CN113824583B - Root element positioning method and device, computer equipment and storage medium - Google Patents

Root element positioning method and device, computer equipment and storage medium Download PDF

Info

Publication number
CN113824583B
CN113824583B CN202110198580.XA CN202110198580A CN113824583B CN 113824583 B CN113824583 B CN 113824583B CN 202110198580 A CN202110198580 A CN 202110198580A CN 113824583 B CN113824583 B CN 113824583B
Authority
CN
China
Prior art keywords
root cause
elements
search space
node
determining
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN202110198580.XA
Other languages
Chinese (zh)
Other versions
CN113824583A (en
Inventor
张静
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Jingdong Technology Holding Co Ltd
Original Assignee
Jingdong Technology Holding Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Jingdong Technology Holding Co Ltd filed Critical Jingdong Technology Holding Co Ltd
Priority to CN202110198580.XA priority Critical patent/CN113824583B/en
Publication of CN113824583A publication Critical patent/CN113824583A/en
Application granted granted Critical
Publication of CN113824583B publication Critical patent/CN113824583B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/06—Management of faults, events, alarms or notifications
    • H04L41/0631—Management of faults, events, alarms or notifications using root cause analysis; using analysis of correlation between notifications, alarms or events based on decision criteria, e.g. hierarchy, tree or time analysis
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/06—Management of faults, events, alarms or notifications
    • H04L41/0677—Localisation of faults

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Mobile Radio Communication Systems (AREA)
  • Telephonic Communication Services (AREA)

Abstract

The application provides a method, a device, a computer device and a storage medium for positioning root cause elements, wherein the method comprises the steps of determining a first search space from call link spaces corresponding to services, determining a first root cause element set of the first search space, wherein the first root cause element set comprises the following components: at least one first cause element, the first cause element being an element of the at least one element that is associated with a failure of the service; determining a plurality of candidate root cause elements from the second search space, wherein attribute dimensions between the elements of the first search space and the elements of the second search space are different; at least one second root cause element is selected from the at least one candidate root cause element based on the first root cause element, the second root cause element being an element associated with a failure of the service within the second search space. The application can effectively improve the positioning accuracy and efficiency of the fault root cause of the service call chain, improve the positioning effect of the fault root cause and realize the root cause positioning of the whole link.

