A business system configuration method, apparatus, equipment and medium

By using pre-set monitoring tools and reinforcement learning algorithms to automatically adjust the configuration of the business system, the problem of determining the optimal parameters under real-time business load changes was solved, improving system performance and stability and reducing manual adjustment costs.

CN116382799BActive Publication Date: 2026-11-10SHANDONG INSPUR SCI RES INST CO LTD
View PDF 1 Cites 0 Cited by

Patent Information

Application Number
CN202310354600.7
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2023-03-31
Publication Date
2026-11-10
Estimated Expiration
2043-03-31

AI Technical Summary

Technical Problem

Existing technologies cannot quickly and accurately find and dynamically adjust the optimal business system configuration parameters based on real-time business load changes, resulting in the inability to guarantee optimal system performance and stability.

Method used

By employing pre-set monitoring tools and reinforcement learning algorithms, the system collects business load information and operating system performance indicators in real time, extracts feature information, calculates reward values ​​using reward functions, and uses reinforcement learning algorithms to determine the optimal configuration parameters, thereby automatically adjusting the system configuration.

Benefits of technology

It enables the system to quickly and accurately find and apply the best configuration parameters based on real-time business load changes, improving the rationality and effectiveness of system configuration parameter optimization and reducing manual adjustment costs.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN116382799B_ABST
    Figure CN116382799B_ABST
Patent Text Reader

Abstract

This application discloses a business system configuration method, apparatus, device, and medium, relating to the field of intelligent operation and maintenance. The method includes: using a preset monitoring tool to collect the business load and target performance indicators in preset application software and operating system at each sampling time according to a preset sampling frequency; extracting target feature information from the target performance indicators and business load; and calculating the reward value at each time using a reward function determined based on the target performance indicators; analyzing and calculating the business load, target performance indicators, target feature information, and reward value at each time using a preset reinforcement learning algorithm to determine the target operating mode and the target parameter configuration scheme corresponding to the target operating mode at each time; and configuring the current business system according to the target operating mode and target parameter configuration scheme corresponding to each sampling time. This invention uses reinforcement learning to optimize the adjustment process of configuration parameters, improving the rationality of parameter optimization.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of intelligent operation and maintenance, and in particular to a business system configuration method, apparatus, equipment and medium. Background Technology

[0002] As businesses expand, servers or large-scale clusters typically run multiple business systems simultaneously, generating varying workloads. These include I / O (Input / Output) intensive storage tasks and compute-intensive processing. Furthermore, different tasks may run at different times depending on the business characteristics, such as supporting high-concurrency user access during the day and performing statistical analysis at night. In complex application scenarios involving high concurrency, big data, and high loads, optimizing the software and underlying operating system based on real-time workloads is crucial to ensuring the performance and reliability of business software and improving the resource utilization of servers and large-scale clusters.

[0003] Traditional optimization methods are primarily based on experience and manual testing. This approach is inefficient, cannot guarantee finding the optimal configuration parameters, and the configuration parameters of common application software and operating systems are static, unable to dynamically adjust according to real-time changes in business load. This results in suboptimal system performance and stability. Manually adjusting configuration parameters under varying business loads is a time-consuming and complex task, requiring experienced technical personnel. Regarding automated performance optimization, domestic scholars have proposed many effective methods. These methods generally focus on optimizing the operating system or a single business scenario, and some utilize artificial intelligence to simplify the manual process, achieving good results. However, when running with multiple types of business loads, they cannot dynamically adjust configuration parameters according to changes in business requirements.

[0004] As can be seen from the above, how to quickly and accurately find and apply the best configuration parameters based on real-time changes in business load during the operation of a business system is a problem that needs to be solved in this field. Summary of the Invention

[0005] In view of this, the purpose of this invention is to provide a business system configuration method, apparatus, device, and medium that can quickly and accurately find and apply the optimal configuration parameters based on real-time changes in business load. The specific solution is as follows:

[0006] Firstly, this application discloses a business system configuration method, including:

[0007] Using a preset monitoring tool, the system collects business load information and target performance indicators in the preset application software and operating system at each sampling time according to a preset sampling frequency. Then, it extracts target feature information from the target performance indicators and the business load information, and calculates the reward value at each sampling time using a reward function determined based on the target performance indicators.

[0008] The service load information, target performance indicators, target feature information and reward value at each sampling time are analyzed and calculated using a preset reinforcement learning algorithm to determine the target operating mode and the target parameter configuration scheme corresponding to each sampling time.

