Robot process automation method based on webpage inspection abnormity monitoring and fault processing

Abstract

According to the robot process automation method based on webpage inspection abnormity monitoring and fault processing, an overall inspection process compensation mechanism is constructed, a specified robot automatically monitors the execution condition of an RPA inspection process, and an idle robot server is coordinated in time in combination with an RPA task real-time scheduling plan and the resource use condition of the robot server; the failure process is automatically re-executed, the frequency of RPA process operation failure caused by objective factors such as basic environment problems and network fluctuation is avoided, and the operation and maintenance efficiency and the system resource utilization rate are remarkably improved.

Description

Technical Field

[0001] This invention relates to the field of anomaly monitoring and fault handling technology for web system inspection, and in particular to a robotic process automation method based on anomaly monitoring and fault handling for web system inspection. Background Technology

[0002] RPA, as an important automation tool for enterprise IT operations and maintenance, is widely used in periodic monitoring scenarios of web-based systems such as OA systems. Related technologies construct a monitoring system based on individual processes through the collaborative operation of detection point components, timestamp verification, and early warning mechanisms. Specifically, this technology covers the entire process from scenario branch configuration to anomaly triggering, including key aspects such as detection point status determination, setting the longest expected interval, and timer task generation. With the development of distributed system architecture, although existing solutions adopt a "one master, two backups" deployment mode, their core logic remains at the level of single-point failure handling, failing to form a collaborative compensation mechanism across server resources, resulting in insufficient system-level fault tolerance.

[0003] However, existing RPA inspection methods directly use independent detection point components for process monitoring without establishing a multi-dimensional data fusion decision model. This can lead to a surge in redundant alarm messages or resource scheduling imbalances. Specifically, when basic environmental issues such as network latency (e.g., instantaneous fluctuations in 24 / 7 service scenarios) or server lag occur, current technologies can only send independent alarm emails through a single robot, failing to incorporate task scheduling plans and server resource status into the compensation decision-making system. Redundant alarm messages can account for as much as 60%, and the lack of dynamic compensation strategies leads to a 40% increase in resource contention and a 35% decrease in system resource utilization. This disconnect between local monitoring and global resource scheduling not only increases the workload of log analysis for operations personnel but also can trigger system-level cascading failures due to blind retries, severely restricting the large-scale application of RPA in complex business environments. Summary of the Invention

[0004] The present invention aims to at least partially solve one of the technical problems in the related art.

[0005] Therefore, the first objective of this invention is to propose a robotic process automation method based on webpage inspection anomaly monitoring and fault handling.

[0006] The second objective of this invention is to propose a robotic process automation device based on webpage inspection anomaly monitoring and fault handling.

[0007] The third objective of this invention is to provide an electronic device.

[0008] The fourth objective of this invention is to provide a computer-readable storage medium.

[0009] The fifth objective of this invention is to provide a computer program product.

[0010] To achieve the above objectives, a first aspect of the present invention proposes a robotic process automation method based on webpage inspection anomaly monitoring and fault handling, comprising: S1 obtains the running result data, real-time task scheduling plan and server resource usage status of each business robot by using the primary and backup monitoring robots deployed on the operation and management platform to perform web page inspection processes; S2, based on the empty string determination rules of the global parameters in the running result data, analyze whether there are non-empty substrings to identify abnormal inspection processes, and locate the business robot corresponding to the abnormal process according to the identity information; S3, combining the real-time task scheduling plan and server resource usage status, dynamically schedule idle servers to perform compensatory execution for abnormal processes. The number of compensatory executions does not exceed a preset threshold and the interval between adjacent executions is configured according to business requirements. S4 summarizes and analyzes the compensation execution results and the original operation results data, sends the processing results to managers and business personnel through the email notification system, and stores the system operation data in the Dameng database to build an operation and maintenance knowledge base.

[0011] Optionally, S1 further includes: Agent_monitor is deployed using a one-master-two-standby architecture. The master server processes the detected running result data first, and when the master server load exceeds a preset threshold, it automatically switches to the standby server for data collection. A heartbeat detection mechanism is used to synchronize data between the primary and backup servers, ensuring the integrity of runtime results data, task scheduling plans, and resource status information during server switchover.

[0012] Optionally, S2 further includes: When a non-empty substring is detected, semantic analysis is performed on the error field according to the preset anomaly classification rules to distinguish between network fluctuation anomalies and system crash anomalies. If the anomaly is determined to be a network fluctuation and the server resource utilization is lower than the preset threshold, the alarm email will be delayed until the compensation is completed.

[0013] Optionally, S3 further includes: Based on the server resource usage status, a load balancing algorithm is used to select idle servers for compensation execution. The algorithm includes:

[0014] in For server The current load, For server Maximum resource capacity; Adjacent compensation execution interval Minimum interval constraint to meet business configuration:

[0015] in and These are the minimum and maximum interval times configured for business personnel, respectively.

[0016] Optional, also includes: An anomaly pattern recognition model is built based on system operation data stored in the Dameng database. By analyzing the distribution characteristics of error fields in historical inspection results data, compensation strategy optimization suggestions are automatically generated and the scheduling rule library of the operation and management platform is updated.

[0017] To achieve the above objectives, a second aspect of the present invention provides a robotic process automation device based on webpage inspection anomaly monitoring and fault handling, comprising: The primary and backup monitoring data acquisition module is used to acquire the running result data, real-time task scheduling plan and server resource usage status of each business robot executing the web inspection process through the primary and backup monitoring robots deployed on the operation and management platform; The anomaly identification and location module is used to analyze whether there are non-empty substrings based on the empty string judgment rules of the global parameters in the running result data to identify the abnormal inspection process, and locate the business robot corresponding to the abnormal process according to the identity information. The dynamic compensation scheduling module is used to dynamically schedule idle servers to perform compensation execution for abnormal processes by combining the real-time task scheduling plan and server resource usage status. The number of compensation executions does not exceed a preset threshold and the interval between adjacent executions is configured according to business requirements. The results processing and storage module is used to summarize and analyze the compensation execution results and the original operation results data, send the processing results to managers and business personnel through the email notification system, and store the system operation data in the Dameng database to build an operation and maintenance knowledge base.

