An analysis method and device of a performance test report

Abstract

The application relates to a performance test report analysis method and device, wherein the method classifies a performance test report to be analyzed, filters test data of the performance test report to be analyzed according to historical performance test reports of the same category screened from a database, generates filtered test data, inputs the filtered test data into a training model to calculate a training value, analyzes the performance test report to be analyzed based on the training value, and generates an analysis result. Thus, the application filters the test data of the performance test report to be analyzed according to the test data of the historical performance test reports of the same category, improves the accuracy of the data, avoids the influence of data with large differences or errors on the analysis result, analyzes the performance test report to be analyzed according to the training value calculated by the training model, avoids the interference of human factors, and improves the analysis quality and efficiency of the performance test report.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to a method and apparatus for analyzing performance test reports. Background Technology

[0002] With the rapid development of the internet, software systems have been applied to enterprises across various industries. The performance of these systems directly impacts the profits of businesses, making system performance testing and analysis increasingly important in the industry. Current technology involves testers using performance testing tools to conduct tests, which automatically generate test reports. Testers then analyze these reports to arrive at the final performance analysis results. However, this analysis is done manually, relying entirely on personal experience, which is not only inefficient but also affects the accuracy of the analysis. Summary of the Invention

[0003] The technical problem to be solved by the present invention is: to provide a method and apparatus for analyzing performance test reports, thereby improving the efficiency and accuracy of performance test report analysis.

[0004] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows:

[0005] In a first aspect, the present invention provides a method for analyzing performance test reports, comprising:

[0006] Receive the performance test reports to be analyzed, classify the performance test reports to be analyzed, and obtain the category of the performance test reports to be analyzed;

[0007] Based on the category of the performance test report to be analyzed, historical performance test reports of the same category are filtered from the database, and historical test data of the historical performance test reports are obtained. Based on the historical test data, the test data of the performance test report to be analyzed is filtered to generate filtered test data.

[0008] The filtered test data is input into the training model to calculate training values. Based on the training values, the performance test report to be analyzed is analyzed to generate analysis results.

[0009] The beneficial effects of this invention are that by classifying the performance test reports to be analyzed, historical performance test reports of the same category can be obtained. The test data of the performance test reports to be analyzed can be filtered using the historical test data of the historical performance test reports, which can improve the accuracy of the data and avoid the impact of large discrepancies or erroneous data on the analysis results. Furthermore, by training the model to calculate the training value of the filtered test data, the performance test reports to be analyzed can be analyzed based on the training value, avoiding interference from human factors and improving the quality and efficiency of performance test report analysis.

[0010] Optionally, receiving the performance test report to be analyzed, classifying the performance test report to be analyzed, and obtaining the category of the performance test report to be analyzed includes:

[0011] Receive performance test reports to be analyzed and retrieve historical performance test reports of known categories from the database;

[0012] The report features of the performance test report to be analyzed and the corresponding report data to be analyzed are obtained. Based on the report features, the historical report data corresponding to the historical performance test report is obtained. The similarity between the report data to be analyzed and the historical report data is calculated using a preset similarity algorithm. The similarity is compared with a first threshold. The historical performance test reports with similarity exceeding the first threshold are used to form a first performance test report sample set.

[0013] Obtain the categories of each performance test report in the first performance test report sample set and count the total number of each category. Take the category with the largest total number of categories as the first main category. Classify the performance test reports to be analyzed according to the first main category to obtain the category of the performance test reports to be analyzed.

[0014] As described above, by calculating the similarity between the performance test report to be analyzed and historical performance test reports of known categories in the database, and statistically analyzing the categories of historical performance test reports with similarity exceeding a first threshold, the category with the largest total number of categories is used as the category of the performance test report to be analyzed, thereby improving the rationality of classifying the performance test report to be analyzed.

[0015] Optionally, calculating the similarity between the report data to be analyzed and the historical report data using a preset similarity algorithm includes:

[0016] Determine whether the type of the report data to be analyzed is a formatted type. If so, calculate the similarity value between the report data to be analyzed and the historical report data using the TF-IDF algorithm. Otherwise, determine whether the type of the report data to be analyzed is an unformatted type. If so, calculate the similarity value between the report data to be analyzed and the historical report data using the SimHash algorithm. Otherwise, determine whether the type of the report data to be analyzed is an image type. If so, calculate the similarity value between the report data to be analyzed and the historical report data using the perceptual hash algorithm.

