Abnormal point location rapid positioning method and system
By using TraceID tracking and log data analysis methods in software testing, we quickly locate and classify software abnormal points, solving the problem of abnormal points positioning difficulties caused by software complexity, and improving testing efficiency and development progress.
Patent Information
- Application Number
- CN202411200605.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-29
- Publication Date
- 2025-02-28
AI Technical Summary
In existing software testing, due to the high complexity of the software, it is difficult for testers to quickly locate abnormal points of the software, resulting in a long test time and affecting the development progress.
By receiving task requests sent by the client, creating a TraceID, and using TraceID to create a link tracking path, collecting the data called by the data call action, and storing it in the positioning unit. After obtaining the query request, read the TraceID, filter the log data, locate the exception points, and classify them.
It realizes the rapid location of abnormal points in the software, improves the speed of handling abnormal problems, and can quickly identify and classify abnormal problems through TraceID tracking and log data analysis.
Smart Images
Figure CN119537199A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of computer technology, and in particular to a method and system for quickly locating abnormal points. Background Art
[0002] With the development of science and technology, software is becoming more and more integrated into every aspect of people's lives. In order to improve the stability of the software, it is necessary to test the software to detect any problems in the software in advance.
[0003] Software testing was born along with the emergence of software. In the early software development process, the software was small in scale and low in complexity. The software development process was chaotic and quite random. The purpose of software testing was to correct known faults in the software, and this part of the work was often completed by the developers themselves. As software tends to be large-scale and highly complex, the quality of software is becoming more and more important. People began to design various processes and management methods for software development, and the way of software development has gradually transitioned from a chaotic and disordered development process to a structured development process, characterized by structured analysis and design, structured review, structured programming, and structured testing.
[0004] During the software testing process, testers are required to be able to quickly locate abnormal points in the software. However, due to the high complexity of existing software, relying solely on testers to locate abnormal points in the software often takes a lot of time, making it difficult to quickly locate abnormal points in the software, and even affecting the software development progress. Summary of the invention
[0005] Based on this, it is necessary to provide a method for quickly locating abnormal points in view of the high complexity of existing software. Relying solely on testers to locate abnormal points of the software often takes a lot of time, making it difficult to quickly locate abnormal points of the software, and even affecting the progress of software development.
[0006] On the one hand, the present application provides a method for quickly locating an abnormal point, comprising:
[0007] Receive task requests sent by the client;
[0008] Create a TraceID, and create a link tracking path by combining the request information and TraceID contained in the client's task request;
[0009] Execute a data calling action corresponding to the task request of the client, and when executing the data calling action corresponding to the task request of the client, collect the data called by the data calling action by using the link tracking path;
[0010] The data called by the data calling action is used as the tracking data, and the tracking data is stored in the positioning unit corresponding to the TraceID;
[0011] Obtain the query request sent by the client, and read the query information contained in the query request;
[0012] Read the TraceID in the query information, and filter out the log data corresponding to the TraceID in the query information;
[0013] The abnormal points are located according to the log data, and the abnormal points are graded based on the abnormal point location results of the log data.
[0014] On the other hand, the present application also provides a system for rapidly locating abnormal points, the system comprising:
[0015] Client; sending request information through the client;
[0016] The server is connected to the client for communication, and includes a storage module for storing log data. The server receives request information sent by the client and executes the method for rapid location of abnormal points mentioned in the above content.
[0017] The present application relates to a method and system for quickly locating an abnormal point. The method for quickly locating an abnormal point receives a task request sent by a client and creates a TraceID, thereby creating a link tracing path, then uses the link tracing path to collect data called by a data call action, uses the data called by the data call action as tracing data, and stores the tracing data in a positioning unit corresponding to the TraceID. After obtaining a query request sent by the client, the TraceID in the query information is read, and log data corresponding to the TraceID in the query information is screened out. Finally, the abnormal point is located according to the log data, and the abnormal point is graded in combination with the abnormal point location result of the log data. The present application can quickly locate the abnormal location in the task request by adding a TraceID to the task request and according to the TraceID, and the abnormal problem corresponding to the abnormal point is graded according to the position of the abnormal point, thereby improving the processing speed of the abnormal problem. BRIEF DESCRIPTION OF THE DRAWINGS
[0018] Figure 1 A method flow chart of a method for quickly locating an abnormal point provided in one embodiment of the present application.
