Server test platform, method, device and equipment and storage medium

Abstract

The embodiment of the invention provides a server test platform, method, device and equipment and a storage medium, an interaction analysis system of the server test platform is used for receiving to-be-tested information input by a user, performing semantic analysis on the to-be-tested information to obtain a semantic analysis result, and sending the semantic analysis result to the server test platform; the semantic analysis result is used for determining a target test to be executed on the server test platform; displaying a test result of the target test in the interactive analysis system according to a target form through the target abnormal mode and the multi-dimensional dynamic association data; the three-dimensional association model is used for extracting multi-dimensional dynamic association data about the target test in the process of executing the target test; identifying a first abnormal mode appearing in the target test through the multi-dimensional dynamic association data; the prediction analysis model is used for predicting a second abnormal mode occurring in a target test corresponding to the multi-dimensional dynamic association data; the knowledge management system is used for determining a diagnosis scheme of a test result through the knowledge graph, so that the test efficiency and the test effect are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of computers, and in particular to a server test platform, method, device, equipment and storage medium. BACKGROUND

[0002] The current server automatic test platform constructs a hierarchical architecture of a technical ecosystem, which is collaboratively operated by three core modules: Jenkins as a control center, responsible for task scheduling and process arrangement, and realizing continuous integration and distributed testing through a rich plug-in system, such as Secure Shell (SSH) remote connection and Pipeline flow process design; the server connection layer adopts a dual-channel design of SSH and Baseboard Management Controller (BMC), the SSH realizes command execution and file deployment based on an encryption protocol, and the BMC provides hardware-level management through an Intelligent Platform Management Interface (IPMI) protocol, forming physical isolation of the business and control network; the test execution layer is based on the Python technology stack, uses the Unit Testing Framework (unittest) / Python Testing Framework (pytest) framework to support modular test cases, covering scenarios such as function verification and performance testing, and integrates tool chains such as log analysis and data generation.

[0003] However, the different dimensions of data of the server automatic test platform are independent of each other, forming an information island, making performance analysis difficult, not only affecting test efficiency, but also restricting the improvement of the entire research and development efficiency; in addition, the test platform does not effectively use historical test data, causing resource waste; in terms of report visualization, the system provides an overly single and rigid display form, which cannot meet different display requirements, reducing the decision-making reference value of test results in actual projects. SUMMARY

[0004] The purpose of the embodiments of the present application is to provide a server test platform, method, device, equipment and storage medium, which solves the problems that the different dimensions of data of the current server automatic test platform are independent of each other, performance analysis is difficult, historical test data is not effectively used, and the display form is overly single and rigid, and the specific technical solutions are as follows:

[0005] In the first aspect of the present application, a server test platform is first provided, which comprises a three-dimensional correlation model, an interactive analysis system, a prediction analysis model and a knowledge management system.

[0006] The interaction analysis system is configured to receive user inputted information to be tested, and perform semantic analysis on the information to be tested through an integrated natural language processing engine to obtain a semantic analysis result, the semantic analysis result being used to determine a target test to be executed on the server test platform, and through a target abnormal mode and multi-dimensional dynamic correlation data, a test result of the target test is displayed in the interaction analysis system in a target form;

[0007] The three-dimensional correlation model is configured to extract multi-dimensional dynamic correlation data about the target test in the process of executing the target test, and identify a first abnormal mode of the target test through the multi-dimensional dynamic correlation data;

[0008] The prediction analysis model is configured to predict a second abnormal mode of the target test corresponding to the multi-dimensional dynamic correlation data, the first abnormal mode and the second abnormal mode belonging to a target abnormal mode;

[0009] The knowledge management system is configured to determine a diagnosis scheme of the test result through a knowledge graph.

[0010] In the second aspect of the embodiment of the present application, a server test method is also provided, which is applied to a server test platform, the server test platform comprising a three-dimensional correlation model, an interaction analysis system, a prediction analysis model and a knowledge management system, the method comprising:

[0011] acquiring user inputted information to be tested in the interaction analysis system, and performing semantic analysis on the information to be tested to obtain a semantic analysis result;

[0012] determining a target test to be executed on the server test platform through the semantic analysis result;

[0013] extracting multi-dimensional dynamic correlation data about the target test from the three-dimensional correlation model in the process of executing the target test;

[0014] identifying a first abnormal mode of the target test through the multi-dimensional dynamic correlation data, or predicting a second abnormal mode of the target test corresponding to the multi-dimensional dynamic correlation data through the prediction analysis model, the first abnormal mode and the second abnormal mode belonging to a target abnormal mode;

[0015] displaying a test result of the target test in the interaction analysis system in a target form through the target abnormal mode and the multi-dimensional dynamic correlation data, and determining a diagnosis scheme of the test result through a knowledge graph of the knowledge management system.

[0016] Optionally, the first abnormal pattern of the target test is identified by the multi-dimensional dynamic correlation data, or a second abnormal pattern of the target test is predicted by a prediction analysis model based on the multi-dimensional dynamic correlation data, and the method comprises the following steps of:

[0017] The multi-dimensional dynamic correlation data is searched within a preset numerical range.

[0018] If the multi-dimensional dynamic correlation data exceeds the preset numerical range, it is determined that the target test has an abnormal pattern.

[0019] The target data type of the multi-dimensional dynamic correlation data exceeding the preset numerical range is identified.

[0020] It is determined that the abnormal pattern of the target test is a first abnormal pattern based on the target data type.

[0021] If the multi-dimensional dynamic correlation data does not exceed the preset numerical range, it is determined that the target test does not have an abnormal pattern.

[0022] After the multi-dimensional dynamic correlation data is normalized, the multi-dimensional dynamic correlation data is divided according to a preset fixed time window to obtain a plurality of groups of multi-dimensional dynamic correlation data.

[0023] For any group of multi-dimensional dynamic correlation data, the key features of the multi-dimensional dynamic correlation data are obtained, and the key features include at least one of time domain features, frequency domain features or statistical features.

[0024] The key features are input into a prediction analysis model to predict a second abnormal pattern of the target test corresponding to the multi-dimensional dynamic correlation data.

