Application program performance analysis method and device based on large language model

Abstract

The invention provides an application program performance analysis method and device based on a large language model, electronic equipment and a readable storage medium, and relates to the technical field of artificial intelligence such as large language models, big data and cloud services. The application program performance analysis method based on the large language model comprises the steps of obtaining target index data, corresponding to a target data dimension, of a target application program running on a cloud server; obtaining target summary information, corresponding to the target data dimension, of the target application program according to the target index data by using a large language model; obtaining target performance analysis knowledge information according to the target data dimension; and obtaining a performance analysis result, corresponding to the target data dimension, of the target application program by using the large language model according to the target summary information and the target performance analysis knowledge information. According to the method and the device, the efficiency and accuracy of performing performance analysis on the target application program can be improved.

Description

Technical Field

[0001] The present disclosure relates to the field of computer technology, and in particular to the field of artificial intelligence technology such as large language models, big data, and cloud services. A method, device, electronic device, and readable storage medium for analyzing application performance based on a large language model are provided. Background Art

[0002] In the prior art, when performing performance analysis on applications running on cloud servers, a method generally adopted is to manually analyze a large amount of indicator data collected during the operation of the application, and then obtain the performance analysis results of the application based on the manual analysis results. Therefore, when performing performance analysis on applications, the prior art has technical problems such as low analysis efficiency, poor analysis accuracy, and high analysis cost. Summary of the invention

[0003] According to a first aspect of the present disclosure, a method for analyzing application performance based on a large language model is provided, comprising: obtaining target indicator data corresponding to a target data dimension of a target application running on a cloud server; using a large language model to obtain target summary information corresponding to the target data dimension based on the target indicator data; obtaining target performance analysis knowledge information based on the target data dimension; using the large language model to obtain a performance analysis result of the target application corresponding to the target data dimension based on the target summary information and the target performance analysis knowledge information.

[0004] According to a second aspect of the present disclosure, there is provided an application performance analysis device based on a large language model, comprising: a first acquisition unit, used to acquire target indicator data corresponding to a target data dimension of a target application running on a cloud server; a summary unit, used to use a large language model to obtain target summary information of the target application corresponding to the target data dimension according to the target indicator data; a second acquisition unit, used to acquire target performance analysis knowledge information according to the target data dimension; and an analysis unit, used to use the large language model to obtain a performance analysis result of the target application corresponding to the target data dimension according to the target summary information and the target performance analysis knowledge information.

[0005] According to a third aspect of the present disclosure, an electronic device is provided, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the method as described above.

[0006] According to a fourth aspect of the present disclosure, a non-transitory computer-readable storage medium storing computer instructions is provided, wherein the computer instructions are used to cause the computer to execute the method as described above.

[0007] According to a fifth aspect of the present disclosure, a computer program product is provided, comprising a computer program, wherein the computer program implements the method as described above when executed by a processor.

[0008] It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present disclosure, nor is it intended to limit the scope of the present disclosure. Other features of the present disclosure will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] The accompanying drawings are used to better understand the present solution and do not constitute a limitation of the present disclosure.

[0010] Figure 1 is a schematic diagram according to a first embodiment of the present disclosure;

[0011] Figure 2 is a schematic diagram according to a second embodiment of the present disclosure;

[0012] Figure 3 is a schematic diagram according to a third embodiment of the present disclosure;

[0013] Figure 4 It is a block diagram of an electronic device used to implement the application performance analysis method based on a large language model according to an embodiment of the present disclosure. DETAILED DESCRIPTION

[0014] The following is a description of exemplary embodiments of the present disclosure in conjunction with the accompanying drawings, including various details of the embodiments of the present disclosure to facilitate understanding, which should be considered as merely exemplary. Therefore, it should be recognized by those of ordinary skill in the art that various changes and modifications may be made to the embodiments described herein without departing from the scope and spirit of the present disclosure. Similarly, for the sake of clarity and conciseness, the description of well-known functions and mechanisms is omitted in the following description.

[0015] Figure 1 Schematic diagram of the first embodiment of the present disclosure. Figure 1 As shown, the application performance analysis method based on the large language model of this embodiment specifically includes the following steps:

[0016] S101, obtaining target indicator data corresponding to a target data dimension of a target application running on a cloud server;

[0017] S102, using a large language model, obtaining target summary information of the target application corresponding to the target data dimension according to the target indicator data;

[0018] S103, acquiring target performance analysis knowledge information according to the target data dimension;

[0019] S104: Using the large language model, according to the target summary information and the target performance analysis knowledge information, obtain a performance analysis result of the target application corresponding to the target data dimension.