[0018] To achieve the above objectives, a third aspect of the present invention provides an electronic device, comprising: a processor, and a memory communicatively connected to the processor; The memory stores computer-executed instructions; The processor executes computer execution instructions stored in the memory to implement the method as described in any one of the first aspects.

[0019] To achieve the above objectives, a fourth aspect of the present invention provides a computer-readable storage medium storing computer-executable instructions that, when executed by a processor, are used to implement the method as described in any one of the first aspects.

[0020] To achieve the above objectives, a fifth aspect of the present invention provides a computer program product that, when executed by a processor, implements the method described in any one of the first aspects.

[0021] The technical solutions provided by the embodiments of the present invention bring at least the following beneficial effects: they can realize global anomaly monitoring and intelligent compensation processing for web page inspection RPA processes, effectively reduce redundant alarms caused by basic environmental problems such as network fluctuations or server lag, and improve system resource utilization and inspection process success rate.

[0022] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0023] The above and / or additional aspects and advantages of the present invention will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, wherein: Figure 1 This is a flowchart illustrating a robotic process automation method based on webpage inspection anomaly monitoring and fault handling, provided in an embodiment of the present invention. Figure 2 This is a flowchart illustrating another robotic process automation method based on webpage inspection anomaly monitoring and fault handling provided in an embodiment of the present invention. Figure 3 This is a flowchart illustrating another robotic process automation method based on webpage inspection anomaly monitoring and fault handling provided in an embodiment of the present invention. Detailed Implementation

[0024] Embodiments of the present invention are described in detail below, examples of which are illustrated in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain the present invention, and should not be construed as limiting the present invention.

[0025] Figure 1 This is a flowchart illustrating a robotic process automation method based on webpage inspection anomaly monitoring and fault handling, provided in an embodiment of the present invention. (Refer to...) Figure 1 , Figure 2 and Figure 3 The method includes the following steps: S1 obtains the operational results data, real-time task scheduling plans, and server resource usage status of each business robot executing the webpage inspection process through a primary and backup monitoring robot deployed on the operation and management platform. Specifically, in some implementations, acquiring the operational results data, real-time task scheduling plans, and server resource usage status of each business robot executing the webpage inspection process through a primary / backup monitoring robot (Agent_monitor) deployed on the operation and management platform is a key step in realizing the overall inspection process compensation mechanism in this invention. This step is based on a distributed monitoring architecture and adopts a primary / backup redundant deployment method to ensure that the system can continue to operate stably under abnormal conditions such as high concurrency, server failure, or network fluctuations.

[0026] Specifically, the Agent_monitor robot is deployed on three application servers of the operation and management platform, adopting a "one master, two backups" architecture. The master node is responsible for real-time data collection and processing, while the backup nodes automatically take over the tasks when the master node fails or is overloaded. Uninterrupted service. The monitoring robot subscribes in real time to the execution result data of each business robot (Agent_1 to Agent_N) through API interfaces or message queues (such as Kafka, RabbitMQ), including key fields such as identity information, global parameters (returning an empty string on success and a specific error field on failure), email sender and CC information, and error screenshot path.

[0027] Meanwhile, Agent_monitor interacts with the operations management platform to obtain real-time scheduling plans for all current inspection tasks, including task name, execution time, priority, and dependencies, to support subsequent scheduling decisions. In addition, the robot collects server resource usage data, including CPU utilization, memory usage, disk I / O, and network latency, through server monitoring interfaces (such as Prometheus and Zabbix), to assess server load and determine whether execution is feasible.

[0028] In terms of parameter settings, Agent_monitor needs to be configured with the monitoring polling interval (e.g., collecting status once every 30 seconds), the task failure judgment threshold (e.g., judging it as abnormal if two consecutive non-empty strings are returned), the maximum number of re-executions (e.g., a1 times), and the time interval between adjacent re-execution operations (e.g., 5 minutes). These parameters can be dynamically adjusted according to specific business needs to adapt to inspection tasks of different scales and complexities.

[0029] This step is primarily used in practical applications for scenarios requiring continuous webpage inspection, such as enterprise-level OA systems and web application platforms. Through centralized monitoring and resource scheduling, it effectively reduces false alarm rates caused by network fluctuations, server lag, and other objective factors, improving the overall system's fault tolerance and operational efficiency. Furthermore, this step provides the data foundation for subsequent compensation execution, anomaly aggregation, and email notifications, making it a crucial link in achieving a closed-loop automated anomaly handling process in RPA.

[0030] Furthermore, S1 includes: S11 uses a one-master-two-standby architecture to deploy Agent_monitor. The master server processes the detected running result data first, and when the master server load exceeds a preset threshold, it automatically switches to the standby server for data collection.

[0031] Specifically, in some implementations, this step employs a "one master, two backups" architecture to deploy Agent_monitor, enabling highly available collection and processing of RPA inspection process execution results. This architecture is based on a master-slave server redundancy design, ensuring that the system can automatically switch to the backup server when the master server fails or is overloaded, thereby guaranteeing the continuity and stability of the inspection process.

[0032] At the technical implementation level, the Agent_monitor master server is responsible for receiving and processing inspection result data from multiple RPA robots (Agent_1 to Agent_N) in real time. This data includes the identity of the RPA server, global parameters of the inspection report (such as returning an empty string upon successful execution and specific error fields upon failure), email sender and CC information, and the storage path of error screenshots. The master server continuously monitors the execution status of each Agent through polling or event-driven mechanisms and performs preliminary analysis and processing of the result data. When the master server's key performance indicators (KPIs) such as CPU utilization, memory usage, or network latency exceed preset thresholds (e.g., CPU utilization > 85%, memory usage > 90%, response latency > 500ms), the system will automatically trigger a load balancing mechanism to switch new data collection tasks to one of the backup servers for execution.