Description

Root element positioning method and device, computer equipment and storage medium
Technical Field
The present application relates to the field of internet technologies, and in particular, to a method and apparatus for locating root cause elements, a computer device, and a storage medium.
Background
In the operation and maintenance of internet service, when a certain total index is abnormal, the abnormality of fine granularity combination of which cross dimension needs to be quickly positioned, so that quick damage stopping is achieved. In the related art, the rule of thumb depends on operation and maintenance, and the root cause elements of faults are checked one by one along the call chain of the service.
In this way, root elements that are not on the same call chain cannot be effectively examined.
Disclosure of Invention
The present application aims to solve at least one of the technical problems in the related art to some extent.
Therefore, the application aims to provide a root cause element positioning method, a root cause element positioning device, computer equipment and a storage medium, which can effectively improve the positioning accuracy and the positioning efficiency of the fault root cause of a service call chain, improve the positioning effect of the fault root cause and realize the root cause positioning of a full link.
In order to achieve the above object, a method for locating a root cause element according to an embodiment of a first aspect of the present application includes: determining a first search space from call link spaces corresponding to services, wherein the first search space comprises: a plurality of nodes, the nodes comprising: at least one element, different elements have the same attribute dimension; determining a first set of cause elements for the first search space, the first set of cause elements comprising: at least one first root cause element, the first root cause element being an element of the at least one element that is related to a failure of the service; determining a plurality of candidate root cause elements from a second search space, wherein the first search space and the second search space jointly form the calling link space, and attribute dimensions between the elements of the first search space and the elements of the second search space are different; selecting at least one second root cause element from the at least one candidate root cause element according to the first root cause element, wherein the second root cause element is an element related to the fault of the service in the second search space; wherein the at least one first root cause element and the at least one second root cause element are jointly used as located root cause elements.
According to the method for positioning the root cause element, which is provided by the embodiment of the first aspect of the application, a first search space is determined from the call link space corresponding to the service, and the first search space comprises: a plurality of nodes, the nodes comprising: at least one element, different elements have the same attribute dimension, a first root cause element set of a first search space is determined, and the first root cause element set comprises: the method comprises the steps of determining a plurality of candidate root cause elements from a second search space, wherein the first search space and the second search space jointly form a calling link space, attribute dimensions between the elements of the first search space and the elements of the second search space are different, selecting at least one second root cause element from the at least one candidate root cause elements according to the first root cause element, and the second root cause element is an element related to the failure of the service in the second search space, wherein the at least one first root cause element and the at least one second root cause element are jointly used as the located root cause element, so that the positioning accuracy and the positioning efficiency of the failure root cause of a service calling chain can be effectively improved, the positioning effect of the failure root cause is improved, and the root cause positioning of the whole link is realized.
In order to achieve the above object, a positioning device for a root cause element according to an embodiment of a second aspect of the present application includes: the first determining module is configured to determine a first search space from call link spaces corresponding to services, where the first search space includes: a plurality of nodes, the nodes comprising: at least one element, different elements have the same attribute dimension; a second determining module, configured to determine a first root cause element set of the first search space, where the first root cause element set includes: at least one first root cause element, the first root cause element being an element of the at least one element that is related to a failure of the service; a third determining module, configured to determine a plurality of candidate root cause elements from a second search space, where the first search space and the second search space together form the call link space, and attribute dimensions between the elements of the first search space and the elements of the second search space are different; a selecting module, configured to select at least one second root cause element from the at least one candidate root cause element according to the first root cause element, where the second root cause element is an element related to a fault of the service in the second search space; wherein the at least one first root cause element and the at least one second root cause element are jointly used as located root cause elements.
According to the positioning device of the root cause element provided by the embodiment of the second aspect of the application, a first search space is determined from the call link space corresponding to the service, and the first search space comprises: a plurality of nodes, the nodes comprising: at least one element, different elements have the same attribute dimension, a first root cause element set of a first search space is determined, and the first root cause element set comprises: the method comprises the steps of determining a plurality of candidate root cause elements from a second search space, wherein the first search space and the second search space jointly form a calling link space, attribute dimensions between the elements of the first search space and the elements of the second search space are different, selecting at least one second root cause element from the at least one candidate root cause elements according to the first root cause element, and the second root cause element is an element related to the failure of the service in the second search space, wherein the at least one first root cause element and the at least one second root cause element are jointly used as the located root cause element, so that the positioning accuracy and the positioning efficiency of the failure root cause of a service calling chain can be effectively improved, the positioning effect of the failure root cause is improved, and the root cause positioning of the whole link is realized.
An embodiment of the third aspect of the present application provides a computer device, including a memory, a processor, and a computer program stored in the memory and capable of running on the processor, where the processor implements the method for locating root cause elements according to the embodiment of the first aspect of the present application when the processor executes the program.
An embodiment of a fourth aspect of the present application proposes a non-transitory computer readable storage medium, on which a computer program is stored, which program, when being executed by a processor, implements a method for locating root cause elements as proposed by an embodiment of the first aspect of the present application.
An embodiment of a fifth aspect of the present application proposes a computer program product which, when executed by an instruction processor in the computer program product, performs a method for locating a root cause element as proposed by an embodiment of the first aspect of the present application.
Additional aspects and advantages of the application will be set forth in part in the description which follows and, in part, will be obvious from the description, or may be learned by practice of the application.
Drawings
The foregoing and/or additional aspects and advantages of the application will become apparent and readily appreciated from the following description of the embodiments, taken in conjunction with the accompanying drawings, in which:
FIG. 1 is a flow chart illustrating a method for locating root cause elements according to an embodiment of the present application;
FIG. 2 is a diagram of call link space in an embodiment of the application;
FIG. 3 is a flow chart illustrating a method for locating root cause elements according to another embodiment of the present application;
FIG. 4 is a flow chart illustrating a method for locating root cause elements according to another embodiment of the present application;
FIG. 5 is a schematic illustration of an application of an embodiment of the present application;
FIG. 6 is a schematic structural diagram of a positioning device for root cause elements according to an embodiment of the present application;
FIG. 7 is a schematic structural diagram of a positioning device for root cause elements according to another embodiment of the present application;
FIG. 8 illustrates a block diagram of an exemplary computer device suitable for use in implementing embodiments of the present application.
Detailed Description
Embodiments of the present application are described in detail below, examples of which are illustrated in the accompanying drawings, wherein like or similar reference numerals refer to like or similar elements or elements having like or similar functions throughout. The embodiments described below by referring to the drawings are illustrative only and are not to be construed as limiting the application. On the contrary, the embodiments of the application include all alternatives, modifications and equivalents as may be included within the spirit and scope of the appended claims.
Fig. 1 is a flow chart of a method for locating root cause elements according to an embodiment of the application.
It should be noted that, the execution body of the root cause element positioning method in this embodiment is a root cause element positioning device, and the device may be implemented in a software and/or hardware manner, and the device may be configured in an electronic device, where the electronic device may include, but is not limited to, a terminal, a server, and the like.
As shown in fig. 1, the method for locating the root cause element includes:
S101: determining a first search space from call link spaces corresponding to the services, wherein the first search space comprises: a plurality of nodes, the nodes comprising: at least one element, different elements have the same attribute dimension.
The service may be, for example, an internet service, a service such as some hardware and/or software service for supporting internet functions, specifically, for example, an application service, an applet service, a search engine service, etc., which is not limited.