[0020] The application performance analysis method based on a large language model of this embodiment achieves the purpose of performing performance analysis on applications running on a cloud server by using a large language model, thereby making use of the powerful content understanding and generation capabilities of the large language model to improve the efficiency and accuracy of performance analysis on the target application, thereby improving the efficiency and accuracy of solving performance problems that arise when the target application runs on the cloud server, and ensuring that the application continues to run on the cloud server with high performance.

[0021] In this embodiment, a large language model (LLM) is an artificial intelligence model trained with a large amount of data. It is based on deep learning technology and aims to understand and generate natural language text.

[0022] In this embodiment, the target application runs on a cloud server (the cloud server can be a physical machine or a virtual machine) to provide corresponding services; that is, this embodiment can perform performance analysis on a specific application (i.e., the target application) running on the cloud server to provide corresponding services based on the operation of the input end.

[0023] When executing S101, this embodiment can use at least one data dimension randomly selected from preset data dimensions as the target data dimension; or, based on actual needs, use at least one data dimension selected by the input end as the target data dimension; that is, the target data dimension in this embodiment can be one or more.

[0024] Preferably, in order to improve the comprehensiveness of the performance analysis of the target application, the target data dimensions in this embodiment are multiple.

[0025] In this embodiment, different data dimensions correspond to different indicator data types. Different types of indicator data refer to key data collected during the operation of the application and used to reflect the operating status of the application. The target application running on the cloud server includes one or more processes.

[0026] In this embodiment, the preset data dimensions include a first data dimension and a second data dimension corresponding to a machine (ie, a cloud server), and a third data dimension, a fourth data dimension, a fifth data dimension, and a sixth data dimension corresponding to an application.

[0027] Among them, the indicator data corresponding to the first data dimension may include hardware configuration parameters, system configuration parameters, runtime configuration parameters, etc. of the machine running the target application; the indicator data corresponding to the second data dimension may include the overall CPU usage data, overall memory usage data, overall network usage data, overall disk usage data, etc. of the machine during the running of the target application; the indicator data corresponding to the third data dimension may include CPU usage data, memory usage data, network usage data, disk usage data, etc. corresponding to the process being executed by the target application; the indicator data corresponding to the fourth data dimension may include CPU timeline data corresponding to the process being executed by the target application; the indicator data corresponding to the fifth data dimension may include function data, library data, instruction data, etc. corresponding to the process being executed by the target application; the indicator data corresponding to the sixth data dimension may include CPU top-down micro-architecture analysis (TMA) data corresponding to the process being executed by the target application.

[0028] Specifically, when executing S101 to obtain target indicator data corresponding to the target data dimension of the target application running on the cloud server, the implementation method that can be adopted in this embodiment is: collecting multiple indicator data when the target application is running on the cloud server; dividing the collected multiple indicator data into different data dimensions; obtaining the indicator data divided into the target data dimension as the target indicator data.

[0029] That is to say, this embodiment obtains target indicator data from multiple indicator data collected according to the target data dimension, which can improve the accuracy and flexibility of obtaining the target indicator data, and when there are multiple target data dimensions, it can also improve the efficiency of obtaining the target indicator data.

[0030] It is understandable that if there are multiple target data dimensions, this embodiment will obtain target indicator data corresponding to each target data dimension respectively when executing S101; wherein, the target indicator data corresponding to the same target data dimension may be one or more.

[0031] In addition, when executing S101, this embodiment may also only collect indicator data corresponding to the target data dimension, thereby directly acquiring the collected indicator data as the target indicator data.

[0032] After executing S101 to obtain target indicator data corresponding to the target data dimension, this embodiment executes S102 to use the large language model to obtain target summary information of the target application corresponding to the target data dimension according to the target indicator data.

[0033] The target summary information obtained by executing S102 in this embodiment is specifically the summary information that is output by the large language model based on the input target indicator data to understand the content and is simple, clear, and conclusive, and is used to reflect the operating characteristics of the target application corresponding to the target data dimension when the target application is running on the cloud server.

[0034] For example, if the target indicator data input into the large language model when executing S102 in this embodiment is the CPU usage percentage of a function in the process being executed by the target application, the large language model, based on its own powerful content understanding ability, performs content understanding according to the input "CPU usage percentage", and outputs target summary information such as "the CPU usage percentage of this function is high" or "the CPU usage percentage of this function is normal".

[0035] In actual application scenarios, if the acquired target indicator data is directly displayed to the user at the input end, the user at the input end may not understand the meaning of the target indicator data, or when the number of acquired target indicator data is large, the user at the input end cannot intuitively understand the specific information related to the program operation; therefore, this embodiment uses a large language model to summarize the input target indicator data, so that the large language model outputs target summary information corresponding to one or more target indicator data, and the user at the input end does not need to summarize it himself.