[0017] Based on the reported features, obtain the corresponding feature weights, and combine the feature weights with the similarity values ​​to generate a similarity score.

[0018] As described above, the similarity between the data in the performance test report to be analyzed and the historical report data is composed of the similarity value of the data in the report to be analyzed and the corresponding report feature weights. Moreover, the similarity value calculation method is different for different types of data in the report to be analyzed. The similarity value calculation comprehensively considers the formatted type, unformatted type and image type of report data to improve the accuracy of the similarity value. In addition, the addition of report feature weights improves the accuracy of the similarity, so as to classify the performance test report to be analyzed more accurately.

[0019] Optionally, the category with the largest total number of categories is taken as the first main category, and the performance test reports to be analyzed are classified according to the first main category. The categories of the performance test reports to be analyzed include:

[0020] When there are multiple first main categories, the similarity is compared with the second threshold, and historical performance test reports with similarity exceeding the second threshold are used to form a second performance test report sample set.

[0021] Obtain the categories of each performance test report in the second performance test report sample set and count the total number of each category. Take the category with the largest total number of categories as the second main category. Classify the performance test reports to be analyzed according to the second main category to obtain the category of the performance test reports to be analyzed.

[0022] As described above, when there are multiple primary categories (i.e., multiple categories with the largest total number of categories) and the performance test reports to be analyzed cannot be classified, a second performance test report sample set with different ranges will be regenerated, and the category of the performance test reports in the second performance test report sample set will be recalculated. This will be used to classify the performance test reports to be analyzed. The possible scenarios during classification have been fully considered, improving the rigor and comprehensiveness of the classification.

[0023] Optionally, the step of filtering historical performance test reports of the same category from the database according to the category of the performance test report to be analyzed, obtaining historical test data of the historical performance test reports, and filtering the test data of the performance test report to be analyzed based on the historical test data to generate filtered test data includes:

[0024] Based on the category of the performance test report to be analyzed, a predetermined number of historical performance test reports of the same category are selected from the database in chronological order, and the historical test data of the historical performance test reports are obtained.

[0025] The historical test data is sorted according to a preset sorting rule to generate a historical test data range. Then, the test data in the performance test report to be analyzed that exceeds the historical test data range is filtered according to a preset filtering rule to generate filtered test data.

[0026] As described above, when filtering historical performance test reports according to the category of the performance test reports to be analyzed, the filtering is not only done in chronological order but also selects a predetermined number of historical performance test reports to ensure the timeliness and diversity of the historical performance test reports. This filters the test data of the performance test reports to be analyzed, effectively avoiding situations where false alarms occur in the test data of the performance test reports to be analyzed or where data differences are too large due to unforeseen circumstances, which may affect the analysis results and improve the accuracy of the analysis.

[0027] Optionally, generating filtered test data includes:

[0028] The filtered test data and the historical test data are merged to form a test dataset. The test dataset is then trained using a preset throughput formula to calculate the throughput-related parameters corresponding to the test dataset.

[0029] The throughput-related parameters are substituted into the preset throughput formula to generate a training model.

[0030] As described above, based on filtered test data and historical test data, a test dataset is constructed, and relevant parameters that conform to the throughput formula of the test dataset are calculated. This generates a training model that conforms to the performance test report to be analyzed, thereby improving the accuracy of training.

[0031] Optionally, the step of inputting the filtered test data into the training model to calculate training values, and analyzing the performance test report to be analyzed based on the training values ​​to generate analysis results includes:

[0032] The filtered test data is input into the training model to calculate the training value, and the training value is used as the benchmark value of the performance test data to be analyzed.

[0033] Obtain the throughput of the performance test report to be analyzed, compare the throughput of the performance test report to be analyzed with the benchmark value, and if the throughput of the performance test report to be analyzed is lower than the benchmark value, the performance test report to be analyzed is unqualified and an unqualified analysis result is generated; otherwise, the performance test report to be analyzed is qualified and a qualified analysis result is generated.