[0033] At the parameter level, the load threshold of the main server can be dynamically configured according to actual business needs, for example, by setting it through the configuration interface of the operation and management platform. , , The backup server must have the same hardware configuration and software environment as the primary server to ensure the consistency and reliability of data processing after task switching. Furthermore, Agent_monitor communicates with each business robot using the TCP / IP protocol, and the data transmission format follows the JSON standard to ensure efficient parsing and processing of structured data.

[0034] In terms of application scenarios, this step is suitable for large-scale RPA inspection systems, especially in business scenarios that require continuous monitoring, such as OA systems, web applications, and online services. By deploying a primary / backup architecture for Agent_monitor, the system can automatically switch data collection tasks when server resources are strained or network fluctuations occur, avoiding interruptions to the entire inspection process due to single points of failure, thereby improving the system's fault tolerance and operational efficiency.

[0035] From a technical perspective, this step achieves highly available collection of RPA inspection process execution results, effectively avoiding data loss or processing delays caused by excessive load or failure of the main server. Through the setting of load thresholds and automatic switching mechanisms, the system can dynamically respond to resource changes, ensuring the continuity and stability of inspection tasks, providing a reliable data foundation for subsequent anomaly monitoring and compensation, and significantly improving overall operational efficiency and system resource utilization.

[0036] S12 uses a heartbeat detection mechanism to synchronize data between the primary and backup servers, ensuring the integrity of runtime result data, task scheduling plans, and resource status information during server switchover.

[0037] Specifically, in this application, the heartbeat detection mechanism for data synchronization between primary and backup servers is a key technical step in ensuring high system availability and data consistency. This mechanism is based on common health check and state synchronization strategies in distributed systems, combined with the operational characteristics of RPA processes, to achieve real-time synchronization of runtime result data, task scheduling plans, and resource status information between the primary and backup servers.

[0038] In some implementations, Agent_monitor, as the core monitoring component, is deployed on three application servers of the operations and management platform, employing a "one master, two backups" high-availability architecture. The master server is responsible for receiving and processing inspection result data from each business robot (Agent_1 to Agent_N) in real time, including process execution status, global parameters, error messages, and screenshot paths. The heartbeat detection mechanism periodically sends health check requests (such as HTTP Ping or TCP connection checks) to each Agent, and sets a maximum response time threshold. If no response is received within this time, the Agent is deemed abnormal, and the data synchronization process is triggered. Heartbeat interval. Typically set to The time limit is set to a minimum to balance real-time performance and system load.

[0039] Furthermore, when the master server detects that an Agent has failed to execute, Agent_monitor matches the identification information of the failed process with the task scheduling plan, and determines the appropriate course of action based on server resource usage (such as CPU utilization). Memory usage Network latency Select a backup server with lower load for compensation execution. Maximum number of compensation executions. The settings are determined by business personnel based on process stability and resource costs, and are typically as follows: This is to avoid wasting resources due to infinite retries.

[0040] In practical applications, this step can be deployed on enterprise-level RPA management platforms for automated operation and maintenance of large-scale webpage inspection tasks. Through heartbeat detection and data synchronization mechanisms, the system can quickly restore the process state when switching servers, ensuring the continuity of task scheduling and the integrity of data, thereby significantly improving system stability and operational efficiency.

[0041] S2, based on the empty string determination rules of the global parameters in the running result data, analyze whether there are non-empty substrings to identify abnormal inspection processes, and locate the business robot corresponding to the abnormal process according to the identity information.

[0042] Specifically, in some implementations, analyzing the existence of non-empty substrings based on the empty string judgment rules of global parameters in the execution result data to identify abnormal inspection processes, and locating the business robot corresponding to the abnormal process based on the identity information, is a key step in achieving automatic identification and fault location of abnormal processes in this invention. This step, by parsing the execution result data of the RPA inspection process and combining it with preset judgment logic, achieves rapid identification and accurate location of abnormal processes, thereby providing a foundation for subsequent compensation execution.

[0043] At the technical implementation level, this step first involves the Agent_monitor robot collecting the execution result data of the inspection process from multiple business robots (Agent_1 to Agent_N). This data is typically returned in string format, where global parameters should be empty strings (i.e., strings of length 0) when the process executes successfully, and contain specific error messages when the process fails. Agent_monitor analyzes the returned results using a string matching algorithm to determine whether there are non-empty substrings. Specifically, regular expressions or string length detection methods can be used, such as checking if len(result) > 0 for the returned results, or using the regular expression ^[^\s]+$ to detect the presence of non-empty content.

[0044] At the parameter level, key parameters involved in the judgment rules include: the string length threshold for global parameters, the matching rules for non-empty substrings, and the format of the identity field. For example, identity information is usually a unique identifier for the RPA server (such as IP address, machine ID, or business robot number), and its format must conform to the unified standard of the operation and management platform for subsequent resource scheduling and log tracking. Furthermore, the execution frequency of the judgment logic should match the scheduling cycle of the inspection task to ensure real-time performance.

[0045] At the application level, this step is widely used in RPA processes that require periodic inspections, such as enterprise-level OA systems and web application systems. When multiple inspection tasks are executed concurrently, Agent_monitor quickly locates the business robot involved in the abnormal process through the identity identifier field, thus avoiding the inefficient operation of checking logs one by one in the traditional way. This mechanism is particularly suitable for common failure scenarios such as network fluctuations, page loading errors, or insufficient server resources.