[0019] Figure 2 A schematic diagram of the structure of a system for rapidly locating abnormal points provided in one embodiment of the present application.
[0020] Reference numerals:
[0021] 100, client; 200, server; 201, storage module. DETAILED DESCRIPTION
[0022] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.
[0023] The present application provides a method and system for quickly locating abnormal points.
[0024] like Figure 1 As shown, in one embodiment of the present application, a method for quickly locating an abnormal point is provided, and the method for quickly locating an abnormal point includes:
[0025] S100: receiving a task request sent by a client.
[0026] Specifically, the task request sent by the client is the action performed by the client. For example, when the client performs a login operation, a task request for "login account" will be sent to the server through the client.
[0027] Then the request information of the task request includes the user account and user password.
[0028] S200, creating a TraceID, and creating a link tracing path by combining the request information and the TraceID included in the task request of the client.
[0029] Specifically, TraceID is a globally unique ID assigned to a task request in order to track the task request sent by the client, so that the complete life cycle of the task request can be tracked through TraceID.
[0030] S300, executing a data calling action corresponding to a task request of a client, and when executing the data calling action corresponding to the task request of the client, collecting data called by the data calling action by using a link tracking path.
[0031] S400: The data called by the data calling action is used as the tracing data, and the tracing data is stored in the positioning unit in correspondence with the TraceID.
[0032] S500, obtaining a query request sent by the client, and reading query information contained in the query request.
[0033] S600, reading the TraceID in the query information, and filtering out the log data corresponding to the TraceID in the query information.
[0034] S700: locate abnormal points according to the log data, and classify the abnormal points based on the abnormal point location results of the log data.
[0035] Specifically, the grading results include immediate resolution, high priority, normal level, and low priority from high to low.
[0036] In this embodiment, a link tracing path is created by receiving a task request sent by a client and creating a TraceID, and then the link tracing path is used to collect data called by the data call action, and the data called by the data call action is used as tracing data, and the tracing data is stored in a positioning unit corresponding to the TraceID. After obtaining the query request sent by the client, the TraceID in the query information is read, and the log data corresponding to the TraceID in the query information is filtered out. Finally, the abnormal point is located according to the log data, and the abnormal point is graded in combination with the abnormal point location result of the log data. The present application can quickly locate the abnormal location in the task request according to the TraceID by adding a TraceID to the task request, and the abnormal problem corresponding to the abnormal point is graded according to the position of the abnormal point, thereby improving the processing speed of the abnormal problem.
[0037] In one embodiment of the present application, the S200 includes:
[0038] S210: Read request information in the task request.
[0039] S220, adding TraceID to the request header of the request information and the log data.
[0040] Specifically, before adding the TraceID to the request header of the request information and the log data, it is necessary to define the generation rule of the TraceID.
[0041] For example, the generation rule of TraceID in this embodiment may be: TraceID=server IP+auto-increment sequence+current process number+generation time.
[0042] In this way, a unique TraceID can be generated for each task request, and the server IP and creation time associated with the task request can be directly determined based on the TraceID of the task request, so that in addition to the tracking function, TraceID can also directly record the relevant information of the task request.
[0043] In this embodiment, by adding a TraceID to the request header of the request information in the task request sent by the client, a link tracing path is created according to the TraceID, and the task request is traced according to the TraceID.
[0044] The link tracking path is to restore a task request to a call link and display the call status of the task request, including the server IP of the service node, the request status of the service node, etc.
[0045] In one embodiment of the present application, the S220 includes:
[0046] S221, determining whether the request header of the request information contains TraceID.