[0034] As described above, the filtered test data is input into the training model to calculate the training value, i.e., the throughput value. The throughput value of the performance test report to be analyzed is compared with the throughput value calculated by the training model to analyze whether the performance test report to be analyzed is qualified, thereby improving the rationality of the analysis and judgment.

[0035] Optionally, if the performance test report to be analyzed is unqualified, generating unqualified analysis results includes:

[0036] The test data of the performance test report to be analyzed is obtained and compared with the test threshold. The test data exceeding the test threshold is used to form the first positioning data.

[0037] The request response time of the first location data is obtained, and the request log and CPU monitoring data are located based on the request response time. The server node with the performance bottleneck is located based on the request log and CPU monitoring data.

[0038] Based on the service deployment list, the middleware that causes the performance bottleneck is located according to the server node of the performance bottleneck, and the CPU monitoring data of the corresponding system is matched according to the middleware to generate abnormal data. The abnormal data is then incorporated into the unqualified analysis results.

[0039] As described above, when the analysis result of the performance test report is unqualified, the middleware that is the performance bottleneck will be located, and the abnormal data will be obtained by matching the system operation CPU monitoring data of the corresponding middleware. The abnormal data will be incorporated into the analysis result to improve the completeness of the analysis result and provide a reference for subsequent analysis by maintenance personnel.

[0040] Optionally, incorporating the abnormal data into the unqualified analysis results includes:

[0041] Based on the category of the performance test report to be analyzed, the middleware of the performance bottleneck, and the abnormal data, corresponding optimization suggestions are selected from the database.

[0042] As described above, the database stores previous optimization suggestions. When a performance test report to be analyzed fails, the database will select suitable optimization suggestions for operation and maintenance personnel to refer to, thereby improving work efficiency.

[0043] In a second aspect, the present invention provides a performance test report analysis apparatus, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the performance test report analysis method described in the first aspect.

[0044] The technical effects of the performance test report analysis device provided in the second aspect are described in reference to the performance test report analysis method provided in the first aspect. Attached Figure Description

[0045] Figure 1 A flowchart illustrating a method for analyzing a performance test report provided in an embodiment of the present invention;

[0046] Figure 2 A flowchart illustrating a performance test report analysis method provided in an embodiment of the present invention.

[0047] Figure 3 A schematic diagram of the structure of a performance test report analysis device provided in an embodiment of the present invention.

[0048] [Explanation of Labels in the Attached Image]

[0049] 1. An analysis device for performance test reports;

[0050] 2. Processor;

[0051] 3. Memory. Detailed Implementation

[0052] To better understand the above technical solutions, exemplary embodiments of the present invention will be described in more detail below with reference to the accompanying drawings. Although exemplary embodiments of the present invention are shown in the drawings, it should be understood that the present invention can be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that the present invention can be understood more clearly and thoroughly, and that the scope of the present invention can be fully conveyed to those skilled in the art.

[0053] Example 1

[0054] Please refer to Figures 1 to 2 This invention provides a method for analyzing performance test reports, including the following steps:

[0055] S1. Receive the performance test report to be analyzed, classify the performance test report to be analyzed, and obtain the category of the performance test report to be analyzed;

[0056] In this embodiment, as Figure 2 As shown, after receiving the performance test report to be analyzed, the system will categorize the report to obtain its category.

[0057] At this point, step S1 includes:

[0058] S11. Receive the performance test report to be analyzed and obtain historical performance test reports of known categories from the database;

[0059] S12. Obtain the report features of the performance test report to be analyzed and the report data to be analyzed corresponding to the report features. Obtain the historical report data corresponding to the historical performance test report according to the report features. Calculate the similarity between the report data to be analyzed and the historical report data through a preset similarity algorithm. Compare the similarity with a first threshold. Form a first performance test report sample set from the historical performance test reports whose similarity exceeds the first threshold.

