Application upgrading method, system and equipment based on low-code platform and storage medium

Abstract

The invention provides an application upgrading method, system and device based on a low-code platform and a storage medium, and relates to the technical field of computers.The method comprises the steps that multi-dimensional comprehensive evaluation is conducted according to current version information and target version information of a target application, and a comprehensive evaluation index is obtained; the comprehensive evaluation index is compared with a comprehensive evaluation threshold value, if the comprehensive evaluation index is larger than the comprehensive evaluation threshold value, automatic upgrading operation is executed, and otherwise, an automatic upgrading scheme is adjusted; historical upgrading information and real-time log information in the upgrading process are obtained, risk prediction is conducted on the upgrading process through the random forest risk prediction model, and a prediction result is obtained; and according to a prediction result, continuing to execute an automatic upgrading operation or a secure rollback operation. According to the method, the application upgrading and rollback process is processed through multi-dimensional comprehensive evaluation and the risk prediction algorithm based on the random forest, and it is ensured that the application always keeps high availability and stability.

Description

Technical Field

[0001] This invention relates to the field of computer technology, and specifically to an application upgrade method, system, device, and storage medium based on a low-code platform. Background Technology

[0002] In the digital age, enterprises have a strong need for rapid application development and iteration in order to respond quickly to market changes. Low-code platforms, with their visual development interfaces, eliminate the need for extensive coding. By dragging and dropping components and configuring parameters, applications can be developed and deployed quickly, greatly simplifying processes and shortening cycles. They are widely used in enterprise digital transformation and have become the preferred choice for many enterprises and developers.

[0003] However, business growth necessitates continuous upgrades for applications using low-code platforms to introduce new features, optimize performance, and quickly roll back to a stable version in case of upgrade issues. Upgrades and rollbacks involve complex operations such as database migration and configuration updates. In complex business scenarios, tightly coupled modules and data relationships can easily lead to compatibility problems, preventing new features from functioning correctly. Resource allocation issues can also cause performance degradation, impacting user experience, and even introduce security vulnerabilities, threatening system security. Existing solutions mostly rely on manual methods or simple scripts, making it difficult to guarantee the security and stability of low-code platform applications during upgrades and rollbacks. Summary of the Invention

[0004] In view of the above-mentioned shortcomings of the prior art, the present invention provides an application upgrade method, system, device and storage medium based on a low-code platform, which effectively solves the problem of insufficient security and stability of applications using low-code platforms during the upgrade and rollback process in the prior art.

[0005] In a first aspect, the present invention provides an application upgrade method based on a low-code platform, the method comprising: A comprehensive evaluation is conducted based on the current version information of the target application and the target version information to obtain comprehensive evaluation indicators; The comprehensive evaluation index is compared with the comprehensive evaluation threshold. If the comprehensive evaluation index is greater than the comprehensive evaluation threshold, an automatic upgrade operation is performed; otherwise, the automatic upgrade plan is adjusted. Obtain historical upgrade information and real-time log information during the upgrade process, and use a random forest risk prediction model to predict the risk of the upgrade process and obtain the prediction results; Based on the prediction results, either continue with the automatic upgrade operation or perform a safe rollback operation.

[0006] In an optional implementation, the step of performing a multi-dimensional comprehensive evaluation based on the current version information and target version information of the target application to obtain comprehensive evaluation indicators includes: Calculate the feature compatibility index based on the first feature set of the current version and the second feature set of the target version; The performance change rate index is calculated based on the first performance index data of the current version and the second performance index data of the target version. The performance change rate index includes at least the response time change rate and the processing capacity change rate. Obtain vulnerability set information for the target version, and calculate security risk indicators based on the vulnerability set information; Based on the system resource requirements of the target version and the system resource threshold information, resource bottleneck indicators are calculated. The resource bottleneck indicators include at least CPU indicators, memory indicators, and disk space indicators. The comprehensive evaluation index is obtained by comprehensively calculating the functional compatibility index, the performance change rate index, the security risk index, and the resource bottleneck index.

[0007] In an optional implementation, the formula for calculating the comprehensive evaluation index is as follows:

[0008] In the above formula, S Indicates comprehensive evaluation indicators, C func Indicates functional compatibility metrics. R T Indicates the rate of change of response time. R Q This indicates the rate of change in processing capacity. R sec Indicators representing security risk. max ( R sec This indicates the highest security risk indicator recorded during the historical upgrade process. B cpu Indicates CPU metrics, B mem Indicates memory metrics, B disk This represents a disk space metric. , , , and These represent the first, second, third, fourth, and fifth weight coefficients, respectively, and the sum of all weight coefficients equals 1.

[0009] In an optional implementation, the step of acquiring historical upgrade information and real-time log information during the upgrade process, and using a random forest risk prediction model to predict the risk of the upgrade process and obtain the prediction results includes: The historical upgrade information and the real-time log information are preprocessed to obtain a sample dataset; Based on the sample dataset, a risk prediction model is constructed and trained using the random forest algorithm to obtain the random forest risk prediction model. Real-time sample data is obtained during the upgrade process, and the real-time sample data is input into the random forest risk prediction model to predict the risk of the upgrade process, thereby obtaining the prediction result.

[0010] In an optional implementation, the historical upgrade information and the real-time log information are preprocessed to obtain a sample dataset, including: The historical upgrade information and the real-time log information are cleaned to obtain a first dataset. The data cleaning includes at least missing value handling, outlier handling and data consistency checking. The first dataset is transformed to obtain the second dataset. The data transformation includes at least standardization, classification feature encoding, and temporal feature processing. Feature extraction is performed on the second dataset to obtain the sample dataset. The feature extraction includes at least feature extraction based on business logic, feature extraction based on statistical methods, and feature extraction fused with real-time log information.

[0011] In an optional implementation, based on the prediction result, the automatic upgrade operation continues or a safe rollback operation is performed, including: If the prediction result is low risk, then continue with the automatic upgrade operation; If the prediction result is medium risk, then after manual intervention, an automatic upgrade operation or a safe rollback operation will be performed. If the prediction result is high risk, then the automatic upgrade operation will be suspended and a safe rollback operation will be performed.

[0012] In an optional implementation, performing the automatic upgrade operation includes: Back up and verify the application data, configuration data, code, and metadata of the target application to obtain backup files; After performing compatibility checks and resource assessments on the target version, the target version is deployed to the server and the database is migrated. Start the target version of the service and perform automated testing, which includes at least core function checks and key interface response status checks. The execution of the safe rollback operation includes: Real-time monitoring of the target metrics of the target application is performed to obtain real-time monitoring metrics; Automatic rollback conditions are triggered based on the real-time monitoring metrics or automated test results. Stop the target version of the service, restore the data based on the backup file, and start the current version of the service.