[0036] In addition, when the present embodiment uses the large language model during execution S102 to obtain target summary information of the target application corresponding to the target data dimension based on the target indicator data, the following method may also be used: obtaining a first prompt text (the prompt text is prompt), the first prompt text in the present embodiment is a prompt text used by the large language model to complete the task of summarizing the target indicator data; inputting the target indicator data and the first prompt text into the large language model, and obtaining the target summary information of the target application corresponding to the target data dimension based on the output result of the large language model.

[0037] In this embodiment, the first prompt text may be "understand the input target indicator data and output summary information that can reflect the operating characteristics of the application program."

[0038] That is to say, this embodiment makes use of the powerful content understanding ability of the large language model to obtain corresponding target summary information based on the target indicator data. It does not require the user at the input end to have professional knowledge, thereby reducing the difficulty in obtaining the target summary information; and when the acquired target indicator data is relatively complex, it can also improve the efficiency in obtaining the target summary information.

[0039] After executing S102 to obtain the target summary information, this embodiment executes S103 to obtain the target performance analysis knowledge information according to the target data dimension.

[0040] In this embodiment, the target performance analysis knowledge information is expert knowledge information corresponding to the target data dimension; wherein the expert knowledge information includes the operation bottleneck information of the application program and the optimization suggestions corresponding to the operation bottleneck information.

[0041] Specifically, when executing S103, this embodiment obtains target performance analysis knowledge information according to the target data dimension, the implementation method that can be adopted is: matching the target summary information corresponding to the target data dimension in the knowledge base; determining the preset summary information in the knowledge base that matches the target summary information; obtaining the performance analysis knowledge information corresponding to the matched preset summary information in the knowledge base as the target performance analysis knowledge information.

[0042] In this embodiment, the knowledge base is pre-constructed, and the knowledge base includes a plurality of different preset summary information and performance analysis knowledge information corresponding to each preset summary information; this embodiment can construct the knowledge base according to actual application optimization cases.

[0043] When executing S103, this embodiment may first calculate the similarity between the target summary information and different preset summary information in the knowledge base, and then use the preset summary information with the highest similarity calculation result as the preset summary information matching the target summary information.

[0044] That is to say, this embodiment can obtain target performance analysis knowledge information from the knowledge base based on preset summary information that is relatively similar to the target summary information. Since the preset performance analysis knowledge information in the knowledge base corresponds to the actual optimization case, it can improve the accuracy of the acquired target performance analysis knowledge information, thereby improving the accuracy of the performance analysis results obtained by using the target performance analysis knowledge information to guide the large language model.

[0045] In addition, when executing S103, this embodiment may also use the performance analysis knowledge information corresponding to the target data dimension as the target performance analysis knowledge information according to the correspondence between the preset data dimension and the performance analysis knowledge information.

[0046] It is understandable that when executing S103 to obtain target performance analysis knowledge information according to the target data dimension, this embodiment can also adopt the following implementation method: in response to determining that the operation of the target application is abnormal according to the obtained target summary information, obtain the target performance analysis knowledge information according to the target data dimension.

[0047] That is to say, this embodiment can execute the subsequent step of obtaining expert knowledge information according to the target data dimension only when it is determined that the operation of the target application is abnormal according to the target summary information, so that the operation abnormality of the target application can be repaired in time to optimize the performance of the target application when running on the cloud server.

[0048] In this embodiment, in addition to the summary information reflecting the operating characteristics of the target application corresponding to the target data dimension when running on the cloud server, the target summary information may also include the risk level of the target application when running on the cloud server, such as low risk level, medium risk level and high risk level.

[0049] In this embodiment, when executing S103, when it is determined that the risk level in the target summary information is a preset level (eg, a medium risk level or a high risk level), it is determined that the operation of the target application program is abnormal.

[0050] After executing S103 to obtain the target performance analysis knowledge information, this embodiment executes S104 to use the large language model to obtain the performance analysis result of the target application corresponding to the target data dimension according to the target summary information and the target performance analysis knowledge information.

[0051] In this embodiment, the acquired target performance analysis knowledge information is used for reference by the large language model, so that the large language model can combine the expert knowledge in the target performance analysis knowledge information to improve the accuracy of the performance analysis results output by the large language model.