[0009] Configure the current business system according to the target operating mode and target parameter configuration scheme corresponding to each sampling time.

[0010] Optionally, the step of using a preset monitoring tool to collect service load information and target performance indicators in the preset application software and operating system at each sampling time according to a preset sampling frequency includes:

[0011] The Prometheus tool is used to collect service load information and target performance indicators in the application software and operating system at each sampling time according to the preset sampling frequency.

[0012] Optionally, the step of extracting target feature information from the target performance metric and the service load information, and calculating the reward value at each sampling time using the reward function determined based on the target performance metric, includes:

[0013] Target feature information is extracted from the target performance indicators and the business load information, and the target feature information is converted into target state variables. Then, the reward value at each sampling time is calculated using the reward function determined based on the target state variables.

[0014] Optionally, the step of using a preset reinforcement learning algorithm to analyze and calculate the service load information, the target performance index, the target feature information, and the reward value at each sampling time to determine the target operating mode corresponding to each sampling time and the target parameter configuration scheme corresponding to the target operating mode includes:

[0015] Based on a preset reinforcement learning algorithm, the target feature information, the target state variable, and the reward value at each sampling time are analyzed and calculated. The target system parameters are converted into action variables. The target operating mode and the target parameter configuration scheme corresponding to each sampling time are determined by using the action variables, the business load information, the target feature information, and the reward value at each sampling time.

[0016] Optionally, the step of using a preset reinforcement learning algorithm to analyze and calculate the service load information, the target performance index, the target feature information, and the reward value at each sampling time to determine the target operating mode corresponding to each sampling time and the target parameter configuration scheme corresponding to the target operating mode includes:

[0017] The service load information, target performance indicators, target feature information and reward value at each sampling time are analyzed and calculated using a preset reinforcement learning algorithm to determine the target operating mode and the target parameter configuration scheme corresponding to the highest reward value at each sampling time.

[0018] Optionally, determining the target operating mode corresponding to each sampling time and the target parameter configuration scheme corresponding to the target operating mode includes:

[0019] Determine the target operating mode and the target parameter configuration scheme corresponding to each sampling time, and determine the target action variables corresponding to the target system parameters of the target parameter configuration scheme;

[0020] Accordingly, configuring the current business system according to the target operating mode and the target parameter configuration scheme corresponding to each sampling time includes:

[0021] Create an executable script to adjust the target system parameters in the current business system, and use the executable script to configure the current business system according to the target operating mode and the target action variable corresponding to the sampling time.

[0022] Optionally, the creation of an executable script for adjusting target system parameters in the current business system includes:

[0023] Create an executable script to adjust the target system parameters in the current business system using preset system parameter adjustment commands and preset application software parameter adjustment commands.

[0024] Secondly, this application discloses a business system configuration device, comprising:

[0025] The data processing module is used to collect business load information and target performance indicators in the preset application software and operating system at each sampling time using a preset monitoring tool at a preset sampling frequency. Then, it extracts target feature information from the target performance indicators and the business load information, and calculates the reward value at each sampling time using a reward function determined based on the target performance indicators.

[0026] The scheme determination module is used to analyze and calculate the service load information, the target performance index, the target feature information and the reward value at each sampling time using a preset reinforcement learning algorithm, so as to determine the target operation mode and the target parameter configuration scheme corresponding to each sampling time.

[0027] The scheme configuration module is used to configure the current business system according to the target operating mode and the target parameter configuration scheme corresponding to each sampling time.

[0028] Thirdly, this application discloses an electronic device, comprising:

[0029] Memory, used to store computer programs;

[0030] A processor is used to execute the computer program to implement the aforementioned business system configuration method.

[0031] Fourthly, this application discloses a computer storage medium for storing a computer program; wherein, when the computer program is executed by a processor, it implements the steps of the aforementioned disclosed business system configuration method.