It can be understood that in the internet service, the internet service is generally provided externally based on the service modes of information receiving, interface calling, algorithm processing, information identifying and the like, so that when the internet service is provided externally by adopting each service mode, one or more calling links exist.
In the embodiment of the application, the first search space can be determined from the call link space corresponding to the service, and the first search space comprises: a plurality of nodes, the nodes comprising: at least one element, different elements have the same attribute dimension.
The first search space may be considered as a local space in the call link space of the entirety of the service, the first search space including: a plurality of nodes, the nodes comprising: at least one element, different elements have the same attribute dimension, wherein a plurality of nodes in the first search space can be partial nodes in the whole calling link space, each node can correspond to a service mode, the elements included in the nodes can correspond to the attributes in the service mode, one element can correspond to one or more attributes, the attribute dimension refers to the number of the attributes, for example, if one element corresponds to two attributes, the attribute dimension can be two-dimensional, and if one element corresponds to three attributes, the attribute dimension can be three-dimensional.
As shown in fig. 2, fig. 2 is a schematic diagram of a call link space in an embodiment of the present application, where the call link space may include four layers of nodes (only examples, but not limited thereto), attribute dimensions of the nodes at the same layer are the same, each node correspondingly includes a plurality of elements, a node may, for example, correspond to a service mode, elements included in the node may correspond to attributes (attributes, for example, "province" and "city") in the service mode, and the node may be regarded as a collection of elements, and elements that may have the same or similar attribute values are divided into a class of nodes.
Referring also to fig. 2, that is, the call link space in the present application may have the following features: elements with the same attribute dimension (i.e., attribute number) are called as same-layer elements, the attribute number represents the layer number, the same-layer elements are divided into different sets according to attribute values contained in the same-layer elements, the divided sets are regarded as a node, and each set of higher-layer elements (including elements with more attribute number) can be combined by elements in the upper-layer nodes.
While the first search space in the present application may be a partial call link space among the call link spaces shown in fig. 2, the first search space includes: a plurality of nodes, the nodes comprising: at least one element, different elements having the same attribute dimension, for example, elements contained in the first search space may be elements of the same layer in the call link space shown in fig. 2.
In the embodiment of the application, the key performance indicators (Key Performance Indicator, KPI) of the service are considered to be generally influenced by a plurality of dimension attributes, and the number of elements contained in each attribute dimension is various, so that the search space for locating the root cause is huge, and therefore, the first search space is determined from the call link space corresponding to the service, thereby effectively reducing the search space and being capable of assisting in improving the locating efficiency of the fault root cause of the service call chain.
Optionally, in some embodiments, as shown in fig. 3, fig. 3 is a flowchart of a method for locating a root cause element according to another embodiment of the present application, determining a first search space from call link spaces corresponding to services includes:
s301: and determining multiple attribute dimensions corresponding to the service.
The attribute dimension refers to the number of attributes, for example, if one element corresponds to two attributes, the attribute dimension may be two-dimensional, and if one element corresponds to three attributes, the attribute dimension may be three-dimensional, and accordingly, the attribute dimensions may be two-dimensional, three-dimensional, and four-dimensional, which is not limited.
S302: the target attribute dimension is determined from among a plurality of attribute dimensions.
That is, in this embodiment, the number of attribute values that actually have a larger influence on the abnormal change of the key performance indicator KPI of the service population is considered with respect to the total number of attribute values, so in this embodiment, each attribute dimension may be first screened to determine the target attribute dimension, where the element to which the target attribute dimension belongs may be, for example, an element that has a relatively larger influence on the abnormal change of the key performance indicator KPI of the service population.
Optionally, in some embodiments, a plurality of elements respectively associated with the plurality of attribute dimensions may be determined; determining a plurality of real value distribution indexes and a plurality of predicted value distribution indexes which respectively correspond to the elements; determining a plurality of distance index values between a plurality of real value distribution indexes and a plurality of corresponding predicted value distribution indexes; selecting a target distance index value from the plurality of distance index values, wherein the target distance index value indicates the influence degree of a corresponding target element on the key performance index KPI of the service to be greater than the influence degree of other elements on the key performance index KPI of the service, and the target element and the other elements jointly form a plurality of associated elements; and taking the attribute dimension of the target element as the target attribute dimension, so that the influence degree of each element on the key performance index KPI of the overall service and the contribution degree of each element on the distance between the true value distribution index and the predicted value distribution index are both included in the selected consideration index of the target attribute dimension, the first search space is determined in an auxiliary mode, and the accuracy of the positioning range of the root cause element can be effectively improved.
The influence degree of the key performance index KPI of each element service can be determined, a plurality of influence degrees are ordered, and the element with the largest influence degree is selected as the target element.
For example, assuming that the plurality of attribute dimensions may be two-dimensional, three-dimensional, and four-dimensional, the plurality of elements respectively associated with the plurality of attribute dimensions may be an element whose attribute dimension is two-dimensional, an element whose attribute dimension is three-dimensional, and an element whose attribute dimension is four-dimensional, and then the real value distribution index and the plurality of predicted value distribution indexes corresponding to the elements of each dimension are respectively determined.
The actual value corresponding to the element may be, for example, an actual output value of a service mode corresponding to the element in the internet service, for example, assuming that the overall internet service is a, the actual output value of the service mode corresponding to each element mapped by the internet service a may be A1, A2, A3, etc., the output value of the element on each call link cooperates with the overall internet service a, the predicted value corresponding to the element may be, for example, a predicted output value of the service mode corresponding to the element in the internet service, for example, assuming that the overall internet service is a, the predicted output value of the service mode corresponding to each element mapped by the internet service a may be A1', A2', A3', etc., the output value of the element on each call link cooperates with the overall internet service a, which is generally limited to an actual environment, and there may be a certain deviation between the predicted value and the actual value.
The above-mentioned real values can be directly obtained by identification according to the output condition of each element, and the predicted values of each element at the current time point can be obtained by adopting a second-order exponential smoothing method in the related art, which is not limited.
The actual value distribution index may be a distribution for evaluating an actual output value, and the predicted value distribution index may be a distribution for evaluating a predicted output value, without limitation.
After determining the plurality of real value distribution indexes and the plurality of predicted value distribution indexes corresponding to the plurality of elements respectively, the degree of similarity between the corresponding real value distribution index and the corresponding predicted value distribution index can be determined for each element, for example, a plurality of distance index values between the plurality of real value distribution indexes and the corresponding predicted value distribution indexes can be determined; and selecting a target distance index value from the plurality of distance index values, so that the target distance index value indicates that the influence degree of the corresponding target element on the key performance index KPI of the service is greater than the influence degree of other elements on the key performance index KPI of the service, and the target element and the other elements jointly form a plurality of associated elements.
In determining the plurality of distance index values between the plurality of real value distribution indexes and the corresponding plurality of predicted value distribution indexes, the above may specifically be determining a plurality of Jensen shannon (Jensen-Shannon Divergence, JS divergence) divergences between the plurality of real value distribution indexes and the corresponding plurality of predicted value distribution indexes, and using the plurality of Jensen shannon JS divergences as the corresponding plurality of distance index values.
For example, for the actual value distribution index and the predicted value distribution index corresponding to each element, the JS divergence therebetween may be calculated, and the JS divergence may be used as the distance index value corresponding to the element.
In the embodiment of the application, the JS dispersion is adopted as the measurement value of the distance index, so that the error caused by assumption can be effectively reduced, the accuracy of determining the target distance index value can be effectively improved, and the most suitable root element positioning range can be assisted to be determined while assisting in reducing the search space.
S303: and determining the node corresponding to the element associated with the target attribute dimension from the calling link space.
