodiments of the present disclosure may be provided as methods, systems or computer program products. Therefore, the present disclosure may take the form of a complete hardware embodiment, a complete software embodiment or an embodiment combining software and hardware. Moreover, the present disclosure may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.
[0132] The above are only embodiments of the present disclosure and are not intended to limit the present disclosure. For those skilled in the art, the present disclosure may have various modifications and variations. Any modification, equivalent substitution, improvement, etc. made within the spirit and principle of the present disclosure shall be included in the scope of the claims of the present disclosure.
Claims
1. An alarm aggregation method, characterized in that: include: Determining at least one target object that is alerting within a preset time window; Determine the target object nodes corresponding to each of the target objects in the preset object relationship graph, the preset object relationship graph is a knowledge graph constructed according to the association characteristics between the objects, the association characteristics include configuration relationship, call relationship and alarm association relationship, the configuration relationship includes containment relationship, deployment relationship, operation relationship and connection relationship, the preset object relationship graph has object nodes and edges connecting the object nodes, each edge corresponds to a preset weight value, wherein the weight value corresponding to each edge is related to the association characteristics between the two objects connected by the edge; Performing clustering processing using spectral clustering according to weight values of edges connecting the target object nodes to obtain at least one alarm cluster; The process of determining the weight value corresponding to the edge in the preset object relationship graph includes: For an edge between any two object nodes in the preset object relationship graph: using a Gaussian similarity function to calculate a first weight value of the edge; For the edge between the object nodes corresponding to the two objects having the alarm association relationship in the preset object relationship graph: add the first weight value of the edge to the third preset value to obtain the second weight value; if the two object nodes are in the configuration relationship, subtract the fourth preset value from the second weight value to obtain the target weight value; if the two object nodes are in the call relationship, add the fifth preset value to the second weight value to obtain the target weight value; if the two object nodes do not have the configuration relationship and the call relationship, determine the second weight value as the target weight value; The target weight value is determined to be the weight value corresponding to the edge.
2. The method according to claim 1, characterized in that The step of performing clustering processing by using spectral clustering according to the weight values of the edges connecting the target object nodes to obtain at least one alarm cluster includes: Taking each of the target object nodes as a sample set; According to the weight values of the edges connecting the target object nodes, an adjacency graph corresponding to the sample set is constructed using a full connection method to obtain an adjacency matrix and a degree matrix corresponding to the adjacency graph; According to the adjacency matrix and the degree matrix, a Laplace matrix is calculated and normalized; Performing eigenvalue decomposition on the normalized Laplace matrix to obtain a characteristic matrix; Each row vector in the feature matrix is clustered using a k-means clustering algorithm to obtain at least one alarm cluster.
3. The method according to claim 1, characterized in that After obtaining at least one alarm cluster, the method further includes: Outputting alarm information according to the alarm cluster.
4. The method according to claim 1, characterized in that: The process of determining the weight value corresponding to the edge in the preset object relationship graph includes: For an edge between any two object nodes in the preset object relationship graph: using a Gaussian similarity function to calculate an initial weight value of the edge; If the two object nodes are in the configuration relationship, the first preset value is subtracted from the initial weight value to obtain the target weight value; if the two object nodes are in the call relationship, the second preset value is added to the initial weight value to obtain the target weight value; The target weight value is determined to be the weight value corresponding to the edge.
5. The method according to claim 1, characterized in that The process of determining the alarm association relationship includes: Obtaining historical alarm records, wherein the historical alarm records include alarm records of at least one object; The historical alarm records are analyzed using an association rule analysis algorithm to determine whether the objects have the alarm association relationship.
6. The method according to claim 1, characterized in that The objects include application systems, clusters, middleware, physical devices, virtual devices, databases, storage and networks in the Internet system.
7. An alarm aggregation device, characterized in that: include: a target object determination unit, a target object node determination unit and an alarm cluster acquisition unit, The target object determination unit is used to determine at least one target object that is warning within a preset time window; The target object node determination unit is used to respectively determine the target object nodes corresponding to each of the target objects in a preset object relationship graph, the preset object relationship graph is a knowledge graph constructed according to the association characteristics between the objects, the association characteristics include configuration relationships, call relationships and alarm association relationships, the configuration relationships include inclusion relationships, deployment relationships, operation relationships and connection relationships, the preset object relationship graph has object nodes and edges connecting the object nodes, each edge corresponds to a preset weight value, wherein the weight value corresponding to each edge is related to the association characteristics between the two objects connected by the edge; The alarm cluster obtaining unit is used to perform clustering processing by using spectral clustering according to the weight values of the edges connecting the target object nodes to obtain at least one alarm cluster; The alarm aggregation device further comprises: a second weight value determining unit, The second weight value determination unit is used to calculate the first weight value of the edge between any two object nodes in the preset object relationship graph using a Gaussian similarity function; for the edge between the object nodes corresponding to two objects having the alarm association relationship in the preset object relationship graph: add the first weight value of the edge to the third preset value to obtain a second weight value; if the two object nodes are in the configuration relationship, subtract the fourth preset value from the second weight value to obtain a target weight value; if the two object nodes are in the calling relationship, add the fifth preset value to the second weight value to obtain the target weight value; if the two object nodes do not have the configuration relationship and the calling relationship, determine the second weight value to be the target weight value; and determine the target weight value to be the weight value corresponding to the edge.
8. A computer-readable storage medium having a program stored thereon, characterized in that: When the program is executed by a processor, the alarm aggregation method according to any one of claims 1 to 6 is implemented.
9. An electronic device, comprising at least one processor, and at least one memory and bus connected to the processor; wherein: The processor and the memory communicate with each other via the bus; The processor is used to call the program instructions in the memory to execute the alarm aggregation method according to any one of claims 1 to 6.
Citation Information
Patent Citations
System alarm clustering analysis method based on co-occurrence graph
CN110399261A
Network security anomaly detection algorithm and detection system based on clustering graph neural network
CN112165496A