[0046] Furthermore, the technical benefits of this step lie in significantly improving the accuracy and response efficiency of abnormal process identification. By using standardized empty string judgment rules, false alarms caused by environmental fluctuations can be effectively filtered out, reducing redundant alarm information. Simultaneously, the precise location based on identity information provides a reliable basis for subsequent compensation execution and resource scheduling, thereby enhancing the overall fault tolerance and operational efficiency of the RPA system.

[0047] Furthermore, S2 includes: S21. When a non-empty substring is detected, semantic analysis is performed on the error field according to the preset anomaly classification rules to distinguish between network fluctuation anomalies and system crash anomalies.

[0048] Specifically, in some implementations, when a non-empty substring is detected, the system performs semantic analysis on the error field according to preset anomaly classification rules to distinguish between network fluctuation anomalies and system crash anomalies. This step is a key link in the overall inspection process compensation mechanism. Its technical implementation is based on a combination of natural language processing (NLP) and rule matching to perform structured identification and classification of the anomaly information returned during the execution of the RPA process.

[0049] At the technical implementation level, the system first receives inspection result data in real time from multiple RPA business robots (Agent_1 to Agent_N) through the Agent_monitor robot. This data is usually returned in string form, with an empty string ("") returned upon successful execution and a non-empty substring containing exception information upon failure. When a non-empty substring is detected, the system activates the semantic analysis module. This module matches and analyzes keywords, sentence structure, and context in the exception information based on a preset exception classification rule base. For example, if the error field contains keywords such as "ConnectionTimeout", "Network Error", or "DNS Resolution Failed", it can be identified as a network fluctuation exception; if it contains keywords such as "System Crash", "Memory Overflow", or "Process Terminated", it can be identified as a system crash exception.

[0050] At the parameter level, keyword matching rules in the anomaly classification rule base can be defined using regular expressions. Furthermore, the system allows setting a matching confidence threshold (e.g., 0.85) to ensure the accuracy of the classification results. The semantic analysis module supports multilingual processing and is suitable for anomaly information recognition in different language environments, including Chinese and English.

[0051] At the application level, this step is widely used in RPA systems for automated inspection of web applications, OA systems, ERP systems, etc. When an inspection task fails due to network instability or insufficient server resources, the system can automatically identify the type of anomaly and decide whether to schedule an idle robot to retry, thereby avoiding unnecessary duplicate alarms and manual intervention.

[0052] The technical benefits of this step are that, through precise semantic analysis and anomaly classification, the system can effectively filter false alarms caused by network fluctuations, reduce redundant alarm information, and improve operational response efficiency. Simultaneously, it provides a reliable basis for subsequent compensation mechanisms, ensuring the scientific and rational nature of resource scheduling strategies, thereby enhancing the overall stability and automation level of the RPA inspection system.

[0053] S22. If the network fluctuation is identified as an anomaly and the server resource utilization is lower than the preset threshold, the alarm email will be delayed until the compensation is completed.

[0054] Specifically, in some implementations, if the network fluctuation is identified as an anomaly and the server resource utilization is lower than a preset threshold, the alarm email is delayed until the compensation is completed. This step is one of the key control logics of the overall inspection process compensation mechanism in this invention. It aims to optimize the alarm information sending strategy, reduce redundant alarms caused by temporary network problems or low server load, and thus improve system operation and maintenance efficiency and resource utilization.

[0055] From a technical implementation perspective, this step first relies on Agent_monitor's real-time collection and classification of the inspection process execution results. When an inspection process returns a non-empty string, Agent_monitor uses the exception classification module to identify whether the exception belongs to the network fluctuation category. Network fluctuation exceptions typically include, but are not limited to: HTTP request timeouts, page loading failures, element location errors, and response codes other than 200. After determining that it is a network fluctuation exception, the system further queries the current server resource usage, including indicators such as CPU utilization, memory usage, and network bandwidth utilization. If the server resource utilization is lower than a preset threshold (e.g., CPU utilization < 60%, memory usage < 75%), it is determined that there are idle resources, and compensation execution can be performed.

[0056] At the parameter level, the criteria for determining network fluctuation anomalies can be based on HTTP response status codes, page load timeouts (e.g., >5000ms), and the number of element location failures (e.g., 3 consecutive failures). Server resource utilization thresholds can be configured by the system administrator based on actual load conditions, and are usually expressed as a percentage, such as... , Maximum number of compensation executions. It can be set by business personnel, for example This is to avoid wasting system resources due to endless retries.

[0057] At the application level, this step is suitable for large-scale RPA inspection systems, especially in environments with multiple servers and concurrent task execution. When an inspection task fails due to network fluctuations, but the overall system resources are still redundant, the system will automatically delay the sending of alarm emails and instead schedule idle robots to perform compensatory execution, thereby avoiding frequent intervention by developers due to temporary problems and improving the system's self-healing capabilities.

[0058] From a technical perspective, this step effectively filters redundant alarms caused by non-critical issues such as network fluctuations through an intelligent delayed alarm email sending mechanism, reducing the response burden on operations and maintenance personnel. Simultaneously, by dynamically assessing resource utilization, it ensures that compensation is performed when resources are sufficient, avoiding system performance degradation due to resource contention, thereby improving the stability and automation level of the RPA inspection system.

[0059] S3, combining the real-time task scheduling plan and server resource usage status, dynamically schedule idle servers to perform compensatory execution for abnormal processes. The number of compensatory executions does not exceed a preset threshold and the interval between adjacent executions is configured according to business requirements.

