Method for multi-system operation and maintenance inspection based on artificial intelligence RPA inspection robot

Abstract

The application relates to the field of IT system operation and maintenance management and discloses a kind of inspection methods for multi-system operation and maintenance of RPA inspection robot based on artificial intelligence, the RPA inspection robot is connected with multiple target systems needing inspection through pre-set interface protocol, and login information of a system page is automatically recognized based on process automation RPA software; the verification code information in the login information is automatically extracted, and the verification code recognition service interface is called for verification code recognition; after the verification code recognition, the RPA robot automatically logs in each target system needing inspection, obtains page information, and stores the collected page information into the inspection directory of the RPA robot; according to the inspection directory, the information collected by each target system needing inspection is uniformly collected and processed for abnormality detection, and the alarm information of the page alarm module is analyzed; the alarm information page is intercepted, and the RPA robot sends the alarm information to the operation and maintenance personnel; the application improves work efficiency.

Description

Technical Field

[0001] The present invention relates to the field of IT system operation and maintenance management, and in particular to an inspection method for multi-system operation and maintenance using an artificial intelligence (AI)-based RPA inspection robot. Background Art

[0002] 1. Technical complexity and rapid updates: Operations and maintenance work involves multiple technical fields, including networks, systems, databases, applications, and security. The complexity and professionalism of these fields require operations and maintenance personnel to have comprehensive knowledge and skills. At the same time, the rapid development of technology requires operations and maintenance personnel to constantly learn new knowledge and master new skills, which increases the difficulty of operations and maintenance work.

[0003] 2. Heavy workload and tedious details: Operations and maintenance personnel are responsible for multiple tasks, including daily inspections, monitoring, troubleshooting, and system upgrades. These tasks are onerous and require high concentration. In addition, the work details, such as configuration management, log analysis, and performance tuning, are tedious and require a lot of time and energy, and are prone to errors.

[0004] 3. Monitoring and Troubleshooting: As IT systems become more complex, monitoring becomes increasingly challenging. Real-time monitoring of system status and the prompt identification of potential issues are essential. However, due to the diversity of data sources and the high real-time requirements, monitoring becomes extremely challenging. Troubleshooting is also complicated by the complex system architecture and numerous components, requiring significant time and effort.

[0005] 4. Insufficient Automation and Standardization: Some companies still rely on manual operations for operations and maintenance, resulting in low efficiency and prone to errors. The use of automated operations and maintenance tools can significantly improve operations and maintenance efficiency and reduce the risk of human error, but the current lack of automation is a pain point. In addition, the lack of unified standards leads to a lack of standardization in operations and maintenance.

[0006] 5. Resource Constraints and Pressure: When faced with complex O&M tasks, operations teams are often constrained by resource constraints, including human resources, technical resources, and financial resources. Insufficient resources can hinder the smooth implementation of O&M work and even impact the normal operation of business systems. Furthermore, the O&M team bears the crucial responsibility of ensuring the stable operation of business systems, and the immense pressure and psychological stress of this responsibility further complicates the work.

[0007] ‌6. Cross-departmental collaboration and communication‌: In large enterprises, operation and maintenance work involves collaboration and communication between multiple departments and teams; however, due to information asymmetry between departments and poor communication channels, collaboration efficiency is often low. Summary of the Invention

[0008] The purpose of this invention is to solve the above problems and design an inspection method for multi-system operation and maintenance using an artificial intelligence-based RPA inspection robot.

[0009] The present invention provides an inspection method for performing multi-system operation and maintenance using an artificial intelligence (AI)-based RPA inspection robot. The inspection method comprises the following steps:

[0010] The RPA inspection robot establishes connections with multiple target systems that need to be inspected through a pre-set interface protocol and automatically identifies the login information on the system login page based on the process automation RPA software. The login information includes at least the login user name, password, and verification code information.

[0011] Automatically extract the verification code information from the login information and call the verification code recognition service interface to perform verification code recognition;

[0012] After the verification code is recognized, the RPA robot automatically logs into each target system that needs to be inspected. Based on the operation and maintenance objectives of each system, it obtains information including at least the target page and the system operation status page, and stores the collected page information in the RPA robot's inspection directory. The target systems that need to be inspected include at least the full network behavior management system, operation and maintenance audit and risk control system, network controller system, enterprise-level data backup and recovery system, Internet behavior management system, application delivery management system, firewall system, cloud platform system, and IT operation and maintenance monitoring platform.

[0013] According to the inspection catalog, the information collected from each target system that needs to be inspected is unified and aggregated, and anomaly detection is performed, and the alarm information of the page alarm module is analyzed;

[0014] The alarm information page is captured and stored in the RPA robot, which then sends the alarm information to the operation and maintenance personnel.

[0015] Optionally, in a first implementation of the present invention, the RPA inspection robot establishes connections with multiple target systems that need to be inspected through a pre-set interface protocol, and automatically identifies login information on the system login page based on the process automation RPA software, including:

[0016] Load and read the RPA tool software package and its corresponding development toolkit, and integrate the RPA tool with inspection rules;

[0017] Start the RAP engine, initialize engine parameters and initialize run instructions, where the engine parameters include at least thread pool size, timeout setting, and retry threshold;

[0018] Clarify the hierarchical structure of the inspection directory, analyze storage rules, and determine the storage formats for different types of data, including at least storing performance indicator data in CSV format in time series and storing log files in raw text format in date-named folders;

[0019] The RPA engine is called and the inspection system URL stored in the configuration library is passed to the browser startup parameter, so that the browser automatically navigates to the login page of the target inspection system after startup.

[0020] When a browser encounters a certificate error when loading the inspection system login page, use the thisisunsafe command to ignore the web certificate error and enter the inspection login page.