[0025] Optionally, the test result of the target test is displayed in an interactive analysis system in a target form based on the target abnormal pattern and the multi-dimensional dynamic correlation data, and the method comprises the following steps of:

[0026] The data type of the multi-dimensional dynamic correlation data and the identity type of the user are identified.

[0027] The initial form of the test result of the target test is determined based on the target abnormal pattern and the data type.

[0028] The initial form is adjusted based on the identity type to obtain a target form of the test result of the target test.

[0029] The test result is displayed in the interactive analysis system in the target form.

[0030] Optionally, the multi-dimensional dynamic correlation data comprises at least one of code commit records, performance index data and resource monitoring information.

[0031] Optionally, before determining the diagnosis scheme of the test result by using the knowledge graph of the knowledge management system, the method further comprises:

[0032] collecting a plurality of historical abnormal patterns about server testing and historical diagnosis schemes for the historical abnormal patterns;

[0033] constructing a correlation relationship between the historical abnormal patterns and the corresponding historical diagnosis schemes;

[0034] constructing a knowledge graph of a knowledge management system based on the correlation relationship.

[0035] Optionally, after constructing the knowledge graph of the knowledge management system based on the correlation relationship, the method further comprises:

[0036] setting a confidence weight for the correlation relationship;

[0037] recording a valid use frequency of the correlation relationship;

[0038] adjusting the confidence weight by using the valid use frequency;

[0039] if the confidence weight is less than a preset deletion threshold, deleting the correlation relationship from the knowledge graph.

[0040] In a third aspect of the embodiment of the present application, a server testing device is also provided, which is applied to a server testing platform, the server testing platform comprises a three-dimensional correlation model, an interactive analysis system, a prediction analysis model and a knowledge management system, and comprises:

[0041] a first acquisition module, configured to acquire test information input by a user in the interactive analysis system and perform semantic analysis on the test information to obtain a semantic analysis result;

[0042] a testing module, configured to determine a target test to be executed in the server testing platform by using the semantic analysis result;

[0043] an extraction module, configured to extract multi-dimensional dynamic correlation data about the target test from the three-dimensional correlation model during execution of the target test;

[0044] an identification module, configured to identify a first abnormal pattern of the target test by using the multi-dimensional dynamic correlation data, or predict a second abnormal pattern of the target test corresponding to the multi-dimensional dynamic correlation data by using the prediction analysis model, the first abnormal pattern and the second abnormal pattern belonging to a target abnormal pattern;

[0045] A display and diagnosis module is configured to display the test result of the target test in a target form on the interactive analysis system and determine a diagnosis scheme of the test result by using a knowledge graph of the knowledge management system.

[0046] In a fourth aspect of the embodiments of the present application, a communication device is further provided, which comprises a transceiver, a memory, a processor and a program stored in the memory and executable on the processor.

[0047] The processor is configured to read the program in the memory to implement the server test method.

[0048] In a fifth aspect of the embodiments of the present application, a computer readable storage medium is further provided, which stores instructions, when the instructions are executed on a computer, the computer is caused to execute the server test method.

[0049] The server test platform provided by the embodiments of the present application comprises a three-dimensional correlation model, an interactive analysis system, a prediction analysis model and a knowledge management system. The interactive analysis system is configured to receive user inputted test information, and perform semantic analysis on the test information by using an integrated natural language processing engine to obtain a semantic analysis result, so as to improve the operation efficiency of non-technical personnel, reduce the use threshold of the server test platform, meet the user demand, display the test result of the target test in a target form on the interactive analysis system by using a target abnormal mode and multi-dimensional dynamic correlation data, avoid the display form being too single and rigid, make the test result meet different display requirements, and improve the decision reference value of the test result in an actual project. The three-dimensional correlation model is configured to extract multi-dimensional dynamic correlation data about the target test in the process of executing the target test, and identify a first abnormal mode of the target test by using the multi-dimensional dynamic correlation data. The prediction analysis model is configured to predict a second abnormal mode of the target test corresponding to the multi-dimensional dynamic correlation data by analyzing the multi-dimensional data, avoid forming an information island, and improve the efficiency of performance analysis. The prediction analysis model is trained based on historical test data, so that the historical test data can be effectively used, and resource waste is avoided. The knowledge management system is configured to determine a diagnosis scheme of the test result by using a knowledge graph. The historical problem processing experience can be continuously converted into institutional knowledge assets, and the processing efficiency when the problem reoccurs can be greatly improved. BRIEF DESCRIPTION OF DRAWINGS

[0050] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed to be used in the embodiments or the prior art description will be briefly introduced.

[0051] Figure 1 A server test platform construction step flow chart provided by the embodiment of the present application is provided.

[0052] Figure 2 A server test method step flow chart provided by the embodiment of the present application is provided.

[0053] Figure 3 A server test method step flow chart provided by the embodiment of the present application is provided. Figure 2 A server test method step flow chart provided by the embodiment of the present application is provided.

[0054] Figure 4 A server test method step flow chart provided by the embodiment of the present application is provided.

[0055] Figure 5 A server test method step flow chart provided by the embodiment of the present application is provided. DETAILED DESCRIPTION

[0056] In order to make the purpose, technical scheme and advantages of the embodiments of the present application clearer, the embodiments of the present application will be described in detail below with reference to the drawings. However, those skilled in the art can understand that in the embodiments of the present application, many technical details are proposed in order to make the readers better understand the present application. However, even without these technical details and various changes and modifications based on the following embodiments, the technical scheme claimed by the present application can be realized. The division of the following embodiments is for the convenience of description, and should not constitute any limitation on the specific implementation of the present application, and the embodiments can be combined and referred to each other without contradiction.

[0057] The current server automation test platform constructs a complete technical ecosystem, and the core architecture thereof is composed of three key modules: the Jenkins control hub as the command center of the whole system, the server connection layer for realizing device access through the secure shell (SSH) and the baseboard management controller (BMC) double channel, and the test execution layer based on Python for constituting the business implementation part. The layered design ensures the integrity of the system in which each part performs its own function and cooperates.

[0058] The core technology architecture of Jenkins is developed in Java and designed based on a master-slave distributed model, with cross-platform characteristics. Its key technical features include an extensible plug-in architecture, Pipeline-as-Code, a pipeline definition method that codes the pipeline, and a well-developed REST API. These technical features enable flexible integration into various software development and testing environments. In terms of continuous integration, Jenkins implements a complete workflow of automatic detection of source code changes, triggering of build tasks, and collection and analysis of test results.