[0032] This application first uses a preset monitoring tool to collect business load information and target performance indicators in the preset application software and operating system at each sampling time according to a preset sampling frequency. Then, it extracts target feature information from the target performance indicators and business load information, and calculates the reward value at each sampling time using a reward function determined based on the target performance indicators. A preset reinforcement learning algorithm is then used to analyze and calculate the business load information, target performance indicators, target feature information, and reward value at each sampling time to determine the target operating mode and the target parameter configuration scheme corresponding to each sampling time. Finally, the current business system is configured according to the target operating mode and target parameter configuration scheme corresponding to each sampling time. In this way, this invention extracts feature information by monitoring the key performance indicators of the application software and the underlying supporting operating system, as well as the load of the current business system. The feature information is used as a reward function to calculate the reward value of the current state. Based on this, a reinforcement learning algorithm is used for training. Combined with stored historical data, load distribution, state information features, and historical reward values, the analysis and calculation are performed to select a matching parameter configuration scheme as the optimal adjustment method, enabling the operating system and application software to adapt to the current business load. In this embodiment, a deep reinforcement learning method is used to optimize the adjustment of configuration parameters, which improves the rationality and effectiveness of the system configuration parameter optimization. The entire system configuration parameter adjustment process is fully automated, which greatly reduces the cost of manual adjustment. Attached Figure Description

[0033] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0034] Figure 1 A flowchart of a business system configuration method provided in this application;

[0035] Figure 2 A flowchart illustrating a specific business system configuration method provided in this application;

[0036] Figure 3 A schematic diagram of a business system configuration device provided in this application;

[0037] Figure 4 This application provides a schematic diagram of a business system configuration device structure categorized by functional type.

[0038] Figure 5 This application provides a structural diagram of an electronic device. Detailed Implementation

[0039] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0040] In existing technologies, traditional business system optimization methods are mainly based on experience and manual testing. This approach is inefficient, cannot guarantee finding the optimal configuration parameters, and the configuration parameters of common application software and operating systems are static, unable to dynamically adjust according to real-time changes in business load, resulting in suboptimal system performance and stability. Regarding automated performance optimization, when multiple types of business loads are running, it is impossible to dynamically adjust configuration parameters based on changes in business operations. This application provides an automated optimization method that can quickly and accurately find and apply the optimal configuration parameters based on real-time changes in business load.

[0041] This invention discloses a business system configuration method, see [link to relevant documentation]. Figure 1 The method includes:

[0042] Step S11: Use a preset monitoring tool to collect service load information and target performance indicators in the preset application software and operating system at each sampling time according to a preset sampling frequency. Then, extract target feature information from the target performance indicators and the service load information, and use the reward function determined based on the target performance indicators to calculate the reward value at each sampling time.

[0043] In this embodiment, the preset monitoring tool collects service load information and target performance indicators in the preset application software and operating system at each sampling time according to a preset sampling frequency, and compares the load of each service under the current indicator conditions. Specifically, the preset monitoring tool monitors the key performance indicators of the application software and the underlying supporting operating system, and determines the load of each service under the current indicator conditions. Next, in this embodiment, target feature information is extracted from the target performance indicators and the service load information, and the reward value at each sampling time is calculated using a reward function determined based on the target performance indicators.

[0044] In a specific implementation, the process of calculating the reward value at each sampling moment using the reward function determined based on the target performance index may include: calculating the reward value at each sampling moment using a preset reward value calculation method based on the reward function determined based on the target performance index; wherein the preset reward value calculation method includes, but is not limited to, a calculation method based on the concurrent access volume of WEB (i.e., World Wide Web) services, a calculation method based on the processing time of statistical analysis services, and a calculation method based on the CPU (i.e., Central Processing Unit) utilization rate when processing the same service. Specifically, higher concurrent access volume of WEB services will result in higher rewards, shorter processing time of statistical analysis services will result in higher rewards, and lower CPU utilization rate when processing the same service will result in higher rewards. Furthermore, it is understood that the preset reward value calculation method can be flexibly formulated by the system administrator and modified according to different application scenarios.

[0045] In this embodiment, the step of collecting service load information and target performance indicators in the preset application software and operating system at each sampling time using a preset monitoring tool at a preset sampling frequency may include: using the Prometheus tool to collect service load information and target performance indicators in the preset application software and operating system at each sampling time at a preset sampling frequency. It is understood that, in specific implementations, the preset monitoring tool is a tool used to collect target performance indicators, including but not limited to Prometheus.

[0046] Step S12: Analyze and calculate the service load information, target performance indicators, target feature information and reward value at each sampling time using a preset reinforcement learning algorithm to determine the target operating mode and the target parameter configuration scheme corresponding to each sampling time.