S304: a first search space is formed from at least one node.
Assuming that the determined target attribute dimension is three-dimensional, a node to which an element whose attribute dimension is three-dimensional belongs may be used as a node corresponding to an element associated with the target attribute dimension, and then the first search space is formed according to a local call link between the nodes.
S102: determining a first set of cause elements for the first search space, the first set of cause elements comprising: at least one first root cause element, the first root cause element being an element of the at least one element that is associated with a failure of the service.
The first search space is determined from the call link space corresponding to the service, and the first search space comprises: a plurality of nodes, the nodes comprising: after at least one element and different elements have the same attribute dimension, a first root cause element set of the first search space can be determined, that is, a positioning search of the root cause element can be performed in the first search space, so that an element which possibly affects the service fault is determined to be used as the first root cause element, and elements on a local call link related to the first root cause element can be used as related elements, so that the method is not limited.
S103: and determining a plurality of candidate root cause elements from the second search space, wherein the first search space and the second search space jointly form a calling link space, and attribute dimensions between the elements of the first search space and the elements of the second search space are different.
After determining the first search space from the call link space corresponding to the service and determining the first root cause element set from the first search space, the plurality of candidate root cause elements may be determined from the second search space.
The whole calling link space of the service can be called a second search space except the first search space, the second search space correspondingly comprises a plurality of nodes, each node comprises one or more corresponding elements, the organization and architecture mode of the second search space can be the same as that of the first search space, and attribute dimensions between the elements of the first search space and the elements of the second search space are different.
For example, assuming that the attribute dimension of an element in the first search space is three-dimensional, the attribute dimension of an element in the second search space may be two-dimensional or four-dimensional, or any other possible dimension other than three-dimensional, without limitation.
In an actual application scenario, key performance indicators (Key Performance Indicator, KPIs) of a service are generally affected by a lot of dimension attributes, the KPIs have additivity, the KPIs corresponding to combinations of elements of different attributes are related to each other, for example, the sum of the KPIs is equal to the sum of KPI values corresponding to all elements under attribute a, and the attribute values among different attributes also affect each other, for example, when the KPIs of an element with an attribute value of "beijing" corresponding to a "province" attribute is abnormal, the KPIs with an attribute value of "beijing mobile" and an attribute value of "beijing" are abnormal, so that the KPIs with an attribute value of "mobile" and an attribute value of "beijing" connected "are abnormal, and therefore, it can be seen that a relatively complex relationship exists between the elements including different attributes.
In the embodiment of the present application, a set of elements with the same attribute dimension may be used as a search space (first search space), then a Monte Carlo TREE SEARCH, MCTS algorithm is applied in the first search space to obtain a set of most probable root cause elements (a first root cause element set) with the attribute dimension, and then a node Potential Score of a node to which the root cause element (a second root cause element) with the different attribute dimension belongs is compared, so as to obtain a target root cause element set, where the target root cause element combination is a set including the first root cause element and the second root cause element.
S104: at least one second root cause element is selected from the at least one candidate root cause element based on the first root cause element, the second root cause element being an element associated with a failure of the service within the second search space.
That is, after the first factor element set is located in the first search space, searching may be performed in a second search space with different attribute dimensions to locate at least one candidate factor element, and then, the second factor element may be screened out according to the node Potential Score of the node to which each candidate factor element belongs.
In the embodiment of the present application, the method of locating the first element set in the first search space may be specifically referred to as the following embodiment, and the method of locating the plurality of candidate elements in the second search space may be referred to as the method of locating the elements in the first search space.
In this embodiment, a first search space is determined from call link spaces corresponding to services, where the first search space includes: a plurality of nodes, the nodes comprising: at least one element, different elements have the same attribute dimension, a first root cause element set of a first search space is determined, and the first root cause element set comprises: the method comprises the steps of determining a plurality of candidate root cause elements from a second search space, wherein the first search space and the second search space jointly form a calling link space, attribute dimensions between the elements of the first search space and the elements of the second search space are different, selecting at least one second root cause element from the at least one candidate root cause elements according to the first root cause element, and the second root cause element is an element related to the failure of the service in the second search space, wherein the at least one first root cause element and the at least one second root cause element are jointly used as the located root cause element, so that the positioning accuracy and the positioning efficiency of the failure root cause of a service calling chain can be effectively improved, the positioning effect of the failure root cause is improved, and the root cause positioning of the whole link is realized.
Fig. 4 is a flow chart of a method for locating root cause elements according to another embodiment of the present application.
As shown in fig. 4, the method for locating the root cause element includes:
S401: determining a first search space from call link spaces corresponding to the services, wherein the first search space comprises: a plurality of nodes, the nodes comprising: at least one element, different elements have the same attribute dimension.
The description of S401 may be specifically referred to the above embodiments, and will not be repeated here.
S402: and determining a plurality of node potential scores and a plurality of calling times corresponding to the plurality of nodes respectively, wherein the calling times are the times that a calling link of the service passes through the corresponding node.
In this embodiment, a monte carlo tree search MCTS algorithm is first applied in the first search space, to obtain a set of most probable root elements in this attribute dimension (the first set of root elements).
For example, in the root cause element positioning method, a monte carlo tree search MCTS algorithm may be applied to a set formed by elements with the same dimension attribute, for example, a first search space may be marked as C, a state s (may be referred to as a corresponding node state) of each node in the first search space C corresponds to an ordered set G(s) of elements under the node in the first search space, in this embodiment, a Potential Score algorithm function may be used as an evaluation function of the monte carlo tree search MCTS algorithm, for each node, the node state of the node is scored by using the Potential Score algorithm function to obtain a node Potential Score Q(s) of the node, and the number of times of call of the node is determined, where the number of times of call of the node is N(s) of times that a call link of a service passes through the corresponding node, and a value of Q(s) is generally equal to a value of a maximum node Potential Score in all sub-nodes of the current node, where a formula is as follows:
Wherein Q(s) represents the node potential score of the current node, a represents the element of the child node, Q (s, a) represents the node potential score of the child node, a(s) represents the set of node states, ps (G (s)) represents the potential score contributed to the node to which each element in the ordered set of elements G(s) under the node belongs.
S403: according to the potential scores and the calling times of the nodes, constructing a plurality of search trees respectively corresponding to the nodes by combining a Monte Carlo tree search MCTS algorithm, wherein the search trees comprise: the root node and the prediction node based on the root node are obtained, and the root node is a node corresponding to the search tree.
After determining the potential scores and the call times of the nodes corresponding to the nodes respectively, the MCTS algorithm may be combined to construct a plurality of search trees corresponding to the nodes respectively.
For example, each node may be used as a root node, at least one iteration is performed through a MCTS search, and according to the principle of the MCTS search, after each iteration, one node is added to the search tree.
If and only if the iteration times reach a set maximum value M, or no new node which is not opened up exists in the whole search tree, the iteration end can be determined, according to the iteration process of the MCTS, the node potential score Q value of the root node is always equal to the maximum Q value in the whole search tree, after the iteration is completed, the node which is equal to the Q value of the root node can be searched, and the element set corresponding to the node is the root cause element set which is most likely to be the root cause of the fault in the first search space.
S404: and determining a target node from the predicted nodes included in the plurality of search trees, wherein the node potential score of the target node is the same as the node potential score of the root node to which the target node belongs, and the layer number to which the target node belongs is the lowest layer number among the plurality of search trees.
For example, assuming that each node corresponds to one search tree, the prediction node with the same node potential score as the root node may be determined from each search tree, the prediction node is determined to correspond to the number of layers in the search tree, and the prediction node with the lowest number of layers is used as the target node.