[0021] Drive the RPA engine to automatically identify page elements and determine whether the page has been loaded by judging whether the page elements exist. When the page elements are loaded, the user name input box on the login page is automatically entered into the user name, the password input box on the login page is automatically entered into the password, and the consent selection box on the login page is automatically checked. Then, the login button is identified to complete the system login.

[0022] Optionally, in a second implementation of the present invention, automatically extracting verification code information from the login information and calling a verification code recognition service interface to perform verification code recognition includes:

[0023] Start the RPA browser plug-in runtime environment, load the pre-configured page element recognition rule set, traverse the HTML structure of the page using DOM parsing technology, locate the node where the verification code information element is located, and identify the verification code code;

[0024] Send an HTTPS request to the server and monitor the response status of the request in real time. If a successful response is received, obtain a call token and use the call token to read the image recognition API access information stored in the local configuration file;

[0025] Extract the returned JSON value, parse the JSON value to obtain the recognition result, pass the recognition result to the system login, automatically fill in the verification code input box during the login process, and complete the entire verification code processing flow.

[0026] Optionally, in a third implementation of the present invention, after the verification code is recognized, the RPA robot automatically logs in to each target system that needs to be inspected, obtains information including at least the target page and the system operation status page based on the operation and maintenance goals of each system, and stores the collected page information in the RPA robot's inspection directory, including:

[0027] After the verification code is successfully recognized and automatically filled in, the RPA robot uses the previously obtained login account and password information to simulate a manual click of the login button and send a login request to the target system server.

[0028] Continuously monitor the response information returned by the server and confirm whether the login is successful by analyzing the response status code, page jump indicator, and login success prompt element. If a successful login signal is not detected within 10 seconds, the RPA robot will re-initiate the login process according to the preset retry mechanism until it successfully logs into each target system that needs to be inspected.

[0029] After successful login, the RPA robot starts the built-in page collection module, allocates memory space as a temporary data buffer, and reads the page collection rule file for the target system. The page collection rules include at least the collection priority, collection frequency, and data storage format for different page elements.

[0030] Initialize the parameters of the page collection module according to the page collection rules. The RPA robot uses the built-in DOM parsing engine to deeply analyze the target page. By identifying the HTML tag structure, CSS style class names, and JavaScript event triggering areas of the page, it locates the page elements containing key information.

[0031] The RPA robot automatically extracts the data content from the page elements where key information is located and pre-processes the data content;

[0032] Based on the URL address or page redirection configuration of the system operation status page, the RPA robot is driven to jump from the current page to the system operation status page. After the system operation status page is loaded, the RPA robot mines the detailed system operation status data on the page.

[0033] Integrate the data content of the page elements containing key information temporarily stored in the memory cache and the detailed data on system operation status. Use the RPA robot to store the integrated data in the inspection directory according to the storage path.

[0034] Optionally, in a fourth implementation of the present invention, the RPA robot mines detailed system operation status data on the page, including:

[0035] Use underlying network request sniffing technology to capture the AJAX requests that interact with the server behind the page, parse the returned data, and obtain system performance indicators pushed in real time by the server. System performance indicators include at least CPU usage, memory usage, and network bandwidth utilization.

[0036] For system operation trend data displayed in the form of visual charts on the page, we combine OCR technology with chart parsing algorithms to convert the data points in the chart into a structured numerical sequence;

[0037] The collected system operation status data is temporarily stored in the memory cache area and preliminarily sorted according to the preset data classification rules, wherein the preset data classification rules at least include data type classification, timestamp classification and data source classification.

[0038] Optionally, in a fifth implementation of the present invention, based on the inspection catalog, information collected from each target system requiring inspection is unified and aggregated, anomaly detection is performed, and alarm information of a page alarm module is analyzed, including:

[0039] Applying data mining algorithms to extract key features from the processed data, inputting the key features into an anomaly detection model to perform anomaly detection, and outputting the detection results;

[0040] When an exception occurs, obtain the alarm information issued by the target system page alarm module, wherein the alarm information at least includes an alarm code, an alarm level and an alarm description;

[0041] Natural language processing technology is used to pre-process the alarm description, extract key semantic information, and locate the problem by combining the alarm code and alarm level.

[0042] Optionally, in a sixth implementation of the present invention, extracting key features from the processed data using a data mining algorithm includes:

[0043] Calculate the Pearson correlation coefficient between each feature in the processed data and the system's target operation and maintenance status, and obtain features whose absolute value of the correlation is higher than the set threshold, where the threshold is 0.3;

[0044] Variance analysis was used to test the mean differences of each feature under different categories of operation and maintenance status, and the features under different system health conditions were screened out to obtain highly correlated feature data;

[0045] Highly correlated feature data is used as input and imported into the LDA model framework. The target dimension of the LDA model is set according to the actual number of system operation and maintenance categories. The high-dimensional feature space is projected into a low-dimensional discriminant subspace, maximizing the inter-class distances of different operation and maintenance status categories within the subspace while minimizing the intra-class variance. The system operation and maintenance categories include at least normal operation, minor failure, and major failure.

[0046] By iteratively optimizing the projection matrix, samples with different operation and maintenance status are divided in the new feature space after projection, and key features are extracted.

[0047] Optionally, in a seventh implementation of the present invention, inputting the key features into an anomaly detection model to perform anomaly detection and outputting a detection result includes:

[0048] A bidirectional RNN is used as the basic model of the anomaly detection model. The key features are input into the anomaly detection model. In the forward propagation phase of the model, starting from the starting time step of the sequence, the forward RNN unit reads the key feature vectors one by one in sequence, and calculates the hidden layer output at the current moment through the activation function based on the current input features and the output state of the hidden layer at the previous moment;

[0049] The reverse RNN unit starts from the last time step of the sequence, reads the feature vector in reverse, and calculates the output based on the input and the reverse hidden layer state at the previous moment. It transmits the key information at the end of the sequence in reverse and merges it with the forward information.