[0059] The distributed build capability of Jenkins enables parallel control of multiple test servers by establishing multiple execution nodes. Jenkins' task scheduling algorithm can intelligently distribute test tasks to available nodes, optimizing resource utilization. Its persistent storage mechanism records the historical data of each test execution, providing a data foundation for test result analysis and trend prediction. Therefore, in the current server automation testing platform, the Jenkins control center not only provides a powerful continuous integration engine to ensure task orchestration and scheduling capabilities, but also extends the functional boundaries through a rich plug-in ecosystem. For example, the included SSH plug-in enables secure connection to remote servers, the Pipeline plug-in supports visual orchestration of complex test processes, and various report plug-ins are responsible for the visual presentation of test results. In addition, the Webhook mechanism is used to achieve deep integration with version control systems, establishing a rapid response mechanism for code change-triggered automated testing.

[0060] The server connection layer adopts a dual-channel design. The SSH connection scheme is based on the SSH-2 protocol to build a secure communication pipeline, ensuring connection security through a key authentication system, and the Jenkins built-in credential management system provides reliable protection for certificate hosting. This channel mainly implements three major functions: remote command execution, test script deployment, and interactive session management. At the same time, the BMC out-of-band management is based on the IPMI standard protocol, building an independent management interface at the hardware level, and achieving physical isolation from the business network through a dedicated management network channel. This system mainly provides basic support functions such as power state control, hardware health monitoring, and firmware-level diagnostics, complementing the functions of the SSH channel.

[0061] The test execution layer adopts the Python technology stack to build a flexible test framework system. The test organization structure based on the unit testing framework (unittest) / Python testing framework (pytest) ensures the normativity of the use case management, and the modular design idea improves the reuse ability of the test components. The layer supports parallel execution of multiple test types, including system configuration verification, service function testing, performance benchmark testing, and abnormal scene simulation, and the like complete test scenarios. The supporting tool chain also provides enhanced functions such as log collection and analysis, test data generation, and result comparison, forming a complete test solution.

[0062] However, the above-mentioned server automatic test platform has obvious functional short board in the data analysis and report generation link, which seriously restricts the value mining and application effect of the test data. At the data analysis level, the system can only provide simple statistical indicators such as pass rate and failure rate, and lacks deep analysis ability of the test results. The performance test data and server resource monitoring indicators cannot be effectively associated, making it difficult to analyze the performance bottleneck. The lack of performance baseline management function makes the regression test lack reliable reference standard. The integration between the test data and the project management system is insufficient, forming an information island. These defects not only affect the test efficiency, but also hinder the establishment of continuous improvement mechanism, and ultimately restrict the improvement of the entire research and development efficiency. More notably, the system cannot construct a trend prediction model based on historical test data, missing the opportunity to realize test optimization through data mining. In the report generation link, the system provides a single and rigid display form. The fixed mode report template cannot meet different differentiated needs. Moreover, the existing report lacks interactive analysis function, and the user cannot perform data drilling and multi-dimensional comparative analysis according to needs.

[0063] Based on this, the embodiment of the present application provides a phased progressive implementation strategy to build a server test platform. Referring to Figure 1In the infrastructure deployment stage, a K8s-based elastic computing platform will be built, a large-scale Spark+Flink hybrid computing cluster will be deployed as a core analysis engine, a distributed storage system and a high-throughput message queue will be configured synchronously, a bottom support platform meeting the high availability requirement will be formed, and the system will be ensured to have elastic expansion capability to adapt to the continuous development needs of the business. In the data governance stage, a standardized model is established to set up an automatic conversion pipeline, realizing seamless conversion from raw data to analysis-ready data. A metadata management system is set up to automatically track data relationships and change history, and a real-time and batch dual-channel data processing system is built, in which the real-time acquisition channel has high-efficiency event processing capability, and the batch pipeline relies on workflow scheduling to realize precise synchronization. A complete quality monitoring system is configured, in which the quality monitoring system has built-in multi-dimensional detection indicators, which can automatically trigger abnormal alarm and provide repair suggestions. In the core setting stage, a three-dimensional correlation model, an interactive analysis system, a prediction analysis model and a knowledge management system are set up, in which the interactive analysis system is used to receive user input test information, and through the integration of a natural language processing engine, the test information is subjected to semantic analysis to obtain a semantic analysis result; and through target abnormal patterns and multi-dimensional dynamic correlation data, the test results of the target test are displayed in the interactive analysis system according to the target form; the three-dimensional correlation model is used to extract multi-dimensional dynamic correlation data about the target test in the process of performing the target test; and through the multi-dimensional dynamic correlation data, a first abnormal pattern of the target test is identified; the prediction analysis model is used to predict a second abnormal pattern of the target test corresponding to the multi-dimensional dynamic correlation data, the first abnormal pattern and the second abnormal pattern belong to the target abnormal pattern; and the knowledge management system is used to determine a diagnosis scheme of the test result through a knowledge graph. In the system integration stage, a full-business-scenario interface test coverage is set, high-concurrency stress testing and security verification are performed. In the operation stage, a gradual strategy is adopted, and after typical business line pilot verification, a gray release is carried out. An automatic updating operation and maintenance system is set, including model continuous optimization, knowledge base intelligent maintenance and system health monitoring, and in cooperation with key milestone nodes and delivery acceptance mechanism, the whole life cycle management efficiency is ensured.

[0064] Referring to Figure 2 , a step flowchart of a server test method provided by an embodiment of the present application is shown, which can include:

[0065] In step 101, test information input by a user in an interactive analysis system is obtained, and semantic analysis is performed on the test information to obtain a semantic analysis result.

[0066] The server test method of the embodiment of the present application is applied to a server test platform. The server test platform comprises a three-dimensional correlation model, an interactive analysis system, a prediction analysis model and a knowledge management system. The embodiment of the present application introduces a natural language processing engine into the interactive analysis system, and develops a semantic parser with context understanding capability, so that non-technical personnel can also input query information or test information to be tested in the interactive analysis system. Then, the natural language processing engine performs semantic analysis on the input information to obtain a semantic analysis result.