[0052] Specifically, when this embodiment uses a large language model when executing S104 to obtain the performance analysis result of the target application corresponding to the target data dimension based on the target summary information and the target performance analysis knowledge information, the implementation method that can be adopted is: obtaining a second prompt text, the second prompt text in this embodiment is the prompt text used by the large language model to complete the performance analysis result generation task; inputting the target summary information, the target performance analysis knowledge information and the second prompt text into the large language model, and obtaining the performance analysis result of the target application corresponding to the target data dimension based on the output result of the large language model.

[0053] In this embodiment, the second prompt text may be "take the input target performance analysis knowledge information as a reference and generate a performance analysis result according to the input target summary information."

[0054] The performance analysis result obtained by executing S104 in this embodiment includes operation bottleneck information corresponding to the target data dimension when the target application is running on the cloud server, and optimization suggestions corresponding to the operation bottleneck information.

[0055] For example, in the performance analysis results obtained by executing S104 in this embodiment, the operating bottleneck information included is "the function library is not the latest version", that is, the operating performance of the target application is limited by the version of the function library, and the optimization suggestion corresponding to the operating bottleneck information may be "update the version of the function library to the latest version"; for another example, in the performance analysis results obtained by executing S104 in this embodiment, the operating bottleneck information included is "cross-NUMA (Non-Uniform Memory Access) memory usage", and the optimization suggestion corresponding to the operating bottleneck information may be "turn off cross-NUMA startup".

[0056] After executing S104 to obtain the performance analysis result of the target application corresponding to the target data dimension, this embodiment may further include the following contents: obtaining optimization suggestions from the performance analysis result; and optimizing the target application using the obtained optimization suggestions.

[0057] That is, this embodiment can optimize the target application in a timely manner according to the optimization suggestions included in the performance analysis results, thereby avoiding affecting the running performance of the target application.

[0058] In addition, after executing S104 and using the obtained optimization suggestions to optimize the target application, this embodiment may also include the following contents: storing the target summary information and performance analysis results corresponding to the target data dimension in the knowledge base; that is, this embodiment uses the performance analysis results output by the large language model this time as the expert knowledge information corresponding to the target summary information or the target data dimension, thereby continuously supplementing the expert knowledge in the knowledge base to further improve the knowledge base.

[0059] It can be understood that when executing S104 to store the target summary information and performance analysis results corresponding to the target data dimension in the knowledge base, this embodiment can also include the following contents: obtaining the optimization results after optimizing the target application using the optimization suggestions; in response to determining that the optimization results meet the preset conditions, storing the target summary information and performance analysis results corresponding to the target data dimension in the knowledge base.

[0060] For example, if the time consumed after optimization when the target application executes the process is reduced by 5% compared with the time consumed before optimization when the target application executes the process, this embodiment will not store the corresponding information in the knowledge base; if the time consumed after optimization is reduced by 20% compared with the time consumed before optimization, this embodiment will perform the operation of storing the corresponding information in the knowledge base.

[0061] Figure 2 is a schematic diagram according to the second embodiment of the present disclosure. Figure 2 As shown in , this embodiment shows a flowchart of application performance analysis based on a large language model: S201, obtaining indicator data corresponding to multiple data dimensions when the target application is running on a cloud server, such as all preset data dimensions; S202, inputting the indicator data corresponding to multiple data dimensions into the large language model respectively, and obtaining summary information output by the large language model for each data dimension; S203, combining with the knowledge base, and obtaining performance analysis knowledge information corresponding to different data dimensions according to the summary information corresponding to different data dimensions; S204, inputting the summary information corresponding to multiple data dimensions and the performance analysis knowledge information into the large language model respectively, and obtaining the performance analysis result output by the large language model for each data dimension.

[0062] Among them, this embodiment can provide a performance analysis page to the user of the input end, and the performance analysis page includes at least one application corresponding to the user of the input end. This embodiment can use the application selected by the user of the input end from the at least one application as the target application; in addition, the performance analysis page also includes multiple data dimensions. This embodiment can use the data dimension selected by the user of the input end from the multiple data dimensions as the target data dimension; when it is detected that the user of the input end has completed the selection of the target application and the target data dimension and clicked the performance analysis button in the performance analysis page, the above-mentioned application performance analysis process based on the large language model is executed.

[0063] Figure 3 is a schematic diagram according to the third embodiment of the present disclosure. Figure 3 As shown, the application performance analysis device 300 based on the large language model of this embodiment includes:

[0064] The first acquisition unit 301 is used to acquire target indicator data corresponding to a target data dimension of a target application running on a cloud server;

[0065] A summarizing unit 302 is configured to use a large language model to obtain target summary information of the target application corresponding to the target data dimension according to the target indicator data;

[0066] A second acquisition unit 303 is used to acquire target performance analysis knowledge information according to the target data dimension;

[0067] The analysis unit 304 is configured to use the large language model to obtain a performance analysis result of the target application corresponding to the target data dimension according to the target summary information and the target performance analysis knowledge information.