[0050] The detection results are output through the anomaly detection model and transmitted to the RPA inspection robot in real time through the communication protocol.

[0051] In the technical solution provided by the present invention, RPA automated robots log in to different systems of the enterprise according to the enterprise's own situation, automatically capture the pages that focus on and reflect the system operation status, and uniformly collect, process and store them for review by operation and maintenance personnel, so as to replace the operation and maintenance personnel's manual login to multiple systems at regular intervals and locations to perform inspection operations; use automated inspection tools to improve work efficiency, ensure that employees can quickly adapt and improve overall work efficiency; the present invention can take inspection screenshots of the following systems to assist operation and maintenance personnel in performing operation and maintenance operations. The inspection system includes the following contents: full network behavior management system, operation and maintenance audit and risk control system, network controller system, enterprise-level data backup and recovery system, Internet behavior management system, application delivery management system, firewall system, cloud platform system, and IT operation and maintenance monitoring platform. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Various other advantages and benefits will become apparent to those skilled in the art by reading the following detailed description of the preferred embodiment.The accompanying drawings are only for the purpose of illustrating the preferred embodiment and are not to be considered as limiting the present invention.

[0053] Figure 1 A schematic diagram of an embodiment of a multi-system operation and maintenance inspection method using an artificial intelligence-based RPA inspection robot provided in an embodiment of the present invention;

[0054] Figure 2 Schematic diagram of the login steps of the inspection method for performing multi-system operation and maintenance using an artificial intelligence-based RPA inspection robot provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0055] The terms "first," "second," "third," "fourth," and so forth (if any) in the description and claims of the present invention and in the accompanying drawings are used to distinguish similar items and are not necessarily used to describe a particular order or precedence. It should be understood that the terms used in this manner are interchangeable where appropriate, so that the embodiments described herein can be implemented in an order other than that shown or described herein. In addition, the terms "including" or "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, apparatus, product, or device that includes a series of steps or elements is not necessarily limited to those steps or elements expressly listed, but may include other steps or elements not expressly listed or inherent to such process, method, product, or device.

[0056] For ease of understanding, the specific process of the embodiment of the present invention is described below. Figure 1 An embodiment of the present invention provides a schematic diagram of an embodiment of a multi-system operation and maintenance inspection method using an artificial intelligence-based RPA inspection robot. The method specifically includes the following steps:

[0057] Step 101: The RPA inspection robot establishes connections with multiple target systems that need to be inspected through a pre-set interface protocol, and automatically identifies login information on the system login page based on the process automation RPA software.

[0058] In this embodiment, the login information includes at least the login user name, password and verification code information;

[0059] In this embodiment, the RPA tool software package and its corresponding development toolkit are loaded and read, and the RPA tool and inspection rules are integrated; the RAP engine is started, the engine parameters and the initialization running instructions are initialized, where the engine parameters include at least the thread pool size, the timeout setting and the retry threshold; the hierarchical structure of the inspection directory is clarified, the storage rules are parsed, and the storage format of different types of data is determined, including at least the performance indicator data is stored in CSV format in time series and the log files are stored in the original text form in a folder named by the date; the RPA engine is called, and the inspection system URL pre-stored in the configuration library is passed to the browser startup parameters, so that the browser automatically navigates to the login page of the target inspection system after startup; when the browser encounters a certificate error prompt when loading the inspection system login page, the thisisunsafe command is used to ignore the web page certificate error and enter the inspection login page; the RPA engine is driven to automatically identify page elements, and determine whether the page is loaded by judging whether the page elements exist. When the page elements are loaded, the username input box on the login page is automatically input with the username, the password input box on the login page is automatically input with the password, and the consent selection box on the login page is automatically checked, and then the login button is identified to complete the system login.

[0060] Step 102: Automatically extract the verification code information from the login information and call the verification code recognition service interface to perform verification code recognition;

[0061] In this embodiment, the RPA browser plug-in operating environment is started, the pre-configured page element recognition rule set is loaded, the HTML structure of the page is traversed through DOM parsing technology, the node where the verification code information element is located is located, and the verification code code is identified; an HTTPS request is sent to the server, and the response status of the request is monitored in real time. If a successful response is received, a call token is obtained, and the image recognition API access information stored in the local configuration file is read through the call token; the returned JSON value is extracted, the JSON value is parsed to obtain the recognition result, and the recognition result is passed to the system login. The verification code input box is automatically filled in during the login process to complete the entire verification code processing flow.

[0062] Step 103: After the verification code is recognized, the RPA robot automatically logs in to each target system that needs to be inspected. Based on the maintenance objectives of each system, the robot obtains information including at least the target page and the system operation status page, and stores the collected page information in the RPA robot's inspection directory.

[0063] In this embodiment, the target systems that need to be inspected include at least the entire network behavior management system, operation and maintenance audit and risk control system, network controller system, enterprise-level data backup and recovery system, Internet behavior management system, application delivery management system, firewall system, cloud platform system, and IT operation and maintenance monitoring platform.

[0064] In this embodiment, after the verification code is successfully recognized and automatically filled in, the RPA robot is driven to use the previously obtained login account and password information to simulate the manual operation of clicking the login button and send a login request to the target system server; the response information returned by the server is continuously monitored, and the login is confirmed to be successful by parsing the response status code, page jump identifier and login success prompt element. If a successful login signal is not detected within 10 seconds, the RPA robot will re-initiate the login process according to the preset retry mechanism until it successfully logs in to each target system that needs to be inspected; after successful login, the built-in page acquisition module of the RPA robot is started, and memory space is allocated as a temporary data cache area, and the page acquisition rule file for the target system is read, where the page acquisition rule includes at least the acquisition priority, acquisition frequency and data storage format of different page elements; the page is initialized according to the page acquisition rule The RPA robot collects the module parameters and uses the built-in DOM parsing engine to deeply analyze the target page. By identifying the HTML tag structure, CSS style class name, and JavaScript event triggering area of ​​the page, it locates the page element where the key information is located. The RPA robot automatically extracts the data content in the page element where the key information is located and pre-processes the data content. Based on the URL address of the system operation status page or the page jump path configuration information, the RPA robot is driven to jump from the current page to the system operation status page. After the system operation status page is loaded, the RPA robot mines the detailed system operation status data on the page. The data content of the page element where the key information is located and the detailed system operation status data temporarily stored in the memory cache are integrated. According to the storage path, the RPA robot stores the integrated data in the inspection directory.

