Application exception processing method, system and equipment based on agent arrangement

Abstract

The invention relates to the technical field of computers, in particular to an application exception handling method, system and equipment based on agent arrangement, which can efficiently handle application exception problems. The application exception handling method based on agent arrangement comprises the steps that by means of agent arrangement, an exception handling workflow comprising a plurality of agent units capable of being arranged is formed in a visual mode for an application exception condition, and a flow engine of the exception handling workflow comprises a control node and a task node; for the captured application exception information, exception information features are extracted through exception analysis logic by using an exception recognition agent unit, and the influence range of the application exception is delimited through an entity relationship graph to form an exception analysis result to be output to a front-end visual interface or pushed to a collaborative agent unit; and establishing an agent co-processing system comprising a co-processing agent unit based on the exception analysis result, matching a processing scheme for the application exception information according to a co-processing strategy, and processing the exception condition.

Description

Technical Field

[0001] One or more embodiments of the present invention relate to the field of computer technology, and more particularly to an application exception handling method, system and device based on intelligent agent orchestration. Background Technology

[0002] Current application systems generally face the problem of inefficient anomaly handling. According to Gartner's 2024 Operations Report, more than 73% of IT failures have a mean time to repair (MTTR) of more than 4 hours, with 42% of that time spent on anomaly localization.

[0003] Traditional handling methods have the following drawbacks: (1) Most abnormal problem location relies on manual viewing of background logs, operation indicators or auxiliary call chain monitoring, lacking multi-dimensional automated correlation analysis, and complex abnormalities often require multiple people to participate in the handling; (2) The capabilities of operation and maintenance personnel vary, and the handling solutions accumulated based on past experience have not been included in the experience library or are scattered in various documents, making it impossible to form a structured knowledge graph; (3) Existing monitoring systems can often only trigger basic alarms based on indicator parameters, lacking the ability to automatically locate root causes and generate handling suggestions. Summary of the Invention

[0004] One or more embodiments of the present invention describe an application anomaly handling method, system and device based on intelligent agent orchestration, which can efficiently handle application anomaly problems.

[0005] According to an embodiment of a first aspect of the present invention, an application anomaly handling method based on agent orchestration is provided, comprising:

[0006] Through intelligent agent orchestration, for application system anomalies, an anomaly handling workflow consisting of multiple orchestratable intelligent agent units is formed in a visual manner. The workflow engine of the anomaly handling workflow includes a control node and multiple task nodes.

[0007] For the application anomaly information of the captured application system, the anomaly identification intelligent agent unit extracts the features of the anomaly information through anomaly analysis logic and delimits the scope of influence of the application anomaly through entity relationship graph to form anomaly analysis results for the application anomaly information. Then, the anomaly analysis results are output to the front-end visualization interface or pushed to the collaborative intelligent agent unit associated with the anomaly identification intelligent agent unit.

[0008] Based on the anomaly analysis results, an intelligent agent collaborative handling system is established, which includes the collaborative intelligent agent unit. According to the collaborative handling strategy, a corresponding handling scheme is matched for the application anomaly information, and the anomaly situation related to the application anomaly information is handled.

[0009] Preferably, in any embodiment, the step of analyzing the captured application anomaly information includes:

[0010] The application system's exception logs are collected through a real-time capture layer, and regular expressions and data cleaning algorithms are used to standardize the format of unstructured information in the application exception information.

[0011] A pre-trained language model is used to perform semantic understanding and intent recognition on the text of the application's anomaly information, and standardized anomaly type labels are output.

[0012] Using natural language processing tools as word segmentation tools, core abnormal feature keywords are extracted from the application's abnormal information through part-of-speech tagging and named entity recognition technologies, and redundant noise is eliminated to construct an abnormal information feature vector.

[0013] Preferably, in any embodiment, the step of analyzing the captured application anomaly information includes:

[0014] A real-time analysis engine is built based on an open-source distributed search and analysis engine, and the impact range of the application anomalies is visualized and delimited through an entity relationship graph.

[0015] Preferably, in any embodiment,

[0016] The steps of the intelligent agent orchestration include: using the mxgraph graphics library to build a JavaScript drawing component to realize a visual orchestration tool and process engine for intelligent agent orchestration.

[0017] Preferably, in any embodiment,

[0018] The steps for arranging intelligent agents include: arranging intelligent agents through drag-and-drop interaction.