[0013] Secondly, the present invention provides an application upgrade system based on a low-code platform, the system comprising: The comprehensive evaluation module is used to perform a multi-dimensional comprehensive evaluation based on the current version information and target version information of the target application to obtain comprehensive evaluation indicators; The automatic upgrade module is used to compare the comprehensive evaluation index with the comprehensive evaluation threshold. If the comprehensive evaluation index is greater than the comprehensive evaluation threshold, an automatic upgrade operation is performed; otherwise, the automatic upgrade plan is adjusted. The risk prediction module is used to obtain historical upgrade information and real-time log information during the upgrade process, and to predict the risks of the upgrade process using a random forest risk prediction model to obtain the prediction results. The safety rollback module is used to either continue the automatic upgrade operation or perform a safety rollback operation based on the prediction result.

[0014] Thirdly, the present invention provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the application upgrade method based on a low-code platform as described in the first aspect of the present invention.

[0015] Fourthly, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the application upgrade method based on a low-code platform as described in the first aspect of the present invention.

[0016] This invention provides an application upgrade method, system, device, and storage medium based on a low-code platform. It processes the application upgrade and rollback process through multi-dimensional comprehensive evaluation and a random forest-based risk prediction algorithm. The multi-dimensional evaluation comprehensively analyzes the impact of the upgrade from multiple key levels, including application performance, compatibility, and business logic, identifying potential risk points and providing detailed evidence for decision-making. The random forest risk prediction algorithm, with its powerful data processing and pattern recognition capabilities, accurately predicts potential problems and anticipates possible faults and anomalies during the upgrade process. This achieves automated intelligent decision-making, eliminating the need for cumbersome manual intervention and quickly and accurately deciding on upgrade or rollback operations. This not only greatly improves upgrade efficiency and shortens business downtime but also effectively avoids risks caused by human error. Through automated and secure upgrade and rollback processes, it ensures that applications maintain high availability and stability, providing a solid guarantee for the continuous and stable operation of low-code platform applications. Attached Figure Description

[0017] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a first schematic diagram of the application upgrade method based on a low-code platform provided in an embodiment of the present invention; Figure 2 This is a second schematic diagram of the application upgrade method based on a low-code platform provided in this embodiment of the invention; Figure 3 This is the third schematic diagram of the application upgrade method based on a low-code platform provided in this embodiment of the invention; Figure 4 This is the fourth schematic diagram of the application upgrade method based on a low-code platform provided in this embodiment of the invention; Figure 5 This is a schematic diagram of the application upgrade system structure based on a low-code platform provided in an embodiment of the present invention; Figure 6 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention.

[0019] Explanation of key component symbols: 200. Application upgrade system based on low-code platform; 210. Comprehensive evaluation module; 220. Automatic upgrade module; 230. Risk prediction module; 240. Security rollback module; 300. Electronic device; 310. Processor; 320. Communication interface; 330. Memory; 340. Communication bus. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be further described clearly and completely below with reference to the accompanying drawings of the embodiments of this invention. It should be noted that the described embodiments are merely some embodiments of this invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0021] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0022] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the technical field to which this application belongs. The terms used in the specification of the present invention are for the purpose of describing specific embodiments only and are not intended to limit the present invention.

[0023] In the current context of rapid business development, enterprises' demand for applications using low-code platforms continues to climb, and applications need to be continuously upgraded to introduce novel functions and optimize overall performance. Upgrade rollback involves delicate and complex operations such as database migration and configuration updates. In complex business scenarios, various modules are tightly coupled and data associations are intricate. Once a problem occurs in a certain link, it is extremely easy to trigger compatibility failures, resulting in the new functions not being able to run properly as expected and causing business chaos. At the same time, if resource allocation is unreasonable during the upgrade process, it may cause a sharp decline in system performance, slow page loading, and sluggish operation responses, seriously affecting the user experience and reducing user satisfaction. More seriously, security vulnerabilities may also be hidden during the upgrade process, threatening the data security and stable operation of the system. However, existing solutions mostly rely on manual operations or simple scripts, lacking systematicness and intelligence, and are difficult to comprehensively address various problems in complex scenarios, and cannot effectively guarantee the security and stability of low-code platform applications during the upgrade rollback process.

[0024] Embodiment 1 The embodiment of the present invention provides an application upgrade method based on a low-code platform, which effectively solves the problem of insufficient security and stability of applications using low-code platforms during the upgrade rollback process in the prior art. Figure 1 It is the first schematic diagram of the application upgrade method based on a low-code platform provided by the embodiment of the present invention. As Figure 1 shown, the method includes the following steps: S100. Perform multi-dimensional comprehensive evaluation based on the current version information and target version information of the target application to obtain a comprehensive evaluation index.

[0025] In the embodiment of the present invention, the upgrade of an application based on a low-code platform can be comprehensively evaluated from four dimensions: function compatibility, performance impact, security risk, and system resource occupancy, providing a quantitative basis for upgrade decisions. Figure 2 It is the second schematic diagram of the application upgrade method based on a low-code platform provided by the embodiment of the present invention. As Figure 2 shown, the multi-dimensional comprehensive evaluation specifically includes the following steps: S110. Calculate the function compatibility index based on the first function set of the current version and the second function set of the target version.

[0026] The function set changes due to the application type and version iteration target. If the first function set of the current version is Fcurrent ={ f 1, f 2,…, f n The second feature set of the target version is: F target ={ g 1, g 2,…, g n In this embodiment of the invention, a functional compatibility index is defined to measure the backward compatibility of new version features, that is, the degree of matching between the features of the current version and the features of the target version. The formula for calculating the functional compatibility index is as follows:

[0027] In the above formula, C func Indicates functional compatibility metrics. F current This represents the first set of features in the current version. F target This represents the second feature set of the target version, where |·| represents the number of elements in the set.

[0028] In an embodiment of the present invention, C func The closer the value is to 1, the better the compatibility between the target version's features and the current version. C func If the value is less than the preset compatibility threshold, the target version is considered to have compatibility issues, and the upgrade risk is high. Optionally, this preset compatibility threshold can be set based on business needs, historical data, statistical analysis, and human experience.

[0029] S120. Calculate the performance change rate index based on the first performance index data of the current version and the second performance index data of the target version. The performance change rate index includes at least the response time change rate and the processing capacity change rate.

[0030] In this embodiment of the invention, performance metrics data of the current version and the target version are collected under the same test environment. These performance metrics include, but are not limited to, response time and processing capacity. Response time refers to the total time from when a user initiates a request to when a result is returned, and processing capacity represents the number of requests processed per unit time. Accordingly, the performance change rate metric includes at least the response time change rate and the processing capacity change rate, used to assess the impact of the target version on system performance. The formula for calculating the response time change rate is as follows:

[0031] In the above formula, R TIndicates the rate of change of response time. T current Indicates the response time of the current version. T target Indicates the response time of the target version.