The method comprises the steps of determining a plurality of node potential scores and a plurality of calling times corresponding to a plurality of nodes respectively, constructing a plurality of search trees corresponding to the plurality of nodes respectively according to the plurality of node potential scores and the plurality of calling times by combining a Monte Carlo tree search MCTS algorithm, determining a target node from prediction nodes included in the plurality of search trees, wherein the node potential scores of the target node are the same as the node potential scores of root nodes to which the target node belongs, the layer number of the target node is the lowest layer number among the plurality of search trees, and accurately and rapidly identifying a root cause element set which is most likely to be a fault root cause in a first search space, so that the positioning effect of the root cause element is improved.
S405: a first set of root cause elements is formed from the elements contained by the target node.
In the embodiment of the application, in order to enable the positioning method of the root cause element to effectively adapt to the actual application scene, if the number of the target nodes is multiple, the number of the elements corresponding to the target nodes is determined, a first root cause element set is formed according to the elements contained in the first target node, and the first target node is the target node with the minimum number of elements among the target nodes.
That is, in the embodiment of the present application, the root cause is corrected by using an oldham razor algorithm, which is illustrated in the example of the oldham razor algorithm: when the two target nodes have the same potential scores, the target node with the least element number is determined, so that a first root cause element set is formed according to the elements contained in the first target node, and the accuracy of root cause element determination is further improved.
S406: and determining a plurality of candidate root cause elements from the second search space, wherein the first search space and the second search space jointly form a calling link space, and attribute dimensions between the elements of the first search space and the elements of the second search space are different.
The method for locating multiple candidate root cause elements in the second search space may refer to a method for locating the first root cause element in the first search space, that is, the second search space in the embodiment of the present application may be one or more, the attribute dimension corresponding to the second search space and the attribute dimension of the first search space are different, if the number of the second search spaces is multiple, the attribute dimensions between the different second search spaces may be different, in this embodiment, the method may implement that the MCTS algorithm is searched by using the monte carlo tree for the search space with the same attribute dimension to determine multiple root cause elements as candidate root cause elements, and then, the first root cause element in the first search space is used to assist in locating the second root cause element from among the multiple candidate root cause elements.
Because the first root cause element is an element with a larger influence degree on the key performance indicator KPI of the service, the second root cause element determined by referring to the first root cause element is also generally a more critical root cause element, and the first root cause element in the first search space is adopted in the embodiment to assist in locating the second root cause element from a plurality of candidate root cause elements, so that the locating accuracy of the root cause element with the same attribute dimension can be improved, and the locating accuracy of the root cause element with different attribute dimensions can be effectively improved.
S407: a first node potential score for the node to which the first root cause element belongs is determined.
For example, the Potential Score of the first node of the node to which the first root cause element belongs may be calculated using the Potential Score algorithm function described above.
S408: and determining a second node potential score of the node to which the candidate root cause element belongs.
For example, the Potential Score of the second node of the node to which the candidate root cause element belongs may be calculated using the Potential Score algorithm function described above.
S409: the first root cause element, the candidate root cause element, the first node potential score, and the second node potential score are input into a pre-trained reinforcement learning model to obtain at least one second root cause element output by the reinforcement learning model.
Wherein the second root element is an element related to a failure of the service within the second search space, wherein the at least one first root element and the at least one second root element are jointly used as located root elements.
After determining the first node potential score of the node to which the first root element belongs and determining the second node potential score of the node to which the candidate root element belongs, the first root element, the candidate root element, the first node potential score, and the second node potential score may be input into a pre-trained reinforcement learning model to obtain at least one second root element output by the reinforcement learning model.
The reinforcement learning model may be obtained by training based on massive training data in advance, and the reinforcement learning model may be based on learning a root cause element with the greatest degree of potential scoring correlation with the first node of the first root cause element, so that the root cause element with the greatest degree of potential scoring correlation is used as the second root cause element, which is not limited.
According to the method for positioning the root cause element, provided by the embodiment of the application, the target dimension attribute is determined by adopting the adtributor algorithm, then the adtributor algorithm is combined with the Monte Carlo tree search MCTS algorithm, and the nodes in the first search space determined based on the target dimension attribute are used as initial searches of the Monte Carlo tree search MCTS algorithm, so that the timeliness of positioning the root cause element is improved to 1.4 seconds, and the root cause positioning range is more accurate.
In the embodiment of the application, the identification of the first root cause element and the second root cause element contained in the positioning result of the Monte Carlo tree search MCTS algorithm can be matched with the alarm template corresponding to the application associated with the first root cause element in real time by combining the operation and maintenance expert experience library and the operation and maintenance knowledge graph so as to verify the accuracy of the positioning result of the root cause element, and the identification of the high-frequency alarm mail association application can be performed by combining a hot spot analysis algorithm in the related technology, so that the full-link root cause positioning of the multidimensional search is realized.
In this embodiment, the positioning accuracy and the positioning efficiency of the fault root cause of the service call chain can be effectively improved, the positioning effect of the fault root cause is improved, and the root cause positioning of the full link is realized. And the positioning method of the root cause element can be effectively adapted to the actual application scene. In addition, since the first root cause element is an element with a larger influence degree on the key performance indicator KPI of the service, the second root cause element determined by referring to the first root cause element is also generally a more critical root cause element, so that the first root cause element in the first search space is adopted to assist in locating the second root cause element from a plurality of candidate root cause elements in the embodiment, the locating accuracy of the root cause element with the same attribute dimension can be improved, and the locating accuracy of the root cause element with different attribute dimensions can be effectively improved.
Fig. 5 is a schematic diagram of an application of the embodiment of the present application, as shown in fig. 5.
Fig. 6 is a schematic structural diagram of a positioning device for root cause elements according to an embodiment of the present application.
As shown in fig. 6, the positioning device 60 of the root cause element includes:
the first determining module 601 is configured to determine a first search space from call link spaces corresponding to services, where the first search space includes: a plurality of nodes, the nodes comprising: at least one element, different elements have the same attribute dimension;
the second determining module 602 is configured to determine a first root cause element set of the first search space, where the first root cause element set includes: at least one first cause element, the first cause element being an element of the at least one element that is associated with a failure of the service;
A third determining module 603, configured to determine a plurality of candidate root cause elements from a second search space, where the first search space and the second search space together form a call link space, and attribute dimensions between the elements of the first search space and the elements of the second search space are different;
A selection module 604, configured to select at least one second root element from at least one candidate root element according to the first root element, where the second root element is an element related to a failure of the service in the second search space;
wherein at least one first root element and at least one second root element are jointly used as located root elements.
In some embodiments of the present application, the second determining module 602 is specifically configured to:
determining a plurality of node potential scores and a plurality of calling times corresponding to the plurality of nodes respectively, wherein the calling times are times when a calling link of the service passes through the corresponding node;
according to the potential scores and the calling times of the nodes, constructing a plurality of search trees respectively corresponding to the nodes by combining a Monte Carlo tree search MCTS algorithm, wherein the search trees comprise: the root node is a node corresponding to the search tree;
Determining a target node from the predicted nodes included in the plurality of search trees, wherein the node potential score of the target node is the same as the node potential score of the root node to which the target node belongs, and the layer number to which the target node belongs is the lowest layer number among the plurality of search trees;
a first set of root cause elements is formed from the elements contained by the target node.
In some embodiments of the present application, the number of target nodes is at least one, and the second determining module 602 is specifically configured to:
If the number of the target nodes is a plurality of, determining the element number corresponding to the plurality of target nodes respectively;