[0047] Specifically, the request header is an important part of the HTTP request. It contains additional information about the request, such as User-Agent, Accept, etc. This information is very helpful for the server to understand the content and context of the request.
[0048] The request header provides additional information about the client, the purpose of the request, the language received, etc., to help the server better handle the request. For example, through the User-Agent header, the server can identify the client's browser type and version. The Accept header tells the server the type of content that the client can receive. This information is essential for implementing features such as cache control, authentication, and content negotiation.
[0049] S222: If the request header of the request information does not include a TraceID, read the TraceID generation rule, create a TraceID according to the TraceID generation rule, and add the TraceID to the request header of the request information sent by the client.
[0050] S223: If the request header of the request information includes a TraceID, the TraceID is extracted and the extracted TraceID is recorded as a target TraceID.
[0051] S224, add the target TraceID to each log.
[0052] In this embodiment, after receiving and reading the task request sent by the client, it is first determined whether the request header of the task request contains a TraceID. If the request header is empty, a TraceID is generated according to the TraceID creation rule and added to the request header of the task request.
[0053] More specifically, since software testing is mainly aimed at the web side and the software side, there are certain differences in the testing process between the web side and the software side.
[0054] When testing tasks on the web page, the task request will first reach the gateway. The interceptor in the gateway will determine whether the request header of the task request contains TraceID. If the request header does not contain TraceID, a TraceID will be generated according to the creation rule and added to the request header when the gateway forwards the request.
[0055] When testing a task on the software side, the task request will reach the server side. The interceptor in the software application will determine whether the request header of the task request contains TraceID. If the request header does not contain TraceID, a TraceID will be generated according to the creation rule and added to the request header when calling the service.
[0056] In one embodiment of the present application, the S600 includes:
[0057] S610: Obtain the TraceID included in the request header of the query information.
[0058] S620: Using the TraceID included in the request header of the query information as an index, query the tracing data corresponding to the TraceID in the positioning unit.
[0059] S630: Obtain log data from the tracking data.
[0060] In this embodiment, since the task request has a unique TraceID, the TraceID can be added to the text of the query information, and the task request can be traced by retrieving the trace data corresponding to the TraceID in the positioning unit.
[0061] In one embodiment of the present application, the S600 further includes:
[0062] S640: Display the tracking data and locate the abnormal point according to the log data in the tracking data.
[0063] In this embodiment, after the tracing data of the task request is tracked according to TraceID, the tracing data is displayed to the user through the client, and the tester can also locate the abnormal point according to the tracing data, and determine the type of abnormal problem corresponding to the abnormal point according to the link or step where the abnormal point is located, so as to facilitate the subsequent calculation of the abnormal score according to the type of abnormal problem, thereby grading the abnormal problem.
[0064] In one embodiment of the present application, the S700 includes:
[0065] S710, obtaining the abnormal point location result corresponding to the log data to obtain the abnormal point.
[0066] S720, calculating the abnormal point score of the abnormal point using Formula 1.
[0067] P=A×M+B×N Formula 1.
[0068] Among them, P is the abnormal point score of the abnormal point, A is the positioning time of the abnormal point, B is the reference score of the type of abnormal problem corresponding to the abnormal point, M is the weight coefficient of the positioning time of the abnormal point, and N is the weight coefficient of the type of abnormal problem corresponding to the abnormal point.
[0069] Specifically, before executing S720, it is necessary to set the reference score corresponding to each type of abnormal question. The correspondence between abnormal questions and reference scores is shown in Table 1 Common Question-Reference Score Correspondence Table.
[0070] Table-1
[0071] Abnormal problems Functional Error Interface Error Compatibility Errors Usability Errors Reference score 40 10 30 20
[0072] The time period from when the query information is sent to when the tracking data is displayed is recorded as the positioning time of the anomaly point. Since the higher the difficulty or the higher the complexity of the anomaly point, the longer the positioning time is generally required, the positioning time of the anomaly point is used as a reference index of the complexity of the anomaly point.