[0032] The formula for calculating the rate of change in processing capacity is as follows:

[0033] In the above formula, R Q This indicates the rate of change in processing capacity. Q current This indicates the processing power of the current version. Q target This indicates the processing capability of the target version.

[0034] In an embodiment of the present invention, when R T Less than the preset response time threshold or R Q If the response time threshold exceeds the preset processing capacity threshold, the target version is considered to have a significant impact on system performance, requiring further analysis and optimization. Optionally, the preset response time threshold and preset processing capacity threshold can be set based on business needs, historical data, statistical analysis, and human experience. For example, assuming a preset response time threshold of -10% and a preset processing capacity threshold of 15%, if the response time change rate is -25%, which is less than the preset response time threshold, it indicates a significant improvement in response time. However, if it is a positive change, i.e., a response time change rate of 25%, which is greater than the preset response time threshold, it indicates a significant deterioration in response time. If the processing capacity change rate is 20%, which is greater than the preset processing capacity threshold, it indicates a significant improvement in processing capacity.

[0035] S130. Obtain vulnerability set information for the target version and calculate security risk indicators based on the vulnerability set information.

[0036] In this embodiment of the invention, a vulnerability scanning tool is used to perform a comprehensive scan of the target version to obtain a vulnerability set. V ={ v 1, v 2,…, v k For each vulnerability v i Define a risk weight W i Risk weights can be determined based on factors such as the severity and difficulty of exploitation of vulnerabilities. Security risk indicators can be calculated based on vulnerability set information, using the following formula:

[0037] In the above formula, R sec Indicators representing security risk. i Indicates the vulnerability number. k This indicates the total number of vulnerabilities. W i Indicates the first i Risk weights for each vulnerability.

[0038] In an embodiment of the present invention, R sec A higher value indicates a higher security risk for the target version. R sec If the risk level exceeds the preset security risk threshold, the target version is considered to have a high security risk and needs to be patched before upgrading. Optionally, the preset security risk threshold can be set based on business needs, historical data, statistical analysis, and human experience.

[0039] S140. Calculate resource bottleneck indicators based on the target version's system resource requirements and system resource threshold information. Resource bottleneck indicators include at least CPU indicators, memory indicators, and disk space indicators.

[0040] In this embodiment of the invention, the system resource requirements of the target version are analyzed, including but not limited to CPU utilization, memory utilization, and disk space utilization. Simultaneously, corresponding system resource threshold information is obtained, including but not limited to preset CPU availability, memory availability, and disk space availability. Resource bottleneck indicators are defined to determine whether the target version will cause a resource bottleneck. These resource bottleneck indicators include, but are not limited to, CPU indicators, memory indicators, and disk space indicators, and their calculation formulas are as follows:

[0041] In the above formula, B cpu Indicates CPU metrics, U cpu This indicates the CPU utilization of the target version. A cpu This indicates the preset CPU availability of system resources. B cpu A value of 1 indicates that the CPU is overloaded and requires expansion or optimization. B cpu A value of 0 indicates that the CPU is not overloaded and is in a normal state.

[0042]

[0043] In the above formula, B mem Indicates memory metrics,U mem Indicates the memory usage of the target version. A mem This indicates the preset memory availability of system resources. B mem A value of 1 indicates memory overload, which may trigger a memory overflow or performance degradation. B mem A value of 0 indicates that the memory is not overloaded and is in a normal state.

[0044]

[0045] In the above formula, B disk This represents a disk space metric. U disk This indicates the disk space usage of the target version. A disk This indicates the preset disk space availability of system resources. B disk A value of 1 indicates insufficient disk space, which may affect writing or log recording. B disk A value of 0 indicates that there is sufficient disk space and the system is in a normal state.

[0046] In an embodiment of the present invention, when B cpu =1 or B mem =1 or B disk When the value is 1, it is believed that the target version may cause corresponding resource bottlenecks, and resource expansion or application code optimization should be carried out in advance.

[0047] S150. A comprehensive evaluation index is obtained by comprehensively calculating based on functional compatibility index, performance change rate index, security risk index, and resource bottleneck index.

[0048] In this embodiment of the invention, a multi-dimensional comprehensive evaluation index is further calculated based on functional compatibility indicators, performance change rate indicators, security risk indicators, and resource bottleneck indicators, thereby achieving a comprehensive and detailed evaluation of application upgrades. The calculation formula for the comprehensive evaluation index is as follows:

[0049] In the above formula, S Indicates comprehensive evaluation indicators, C func Indicates functional compatibility metrics. R T Indicates the rate of change of response time. R QThis indicates the rate of change in processing capacity. R sec Indicators representing security risk. max ( R sec This indicates the highest security risk indicator recorded during historical upgrades, used to assess... R sec Normalization is performed. B cpu Indicates CPU metrics, B mem Indicates memory metrics, B disk This represents a disk space metric. , , , and These represent the first, second, third, fourth, and fifth weight coefficients, respectively. The sum of all weight coefficients equals 1, and each weight coefficient can be adjusted according to the actual application scenario and business needs.

[0050] In this embodiment of the invention, functional compatibility assessment can identify potential functional conflicts that may arise from upgrades in advance, ensuring stable application operation; performance impact analysis can accurately pinpoint changes in performance indicators such as response time and processing capacity caused by upgrades, helping to optimize upgrade plans; security risk assessment can identify potential security vulnerabilities, protecting user data and system security; and system resource usage assessment can rationally plan resources, avoiding resource exhaustion due to upgrades. A comprehensive quantitative assessment across these four dimensions provides a scientific and comprehensive basis for application upgrade decisions, reducing upgrade risks and increasing upgrade success rates.

[0051] S200. Compare the comprehensive evaluation index with the comprehensive evaluation threshold. If the comprehensive evaluation index is greater than the comprehensive evaluation threshold, then execute the automatic upgrade operation; otherwise, adjust the automatic upgrade plan.

[0052] Optionally, preset comprehensive evaluation thresholds can be set by combining business needs, historical data, statistical analysis, and human experience. If the comprehensive evaluation indicators... S If the value exceeds the preset comprehensive evaluation threshold, the application upgrade is considered feasible, and an automatic upgrade operation is performed, which includes the following process: First, the application data, configuration data, code, and metadata of the target application are backed up and verified to obtain backup files. Application data includes, but is not limited to, all tables, table data, and file storage corresponding to the database application. Configuration data includes, but is not limited to, environment variables, permission rules, API keys, and workflow definitions. Code and metadata include, but is not limited to, code generated by the low-code platform, page layouts, and logic scripts. Specifically, when an automatic upgrade operation is triggered, a full automatic backup is triggered simultaneously. Different methods are used for different data types. The full database files of the application are exported from the database. For file storage, the file directories that the application depends on are packaged. For configuration and code, the platform configuration JSON files are exported and the generated code directory is packaged. After collecting the backup content, the backup files are encrypted and signed. AES-256 encryption can be used, and a SHA-256 signature is generated to prevent the backup files from being tampered with. The backup files are then stored off-site, such as via Alibaba Cloud OSS off-site backup, with file naming following the format "application_full_backup_YYYYMMDD_HHMMSS.tar.gz.enc". Finally, backup verification is performed. Through automated checks, the backup file size is verified to be greater than 0 to avoid empty files. After decrypting the backup file, the database SQL file is verified to ensure that it can be successfully imported into the test environment, thus ensuring that the backup is valid and usable.