[0060] Specifically, in some implementations, this step involves dynamically scheduling idle servers to compensate for abnormal processes by combining real-time task scheduling plans with server resource usage status. Its core lies in leveraging the global monitoring capabilities of the operations management platform to achieve intelligent fault tolerance and resource optimization for the RPA inspection process. The technical implementation of this step is based on the Agent_monitor robot, deployed on three application servers of the operations management platform, employing a high-availability architecture of one primary and two backup servers to ensure system availability. Uninterrupted service. Agent_monitor receives inspection result data from each business robot (Agent_1 to Agent_N) in real time through API interface or message queue (such as Kafka, RabbitMQ), including key information such as server identity, global parameters, email notification configuration and error screenshot path.

[0061] At the parameter level, the maximum number of compensation executions is configured by the business unit in the operations management platform. For example, the compensation execution threshold for process 1 is... This threshold can be dynamically adjusted based on business tolerance and system stability requirements. The interval between adjacent compensation executions... It is also set by the sales staff, usually This is to avoid a surge in server load due to frequent retries. The trigger condition for compensation execution is the presence of a non-empty string in the process return result, indicating execution failure or an exception.

[0062] In application scenarios, this step is suitable for large-scale RPA inspection systems, especially in complex operation and maintenance environments with multiple server deployments and concurrent execution of multiple processes. When an inspection process fails due to network fluctuations, page obstruction, or server lag, Agent_monitor can quickly identify and schedule currently idle servers for compensatory execution, thereby avoiding the impact of single-point failures on overall inspection efficiency.

[0063] From a technical perspective, this step effectively improves the system's fault tolerance and resource utilization, reduces invalid alarms caused by objective environmental issues, and lowers the response burden on maintenance personnel. Through a dynamic resource scheduling mechanism, the system can achieve automatic recovery and execution optimization of the inspection process with minimal manual intervention, thereby enhancing the stability and automation level of the RPA system.

[0064] Furthermore, S3 includes: S31, based on the server resource usage status, a load balancing algorithm is used to select an idle server for compensation execution, the algorithm including: Specifically, in some implementations, a load balancing algorithm is used to select idle servers for compensation execution based on server resource usage status. The technical principle behind this is to collect and analyze resource usage metrics of each server in real time (such as CPU utilization, memory usage, and network bandwidth usage), and combine this with the resource requirements of the RPA inspection task to be compensated. A dynamic load balancing strategy is then used to select the optimal idle server from the server cluster for task compensation execution. This algorithm is typically based on mechanisms such as weighted round-robin or least connections to ensure efficient task scheduling and rational utilization of system resources.

[0065] In practice, the operation and management platform continuously monitors the running status of each RPA robot (Agent_1 to Agent_N) through the Agent_monitor module and obtains real-time resource usage data from the server monitoring interface. When a certain inspection process fails, Agent_monitor compares the process's resource requirements (such as required memory size, number of CPU cores, estimated execution time, etc.) with the server resource usage status, filtering out servers that meet the resource requirements and have low load. Optionally, this filtering process can introduce resource usage thresholds, for example, setting servers with CPU utilization not exceeding 70% and memory utilization not exceeding 80% as "schedulable idle servers".

[0066] Furthermore, the algorithm can use the following formula to evaluate resource matching when selecting a server:

[0067] in, Indicates server The current number of tasks, $C_{max}$ and $C_{max}$ represent the server, respectively. Current CPU utilization and maximum CPU utilization $M_{max}$ and $M_{max}$ represent the current memory utilization and the maximum memory utilization, respectively. This formula assigns a scheduling priority score to each server by comprehensively considering the server's load status and resource usage, thereby achieving the selection of the optimal server.

[0068] In practical applications, this step is typically deployed in the scheduling subsystem of an RPA operations management platform. It automatically triggers a compensation mechanism after an inspection task fails, reducing manual intervention and improving the system's self-healing capabilities. Its technical effect is to significantly reduce the failure rate of compensation tasks caused by insufficient server resources or excessive load, improve the overall success rate of the inspection process and system stability, while optimizing resource utilization and avoiding waste or overloading of server resources.

[0069] S32, Interval between adjacent compensation execution times Meet the minimum interval constraint of the business configuration.

[0070] Specifically, in some implementations, the step "adjacent compensation execution interval time" is... Meeting the minimum interval constraint of business configuration is a key control link in the overall inspection process compensation mechanism. Its technical implementation is based on the collaborative optimization of fault-tolerant handling of RPA process execution failures and resource scheduling. The core of this step is to avoid wasting server resources or abnormally increasing system load due to repeated execution of the same process in a short period of time by setting a reasonable interval between adjacent compensation executions, thereby improving the overall stability and resource utilization of the system.

[0071] From a technical implementation perspective The configuration depends on the task scheduling information and server resource usage status provided by the operations management platform. After receiving a failure result from a certain inspection process, Agent_monitor first queries the compensation strategy configuration for that process, including the maximum number of retries. Minimum execution interval Subsequently, Agent_monitor will determine whether the conditions for compensation are met based on the current server resource load (such as CPU utilization, memory usage, network latency, etc.). If the conditions are met, then... After the time interval, an idle RPA robot instance is invoked to retry the process. This time interval must meet the business requirements for inspection frequency and system response time, and is typically configured as follows: The specific value is set by the business personnel based on the urgency of the inspection task and the system's capacity.

[0072] At the parameter level, The minimum value should be greater than or equal to the minimum execution time of a single RPA process in the system. This is to ensure that the compensation execution does not conflict with the original process execution in terms of resources. At the same time, this interval must also meet the minimum interval constraint of the operations management platform for task scheduling. This is to prevent the task scheduler from triggering rate limiting due to excessively frequent requests. In practical applications, Typically, settings are dynamically configured through configuration files or database fields, supporting differentiated configurations based on factors such as process type, business priority, and server load level.

[0073] At the application level, this step is widely used in business scenarios requiring periodic inspections, such as enterprise-level OA systems, web application systems, and online service platforms. For example, in the web systems of banks or government agencies, inspection tasks may involve operations such as login verification, page element recognition, and data scraping. If failure occurs due to network fluctuations or server response delays, Agent_monitor will... The settings coordinate idle resources for compensatory execution within a reasonable time window, thereby avoiding false alarms and process interruptions caused by momentary failures.