[0047] In this embodiment, the step of using a preset reinforcement learning algorithm to analyze and calculate the service load information, target performance index, target feature information, and reward value at each sampling time to determine the target operating mode and the target parameter configuration scheme corresponding to the target operating mode at each sampling time may include: using a preset reinforcement learning algorithm to analyze and calculate the service load information, target performance index, target feature information, and reward value at each sampling time to determine the target operating mode corresponding to the highest reward value at each sampling time and the target parameter configuration scheme corresponding to the target operating mode. That is, in this embodiment, a reinforcement learning algorithm can be used to determine a matching parameter configuration scheme with a high reward value based on the data corresponding to each sampling time. It can be understood that the preset reinforcement learning algorithm can identify the target operating mode corresponding to each sampling time in multiple analysis calculations and determine the target parameter configuration scheme that best matches the current business requirements when running in the current target operating mode at each sampling time.

[0048] In specific implementations, the business system may maintain different operating modes at different times. For example, during the day, it supports high-concurrency web page access to ensure normal business operation, while at night, the concurrency of business access decreases, but data analysis, statistics, and report generation are still required. In this case, the daytime is the first high-concurrency operating mode, and the nighttime is the second analysis and statistics mode. In this embodiment, the target operating mode of the business system at each sampling time can be identified based on the load information corresponding to each sampling time and the operating information in the application software and operating system, and the optimal target parameter configuration scheme corresponding to the target operating mode at the current time can be determined.

[0049] In this embodiment, the target parameter configuration scheme corresponding to each sampling time includes the optimal configuration parameters of the target system parameters at the current sampling time. Specifically, the target parameter configuration scheme corresponding to each sampling time includes the optimal operating system configuration parameters and application software configuration parameters of the target system parameters at the current sampling time.

[0050] In specific implementation, this embodiment can also save the business load information corresponding to each moment and the target performance indicators, target feature information and reward value in the preset application software and operating system, so that when determining the target parameter configuration scheme corresponding to each moment in the future, the above data corresponding to each moment can be analyzed and calculated as historical data to determine the target parameter configuration scheme corresponding to each moment in real time, thereby facilitating the adaptive configuration of the business system in real time.

[0051] Step S13: Configure the current business system according to the target operating mode and the target parameter configuration scheme corresponding to each sampling time.

[0052] In this embodiment, after determining the target operating mode of the business system at each sampling time and the optimal target parameter configuration scheme corresponding to the target operating mode at the current time based on the load information corresponding to each sampling time and the running information in the application software and operating system, the current business system can be automatically configured according to the target operating mode and the target parameter configuration scheme corresponding to each sampling time.

[0053] In this embodiment, a preset monitoring tool is first used to collect business load information and target performance indicators in the preset application software and operating system at each sampling time according to a preset sampling frequency. Then, target feature information is extracted from the target performance indicators and business load information, and a reward value is calculated for each sampling time using a reward function determined based on the target performance indicators. A preset reinforcement learning algorithm is used to analyze and calculate the business load information, target performance indicators, target feature information, and reward value for each sampling time to determine the target operating mode and the target parameter configuration scheme corresponding to the target operating mode for each sampling time. The current business system is then configured according to the target operating mode and target parameter configuration scheme corresponding to each sampling time. In this way, this embodiment extracts feature information by monitoring the key performance indicators of the application software and the underlying supporting operating system and the load of the current business system. The feature information is used as a reward function to calculate the reward value of the current state. Based on this, a reinforcement learning algorithm is used for training. Combined with stored historical data, load distribution, state information features, and historical reward values, the algorithm is analyzed and calculated to select a matching parameter configuration scheme with a high reward value as the current optimal adjustment method, so that the operating system and application software can adapt to the current business load. In this embodiment, a deep reinforcement learning method is used to optimize the adjustment of configuration parameters, which improves the rationality and effectiveness of the system configuration parameter optimization. The entire system configuration parameter adjustment process is fully automated, which greatly reduces the cost of manual adjustment.

[0054] Figure 2A flowchart illustrating a specific business system configuration method provided in this application embodiment. See also... Figure 2 As shown, the method includes:

[0055] Step S21: Use a preset monitoring tool to collect service load information and target performance indicators in the preset application software and operating system at each sampling time according to a preset sampling frequency. Then extract target feature information from the target performance indicators and the service load information, and convert the target feature information into target state variables. Then use the reward function determined based on the target state variables to calculate the reward value at each sampling time.