[0019] Preferably, in any embodiment,

[0020] The control nodes include: a start node that serves as the entry point for triggering a workflow, an end node that serves as the termination identifier for a workflow, and a gateway node that supports branching or aggregation logic control.

[0021] Preferably, in any embodiment,

[0022] The task node includes at least one of the following: a form configuration component for structured data entry, an executor selector for third-party service calls, a JavaScript code editor for custom logic implementation, an atomic capability component for microservice interface encapsulation, a large model interaction component for artificial intelligence capability integration, and a barcode recognition component for physical world interaction interfaces.

[0023] Preferably, in any embodiment, the collaborative processing strategy includes at least one of the following:

[0024] For application anomalies that the system can repair autonomously, the anomaly analysis results are sent to the automatic processing intelligent agent unit with preset automatic processing logic for automatic processing.

[0025] For application anomalies that require manual intervention or tool assistance, the anomaly analysis results are sent to the human-computer interaction intelligent agent unit for human-computer collaborative processing.

[0026] For complex application anomalies that lack online solutions, the anomaly analysis results are sent to the operation and maintenance interactive intelligent agent unit for collaborative processing with the assistance of the operation and maintenance system.

[0027] According to an embodiment of a second aspect of the present invention, an application exception handling system based on agent orchestration is provided for implementing the application exception handling method based on agent orchestration as described above, comprising:

[0028] The intelligent agent orchestration module is used to form an exception handling workflow, which includes multiple orchestratable intelligent agent units, in a visual manner for application exceptions of the application system through intelligent agent orchestration. The workflow engine of the exception handling workflow includes a control node and multiple task nodes.

[0029] An anomaly analysis module is used to extract features of the anomaly information captured by the application system using an anomaly identification intelligent agent unit through anomaly analysis logic and to define the scope of influence of the application anomaly through an entity relationship graph, so as to form an anomaly analysis result for the application anomaly information. Then, the anomaly analysis result is output to the front-end visualization interface or pushed to the collaborative intelligent agent unit associated with the anomaly identification intelligent agent unit.

[0030] The collaborative handling module is used to establish an intelligent agent collaborative handling system including the collaborative intelligent agent unit based on the anomaly analysis results, match the corresponding handling scheme for the application anomaly information according to the collaborative handling strategy, and handle the anomaly situation related to the application anomaly information.

[0031] According to an embodiment of a third aspect of the present invention, an application exception handling device based on agent orchestration is provided, comprising a memory and a processor, wherein the memory stores executable code, and the processor executes the executable code to implement the application exception handling method based on agent orchestration as described above.

[0032] The application anomaly handling method, system, and device based on intelligent agent orchestration provided by one or more embodiments of the present invention can efficiently handle application anomaly problems. Attached Figure Description

[0033] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments of the present invention or the prior art will be briefly introduced below. Obviously, the drawings described below relate to some embodiments of the present invention. However, for those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0034] Figure 1 This is a flowchart illustrating an application exception handling method based on intelligent agent orchestration according to an embodiment of the present invention.

[0035] Figure 2 This is a schematic diagram of the structure of an application exception handling system based on intelligent agent orchestration according to an embodiment of the present invention. Detailed Implementation

[0036] The technical solutions of various embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments described in the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0037] One or more embodiments of the present invention describe an application anomaly handling method, system and device based on intelligent agent orchestration, which can efficiently handle application anomaly problems.

[0038] According to an embodiment of a first aspect of the present invention, an application anomaly handling method based on agent orchestration is provided, comprising:

[0039] Through intelligent agent orchestration, for application system anomalies, an anomaly handling workflow consisting of multiple orchestratable intelligent agent units is formed in a visual manner. The workflow engine of the anomaly handling workflow includes a control node and multiple task nodes.

[0040] For the application anomaly information of the captured application system, the anomaly identification intelligent agent unit extracts the features of the anomaly information through anomaly analysis logic and delimits the scope of influence of the application anomaly through entity relationship graph to form anomaly analysis results for the application anomaly information. Then, the anomaly analysis results are output to the front-end visualization interface or pushed to the collaborative intelligent agent unit associated with the anomaly identification intelligent agent unit.

[0041] Based on the anomaly analysis results, an intelligent agent collaborative handling system is established, which includes the collaborative intelligent agent unit. According to the collaborative handling strategy, a corresponding handling scheme is matched for the application anomaly information, and the anomaly situation related to the application anomaly information is handled.