[0060] In this embodiment, as Figure 2 As shown, while receiving the performance test report to be analyzed, the system also retrieves historical performance test reports of known categories already stored in the database. These historical reports are used as the basis for classifying the performance test reports to be analyzed. The system obtains the report features of the performance test report to be analyzed and the corresponding report data. In this application, the report features include, but are not limited to: project name, project type, use case name, interface information, configuration resource information, test images, operation steps, and test result data. The system calculates the similarity between the report data to be analyzed corresponding to these report features and the historical report data of the same report features of the historical performance test reports of known categories. The calculated similarity is compared with a first threshold. Historical performance test reports with a similarity exceeding the first threshold form the first performance test report sample set. At this time, the first threshold is set to 85%, that is, historical performance test reports with a similarity exceeding 85% form the first performance test report sample set. The value of the threshold can be set according to the actual situation.

[0061] At this point, step S12, which involves calculating the similarity between the report data to be analyzed and the historical report data using a preset similarity algorithm, includes:

[0062] S121. Determine whether the type of the report data to be analyzed is a formatted type. If yes, calculate the similarity value between the report data to be analyzed and the historical report data using the TF-IDF algorithm. Otherwise, determine whether the type of the report data to be analyzed is an unformatted type. If yes, calculate the similarity value between the report data to be analyzed and the historical report data using the SimHash algorithm. Otherwise, determine whether the type of the report data to be analyzed is an image type. If yes, calculate the similarity value between the report data to be analyzed and the historical report data using the perceptual hash algorithm.

[0063] In this embodiment, as Figure 2As shown, when calculating the similarity between the report data to be analyzed and the historical report data using a preset similarity algorithm, different similarity algorithms are used depending on the type of the report data to be analyzed. When the type of the report data to be analyzed is formatted, the TF-IDF algorithm is used to perform full-text matching and retrieval of the report data to be analyzed and the historical report data, and scores are given based on the frequency of occurrence of fields in the data to calculate the similarity value. For the unformatted type of report data to be analyzed, the SimHash algorithm is used to convert both the report data to be analyzed and the historical report data into 64-bit bytes, and the similarity value is obtained by calculating the distance between two bytes. When the type of the report data to be analyzed is an image, the perceptual hash algorithm is used to generate corresponding image fingerprint strings for the image of the report data to be analyzed and the image of the historical report data, and the similarity value is calculated by comparing whether the image fingerprint strings are consistent.

[0064] S122. Obtain the corresponding feature weights based on the reported features, and combine the feature weights with the similarity values ​​to generate a similarity score.

[0065] In this embodiment, as Figure 2 As shown, after calculating the similarity value between the report data to be analyzed and the historical report data in step S121, the corresponding report feature weights will be obtained. The weights and similarity values ​​will be combined and the similarity value will be calculated using the preset similarity formula: Similarity = Weight 1 * Similarity value 1 + Weight 2 * Similarity value 2 + Weight 3 * Similarity value 3 + Weight N * Similarity value N.

[0066] In one specific embodiment, the features of the performance test report to be analyzed include: project name, project type, test case name, interface information, configuration resource information, test images, operation steps, and test result data. The weights corresponding to the report features are 5%, 5%, 5%, 15%, 5%, 10%, 25%, and 30%, respectively. The similarity values ​​of the report data corresponding to the report features are 70%, 80%, 75%, 86%, 88%, 77%, 60%, 90%, and 95%, respectively. Therefore, the similarity is 5%*70% + 5%*80% + 5%*75% + 15%*86% + 5%*88% + 10%*77% + 25%*90% + 30%*95% = 87.25%, which means the similarity is 87.25%.

[0067] S13. Obtain the categories of each performance test report in the first performance test report sample set and count the total number of each category. Take the category with the largest total number of categories as the first main category. Classify the performance test reports to be analyzed according to the first main category to obtain the category of the performance test reports to be analyzed.

[0068] In this embodiment, as Figure 2As shown, after generating the first performance test report sample set in step S12, the categories of each performance test report in the first performance test report sample set will be counted, and the total number of the same category will be calculated. The category with the largest total number of categories will be used as the category of the performance test report to be analyzed, thereby realizing the classification of the performance test reports to be analyzed.