[0056] In a specific implementation, the current business system can be a high-concurrency web system that supports high-concurrency web page access during the day to ensure normal business operations. At night, the concurrency of business access decreases, but data analysis, statistics, and report generation are still required. This web system is deployed on a Linux operating system, using Nginx as a reverse proxy server and employing Nginx for load balancing to improve service availability and performance. Business data is stored in a MySQL database, and MySQL (Relational Database Management System) is used for data storage and management. At night, the system may need to execute a large number of SQL (Structured Query Language) operations for data statistics and report generation to meet business needs. Therefore, this step specifically utilizes the Prometheus monitoring tool to collect metrics information from Nginx, MySQL, and the Linux operating system. Key Nginx metrics can reflect the real-time load of concurrent web access, while MySQL metrics can reflect the business load of statistical analysis and report generation at night. Linux operating system metrics can also reflect the usage and stress of CPU, memory, file system, network, etc. Table 1 below lists some performance feedback metrics. The table provides various metric names, the types of metric names, and the descriptions of the metric names.

[0057] Table 1

[0058]

[0059] Specifically, in this embodiment, the target feature information is converted into target state variables, and then the reward value at each sampling time is calculated using a reward function determined based on the target state variables. In specific implementation, the format of the reward value calculation formula can be: Among them, W n The weight of the nth state variable. This is the normalized value of the nth state variable.

[0060] Step S22: Based on a preset reinforcement learning algorithm, analyze and calculate the target feature information, the target state variable, and the reward value at each sampling time, convert the target system parameters into action variables, and use the action variables, the business load information, the target feature information, and the reward value at each sampling time to determine the target operating mode and the target parameter configuration scheme corresponding to the target operating mode, and determine the target action variables corresponding to the target system parameters corresponding to the target parameter configuration scheme.

[0061] In this embodiment, a preset reinforcement learning algorithm is used to analyze and calculate the target feature information, the target state variables, and the reward value at each sampling time, while simultaneously converting adjustable target system parameters into action variables. Specifically, the target system parameters include, but are not limited to:

[0062] 1. `net.ipv4.tcp_tw_reuse`: This system parameter specifies whether TCP connections in the TIME-WAIT state are allowed to be reused for new connections. In high-concurrency Nginx scenarios, it is recommended to enable this feature to improve connection reuse. However, in MySQL scenarios handling complex business logic and statistical analysis, it is recommended to disable this feature to avoid problems caused by excessive TIME-WAIT connections.

[0063] 2. `net.ipv4.tcp_syncookies`: This system parameter specifies whether to enable TCP (Transmission Control Protocol) SYN (Synchronize Sequence Number) cookies. In high-concurrency Nginx scenarios, it is recommended to enable this feature to improve system security and resistance to SYN attacks. However, in MySQL scenarios involving complex business logic and statistical analysis, it is recommended to disable this feature to avoid impacting the system.

[0064] 3. `vm.swappiness`: This system parameter specifies the level of swap usage. In high-concurrency Nginx scenarios, it is recommended to set a lower value to avoid performance issues caused by excessive memory swapping. However, in MySQL scenarios involving complex business logic and statistical analysis, it is recommended to set a higher value to improve system performance.

[0065] 4. `kernel.sem`: This system parameter specifies the kernel's System V semaphore configuration. In high-concurrency Nginx scenarios, it is recommended to increase the number of semaphores to improve system concurrency performance. However, in MySQL scenarios involving complex business logic and statistical analysis, it is recommended to adjust the configuration according to the actual situation to avoid performance issues caused by insufficient semaphores.

[0066] 5. `net.ipv4.tcp_max_syn_backlog`: This system parameter specifies the maximum length of the TCP SYN queue. In high-concurrency Nginx scenarios, it is recommended to increase this value to improve system concurrency performance. However, in MySQL scenarios involving complex business logic and statistical analysis, it is recommended to appropriately reduce this value to avoid performance issues caused by excessive connections.

[0067] 6. `net.ipv4.tcp_max_tw_buckets`: This system parameter specifies the maximum TIME-WAIT queue length for TCP connections. In high-concurrency Nginx scenarios, it is recommended to increase this value to improve system concurrency performance. However, in MySQL scenarios handling complex business logic and statistical analysis, it is recommended to appropriately reduce this value to avoid performance issues caused by excessive TIME-WAIT connections.