[0067] It should be understood that the natural language processing engine introduced in the embodiment of the present application is not general, but is customized for the server test field, and can accurately understand test-related terms (such as "API response time", "version comparison", etc.). Moreover, the natural language processing engine is different from simple keyword matching, and can understand the intention of a complex query through context understanding. For example, when the user inputs "compare the API response time of A version and B version", the specific meaning of "response time" in performance test (rather than "reaction time" in ordinary conversation) can be distinguished, and the specific software version numbers referred to by "A version" and "B version" can also be automatically associated.

[0068] In step 102, the target test to be executed on the server test platform is determined through the semantic analysis result.

[0069] The embodiment of the present application determines the test object of the target test to be executed, the test tool to be used and the test environment to be configured through the semantic analysis result, and then executes the target test through the server test platform, monitors the test process in real time, synchronously generates a test log, and facilitates subsequent problem positioning and analysis.

[0070] In step 103, multi-dimensional dynamic correlation data about the target test is extracted from the three-dimensional correlation model during execution of the target test.

[0071] The multi-dimensional dynamic correlation data in the embodiment of the present application comprises at least one of code submission records, performance index data and resource monitoring information. The three-dimensional correlation model of the embodiment of the present application constructs a multi-layer correlation network of code library, test case and performance index, which can dynamically correlate code submission records, performance index data and resource monitoring information. When the target test is executed, the multi-dimensional dynamic correlation data related to this target test can be obtained. By obtaining the multi-dimensional dynamic correlation data, the cause of an exception can be analyzed when the exception occurs, and a reliable reference standard can be found when subsequent regression testing is performed.

[0072] For example, the embodiment of the present application takes each submitted code as a new node, synchronously records the performance index data and resource monitoring data in the code execution process when performing the test according to the newly submitted code, and can obtain the performance index data and resource monitoring data of this test to analyze the error cause when the test reports an error. After the code is modified (including repairing defects and adding new functions), the performance index data and resource monitoring data of the execution code before modification can be taken as a reference standard to compare with the performance index data and resource monitoring data of the regression test, so as to determine whether the modified code destroys the correctness of the existing function. The steps involved can include:

[0073] In the case of determining to perform regression testing on the target code, the original code corresponding to the target code and the original performance index data and original resource monitoring data in the test process of the original code are obtained;

[0074] The actual performance index data and actual resource monitoring data in the regression test of the target code are obtained;

[0075] The original performance index data and the actual performance index data, and the original resource monitoring data and the actual resource monitoring data are compared respectively;

[0076] If the errors of the original performance index data and the actual performance index data, and the original resource monitoring data and the actual resource monitoring data are within the preset range, it is determined that the regression test of the target code is passed.

[0077] The above steps realize the quantitative evaluation of the regression test through systematic data collection and comparison: by comparing the performance index (such as response time and throughput) and resource monitoring data (such as CPU and memory occupation) of the new and old codes, the performance degradation and resource anomaly introduced by code modification can be efficiently identified, and the preset error range provides a clear passing standard for automatic decision-making.

[0078] In addition to judging whether the regression test is passed through the above data comparison, it can also be judged whether all the automatic test cases in the regression test are passed, whether the core business process verification is passed, whether the upstream and downstream service indexes are affected, whether the database transaction is complete, etc., which are not limited in the present application.

[0079] In step 104, a first abnormal mode of the target test is identified through multi-dimensional dynamic correlation data, or a second abnormal mode of the target test corresponding to the multi-dimensional dynamic correlation data is predicted through a prediction analysis model, and the first abnormal mode and the second abnormal mode belong to the target abnormal mode.

[0080] The three-dimensional correlation model of the embodiment of the present application not only correlates multi-dimensional data, but also integrates an abnormality detection subsystem based on deep learning. The subsystem integrates a Long Short-Term Memory (LSTM) neural network and an adaptive threshold algorithm, and can accurately identify various abnormal patterns of the target test, thereby realizing millisecond-level real-time problem warning. In addition, in order to avoid resource waste of historical test data, the embodiment of the present application further sets a prediction analysis model, which predicts the probability of abnormality of the target test and the abnormal pattern that will occur through multi-dimensional dynamic correlation data of the target test.

[0081] The step 104 comprises the following sub-steps:

[0082] The sub-step 1041 searches the multi-dimensional dynamic correlation data in a preset numerical range.

[0083] The sub-step 1042 determines that the target test has appeared abnormality if the multi-dimensional dynamic correlation data exceeds the preset numerical range.

[0084] The sub-step 1043 identifies the target data type of the multi-dimensional dynamic correlation data exceeding the preset numerical range.

[0085] The sub-step 1044 determines that the abnormal pattern of the target test is a first abnormal pattern through the target data type.

[0086] The sub-step 1045 determines that the target test has not appeared abnormality if the multi-dimensional dynamic correlation data does not exceed the preset numerical range.

[0087] The sub-step 1046 divides the multi-dimensional dynamic correlation data into several groups after normalizing the multi-dimensional dynamic correlation data according to a preset fixed time window.

[0088] The sub-step 1047 obtains key features of the multi-dimensional dynamic correlation data, including at least one of time domain features, frequency domain features or statistical features, for any group of multi-dimensional dynamic correlation data.

[0089] The sub-step 1048 inputs the key features into the prediction analysis model to predict a second abnormal pattern of the target test corresponding to the multi-dimensional dynamic correlation data.

[0090] For example, assuming that the acquired multi-dimensional dynamic correlation data is API response time and CPU occupancy, the preset numerical range set for the API response time is: API average response time ≤ 200 ms, and the preset numerical range set for the CPU occupancy is: CPU occupancy ≤ 70%, it is detected that the API response time in a certain period of time is 260 ms > 200 ms, at this time, it is determined that the target test is abnormal, and the first abnormal mode may be performance degradation, if the multi-dimensional dynamic correlation data does not exceed the preset numerical range, at this time, the data can be normalized to ensure the stability of the input numerical value, the multi-dimensional dynamic correlation data is divided through a preset fixed time window (for example, the multi-dimensional dynamic correlation data in the past 24 hours is divided according to a 5-minute window), this operation can construct the required sequence input, the characteristics of the data can be fully characterized from different dimensions through feature extraction, and the limitation of a single feature type is compensated. Then, the extracted key features are input into the prediction analysis model, and the abnormal analysis result of the target test corresponding to the multi-dimensional dynamic correlation data is output, the abnormal analysis result includes the probability of the target test being abnormal and the second abnormal mode (for example, memory leakage causes continuous deterioration of the response time).