[0073] Different types of abnormal problems have different importance, so their weight coefficients are also different, and the focus of each test is also different. For example, this test is mainly for the backend, so the weight coefficient of the backend type of abnormal problems can be increased, making this test more efficient.
[0074] Optionally, reference data for calculating the abnormal point score and corresponding weight coefficients may be added according to actual test requirements, thereby improving the reliability of the abnormal point score calculation.
[0075] S730: Determine the abnormal point level to which the abnormal point belongs based on the abnormal point score of the abnormal point.
[0076] In this embodiment, the score of the abnormal problem can be calculated according to the abnormal problem type corresponding to the abnormal point, so that the abnormal problem can be graded according to the score of the abnormal problem.
[0077] Abnormal problems can be graded according to the number-level correspondence table in Table-2.
[0078] Table-2
[0079] Fraction 0-25 26-50 51-75 76-100 grade Low Priority Normal level High priority Immediate solution
[0080] In one embodiment of the present application, the method for quickly locating an abnormal point further includes:
[0081] S801, creating a hierarchical model.
[0082] S802, obtaining multiple log data with abnormal points.
[0083] S803: For each piece of log data with abnormal points, determine the abnormal point level to which the abnormal points in each piece of log data with abnormal points belong, as the abnormal point level of each piece of log data with abnormal points.
[0084] S804: Divide the multiple log data into training sets and test sets in random proportions.
[0085] S805: Randomly select a set of log data and the abnormal point level of the log data from the training set.
[0086] S806: Create a log vector and an abnormal point vector based on the selected log data and the abnormal point level of the selected log data.
[0087] Specifically, word2vec is set to represent the log vector and the abnormal point vector, that is, word2vec = {word1, word2, ..., wordn}, where n is the total amount of selected log data.
[0088] Word2vec is a family of related models used to generate word embeddings. These models are shallow, two-layer neural networks that are trained to reconstruct linguistic word texts.
[0089] S806, calculating the TF-IDF value of the log vector and the abnormal point vector, and using the TF-IDF value as the TF-IDF value of the selected log data.
[0090] Specifically, TF-IDF is a commonly used weighting technique for information retrieval and data mining. TF is term frequency, and IDF is inverse document frequency index.
[0091] TF-IDF is a statistical method used to assess the importance of a word to a document in a collection of documents or a corpus. The importance of a word increases in direct proportion to the number of times it appears in a document, but decreases in inverse proportion to the frequency of its occurrence in the corpus. Various forms of TF-IDF weighting are often used by search engines as a measure or rating of the relevance of a document to a user query.
[0092] In this embodiment, since similar or identical texts may appear in the logs of abnormal problems, the degree of association between specific words and classification is calculated through TF-IDF, so that the classification model can directly classify abnormal tasks through the text information of the request task.
[0093] For example, there are 200 log data, of which 136 log data contain the text information 000200100011A, and the abnormal problem levels of these 136 log data are all high priority, and the text information of 58 log data contains the text information 010100100010C, and the abnormal problem levels of these 58 log data are all normal levels. Then, by calculating the TF-IDF value between 000200100011A and the high priority, and calculating the TF-IDF value between 010100100010C and the normal level, it can be considered that the request task containing the text information 000200100011A is more likely to have high priority abnormal problems, and the request task containing the text information 010100100010C is more likely to have normal level abnormal problems.
[0094] S807, returning the randomly selected log data from the training set and the abnormal point level of the log data, until the TF-IDF value of each log data is obtained.
[0095] S808, clustering is performed based on each log data and the TF-IDF value of each log data to obtain a coupling relationship between each log data and the abnormal point level of the log data.
[0096] Specifically, the log data and the TF-IDF values of the log data are clustered by K-means clustering.
[0097] S809, input each log data, the abnormal point level of each log data, and the coupling relationship between each log data and the abnormal point level of the log data as training data into the classification model, train the classification model, and obtain the trained classification model.