the first root cause element set is formed according to elements contained in a first target node, and the first target node is the target node with the minimum element number among a plurality of target nodes.
In some embodiments of the present application, the selecting module 604 is specifically configured to:
determining a first node potential score of a node to which the first root cause element belongs;
Determining a second node potential score of the node to which the candidate root cause element belongs;
The first root cause element, the candidate root cause element, the first node potential score, and the second node potential score are input into a pre-trained reinforcement learning model to obtain at least one second root cause element output by the reinforcement learning model.
In some embodiments of the present application, as shown in fig. 7, the first determining module 601 includes:
A first determining submodule 6011, configured to determine multiple attribute dimensions corresponding to the service;
A second determining submodule 6012, configured to determine a target attribute dimension from a plurality of attribute dimensions;
A third determining submodule 6013, configured to determine, from among the call link spaces, a node corresponding to the element associated with the target attribute dimension;
a generation submodule 6014 is used for forming a first search space according to at least one node.
In some embodiments of the present application, the second determining submodule 6012 is specifically configured to:
determining a plurality of elements respectively associated with the attribute dimensions;
Determining a plurality of real value distribution indexes and a plurality of predicted value distribution indexes which respectively correspond to the elements;
Determining a plurality of distance index values between a plurality of real value distribution indexes and a plurality of corresponding predicted value distribution indexes;
selecting a target distance index value from the plurality of distance index values, wherein the target distance index value indicates the influence degree of a corresponding target element on the key performance index KPI of the service to be greater than the influence degree of other elements on the key performance index KPI of the service, and the target element and the other elements jointly form a plurality of associated elements;
And taking the attribute dimension of the target element as the target attribute dimension.
In some embodiments of the present application, the second determining submodule 6012 is specifically configured to:
And determining a plurality of jensen shannon JS divergences between the plurality of real value distribution indexes and the corresponding plurality of predicted value distribution indexes, and taking the plurality of jensen shannon JS divergences as a corresponding plurality of distance index values.
Corresponding to the method for positioning the root element provided by the embodiments of fig. 1 to 5, the present application further provides a device for positioning the root element, and since the device for positioning the root element provided by the embodiments of the present application corresponds to the method for positioning the root element provided by the embodiments of fig. 1 to 5, the implementation of the method for positioning the root element is also applicable to the device for positioning the root element provided by the embodiments of the present application, which is not described in detail in the embodiments of the present application.
In this embodiment, a first search space is determined from call link spaces corresponding to services, where the first search space includes: a plurality of nodes, the nodes comprising: at least one element, different elements have the same attribute dimension, a first root cause element set of a first search space is determined, and the first root cause element set comprises: the method comprises the steps of determining a plurality of candidate root cause elements from a second search space, wherein the first search space and the second search space jointly form a calling link space, attribute dimensions between the elements of the first search space and the elements of the second search space are different, selecting at least one second root cause element from the at least one candidate root cause elements according to the first root cause element, and the second root cause element is an element related to the failure of the service in the second search space, wherein the at least one first root cause element and the at least one second root cause element are jointly used as the located root cause element, so that the positioning accuracy and the positioning efficiency of the failure root cause of a service calling chain can be effectively improved, the positioning effect of the failure root cause is improved, and the root cause positioning of the whole link is realized.
In order to implement the above embodiment, the present application further proposes a computer device including: the method for locating the root cause element according to the embodiment of the application is realized when the processor executes the program.
In order to achieve the above-mentioned embodiments, the present application also proposes a non-transitory computer-readable storage medium having stored thereon a computer program which, when executed by a processor, implements a localization method of a root cause element as proposed in the foregoing embodiments of the present application.
In order to implement the above embodiments, the present application also proposes a computer program product which, when executed by an instruction processor in the computer program product, performs the positioning method of root cause elements as proposed in the previous embodiments of the present application.
FIG. 8 illustrates a block diagram of an exemplary computer device suitable for use in implementing embodiments of the present application. The computer device 12 shown in fig. 8 is merely an example and should not be construed as limiting the functionality and scope of use of embodiments of the present application.
As shown in FIG. 8, the computer device 12 is in the form of a general purpose computing device. Components of computer device 12 may include, but are not limited to: one or more processors or processing units 16, a system memory 28, a bus 18 that connects the various system components, including the system memory 28 and the processing units 16.
Bus 18 represents one or more of several types of bus structures, including a memory bus or memory controller, a peripheral bus, an accelerated graphics port, a processor, and a local bus using any of a variety of bus architectures. By way of example, and not limitation, such architectures include industry Standard architecture (Industry Standard Architecture; hereinafter ISA) bus, micro channel architecture (Micro Channel Architecture; hereinafter MAC) bus, enhanced ISA bus, video electronics standards Association (Video Electronics Standards Association; hereinafter VESA) local bus, and peripheral component interconnect (PERIPHERAL COMPONENT INTERCONNECTION; hereinafter PCI) bus.
Computer device 12 typically includes a variety of computer system readable media. Such media can be any available media that is accessible by computer device 12 and includes both volatile and nonvolatile media, removable and non-removable media.
Memory 28 may include computer system readable media in the form of volatile memory, such as random access memory (Random Access Memory; hereinafter: RAM) 30 and/or cache memory 32. The computer device 12 may further include other removable/non-removable, volatile/nonvolatile computer system storage media. By way of example only, storage system 34 may be used to read from or write to non-removable, nonvolatile magnetic media (not shown in FIG. 8, commonly referred to as a "hard disk drive"). Although not shown in fig. 8, a disk drive for reading from and writing to a removable nonvolatile magnetic disk (e.g., a "floppy disk"), and an optical disk drive for reading from or writing to a removable nonvolatile optical disk (e.g., a compact disk read only memory (Compact Disc Read Only Memory; hereinafter CD-ROM), digital versatile read only optical disk (Digital Video Disc Read Only Memory; hereinafter DVD-ROM), or other optical media) may be provided. In such cases, each drive may be coupled to bus 18 through one or more data medium interfaces. Memory 28 may include at least one program product having a set (e.g., at least one) of program modules configured to carry out the functions of embodiments of the application.
A program/utility 40 having a set (at least one) of program modules 42 may be stored in, for example, memory 28, such program modules 42 including, but not limited to, an operating system, one or more application programs, other program modules, and program data, each or some combination of which may include an implementation of a network environment. Program modules 42 generally perform the functions and/or methods of the embodiments described herein.
The computer device 12 may also communicate with one or more external devices 14 (e.g., keyboard, pointing device, display 24, etc.), one or more devices that enable a user to interact with the computer device 12, and/or any devices (e.g., network card, modem, etc.) that enable the computer device 12 to communicate with one or more other computing devices. Such communication may occur through an input/output (I/O) interface 22. Moreover, the computer device 12 may also communicate with one or more networks such as a local area network (Local Area Network; hereinafter: LAN), a wide area network (Wide Area Network; hereinafter: WAN) and/or a public network such as the Internet via the network adapter 20. As shown, network adapter 20 communicates with other modules of computer device 12 via bus 18. It should be appreciated that although not shown, other hardware and/or software modules may be used in connection with computer device 12, including, but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, data backup storage systems, and the like.
The processing unit 16 executes various functional applications and data processing by running programs stored in the system memory 28, for example, implementing the positioning method of the root cause element mentioned in the foregoing embodiment.
Other embodiments of the application will be apparent to those skilled in the art from consideration of the specification and practice of the application disclosed herein. This application is intended to cover any variations, uses, or adaptations of the application following, in general, the principles of the application and including such departures from the present disclosure as come within known or customary practice within the art to which the application pertains. It is intended that the specification and examples be considered as exemplary only, with a true scope and spirit of the application being indicated by the following claims.
It is to be understood that the application is not limited to the precise arrangements and instrumentalities shown in the drawings, which have been described above, and that various modifications and changes may be effected without departing from the scope thereof. The scope of the application is limited only by the appended claims.