[0091] The above steps can quickly locate the explicit abnormality (the first abnormal mode) through threshold detection, and further discover the potential risk (the second abnormal mode) in combination with the time sequence analysis and the prediction model, so that the upgrade from passive alarm to active prediction is realized, and the system stability is improved.

[0092] In step 105, the test result of the target test is displayed in the interactive analysis system according to the target form through the target abnormal mode and the multi-dimensional dynamic correlation data, and the diagnosis scheme of the test result is determined by using the knowledge graph of the knowledge management system.

[0093] The sub-step of displaying the test result in the interactive analysis system according to the target form through the target abnormal mode and the multi-dimensional dynamic correlation data in the embodiment of the application, as shown in Figure 3 , includes:

[0094] In sub-step 1051, the data type of the multi-dimensional dynamic correlation data and the identity type of the user are identified.

[0095] In sub-step 1052, the initial form of the test result display is determined through the target abnormal mode and the data type.

[0096] In sub-step 1053, the initial form is adjusted through the identity type to obtain the target form of the test result display.

[0097] In sub-step 1054, the test result is displayed in the interactive analysis system according to the target form.

[0098] For example, if the data type of the identified multi-dimensional dynamic correlation data is "API response time", the identity type of the user is an operation and maintenance personnel, and the target abnormal mode is that the API response time exceeds the standard, the initial form of the test result display is a line chart for comparison between the new and old versions. Because the operation and maintenance personnel focus on system stability, the initial form is adjusted to a heat map display form. Because the line chart only shows the global average, it may mask serious problems in some nodes, and the heat map can show the performance differences between cluster nodes, such as intuitively exposing specific nodes and time hotspots. In addition, the database load indicator (color gradient representing query delay) can be superimposed on the heat map to help correlation analysis of the root cause.

[0099] The above steps provide the best visualization form through different data types and different identities of the user, and meet different needs of the user.

[0100] The interactive analysis platform of the present application also supports multi-level data drilling from macro trends to micro details after displaying the test result, so that the analysis efficiency is improved. For example, abnormal annotations are made on the displayed test result, and clicking on the abnormal annotations is supported to display abnormal details. For example, the order service delay is found to be rising on a macro level through a line chart, the payment interface delay is located to have the highest contribution on a meso level, and the payment interface call chain is drilled down to view the third-party API timeout on a micro level.

[0101] In addition, in the embodiment of the present application, before the knowledge graph of the knowledge management system is used to determine the diagnosis scheme of the test result, a professional knowledge graph of the test field needs to be constructed. The steps of construction include:

[0102] Collecting a plurality of historical abnormal modes of server testing and historical diagnosis schemes for the historical abnormal modes;

[0103] Building an association relationship between the historical abnormal modes and the corresponding historical diagnosis schemes;

[0104] Building a knowledge graph of the knowledge management system based on the association relationship.

[0105] The historical abnormal modes and the historical diagnosis schemes for the historical abnormal modes are continuously updated, and therefore the knowledge graph is also continuously updated. When the knowledge graph is used, the target abnormal mode is matched with the historical abnormal modes of the knowledge graph in terms of similarity, and the historical diagnosis scheme corresponding to the historical abnormal mode with a similarity greater than a similarity threshold is taken as the diagnosis scheme of the target test. If there are a plurality of historical abnormal modes with a similarity greater than the similarity threshold, the historical diagnosis scheme corresponding to the historical abnormal mode with the maximum similarity value is selected as the diagnosis scheme of the target test.

[0106] Further, when the knowledge graph of the embodiment of the present application is dynamically updated, in addition to expanding the knowledge graph according to the newly added historical abnormal mode and the historical diagnosis scheme for the historical abnormal mode, the invalid suggestions are also deleted to avoid misleading analysis, and the steps involved include:

[0107] setting a confidence weight for the association relationship;

[0108] recording the effective use times of the association relationship;

[0109] adjusting the confidence weight through the effective use times;

[0110] if the confidence weight is less than a preset deletion threshold, the association relationship is deleted from the knowledge graph.

[0111] For example, assuming that there is an association relationship: "CPU usage rate continuously exceeds 90%" → "diagnosis scheme: check the GC configuration of the Java application", and the confidence weight set for this association relationship is 0.7, when the CPU high load occurs again, the operation and maintenance personnel adopt the scheme and successfully solve the problem, the effective use times +1, the system automatically increases the confidence weight from 0.7 to 0.8 (each successful use increases 0.1), and the scheme fails to solve the problem for two subsequent times (because the new version uses Golang, the GC configuration is irrelevant), the confidence weight gradually decreases to 0.6, and assuming that the preset deletion threshold is 0.65, because 0.6<0.65, the system automatically removes this outdated association relationship from the knowledge graph to avoid misleading subsequent diagnosis.

[0112] By dynamically maintaining the confidence weight and the effective use times, the knowledge graph retains high-frequency and effective diagnosis schemes, eliminates invalid experiences, improves the accuracy of the diagnosis scheme, and at the same time, can reduce the time of the operation and maintenance personnel to screen invalid schemes and focus on high-weight associations.

[0113] The server test method provided by the embodiment of the application is applied to a server test platform, the server test platform comprises a three-dimensional correlation model, an interactive analysis system, a prediction analysis model and a knowledge management system; the user inputted information to be tested in the interactive analysis system is acquired, semantic analysis is performed on the information to be tested, and a semantic analysis result is obtained; the operation efficiency of non-technical personnel is improved, the use threshold of the server test platform is reduced, the user demand is met, the target test to be executed on the server test platform is determined through the semantic analysis result; in the process of executing the target test, multi-dimensional dynamic correlation data about the target test is extracted from the three-dimensional correlation model; through analysis by combining the multi-dimensional data, an information island is avoided, the efficiency of performance analysis is improved, a first abnormal mode of the target test is identified through the multi-dimensional dynamic correlation data, or a second abnormal mode of the target test is predicted through the prediction analysis model, the prediction analysis model is trained based on historical test data, the historical test data can be effectively used, resource waste is avoided, the first abnormal mode and the second abnormal mode belong to a target abnormal mode; through the target abnormal mode and the multi-dimensional dynamic correlation data, the test result of the target test is displayed in the interactive analysis system in a target form, the display form is avoided to be too single and rigid, the test result meets different display requirements, the decision reference value of the test result in an actual project is improved, and a diagnosis scheme of the test result is determined by using a knowledge graph of the knowledge management system, the use of the knowledge graph can continuously convert historical problem processing experience into institutional knowledge assets, and the processing efficiency when a problem reoccurs is greatly improved.