[0042] In this way, by orchestrating intelligent agents, the handling process of application anomalies can be broken down into the work of multiple selectable and orchestratable intelligent agent units. The workflow engine enables the dynamic combination of the handling process, and can be combined with auxiliary tools and technologies such as natural language processing toolkits (e.g., HANLP), open-source distributed search and analysis engines (e.g., Elasticsearch), and knowledge bases to improve the ability to identify application anomaly features and generate matching solutions / handling schemes. This can shorten the problem localization time and efficiently provide practical and effective solutions to application anomalies.

[0043] The purpose of this invention is to solve the problem of low efficiency in handling application anomalies in application systems. By using intelligent agent orchestration, combined with large models and auxiliary tools, it can quickly analyze anomalies and provide solutions, shortening the anomaly handling time and making application system anomaly handling more timely, accurate and clear.

[0044] By using an intelligent agent orchestration platform, various problem-solving logics for anomaly handling can be configured in a visual manner. Through anomaly identification intelligent agent units and multi-dimensional auxiliary means, the accuracy of anomaly identification, analysis and localization and the reliability of optimal solution generation can be improved. Combined with multi-agent unit collaborative technology, accurate solution delivery and closed-loop management of anomaly handling can be achieved.

[0045] Therefore, the application anomaly handling method based on intelligent agent orchestration provided by the embodiments of the present invention can efficiently handle application anomaly problems.

[0046] Preferably, in any embodiment, the steps of intelligent agent orchestration include: using the mxgraph graphics library to build a JavaScript drawing component to realize a visual orchestration tool and process engine for intelligent agent orchestration.

[0047] Preferably, in any embodiment, the agent orchestration step includes: orchestrating agents through a drag-and-drop interactive method. In this case, it can support the visual orchestration of the processing logic of related agent units and the execution order of sub-nodes, thus providing complete flow control capabilities.

[0048] Preferably, in any embodiment, the control node includes: a start node serving as a workflow trigger entry point, an end node serving as a workflow termination identifier, and a gateway node supporting branch or aggregation logic control.

[0049] Preferably, in any embodiment, the task node includes at least one of the following: a form configuration component for structured data entry, an executor selector for third-party service calls, a JavaScript code editor for custom logic implementation, an atomic capability component for microservice interface encapsulation, a large model interaction component for artificial intelligence (AI) capability integration, a barcode recognition component for physical world interaction interfaces, and other functional modules required to perform the processing task.

[0050] Optionally, in any embodiment, the step of analyzing the captured application anomaly information includes: using a natural language processing tool (e.g., HANLP) as a word segmentation tool, and extracting core anomaly feature keywords from the application anomaly information through part-of-speech tagging and named entity recognition technology for anomaly analysis.

[0051] Preferably, in any embodiment, the step of analyzing the captured application anomaly information includes:

[0052] The application system's exception logs are collected through a real-time capture layer, and regular expressions and data cleaning algorithms are used to standardize the format of unstructured information in the application exception information (thereby ensuring the consistency of data input).

[0053] Using a pre-trained language model, semantic understanding and intent recognition are performed on the text of the application's abnormal information, and standardized abnormal type labels are output (in this case, the abnormal situation of the application is classified based on a large model, providing a classification basis for subsequent differentiated processing);

[0054] Using natural language processing tools (such as HANLP) as word segmentation tools, core abnormal feature keywords are extracted from the application's abnormal information through part-of-speech tagging and named entity recognition technologies, and redundant noise is eliminated (e.g., using the TF-IDF algorithm to eliminate redundant noise) to construct an abnormal information feature vector.

[0055] Optionally, in any embodiment, the step of analyzing the captured application anomaly information includes: building a real-time analysis engine based on an open-source distributed search analysis engine (e.g., Elasticsearch), and visually delimiting the scope of influence of the application anomaly through an entity relationship graph.

[0056] Preferably, in any embodiment, the step of analyzing the captured application anomaly information includes: building a real-time analysis engine based on Elasticsearch, and visually delimiting the scope of influence of the application anomaly through an entity relationship graph.

[0057] Optionally, in any embodiment, the visualization of the impact range of the application anomaly supports multi-dimensional association retrieval and root cause localization.

[0058] Optionally, in any embodiment, the step of analyzing the captured application anomaly information includes: using a similarity matching algorithm to achieve an accurate mapping between the application anomaly and its solution / handling plan by using historical cases, expert rules, and handling plans stored in a structured manner in a multimodal knowledge base system.