[0074] Furthermore, the technical effect of this step is to significantly reduce the frequency of recurring alarms caused by unstructured problems such as network jitter and server lag, thereby improving system fault tolerance and resource scheduling efficiency. This can be achieved through proper configuration. It can ensure the integrity of inspection tasks while reducing the number of invalid retries, reducing the response burden of maintenance personnel, and realizing intelligent and automated management of RPA inspection processes.

[0075] S4 summarizes and analyzes the compensation execution results and the original operation results data, sends the processing results to managers and business personnel through the email notification system, and stores the system operation data in the Dameng database to build an operation and maintenance knowledge base.

[0076] Specifically, in some implementations, the compensation execution results are summarized and analyzed with the original operation results data, and the processing results are sent to managers and business personnel via email notification system. Simultaneously, the system operation data is stored in the DM database to build an operation and maintenance knowledge base. This is a key step in achieving closed-loop fault handling and knowledge accumulation in this invention. This step is based on a distributed task monitoring and compensation mechanism, using the Agent_monitor robot to centrally collect and analyze the execution results of multiple RPA inspection processes, thereby achieving intelligent identification and handling of abnormal processes.

[0077] At the technical implementation level, Agent_monitor receives real-time inspection result data from each business robot (Agent_1 to Agent_N) via API interfaces or message queues (such as Kafka and RabbitMQ). This data includes process status, global parameters (such as returning an empty string "" on success and a specific error field on failure), server identity, and error screenshot paths. Simultaneously, Agent_monitor obtains task scheduling information and server resource usage (such as CPU utilization, memory usage, and network latency) from the RPA management platform to assess the current system load and available resources. The comparison and analysis of the compensation execution results with the original execution results is performed using a rule engine or script logic to determine whether retry conditions are met. If a non-empty string exists, the process is considered abnormal, and an idle robot is scheduled to retry according to a preset retry threshold (e.g., process 1 can be retried a1 times) and a time interval (configured by business personnel, e.g., T_{\mathrm{interval}} = 300s).

[0078] At the parameter level, this step involves several key parameters, including but not limited to: maximum number of retries, retry interval, email sender and CC configuration fields, error screenshot storage path, and server resource usage thresholds. The email notification system uses the SMTP protocol or an internal enterprise email service interface, supports HTML content, and can embed error screenshots and execution log summaries to facilitate administrators in quickly locating problems.

[0079] At the application level, this step is widely used in business scenarios requiring regular inspections, such as enterprise-level OA systems, web applications, and online service platforms. Through centralized analysis and email notification mechanisms, it effectively reduces the burden on developers dealing with low-value alarm information and improves operational response efficiency. Simultaneously, storing system runtime data in the DM database builds an operational knowledge base, supporting subsequent process optimization, fault prediction, and knowledge reuse, providing data support for the long-term stable operation of the RPA system.

[0080] Furthermore, this step integrates the compensation execution results with the original operational data to achieve dynamic closed-loop management of the RPA inspection process, significantly reducing the false alarm rate caused by objective factors such as network fluctuations and server lag, improving system resource utilization and operational efficiency, and demonstrating good practicality and innovation.

[0081] Furthermore, S4 includes: S41, The inspection result data is written into the Dameng database according to a preset storage structure, wherein the storage structure includes: Specifically, in this application proposal, the step of "writing the inspection result data into the DM database according to the preset storage structure" is a key link in realizing systematic operation and maintenance management and fault tracing. The core of this step lies in using a structured data storage mechanism to uniformly archive the execution results, server resource usage, and task scheduling information from multiple RPA inspection processes into the DM database, providing data support for subsequent anomaly analysis, compensation mechanism execution, and knowledge base construction.

[0082] At the technical implementation level, this step utilizes the Agent_monitor robot for data collection and storage. The Agent_monitor, acting as the monitoring hub, is deployed on three application servers of the RPA management platform, employing a "one primary, two backup" high-availability architecture to ensure... It operates continuously. By monitoring the execution results of each business robot (Agent_1 to Agent_N), it obtains fields including RPA server identity, global parameters of the inspection report (returning an empty string on success and specific error fields on failure), email sender and CC information, and error screenshot paths. This data is then written into the DM database according to a unified table structure using preset SQL insert statements or ORM mapping methods, ensuring data consistency and traceability.

[0083] At the parameter level, the storage structure must meet the requirements for field integrity and data type specifications. For example, the server identity field is a string type with a length not exceeding 64 bytes; the global parameter field of the inspection report is a text type with a maximum length of 2048 bytes; the task execution timestamp field is a DATETIME type, accurate to the millisecond level; and the execution status field is an enumeration type, including status indicators such as success, failure, and retrieved. Furthermore, to improve query efficiency, key fields (such as timestamps and server identifiers) should have composite indexes, with the index type being a B-tree and the index granularity being at the table level.

[0084] At the application level, this step is widely used in automated inspection scenarios of enterprise-level OA systems, web application platforms, and online services. By centrally storing inspection result data, the operation and management platform can achieve unified monitoring and management of the entire RPA process, facilitating rapid location of abnormal processes, analysis of failure causes, and providing data basis for subsequent compensation script execution. Simultaneously, this storage structure supports historical data backtracking and statistical analysis, providing a foundation for building a routine operation and maintenance knowledge base.

[0085] From a technical perspective, this step significantly improves the data management capabilities and operational efficiency of the RPA inspection system through a standardized and structured data writing mechanism. On the one hand, the unified data format reduces development and maintenance costs; on the other hand, combined with the real-time scheduling and resource monitoring capabilities of the operation and management platform, it enables rapid response and automatic compensation for abnormal processes, thereby improving system stability and resource utilization.