[0053] Then, after performing compatibility checks and resource assessments on the target version, it is deployed to the server and the database is migrated. Compatibility checks confirm that the target version is compatible with the old data structure, such as ensuring that database table fields are not deleted, to avoid application errors due to data mismatch after the upgrade. Resource assessment checks whether the server disk space is sufficient, as the upgrade process requires reserving space not only for backup files but also for upgrade files such as the target version's code package; insufficient space may interrupt the upgrade. During the upgrade, to ensure data security and integrity, the current version service must be stopped first to prevent conflicts caused by new data writing. Subsequently, the target version deployment is carried out, uploading the target version code package to the designated location on the server. During the configuration file overwrite stage, critical parameters such as database connections are preserved to prevent service malfunction due to configuration loss. If the target version involves database structure changes, the SQL migration script provided by the new version must be executed, and the database upgrade must be completed step by step according to the script requirements to ensure database compatibility with the target version code, thereby ensuring that the entire application system can be smoothly and successfully upgraded to the target version and put into use.

[0054] Finally, start the target version of the service and perform automated testing. Automated testing should include at least core functionality checks and critical interface response status checks. Specifically, automated testing needs to check each core function one by one. For example, for the user login function, it's necessary to test whether different types of users can log in to the system normally. For the data query function, it's necessary to verify whether various types of data can be displayed accurately and quickly, and whether the query results are complete and error-free. At the same time, professional tools or scripts can be used to send requests to critical interfaces, and the interface response status codes can be carefully examined to ensure that the interfaces can receive and process requests normally.

[0055] If comprehensive evaluation indicators S If the values ​​are less than or equal to the preset comprehensive evaluation threshold, it indicates that further optimization or adjustment of the upgrade plan is needed. Furthermore, the corresponding problems can be further identified based on functional compatibility indicators, performance change rate indicators, security risk indicators, and resource bottleneck indicators. For example, if the functional compatibility indicator is less than the preset compatibility threshold, it indicates that the target version has compatibility issues and the upgrade risk is high. In this case, the process will be prevented from entering the application upgrade and risk prediction and intelligent decision-making stage, and the compatibility issues will be resolved first.

[0056] S300 acquires historical upgrade information and real-time log information during the upgrade process, and uses a random forest risk prediction model to predict the risk of the upgrade process and obtain the prediction results.

[0057] In this embodiment of the invention, a random forest risk prediction model is constructed using historical upgrade information and real-time log information during the upgrade process to predict potential risks during the automatic upgrade process. Figure 3 This is a third schematic diagram of the application upgrade method based on a low-code platform provided in this embodiment of the invention, as shown below. Figure 3 As shown, the specific steps for risk prediction are as follows: S310. Perform data preprocessing on historical upgrade information and real-time log information to obtain a sample dataset.

[0058] In this embodiment of the invention, data preprocessing includes three steps: data cleaning, data transformation, and feature extraction, wherein: The purpose of data cleaning is to remove noise, erroneous values, and inconsistencies from historical upgrade information and real-time log information to obtain a preliminary dataset. This ensures the quality and accuracy of the data, providing a reliable foundation for subsequent data transformation and feature extraction. Data cleaning includes, but is not limited to, handling missing values, handling outliers, and checking data consistency.

[0059] Missing value handling involves iterating through historical upgrade information and real-time logs to check for missing values ​​in each feature of each sample. For example, it checks for null values ​​in features such as "upgrade time," "functional compatibility index," and "performance change rate index." If a sample has many missing features, or if the missing features are crucial to the analysis and cannot be adequately filled, the sample can be deleted. For instance, if a sample is missing both "upgrade time" and "performance change rate index," and these two features are critical to subsequent analysis, then the sample can be deleted. For samples with few missing values, imputation methods can be used. For example, for numerical features, the mean, median, or mode can be used for imputation; for categorical features, the most frequently occurring category can be used.

[0060] Outlier handling uses statistical methods or visualization tools to detect outliers in the data. For example, for numerical features, box plots can be used to identify outliers, and for categorical features, the presence of obviously unreasonable categories can be checked. If outliers are due to data entry or measurement errors, these values ​​can be corrected. For example, if the "Performance Change Rate" metric is recorded as +150%, while the performance change rates of other samples are mostly between ±20%, it is likely a data entry error and can be corrected to a reasonable value. If outliers cannot be corrected and would significantly impact the analysis results, samples containing outliers can be deleted. For example, if a sample's "Resource Bottleneck" metric shows "Memory > 500%", this is clearly unrealistic and cannot be corrected, so the sample can be deleted. In some cases, outliers may contain important information; these samples can be retained but marked for special attention in subsequent analyses. For example, if a sample's "Security Risk" is abnormally high, it may be due to a severe system attack; retaining and marking this sample helps in analyzing specific security risk scenarios.

[0061] Data consistency checks ensure that all data is formatted consistently and check for logical inconsistencies. For example, "upgrade time" should be standardized to a specific date and time format (e.g., YYYY-MM-DD HH:MM:SS), and category names for categorical features should be standardized to avoid spelling errors or inconsistent representations. The "functional compatibility index" should be between 0 and 1; values ​​greater than 1 or less than 0 indicate logical errors in the data that require correction.

[0062] The purpose of data transformation is to convert the raw data in the first dataset into a form suitable for subsequent analysis and modeling, including but not limited to standardization, categorical feature encoding, and temporal feature processing.

[0063] Standardization primarily targets numerical features. Specifically, for numerical features such as "functional compatibility indicators," "performance change rate indicators," "security risk indicators," and "number of error logs," it is necessary to determine the minimum and maximum values ​​for each feature. Then, based on the minimum and maximum values, a standardization formula is applied to each numerical feature. The standardization formula is as follows:

[0064] In the above formula, y ij Represents the standardized first i The first sample j One standard eigenvalue, x ij Indicates the first i The first sample j 1 eigenvalue, min ( x j )and max ( x j ) respectively represent the first j The minimum and maximum values ​​of each feature. For example, for sample 1, the "functional compatibility index" value is 0.85. Assuming the minimum value of the "functional compatibility index" is 0.5 and the maximum value is 1, the standardized value is 0.7.