[0098] In this embodiment, a grading model is created and trained so that the trained grading model can directly determine the grade of the abnormal problem corresponding to the request task according to the input text information of the request task.
[0099] In one embodiment of the present application, the method for quickly locating an abnormal point further includes:
[0100] S901, setting a pass rate threshold.
[0101] Specifically, the higher the pass rate threshold is set, the higher the classification accuracy of the trained classification model is.
[0102] S902, selecting a set of log data and the abnormal point level of the log data from the test set.
[0103] S903: Input the selected log data into the trained classification model, obtain the abnormal point level of the log data output by the trained classification model, and use the abnormal point level as the predicted abnormal point level of the selected log data.
[0104] S904, returning to S902, obtaining the predicted abnormal point level of each log data.
[0105] S905, comparing the abnormal point level of each log data in the test set with the predicted abnormal point level, and judging whether the qualified rate of the predicted abnormal point level of the test set as a whole reaches the qualified rate threshold.
[0106] S906: If the pass rate of the predicted abnormal point level of the entire test set reaches the pass rate threshold, it is determined that the trained classification model is qualified.
[0107] S907, if the pass rate of the predicted abnormal point level of the entire test set does not reach the pass rate threshold, return to the step of obtaining multiple log data with abnormal points, and increase the number of log data with abnormal points.
[0108] In this embodiment, since the training effect of the grading model cannot be determined after the training of the grading model is completed, it is necessary to set a pass rate threshold to determine whether the trained grading model has the ability to determine the level of the abnormal problem corresponding to the request task based on the text information of the input request task.
[0109] If the pass rate threshold is set to 80%, and the test set contains 100 log data, each log data has its corresponding abnormal problem level.
[0110] Input these 100 log data into the trained grading model to obtain the predicted abnormal point level of each log data output by the trained grading model, and compare the predicted abnormal point level of each log data with the abnormal problem level of each log data to determine whether the predicted abnormal point level of each log data is at the same level as the abnormal problem level of each log data. Finally, calculate whether the overall pass rate of the test set is greater than or equal to 80%. When the overall pass rate of the test set is greater than or equal to 80%, the trained grading model is considered qualified.
[0111] In one embodiment of the present application, the S905 includes:
[0112] S905A, set the qualified number N.
[0113] Specifically, the initial qualified number n=0.
[0114] S905B, select a log data from the test set.
[0115] S905C, obtaining the log data abnormal point level corresponding to the log data and the predicted abnormal point level of the log data.
[0116] S905D, determining whether the log data abnormal point level corresponding to the log data and the predicted abnormal point level of the log data are at the same level.
[0117] S905E, if the log data anomaly point level corresponding to the log data and the predicted anomaly point level of the log data are at the same level, then the qualified number is increased by 1, and a log data is selected from the test set, until all the log data in the test set are selected.
[0118] Specifically, when the abnormal point level of the log data corresponding to the first log data and the predicted abnormal point level of the log data are at the same level, the qualified number N=n+1.
[0119] S905F, if the log data abnormal point level corresponding to the log data and the predicted abnormal point level of the log data are not at the same level, then directly return to select a log data from the test set until all the log data in the test set are selected.
[0120] S905G, the pass rate of the test set is calculated using Formula 2.
[0121]
[0122] Among them, Q is the pass rate of the test set, N is the number of qualified log data in the test set, and Z is the total number of log data in the test set.
[0123] S905H, determine whether the pass rate of the test set is greater than or equal to the pass rate threshold.
[0124] In this embodiment, the consistency between the predicted abnormal point level of each log data in the test set and the abnormal point level of the log data is determined, and the number of all qualified log data in the test set is counted to determine whether the trained classification model is qualified.
[0125] The more log data there are in the test set, the higher the pass rate threshold is set, and the higher the requirement for the training effect of the classification model is, so that the classification results output by the trained classification model are more reliable.