[0065] In this embodiment, the underlying network request sniffing technology is used to capture the AJAX requests that interact with the server behind the page, parse the returned data, and obtain system performance indicators pushed in real time by the server, where the system performance indicators include at least CPU usage, memory occupancy, and network bandwidth utilization; for the system operation trend data displayed in the form of visual charts on the page, the OCR technology and chart parsing algorithm are combined to convert the data points in the chart into a structured numerical sequence; the collected system operation status data is temporarily stored in the memory cache area and preliminarily sorted according to preset data classification rules, where the preset data classification rules include at least data type classification, timestamp classification, and data source classification.

[0066] Step 104: Based on the inspection catalog, the information collected from each target system that needs to be inspected is unified and aggregated, and anomaly detection is performed, and the alarm information of the page alarm module is analyzed;

[0067] In this embodiment, a data mining algorithm is used to extract key features from the processed data, and the key features are input into an anomaly detection model to perform anomaly detection and output the detection results. When an anomaly occurs, the alarm information issued by the target system page alarm module is obtained, where the alarm information includes at least an alarm code, an alarm level, and an alarm description. Natural language processing technology is used to preprocess the alarm description, extract key semantic information, and locate the problem in combination with the alarm code and alarm level.

[0068] In this embodiment, the Pearson correlation coefficient between each feature in the processed data and the target operation and maintenance status of the system is calculated to obtain features whose absolute value of the correlation is higher than a set threshold, where the threshold is 0.3; variance analysis is used to test the mean difference of each feature under different categories of operation and maintenance status, and features under different system health conditions are screened out to obtain high-correlation feature data; the high-correlation feature data is used as input and imported into the LDA model framework, and the target dimension of the LDA model is set according to the actual number of categories of system operation and maintenance, and the high-dimensional feature space is projected into the low-dimensional discriminant subspace, so that the inter-class distance of different operation and maintenance status categories in the subspace is maximized, and the intra-class variance is minimized, where the categories of system operation and maintenance include at least normal operation, minor faults and serious faults; by iteratively optimizing the projection matrix, samples of different operation and maintenance status are divided in the new feature space after projection, and key features are extracted.

[0069] In this embodiment, a bidirectional RNN is used as the basic model of the anomaly detection model, and the key features are input into the anomaly detection model. In the forward propagation stage of the model, starting from the starting time step of the sequence, the forward RNN unit reads the key feature vectors one by one in sequence, and calculates the hidden layer output at the current moment through the activation function based on the current input features and the output state of the hidden layer at the previous moment; the reverse RNN unit starts from the last time step of the sequence, reads the feature vector in reverse, and also calculates the output based on the input and the reverse hidden layer state at the previous moment, and transmits the key information at the back end of the sequence in reverse and merges it with the forward transmitted information; the detection results are output through the anomaly detection model, and the detection results are transmitted to the RPA inspection robot in real time through the communication protocol.

[0070] Step 105: Intercept the alarm information page, store it in the RPA robot, and have the RPA robot send the alarm information to the operation and maintenance personnel.

[0071] Another embodiment of a multi-system operation and maintenance inspection method using an artificial intelligence (AI)-based RPA inspection robot includes the following steps:

[0072] System login: Automatically identify the login information of the system login page based on process automation software, including the login user name, password and verification code information;

[0073] Verification code recognition: Automatically extract the page verification code image based on machine process automation, and call the general verification code recognition service interface to perform verification code recognition;

[0074] System inspection: Log in to each inspection system, focus on the operation and maintenance of each system, intercept the page of interest, reflect the system operation status and other page information, and save it in the designated inspection directory;

[0075] Data collection: unified processing of inspection data collected from multiple different systems;

[0076] Alarm analysis: After logging into the system, the system automatically analyzes the alarm information of the page alarm module, captures the alarm information page, stores it in the designated inspection directory, and stores the image according to the screenshot naming rules, so that operation and maintenance personnel can confirm the system alarm information through the file name.

[0077] See also Figure 2 The figure shows a login process diagram of a multi-system operation and maintenance inspection method using an artificial intelligence-based RPA inspection robot, including the following steps:

[0078] Integrate and package the RPA engine: Integrate RPA tools and inspection rules to automate the inspection logic process;

[0079] Initialize the RPA engine: Start the RPA engine and initialize engine parameters and methods.

[0080] Initialize inspection system parameters: define inspection directory and storage rules;

[0081] RPA calls the browser to open the inspection URL: calls the RPA engine, opens the system default browser, and automatically enters the VPN system URL;

[0082] Resolve certificate errors: Use the "thisisunsafe" command to ignore certificate errors, skip the "There is a problem with this website's security certificate" page, and enter the VPN login page.

[0083] RPA calls the browser to check and log in to the system: open the inspection URL: call the RPA engine, automatically identify the page elements through the browser plug-in, and determine whether the page loading is complete by judging whether the page elements exist. When the page elements are loaded, the username input box on the login page is identified to automatically enter the username, the password input box on the login page is identified to automatically enter the password, and the consent selection box on the login page is identified to automatically check the box, and then the login button is identified to complete the VPN system login.