[0065] Categorical feature encoding primarily targets categorical features. First, the categorical features in the data are determined. For example, "resource bottleneck" might have values ​​such as "CPU > 80%", "insufficient memory", and "limited disk space", while "risk level" might have values ​​such as "high", "medium", and "low". Then, one-hot encoding can be used to encode unordered categorical features. For example, for the "resource bottleneck" feature, if there are three possible values, three binary bits can be used to represent it, with each bit corresponding to one value, only one bit being 1 and the rest being 0. Label encoding can be used to encode ordered categorical features. For example, the "risk level" feature can be encoded as 0, 1, and 2, corresponding to low, medium, and high, respectively.

[0066] Time feature processing primarily targets time characteristics. For time features like "upgrade time," year, month, day, hour, and minute information can be extracted as new features. For example, from "2025-07-20 10:30," the year 2025, month 7, day 20, hour 10, and minute 30 can be extracted as new features. Optionally, if it's necessary to analyze the time relationship between the upgrade time and other events, time differences can be calculated; for example, calculating the time interval between two upgrades.

[0067] The purpose of feature extraction is to extract more meaningful and representative features from the second dataset to improve the effectiveness of subsequent analysis and modeling. Feature extraction includes, but is not limited to, feature extraction based on business logic, feature extraction based on statistical methods, and feature extraction based on real-time log information fusion.

[0068] Feature extraction based on business logic is used to extract features related to functional changes, performance, and security. Based on features such as "functional compatibility" and "number of functional changes," some features related to functional changes can be extracted. For example, the impact of a functional change can be measured by the degree of change in functional compatibility, or the frequency of functional changes can be calculated by counting the number of functional changes over a period of time. Combining "performance change rate" and other relevant features, more comprehensive performance-related features are extracted. For example, the stability of performance changes can be measured by counting the standard deviation of the performance change rate over a period of time, or the trend of performance changes can be extracted by performing time series analysis on the performance change rate. Based on features such as "security risk" and "number of error logs," security-related features are extracted. For example, the severity index of a security risk can be calculated by comprehensively considering the value of the security risk and the number of error logs, or the frequency of security events can be extracted by counting the number of security events (such as high-security-risk escalations) over a period of time.

[0069] Feature extraction based on statistical methods is used to extract aggregated and correlated features. Historical upgrade data is aggregated to extract aggregated features. For example, the average, maximum, and minimum number of upgrades within each time period (e.g., daily and weekly) can be calculated; the average performance change rate and average security risk for each functional module can also be calculated. The correlations between various features are analyzed to extract combinations of features with strong correlations. For example, if a strong positive correlation is found between "number of error logs" and "security risk," they can be combined into a new feature to represent the overall risk status of the system.

[0070] Real-time log information fusion feature extraction is used to extract real-time error features and real-time performance features. Real-time error features, such as the quantity and distribution of error types, are extracted from real-time log information. These real-time error features can be fused with relevant features from historical upgrade data; for example, the rate of change in the number of real-time error logs can be calculated over a period before and after an upgrade. Real-time performance features, such as the system's real-time response time and throughput, are extracted. These real-time performance features can be compared and analyzed with features such as performance change rates in historical upgrade data to extract more meaningful performance features; for example, the degree of deviation between real-time performance and historical average performance can be calculated.

[0071] After the above data cleaning, data transformation, and feature extraction steps, a sample dataset containing multiple samples can be obtained. D ={ d 1, d 2,…, d n}, where each sample d i This includes multiple features after data processing and feature extraction. For example, a sample might include the following features: standardized upgrade time-related features, standardized functional compatibility metrics, standardized performance change rate metrics, coded security risks, coded resource bottlenecks, standardized error log counts, and other extracted relevant features (such as the scope of impact of functional changes and the stability of performance changes). Combining all the features from these samples constitutes the final sample dataset, which can be used for subsequent data analysis, modeling, and prediction tasks.

[0072] S320. Based on the sample dataset, construct and train a risk prediction model using the random forest algorithm to obtain the random forest risk prediction model.

[0073] In this embodiment of the invention, a random forest algorithm is used to construct and train a risk prediction model, thereby obtaining a random forest risk prediction model to predict potential risks during the upgrade process. The random forest algorithm improves the accuracy and stability of the model by constructing multiple decision trees and combining their prediction results. The training steps of the risk prediction model are as follows: First, samples are drawn from the sample dataset using a bootstrap sampling method with replacement. n One sample, construct n There are 10 training subsets. For each training subset, a decision tree is constructed. During the construction of the decision tree, for each node split, features are randomly selected from all features. m From each feature, a feature subset is obtained, where... m It is much smaller than the total number of features.

[0074] Then, based on metrics such as information gain and Gini index, the optimal feature is selected from a randomly selected feature subset to split the node, thereby enabling the decision tree to better divide the data and improve the model's classification or prediction capabilities.

[0075] In this embodiment of the invention, information gain is used to measure the degree of reduction in information uncertainty after partitioning a dataset using a certain feature. A larger information gain indicates a better partitioning effect of the feature and makes it more suitable as a splitting node. The calculation steps for information gain are as follows: First, calculate the information entropy of the sample dataset. Information entropy is a metric for the purity of a sample dataset; the more disordered the dataset, the higher the information entropy, and vice versa. The formula for calculating information entropy is as follows:

[0076] In the above formula, H(D) Represents the sample dataset D Information entropy K Indicates the number of categories in the sample dataset. p k Indicates the first k Class samples in sample dataset D The proportion it accounts for.

[0077] Then, the conditional entropy after partitioning the sample dataset using the target feature is calculated. Conditional entropy represents the uncertainty of the sample dataset given the values ​​of the target feature. Assume the target feature... A have V A number of different values ​​{ a 1, a 2,…, a v}, based on target features A Sample dataset D Divided into V Subset D 1, D 2,…, Dv Then the formula for calculating conditional entropy is:

[0078] In the above formula, Representing target features A Split sample dataset D The subsequent conditional entropy, V Indicates the number of subsets, Representing a subset D v The number of samples in the middle Represents the sample dataset D The number of samples in the middle Representing a subset D v Information entropy.

[0079] Finally, subtract the conditional entropy after partitioning the sample dataset using the target feature from the information entropy of the sample dataset to obtain the information gain of the target feature, i.e.:

[0080] If there are multiple target features, calculate the information gain using the same method, compare the information gain of each target feature, and select the target feature with the largest information gain as the splitting feature of the current node.