[0114] Referring to Figure 4 , a structure schematic diagram of a server test device provided by the embodiment of the application is shown, the device is applied to a server test platform, the server test platform comprises a three-dimensional correlation model, an interactive analysis system, a prediction analysis model and a knowledge management system, as Figure 4 indicated, the device can comprise:

[0115] A first acquisition module 201 is configured to acquire user inputted information to be tested in an interactive analysis system, perform semantic analysis on the information to be tested, and obtain a semantic analysis result.

[0116] A test module 202 is configured to determine a target test to be executed on a server test platform through a semantic analysis result.

[0117] An extraction module 203 is configured to extract multi-dimensional dynamic correlation data about the target test from a three-dimensional correlation model in the process of executing the target test.

[0118] The identification module 204 is configured to identify a first abnormal pattern of the target test from the multidimensional dynamic correlation data, or predict a second abnormal pattern of the target test from the multidimensional dynamic correlation data by using a prediction analysis model, wherein the first abnormal pattern and the second abnormal pattern belong to a target abnormal pattern.

[0119] The display and diagnosis module 205 is configured to display the test result of the target test in a target form in the interactive analysis system according to the target abnormal pattern and the multidimensional dynamic correlation data, and determine a diagnosis scheme of the test result by using a knowledge graph of the knowledge management system.

[0120] Optionally, the identification module 204 specifically includes:

[0121] The retrieval sub-module is configured to retrieve the multidimensional dynamic correlation data in a preset numerical range.

[0122] The first abnormality determination sub-module is configured to determine that the target test has an abnormality if the multidimensional dynamic correlation data exceeds the preset numerical range.

[0123] The first identification sub-module is configured to identify a target data type of the multidimensional dynamic correlation data exceeding the preset numerical range.

[0124] The mode determination sub-module is configured to determine that the abnormal pattern of the target test is the first abnormal pattern by using the target data type.

[0125] The second abnormality determination sub-module is configured to determine that the target test has no abnormality if the multidimensional dynamic correlation data does not exceed the preset numerical range.

[0126] The data division sub-module is configured to divide the multidimensional dynamic correlation data according to a preset fixed time window after normalizing the multidimensional dynamic correlation data, to obtain a plurality of groups of multidimensional dynamic correlation data.

[0127] The acquisition sub-module is configured to acquire a key feature of the multidimensional dynamic correlation data, wherein the key feature includes at least one of a time domain feature, a frequency domain feature or a statistical feature.

[0128] The prediction sub-module is configured to input the key feature into a prediction analysis model to predict the second abnormal pattern of the target test from the multidimensional dynamic correlation data.

[0129] Optionally, the display and diagnosis module 205 specifically includes:

[0130] The second identification sub-module is configured to identify a data type of the multidimensional dynamic correlation data and an identity type of a user.

[0131] The determination sub-module is configured to determine an initial form of the test result of the target test according to the target abnormal pattern and the data type.

[0132] an adjusting sub-module, configured to adjust the initial form by the identity type to obtain a target form of the test result display of the target test.

[0133] a display sub-module, configured to display the test result in the interactive analysis system according to the target form.

[0134] Optionally, the multi-dimensional dynamic correlation data comprises at least one of code submission records, performance index data and resource monitoring information.

[0135] Optionally, the server testing apparatus further comprises:

[0136] a collecting module, configured to collect a plurality of historical abnormal patterns about server testing and historical diagnosis schemes for the historical abnormal patterns.

[0137] a relationship constructing module, configured to construct a correlation relationship between the historical abnormal patterns and the corresponding historical diagnosis schemes.

[0138] a knowledge graph constructing module, configured to construct a knowledge graph of a knowledge management system based on the correlation relationship.

[0139] Optionally, the server testing apparatus further comprises:

[0140] a weight setting module, configured to set a confidence weight for the correlation relationship.

[0141] a recording module, configured to record an effective use frequency of the correlation relationship.

[0142] an adjusting module, configured to adjust the confidence weight by the effective use frequency.

[0143] a deleting module, configured to delete the correlation relationship from the knowledge graph if the confidence weight is less than a preset deleting threshold.

[0144] The server testing apparatus provided in this invention is applied to a server testing platform, which includes a three-dimensional correlation model, an interactive analysis system, a predictive analysis model, and a knowledge management system. It acquires the test information input by the user in the interactive analysis system and performs semantic analysis on the test information to obtain semantic analysis results. This improves the operational efficiency of non-technical personnel, lowers the barrier to entry for using the server testing platform, meets user needs, and determines the target test to be executed on the server testing platform based on the semantic analysis results. During the execution of the target test, it extracts multi-dimensional dynamic correlation data about the target test from the three-dimensional correlation model. By combining and analyzing the multi-dimensional data, it avoids the formation of information silos, improves the efficiency of performance analysis, and identifies the first abnormal pattern appearing in the target test through the multi-dimensional dynamic correlation data. The predictive analytics model predicts the second abnormal pattern of the target test corresponding to multidimensional dynamic correlation data. The predictive analytics model is trained on historical test data, which can be effectively used to avoid resource waste. The first and second abnormal patterns belong to the target abnormal patterns. Through the target abnormal patterns and multidimensional dynamic correlation data, the test results of the target test are displayed in the interactive analysis system in the target format, avoiding the presentation format being too monotonous and rigid. This allows the test results to meet different display needs, while improving the decision-making reference value of the test results in actual projects. The knowledge graph of the knowledge management system is used to determine the diagnostic solution of the test results. The use of the knowledge graph can continuously transform historical problem-solving experience into organizational knowledge assets, which can significantly improve the processing efficiency when the problem is reproduced.