[0059] Optionally, in any embodiment, the step of analyzing the captured application anomaly information includes: using a large model as a verification layer, performing a secondary evaluation of the matching anomaly handling scheme through logical reasoning and compliance checks, outputting a confidence score and correction suggestions to reduce the risk of misjudgment.

[0060] Optionally, in any embodiment, the step of analyzing the captured application anomaly information includes: automatically triggering related data query and context completion operations by connecting to the backend capability gateway, and dynamically optimizing the executability and completeness of the anomaly handling scheme.

[0061] Optionally, in any embodiment, the step of analyzing the captured application anomaly information includes: using the text generation capability of the large model to integrate the analysis process, processing suggestions and execution basis of the application anomaly information into a standard language format or report document, and outputting it to the front end.

[0062] Optionally, in any embodiment, the anomaly recognition intelligent agent unit is used to perform two-stage intelligent agent unit orchestration and processing: First, in the application anomaly information capture stage, anomaly data input is acquired in real time and the application anomaly status can be updated to "processing" synchronously; then, in the intelligent analysis and processing stage, the application anomaly information is processed by the anomaly analysis logic built into the anomaly recognition intelligent agent unit, and the anomaly analysis results are output to the front-end visualization interface or pushed to the collaborative intelligent agent unit associated with the anomaly recognition intelligent agent unit.

[0063] Preferably, in any embodiment, the collaborative processing strategy includes at least one of the following:

[0064] For application anomalies that the system can repair autonomously, the anomaly analysis results are sent to the automatic processing intelligent agent unit with preset automatic processing logic for automatic processing.

[0065] For application anomalies that require manual intervention or tool assistance, the anomaly analysis results are sent to the human-computer interaction intelligent agent unit for human-computer collaborative processing.

[0066] For complex application anomalies that lack online solutions, the anomaly analysis results are sent to the operation and maintenance interactive intelligent agent unit for collaborative processing with the assistance of the operation and maintenance system.

[0067] Optionally, in any embodiment, when the anomaly analysis result is sent to the automatic processing intelligent agent unit with preset automatic processing logic for automatic processing, the automatic processing intelligent agent unit dynamically calls the underlying atomic capability interface to realize the fully automated closed-loop processing of the detection, analysis and repair of application anomalies.

[0068] Optionally, in any embodiment, when the anomaly analysis results are sent to the human-computer interaction intelligent agent unit for human-computer collaborative processing, the human-computer interaction intelligent agent unit configures a visual operation interface and decision support tools on the orchestration platform according to the business scenario characteristics related to the application anomaly, to assist in completing the anomaly handling.

[0069] Optionally, in any embodiment, when the anomaly analysis results are sent to the operation and maintenance interactive intelligent agent unit for collaborative processing with the assistance of the operation and maintenance system, the operation and maintenance interactive intelligent agent unit provides operation and maintenance processing suggestions based on the solutions provided by the anomaly identification intelligent agent unit, so as to assist the operation and maintenance system in collaboratively resolving the application anomaly.

[0070] Optionally, in any embodiment, when the anomaly analysis result is sent to the operation and maintenance interactive intelligent agent unit for collaborative processing with the assistance of the operation and maintenance system, the distributed inspection intelligent agent cluster is configured to poll to determine whether the application anomaly has been recovered. If the application anomaly has been recovered, other collaborative intelligent agent units are promptly notified to perform the corresponding order cancellation operation.

[0071] In a preferred embodiment of the present invention, the agent orchestration step includes: constructing a JavaScript drawing component using the mxgraph graphics library to realize a visual orchestration tool and process engine for agent orchestration. Preferably, a drag-and-drop interactive method is used to support the visual orchestration of agent processing logic and the execution order of child nodes, providing complete process control capabilities.

[0072] In a preferred embodiment of the present invention, the core node types of the process engine are divided into two main categories: control nodes and task nodes.

[0073] In a preferred embodiment of the present invention, the control node includes a start node (process trigger entry point), an end node (process termination identifier), and a gateway node (supporting logical control such as branching / aggregation).

[0074] In a preferred embodiment of the present invention, the task node, as a specific processing unit, provides a form configuration component (structured data entry), an executor selector (third-party service call), a JavaScript code editor (custom logic implementation), an atomic capability component (microservice interface encapsulation), a large model interaction component (AI capability integration), a barcode recognition component (physical world interaction interface), and other related functional modules.

