Fault dependency graph construction method, electronic device and storage medium

Abstract

The present invention relates to the field of computer technology applications, and provides a method for constructing a fault dependence graph, an electronic device, and a storage medium, including: obtaining basic information I corresponding to the current fault of the system to be monitored; based on Q1 to Q n Construct a fully connected network graph; obtain a causal connection network graph based on the fully connected network graph; construct a fault dependence graph corresponding to the current fault based on I and the causal connection network graph. When a fault occurs in the system, the present invention can display the fault indicators and the indicators corresponding to the causes of the fault, so as to visually display the fault, improve the user experience, and help improve the accuracy and interpretability of root cause analysis.

Description

Technical Field

[0001] The present invention relates to the field of computer technology applications, and in particular to a fault dependency graph construction method, electronic equipment and storage medium. Background Art

[0002] Currently, for online service systems, due to the large amount and variety of monitoring data of online service systems and the complex dependencies between components (such as services or databases), in order to intuitively know the components that have failed, fault dependency graphs are generally used. The fault dependency graph uses faulty units as nodes and connects the corresponding nodes according to the call relationship and deployment relationship between components. However, in this fault dependency graph, the lines between nodes are undirected, which means that the causal relationship and propagation direction between faults cannot be directly represented. In real systems, the propagation of faults is usually directional, and some faults may be the result of other faults. Due to the lack of directional information, traditional fault dependency graphs may ignore the complex interactions between faults, resulting in limited accurate identification of root causes. Summary of the invention

[0003] In view of the above technical problems, the technical solution adopted by the present invention is:

[0004] An embodiment of the present invention provides a method for constructing a fault dependency graph, the method comprising the following steps:

[0005] S100, obtaining basic information corresponding to the current fault of the system to be monitored I=(I 1 , I 2 , ..., I i , ..., I n ); the i-th basic information I i =(P i , Q i ), P i is the ID of the i-th component corresponding to the current fault, i ranges from n to n, and n is the number of components corresponding to the current fault; Q i P i The corresponding indicator information set, Q i =(Q i1 , Q i2 , ..., Q ij , ..., Q if（i） ), Q ij P i The corresponding indicator information of the jth indicator, the value of j ranges from 1 to f(i), and f(i) is P i The corresponding number of indicators; Q ij =(q ij , d 1 ij , d 2ij , ..., d r ij , ..., d m ij ), q ij P i The ID of the corresponding i-th indicator, d r ij is the q collected at the rth collection time corresponding to the current fault ij The index value of r ranges from 1 to m, and m is the number of collection moments.

[0006] S200, based on Q 1 To Q n Construct a fully connected network graph; the nodes of the fully connected network graph are composed of indicators.

[0007] S300, obtaining a causal connection network diagram based on the fully connected network diagram; nodes in the causal connection network diagram that have connection relationships are connected through corresponding directed connection lines.

[0008] S400, constructing a fault dependency graph corresponding to the current fault based on I and the causal connection network graph; the nodes of the fault dependency graph are composed of components and corresponding indicators, and the indicators with connection relationships are connected through corresponding directed connection lines.

[0009] The present invention has at least the following beneficial effects:

[0010] The dependency graph construction method provided in the embodiment of the present invention constructs a fault dependency graph based on the components of the system to be monitored and the corresponding indicators. For each node in the fault dependency graph, the indicator data is analyzed for causal relationships, and directed edges are introduced to represent the causal relationships between faults, thereby more accurately reflecting the propagation direction and mutual correlation of the faults. Therefore, when a system fault occurs, the fault indicator and the corresponding indicator causing the cause of the fault can be displayed, thereby making it possible to intuitively display the fault, improve user experience, and help improve the accuracy and explainability of root cause analysis. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0012] Figure 1 A flowchart of a method for constructing a fault dependency graph provided by an embodiment of the present invention;

[0013] Figure 2A schematic diagram showing a collision structure;

[0014] Figures 3a to 6b A schematic diagram showing the setting of directed edges;

[0015] Figure 7 A schematic diagram of a fault dependency graph according to an embodiment of the present invention. DETAILED DESCRIPTION

[0016] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0017] An embodiment of the present invention provides a fault dependency graph construction method, which is used to construct a fault dependency graph in a monitoring model for monitoring faults of a system to be monitored.