[0068] The first acquisition unit 301 may use at least one data dimension randomly selected from the preset data dimensions as the target data dimension; or, based on actual needs, may use at least one data dimension selected by the input end as the target data dimension; that is, the target data dimension in this embodiment may be one or more.

[0069] Preferably, in order to improve the comprehensiveness of the performance analysis of the target application, the target data dimensions in this embodiment are multiple.

[0070] In this embodiment, different data dimensions correspond to different indicator data types. Different types of indicator data refer to key data collected during the operation of the application and used to reflect the operating status of the application. The target application running on the cloud server includes one or more processes.

[0071] In this embodiment, the preset data dimensions include a first data dimension and a second data dimension corresponding to a machine (ie, a cloud server), and a third data dimension, a fourth data dimension, a fifth data dimension, and a sixth data dimension corresponding to an application.

[0072] Specifically, when the first acquisition unit 301 acquires the target indicator data corresponding to the target data dimension of the target application running on the cloud server, the implementation method that can be adopted is: collecting multiple indicator data when the target application is running on the cloud server; dividing the collected multiple indicator data into different data dimensions; and acquiring the indicator data divided into the target data dimension as the target indicator data.

[0073] That is to say, the first acquisition unit 301 acquires the target indicator data from the collected multiple indicator data according to the target data dimension, which can improve the acquisition accuracy and flexibility of the target indicator data, and in the case of multiple target data dimensions, can also improve the acquisition efficiency of the target indicator data.

[0074] It is understandable that if there are multiple target data dimensions, the first acquisition unit 301 will respectively acquire the target indicator data corresponding to each target data dimension; wherein, the target indicator data corresponding to the same target data dimension may be one or more.

[0075] In addition, the first acquisition unit 301 may also only collect indicator data corresponding to the target data dimension, thereby directly acquiring the collected indicator data as the target indicator data.

[0076] In this embodiment, after the first acquisition unit 301 acquires the target indicator data corresponding to the target data dimension, the summary unit 302 uses the large language model to obtain target summary information of the target application corresponding to the target data dimension according to the target indicator data.

[0077] The target summary information obtained by the summary unit 302 is specifically the summary information output by the large language model based on the input target indicator data to understand the content, which is simple, clear, and conclusive, and is used to reflect the operating characteristics of the target application corresponding to the target data dimension when running on the cloud server.

[0078] In actual application scenarios, if the acquired target indicator data is directly displayed to the user at the input end, the user at the input end may not understand the meaning of the target indicator data, or when the number of acquired target indicator data is large, the user at the input end cannot intuitively understand the specific information related to the program operation; therefore, this embodiment uses a large language model to summarize the input target indicator data, so that the large language model outputs target summary information corresponding to one or more target indicator data, and the user at the input end does not need to summarize it himself.

[0079] In addition, when the summary unit 302 uses the large language model to obtain the target summary information of the target application corresponding to the target data dimension according to the target indicator data, the following method can be adopted: obtaining a first prompt text (the prompt text is prompt), and the first prompt text in this embodiment is the prompt text used by the large language model to complete the task of summarizing the target indicator data; inputting the target indicator data and the first prompt text into the large language model, and obtaining the target summary information of the target application corresponding to the target data dimension according to the output result of the large language model.

[0080] In this embodiment, the first prompt text may be "understand the input target indicator data and output summary information that can reflect the operating characteristics of the application program."

[0081] That is to say, the summary unit 302 makes use of the powerful content understanding ability of the large language model to obtain corresponding target summary information based on the target indicator data, without requiring the user at the input end to have professional knowledge, thereby reducing the difficulty in obtaining the target summary information; and when the acquired target indicator data is relatively complex, it can also improve the efficiency in obtaining the target summary information.

[0082] In this embodiment, after the summarizing unit 302 obtains the target summary information, the second obtaining unit 303 obtains the target performance analysis knowledge information according to the target data dimension.

[0083] In this embodiment, the target performance analysis knowledge information is expert knowledge information corresponding to the target data dimension; wherein the expert knowledge information includes the operation bottleneck information of the application program and the optimization suggestions corresponding to the operation bottleneck information.

[0084] Specifically, when the second acquisition unit 303 acquires the target performance analysis knowledge information according to the target data dimension, the implementation method that can be adopted is: matching the target summary information corresponding to the target data dimension in the knowledge base; determining the preset summary information in the knowledge base that matches the target summary information; acquiring the performance analysis knowledge information corresponding to the matched preset summary information in the knowledge base as the target performance analysis knowledge information.