[0075] In a preferred embodiment of the present invention, the step of analyzing the captured application anomaly information includes:

[0076] (1) Standardized processing of abnormal information

[0077] The application system's exception logs are collected through a real-time capture layer. Regular expressions and data cleaning algorithms are used to standardize the format of unstructured information, ensuring the consistency of data input.

[0078] (2) Anomaly classification based on large models

[0079] By using a pre-trained language model to perform semantic understanding and intent recognition on abnormal text, standardized abnormal type labels are output to provide a classification basis for subsequent differential processing.

[0080] (3) Abnormal information feature extraction

[0081] Using natural language processing tools (such as HANLP) as word segmentation tools, core abnormal keywords are extracted through part-of-speech tagging and named entity recognition technologies, and redundant noise is eliminated by combining the TF-IDF algorithm to construct high-value information feature vectors.

[0082] (4) Distributed anomaly location analysis

[0083] A real-time analysis engine is built based on an open-source distributed search and analysis engine (such as Elasticsearch). It uses an entity relationship graph to visualize and delimit the scope of anomaly impact, and supports multi-dimensional association retrieval and root cause localization.

[0084] (5) Knowledge-driven solution matching

[0085] Establish a multimodal knowledge base system to structurally store historical cases, expert rules, and contingency plans, and use a similarity matching algorithm to achieve accurate mapping between abnormal scenarios and solutions.

[0086] (6) Intelligent calibration mechanism of the scheme

[0087] A large model is introduced as a verification layer. The matching scheme is evaluated a second time through logical reasoning and compliance checks, and confidence scores and correction suggestions are output to reduce the risk of misjudgment.

[0088] (7) Background processing capabilities execution

[0089] By connecting to the backend capability gateway, it automatically triggers related data queries and context completion operations, dynamically optimizing the executability and completeness of the processing solution.

[0090] (8) Generation of structured results

[0091] Leveraging the text generation capabilities of large models, the analysis process, processing suggestions, and execution basis are integrated into a standard language format or report document and output to the front end.

[0092] In this way, the anomaly identification intelligent agent unit based on large model technology adopts a full-link analysis logic of standardized processing → large model classification → feature extraction → distributed localization → knowledge matching → intelligent calibration → background execution → structured output, thereby realizing an efficient application anomaly handling process.

[0093] In a preferred embodiment of the present invention, the step of processing the abnormal situation related to the application abnormal information according to the collaborative handling strategy includes:

[0094] (1) For abnormal events that the system can repair autonomously, push them to the automatic processing intelligent agent module. This module uses the intelligent agent orchestration platform to pre-set processing logic and dynamically call the underlying atomic capability interfaces to realize a fully automated closed loop of anomaly detection, analysis and repair.

[0095] (2) Abnormal events requiring manual intervention or tool assistance are handled by the human-interactive intelligent agent. This module configures a visual operation interface and decision support tools on the orchestration platform according to the characteristics of the business scenario to assist in the handling of abnormal events.

[0096] (3) In cases of complex anomalies or lack of online solutions, the issue is transferred to the operations and maintenance (O&M) interactive intelligent agent for handling. This module provides O&M handling suggestions based on the solutions offered by the anomaly intelligent identification agent, assisting O&M in resolving the anomaly. A distributed inspection intelligent agent cluster is configured to poll the system to determine whether the anomaly has been resolved. If the anomaly has been resolved, other collaborative intelligent agents are promptly notified to perform the corresponding order cancellation operation.

[0097] Figure 1 This is a flowchart illustrating an application exception handling method based on intelligent agent orchestration according to an embodiment of the present invention.

[0098] exist Figure 1 An application anomaly handling method based on agent orchestration can be seen in the embodiments, including:

[0099] 110: Through intelligent agent orchestration, for application anomalies of the application system, an anomaly handling workflow including multiple orchestratable intelligent agent units is formed in a visual manner. The workflow engine of the anomaly handling workflow includes a control node and multiple task nodes.

[0100] 120: For the captured application anomaly information of the application system, the anomaly identification intelligent agent unit extracts the features of the anomaly information through anomaly analysis logic and delimits the scope of influence of the application anomaly through entity relationship graph to form anomaly analysis results for the application anomaly information. Then, the anomaly analysis results are output to the front-end visualization interface or pushed to the collaborative intelligent agent unit associated with the anomaly identification intelligent agent unit.