[0068] After converting the target system parameters into action variables, for the action variables corresponding to the application software, the performance and reliability of the application software can be improved by modifying its configuration parameters. For the action variables corresponding to the operating system, CPU, memory, disk I / O, and network are among the core resources of the system. CPU utilization, memory utilization, disk I / O, and network throughput can reflect the system's load and performance bottlenecks. We can optimize CPU utilization by adjusting parameters such as process priority, CPU affinity, and CPU limits; optimize memory utilization by adjusting parameters such as process memory limits, memory swapping, and memory compression; optimize disk I / O performance by adjusting parameters such as the disk scheduler and file system cache; and optimize network bandwidth and concurrent connections by adjusting parameters such as network buffers and TCP kernel configurations, etc.

[0069] Step S23: Create an executable script for adjusting the target system parameters in the current business system, and use the executable script to configure the current business system according to the target operating mode and the target action variable corresponding to the sampling time.

[0070] In this embodiment, creating an executable script for adjusting target system parameters in the current business system may include: creating an executable script for adjusting target system parameters in the current business system using preset system parameter adjustment instructions and preset application software parameter adjustment instructions.

[0071] In this embodiment, the optimal configuration parameters can be calculated based on the scheme obtained from reinforcement learning simulation, and an executable script or command for adjusting the configuration can be generated. In this step, the system configuration parameters can be adjusted according to the script or command. For example, system parameters can be made effective by the sysctl command, and application software such as the ceph distributed file system can be made effective for the entire software cluster by the ceph command. Configuration parameters that need to be modified through configuration files can be modified by a customized strategy within a specified time and the service can be restarted to avoid the impact of restarting the service on business operation, thereby changing the resource allocation and optimizing the system.

[0072] This embodiment proposes using a preset monitoring tool to collect service load information and target performance indicators in preset application software and operating system at each sampling time according to a preset sampling frequency. Then, target feature information is extracted from the target performance indicators and service load information, and the target feature information is converted into target state variables. Next, a reward function determined based on the target state variables is used to calculate the reward value at each sampling time. Then, a preset reinforcement learning algorithm is used to analyze and calculate the target feature information, target state variables, and reward values ​​at each sampling time, converting target system parameters into action variables. Using the action variables, service load information, target feature information, and reward values ​​at each sampling time, the target operating mode and the target parameter configuration scheme corresponding to the target operating mode are determined, and the target action variables corresponding to the target system parameters of the target parameter configuration scheme are determined. Finally, an executable script is created to adjust the target system parameters in the current business system, and the executable script is used to configure the current business system according to the target operating mode and target action variables corresponding to the sampling time. In this way, the deep reinforcement learning method used in this embodiment optimizes the adjustment of configuration parameters, improves the rationality and effectiveness of the system configuration parameter optimization, and realizes the full automation of the entire system configuration parameter adjustment process, which greatly reduces the cost of manual adjustment.

[0073] See Figure 3 As shown in the figure, this application discloses a business system configuration device, which may specifically include:

[0074] The data processing module 11 is used to collect business load information and target performance indicators in the preset application software and operating system at each sampling time using a preset monitoring tool at a preset sampling frequency, then extract target feature information from the target performance indicators and the business load information, and calculate the reward value at each sampling time using a reward function determined based on the target performance indicators.

[0075] The scheme determination module 12 is used to analyze and calculate the service load information, the target performance index, the target feature information and the reward value at each sampling time using a preset reinforcement learning algorithm, so as to determine the target operation mode and the target parameter configuration scheme corresponding to each sampling time.

[0076] The scheme configuration module 13 is used to configure the current business system according to the target operating mode and the target parameter configuration scheme corresponding to each sampling time.

[0077] This application first uses a preset monitoring tool to collect business load information and target performance indicators in the preset application software and operating system at each sampling time according to a preset sampling frequency. Then, it extracts target feature information from the target performance indicators and business load information, and calculates the reward value at each sampling time using a reward function determined based on the target performance indicators. A preset reinforcement learning algorithm is then used to analyze and calculate the business load information, target performance indicators, target feature information, and reward value at each sampling time to determine the target operating mode and the target parameter configuration scheme corresponding to each sampling time. Finally, the current business system is configured according to the target operating mode and target parameter configuration scheme corresponding to each sampling time. In this way, this invention extracts feature information by monitoring the key performance indicators of the application software and the underlying supporting operating system, as well as the load of the current business system. The feature information is used as a reward function to calculate the reward value of the current state. Based on this, a reinforcement learning algorithm is used for training. Combined with stored historical data, load distribution, state information features, and historical reward values, the analysis and calculation are performed to select a matching parameter configuration scheme as the optimal adjustment method, enabling the operating system and application software to adapt to the current business load. In this embodiment, a deep reinforcement learning method is used to optimize the adjustment of configuration parameters, which improves the rationality and effectiveness of the system configuration parameter optimization. The entire system configuration parameter adjustment process is fully automated, which greatly reduces the cost of manual adjustment.