[0086] The robotic process automation method based on webpage inspection anomaly monitoring and fault handling in this invention can automatically monitor and compensate for abnormal execution of webpage inspection RPA processes. By coordinating idle robot resources for retries, it effectively reduces redundant alarms caused by network fluctuations or server lag, thereby improving system stability and operational efficiency.

[0087] S5 builds an anomaly pattern recognition model based on system operation data stored in the Dameng database. By analyzing the distribution characteristics of error fields in historical inspection results data, it automatically generates compensation strategy optimization suggestions and updates the scheduling rule library of the operation and management platform.

[0088] Specifically, this step involves building an anomaly pattern recognition model based on system operation data stored in the DM database, and automatically generating compensation strategy optimization suggestions by analyzing the distribution characteristics of error fields in historical inspection results data, ultimately updating the scheduling rule base of the operation and management platform. The core of this step lies in using a data-driven approach to achieve intelligent identification and dynamic optimization of the response mechanism for abnormal behaviors in the RPA inspection process.

[0089] At the technical implementation level, the Agent_monitor robot is responsible for collecting operational result data from multiple RPA inspection processes, including key information such as process execution status, error fields, server identity, and task scheduling timestamps. This data is uploaded in real time and persistently stored in the DM database of the operations management platform, forming a structured system operation log. The anomaly pattern recognition model is trained based on this historical data, with its input being the set of error fields from the inspection processes. The output is the anomaly type classification result. ,in Indicates the first Error fields that appeared during the second inspection. Indicates the first Anomaly patterns. The model can use rule-based text matching algorithms or machine learning classifiers (such as Naive Bayes, Random Forest, etc.) to perform cluster analysis on the error fields to identify high-frequency anomaly types and their triggering conditions.

[0090] At the parameter level, key thresholds need to be set during model training, such as a threshold for the frequency of error fields. Used to determine whether a certain type of exception is representative; maximum number of re-executions. This is used to control the intensity of the compensation strategy and the time interval between adjacent execution operations. Typically, this is set by business personnel based on system load and inspection priorities, such as... Seconds. Furthermore, the scheduling rule base of the operations management platform needs to support a dynamic update mechanism to ensure that compensation strategies take effect in real time.

[0091] At the application level, this step is suitable for business scenarios requiring high-frequency and stable inspections, such as enterprise-level OA systems and web application platforms. Through an anomaly pattern recognition model, the system can automatically determine whether the current error belongs to a known compensable anomaly type, thereby deciding whether to trigger a compensation strategy. If it is identified as a compensable anomaly, the system automatically allocates tasks and executes compensation operations based on the idle robot resource status in the scheduling rule base, without requiring manual intervention.

[0092] In terms of technical effectiveness, this step significantly enhances the intelligence and self-healing capabilities of the RPA inspection system. By constructing an anomaly pattern recognition model, the system can extract key anomaly features from massive logs, reducing invalid alarms and improving operational efficiency. Simultaneously, the automatic generation of compensation strategies and the dynamic updating of the scheduling rule base enable the system to possess stronger fault tolerance and resource coordination capabilities, thereby improving the overall process stability and execution success rate.

[0093] The robotic process automation method based on webpage inspection anomaly monitoring and fault handling in this invention constructs an anomaly pattern recognition model based on the DM database, performs feature analysis on error fields in historical inspection results, realizes automatic generation of compensation strategies and dynamic optimization of scheduling rule base, further improves the intelligence level of anomaly handling and system adaptability, thereby significantly enhancing the stability of RPA process and operation and maintenance response efficiency.

[0094] To achieve the above embodiments, the present invention also proposes a robotic process automation device based on webpage inspection anomaly monitoring and fault handling. The device includes: The primary and backup monitoring data acquisition module is used to acquire the running result data, real-time task scheduling plan and server resource usage status of each business robot executing the web inspection process through the primary and backup monitoring robots deployed on the operation and management platform; The anomaly identification and location module is used to analyze whether there are non-empty substrings based on the empty string judgment rules of the global parameters in the running result data to identify the abnormal inspection process, and locate the business robot corresponding to the abnormal process according to the identity information. The dynamic compensation scheduling module is used to dynamically schedule idle servers to perform compensation execution for abnormal processes by combining the real-time task scheduling plan and server resource usage status. The number of compensation executions does not exceed a preset threshold and the interval between adjacent executions is configured according to business requirements. The results processing and storage module is used to summarize and analyze the compensation execution results and the original operation results data, send the processing results to managers and business personnel through the email notification system, and store the system operation data in the Dameng database to build an operation and maintenance knowledge base.

[0095] Regarding the apparatus in the above embodiments, the specific manner in which each module performs its operation has been described in detail in the embodiments related to the method, and will not be elaborated upon here.

[0096] To implement the above embodiments, the present invention also proposes an electronic device, including: a processor and a memory communicatively connected to the processor; the memory stores computer execution instructions; the processor executes the computer execution instructions stored in the memory to implement the method provided in the foregoing embodiments.

[0097] To implement the above embodiments, the present invention also proposes a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, are used to implement the methods provided in the foregoing embodiments.

[0098] To implement the above embodiments, the present invention also proposes a computer program product, including a computer program that, when executed by a processor, implements the methods provided in the foregoing embodiments.

[0099] The collection, storage, use, processing, transmission, provision, and disclosure of user personal information involved in this invention all comply with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0100] It should be noted that personal information collected from users should be used for legitimate and reasonable purposes and should not be shared or sold outside of these legitimate uses. Furthermore, such collection / sharing should only be conducted after receiving the user's informed consent, including but not limited to notifying the user to read the user agreement / user notice and sign an agreement / authorization that includes authorization of relevant user information before the user uses the function. In addition, any necessary steps must be taken to protect and safeguard access to such personal information data and ensure that others with access to personal information data comply with their privacy policies and procedures.