[0101] 130: Based on the anomaly analysis results, establish an agent collaborative handling system including the collaborative agent unit, match corresponding handling schemes for the application anomaly information according to the collaborative handling strategy, and handle the anomaly situation related to the application anomaly information.

[0102] According to an embodiment of a second aspect of the present invention, an application exception handling system based on agent orchestration is provided for implementing the application exception handling method based on agent orchestration as described above, comprising:

[0103] The intelligent agent orchestration module is used to form an exception handling workflow, which includes multiple orchestratable intelligent agent units, in a visual manner for application exceptions of the application system through intelligent agent orchestration. The workflow engine of the exception handling workflow includes a control node and multiple task nodes.

[0104] An anomaly analysis module is used to extract features of the anomaly information captured by the application system using an anomaly identification intelligent agent unit through anomaly analysis logic and to define the scope of influence of the application anomaly through an entity relationship graph, so as to form an anomaly analysis result for the application anomaly information. Then, the anomaly analysis result is output to the front-end visualization interface or pushed to the collaborative intelligent agent unit associated with the anomaly identification intelligent agent unit.

[0105] The collaborative handling module is used to establish an intelligent agent collaborative handling system including the collaborative intelligent agent unit based on the anomaly analysis results, match the corresponding handling scheme for the application anomaly information according to the collaborative handling strategy, and handle the anomaly situation related to the application anomaly information.

[0106] Optionally, in any embodiment, the intelligent agent collaborative processing system in the collaborative processing module enables the collaborative linkage of multiple related intelligent agent units.

[0107] Optionally, in any embodiment, the intelligent agent collaborative processing system controls the automatic processing intelligent agent unit to perform closed-loop automatic processing, controls the human-machine interaction intelligent agent unit to assist the user in human-machine collaborative processing, controls the operation and maintenance interaction intelligent agent unit to perform collaborative processing with the assistance of the operation and maintenance system, and controls the inspection intelligent agent unit to poll to ensure closed-loop management for application anomaly recovery.

[0108] Figure 2 This is a schematic diagram of the structure of an application exception handling system based on intelligent agent orchestration according to an embodiment of the present invention.

[0109] exist Figure 2 The illustrated embodiment shows an application exception handling system based on agent orchestration, used to implement the application exception handling method based on agent orchestration as described above, including:

[0110] The intelligent agent orchestration module 201 is used to form an exception handling workflow including multiple orchestratable intelligent agent units in a visual manner for application exceptions of the application system through intelligent agent orchestration. The workflow engine of the exception handling workflow includes a control node and multiple task nodes.

[0111] The anomaly analysis module 202 is used to extract the features of the anomaly information by using the anomaly identification intelligent agent unit through anomaly analysis logic and delineate the scope of influence of the application anomaly through the entity relationship graph for the application anomaly information captured by the application system, so as to form an anomaly analysis result for the application anomaly information, and then output the anomaly analysis result to the front-end visualization interface or push it to the collaborative intelligent agent unit associated with the anomaly identification intelligent agent unit.

[0112] The collaborative handling module 203 is used to establish an intelligent agent collaborative handling system including the collaborative intelligent agent unit based on the anomaly analysis results, match the corresponding handling scheme for the application anomaly information according to the collaborative handling strategy, and handle the anomaly situation related to the application anomaly information.

[0113] According to an embodiment of a third aspect of the present invention, an application exception handling device based on agent orchestration is provided, comprising a memory and a processor, wherein the memory stores executable code, and the processor executes the executable code to implement the application exception handling method based on agent orchestration as described above.

[0114] One embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed in a computer, causes the computer to perform an application exception handling method based on intelligent agent orchestration as described in any embodiment of the specification.

[0115] The application anomaly handling method, system, and device based on intelligent agent orchestration provided by one or more embodiments of the present invention are mainly aimed at addressing the problem of inability to quickly locate anomalies in various application systems running on the network, and the lack of effective means to assist in resolving application anomalies. The application anomaly handling method, system, and device based on intelligent agent orchestration provided by one or more embodiments of the present invention utilize intelligent agent orchestration to enable multiple intelligent agent units to cooperate and combine with natural language processing toolkits (such as HANLP), open-source distributed search and analysis engines (such as Elasticsearch), large language models, and knowledge bases to achieve rapid analysis and location of application anomaly problems, and match and generate optimal solutions, thereby efficiently resolving application anomaly problems and ensuring high availability and business continuity of the system.