[0018] In the embodiment of the present invention, the system to be monitored may be any system that provides data services. For example, in an exemplary embodiment, the system may be an online service system.

[0019] Furthermore, if Figure 1 As shown, the fault dependency graph construction method provided by the embodiment of the present invention may include the following steps:

[0020] S100, obtaining basic information corresponding to the current fault of the system to be monitored I=(I 1 , I 2 , ..., I i , ..., I n ); the i-th basic information I i =(P i , Q i ), P i is the ID of the i-th component corresponding to the current fault, i ranges from n to n, and n is the number of components corresponding to the current fault; Q i P i The corresponding indicator information set, Q i =(Q i1 , Q i2 , ..., Q ij , ..., Q if（i） ), Q ij P i The corresponding indicator information of the jth indicator, the value of j ranges from 1 to f(i), and f(i) is P i The corresponding number of indicators; Q ij =(q ij , d1 ij , d 2 ij , ..., d r ij , ..., d m ij ), q ij P i The ID of the corresponding i-th indicator, d r ij is the q collected at the rth collection time corresponding to the current fault ij The index value of r ranges from 1 to m, and m is the number of collection moments.

[0021] In an embodiment of the present invention, the ID of a component may be the name of the component, and the ID of an indicator may be the name of the indicator. A component may be a service, a database, a container, etc. provided by the system. An indicator may be a parameter representing the performance of a component, and the indicators of a component may provide a better understanding of what is happening inside the system. Components and corresponding indicators may be determined based on actual conditions. For example, service indicators may include average response time, success rate, process time, etc. Database indicators may include total memory, used memory, etc.; container indicators may include CPU usage, memory usage, etc.

[0022] In the embodiment of the present invention, the index value can be collected according to the set sampling interval, that is, the interval between two sampling moments. The set sampling interval can be set based on actual needs. The value of m is determined based on the time interval between the last fault and the current fault. △t is the time interval between the current fault and the last fault, t0 is the set sampling interval, Indicates rounding down.

[0023] In the embodiment of the present invention, the component corresponding to the current fault is the component involved in the fault currently occurring in the system, and the specific involved component can be obtained based on the specific fault type. Different faults correspond to different components.

[0024] S200, based on Q 1 To Q n Construct a fully connected network graph; the nodes of the fully connected network graph are composed of indicators, that is, one node in the fully connected network graph represents one indicator. In an embodiment of the present invention, the nodes of the fully connected network graph can be represented by a custom shape, such as a circle or a square. Each node can be set with an ID of a corresponding indicator.

[0025] In S200, each indicator corresponding to the current fault is taken as a node to construct a fully connected network graph.

[0026] S300, obtaining a causal connection network diagram based on the fully connected network diagram; nodes in the causal connection network diagram that have connection relationships are connected through corresponding directed connection lines.

[0027] Furthermore, S300 may specifically include:

[0028] S310, obtaining an association value between any two nodes in the fully connected network graph, and determining whether to delete a connection line between the two nodes based on the obtained association value; and obtaining an intermediate connection network graph.

[0029] S320, analyzing the causal relationship between any two nodes in the intermediate connection network diagram to obtain a connection network diagram with a causal relationship added, thereby obtaining the causal connection network diagram.

[0030] Further, in the embodiment of the present invention, S310 may specifically include:

[0031] S3101, traverse the fully connected network graph, and for any node pair (a, b) in the fully connected network graph, obtain all nodes connected to node a as the adjacent node set Aa of node a; the values ​​of a and b are 1 to N, N is the number of nodes in the fully connected network graph, and a≠b. Obviously, N=f(1)+f(2)+…+f(i)…+f(n).