Claims (9)

1. A method of locating root cause elements, the method comprising:
Determining a first search space from call link spaces corresponding to services, wherein the first search space comprises: a plurality of nodes, the nodes comprising: at least one element, different elements have the same attribute dimension;
Determining a first set of cause elements for the first search space, the first set of cause elements comprising: at least one first root cause element, the first root cause element being an element of the at least one element that is related to a failure of the service;
determining a plurality of candidate root cause elements from a second search space, wherein the first search space and the second search space jointly form the calling link space, and attribute dimensions between the elements of the first search space and the elements of the second search space are different;
Selecting at least one second root cause element from the at least one candidate root cause element according to the first root cause element, wherein the second root cause element is an element related to the fault of the service in the second search space;
wherein the at least one first root cause element and the at least one second root cause element are jointly used as located root cause elements;
wherein said selecting at least one second root cause element from said at least one candidate root cause element based on said first root cause element comprises:
determining a first node potential score of a node to which the first root cause element belongs;
Determining a second node potential score of the node to which the candidate root cause element belongs;
Inputting the first root cause element, the candidate root cause element, the first node potential score, and the second node potential score into a pre-trained reinforcement learning model to obtain the at least one second root cause element output by the reinforcement learning model.
2. The method of claim 1, wherein the determining the first set of root cause elements for the first search space comprises:
Determining a plurality of node potential scores and a plurality of calling times corresponding to the plurality of nodes respectively, wherein the calling times are times when calling links of the service pass through the corresponding nodes;
Constructing a plurality of search trees respectively corresponding to the plurality of nodes according to the potential scores of the plurality of nodes and the plurality of calling times by combining a Monte Carlo tree search MCTS algorithm, wherein the search trees comprise: a root node and a predicted node derived based on the root node, the root node being the node corresponding to the search tree;
Determining a target node from the predicted nodes included in the plurality of search trees, wherein the node potential score of the target node is the same as the node potential score of the root node to which the target node belongs, and the layer number to which the target node belongs is the lowest layer number among the plurality of search trees;
And forming the first root cause element set according to the elements contained in the target node.
3. The method of claim 2, wherein the number of target nodes is at least one, the forming the first set of root cause elements from elements contained by the target nodes comprises:
If the number of the target nodes is a plurality of, determining the element number corresponding to the target nodes respectively;
the first root cause element set is formed according to elements contained in a first target node, and the first target node is the target node with the minimum element number among a plurality of target nodes.
4. The method of claim 1, wherein determining the first search space from the call link spaces corresponding to the services comprises:
determining multiple attribute dimensions corresponding to the service;
Determining a target attribute dimension from the plurality of attribute dimensions;
determining a node corresponding to the element associated with the target attribute dimension from the calling link space;
The first search space is formed from at least one of the nodes.
5. The method of claim 4, wherein said determining a target attribute dimension from among said plurality of attribute dimensions comprises:
Determining a plurality of elements respectively associated with the plurality of attribute dimensions;
Determining a plurality of real value distribution indexes and a plurality of predicted value distribution indexes which respectively correspond to the elements;
Determining a plurality of distance index values between the plurality of real value distribution indexes and the corresponding plurality of predicted value distribution indexes;
Selecting a target distance index value from the plurality of distance index values, wherein the target distance index value indicates that the influence degree of a corresponding target element on the key performance index KPI of the service is greater than that of other elements on the key performance index KPI of the service, and the target element and the other elements jointly form the associated plurality of elements;
and taking the attribute dimension of the target element as the target attribute dimension.
6. The method of claim 5, wherein the determining a plurality of distance index values between the plurality of real value distribution indices and a plurality of predicted value distribution indices further comprises:
and determining a plurality of jensen shannon JS divergences between the plurality of real value distribution indexes and the corresponding plurality of predicted value distribution indexes, and taking the plurality of jensen shannon JS divergences as the corresponding plurality of distance index values.
7. A root cause element positioning device, the device comprising:
the first determining module is configured to determine a first search space from call link spaces corresponding to services, where the first search space includes: a plurality of nodes, the nodes comprising: at least one element, different elements have the same attribute dimension;
A second determining module, configured to determine a first root cause element set of the first search space, where the first root cause element set includes: at least one first root cause element, the first root cause element being an element of the at least one element that is related to a failure of the service;
A third determining module, configured to determine a plurality of candidate root cause elements from a second search space, where the first search space and the second search space together form the call link space, and attribute dimensions between the elements of the first search space and the elements of the second search space are different;
A selecting module, configured to select at least one second root cause element from the at least one candidate root cause element according to the first root cause element, where the second root cause element is an element related to a fault of the service in the second search space;
wherein the at least one first root cause element and the at least one second root cause element are jointly used as located root cause elements;
the selecting module is specifically configured to:
determining a first node potential score of a node to which the first root cause element belongs;
Determining a second node potential score of the node to which the candidate root cause element belongs;
Inputting the first root cause element, the candidate root cause element, the first node potential score, and the second node potential score into a pre-trained reinforcement learning model to obtain the at least one second root cause element output by the reinforcement learning model.
8. A computer device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, the processor implementing the method of any of claims 1-6 when the program is executed.
9. A storage medium, which when executed by a processor of an electronic device, enables the electronic device to perform the method of any one of claims 1 to 6.
CN202110198580.XA 2021-02-22 2021-02-22 Root element positioning method and device, computer equipment and storage medium Active CN113824583B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110198580.XA CN113824583B (en) 2021-02-22 2021-02-22 Root element positioning method and device, computer equipment and storage medium