[0078] like Figure 4 As shown, the business system configuration device in this invention can also be divided into a performance indicator acquisition module, a reinforcement learning module, and a configuration parameter adjustment module according to their functional types. The performance indicator acquisition module can perform feature extraction, indicator conversion, and reward value calculation on the acquired performance indicators of the business system, application software, and Linux operating system. In the reinforcement learning module, the reinforcement learning algorithm is used to perform algorithm execution based on the information in the performance indicator acquisition module to finally determine the target running mode corresponding to each sampling time and the target parameter configuration scheme corresponding to the target running mode. Finally, the configuration parameter adjustment module generates an executable file or command to complete the adjustment of the configuration parameters.

[0079] Furthermore, embodiments of this application also disclose an electronic device, Figure 5 This is a structural diagram of an electronic device 20 according to an exemplary embodiment. The content of the diagram should not be construed as limiting the scope of use of this application.

[0080] Figure 5 This is a schematic diagram of the structure of an electronic device 20 provided in an embodiment of this application. Specifically, the electronic device 20 may include: at least one processor 21, at least one memory 22, a power supply 23, a display screen 24, an input / output interface 25, a communication interface 26, and a communication bus 27. The memory 22 stores a computer program, which is loaded and executed by the processor 21 to implement the relevant steps in the business system configuration method disclosed in any of the foregoing embodiments. Furthermore, the electronic device 20 in this embodiment may specifically be an electronic computer.

[0081] In this embodiment, the power supply 23 is used to provide operating voltage for each hardware device on the electronic device 20; the communication interface 26 can create a data transmission channel between the electronic device 20 and external devices, and the communication protocol it follows can be any communication protocol applicable to the technical solution of this application, and is not specifically limited here; the input / output interface 25 is used to acquire external input data or output data to the outside world, and its specific interface type can be selected according to specific application needs, and is not specifically limited here.

[0082] Furthermore, the memory 22, as a carrier for resource storage, can be a read-only memory, random access memory, disk, or optical disk, etc. The resources stored on it can include an operating system 221, computer programs 222, and virtual machine data 223, etc. The virtual machine data 223 can include various types of data. The storage method can be temporary storage or permanent storage.

[0083] The operating system 221 is used to manage and control the various hardware devices on the electronic device 20 and the computer program 222, which may be Windows Server, Netware, Unix, Linux, etc. In addition to including a computer program capable of performing the business system configuration method executed by the electronic device 20 as disclosed in any of the foregoing embodiments, the computer program 222 may further include a computer program capable of performing other specific tasks.

[0084] Furthermore, this application also discloses a computer-readable storage medium, which includes random access memory (RAM), main memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disks, magnetic disks, optical disks, or any other form of storage medium known in the art. The computer program, when executed by a processor, implements the aforementioned disclosed business system configuration method. Specific steps of this method can be found in the corresponding content disclosed in the foregoing embodiments, and will not be repeated here.

[0085] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatuses disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the descriptions are relatively simple; relevant parts can be referred to the method section. Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in terms of function in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this application.

[0086] The steps of the methods or algorithms described in conjunction with the embodiments disclosed herein can be implemented directly by hardware, a software module executed by a processor, or a combination of both. The software module can be located in random access memory (RAM), main memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium known in the art.