[0116] In this way, by orchestrating intelligent agents, the handling process of application anomalies can be broken down into the work of multiple selectable and orchestratable intelligent agent units. The workflow engine enables the dynamic combination of the handling process, and can be combined with auxiliary tools and technologies such as natural language processing toolkits (e.g., HANLP), open-source distributed search and analysis engines (e.g., Elasticsearch), and knowledge bases to improve the ability to identify application anomaly features and generate matching solutions / handling schemes. This can shorten the problem localization time and efficiently provide practical and effective solutions to application anomalies.

[0117] The purpose of this invention is to solve the problem of low efficiency in handling application anomalies in application systems. By using intelligent agent orchestration, combined with large models and auxiliary tools, it can quickly analyze anomalies and provide solutions, shortening the anomaly handling time and making application system anomaly handling more timely, accurate and clear.

[0118] By using an intelligent agent orchestration platform, various problem-solving logics for anomaly handling can be configured in a visual manner. Through anomaly identification intelligent agent units and multi-dimensional auxiliary means, the accuracy of anomaly identification, analysis and localization and the reliability of optimal solution generation can be improved. Combined with multi-agent unit collaborative technology, accurate solution delivery and closed-loop management of anomaly handling can be achieved.

[0119] The application exception handling method, system, and device based on intelligent agent orchestration provided by one or more embodiments of the present invention can provide at least one of the following beneficial effects:

[0120] 1) Intelligent agent orchestration makes exception handling more flexible: By adopting intelligent agent orchestration, the exception handling logic configuration is more flexible and can be adapted to different exception scenarios more quickly and effectively.

[0121] 2) Intelligent anomaly identification makes system anomaly location and handling / solution generation more accurate: Through multi-dimensional and multi-round anomaly identification and screening schemes, high-quality anomaly information identification and anomaly handling scheme matching are achieved, which can more clearly and reliably recommend anomaly handling schemes to effectively solve application anomaly problems.

[0122] 3) Multi-agent unit collaboration makes the system's anomaly closed-loop handling more complete: Through multi-agent unit collaboration, the anomaly intelligent analysis agent unit pushes the anomaly handling solution more accurately to the relevant parties, improving the efficiency of application anomaly handling and ensuring the success rate of application anomaly handling.

[0123] In summary, the application anomaly handling method, system, and device based on intelligent agent orchestration provided by one or more embodiments of the present invention can efficiently handle application anomaly problems.

[0124] It should be noted that the terminology used in the embodiments of this invention is for the purpose of describing particular embodiments only and is not intended to limit the invention. The singular forms “a,” “the,” and “the” used in the embodiments of this invention and the appended claims are also intended to include the plural forms, unless the context clearly indicates otherwise.

[0125] It should be understood that the term "and / or" used in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone. Additionally, the character " / " in this article generally indicates that the preceding and following related objects have an "or" relationship.

[0126] It is understood that the structures illustrated in the embodiments of the present invention do not constitute a specific limitation on the apparatus of the embodiments of the present invention. In other embodiments of the specification, the above-described apparatus may include more or fewer components than illustrated, or combine some components, or split some components, or have different component arrangements. The illustrated components may be implemented in hardware, software, or a combination of software and hardware.

[0127] The various embodiments in this invention are described in a progressive manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the device embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions of the method embodiments.

[0128] Those skilled in the art will recognize that, in one or more of the examples above, the functions described in this invention can be implemented using hardware, software, widgets, or any combination thereof. When implemented in software, these functions can be stored in a computer-readable medium or transmitted as one or more instructions or code on a computer-readable medium.

[0129] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made on the basis of the technical solution of the present invention should be included within the scope of protection of the present invention.

Claims

1. An application anomaly handling method based on agent orchestration, characterized in that, include: Through intelligent agent orchestration, for application system anomalies, an anomaly handling workflow consisting of multiple orchestratable intelligent agent units is formed in a visual manner. The workflow engine of the anomaly handling workflow includes a control node and multiple task nodes. For the application anomaly information of the captured application system, the anomaly identification intelligent agent unit extracts the features of the anomaly information through anomaly analysis logic and delimits the scope of influence of the application anomaly through entity relationship graph to form anomaly analysis results for the application anomaly information. Then, the anomaly analysis results are output to the front-end visualization interface or pushed to the collaborative intelligent agent unit associated with the anomaly identification intelligent agent unit. Based on the anomaly analysis results, an intelligent agent collaborative handling system is established, which includes the collaborative intelligent agent unit. According to the collaborative handling strategy, a corresponding handling scheme is matched for the application anomaly information, and the anomaly situation related to the application anomaly information is handled.