[0085] In this embodiment, the knowledge base is pre-constructed, and the knowledge base includes a plurality of different preset summary information and performance analysis knowledge information corresponding to each preset summary information; this embodiment can construct the knowledge base according to actual application optimization cases.

[0086] The second acquisition unit 303 may first calculate the similarity between the target summary information and different preset summary information in the knowledge base, and then select the preset summary information with the highest similarity calculation result as the preset summary information matching the target summary information.

[0087] That is to say, the second acquisition unit 303 can obtain the target performance analysis knowledge information from the knowledge base based on the preset summary information that is relatively similar to the target summary information. Since the preset performance analysis knowledge information in the knowledge base corresponds to the actual optimization case, it can improve the accuracy of the acquired target performance analysis knowledge information, and thus improve the accuracy of the performance analysis results obtained by using the target performance analysis knowledge information to guide the large language model.

[0088] In addition, the second acquisition unit 303 may also use the performance analysis knowledge information corresponding to the target data dimension as the target performance analysis knowledge information according to the correspondence between the preset data dimension and the performance analysis knowledge information.

[0089] It is understandable that when the second acquisition unit 303 acquires the target performance analysis knowledge information according to the target data dimension, the following implementation method can also be adopted: in response to determining that the operation of the target application is abnormal according to the obtained target summary information, the target performance analysis knowledge information is acquired according to the target data dimension.

[0090] That is to say, the second acquisition unit 303 can execute the subsequent step of acquiring expert knowledge information according to the target data dimension only when it is determined that the operation of the target application is abnormal according to the target summary information, so that the operation abnormality of the target application can be repaired in time to optimize the performance of the target application when running on the cloud server.

[0091] In this embodiment, in addition to the summary information reflecting the operating characteristics of the target application corresponding to the target data dimension when running on the cloud server, the target summary information may also include the risk level of the target application when running on the cloud server, such as low risk level, medium risk level and high risk level.

[0092] The second acquisition unit 303 may determine that the operation of the target application program is abnormal when it is determined that the risk level in the target summary information is a preset level (eg, a medium risk level or a high risk level).

[0093] In this embodiment, after the second acquisition unit 303 acquires the target performance analysis knowledge information, the analysis unit 304 uses the large language model to obtain the performance analysis result of the target application corresponding to the target data dimension according to the target summary information and the target performance analysis knowledge information.

[0094] In this embodiment, the acquired target performance analysis knowledge information is used for reference by the large language model, so that the large language model can combine the expert knowledge in the target performance analysis knowledge information to improve the accuracy of the performance analysis results output by the large language model.

[0095] Specifically, when the analysis unit 304 uses the large language model to obtain the performance analysis result of the target application corresponding to the target data dimension according to the target summary information and the target performance analysis knowledge information, the implementation method that can be adopted is: obtaining a second prompt text, the second prompt text in this embodiment is the prompt text used by the large language model to complete the performance analysis result generation task; inputting the target summary information, the target performance analysis knowledge information and the second prompt text into the large language model, and obtaining the performance analysis result of the target application corresponding to the target data dimension according to the output result of the large language model.

[0096] In this embodiment, the second prompt text may be "take the input target performance analysis knowledge information as a reference and generate a performance analysis result according to the input target summary information."

[0097] The performance analysis result obtained by the analysis unit 304 includes operation bottleneck information corresponding to the target data dimension when the target application is running on the cloud server, and optimization suggestions corresponding to the operation bottleneck information.

[0098] The application performance analysis device 300 based on a large language model of this embodiment may also include an optimization unit 305, which is used to perform the following contents after the analysis unit 304 obtains the performance analysis results of the target application corresponding to the target data dimension: obtain optimization suggestions from the performance analysis results; and optimize the target application using the obtained optimization suggestions.

[0099] That is, the optimization unit 305 can optimize the target application in a timely manner according to the optimization suggestions included in the performance analysis result, so as to avoid affecting the running performance of the target application.

[0100] In addition, after the optimization unit 305 uses the obtained optimization suggestions to optimize the target application, it can also include the following contents: storing the target summary information and performance analysis results corresponding to the target data dimension in the knowledge base; that is, this embodiment uses the performance analysis results output by the large language model this time as the expert knowledge information corresponding to the target summary information or the target data dimension, thereby continuously supplementing the expert knowledge in the knowledge base to further improve the knowledge base.

[0101] It can be understood that when the optimization unit 305 stores the target summary information and performance analysis results corresponding to the target data dimension in the knowledge base, it can also include the following contents: obtaining the optimization results after optimizing the target application using the optimization suggestions; in response to determining that the optimization results meet the preset conditions, storing the target summary information and performance analysis results corresponding to the target data dimension in the knowledge base.