[0084] Based on the different information and key points of different systems, the methods and steps for each system to conduct inspections are as follows:

[0085] 1. Inspection of the entire network behavior management system:

[0086] Use the default browser to open the full-network behavior management system: Call the RPA engine, open the system's default browser, and automatically enter the full-network behavior management system URL;

[0087] Resolve certificate errors: Use the "thisisunsafe" command to ignore certificate errors, skip the "There is a problem with this website's security certificate" page, and enter the full network behavior management system login page;

[0088] Login to the entire network behavior management system: Call the RPA engine and automatically identify page elements through the browser plug-in. Determine whether the page has been loaded by judging whether the page elements exist. When the page elements are loaded, the user name input box on the login page is automatically entered into the user name, the password input box on the login page is automatically entered into the password, and the consent selection box on the login page is automatically checked. Then, the login button is identified to complete the login to the entire network behavior management system.

[0089] Capture the homepage screen: The RPA engine is invoked to automatically identify page elements. The system checks the presence of these elements to determine if the page has loaded successfully. Once the elements have loaded, the system captures the homepage screen and saves the screenshot to the inspection directory. The screenshot should include information reflecting network behavior, such as device status icons, management status information, traffic analysis (interface throughput line chart), user traffic rankings, and application traffic rankings.

[0090] 2. Operation and maintenance audit and risk control system inspection:

[0091] Use the default browser to open the operation and maintenance audit and risk control system: Call the RPA engine, open the system default browser, and automatically enter the operation and maintenance audit and risk control system URL;

[0092] Resolve certificate errors: Use the "thisisunsafe" command to ignore certificate errors, skip the "There is a problem with this website's security certificate" page, and enter the operation and maintenance audit and risk control system login page;

[0093] Login to the entire network behavior management system: Call the RPA engine and automatically identify page elements through browser plug-ins. Determine whether the page has been loaded by judging whether the page elements exist. When the page elements are loaded, identify the username input box on the login page and automatically enter the username, identify the password input box on the login page and automatically enter the password, identify the consent selection box on the login page and automatically check it. Then, call the Baidu image recognition API to identify the login verification code, automatically parse the verification code recognition result, and automatically fill the verification code into the verification code input box. Finally, identify the login button to complete the operation and maintenance audit and risk control system login.

[0094] Capture the main interface screen: The RPA engine is invoked to automatically identify page elements. The page load is determined by determining whether the elements are present. Once the elements are loaded, a screenshot of the main interface is captured and saved in the inspection directory. The screenshot should include information such as operation and maintenance statistics, operation and maintenance statistics, real-time monitoring information, system operation status, and license information.

[0095] Capture the network configuration screen: This calls the RPA engine, automatically enters the network configuration page URL, opens the page, automatically identifies page elements, and determines whether the page has loaded by checking whether the elements exist. Once the elements have loaded, the network configuration screen is captured and saved in the inspection directory. The screenshot should include a detailed list of interface information, among other things.

[0096] 3. NAC network controller system inspection:

[0097] Use the default browser to open the network controller system: Call the RPA engine, open the system default browser, and automatically enter the network controller system URL;

[0098] Resolve certificate errors: Use the "thisisunsafe" command to ignore certificate errors, skip the "There is a problem with this website's security certificate" page, and enter the operation and maintenance audit and risk control system login page;

[0099] Network controller system login: The RPA engine is called and the browser plug-in is used to automatically identify page elements. The page loading is determined by judging whether the page elements exist. When the page elements are loaded, the user name input box on the login page is automatically entered, the password input box on the login page is automatically entered, the consent check box on the login page is automatically checked, and the login button is finally identified to complete the network controller system login.

[0100] Capture the main interface screen: The RPA engine is invoked to automatically identify page elements. The system checks the presence of these elements to determine if the page has loaded successfully. Once the page has loaded successfully, the system captures the main interface screen and saves the screenshot to the inspection directory. The screenshot should include information such as device resources, system status, network port status, access point status, and a total throughput chart.

[0101] Capture the address pool status screen: The RPA engine is invoked, and the URL of the address pool status page is automatically entered. The page opens and automatically identifies page elements. The page loads successfully by checking whether the elements are present. When the elements are loaded, the RPA captures the address pool status screen and saves the screenshot to the inspection directory. The screenshot should include a detailed wireless address pool status chart and other information.

[0102] 4. Enterprise-level data backup and recovery system inspection:

[0103] Use the default browser to open the enterprise-level data backup and recovery system: Invoke the RPA engine, open the system's default browser, and automatically enter the URL of the enterprise-level data backup and recovery system;

[0104] Resolve certificate errors: Use the "thisisunsafe" command to ignore certificate errors, skip the "There is a problem with this website's security certificate" page, and enter the enterprise-level data backup and recovery system login page;

[0105] Enterprise-level data backup and recovery system login: The RPA engine is called to automatically identify page elements by browsing the browser plug-in. The page loading is determined by judging whether the page elements exist. When the page elements are loaded, the user name input box on the login page is automatically entered into the user name, the password input box on the login page is automatically entered into the password, the consent selection box on the login page is automatically checked, and finally the login button is identified to complete the enterprise-level data backup and recovery system login.

[0106] Capture the main interface screen: The RPA engine is invoked to automatically identify page elements. The system checks the presence of these elements to determine if the page has loaded successfully. Once the page has loaded successfully, the system captures the main interface screen and saves the screenshot to the inspection directory. The screenshot should include information such as system time, cumulative run time, cumulative protected data, virtualization center, backup storage, backup node chart, network traffic chart, and current task list.

[0107] Alarm information reminder: Call the RPA engine to automatically parse the content of the page alarm information node. When an alarm is found in the system, the main screen screenshot information is named as a picture name containing "alarm". When the operation and maintenance personnel find a picture with the file name containing "alarm", they log in to the operation and maintenance system to check the system alarm.