[0126] If the pass rate of the test set is greater than or equal to the pass rate threshold, it is determined that the pass rate of the predicted abnormal point level of the entire test set reaches the pass rate threshold, and S906 is subsequently executed.
[0127] If the pass rate of the test set is less than the pass rate threshold, the pass rate of the predicted abnormal point level of the entire test set does not reach the pass rate threshold, and S907 is subsequently executed.
[0128] like Figure 2 As shown, in one embodiment of the present application, a system for rapid positioning of abnormal points is also provided, and the system for rapid positioning of abnormal points includes a client 100 and a server 200.
[0129] The request information is sent through the client 100. The server 200 is in communication connection with the client 100, and the server 200 includes a storage module 201, and the storage module 201 is used to store log data, and the request information sent by the client 100 is received through the server 200, and the abnormal point rapid positioning method mentioned in any one of the aforementioned embodiments is executed.
[0130] In this embodiment, the user sends various task requests to the server 200 through the client 100. After the server 200 receives the task request sent by the client 100, it will execute the abnormal point rapid positioning method described in the above embodiment for the received task request, thereby assisting the user to quickly locate the abnormal points in the software and classify the abnormal points.
[0131] The technical features of the above-described embodiments may be arbitrarily combined, and the execution order of the method steps is not limited. To make the description concise, not all possible combinations of the technical features in the above-described embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.
[0132] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be construed as limiting the scope of the present application. It should be noted that, for a person of ordinary skill in the art, several modifications and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.
Claims
1. A method for quickly locating abnormal points, characterized in that: The method for quickly locating abnormal points includes: Receive task requests sent by the client; Create a TraceID, and create a link tracking path by combining the request information and TraceID contained in the client's task request; Execute a data calling action corresponding to the task request of the client, and when executing the data calling action corresponding to the task request of the client, collect the data called by the data calling action by using the link tracking path; The data called by the data calling action is used as the tracking data, and the tracking data is stored in the positioning unit corresponding to the TraceID; Obtain the query request sent by the client, and read the query information contained in the query request; Read the TraceID in the query information, and filter out the log data corresponding to the TraceID in the query information; The abnormal points are located according to the log data, and the abnormal points are graded based on the abnormal point location results of the log data.
2. The method for rapid positioning of abnormal points according to claim 1, characterized in that: The step of creating a TraceID, combining the client's request information and the TraceID to create a link tracking path, includes: Read the request information in the task request; Add TraceID to the request header of the request information and the log data.
3. The method for rapidly locating abnormal points according to claim 2, characterized in that: The adding of TraceID in the request header of the request information and in the log data includes: Determine whether the request header of the request information contains TraceID; If the request header of the request information does not contain TraceID, read the TraceID generation rule, create a TraceID according to the TraceID generation rule, and add the TraceID to the request header of the request information sent by the client; If the request header of the request information contains TraceID, the TraceID is extracted; the extracted TraceID is recorded as the target TraceID; Add the target TraceID to each log.
4. The method for rapid positioning of abnormal points according to claim 3 is characterized in that: The step of reading the TraceID in the query information and filtering out the log data corresponding to the TraceID in the query information includes: Get the TraceID contained in the request header of the query information; Using the TraceID contained in the request header of the query information as an index, query the tracing data corresponding to the TraceID in the positioning unit; Get the log data in the trace data.
5. The method for rapid positioning of abnormal points according to claim 4, characterized in that: The obtaining of the query information of the client and filtering out the log data corresponding to the TraceID of the query information of the client also includes: Display tracking data and locate abnormal points based on the log data in the tracking data.
6. The method for rapid positioning of abnormal points according to claim 5, characterized in that: The abnormal point location result combined with the log data is used to classify the abnormal point, including: Obtain the abnormal point location result corresponding to the log data and obtain the abnormal point; The abnormal point score of the abnormal point is calculated by formula 1; P = A × M + B × N Formula 1; Among them, P is the abnormal point score of the abnormal point, A is the location time of the abnormal point, B is the reference score of the type of abnormal problem corresponding to the abnormal point, M is the weight coefficient of the location time of the abnormal point, and N is the weight coefficient of the type of abnormal problem corresponding to the abnormal point; Based on the abnormal point score of the abnormal point, the abnormal point level to which the abnormal point belongs is determined.