[0087] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0088] The above provides a detailed description of the business system configuration method, apparatus, device, and storage medium provided by the present invention. Specific examples have been used to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of the present invention. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A business system configuration method, characterized in that, include: Using a preset monitoring tool, the system collects business load information and target performance indicators in the preset application software and operating system at each sampling time according to a preset sampling frequency. Then, it extracts target feature information from the target performance indicators and the business load information, and calculates the reward value at each sampling time using a reward function determined based on the target performance indicators. The service load information, target performance indicators, target feature information and reward value at each sampling time are analyzed and calculated using a preset reinforcement learning algorithm to determine the target operating mode and the target parameter configuration scheme corresponding to each sampling time. Configure the current business system according to the target operating mode and target parameter configuration scheme corresponding to each sampling time; The step of extracting target feature information from the target performance metric and the service load information, and calculating the reward value at each sampling time using the reward function determined based on the target performance metric, includes: Target feature information is extracted from the target performance indicators and the business load information, and the target feature information is converted into target state variables. Then, the reward value at each sampling time is calculated using the reward function determined based on the target state variables. The step of using a preset reinforcement learning algorithm to analyze and calculate the service load information, target performance indicators, target feature information, and reward value at each sampling time to determine the target operating mode and the target parameter configuration scheme corresponding to the target operating mode at each sampling time includes: Based on a preset reinforcement learning algorithm, the target feature information, the target state variable and the reward value at each sampling time are analyzed and calculated. The target system parameters are converted into action variables. The target operating mode and the target parameter configuration scheme corresponding to each sampling time are determined by using the action variables, the business load information, the target feature information and the reward value at each sampling time. The determination of the target operating mode corresponding to each sampling time and the target parameter configuration scheme corresponding to the target operating mode includes: Determine the target operating mode and the target parameter configuration scheme corresponding to each sampling time, and determine the target action variables corresponding to the target system parameters of the target parameter configuration scheme; Accordingly, configuring the current business system according to the target operating mode and the target parameter configuration scheme corresponding to each sampling time includes: Create an executable script to adjust the target system parameters in the current business system, and use the executable script to configure the current business system according to the target operating mode and the target action variable corresponding to the sampling time.

2. The business system configuration method according to claim 1, characterized in that, The process of collecting service load information and target performance indicators in the application software and operating system at each sampling time using a preset monitoring tool at a preset sampling frequency includes: The Prometheus tool is used to collect service load information and target performance indicators in the application software and operating system at each sampling time according to the preset sampling frequency.

3. The business system configuration method according to claim 1, characterized in that, The step of using a preset reinforcement learning algorithm to analyze and calculate the service load information, target performance indicators, target feature information, and reward value at each sampling time to determine the target operating mode and the target parameter configuration scheme corresponding to the target operating mode at each sampling time includes: The service load information, target performance indicators, target feature information and reward value at each sampling time are analyzed and calculated using a preset reinforcement learning algorithm to determine the target operating mode and the target parameter configuration scheme corresponding to the highest reward value at each sampling time.

4. The business system configuration method according to claim 1, characterized in that, The creation of an executable script for adjusting target system parameters in the current business system includes: Create an executable script to adjust the target system parameters in the current business system using preset system parameter adjustment commands and preset application software parameter adjustment commands.

5. A business system configuration device, characterized in that, include: The data processing module is used to collect business load information and target performance indicators in the preset application software and operating system at each sampling time using a preset monitoring tool at a preset sampling frequency. Then, it extracts target feature information from the target performance indicators and the business load information, and calculates the reward value at each sampling time using a reward function determined based on the target performance indicators. The scheme determination module is used to analyze and calculate the service load information, the target performance index, the target feature information and the reward value at each sampling time using a preset reinforcement learning algorithm, so as to determine the target operation mode and the target parameter configuration scheme corresponding to each sampling time. The scheme configuration module is used to configure the current business system according to the target operating mode and the target parameter configuration scheme corresponding to each sampling time. The data processing module is specifically used to extract target feature information from the target performance indicators and the business load information, convert the target feature information into target state variables, and then use the reward function determined based on the target state variables to calculate the reward value at each sampling time. The scheme determination module is specifically used to analyze and calculate the target feature information, the target state variable and the reward value at each sampling time based on a preset reinforcement learning algorithm, convert the target system parameters into action variables, and use the action variables, the business load information, the target feature information and the reward value at each sampling time to determine the target operation mode and the target parameter configuration scheme corresponding to the target operation mode at each sampling time. The scheme determination module is specifically used to determine the target operating mode and the target parameter configuration scheme corresponding to each sampling time, and to determine the target action variables corresponding to the target system parameters of the target parameter configuration scheme. The scheme configuration module is specifically used to create an executable script for adjusting the target system parameters in the current business system, and to configure the current business system according to the target operating mode and the target action variables corresponding to the sampling time using the executable script.

6. An electronic device, characterized in that, It includes a processor and a memory; wherein, when the processor executes a computer program stored in the memory, it implements the business system configuration method as described in any one of claims 1 to 4.

7. A computer-readable storage medium, characterized in that, Used to store computer programs; wherein, when the computer programs are executed by a processor, they implement the business system configuration method as described in any one of claims 1 to 4.

Citation Information

Patent Citations

  • Ceph system performance optimization strategy and system based on deep reinforcement learning

    CN111581178A