2. The application anomaly handling method based on agent orchestration according to claim 1, characterized in that, The steps for analyzing the captured application anomaly information include: The application system's exception logs are collected through a real-time capture layer, and regular expressions and data cleaning algorithms are used to standardize the format of unstructured information in the application exception information. A pre-trained language model is used to perform semantic understanding and intent recognition on the text of the application's anomaly information, and standardized anomaly type labels are output. Using natural language processing tools as word segmentation tools, core abnormal feature keywords are extracted from the application's abnormal information through part-of-speech tagging and named entity recognition technologies, and redundant noise is eliminated to construct an abnormal information feature vector.

3. The application anomaly handling method based on agent orchestration according to claim 2, characterized in that, The steps for analyzing the captured application anomaly information include: A real-time analysis engine is built based on an open-source distributed search and analysis engine, and the impact range of the application anomalies is visualized and delimited through an entity relationship graph.

4. The application anomaly handling method based on agent orchestration according to claim 1, characterized in that, The steps of the intelligent agent orchestration include: using the mxgraph graphics library to build a JavaScript drawing component to realize a visual orchestration tool and process engine for intelligent agent orchestration.

5. The application anomaly handling method based on agent orchestration according to claim 1, characterized in that, The steps for arranging intelligent agents include: arranging intelligent agents through drag-and-drop interaction.

6. The application anomaly handling method based on agent orchestration according to claim 1, characterized in that, The control nodes include: a start node that serves as the entry point for triggering a workflow, an end node that serves as the termination identifier for a workflow, and a gateway node that supports branching or aggregation logic control.

7. The application anomaly handling method based on agent orchestration according to claim 1, characterized in that, The task node includes at least one of the following: a form configuration component for structured data entry, an executor selector for third-party service calls, a JavaScript code editor for custom logic implementation, an atomic capability component for microservice interface encapsulation, a large model interaction component for artificial intelligence capability integration, and a barcode recognition component for physical world interaction interfaces.

8. The application anomaly handling method based on agent orchestration according to claim 1, characterized in that, The collaborative processing strategy includes at least one of the following: For application anomalies that the system can repair autonomously, the anomaly analysis results are sent to the automatic processing intelligent agent unit with preset automatic processing logic for automatic processing. For application anomalies that require manual intervention or tool assistance, the anomaly analysis results are sent to the human-computer interaction intelligent agent unit for human-computer collaborative processing. For complex application anomalies that lack online solutions, the anomaly analysis results are sent to the operation and maintenance interactive intelligent agent unit for collaborative processing with the assistance of the operation and maintenance system.

9. An application exception handling system based on intelligent agent orchestration, characterized in that, The method for implementing the application anomaly handling method based on agent orchestration according to any one of claims 1-8 includes: The intelligent agent orchestration module is used to form an exception handling workflow, which includes multiple orchestratable intelligent agent units, in a visual manner for application exceptions of the application system through intelligent agent orchestration. The workflow engine of the exception handling workflow includes a control node and multiple task nodes. An anomaly analysis module is used to extract features of the anomaly information captured by the application system using an anomaly identification intelligent agent unit through anomaly analysis logic and to define the scope of influence of the application anomaly through an entity relationship graph, so as to form an anomaly analysis result for the application anomaly information. Then, the anomaly analysis result is output to the front-end visualization interface or pushed to the collaborative intelligent agent unit associated with the anomaly identification intelligent agent unit. The collaborative handling module is used to establish an intelligent agent collaborative handling system including the collaborative intelligent agent unit based on the anomaly analysis results, match the corresponding handling scheme for the application anomaly information according to the collaborative handling strategy, and handle the anomaly situation related to the application anomaly information.

10. An application anomaly handling device based on intelligent agent orchestration, characterized in that, It includes a memory and a processor, wherein the memory stores executable code, and when the processor executes the executable code, it implements the application exception handling method based on intelligent agent orchestration according to any one of claims 1-8.

Images