[0108] 5. VPN Internet Behavior Management System Inspection:

[0109] Use the default browser to open the Internet behavior management system: Call the RPA engine, open the system's default browser, and automatically enter the URL of the Internet behavior management system;

[0110] Resolve certificate errors: Use the "thisisunsafe" command to ignore certificate errors, skip the "There is a problem with this website's security certificate" page, and enter the Internet Behavior Management System login page;

[0111] Logging in to the Internet Behavior Management System: The RPA engine is called to automatically identify page elements by browsing the browser plug-in. The page loading is determined by judging whether the page elements exist. When the page elements are loaded, the user name input box on the login page is automatically entered into the user name, the password input box on the login page is automatically entered into the password, the consent check box on the login page is automatically checked, and finally the login button is identified to complete the login to the Internet Behavior Management System.

[0112] Capture the main interface screen: The RPA engine is invoked to automatically identify page elements. The system then checks the presence of these elements to determine if the page has loaded successfully. Once the page has loaded successfully, the system captures the main interface screen and saves the screenshot to the inspection directory. The screenshot should include a line chart of system information, a real-time network throughput line chart, a real-time concurrent session line chart, a line chart of line status, a real-time user concurrency trend line chart, and a real-time stream cache status line chart.

[0113] 6. Application delivery management system inspection:

[0114] Use the default browser to open the application delivery management system: Call the RPA engine, open the system's default browser, and automatically enter the application delivery management system URL;

[0115] Resolve certificate errors: Use the "thisisunsafe" command to ignore certificate errors, skip the "There is a problem with this website's security certificate" page, and enter the application delivery management system login page;

[0116] Application Delivery Management System Login: The RPA engine is called to automatically identify page elements by browsing the browser plug-in. The page loading is determined by determining whether the page elements exist. When the page elements are loaded, the RPA engine automatically enters the user name in the login page's user name input box, automatically enters the password in the login page's password input box, automatically selects the consent check box in the login page, and finally identifies the login button to complete the application delivery management system login.

[0117] Capture application load screen information: The RPA engine is invoked to automatically identify page elements and determine whether the page has loaded by checking whether the elements exist. Once the page elements have loaded, the application load screen information is captured and saved in the inspection directory. The screenshot should include: an overall application load chart, device operation information (CPU, memory, power supply, temperature, fan), a line chart of system throughput, a line chart of the number of new system connections, and a line chart of concurrent system connections.

[0118] Capture link load screen information: The RPA engine is invoked, and the link load page URL is automatically entered. The link load page opens and automatically identifies page elements. The page load is determined by determining whether the elements exist. Once the elements are loaded, the link load screen information is captured and saved in the inspection directory. The screenshot should include information such as link status, a line chart showing the real-time link load status, a line chart showing stability monitoring, a bar chart showing application traffic share, a line chart showing system throughput, a line chart showing the number of new system connections, and a line chart showing concurrent system connections.

[0119] 7. Firewall system inspection:

[0120] Use the default browser to open the firewall system: Call the RPA engine, open the system default browser, and automatically enter the firewall system URL;

[0121] Resolve certificate errors: Use the "thisisunsafe" command to ignore certificate errors, skip the "There is a problem with this website's security certificate" page, and enter the firewall system login page;

[0122] Firewall system login: The RPA engine is called to automatically identify page elements by browsing the browser plug-in. The page loading is determined by judging whether the page elements exist. When the page elements are loaded, the user name input box on the login page is automatically entered into the user name, the password input box on the login page is automatically entered into the password, the consent selection box on the login page is automatically checked, and finally the login button is identified to complete the firewall system login.

[0123] Capture the homepage screen: The RPA engine is invoked to automatically identify page elements. The presence of these elements determines whether the page has loaded successfully. Once the elements have loaded, a screenshot is captured and saved in the inspection directory. The screenshot should include information such as business security risk analysis, attack trends, and hot event charts; user security risk analysis, attack trends, and hot event charts; device and system operations information such as device status, interface status, and security capabilities; and network operations information such as concurrent sessions and interface throughput.

[0124] Capture network interface screen information: The RPA engine is invoked, the web interface page URL is automatically entered, the web interface page is opened, and the page elements are automatically identified. The page load is determined by checking whether the elements exist. When the page elements are loaded, the network interface screen information is captured and saved in the inspection directory. The screenshot should include information such as detailed physical interface information and a status list.

[0125] 8. Cloud platform system inspection:

[0126] Use the default browser to open the cloud platform system: Call the RPA engine, open the system default browser, and automatically enter the cloud platform system URL;

[0127] Resolve certificate errors: Use the "thisisunsafe" command to ignore certificate errors, skip the "There is a problem with this website's security certificate" page, and enter the cloud platform system login page;

[0128] Cloud platform system login: The RPA engine is called to automatically identify page elements by browsing the browser plug-in. The page loading is determined by judging whether the page elements exist. When the page elements are loaded, the user name input box on the login page is automatically entered into the user name, the password input box on the login page is automatically entered into the password, the consent check box on the login page is automatically checked, and finally the login button is identified to complete the cloud platform system login.

[0129] Capture the homepage screen: The RPA engine is invoked to automatically identify page elements and determine whether the page has loaded by checking whether the elements are present. Once the page elements have loaded, a screenshot is captured and saved in the inspection directory. The screenshot should include information such as the overall status of cluster resource scheduling and load balancing, physical machine and storage status (CPU, physical memory, and a detailed list of configured memory).

[0130] Capturing application delivery screen information: This system invokes the RPA engine, automatically inputs the application delivery page URL, opens the application delivery page, and automatically identifies page elements. It then determines whether the page has loaded successfully by checking whether the elements exist. Once the elements have loaded, it captures the application delivery screen information and saves the screenshot to the inspection directory. The screenshot should include information such as a device status list, session trend chart, connection status list, and network status trend chart.