[0145] This invention also provides a communication device, such as... Figure 5 As shown, it includes a processor 301, a communication interface 302, a memory 303, and a communication bus 304, wherein the processor 301, the communication interface 302, and the memory 303 communicate with each other through the communication bus 304.

[0146] Memory 303 is used to store computer programs;

[0147] When processor 301 executes a program stored in memory 303, it performs the following steps:

[0148] Obtain the test information input by the user in the interaction analysis system, and perform semantic analysis on the test information to obtain the semantic analysis results;

[0149] The semantic analysis results are used to determine the target test to be executed on the server testing platform.

[0150] During the execution of the target test, multidimensional dynamic correlation data about the target test is extracted from the three-dimensional correlation model;

[0151] The first abnormal pattern of the target test appearing is identified through the multi-dimensional dynamic correlation data, or a second abnormal pattern of the target test appearing is predicted through a prediction analysis model, and the first abnormal pattern and the second abnormal pattern belong to a target abnormal pattern;

[0152] Through the target abnormal pattern and the multi-dimensional dynamic correlation data, a test result of the target test is displayed in a target form in an interactive analysis system, and a diagnosis scheme of the test result is determined by using a knowledge graph of a knowledge management system.

[0153] The communication bus mentioned above can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. The communication bus can be divided into a physical address bus, a data bus, a control bus, etc. For the convenience of representation, only one thick line is used in the figure, but it does not mean that there is only one bus or only one type of bus.

[0154] The communication interface is used for communication between the terminal and other devices.

[0155] The memory can include a Random Access Memory (RAM) and can also include a non-volatile memory, for example, at least one disk memory. Optionally, the memory can also be at least one storage device located away from the aforementioned processor.

[0156] The processor mentioned above can be a general-purpose processor, including a Central Processing Unit (CPU), a Network Processor (NP), etc. It can also be a Digital Signal Processing (DSP), an Application Specific Integrated Circuit (ASIC), a Field-Programmable Gate Array (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components.

[0157] The application also provides a readable storage medium, when instructions in the storage medium are executed by a processor of an electronic device, the electronic device can execute the server test method of the foregoing embodiments.

[0158] For device embodiments, which are substantially similar to the method embodiments, the description is relatively simple, and reference is made to the description of the method embodiments where relevant.

[0159] The algorithms and displays presented herein are not inherently related to any particular computer, virtual apparatus, or other apparatus. Various structural requirements are determined to construct such a computer, virtual apparatus, or other apparatus according to the description given above. Moreover, the present application is not intended to be bound by any particular programming language. It will be appreciated that there are many alternative ways of implementing the present application described herein, and implementations of the present application described herein can be realized using a variety of programming languages. The description above of a particular language is merely provided to disclose a best mode of the present application.

[0160] In the description provided herein, numerous specific details are set forth. However, it is understood that embodiments of the application can be practiced without these specific details. In some instances, well-known methods, structures and techniques have not been described in detail in order to not obscure the understanding of this description.

[0161] Similarly, it is to be understood that the mechanical components of the application can be placed in any desired configuration according to the specific requirements of a given application. Further, it is to be understood that the application can employ various computer-implemented processes which include routines executed by computers. These computer-implemented processes and routines can be used to implement various aspects of the application. Additionally, it is to be understood that the application can be used to support one or more computer systems, devices, or things which perform one or more operations. Further, it is to be understood that at least some of the processes can be performed under the control of one or more computer systems or devices configured with executable instructions (which can be interpreted as a "processing device" component of a computer system or device implementing the application). One or more computer systems or devices configured with executable instructions can be used to implement various aspects of the application and can include at least one processing device component.

[0162] Those skilled in the art will appreciate that the modules in the apparatuses in the embodiments can be adapted and placed in one or more apparatuses other than the embodiments. The modules or units or components in the embodiments can be combined into one module or unit or component, and further can be split into more sub-modules or sub-units or sub-components. Any combination of all the features disclosed in the specification (including the accompanying claims, abstract and drawings), and any method or process or steps of any method or process so disclosed, can be made unless specifically stated otherwise. Each feature disclosed in the description (including the accompanying claims, abstract and drawings) can be replaced by alternative features serving the same, equivalent or similar purpose, unless expressly stated otherwise.

[0163] Various component embodiments of the present application can be implemented in hardware, or as software modules running in one or more processors, or in combinations thereof. As will be apparent to those of ordinary skill in the art, some or all of the functionality of some or all of the components in the sequencing apparatus according to the present application can be implemented using a microprocessor or a digital signal processor (DSP) in practice. The present application can also be implemented as a program for executing, in whole or in part, the methods described herein on a device or apparatus. Such a program embodying the present application can be stored on a computer-readable medium, or can be in the form of signals on one or more carriers. Such signals can be downloaded from an Internet website, or provided on a carrier, or in any other form.

[0164] It should be noted that the above-mentioned embodiments illustrate rather than limit the application, and that one skilled in the art will be able to design many alternative embodiments without departing from the scope of the appended claims. In the claims, any reference signs placed between parentheses shall not be construed as limiting the claim. The word 'comprising' does not exclude the presence of elements or steps not listed in a claim. The word 'a' or 'an' preceding an element does not exclude the presence of a plurality of such elements. The application can be implemented by means of hardware comprising several distinct elements, and by means of a suitably programmed computer. In the claims, the word 'first','second', 'third', etc. does not limit the number for these elements. These words are only used to distinguish between alternative claims. The mere fact that different claims depend on the same independent claim is not a reason for assuming that the dependent claims are mutually identical.

[0165] Those skilled in the art can clearly understand that, for the convenience and brevity of the description, the specific working processes of the devices, apparatuses and units described above can refer to the corresponding processes in the foregoing method embodiments, which will not be described here.

[0166] The above only describes the preferred embodiments of the present application and is not intended to limit the present application. Any modification, equivalent replacement and improvement made within the spirit and principle of the present application shall be included in the protection scope of the present application.

[0167] The above only describes the preferred embodiments of the present application and is not intended to limit the present application. Any modification, equivalent replacement and improvement made within the spirit and principle of the present application shall be included in the protection scope of the present application.