[0069] In one specific embodiment, the known categories of historical performance test reports in the first performance test report sample set are: historical 1, historical 2, historical 3, and historical 4. Historical 1 belongs to category A, historical 2 belongs to category B, historical 3 belongs to category A, and historical 4 belongs to category A. Therefore, category A has the largest number of reports, so category A is used as the category of the performance test reports to be analyzed.

[0070] At this point, step S13, which involves selecting the category with the largest total number of categories as the first primary category and classifying the performance test report to be analyzed according to the first primary category, includes obtaining the categories of the performance test report to be analyzed:

[0071] S131. When there are multiple first main categories, the similarity is compared with the second threshold, and the historical performance test reports with similarity exceeding the second threshold are used to form a second performance test report sample set.

[0072] S132. Obtain the categories of each performance test report in the second performance test report sample set and count the total number of each category. Take the category with the largest total number of categories as the second main category. Classify the performance test reports to be analyzed according to the second main category to obtain the category of the performance test reports to be analyzed.

[0073] In this embodiment, when multiple primary categories exist (i.e., when multiple categories with the highest total number of categories appear in the first performance test report sample set), the performance test reports to be analyzed cannot be classified. In this case, the first performance test report sample set needs to be re-divided, that is, the similarity is compared with the second threshold. Historical performance test reports with similarity exceeding the second threshold are combined into the second performance test report sample set. The second threshold can be set to be larger or smaller than the first threshold. When the second threshold is smaller than the first threshold, the scope of the first performance test report sample set is narrowed to form the second performance test report sample set. When the second threshold is larger than the first threshold, the scope of the first performance test report sample set is expanded to form the second performance test report sample set. In this case, the second threshold is set to 90%, that is, historical performance test reports with similarity exceeding 90% are combined into the second performance test report sample set. The value of the second threshold can be set according to the actual situation.

[0074] S2. Based on the category of the performance test report to be analyzed, filter out historical performance test reports of the same category from the database, obtain the historical test data of the historical performance test reports, filter the test data of the performance test report to be analyzed based on the historical test data, and generate filtered test data.

[0075] In this embodiment, after the performance test reports to be analyzed are classified, historical performance test reports of the same category are filtered from the database according to the category of the performance test reports to be analyzed, and historical test data of the historical performance test reports of the same category are obtained. The test data of the performance test reports to be analyzed is then filtered using the historical test data.

[0076] At this point, step S2 includes:

[0077] S21. Based on the category of the performance test report to be analyzed, filter out a predetermined number of historical performance test reports of the same category from the database in chronological order, and obtain the historical test data of the historical performance test reports;

[0078] In this embodiment, as Figure 2 As shown, when filtering historical performance test reports of the same category according to the category of the performance test report to be analyzed, a predetermined number of historical performance test reports will be filtered in chronological order to ensure the timeliness and diversity of the historical test data of the obtained historical performance test reports. In this application, the test data includes, but is not limited to: number of threads, error rate, average response time, minimum response time, maximum response time, 90% request response time, 95% request response time, 99% request response time, throughput, CPU utilization, memory utilization, etc.

[0079] S22. The historical test data is sorted according to a preset sorting rule to generate a historical test data range, and the test data in the performance test report to be analyzed that exceeds the historical test data range is filtered according to a preset filtering rule to generate filtered test data.

[0080] In this embodiment, as Figure 2 As shown, the historical test data obtained in step S21 will be sorted according to the preset sorting rules: sorted from smallest to largest according to "90% request response time" to obtain the range of historical test data. Then, the test data to be analyzed in the performance test report will be filtered by the preset filtering rules: 10% of the test data smaller than the minimum value of the historical test data will be removed, and 10% of the test data larger than the maximum value of the historical test data will be removed. The preset sorting rules and preset filtering rules can be set according to the actual situation.

[0081] In one specific embodiment, 10 historical performance test reports correspond to 10 historical performance test data sets. The 90% request response time of each historical performance test data set is obtained and sorted from smallest to largest. The sorted historical performance test data sets are: 15, 18, 23, 25, 30, 45. 10% of the test data sets smaller than the minimum historical test data set are removed. If there are 30 test data sets smaller than the minimum historical test data set in the performance test reports to be analyzed, then 30 * 10% = 3, meaning 3 test data sets smaller than the minimum historical test data set are removed from the test data in the performance test reports to be analyzed. If there are 40 test data sets larger than the maximum historical test data set, then 40 * 10% = 4, meaning 4 test data sets larger than the maximum historical test data set are removed from the test data in the performance test reports to be analyzed.