7. The method for rapid positioning of abnormal points according to claim 6, characterized in that: Before receiving the task request sent by the client, the method for quickly locating an abnormal point further includes: Create hierarchical models; Obtain multiple log data with abnormal points; For each piece of log data with abnormal points, determine the abnormal point level to which the abnormal points in each piece of log data with abnormal points belong, as the abnormal point level of each piece of log data with abnormal points; Divide multiple log data into training sets and test sets in random proportions; Randomly select a piece of log data and the abnormal point level of the log data from the training set; Creating a log vector and an abnormal point vector based on the selected log data and the abnormal point level of the selected log data; Calculate the TF-IDF value of the log vector and the abnormal point vector, and use the TF-IDF value as the TF-IDF value of the selected log data; Return the randomly selected log data from the training set and the abnormal point level of the log data until the TF-IDF value of each log data is obtained; Clustering is performed based on each log data and the TF-IDF value of each log data to obtain the coupling relationship between each log data and the abnormal point level of the log data; Each log data, the abnormal point level of each log data, and the coupling relationship between each log data and the abnormal point level of the log data are input into the classification model as training data, and the classification model is trained to obtain a trained classification model.
8. The method for rapid positioning of abnormal points according to claim 7, characterized in that: Before receiving the task request sent by the client, the method for quickly locating an abnormal point further includes: Set pass rate threshold; Select a piece of log data and the abnormal point level of the log data from the test set; Input the selected log data into the trained classification model, obtain the abnormal point level of the log data output by the trained classification model, and use the abnormal point level as the predicted abnormal point level of the selected log data; Return a log data and the abnormal point level of the log data selected from the test set, and obtain the predicted abnormal point level of each log data; The abnormal point level of each log data in the test set is compared with the predicted abnormal point level to determine whether the qualified rate of the predicted abnormal point level of the test set as a whole reaches the qualified rate threshold; If the pass rate of the predicted abnormal point level of the test set as a whole reaches the pass rate threshold, the trained classification model is determined to be qualified; If the pass rate of the predicted abnormal point level of the entire test set does not reach the pass rate threshold, the step of obtaining multiple log data with abnormal points is returned, and the number of log data with abnormal points is increased.
9. The method for rapid positioning of abnormal points according to claim 8, characterized in that: The abnormal point level of each log data in the statistical test set is compared with the predicted abnormal point level to determine whether the qualified rate of the predicted abnormal point level of the test set as a whole reaches the qualified rate threshold, including: Set the qualified number N; Select a log data from the test set; Obtain the log data abnormal point level and the log data predicted abnormal point level corresponding to the log data; Determine whether the log data abnormal point level corresponding to the log data and the predicted abnormal point level of the log data are at the same level; If the log data anomaly point level corresponding to the log data and the predicted anomaly point level of the log data are at the same level, the qualified number is increased by 1, and a log data is selected from the test set until all the log data in the test set are selected; If the log data anomaly point level corresponding to the log data and the predicted anomaly point level of the log data are not at the same level, then directly return to select a log data from the test set until all the log data in the test set are selected; The pass rate of the test set is calculated by formula 2; Among them, Q is the pass rate of the test set, N is the number of qualified log data in the test set, and Z is the total number of log data in the test set; Determine whether the pass rate of the test set is greater than or equal to the pass rate threshold.
10. A system for rapid positioning of abnormal points, characterized in that: The abnormal point rapid positioning system comprises: Client; sending request information through the client; A server, the server is communicatively connected to the client, the server includes a storage module, the storage module is used to store log data, the request information sent by the client is received through the server, and the method for rapid positioning of abnormal points as described in any one of claims 1 to 9 is executed.