[0131] Capture web management screen information: The RPA engine is invoked, automatically inputting the web management page URL, opening the web management page, and automatically identifying page elements. The page load is determined by determining whether the elements are present. Once the elements are loaded, the web management screen is captured and saved in the inspection directory. The screenshot should include: a chart showing the overall application load; device operation information including CPU, memory, power supply, temperature, fan speed; a line chart showing system throughput, a line chart showing the number of new connections, and a line chart showing concurrent connections.

[0132] Capture firewall screen information: This calls the RPA engine, automatically enters the firewall page URL, opens the firewall page, automatically identifies page elements, and determines whether the page has loaded by checking whether the elements exist. Once the page elements have loaded, the firewall screen information is captured and saved in the inspection directory. The screenshot information should include: device status list, session trend chart, connection status list, network status trend chart, and other information.

[0133] 9. Inspection of IT operation and maintenance monitoring platform:

[0134] Use the default browser to open the IT operation and maintenance monitoring platform: Call the RPA engine, open the system default browser, and automatically enter the IT operation and maintenance monitoring platform URL;

[0135] Resolve certificate errors: Use the "thisisunsafe" command to ignore certificate errors, skip the "There is a problem with this website's security certificate" page, and enter the IT operation and maintenance monitoring platform login page;

[0136] IT operation and maintenance monitoring platform system login: Call the RPA engine, browse the browser plug-in, automatically identify page elements, and determine whether the page is loaded by judging whether the page elements exist. When the page elements are loaded, identify the username input box on the login page and automatically enter the username, identify the password input box on the login page and automatically enter the password, identify the consent selection box on the login page and automatically check it, and finally identify the login button to complete the cloud platform system login.

[0137] Capture dashboard screen information: The RPA engine is invoked to automatically identify page elements and determine whether the page has loaded by checking whether the elements exist. Once the page elements have loaded, a screenshot of the dashboard screen is captured and saved in the inspection directory. The screenshot should include summary charts, SLA status, monitor status, alarm status, and a device status list.

[0138] Capture core switch screen information: The RPA engine is invoked to automatically enter the core switch page URL, open the core switch page, and automatically identify page elements. The page load is determined by checking whether the elements exist. Once the page elements are loaded, the core switch screen information is captured and saved in the inspection directory. The screenshot should include information such as the core switch name, network port, memory pool, IP address, CPU, and file list.

[0139] Capture the main switch screen: This calls the RPA engine, automatically enters the main switch page URL, opens the main switch page, automatically identifies page elements, and determines whether the page has loaded by checking whether the elements are present. Once the page elements have loaded, the main switch screen is captured and saved in the inspection directory. The screenshot should include information such as the main switch name, main switch location, network ports, fans, stacking modules, module humidity, power supply, IP address, CPU, and memory.

[0140] Through the implementation of the above solution, automated inspection tasks can save a lot of time compared to manual work, reduce the probability of missed inspections and false inspections caused by human negligence, and ensure that subtle system anomalies can be discovered in a timely manner, such as occasional error messages hidden in system logs; real-time warnings provide operation and maintenance personnel with comprehensive insights into the health status of the system, helping to accurately locate the root cause of the problem, formulate optimization strategies, and ensure business continuity of the enterprise.

[0141] The above shows and describes the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The above embodiments and descriptions are merely preferred examples of the present invention and are not intended to limit the present invention. Various changes and improvements may be made to the present invention without departing from the spirit and scope of the present invention. Such changes and improvements fall within the scope of the present invention. The scope of protection claimed in the present invention is defined by the appended claims and their equivalents.

Claims

1. The inspection method for multi-system operation and maintenance using an artificial intelligence-based RPA inspection robot is characterized by: The inspection method includes the following steps: The RPA inspection robot establishes connections with multiple target systems that need to be inspected through a pre-set interface protocol, and automatically identifies the login information of the login system page based on the process automation RPA software, including: loading and reading the RPA tool software package and its corresponding development toolkit, integrating the RPA tool and inspection rules; starting the RAP engine, initializing the engine parameters and initializing the running instructions, where the engine parameters at least include the thread pool size, timeout setting and retry count threshold; clarifying the hierarchical structure of the inspection directory, parsing the storage rules, and determining the storage format of different types of data, including at least storing performance indicator data in CSV format in time series and storing log files in original text form in date-named folders; calling the RPA engine The engine passes the inspection system URL pre-stored in the configuration library to the browser startup parameters, so that the browser automatically navigates to the login page of the target inspection system after startup; when the browser encounters a certificate error prompt when loading the inspection system login page, it ignores the web page certificate error through the thisisunsafe command to enter the inspection login page, drives the RPA engine, automatically identifies the page elements, and determines whether the page is loaded by judging whether the page elements exist. It automatically enters the user name in the username input box of the login page, automatically enters the password in the password input box of the login page, and automatically ticks the consent selection box on the login page, and then identifies the login button to complete the system login; wherein the login information at least includes the login user name, password and verification code information; Automatically extract the verification code information from the login information and call the verification code recognition service interface to perform verification code recognition; After the verification code is recognized, the RPA robot automatically logs into each target system that needs to be inspected. Based on the operation and maintenance objectives of each system, it obtains information including at least the target page and the system operation status page, and stores the collected page information in the RPA robot's inspection directory. The target systems that need to be inspected include at least the full network behavior management system, operation and maintenance audit and risk control system, network controller system, enterprise-level data backup and recovery system, Internet behavior management system, application delivery management system, firewall system, cloud platform system, and IT operation and maintenance monitoring platform. According to the inspection catalog, the information collected from each target system that needs to be inspected is unified and aggregated, and anomaly detection is performed, and the alarm information of the target system page alarm module is analyzed; The alarm information page is captured and stored in the RPA robot, which then sends the alarm information to the operation and maintenance personnel.