[0082] At this point, generating the filtered test data in step S22 includes:

[0083] S221. Merge the filtered test data and the historical test data to form a test dataset, train the test dataset using a preset throughput formula, and calculate the throughput-related parameters corresponding to the test dataset.

[0084] S222. Substitute the throughput-related parameters into the preset throughput formula to generate a training model.

[0085] In this embodiment, as Figure 2 As shown, the filtered test data generated in step S22 is merged with historical test data to form a test dataset. The test dataset is then input into a preset throughput formula: Throughput = a0 + a1 * 90% request response time + a2 * 95% request response time + a3 * 99% request response time + a4 * CPU utilization + a5 * memory utilization + a6 * average response time. Since each test data has its own 90% request response time, 95% request response time, 99% request response time, CPU utilization, memory utilization, and average response time, inputting each test data into the throughput formula can calculate the throughput-related parameters, namely a0, a1, a2, a3, a4, a5, and a6. Substituting the throughput-related parameters into the preset throughput formula establishes a training model, thereby generating a throughput formula, i.e., a training model, that conforms to the test data in the performance test report to be analyzed.

[0086] S3. Input the filtered test data into the training model to calculate the training value, and analyze the performance test report to be analyzed based on the training value to generate the analysis result.

[0087] In this embodiment, as Figure 2As shown, after the training model is formed, the filtered test data will be input into the training model to calculate the training value, which is the estimated value of the performance test report to be analyzed. The performance test report to be analyzed will then be analyzed to generate the analysis results.

[0088] At this point, step S3 includes:

[0089] S31. Input the filtered test data into the training model to calculate the training value, and use the training value as the benchmark value of the performance test data to be analyzed.

[0090] S32. Obtain the throughput of the performance test report to be analyzed, compare the throughput of the test report to be analyzed with the benchmark value, and if the throughput of the performance test report to be analyzed is lower than the benchmark value, the performance test report to be analyzed is unqualified and an unqualified analysis result is generated; otherwise, the performance test report to be analyzed is qualified and a qualified analysis result is generated.

[0091] In this embodiment, as Figure 2 As shown, the filtered test data is input into the training model to calculate the corresponding throughput value. The calculated throughput value is used as the benchmark value for the performance test report to be analyzed. The throughput of the performance test report to be analyzed is obtained and compared with the benchmark value. If the throughput is lower than the benchmark value, it means that the performance test report to be analyzed has not met the expected performance requirements, and an unqualified analysis result is generated. If the throughput is higher than or equal to the benchmark value, it means that the performance test report to be analyzed has met the expected performance requirements, and a qualified analysis result is generated.

[0092] At this point, the performance test report to be analyzed described in step S32 is unqualified, and the generated unqualified analysis results include:

[0093] S321. Obtain the test data of the performance test report to be analyzed and compare the test data with the test threshold, and form the test data exceeding the test threshold into the first positioning data;

[0094] In this embodiment, as Figure 2 As shown, when the performance test report to be analyzed is unqualified, the test data of the performance test report to be analyzed will be obtained and compared with the test thresholds. For example, the average request response time and CPU utilization in the test data will be obtained and compared with the preset average request response time test threshold, and the CPU utilization will be compared with the preset CPU utilization test threshold. The specific test data to be obtained and the preset test threshold values ​​can be selected and set according to the actual situation. Exceeding the test threshold indicates that there is a bottleneck. Therefore, the test data that exceeds the test threshold is composed of the first positioning data.

[0095] S322. Obtain the request response time of the first location data and locate the request log and CPU monitoring data according to the request response time, and locate the server node of the performance bottleneck according to the request log and CPU monitoring data.

[0096] In this embodiment, as Figure 2 As shown, by obtaining the request response time of the first location data, the request logs and CPU monitoring data corresponding to the request response time can be retrieved from the server cluster, thereby locating the server node with a performance bottleneck.