[0032] S3102, remove node b from Aa as the condition set Ka of node a, and obtain the associated value set C of node a and node b under the condition Ka ab .

[0033] In the embodiment of the present invention, C ab ={C 1 ab , C 2 ab , ..., C h ab , ..., C z（a） ab}, C h ab C ab The hth associated value in , where h ranges from 1 to z(a), and z(a) is the number of nodes in Ka.

[0034] Among them, C h ab Based on the following steps:

[0035] S10, obtain the correlation coefficient g of node a and node b under node h h ab =(g ab -g ah ×gbh ) / [(1-(g ah ) 2 )(1-(g bh ) 2 ))] 1 / 2 ;g ab is the correlation coefficient between node a and node b, g ah is the correlation coefficient between node a and node h, g bh is the correlation coefficient between node b and node h.

[0036] In the embodiment of the present invention, the correlation coefficient g between node x and node y is xy It can be obtained based on the existing correlation analysis method. In an exemplary embodiment, it can be obtained based on the Pearson correlation coefficient analysis method, that is, g xy =[∑ m j=1 (d xj ×d yj )-∑ m j=1 d xj ∑ m j=1 d yj ] / [(∑ m j=1 (d xj ) 2 -(∑ m j=1 d xj ) 2 / m)((∑ m j=1 (d yj ) 2 -(∑ m j=1 d yj ) 2 )] 1 / 2 ;d xj is the index value corresponding to node x at the jth collection time, d yj is the index value corresponding to node y at the jth collection time. When x is a, y is b or h; when x is b, y is h.

[0037] S12, get g h ab The intermediate correlation value p h ab =(0.5×ln(1+p h ab )) / (1+p h ab ).

[0038] S14, based on |p hab ∣, consult the F test critical value table and get the corresponding correlation value C h ab . ∣ ∣ represents an absolute value.

[0039] Those skilled in the art know that the F-test critical value table may be existing content.

[0040] S3103, if C ab If there is a value less than C0 in , it means that nodes a and b are unrelated. Then the connection line between nodes a and b is deleted, and Ka is used as the separation set S of nodes a and b. ab , and added to the current separation set S; obtain the intermediate connection network diagram, the initial value of S is an empty set, and C0 is the set association value threshold.

[0041] In the embodiment of the present invention, C0 may be an empirical value. In an illustrative embodiment, C0=0.005.

[0042] Those skilled in the art know that if C ab There is no value less than C0, that is, all the associated values ​​are greater than or equal to C0, which means that node a and node b are related, and the connection line between them is not deleted. It can be seen that the intermediate connection network graph is a network graph in which the connection lines between nodes that do not have a connection relationship are deleted from the fully connected network graph.

[0043] Further, in the embodiment of the present invention, S320 may specifically include:

[0044] S3201, obtain the node string set CN in the intermediate connection network graph = {CN 1 , CN 2 ,……,CN e ,……,CN R}, CN e is the e-th node string in CN, CN e It is formed by three nodes e1 to e3 connected in sequence; the value of e is 1 to R, and R is the number of node strings in CN.

[0045] S3202, traverse CN, if CN e If the nodes in meet the preset conditions, the connection line between node e1 and node e2 is set to be a directed connection line and the direction of the directed connection line points to node e2, and the connection line between node e3 and node e2 is set to be a directed connection line and the direction of the directed connection line points to node e2, that is, node e1 and node e3 are both causes of node e2, CN e For collision structures, such as Figure 2As shown; execute S3203, otherwise, it means that the node that meets the preset conditions has not been found, and it is necessary to continue searching and execute S3202.