[0102] In the technical solution disclosed herein, the acquisition, storage and application of user personal information involved are in compliance with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0103] According to an embodiment of the present disclosure, the present disclosure also provides an electronic device, a readable storage medium and a computer program product.

[0104] like Figure 4 , is a block diagram of an electronic device according to an application performance analysis method based on a large language model according to an embodiment of the present disclosure. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workbenches, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present disclosure described and / or required herein.

[0105] like Figure 4 As shown, the device 400 includes a computing unit 401, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 402 or a computer program loaded from a storage unit 408 into a random access memory (RAM) 403. In the RAM 403, various programs and data required for the operation of the device 400 can also be stored. The computing unit 401, the ROM 402, and the RAM 403 are connected to each other via a bus 404. An input / output (I / O) interface 405 is also connected to the bus 404.

[0106] A number of components in the device 400 are connected to the I / O interface 405, including: an input unit 406, such as a keyboard, a mouse, etc.; an output unit 407, such as various types of displays, speakers, etc.; a storage unit 408, such as a disk, an optical disk, etc.; and a communication unit 409, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 409 allows the device 400 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0107] The computing unit 401 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the computing unit 401 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, digital signal processors (DSPs), and any appropriate processors, controllers, microcontrollers, etc. The computing unit 401 performs the various methods and processes described above, such as an application performance analysis method based on a large language model. For example, in some embodiments, the application performance analysis method based on a large language model may be implemented as a computer software program, which is tangibly contained in a machine-readable medium, such as a storage unit 408.

[0108] In some embodiments, part or all of the computer program may be loaded and / or installed on the device 400 via the ROM 402 and / or the communication unit 409. When the computer program is loaded into the RAM 403 and executed by the computing unit 401, one or more steps of the application performance analysis method based on the large language model described above may be performed. Alternatively, in other embodiments, the computing unit 401 may be configured to perform the application performance analysis method based on the large language model in any other appropriate manner (e.g., by means of firmware).

[0109] Various implementations of the systems and techniques described herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), complex programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0110] The program code for implementing the method of the present disclosure may be written in any combination of one or more programming languages. These program codes may be provided to a processor or controller of a general-purpose computer, a special-purpose computer, or other programmable large language model-based application performance analysis device, so that the program code, when executed by the processor or controller, enables the functions / operations specified in the flow chart and / or block diagram to be implemented. The program code may be executed entirely on the machine, partially on the machine, partially on the machine as a stand-alone software package and partially on a remote machine, or entirely on a remote machine or server.

[0111] In the context of the present disclosure, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or equipment, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0112] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or trackball) through which the user can provide input to the computer. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0113] The systems and techniques described herein may be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes front-end components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such back-end components, middleware components, or front-end components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), and the Internet.

[0114] A computer system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The relationship between the client and the server is generated by computer programs running on the corresponding computers and having a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or a cloud host, which is a host product in the cloud computing service system to solve the defects of difficult management and weak business scalability in traditional physical hosts and VPS services ("Virtual Private Server", or "VPS" for short). The server may also be a server for a distributed system, or a server combined with a blockchain.

[0115] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps recorded in this disclosure can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solutions disclosed in this disclosure can be achieved, and this document does not limit this.

[0116] The above specific implementations do not constitute a limitation on the protection scope of the present disclosure. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present disclosure shall be included in the protection scope of the present disclosure.

Claims

1. A method for analyzing application performance based on a large language model, comprising: Obtain target indicator data corresponding to a target data dimension of a target application running on the cloud server; Using a large language model, obtaining target summary information of the target application corresponding to the target data dimension according to the target indicator data; According to the target data dimension, obtaining target performance analysis knowledge information; The large language model is used to obtain a performance analysis result of the target application corresponding to the target data dimension according to the target summary information and the target performance analysis knowledge information.

2. The method according to claim 1, wherein: The step of obtaining target indicator data corresponding to a target data dimension of a target application program running on the cloud server includes: Collecting multiple indicator data when the target application is running on the cloud server; Dividing the multiple indicator data into different data dimensions; The indicator data divided into the target data dimension is obtained as the target indicator data.

3. The method according to claim 1, wherein: The acquiring target performance analysis knowledge information according to the target data dimension includes: In response to determining, according to the target summary information, that the operation of the target application program is abnormal, target performance analysis knowledge information is acquired according to the target data dimension.