[0097] S323. Based on the service deployment list, locate the middleware of the performance bottleneck according to the server node of the performance bottleneck, and generate abnormal data by matching the CPU monitoring data of the corresponding system running according to the middleware, and incorporate the abnormal data into the unqualified analysis results.

[0098] In this embodiment, as Figure 2 As shown, after locating the server node with the performance bottleneck through step S322, the middleware with the performance bottleneck is retrieved according to the service deployment list, and the system running CPU monitoring data that matches the middleware is matched to obtain abnormal data. The abnormal data is then incorporated into the unqualified analysis results to improve the analysis results.

[0099] At this point, the step S323 of incorporating the abnormal data into the unqualified analysis results includes:

[0100] S3231. Based on the category of the performance test report to be analyzed, the middleware of the performance bottleneck, and the abnormal data, corresponding optimization suggestions are selected from the database.

[0101] In this embodiment, as Figure 2 As shown, the database stores corresponding optimization suggestions based on the category of the performance test report, the middleware information of the performance bottleneck, and the abnormal data information. Therefore, when the performance test report to be analyzed is unqualified, the database will filter out the corresponding optimization suggestions based on the category of the performance test report to be analyzed, the middleware of the performance bottleneck, and the abnormal data, and provide them to the operation and maintenance personnel for reference.

[0102] Example 2

[0103] Please refer to Figure 3 An analysis device 1 for a performance test report includes a memory 3, a processor 2, and a computer program stored in the memory 3 and executable on the processor 2. When the processor 2 executes the computer program, it implements the steps in Embodiment 1 above.

[0104] Since the systems / devices described in the above embodiments of the present invention are systems / devices used to implement the methods of the above embodiments of the present invention, those skilled in the art can understand the specific structure and modifications of the systems / devices based on the methods described in the above embodiments of the present invention, and therefore will not be repeated here. All systems / devices used in the methods of the above embodiments of the present invention fall within the scope of protection of the present invention.

[0105] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0106] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, as well as combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions.

[0107] It should be noted that any reference numerals placed between parentheses in the claims should not be construed as limiting the claims. The word "comprising" does not exclude the presence of components or steps not listed in the claims. The word "a" or "an" preceding a component does not exclude the presence of a plurality of such components. The invention can be implemented by means of hardware comprising several different components and by means of a suitably programmed computer. In claims that enumerate several means, several of these means may be embodied by the same hardware. The use of the terms first, second, third, etc., is merely for convenience of expression and does not indicate any order. These terms can be understood as part of the component names.

[0108] Furthermore, it should be noted that in the description of this specification, the terms "one embodiment," "some embodiments," "embodiment," "example," "specific example," or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Furthermore, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0109] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the claims should be interpreted to include both the preferred embodiments and all changes and modifications falling within the scope of the invention.

[0110] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, then this invention should also include these modifications and variations.

Claims

1. A method for analyzing performance test reports, characterized in that, include: Receive the performance test reports to be analyzed, classify the performance test reports to be analyzed, and obtain the category of the performance test reports to be analyzed; Based on the category of the performance test report to be analyzed, historical performance test reports of the same category are filtered from the database, and historical test data of the historical performance test reports are obtained. Based on the historical test data, the test data of the performance test report to be analyzed is filtered to generate filtered test data. The filtered test data is input into the training model to calculate training values. Based on the training values, the performance test report to be analyzed is analyzed to generate analysis results. The process of receiving performance test reports to be analyzed, classifying the performance test reports to be analyzed, and obtaining the categories of the performance test reports to be analyzed includes: Receive performance test reports to be analyzed and retrieve historical performance test reports of known categories from the database; The report features of the performance test report to be analyzed and the corresponding report data to be analyzed are obtained. Based on the report features, the historical report data corresponding to the historical performance test report is obtained. The similarity between the report data to be analyzed and the historical report data is calculated using a preset similarity algorithm. The similarity is compared with a first threshold. The historical performance test reports with similarity exceeding the first threshold are used to form a first performance test report sample set. Obtain the categories of each performance test report in the first performance test report sample set and count the total number of each category. Take the category with the largest total number of categories as the first main category. Classify the performance test reports to be analyzed according to the first main category to obtain the category of the performance test reports to be analyzed. The step of filtering historical performance test reports of the same category from the database according to the category of the performance test report to be analyzed, obtaining historical test data of the historical performance test reports, and filtering the test data of the performance test report to be analyzed based on the historical test data to generate filtered test data includes: Based on the category of the performance test report to be analyzed, a predetermined number of historical performance test reports of the same category are selected from the database in chronological order, and the historical test data of the historical performance test reports are obtained. The historical test data is sorted according to a preset sorting rule to generate a historical test data range. The test data in the performance test report to be analyzed that exceeds the historical test data range is filtered according to a preset filtering rule to generate filtered test data. The generated filtered test data includes: The filtered test data and the historical test data are merged to form a test dataset. The test dataset is then trained using a preset throughput formula to calculate the throughput-related parameters corresponding to the test dataset. The throughput-related parameters are substituted into the preset throughput formula to generate a training model.