[0046] In the embodiment of the present invention, the preset condition is: node e2 is not in the separation set corresponding to node e1 and node e2, and there is no connection line between node e1 and node e3. In the embodiment of the present invention, as long as a node string that meets the preset condition is found in CN, S3204 is executed, that is, only a node string that meets the preset condition needs to be found.

[0047] S3203, if there are nodes in the current connection network diagram that meet any of the first to fourth conditions, execute the corresponding operation to update the current connection network diagram, and use the updated connection network diagram as the current connection network diagram, execute S3203, that is, continue to determine whether there are nodes in the current connection network diagram that meet any of the first to fourth conditions; otherwise, it means that there are no nodes in the current connection network that can be connected in a directed manner, exit the current control program, and obtain the causal connection network diagram.

[0048] In the embodiment of the present invention, the first condition is: there is a directed connection line between node u1 and node u2, and node u2 and node u3 are directly connected but the connection line between node u2 and node u3 is not a directed connection line, and the direction of the directed connection line between node u1 and node u2 points to node u2, and there is no connection line between node u1 and node u3, that is, node u1 and node u3 are not connected, such as Figure 3a .

[0049] The operation corresponding to the first condition is: set the connection line between node u2 and node u3 to be a directed connection line, and the direction of the directed connection line between node u2 and node u3 points to node u3, such as Figure 3b shown.

[0050] The operation corresponding to the first condition is to avoid new collision structures. Because if a new collision structure is introduced in the analysis, it may lead to the erroneous interpretation of correlation as causality. Therefore, in order to ensure the accuracy and reliability of causal inference, it is necessary to consider that there should be no new collision structures in subsequent analysis. Only by inferring causal relationships based on the already determined collision structure can we more accurately understand and interpret the causal relationship in the data.

[0051] The second condition is: there is a directed connection line between node u1 and node u2 and there is a directed connection line between node u2 and node u3, and the direction of the directed connection line between node u1 and node u2 points to node u2, the direction of the directed connection line between node u2 and node u3 points to node u3, and node u1 and node u3 are directly connected, that is, directly connected by the connection line, and there is no other node in between, such as Figure 4a shown.

[0052] The operation corresponding to the second condition is: set the direction of the directed connection line between node u1 and node u3 to point to node u3, such as Figure 4b The third condition is: node u1 and node u2 are directly connected but the connection line between node u1 and node u2 is not a directed connection line, the connection line between node u2 and node u3 is a directed connection line, and the direction of the directed connection line between node u2 and node u3 points to node u3, node u1 and node u4 are directly connected but the connection line between node u1 and node u4 is not a directed connection line, the connection line between node u4 and node u3 is a directed connection line, and node u1 and node u3 are directly connected, as shown in FIG. Figure 5a shown.

[0053] The operation corresponding to the third condition is: set the connection line between node u1 and node u3 to be a directed connection line, and the direction of the directed connection line between node u1 and node u3 points to node u3, such as Figure 5b shown.

[0054] The fourth condition is: the node u1 and the node u2 are directly connected but the connection line between the node u1 and the node u2 is not a directed connection line, the connection line between the node u2 and the node u3 is a directed connection line, and the direction of the directed connection line between the node u2 and the node u3 points to the node u3, the connection line between the node u2 and the node u4 is a directed connection line, and the direction of the directed connection line between the node u2 and the node u4 points to the node u4; the connection line between the node u3 and the node u4 is a directed connection line, and the direction of the directed connection line between the node u3 and the node u4 points to the node u4; and the node u1 and the node u4 are directly connected, such as Figure 6a shown.

[0055] The operation corresponding to the fourth condition is: setting the connection line between node u1 and node u4 to a directed connection line, and the direction of the directed connection line between node u1 and node u4 points to node u4, such as Figure 6b shown.

[0056] In the embodiment of the present invention, the operations corresponding to the first condition to the fourth condition are to prevent the currently connected network from forming a ring.

[0057] In the embodiment of the present invention, the direction of the directed connection line is the effect, and the corresponding direction is the cause, that is, the cause points to the effect.