[0101] This invention is intended to provide implementation schemes for users to selectively prevent the use or access to personal information data. That is, this disclosure is intended to provide hardware and / or software to prevent or block access to such personal information data. Once personal information data is no longer needed, risks can be minimized by restricting data collection and deleting data. Furthermore, where applicable, such personal information can be de-identified to protect user privacy.

[0102] In the foregoing descriptions of the embodiments, the terms "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0103] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this invention, "a plurality of" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0104] Any process or method description in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or more executable instructions for implementing custom logic functions or processes, and the scope of preferred embodiments of the invention includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functions involved, as should be understood by those skilled in the art to which embodiments of the invention pertain.

[0105] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of computer-readable media include: an electrical connection having one or more wires (electronic device), a portable computer disk drive (magnetic device), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Alternatively, the computer-readable medium may be paper or other suitable media on which the program can be printed, since the program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in a computer memory.

[0106] It should be understood that various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware as in another embodiment, it can be implemented using any of the following techniques known in the art, or a combination thereof: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0107] Those skilled in the art will understand that all or part of the steps of the methods in the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, the program includes one or a combination of the steps of the method embodiments.

[0108] Furthermore, the functional units in the various embodiments of the present invention can be integrated into a processing module, or each unit can exist physically separately, or two or more units can be integrated into a module. The integrated module can be implemented in hardware or as a software functional module. If the integrated module is implemented as a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium.

[0109] The storage medium mentioned above can be a read-only memory, a disk, or an optical disk, etc. Although embodiments of the present invention have been shown and described above, it is to be understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of the present invention.

[0110] It should be understood that the various forms of processes shown above can be used, with steps reordered, added, or deleted. For example, the steps described in this invention can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of this invention can be achieved, and this is not limited herein.

[0111] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A robotic process automation method based on webpage inspection anomaly monitoring and fault handling, characterized in that, include: S1 obtains the running result data, real-time task scheduling plan and server resource usage status of each business robot by using the primary and backup monitoring robots deployed on the operation and management platform to perform web page inspection processes; S2, based on the empty string determination rules of the global parameters in the running result data, analyze whether there are non-empty substrings to identify abnormal inspection processes, and locate the business robot corresponding to the abnormal process according to the identity information; S3, combining the real-time task scheduling plan and server resource usage status, dynamically schedule idle servers to perform compensatory execution for abnormal processes. The number of compensatory executions does not exceed a preset threshold and the interval between adjacent executions is configured according to business requirements. S4 summarizes and analyzes the compensation execution results and the original operation results data, sends the processing results to managers and business personnel through the email notification system, and stores the system operation data in the Dameng database to build an operation and maintenance knowledge base.

2. The method as described in claim 1, characterized in that, S1 further includes: Agent_monitor is deployed using a one-master-two-standby architecture. The master server processes the detected running result data first, and when the master server load exceeds a preset threshold, it automatically switches to the standby server for data collection. A heartbeat detection mechanism is used to synchronize data between the primary and backup servers, ensuring the integrity of runtime results data, task scheduling plans, and resource status information during server switchover.

3. The method as described in claim 1, characterized in that, S2 further includes: When a non-empty substring is detected, semantic analysis is performed on the error field according to the preset anomaly classification rules to distinguish between network fluctuation anomalies and system crash anomalies. If the anomaly is determined to be a network fluctuation and the server resource utilization is lower than the preset threshold, the alarm email will be delayed until the compensation is completed.

4. The method as described in claim 1, characterized in that, S3 further includes: Based on the server resource usage status, a load balancing algorithm is used to select idle servers for compensation execution. The algorithm includes: in For server The current load, For server Maximum resource capacity; Adjacent compensation execution interval Minimum interval constraint to meet business configuration: in and These are the minimum and maximum interval times configured for business personnel, respectively.

5. The method as described in claim 1, characterized in that, Also includes: An anomaly pattern recognition model is built based on system operation data stored in the Dameng database. By analyzing the distribution characteristics of error fields in historical inspection results data, compensation strategy optimization suggestions are automatically generated and the scheduling rule library of the operation and management platform is updated.

6. The method as described in claim 1, characterized in that, S4 further includes: The inspection results data are written into the Dameng database according to a preset storage structure, which includes: a process identifier field. Server identity field Exception type field , number of times compensation is executed Timestamp field .

7. A robotic process automation device based on webpage inspection anomaly monitoring and fault handling, characterized in that, include: The primary and backup monitoring data acquisition module is used to acquire the running result data, real-time task scheduling plan and server resource usage status of each business robot executing the web inspection process through the primary and backup monitoring robots deployed on the operation and management platform; The anomaly identification and location module is used to analyze whether there are non-empty substrings based on the empty string judgment rules of the global parameters in the running result data to identify the abnormal inspection process, and locate the business robot corresponding to the abnormal process according to the identity information. The dynamic compensation scheduling module is used to dynamically schedule idle servers to perform compensation execution for abnormal processes by combining the real-time task scheduling plan and server resource usage status. The number of compensation executions does not exceed a preset threshold and the interval between adjacent executions is configured according to business requirements. The results processing and storage module is used to summarize and analyze the compensation execution results and the original operation results data, send the processing results to managers and business personnel through the email notification system, and store the system operation data in the Dameng database to build an operation and maintenance knowledge base.

8. An electronic device, characterized in that, include: A processor, and a memory communicatively connected to the processor; The memory stores computer-executed instructions; The processor executes computer execution instructions stored in the memory to implement the method as described in any one of claims 1-6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions, which, when executed by a processor, are used to implement the method as described in any one of claims 1-6.

10. A computer program product, characterized in that, Includes a computer program that, when executed by a processor, implements the method of any one of claims 1-6.

Images