[0081] For example, a sample dataset for predicting upgrade risks contains 10 samples, categorized into "high risk" and "low risk." There are 4 "high risk" samples and 6 "low risk" samples. The dataset is partitioned using the feature of "error log count," which has two values: "more" and "less." "More" represents 3 or more error log entries, and "less" represents fewer than 3. The information gain of the "error log count" is calculated as follows: The information entropy of the sample dataset is:

[0082] The sample dataset was divided based on the number of error logs. There were 3 samples with a "high" number of error logs, including 2 "high-risk" samples and 1 "low-risk" sample. The subset with a "high" number of error logs... D The information entropy of 1 is:

[0083] There are 7 samples with "few error log entries", of which 2 are "high-risk" and 5 are "low-risk". This represents a subset of samples with "few error log entries". D The information entropy of 2 is:

[0084] The conditional entropy after dividing the sample dataset according to the "number of error logs" is:

[0085] The information gain of "error log count" is:

[0086] Optionally, the Gini index can also be used as a measure of the purity of a sample dataset; a larger Gini index indicates a more chaotic dataset. In a decision tree, the target feature that minimizes the Gini index is selected as the splitting feature. The steps for calculating the Gini index are as follows: First, calculate the Gini coefficient of the sample dataset using the following formula:

[0087] In the above formula, Gini ( D ) represents the sample dataset D The Gini coefficient, KIndicates the number of categories in the sample dataset. p k Indicates the first k Class samples in sample dataset D The proportion it accounts for.

[0088] Then, the Gini index is calculated after partitioning the sample dataset using the target feature, assuming the target feature is... A have V A number of different values ​​{ a 1, a 2,…, a v}, based on target features A Sample dataset D Divided into V Subset D 1, D 2,…, Dv The formula for calculating the Gini index is as follows:

[0089] In the above formula, Gini ( D,A ) indicates the use of target features A Split sample dataset D The Gini index after that. V Indicates the number of subsets, Representing a subset D v The number of samples in the middle Represents the sample dataset D The number of samples in the middle Representing a subset D v The Gini coefficient.

[0090] Finally, if there are multiple target features, calculate the Gini index of each target feature in the same way, and select the target feature that minimizes the Gini index as the splitting feature of the current node, because the smaller the Gini index, the higher the purity of the sample dataset after splitting.

[0091] For example, taking a sample dataset containing 10 samples related to upgrade risk prediction, the Gini index calculated after partitioning the dataset using the "number of error logs" feature is as follows: The Gini coefficient of the sample dataset is:

[0092] The sample dataset was divided based on the "number of error logs" feature, and the Gini coefficient of the subsets was calculated as follows: The subset where "number of error logs" is "high". D1. This subset contains 2 "high-risk" samples and 1 "low-risk" sample. D The Gini coefficient of 1 is:

[0093] The subset where "number of error logs" is "few" D 2, of which there are 2 "high-risk" samples and 5 "low-risk" samples, this subset D The Gini coefficient of 2 is:

[0094] The Gini index is calculated by partitioning the sample dataset using the "number of error logs":

[0095] Finally, for n Repeat the above steps for each training subset, selecting the optimal feature from a randomly chosen feature subset based on metrics such as information gain and Gini index to split nodes, and construct... n Decision trees are used to form a random forest, which is then used to obtain a random forest risk prediction model.

[0096] S330. Obtain real-time sample data during the upgrade process, input the real-time sample data into the random forest risk prediction model to predict the risk of the upgrade process, and obtain the prediction results.

[0097] In this embodiment of the invention, real-time log information during the upgrade process undergoes data preprocessing, including data cleaning, data transformation, and feature extraction, to obtain real-time sample data. This real-time sample data is then input into a trained random forest risk prediction model to predict the risk of the upgrade process, yielding prediction results. The prediction result of the random forest risk prediction model is the voting result of the predictions from each decision tree. In this embodiment, the risk of the upgrade process is divided into three levels: low risk, medium risk, and high risk. The random forest risk prediction model outputs the probability of each level. P low , P medium and P high The final risk level is determined based on the principle of maximum probability, i.e., if... P high If the maximum value is found, the prediction result is high risk. P medium If the maximum is reached, the predicted risk is medium. P low If the value is the largest, the prediction result is low risk.

[0098] In this embodiment of the invention, by continuously collecting historical upgrade information and real-time log information, a trained random forest risk prediction model is used to predict potential risks during the upgrade process. Even if the multi-dimensional comprehensive evaluation process has been passed and the upgrade is deemed feasible at the macro level, the risk prediction and intelligent decision-making process can still accurately predict potential problems that may occur during the upgrade process from a more micro and dynamic perspective, combining historical upgrade data and real-time log information.

[0099] S400: Based on the prediction results, continue with the automatic upgrade operation or perform a safe rollback operation.

[0100] In this embodiment of the invention, corresponding processing operations are performed based on the prediction results of the random forest risk prediction model. Figure 4 This is the fourth schematic diagram of the application upgrade method based on a low-code platform provided in this embodiment of the invention, as shown below. Figure 4 As shown, the specific steps for performing corresponding processing operations based on the prediction results include the following: S410. If the prediction result is low risk, continue to perform the automatic upgrade operation.

[0101] In this embodiment of the invention, if the prediction result is low risk, the automatic upgrade operation continues according to step S200. During the upgrade process, key indicators are monitored in real time, including but not limited to database connection count, API request error rate, server CPU utilization, and memory utilization.

[0102] S420. If the prediction result is medium risk, then after manual intervention, either perform an automatic upgrade operation or a safe rollback operation.

[0103] In this embodiment of the invention, if the predicted risk is medium, a warning message is issued, and a decision is made after manual intervention to determine whether to continue the automatic upgrade operation or perform a security rollback operation. Specifically, a comprehensive judgment can be made based on the results of a multi-dimensional comprehensive evaluation and the real-time log information generated during the upgrade process.

[0104] S430. If the prediction result is high risk, then pause the automatic upgrade operation and perform a safe rollback operation.

[0105] In this embodiment of the invention, if the prediction result is high risk, it indicates that an upgrade failure or an upgrade anomaly will occur. In this case, the automatic upgrade operation is paused and a safe rollback operation is performed. The safe rollback operation specifically includes the following steps: First, the target metrics of the target application are monitored in real time to obtain real-time monitoring metrics, including but not limited to database connection count, API request error rate, server CPU utilization, and memory utilization.

[0106] Then, based on real-time monitoring metrics or automated test results, automatic rollback conditions are triggered. Specifically, each real-time monitoring metric has a corresponding threshold. If a real-time monitoring metric exceeds the corresponding threshold and the duration reaches a preset time threshold, automatic rollback conditions are triggered. Simultaneously, after starting the target version of the service, automated tests will be performed. If the automated test failure rate exceeds a preset threshold, automatic rollback conditions are also triggered.

[0107] Finally, stop the target version of the service, restore the data based on the backup file, and start the current version of the service. Specifically, first stop the target version of the service to avoid affecting user operations and data consistency. Next, perform data recovery. For the database, execute the backup file in a secure environment to overwrite the current database. For file storage, decompress the backup file and accurately overwrite the application directory to ensure file integrity. In the configuration phase, remove the old version... config.json Restore to the specified location. Carefully check that all data and configurations are accurate. Finally, start the current version of the service and closely monitor its startup status and operation to ensure that business operations are quickly restored to normal.