[0058] S400, construct a fault dependency graph corresponding to the current fault based on I and the causal connection network graph; the nodes of the fault dependency graph are composed of components and corresponding indicators, a node of the set fault dependency graph represents a component and a corresponding indicator, and the indicators with connection relationships are connected through corresponding directed connection lines.

[0059] In the embodiment of the present invention, the connection relationship between the nodes of the fault dependency graph is obtained based on the call relationship and deployment relationship between components. The nodes of the fault dependency graph can be represented by boxes, and each node is provided with the ID of the component and the ID of the corresponding indicator.

[0060] In a specific embodiment, a basic network diagram without connection lines can be constructed based on the components of the system to be monitored and the corresponding indicators, that is, no connection between any two nodes. Afterwards, each time a system failure occurs, a causal connection network diagram is constructed based on the indicators involved in the current failure, and then the corresponding causal edges are constructed in the basic network diagram based on the causal connection network diagram to obtain the fault dependency graph of the current failure.

[0061] In an exemplary embodiment, the structure of the fault dependency graph in the embodiment of the present invention can be as follows: Figure 7 It should be noted that Figure 7 Only two components are shown, and each component shows three indicators, which is only for illustrative purposes and not for limiting purposes.

[0062] Furthermore, in the embodiment of the present invention, the method may further include the following steps:

[0063] In response to a failure detected in any indicator of the system, a visual display is performed on the failure dependency graph.

[0064] In an embodiment of the present invention, the visual display includes displaying the indicator where the fault occurs in a set color and the direction of the directed connection line between the indicator and the indicator is the indicator pointing to the indicator. The set color can be set based on actual needs, for example, red.

[0065] One application scenario of the fault dependency graph construction method provided by the embodiment of the present invention is to construct a fault dependency graph used in a fault monitoring model that uses a root cause location method to perform fault monitoring. For a specific online service system, the fault monitoring model takes the historical faults and dependencies in the system as input, and trains the root cause location model offline; for the input fault, the model will give the location where the fault occurred (i.e., the faulty component) and the type of fault that occurred (i.e., the indicative metric group).

[0066] Those skilled in the art know that the specific structure of the fault monitoring model can be an existing structure. The fault dependency graph constructed by the fault dependency graph construction method provided in the embodiment of the present invention is introduced into the fault monitoring model. Experimental results on a specific data set show that the accuracy and efficiency of root cause location can be improved.

[0067] An embodiment of the present invention also provides a non-transitory computer-readable storage medium, which can be set in an electronic device to store at least one instruction or at least one program related to implementing a method in a method embodiment. The at least one instruction or the at least one program is loaded and executed by the processor to implement the method provided in the above embodiment.

[0068] In an embodiment of the present invention, a non-transient computer-readable storage medium may include a storage program area and a storage data area, wherein the storage program area may store an operating system, an application required for at least one function; the storage data area may store data created according to a server hard disk security measurement method, etc. In addition, the memory may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other non-volatile solid-state storage devices. In some embodiments, a non-transient computer-readable storage medium may optionally include a memory remotely arranged relative to the processor, and these remote memories may be connected to a local module via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof such as read-only memory, flash memory, a hard disk drive (HDD) or a solid-state drive (SSD), or a memory is any other medium that can be used to carry or store a desired program code in the form of an instruction or data structure and can be accessed by a computer, but is not limited thereto. The non-transient computer-readable storage medium may be a combination of the above-mentioned memories.

[0069] An embodiment of the present invention further provides an electronic device, comprising a processor and the aforementioned non-transitory computer-readable storage medium.

[0070] Although some specific embodiments of the present invention have been described in detail by way of example, it should be understood by those skilled in the art that the above examples are for illustration only and are not intended to limit the scope of the present invention. It should also be understood by those skilled in the art that various modifications may be made to the embodiments without departing from the scope and spirit of the present invention. The scope of the present invention is defined by the appended claims.