2. The inspection method for performing multi-system operation and maintenance using an artificial intelligence-based RPA inspection robot according to claim 1, characterized in that: The automatic extraction of the verification code information in the login information and calling the verification code recognition service interface to perform verification code recognition include: Start the RPA browser plug-in runtime environment, load the pre-configured page element recognition rule set, traverse the HTML structure of the page using DOM parsing technology, locate the node where the verification code information element is located, and identify the verification code code; Send an HTTPS request to the server and monitor the response status of the request in real time. If a successful response is received, obtain a call token and use the call token to read the image recognition API access information stored in the local configuration file; Extract the returned JSON value, parse the JSON value to obtain the recognition result, pass the recognition result to the system login, automatically fill in the verification code input box during the login process, and complete the entire verification code processing flow.

3. The inspection method for performing multi-system operation and maintenance using an artificial intelligence-based RPA inspection robot according to claim 1, characterized in that: After the verification code is recognized, the RPA robot automatically logs in to each target system that needs to be inspected. Based on the operation and maintenance goals of each system, it obtains information including at least the target page and the system operation status page, and stores the collected page information in the RPA robot's inspection directory, including: After the verification code is successfully recognized and automatically filled in, the RPA robot uses the previously obtained login account and password information to simulate a manual click of the login button and send a login request to the target system server. Continuously monitor the response information returned by the server and confirm whether the login is successful by analyzing the response status code, page jump indicator, and login success prompt element. If a successful login signal is not detected within 10 seconds, the RPA robot will re-initiate the login process according to the preset retry mechanism until it successfully logs into each target system that needs to be inspected. After successful login, the RPA robot starts the built-in page collection module, allocates memory space as a temporary data buffer, and reads the page collection rule file for the target system. The page collection rules include at least the collection priority, collection frequency, and data storage format for different page elements. Initialize the parameters of the page collection module according to the page collection rules. The RPA robot uses the built-in DOM parsing engine to deeply analyze the target page. By identifying the HTML tag structure, CSS style class names, and JavaScript event triggering areas of the page, it locates the page elements containing key information. The RPA robot automatically extracts the data content from the page elements where key information is located and pre-processes the data content; Based on the URL address or page redirection configuration of the system operation status page, the RPA robot is driven to jump from the current page to the system operation status page. After the system operation status page is loaded, the RPA robot mines the detailed system operation status data on the page. Integrate the data content of the page elements containing key information temporarily stored in the memory cache and the detailed data on system operation status. Use the RPA robot to store the integrated data in the inspection directory according to the storage path.

4. The inspection method for performing multi-system operation and maintenance using an artificial intelligence-based RPA inspection robot according to claim 3 is characterized in that: The RPA robot mining page contains detailed system operation status data, including: Use underlying network request sniffing technology to capture the AJAX requests that interact with the server behind the page, parse the returned data, and obtain system performance indicators pushed in real time by the server. System performance indicators include at least CPU usage, memory usage, and network bandwidth utilization. For system operation trend data displayed in the form of visual charts on the page, we combine OCR technology with chart parsing algorithms to convert the data points in the chart into a structured numerical sequence; The collected system operation status data is temporarily stored in the memory cache area and preliminarily sorted according to the preset data classification rules, wherein the preset data classification rules at least include data type classification, timestamp classification and data source classification.

5. The inspection method for performing multi-system operation and maintenance using an artificial intelligence-based RPA inspection robot according to claim 1, characterized in that: According to the inspection catalog, the information collected from each target system that needs to be inspected is unified and aggregated, and anomaly detection is performed. The alarm information of the page alarm module is analyzed, including: Applying data mining algorithms to extract key features from the processed data, inputting the key features into an anomaly detection model to perform anomaly detection, and outputting the detection results; When an exception occurs, obtain the alarm information issued by the target system page alarm module, wherein the alarm information at least includes an alarm code, an alarm level and an alarm description; Natural language processing technology is used to pre-process the alarm description, extract key semantic information, and locate the problem by combining the alarm code and alarm level.

6. The inspection method for performing multi-system operation and maintenance using an artificial intelligence-based RPA inspection robot according to claim 5, characterized in that: The processing data is subjected to a data mining algorithm to extract key features, including: Calculate the Pearson correlation coefficient between each feature in the processed data and the system's target operation and maintenance status, and obtain features whose absolute value of the correlation is higher than the set threshold, where the threshold is 0.3; Variance analysis was used to test the mean differences of each feature under different categories of operation and maintenance status, and the features under different system health conditions were screened out to obtain highly correlated feature data; Highly correlated feature data is used as input and imported into the LDA model framework. The target dimension of the LDA model is set according to the actual number of system operation and maintenance categories. The high-dimensional feature space is projected into a low-dimensional discriminant subspace, maximizing the inter-class distances of different operation and maintenance status categories within the subspace while minimizing the intra-class variance. The system operation and maintenance categories include at least normal operation, minor failure, and major failure. By iteratively optimizing the projection matrix, samples with different operation and maintenance status are divided in the new feature space after projection, and key features are extracted.

7. The inspection method for performing multi-system operation and maintenance using an artificial intelligence-based RPA inspection robot according to claim 5, characterized in that: The key features are input into the anomaly detection model to perform anomaly detection, and the detection results are output, including: A bidirectional RNN is used as the basic model of the anomaly detection model. The key features are input into the anomaly detection model. In the forward propagation phase of the model, starting from the starting time step of the sequence, the forward RNN unit reads the key feature vectors one by one in sequence, and calculates the hidden layer output at the current moment through the activation function based on the current input features and the output state of the hidden layer at the previous moment; The reverse RNN unit starts from the last time step of the sequence, reads the feature vector in reverse, and calculates the output based on the input and the reverse hidden layer state at the previous moment. It transmits the key information at the end of the sequence in reverse and merges it with the forward information. The detection results are output through the anomaly detection model and transmitted to the RPA inspection robot in real time through the communication protocol.

Images