[0108] In this embodiment of the invention, even if a multi-dimensional comprehensive evaluation deems the upgrade feasible, if the random forest risk prediction model predicts a high risk in the upgrade process based on factors such as an increase in the number of error logs in the real-time log information, the automatic upgrade operation will be suspended and a safe rollback operation will be triggered. This provides additional security for the application upgrade and further ensures the security and stability of the system during the upgrade process.

[0109] The application upgrade method based on a low-code platform provided in this invention processes the application upgrade and rollback process through multi-dimensional comprehensive evaluation and a random forest-based risk prediction algorithm. The multi-dimensional evaluation comprehensively analyzes the impact of the upgrade from multiple key levels, including application performance, compatibility, and business logic, identifying potential risk points and providing detailed evidence for decision-making. The random forest risk prediction algorithm, with its powerful data processing and pattern recognition capabilities, accurately predicts potential problems and anticipates possible faults and anomalies during the upgrade process. This achieves automated intelligent decision-making, eliminating the need for cumbersome manual intervention and quickly and accurately deciding on upgrade or rollback operations. This not only greatly improves upgrade efficiency and shortens business interruption time but also effectively avoids risks caused by human error.

[0110] Example 2 Based on the same technical concept as Embodiment 1 above, this embodiment of the invention provides an application upgrade system based on a low-code platform. Figure 5 This is a schematic diagram of the application upgrade system architecture based on a low-code platform provided in an embodiment of the present invention, such as... Figure 5 As shown, the application upgrade system 200 based on the low-code platform includes: The comprehensive evaluation module 210 is used to conduct a multi-dimensional comprehensive evaluation based on the current version information and target version information of the target application to obtain comprehensive evaluation indicators.

[0111] The automatic upgrade module 220 is used to compare the comprehensive evaluation index with the comprehensive evaluation threshold. If the comprehensive evaluation index is greater than the comprehensive evaluation threshold, the automatic upgrade operation is executed; otherwise, the automatic upgrade plan is adjusted.

[0112] The risk prediction module 230 is used to obtain historical upgrade information and real-time log information during the upgrade process, and to predict the risk of the upgrade process through a random forest risk prediction model to obtain the prediction results. The safety rollback module 240 is used to continue the automatic upgrade operation or perform a safety rollback operation based on the prediction results.

[0113] Optionally, the application upgrade system based on a low-code platform provided in this embodiment of the invention adopts a layered architecture design, including a data layer, a service layer, and an application layer. These layers communicate with each other through interfaces to achieve data interaction and function calls. The data layer is responsible for storing application version data, backup data, and log information. The service layer includes core services such as comprehensive evaluation, automatic upgrades, risk prediction and intelligent decision-making, and secure rollback. The application layer provides a user interface to facilitate users in performing upgrade operations and viewing system status.

[0114] The application upgrade system based on a low-code platform provided in this invention ensures that applications always maintain high availability and stability through automated and secure upgrade and rollback processes, providing a solid guarantee for the continuous and stable operation of low-code platform applications.

[0115] It is understood that the implementation method of the application upgrade method based on the low-code platform described in Embodiment 1 above is also applicable to this embodiment and can achieve the same technical effect, so it will not be described again here.

[0116] Example 3 Based on the same concept, embodiments of the present invention also provide an electronic device. Figure 6 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention, such as... Figure 6 As shown, the electronic device 300 may include a processor 310, a communication interface 320, a memory 330, and a communication bus 340, wherein the processor 310, the communication interface 320, and the memory 330 communicate with each other via the communication bus 340. The processor 310 can call logical instructions in the memory 330 to execute the steps of the application upgrade method based on the low-code platform as described in the above embodiments. For example, this includes: S100. Conduct a multi-dimensional comprehensive evaluation based on the current version information and target version information of the target application to obtain comprehensive evaluation indicators; S200. Compare the comprehensive evaluation index with the comprehensive evaluation threshold. If the comprehensive evaluation index is greater than the comprehensive evaluation threshold, then perform the automatic upgrade operation; otherwise, adjust the automatic upgrade plan. S300: Obtain historical upgrade information and real-time log information during the upgrade process, and use the random forest risk prediction model to predict the risk of the upgrade process and obtain the prediction results; S400: Based on the prediction results, continue with the automatic upgrade operation or perform a safe rollback operation.

[0117] The processor 310 can be a central processing unit (CPU). The processor can also be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, or combinations of the above types of chips.

[0118] Furthermore, the logical instructions in the aforementioned memory 330 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, essentially, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0119] The memory 330 may include a program storage area and a data storage area. The program storage area may store the operating system and applications required for at least one function; the data storage area may store data created by the processor, etc. Furthermore, the memory may include high-speed random access memory and non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some embodiments, the memory may optionally include memory remotely located relative to the processor, which can be connected to the processor via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0120] Example 4 Based on the same concept, embodiments of the present invention also provide a computer-readable storage medium storing a computer program. This computer program includes at least one piece of code executable by a master control device to control the master control device to implement the steps of the application upgrade method based on a low-code platform as described in the above embodiments. For example, it includes: S100. Conduct a multi-dimensional comprehensive evaluation based on the current version information and target version information of the target application to obtain comprehensive evaluation indicators; S200. Compare the comprehensive evaluation index with the comprehensive evaluation threshold. If the comprehensive evaluation index is greater than the comprehensive evaluation threshold, then perform the automatic upgrade operation; otherwise, adjust the automatic upgrade plan. S300: Obtain historical upgrade information and real-time log information during the upgrade process, and use the random forest risk prediction model to predict the risk of the upgrade process and obtain the prediction results; S400: Based on the prediction results, continue with the automatic upgrade operation or perform a safe rollback operation.

[0121] Based on the same technical concept, this embodiment of the invention also provides a computer program, which, when executed by a master control device, is used to implement the above-described method embodiments.

[0122] The computer program may be stored, in whole or in part, on a computer-readable storage medium packaged with the processor, or in part or in whole on a memory not packaged with the processor.

[0123] Based on the same technical concept, embodiments of the present invention also provide a processor for implementing the above-described method embodiments. The processor may be a chip.