Claims

1. A method for constructing a fault dependency graph, characterized in that: The method comprises the following steps: S100, obtaining basic information corresponding to the current fault of the system to be monitored I=(I1, I2, ..., I i , ..., I n )。 ); The i-th basic information I i =(P i , Q i ), P i is the ID of the i-th component corresponding to the current fault, i ranges from n to n, and n is the number of components corresponding to the current fault; Q i P i The corresponding indicator information set, Q i =(Q i1 , Q i2 , ..., Q ij , ..., Q if（i） ), Q ij P i The corresponding indicator information of the jth indicator, the value of j ranges from 1 to f(i), and f(i) is P i The corresponding number of indicators; Q ij =(q ij , d 1 ij , d 2 ij , ..., d r ij , ..., d m ij ), q ij P i The ID of the corresponding i-th indicator, d r ij is the q collected at the rth collection time corresponding to the current fault ij The index value of r ranges from 1 to m, where m is the number of collection moments; S200, based on Q1 to Q n Constructing a fully connected network graph; the nodes of the fully connected network graph are composed of indicators; S300, obtaining a causal connection network diagram based on the fully connected network diagram; nodes in the causal connection network diagram that have connection relationships are connected through corresponding directed connection lines; S400, constructing a fault dependency graph corresponding to the current fault based on I and the causal connection network graph; The nodes of the fault dependency graph are composed of components and corresponding indicators, and the indicators with connection relationships are connected through corresponding directed connection lines; S300 specifically includes: S310, obtaining an association value between any two nodes in the fully connected network graph, and determining whether to delete a connection line between the two nodes based on the obtained association value; obtaining an intermediate connection network graph; S320, analyzing the causal relationship between any two nodes in the intermediate connection network diagram to obtain a connection network diagram with a causal relationship added, thereby obtaining the causal connection network diagram; S310 specifically includes: S3101, traverse the fully connected network graph, and for any node pair (a, b) in the fully connected network graph, obtain all nodes connected to node a as the adjacent node set Aa of node a; the values ​​of a and b are 1 to N, N is the number of nodes in the fully connected network graph, and a≠b; S3102, remove node b from Aa as the condition set Ka of node a, and obtain the associated value set C of node a and node b under the condition Ka ab ; S3103, if C ab If there is a value less than C0 in , the connection line between node a and node b is deleted, and Ka is used as the separation set S of node a and node b. ab , and added to the current separation set S; obtain the intermediate connection network diagram, the initial value of S is an empty set, and C0 is the set association value threshold; C ab ={C 1 ab , C 2 ab , ..., C h ab , ..., C z（a） ab }, C h ab C ab The hth associated value in , where h ranges from 1 to z(a), and z(a) is the number of nodes in Ka; Among them, C h ab Based on the following steps: S10, obtain the correlation coefficient g of node a and node b under node h h ab =(g ab -g ah ×g bh ) / [(1-(g ah ) 2 )(1-(g bh ) 2 ))] 1 / 2 ;g ab is the correlation coefficient between node a and node b, g ah is the correlation coefficient between node a and node h, g bh is the correlation coefficient between node b and node h; where the correlation coefficient g between node x and node y xy =[∑ m j=1 (d xj ×d yj )-∑ m j=1 d xj ∑ m j=1 d yj ] / [(∑ m j=1 (d xj ) 2 -(∑ m j=1 d xj ) 2 / m)((∑ m j=1 (d yj ) 2 -(∑ m j=1 d yj ) 2 )] 1 / 2 ;d xj is the index value corresponding to node x at the jth collection time, d yj is the index value corresponding to node y at the jth collection time. When x is a, y is b or h; when x is b, y is h; S12, get g h ab The intermediate correlation value p h ab =(0.5×ln(1+p h ab )) / (1+p h ab ); S14, based on |p h ab ∣, consult the F test critical value table and get the corresponding correlation value C h ab .