4. The method according to claim 1 or 3, wherein: The acquiring target performance analysis knowledge information according to the target data dimension includes: Matching the target summary information corresponding to the target data dimension in the knowledge base; Determining preset summary information in the knowledge base that matches the target summary information; The performance analysis knowledge information corresponding to the matched preset summary information in the knowledge base is obtained as the target performance analysis knowledge information.

5. The method according to claim 1, wherein: The using of the large language model to obtain target summary information of the target application corresponding to the target data dimension according to the target indicator data includes: Get the first prompt text; Inputting the target indicator data and the first prompt text into the large language model; According to the output result of the large language model, target summary information of the target application corresponding to the target data dimension is obtained.

6. The method according to claim 1, wherein: The using the large language model to obtain the performance analysis result of the target application corresponding to the target data dimension according to the target summary information and the target performance analysis knowledge information includes: Get the second prompt text; Inputting the target summary information, the target performance analysis knowledge information and the second prompt text into the large language model; According to the output result of the large language model, a performance analysis result of the target application corresponding to the target data dimension is obtained.

7. The method according to claim 1, further comprising: After obtaining the performance analysis result of the target application corresponding to the target data dimension, obtaining optimization suggestions from the performance analysis result; The target application running on the cloud server is optimized using the optimization suggestion.

8. The method according to claim 7, further comprising: After the target application running on the cloud server is optimized using the optimization suggestion, the target summary information and performance analysis results corresponding to the target data dimension are stored in the knowledge base.

9. An application performance analysis device based on a large language model, comprising: A first acquisition unit is used to acquire target indicator data corresponding to a target data dimension of a target application program running on the cloud server; A summary unit, configured to obtain target summary information of the target application corresponding to the target data dimension according to the target indicator data using a large language model; A second acquisition unit, configured to acquire target performance analysis knowledge information according to the target data dimension; The analysis unit is used to use the large language model to obtain a performance analysis result of the target application corresponding to the target data dimension according to the target summary information and the target performance analysis knowledge information.

10. The device according to claim 9, wherein: When the first acquisition unit acquires the target indicator data corresponding to the target data dimension of the target application running on the cloud server, the first acquisition unit specifically performs: Collecting multiple indicator data when the target application is running on the cloud server; Dividing the multiple indicator data into different data dimensions; The indicator data divided into the target data dimension is obtained as the target indicator data.

11. The device according to claim 9, wherein: When the second acquisition unit acquires the target performance analysis knowledge information according to the target data dimension, the second acquisition unit specifically performs: In response to determining, according to the target summary information, that the operation of the target application program is abnormal, target performance analysis knowledge information is acquired according to the target data dimension.

12. The device according to claim 9 or 11, wherein: When the second acquisition unit acquires the target performance analysis knowledge information according to the target data dimension, the second acquisition unit specifically performs: Matching the target summary information corresponding to the target data dimension in the knowledge base; Determining preset summary information in the knowledge base that matches the target summary information; The performance analysis knowledge information corresponding to the matched preset summary information in the knowledge base is obtained as the target performance analysis knowledge information.

13. The device according to claim 9, wherein: When the summary unit uses the large language model to obtain the target summary information of the target application corresponding to the target data dimension according to the target indicator data, specifically: Get the first prompt text; Inputting the target indicator data and the first prompt text into the large language model; According to the output result of the large language model, target summary information of the target application corresponding to the target data dimension is obtained.

14. The device according to claim 9, wherein: When the analysis unit uses the large language model to obtain the performance analysis result of the target application corresponding to the target data dimension according to the target summary information and the target performance analysis knowledge information, the analysis unit specifically performs: Get the second prompt text; Inputting the target summary information, the target performance analysis knowledge information and the second prompt text into the large language model; According to the output result of the large language model, a performance analysis result of the target application corresponding to the target data dimension is obtained.

15. The apparatus according to claim 9, further comprising an optimization unit, configured to perform: After the analysis unit obtains the performance analysis result of the target application corresponding to the target data dimension, obtaining optimization suggestions from the performance analysis result; The target application running on the cloud server is optimized using the optimization suggestion.

16. The apparatus according to claim 15, wherein the optimization unit is further configured to perform: After the target application running on the cloud server is optimized using the optimization suggestion, the target summary information and performance analysis results corresponding to the target data dimension are stored in the knowledge base.

17. An electronic device comprising: at least one processor; as well as a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method according to any one of claims 1 to 8.

18. A non-transitory computer-readable storage medium storing computer instructions, wherein: The computer instructions are used to cause the computer to execute the method according to any one of claims 1 to 8.

19. A computer program product, comprising a computer program, which, when executed by a processor, implements the method according to any one of claims 1 to 8.

Images