[0124] In summary, the application upgrade method, system, device, and storage medium based on a low-code platform provided by this invention process the application upgrade and rollback process through multi-dimensional comprehensive evaluation and a risk prediction algorithm based on random forests. The multi-dimensional evaluation comprehensively analyzes the impact of upgrades from multiple key levels, including application performance, compatibility, and business logic, identifying potential risk points and providing detailed evidence for decision-making. The random forest risk prediction algorithm, with its powerful data processing and pattern recognition capabilities, accurately predicts potential problems and anticipates possible faults and anomalies during the upgrade process. This achieves automated intelligent decision-making, eliminating the need for cumbersome manual intervention and quickly and accurately deciding on upgrade or rollback operations. This not only greatly improves upgrade efficiency and shortens business downtime but also effectively avoids risks caused by human error. Through automated and secure upgrade and rollback processes, the application maintains high availability and stability, providing a solid guarantee for the continuous and stable operation of low-code platform applications.

[0125] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0126] The embodiments described above are merely examples of several implementations of the present invention, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the scope of protection of the present invention.

[0127] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. An application upgrade method based on a low-code platform, characterized in that, The method includes: A comprehensive evaluation is conducted based on the current version information of the target application and the target version information to obtain comprehensive evaluation indicators; The comprehensive evaluation index is compared with the comprehensive evaluation threshold. If the comprehensive evaluation index is greater than the comprehensive evaluation threshold, an automatic upgrade operation is performed; otherwise, the automatic upgrade plan is adjusted. Obtain historical upgrade information and real-time log information during the upgrade process, and use a random forest risk prediction model to predict the risk of the upgrade process and obtain the prediction results; Based on the prediction results, either continue the automatic upgrade operation or perform a safe rollback operation; The step of performing a multi-dimensional comprehensive evaluation based on the current version information and target version information of the target application to obtain comprehensive evaluation indicators includes: Calculate the feature compatibility index based on the first feature set of the current version and the second feature set of the target version; The performance change rate index is calculated based on the first performance index data of the current version and the second performance index data of the target version. The performance change rate index includes at least the response time change rate and the processing capacity change rate. Obtain vulnerability set information for the target version, and calculate security risk indicators based on the vulnerability set information; Based on the system resource requirements of the target version and the system resource threshold information, resource bottleneck indicators are calculated. The resource bottleneck indicators include at least CPU indicators, memory indicators, and disk space indicators. The comprehensive evaluation index is obtained by comprehensively calculating the functional compatibility index, the performance change rate index, the security risk index, and the resource bottleneck index.

2. The application upgrade method based on a low-code platform according to claim 1, characterized in that, The formula for calculating the comprehensive evaluation index is as follows: In the above formula, S Indicates comprehensive evaluation indicators, C func Indicates functional compatibility metrics. R T Indicates the rate of change of response time. R Q This indicates the rate of change in processing capacity. R sec Indicators representing security risk. max ( R sec This indicates the highest security risk indicator recorded during the historical upgrade process. B cpu Indicates CPU metrics, B mem Indicates memory metrics, B disk This represents a disk space metric. , , , and These represent the first, second, third, fourth, and fifth weight coefficients, respectively, and the sum of all weight coefficients equals 1.

3. The application upgrade method based on a low-code platform according to claim 1, characterized in that, The process of acquiring historical upgrade information and real-time log information during the upgrade process, and using a random forest risk prediction model to predict the risk of the upgrade process and obtain prediction results includes: The historical upgrade information and the real-time log information are preprocessed to obtain a sample dataset; Based on the sample dataset, a risk prediction model is constructed and trained using the random forest algorithm to obtain the random forest risk prediction model. Real-time sample data is obtained during the upgrade process, and the real-time sample data is input into the random forest risk prediction model to predict the risk of the upgrade process, thereby obtaining the prediction result.

4. The application upgrade method based on a low-code platform according to claim 3, characterized in that, The historical upgrade information and the real-time log information are preprocessed to obtain a sample dataset, including: The historical upgrade information and the real-time log information are cleaned to obtain a first dataset. The data cleaning includes at least missing value handling, outlier handling and data consistency checking. The first dataset is transformed to obtain the second dataset. The data transformation includes at least standardization, classification feature encoding, and temporal feature processing. Feature extraction is performed on the second dataset to obtain the sample dataset. The feature extraction includes at least feature extraction based on business logic, feature extraction based on statistical methods, and feature extraction fused with real-time log information.

5. The application upgrade method based on a low-code platform according to claim 1, characterized in that, Based on the prediction results, either continue the automatic upgrade operation or perform a safe rollback operation, including: If the prediction result is low risk, then continue with the automatic upgrade operation; If the prediction result is medium risk, then after manual intervention, an automatic upgrade operation or a safe rollback operation will be performed. If the prediction result is high risk, then the automatic upgrade operation will be suspended and a safe rollback operation will be performed.

6. The application upgrade method based on a low-code platform according to claim 5, characterized in that, The automatic upgrade operation includes: Back up and verify the application data, configuration data, code, and metadata of the target application to obtain backup files; After performing compatibility checks and resource assessments on the target version, the target version is deployed to the server and the database is migrated. Start the target version of the service and perform automated testing, which includes at least core function checks and key interface response status checks. The execution of the safe rollback operation includes: Real-time monitoring of the target metrics of the target application is performed to obtain real-time monitoring metrics; Automatic rollback conditions are triggered based on the real-time monitoring metrics or automated test results. Stop the target version of the service, restore the data based on the backup file, and start the current version of the service.

7. An application upgrade system based on a low-code platform, characterized in that, The system includes: The comprehensive evaluation module is used to perform a multi-dimensional comprehensive evaluation based on the current version information and target version information of the target application to obtain comprehensive evaluation indicators; The automatic upgrade module is used to compare the comprehensive evaluation index with the comprehensive evaluation threshold. If the comprehensive evaluation index is greater than the comprehensive evaluation threshold, an automatic upgrade operation is performed; otherwise, the automatic upgrade plan is adjusted. The risk prediction module is used to obtain historical upgrade information and real-time log information during the upgrade process, and to predict the risks of the upgrade process using a random forest risk prediction model to obtain the prediction results. The safety rollback module is used to either continue the automatic upgrade operation or perform a safety rollback operation based on the prediction result. The step of performing a multi-dimensional comprehensive evaluation based on the current version information and target version information of the target application to obtain comprehensive evaluation indicators includes: Calculate the feature compatibility index based on the first feature set of the current version and the second feature set of the target version; The performance change rate index is calculated based on the first performance index data of the current version and the second performance index data of the target version. The performance change rate index includes at least the response time change rate and the processing capacity change rate. Obtain vulnerability set information for the target version, and calculate security risk indicators based on the vulnerability set information; Based on the system resource requirements of the target version and the system resource threshold information, resource bottleneck indicators are calculated. The resource bottleneck indicators include at least CPU indicators, memory indicators, and disk space indicators. The comprehensive evaluation index is obtained by comprehensively calculating the functional compatibility index, the performance change rate index, the security risk index, and the resource bottleneck index.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, The processor executes the computer program to implement the application upgrade method based on a low-code platform as described in any one of claims 1-6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the application upgrade method based on a low-code platform as described in any one of claims 1-6.

Images