2. The method according to claim 1, characterized in that S320 specifically includes: S3201, obtaining a node string set CN={CN1, CN2, ..., CN e ,……,CN R }, CN e is the e-th node string in CN, CN e It is formed by three nodes e1 to e3 connected in sequence; the value of e ranges from 1 to R, and R is the number of node strings in CN; S3202, traverse CN, if CN e If the nodes in the table satisfy the preset conditions, then the connection line between node e1 and node e2 is set to be a directed connection line and the direction of the directed connection line points to node e2, and the connection line between node e3 and node e2 is set to be a directed connection line and the direction of the directed connection line points to node e2; execute S3203; otherwise, execute S3202; S3203, if there is a node in the current connection network diagram that satisfies any of the first condition to the fourth condition, then execute the corresponding operation to update the current connection network diagram, and use the updated connection network diagram as the current connection network diagram to execute S3203; otherwise, exit the current control program to obtain the causal connection network diagram; The first condition is: there is a directed connection line between node u1 and node u2, and node u2 and node u3 are directly connected but the connection line between node u2 and node u3 is not a directed connection line, and the direction of the directed connection line between node u1 and node u2 points to node u2, and there is no connection line between node u1 and node u3; The operation corresponding to the first condition is: setting the connection line between the node u2 and the node u3 to be a directed connection line, and the direction of the directed connection line between the node u2 and the node u3 points to the node u3; The second condition is: there is a directed connection line between node u1 and node u2 and there is a directed connection line between node u2 and node u3, and the direction of the directed connection line between node u1 and node u2 points to node u2, the direction of the directed connection line between node u2 and node u3 points to node u3, and node u1 and node u3 are directly connected; The operation corresponding to the second condition is: setting the direction of the directed connection line between node u1 and node u3 to point to node u3; The third condition is: the node u1 and the node u2 are directly connected but the connection line between the node u1 and the node u2 is not a directed connection line, the connection line between the node u2 and the node u3 is a directed connection line, and the direction of the directed connection line between the node u2 and the node u3 points to the node u3, the node u1 and the node u4 are directly connected but the connection line between the node u1 and the node u4 is not a directed connection line, the connection line between the node u4 and the node u3 is a directed connection line, and the node u1 and the node u3 are directly connected; The operation corresponding to the third condition is: setting the connection line between the node u1 and the node u3 to be a directed connection line, and the direction of the directed connection line between the node u1 and the node u3 points to the node u3; The fourth condition is: the node u1 and the node u2 are directly connected but the connection line between the node u1 and the node u2 is not a directed connection line, the connection line between the node u2 and the node u3 is a directed connection line, and the direction of the directed connection line between the node u2 and the node u3 points to the node u3, the connection line between the node u2 and the node u4 is a directed connection line, and the direction of the directed connection line between the node u2 and the node u4 points to the node u4; the connection line between the node u3 and the node u4 is a directed connection line, and the direction of the directed connection line between the node u3 and the node u4 points to the node u4; and the node u1 and the node u4 are directly connected; The operation corresponding to the fourth condition is: setting the connection line between the node u1 and the node u4 to be a directed connection line, and the direction of the directed connection line between the node u1 and the node u4 points to the node u4.

3. The method according to claim 2, characterized in that The preset condition is that the node e2 is not in the separation set corresponding to the node e1 and the node e2, and there is no connection line between the node e1 and the node e3.

4. The method according to claim 1, characterized in that The system is an online service system.

5. A non-transitory computer-readable storage medium, wherein at least one instruction or at least one program is stored in the storage medium, characterized in that: The at least one instruction or the at least one program is loaded and executed by the processor to implement the method according to any one of claims 1 to 4.

6. An electronic device, characterized in that: The invention comprises a processor and the non-transitory computer-readable storage medium as claimed in claim 5.

Images