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110198580.XA CN113824583B (en) 2021-02-22 2021-02-22 Root element positioning method and device, computer equipment and storage medium

Publications (2)

Publication Number Publication Date
CN113824583A CN113824583A (en) 2021-12-21
CN113824583B true CN113824583B (en) 2024-06-18

Family

ID=78912455

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110198580.XA Active CN113824583B (en) 2021-02-22 2021-02-22 Root element positioning method and device, computer equipment and storage medium

Country Status (1)

Country Link
CN (1) CN113824583B (en)

Family Cites Families (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US7752154B2 (en) * 2007-02-26 2010-07-06 International Business Machines Corporation System and method for deriving a hierarchical event based database optimized for analysis of criminal and security information
EP2241983B1 (en) * 2009-04-17 2012-12-19 Fraunhofer-Gesellschaft zur Förderung der angewandten Forschung e.V. Method for searching objects in a database
US9961571B2 (en) * 2015-09-24 2018-05-01 Futurewei Technologies, Inc. System and method for a multi view learning approach to anomaly detection and root cause analysis
CN111444247B (en) * 2020-06-17 2023-10-17 北京必示科技有限公司 Root cause positioning method, root cause positioning device and storage medium based on KPI (key performance indicator)
CN112187554B (en) * 2020-12-01 2021-03-19 北京蒙帕信创科技有限公司 Operation and maintenance system fault positioning method and system based on Monte Carlo tree search

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
一种调和序列生成及其在故障定位中的应用;叶俊民;谢茜;姜丽;李嵩嵩;许磊;;计算机科学(03);全文 *
多因素集成的分布式应用故障诊断方法;尹殷;李巍;李云春;;计算机工程(20);全文 *

Also Published As

Publication number Publication date
CN113824583A (en) 2021-12-21

Similar Documents

Publication Publication Date Title
US7050950B2 (en) System, method and computer product for incremental improvement of algorithm performance during algorithm development
US9208053B2 (en) Method and system for predicting performance of software applications on prospective hardware architecture
CN111382255B (en) Method, apparatus, device and medium for question-answering processing
EP3418910A1 (en) Big data-based method and device for calculating relationship between development objects
CN107124289B (en) Weblog time alignment method, device and host
CN107168995B (en) Data processing method and server
CN113377567A (en) Distributed system fault root cause tracing method based on knowledge graph technology
CN115577114A (en) Event detection method and device based on time sequence knowledge graph
CN114036826A (en) Model training method, root cause determination method, device, equipment and storage medium
CN113792554A (en) Method and device for evaluating change influence based on knowledge graph
US8650180B2 (en) Efficient optimization over uncertain data
CN116502166A (en) Prediction method, device, equipment and medium based on other equipment data
CN113919510A (en) Sample feature selection method, device, equipment and medium
Rodriguez et al. Leveraging intermediate artifacts to improve automated trace link retrieval
CN113824583B (en) Root element positioning method and device, computer equipment and storage medium
CN112528500B (en) Evaluation method and evaluation equipment for scene graph construction model
CN114240654A (en) Performance capability perspective method, system, equipment and readable storage medium based on transaction link
CN111078440B (en) Disk error detection method, device and storage medium
Davidovsky et al. Agent-Based Implementation for the Discovery of Structural Difference in OWL-DL Ontologies
CN110727538B (en) Fault positioning system and method based on model hit probability distribution
CN113297069A (en) Software testing method and device based on target drive
CN114371950A (en) Root cause positioning method and device for application service abnormity
CN113434193B (en) Root cause change positioning method and device
Wu et al. GMBFL: Optimizing Mutation-Based Fault Localization via Graph Representation
CN117556920B (en) Large model illusion treatment method, device, equipment and storage medium

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant