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

Abstract

The invention relates to the field of IT system operation and maintenance management, and discloses a multi-system operation and maintenance inspection method of an RPA inspection robot based on artificial intelligence, the RPA inspection robot establishes connection with a plurality of target systems needing inspection through a preset interface protocol, and automatically identifies login information of a login system page based on process automation RPA software; automatically extracting verification code information in the login information, and calling a verification code identification service interface to carry out verification code identification; after the verification code passes identification, the RPA robot automatically logs in each target system needing inspection, obtains page information, and stores the collected page information into an inspection directory of the RPA robot; according to the routing inspection catalog, information collected by each target system needing routing inspection is subjected to unified collection processing and anomaly detection, and alarm information of a page alarm module is analyzed; intercepting an alarm information page, and sending alarm information to operation and maintenance personnel by the RPA robot; the working efficiency is improved.

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 performing multi-system operation and maintenance by an artificial intelligence (AI)-based RPA inspection robot. Background Art

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

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

[0004] 3. Monitoring and troubleshooting: With the increasing complexity of IT systems, monitoring becomes more difficult. It is necessary to monitor the system operation status in real time and detect potential problems in a timely manner. However, due to the diversity of data sources and high real-time requirements, monitoring becomes extremely difficult. Troubleshooting is also complicated due to the complex system architecture and numerous components, which requires a lot of time and effort.

[0005] 4. Insufficient automation and standardization: The operation and maintenance work of some enterprises still relies on manual operations, which leads to low work efficiency and prone to errors. The application of automated operation and maintenance tools can significantly improve the efficiency of operation and maintenance and reduce the risk of human errors, but the current lack of automation is a pain point. In addition, the lack of unified standards leads to a lack of standardization in operation and maintenance work.

[0006] 5. Resource limitations and pressure: When facing complex operation and maintenance tasks, the operation and maintenance team is often affected by resource limitations, including human resources, technical resources, and financial resources. Insufficient resources will make it difficult to carry out operation and maintenance work smoothly, and even affect the normal operation of the business system; at the same time, the operation and maintenance team bears the important responsibility of ensuring the stable operation of the business system, and the huge responsibility pressure and psychological pressure make the operation and maintenance work more difficult.

[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 the present invention is to solve the above problems and to design an inspection method for multi-system operation and maintenance by an RPA inspection robot based on artificial intelligence.

[0009] The present invention provides an inspection method for performing multi-system operation and maintenance by an artificial intelligence-based RPA inspection robot, the inspection method comprising 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, where 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 in to each target system that needs to be inspected, obtains at least the target page and system operation status page information according to the operation and maintenance goals of each system, 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 whole 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 collected for processing, and anomaly detection is performed, and the alarm information of the 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.

[0010] 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 preset interface protocol, and automatically identifies login information of a login system page based on the process automation RPA software, including: Load and read the RPA tool software package and its corresponding development tool package, and integrate the RPA tool and inspection rules; Start the RAP engine, initialize engine parameters and initialize running instructions, wherein the engine parameters at least include thread pool size, timeout setting and retry count threshold; Clarify the hierarchical structure of the inspection directory, analyze the storage rules, and determine 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; Call the RPA engine and pass the inspection system URL pre-stored in the configuration library to the browser startup parameter, so that the browser automatically navigates to the login page of the target inspection system after startup; When the browser encounters a certificate error when loading the inspection system login page, you can use the thisisunsafe command to ignore the web certificate error and enter the inspection login page; 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, identify the username input box on the login page to automatically enter the username, identify the password input box on the login page to automatically enter the password, identify the consent selection box on the login page to automatically check the box, and then identify the login button to complete the system login.

[0011] Optionally, in a second implementation of the present invention, the automatically extracting verification code information from the login information and calling a verification code recognition service interface to perform verification code recognition includes: Start the RPA browser plug-in operating environment, load the pre-configured page element recognition rule set, traverse the HTML structure of the page through 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 get 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.

[0012] 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 at least target page and system operation status page information according to the operation and maintenance objectives of each system, and stores the collected page information in the inspection directory of the RPA robot, including: 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; Continuously monitor the response information returned by the server, and confirm whether the login is successful by parsing the response status code, page jump mark, and login success prompt element. If no successful login signal is 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 collection module of the RPA robot is started, and memory space is allocated as a temporary data cache area. The page collection rule file for the target system is read. The page collection rule includes at least the collection priority, collection frequency, and data storage format of different page elements. Initialize the parameters of the page collection module according to the page collection rules, and use the built-in DOM parsing engine of the RPA robot to deeply analyze the target page. By identifying the HTML tag structure, CSS style class name, and JavaScript event triggering area of ​​the page, locate the page elements where the key information is located; The RPA robot automatically extracts the data content in the page elements where the key information is located and pre-processes the data content; According to 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. Integrate the data content in the page elements where the key information temporarily stored in the memory cache is located and the detailed data on the system operation status. Use the RPA robot to store the integrated data into the inspection directory according to the storage path.

[0013] Optionally, in a fourth implementation of the present invention, the RPA robot mines detailed data of system operation status in the page, including: Use the underlying network request sniffing technology to capture the AJAX requests that interact with the server behind the page, parse the returned data, and obtain the system performance indicators pushed in real time by the server. The system performance indicators include at least CPU usage, memory usage, 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 the preset data classification rules, wherein the preset data classification rules at least include data type classification, timestamp classification and data source classification.

[0014] Optionally, in a fifth implementation of the present invention, according to the inspection catalog, the information collected from each target system that needs to be inspected is uniformly collected and processed, and abnormality detection is performed, and 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.

[0015] Optionally, in a sixth implementation of the present invention, the extracting key features from the processed data using a data mining algorithm includes: Calculate the Pearson correlation between each feature in the processed data and the system target operation and maintenance status, and obtain the features whose absolute value of 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 types of operation and maintenance status, and the features under different system health conditions were screened out to obtain highly correlated feature data; The 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 categories of system operation and maintenance. The high-dimensional feature space is projected into a 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. The categories of system operation and maintenance include at least normal operation, minor faults, and major faults. 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.

[0016] Optionally, in a seventh implementation of the present invention, the key features are input into an 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, and 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 according to 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, 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 transmitted to the RPA inspection robot in real time through the communication protocol.

[0017] In the technical solution provided by the present invention, the RPA automated robot logs in to different systems of the enterprise according to the enterprise's own situation, automatically captures the pages that focus on and reflect the system operation status, and uniformly collects, processes and stores them for review by operation and maintenance personnel, so as to replace the operation and maintenance personnel to manually log in to multiple systems at regular intervals and points to perform inspection operations; use automated inspection tools to improve work efficiency and 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

[0018] Various other advantages and benefits will become apparent to those of ordinary skill in the art by reading the following detailed description of the preferred embodiment.The drawings are only for the purpose of illustrating the preferred embodiments and are not to be construed as limiting the invention.

[0019] Figure 1 A schematic diagram of an embodiment of an inspection method for performing multi-system operation and maintenance by an artificial intelligence-based RPA inspection robot provided in an embodiment of the present invention; Figure 2 A schematic diagram of the login steps of the inspection method for performing multi-system operation and maintenance by an artificial intelligence-based RPA inspection robot provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0020] The terms "first", "second", "third", "fourth", etc. (if any) in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments described herein can be implemented in an order other than that illustrated 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, device, product or device that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0021] 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 an inspection method for performing multi-system operation and maintenance by an artificial intelligence-based RPA inspection robot. The method specifically includes the following steps: 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 login system page based on the process automation RPA software; In this embodiment, the login information includes at least the login user name, password and verification code information; In this embodiment, the RPA tool software package and its corresponding development toolkit are loaded and read, and the RPA tool and the inspection rules are integrated; the RAP engine is started, the engine parameters and the initialization running instructions are initialized, wherein the engine parameters at least include the thread pool size, the timeout setting and the retry number 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, at least including the performance indicator data being stored in CSV format in time series and the log files being stored in the original text form in a folder named after 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 when loading the inspection system login page, the web page certificate error is ignored through the thisisunsafe command to enter the inspection login page; the RPA engine is driven to automatically identify the page elements, and 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.

[0022] Step 102: Automatically extract the verification code information in the login information, and call the verification code recognition service interface to perform verification code recognition; 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 the 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, and the verification code input box is automatically filled in during the login process to complete the entire verification code processing flow.

[0023] Step 103: After the verification code is recognized, the RPA robot automatically logs in to each target system that needs to be inspected, obtains at least target page and system operation status page information according to the operation and maintenance objectives of each system, and stores the collected page information in the inspection directory of the RPA robot; 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.

[0024] 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 parameters of the collection module are used to perform in-depth analysis of the target page through the built-in DOM parsing engine of the RPA robot. By identifying the HTML tag structure, CSS style class name and JavaScript event triggering area of ​​the page, the page element where the key information is located 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; according to 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 in the page element where the key information is temporarily stored in the memory cache and the detailed system operation status data are integrated, and according to the storage path, the RPA robot is used to store the integrated data into the inspection directory.

[0025] 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 the system performance indicators pushed in real time by the server, where the system performance indicators at least include CPU usage, memory occupancy and network bandwidth utilization; for the system operation trend data displayed in the form of a visual chart on the page, the OCR technology and the 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 the preset data classification rules, where the preset data classification rules at least include data type classification, timestamp classification and data source classification.

[0026] Step 104: According to the inspection catalog, the information collected from each target system that needs to be inspected is unified and collected for processing, and abnormality detection is performed, and the alarm information of the page alarm module is analyzed; 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, wherein the alarm information includes at least 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 in combination with the alarm code and alarm level.

[0027] In this embodiment, the Pearson correlation between each feature in the processed data and the target operation and maintenance status of the system is calculated to obtain the number of features whose absolute value of 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 the 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, wherein 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.

[0028] 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 according to 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 according to 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.

[0029] 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.

[0030] Another embodiment of the inspection method for performing multi-system operation and maintenance by an artificial intelligence-based RPA inspection robot includes the following steps: System login: Automatically identify the login information of the system login page based on process automation software, including login user name, password and verification code information; 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; System inspection: Log in to each inspection system, focus on the operation and maintenance of each system, intercept the page of interest, the page information reflecting the system operation status, and store it in the designated inspection directory; Data collection: unified processing of inspection data collected from multiple different systems; Alarm analysis: After logging into the system, the alarm information of the page alarm module is automatically analyzed, the alarm information page is captured, stored in the designated inspection directory, and the pictures are stored according to the screenshot naming rules, so that the operation and maintenance personnel can confirm the system alarm information through the file name.

[0031] See also Figure 2 As shown, a schematic diagram of the login steps of an inspection method for performing multi-system operation and maintenance by an artificial intelligence-based RPA inspection robot includes the following steps: Integrate and package the RPA engine: Integrate the RPA tool and inspection rules to automate the inspection logic process; Initialize the RPA engine: Start the RAP engine and initialize engine parameters and methods; Initialize inspection system parameters: define inspection directory and storage rules; RPA calls the browser to open the inspection website: calls the RPA engine, opens the system default browser, and automatically enters the VPN system website; Solve the certificate error: Use the "thisisunsafe" command to ignore the certificate error, skip the "There is a problem with this website's security certificate" page, and enter the VPN login page RPA calls the browser to check and log in to the system: open the inspection website: call the RPA engine, automatically identify the page elements through the browser plug-in, and determine whether the page has been loaded 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.

[0032] According to the different information and key points of different systems, the methods and steps for each system to conduct inspections are as follows: 1. Inspection of the entire network behavior management system: Use the default browser to open the full-network behavior management system: call the RPA engine, open the system default browser, and automatically enter the full-network behavior management system URL; Solve the certificate error: Use the "thisisunsafe" command to ignore the certificate error, skip the "There is a problem with the security certificate of this website" page, and enter the login page of the full network behavior management system; Login to the entire network behavior management system: Call the RPA engine, automatically identify page elements through browser plug-ins, and 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 to automatically enter the username, identify the password input box on the login page to automatically enter the password, identify the consent selection box on the login page to automatically check it, and then identify the login button to complete the login to the entire network behavior management system.

[0033] Capture the homepage screen information: Call the RPA engine 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, capture the homepage screen information and save the screenshot image in the inspection directory. The screenshot information should include: device status icon information, management status information, traffic analysis-interface throughput line chart, user traffic ranking, application traffic ranking, and other information reflecting network behavior.

[0034] 2. Operation and maintenance audit and risk control system inspection: 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 URL of the operation and maintenance audit and risk control system; Solve the certificate error: Use the "thisisunsafe" command to ignore the certificate error, skip the "There is a problem with the security certificate of this website" page, and enter the operation and maintenance audit and risk control system login page; Login to the entire network behavior management system: call the RPA engine, automatically identify page elements through browser plug-ins, and 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, automatically fill the verification code into the verification code input box, and finally identify the login button to complete the operation and maintenance audit and risk control system login.

[0035] Capture the main interface screen information: Call the RPA engine 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, capture the main interface screen information and save the screenshot image in the inspection directory. The screenshot information should include: operation and maintenance statistics chart, operation and maintenance statistics, real-time monitoring information, system operation status, license and other information.

[0036] Capture network configuration screen information: Call the RPA engine, automatically enter the network configuration page URL, open the network configuration page, automatically identify page elements, and determine whether the page is loaded by judging whether the page elements exist. When the page elements are loaded, capture the network configuration screen information and save the screenshot image in the inspection directory. The screenshot information should include: detailed interface information list and other information.

[0037] 3. NAC network controller system inspection: 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; Solve the certificate error: Use the "thisisunsafe" command to ignore the certificate error, skip the "There is a problem with the security certificate of this website" page, and enter the operation and maintenance audit and risk control system login page; Network controller system login: 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 user name input box on the login page and automatically enter the user name, 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 network controller system login.

[0038] Capture the main interface screen information: Call the RPA engine 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, capture the main interface screen information and save the screenshot image in the inspection directory. The screenshot information should include: device resources, system status, network port status, access point status, access point status, total throughput flow chart and other information.

[0039] Capture the address pool status screen information: Call the RPA engine, automatically enter the address pool status page URL, open the address pool status page, automatically identify the page elements, and determine whether the page is loaded by judging whether the page elements exist. When the page elements are loaded, capture the address pool status screen information and save the screenshot image in the inspection directory. The screenshot information should include: detailed wireless address pool status chart and other information.

[0040] 4. Enterprise-level data backup and recovery system inspection: Use the default browser to open the enterprise-level data backup and recovery system: Call the RPA engine, open the system default browser, and automatically enter the URL of the enterprise-level data backup and recovery system; Solve the certificate error: Use the "thisisunsafe" command to ignore the certificate error, skip the "There is a problem with the security certificate of this website" page, and enter the enterprise-level data backup and recovery system login page; Enterprise-level data backup and recovery system login: Call the RPA engine, browse the browser plug-in, automatically identify the 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 enterprise-level data backup and recovery system login.

[0041] Capture the main interface screen information: Call the RPA engine 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, capture the main interface screen information and save the screenshot image in the inspection directory. The screenshot information should include: system time, cumulative running time, cumulative protection data, virtualization center, backup storage, backup node chart, network traffic chart, current task list, current task list and other information.

[0042] 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 "alarm" file name, they log in to the operation and maintenance system to check the system alarm.

[0043] 5. VPN Internet Behavior Management System Inspection: Use the default browser to open the Internet behavior management system: call the RPA engine, open the system default browser, and automatically enter the URL of the Internet behavior management system; Solve the certificate error: Use the "thisisunsafe" command to ignore the certificate error, skip the "There is a problem with the security certificate of this website" page, and enter the Internet behavior management system login page; Login to the Internet Behavior Management System: Call the RPA engine, browse the browser plug-in, automatically identify the 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 login to the Internet Behavior Management System.

[0044] Capture the main interface screen information: call the RPA engine to automatically identify page elements, determine whether the page is loaded by judging whether the page elements exist, and when the page elements are loaded, capture the main interface screen information and save the screenshot image in the inspection directory. The screenshot information should include: system information line chart, real-time network throughput line chart, real-time concurrent session line chart, line status list, real-time user concurrent trend line chart, real-time stream cache status line chart, and other information.

[0045] 6. Application delivery management system inspection: Use the default browser to open the application delivery management system: call the RPA engine, open the system default browser, and automatically enter the application delivery management system URL; Solve the certificate error: Use the "thisisunsafe" command to ignore the certificate error, skip the "There is a problem with the security certificate of this website" page, and enter the application delivery management system login page; Application delivery management system login: Call the RPA engine, browse the browser plug-in, automatically identify the page elements, and determine whether the page is loaded by judging whether the page elements exist. When the page elements are loaded, identify the login page user name input box to automatically enter the user name, identify the login page password input box to automatically enter the password, identify the login page consent selection box to automatically check, and finally identify the login button to complete the application delivery management system login.

[0046] Capture application load screen information: Call the RPA engine to automatically identify page elements, determine whether the page is loaded by judging whether the page elements exist, and when the page elements are loaded, capture the application load screen information and save the screenshot image in the inspection directory. The screenshot information should include: application load overall situation chart, device operation information - CPU, memory, power supply, temperature, fan, system throughput line chart, system new connection number line chart, system concurrent connection line chart and other information.

[0047] Capture link load screen information: Call the RPA engine, automatically enter the link load page URL, open the link load page, automatically identify page elements, and determine whether the page is loaded by judging whether the page elements exist. When the page elements are loaded, capture the link load screen information and save the screenshot image in the inspection directory. The screenshot information should include: link status, link load real-time status line chart, stability monitoring line chart, application traffic share analysis bar chart, system throughput line chart, system new connection number line chart, system concurrent connection line chart, and other information.

[0048] 7. Firewall system inspection: 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; Solve the certificate error: Use the "thisisunsafe" command to ignore the certificate error, skip the "There is a problem with the security certificate of this website" page, and enter the firewall system login page; Firewall system login: Call the RPA engine, browse the browser plug-in, automatically identify the 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 firewall system login.

[0049] Capture the homepage screen information: Call the RPA engine to automatically identify page elements, determine whether the page is loaded by judging whether the page elements exist, and when the page elements are loaded, capture the homepage screen information and save the screenshots in the inspection directory. The screenshot information should include: business security-risk analysis, attack trends, hot event chart information, user security-risk analysis, attack trends, hot event chart information, equipment and system operations-equipment status, interface status, security capability information, network operations-concurrent sessions, interface throughput and other information.

[0050] Capture network interface screen information: Call the RPA engine, automatically enter the network interface page URL, open the network interface page, automatically identify page elements, and determine whether the page is loaded by judging whether the page elements exist. When the page elements are loaded, capture the network interface screen information and save the screenshot image in the inspection directory. The screenshot information should include: physical interface details, status list and other information.

[0051] 8. Cloud platform system inspection: 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; Solve the certificate error: Use the "thisisunsafe" command to ignore the certificate error, skip the "There is a problem with the security certificate of this website" page, and enter the cloud platform system login page; Cloud platform system login: Call the RPA engine, browse the browser plug-in, automatically identify the 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.

[0052] Capture the homepage screen information: Call the RPA engine 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, capture the homepage screen information and save the screenshot image in the inspection directory. The screenshot information should include: the overall situation of cluster resource scheduling and load balancing, the physical machine and storage status - CPU, physical memory, and a detailed list of configured memory.

[0053] Capture application delivery screen information: Call the RPA engine, automatically enter the application delivery page URL, open the application delivery page, automatically identify page elements, and determine whether the page is loaded by judging whether the page elements exist. When the page elements are loaded, capture the application delivery screen information and save the screenshot image 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.

[0054] Capture the web management screen information: call the RPA engine, automatically enter the web management page URL, open the web management page, automatically identify the page elements, and determine whether the page is loaded by judging whether the page elements exist. When the page elements are loaded, capture the web management screen information and save the screenshot image in the inspection directory. The screenshot information should include: the overall application load chart, device operation information - CPU, memory, power supply, temperature, fan, system throughput line chart, system new connection number line chart, system concurrent connection line chart and other information.

[0055] Capture firewall screen information: call the RPA engine, automatically enter the firewall page URL, open the firewall page, automatically identify page elements, and determine whether the page is loaded by judging whether the page elements exist. When the page elements are loaded, capture the firewall screen information and save the screenshot image in the inspection directory. The screenshot information should include: device status list, session trend chart, connection status list, network status trend chart, etc. 9. Inspection of IT operation and maintenance monitoring platform: 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 URL of the IT operation and maintenance monitoring platform; Solve the certificate error: Use the "thisisunsafe" command to ignore the certificate error, skip the "There is a problem with the security certificate of this website" page, and enter the IT operation and maintenance monitoring platform login page; IT operation and maintenance monitoring platform system login: call the RPA engine, browse the browser plug-in, automatically identify the 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.

[0056] Capture the instrument panel screen information: Call the RPA engine 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, capture the instrument panel screen information and save the screenshot image in the inspection directory. The screenshot information should include: summary information charts - SLA status, monitor status, alarm status, device status list and other information.

[0057] Capture the core switch screen information: call the RPA engine, automatically enter the core switch page URL, open the core switch page, automatically identify the page elements, and determine whether the page is loaded by judging whether the page elements exist. When the page elements are loaded, capture the core switch screen information and save the screenshot image in the inspection directory. The screenshot information should include: core switch name, network port, memory pool, IP address, CPU, file list and other information.

[0058] Capture the main switch screen information: call the RPA engine, automatically enter the main switch page URL, open the main switch page, automatically identify the page elements, and determine whether the page is loaded by judging whether the page elements exist. When the page elements are loaded, capture the main switch screen information and save the screenshot image in the inspection directory. The screenshot information should include: main switch name, main switch location, network port, fan, stacking module, module humidity, power supply, IP address, CPU, memory and other list information.

[0059] 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 detection and false detection caused by human negligence, and ensure timely detection of subtle system anomalies, such as occasional error information 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.

[0060] 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 by the above embodiments. The above embodiments and descriptions are only preferred examples of the present invention and are not intended to limit the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. An inspection method for multi-system operation and maintenance using an artificial intelligence-based RPA inspection robot, characterized in that: 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, where 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 in to each target system that needs to be inspected, obtains at least the target page and system operation status page information according to the operation and maintenance goals of each system, 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 whole 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 collected for processing, and anomaly detection is performed, and the alarm information of the 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 by an artificial intelligence-based RPA inspection robot according to claim 1, characterized in that: 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 login system page based on the process automation RPA software, including: Load and read the RPA tool software package and its corresponding development tool package, and integrate the RPA tool and inspection rules; Start the RAP engine, initialize engine parameters and initialize running instructions, wherein the engine parameters at least include thread pool size, timeout setting and retry count threshold; Clarify the hierarchical structure of the inspection directory, analyze the storage rules, and determine 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; Call the RPA engine and pass the inspection system URL pre-stored in the configuration library to the browser startup parameter, so that the browser automatically navigates to the login page of the target inspection system after startup; When the browser encounters a certificate error when loading the inspection system login page, you can use the thisisunsafe command to ignore the web certificate error and enter the inspection login page; 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, identify the username input box on the login page to automatically enter the username, identify the password input box on the login page to automatically enter the password, identify the consent selection box on the login page to automatically check the box, and then identify the login button to complete the system login.

3. The inspection method for performing multi-system operation and maintenance by an artificial intelligence-based RPA inspection robot according to claim 1, characterized in that: The automatic extraction of 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 operating environment, load the pre-configured page element recognition rule set, traverse the HTML structure of the page through 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 get 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.

4. The inspection method for performing multi-system operation and maintenance by 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, obtains at least target page and system operation status page information according to the operation and maintenance objectives of each system, 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 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; Continuously monitor the response information returned by the server, and confirm whether the login is successful by parsing the response status code, page jump mark, and login success prompt element. If no successful login signal is 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 collection module of the RPA robot is started, and memory space is allocated as a temporary data cache area. The page collection rule file for the target system is read. The page collection rule includes at least the collection priority, collection frequency, and data storage format of different page elements. Initialize the parameters of the page collection module according to the page collection rules, and use the built-in DOM parsing engine of the RPA robot to deeply analyze the target page. By identifying the HTML tag structure, CSS style class name, and JavaScript event triggering area of ​​the page, locate the page elements where the key information is located; The RPA robot automatically extracts the data content in the page elements where the key information is located and pre-processes the data content; According to 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. Integrate the data content in the page elements where the key information temporarily stored in the memory cache is located and the detailed data on the system operation status. Use the RPA robot to store the integrated data into the inspection directory according to the storage path.

5. The inspection method for performing multi-system operation and maintenance by an artificial intelligence-based RPA inspection robot according to claim 4, characterized in that: The RPA robot mines detailed data on the system operation status on the page, including: Use the underlying network request sniffing technology to capture the AJAX requests that interact with the server behind the page, parse the returned data, and obtain the system performance indicators pushed in real time by the server. The system performance indicators include at least CPU usage, memory usage, 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 the preset data classification rules, wherein the preset data classification rules at least include data type classification, timestamp classification and data source classification.

6. The inspection method for performing multi-system operation and maintenance by 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 collected for abnormal detection, and 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.

7. The inspection method for performing multi-system operation and maintenance by an artificial intelligence-based RPA inspection robot according to claim 1, characterized in that: The process of extracting key features from the processed data using a data mining algorithm includes: Calculate the Pearson correlation between each feature in the processed data and the system target operation and maintenance status, and obtain the features whose absolute value of 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; The 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 categories of system operation and maintenance. The high-dimensional feature space is projected into a 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. The categories of system operation and maintenance include at least normal operation, minor faults, and major faults. 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.

8. The inspection method for performing multi-system operation and maintenance by an artificial intelligence-based RPA inspection robot according to claim 1, 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, and 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 according to 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, 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 transmitted to the RPA inspection robot in real time through the communication protocol.

Images