[0168] It should be noted that the above-mentioned embodiments of the present application acquire various data related processes, which are carried out under the premise of complying with the corresponding data protection regulations and policies of the country where the location is located, and with the authorization given by the corresponding device owner.

Claims

1. A server testing platform, characterized in that, The server testing platform includes: a 3D association model, an interactive analysis system, a predictive analysis model, and a knowledge management system; The interactive analysis system is used to receive test information input by the user, and to perform semantic analysis on the test information by integrating a natural language processing engine to obtain semantic analysis results. The semantic analysis results are used to determine the target test to be executed on the server test platform; and to display the test results of the target test in the interactive analysis system in the target format by using the target anomaly pattern and multi-dimensional dynamic correlation data. The three-dimensional correlation model is used to extract multi-dimensional dynamic correlation data about the target test during the execution of the target test; and to identify the first abnormal pattern of the target test through the multi-dimensional dynamic correlation data. The predictive analysis model is used to predict the second abnormal pattern that will occur in the target test corresponding to the multidimensional dynamic correlation data, where the first abnormal pattern and the second abnormal pattern are target abnormal patterns. The knowledge management system is used to determine a diagnostic plan for the test results through a knowledge graph.

2. A server testing method, characterized in that, Applied to a server testing platform, the server testing platform includes a 3D association model, an interactive analysis system, a predictive analysis model, and a knowledge management system, the method includes: Obtain the test information input by the user in the interaction analysis system, and perform semantic analysis on the test information to obtain the semantic analysis results; The semantic analysis results are used to determine the target test to be executed on the server testing platform. During the execution of the target test, multidimensional dynamic correlation data about the target test is extracted from the three-dimensional correlation model; The first abnormal pattern of the target test is identified by the multidimensional dynamic correlation data, or the second abnormal pattern of the target test corresponding to the multidimensional dynamic correlation data is predicted by the predictive analysis model. The first abnormal pattern and the second abnormal pattern belong to the target abnormal pattern. Based on the target anomaly pattern and the multidimensional dynamic correlation data, the test results of the target test are displayed in the interactive analysis system according to the target format, and the diagnostic scheme of the test results is determined using the knowledge graph of the knowledge management system.

3. The method according to claim 2, characterized in that, The step of identifying a first abnormal pattern in the target test through the multidimensional dynamic correlation data, or predicting a second abnormal pattern in the target test corresponding to the multidimensional dynamic correlation data through a predictive analysis model, includes: The multidimensional dynamic correlation data is retrieved within a preset numerical range; If the multidimensional dynamic correlation data exceeds the preset value range, it is determined that the target test has an anomaly. Identify the target data type of the multidimensional dynamic correlation data that exceeds the preset numerical range; Based on the target data type, the abnormal pattern that occurred in the target test is determined to be the first abnormal pattern; If the multidimensional dynamic correlation data does not exceed the preset value range, it is determined that the target test has not shown any abnormalities. After normalizing the multidimensional dynamic correlation data, the multidimensional dynamic correlation data is divided according to a preset fixed time window to obtain several sets of multidimensional dynamic correlation data. For any set of multidimensional dynamic correlation data, obtain the key features of the multidimensional dynamic correlation data, wherein the key features include at least one of time domain features, frequency domain features, or statistical features; The key features are input into the predictive analysis model to predict the second abnormal pattern that will occur in the target test corresponding to the multidimensional dynamic correlation data.

4. The method according to claim 2, characterized in that, The step of displaying the test results of the target test in the interactive analysis system according to the target format using the target anomaly pattern and the multi-dimensional dynamic correlation data includes: Identify the data type of the multidimensional dynamic correlation data and the user's identity type; Based on the target anomaly pattern and the data type, determine the initial format for displaying the test results of the target test; The initial form is adjusted based on the identity type to obtain the target form for displaying the test results of the target test; The test results will be displayed in the interactive analysis system according to the target format.

5. The method according to claim 2, characterized in that, The multidimensional dynamic correlation data includes at least one of code commit records, performance index data, and resource monitoring information.

6. The method according to claim 2, characterized in that, Before determining the diagnostic plan for the test results using the knowledge graph of the knowledge management system, the method further includes: Collect several historical anomaly patterns related to server testing and historical diagnostic solutions for these anomaly patterns; Establish a correlation between the historical anomaly patterns and the corresponding historical diagnostic schemes; A knowledge graph for the knowledge management system is constructed based on the aforementioned relationships.

7. The method according to claim 6, characterized in that, After constructing the knowledge graph of the knowledge management system based on the aforementioned relationships, the system further includes: Set confidence weights for the relationships; Record the number of times the aforementioned association is used effectively; The confidence weight is adjusted based on the number of valid uses. If the confidence weight is less than the preset deletion threshold, the association will be deleted from the knowledge graph.

8. A server testing device, characterized in that, This is applied to a server testing platform, which includes a 3D relational model, an interactive analysis system, a predictive analysis model, and a knowledge management system, comprising: The first acquisition module is used to acquire the test information input by the user in the interaction analysis system, and to perform semantic analysis on the test information to obtain the semantic analysis result; The testing module is used to determine the target test to be executed on the server testing platform based on the semantic analysis results. An extraction module is used to extract multidimensional dynamic correlation data about the target test from the three-dimensional correlation model during the execution of the target test; The identification module is used to identify a first abnormal pattern of the target test through the multidimensional dynamic correlation data, or to predict a second abnormal pattern of the target test corresponding to the multidimensional dynamic correlation data through a predictive analysis model, wherein the first abnormal pattern and the second abnormal pattern belong to the target abnormal pattern. The display and diagnosis module is used to display the test results of the target test in the interactive analysis system according to the target format through the target anomaly pattern and the multi-dimensional dynamic correlation data, and to determine the diagnosis plan of the test results using the knowledge graph of the knowledge management system.

9. A communication device, characterized in that, include: A transceiver, a memory, a processor, and a program stored in the memory and executable on the processor; The processor is used to read a program from the memory to implement the steps in the server testing method as described in any one of claims 2-7.

10. A readable storage medium for storing a program, characterized in that, When the stored program is executed by the processor, it implements the steps of the server testing method as described in any one of claims 2-7.

Images