2. The method for analyzing a performance test report as described in claim 1, characterized in that, The calculation of the similarity between the report data to be analyzed and the historical report data using a preset similarity algorithm includes: Determine whether the type of the report data to be analyzed is a formatted type. If so, calculate the similarity value between the report data to be analyzed and the historical report data using the TF-IDF algorithm. Otherwise, determine whether the type of the report data to be analyzed is an unformatted type. If so, calculate the similarity value between the report data to be analyzed and the historical report data using the SimHash algorithm. Otherwise, determine whether the type of the report data to be analyzed is an image type. If so, calculate the similarity value between the report data to be analyzed and the historical report data using the perceptual hash algorithm. Based on the reported features, obtain the corresponding feature weights, and combine the feature weights with the similarity values ​​to generate a similarity score.

3. The method for analyzing a performance test report as described in claim 1, characterized in that, The category with the largest number of categories is taken as the first main category. The performance test reports to be analyzed are then categorized according to the first main category. The categories of the performance test reports to be analyzed include: When there are multiple first main categories, the similarity is compared with the second threshold, and historical performance test reports with similarity exceeding the second threshold are used to form a second performance test report sample set. Obtain the categories of each performance test report in the second performance test report sample set and count the total number of each category. Take the category with the largest total number of categories as the second main category. Classify the performance test reports to be analyzed according to the second main category to obtain the category of the performance test reports to be analyzed.

4. The method for analyzing a performance test report as described in claim 1, characterized in that, The step of inputting the filtered test data into the training model to calculate training values, and analyzing the performance test report to be analyzed based on the training values ​​to generate analysis results includes: The filtered test data is input into the training model to calculate the training value, and the training value is used as the benchmark value of the performance test report to be analyzed. Obtain the throughput of the performance test report to be analyzed, compare the throughput of the test report to be analyzed with the benchmark value, and if the throughput of the performance test report to be analyzed is lower than the benchmark value, the performance test report to be analyzed is unqualified and an unqualified analysis result is generated; otherwise, the performance test report to be analyzed is qualified and a qualified analysis result is generated.

5. The method for analyzing a performance test report as described in claim 4, characterized in that, The performance test report to be analyzed is unqualified, and the resulting unqualified analysis results include: The test data of the performance test report to be analyzed is obtained and compared with the test threshold. The test data exceeding the test threshold is used to form the first positioning data. The request response time of the first location data is obtained, and the request log and CPU monitoring data are located based on the request response time. The server node with the performance bottleneck is located based on the request log and CPU monitoring data. Based on the service deployment list, the middleware that causes the performance bottleneck is located according to the server node of the performance bottleneck, and the CPU monitoring data of the corresponding system is matched according to the middleware to generate abnormal data. The abnormal data is then incorporated into the unqualified analysis results.

6. The method for analyzing a performance test report as described in claim 5, characterized in that, The inclusion of the abnormal data in the unqualified analysis results includes: Based on the category of the performance test report to be analyzed, the middleware of the performance bottleneck, and the abnormal data, corresponding optimization suggestions are selected from the database.

7. An analysis apparatus for performance test reports, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the method as described in any one of claims 1 to 6.

Images