Non-functional governance system and method for financial application system

Abstract

The invention provides a non-functional governance system and method for a financial application system, and the method comprises the steps: obtaining input information from any task agent in a task agent set through a knowledge service agent, and retrieving the non-functional governance related information of the input information; the index construction agent determines non-functional governance index information of the financial application system; the diagnosis detection agent determines a non-functional treatment diagnosis treatment strategy of the financial application system; the operation processing agent determines non-functional treatment task planning information of the financial application system; and the tool calling agent determines non-functional governance tool configuration information of the financial application system. According to the technical scheme, full-process automation of non-functional treatment from index design to problem handling is constructed through a multi-agent collaborative architecture, and the non-functional treatment efficiency and accuracy of a financial application system are improved.

Description

Technical field

[0001] The present application relates to the field of non-functional governance technology, and in particular to a non-functional governance system and method for financial application systems. [Background Technology]

[0002] With the rapid development of the digital economy and the widespread application of financial technology, the architecture of financial application systems is becoming increasingly complex. The widespread adoption of technologies like microservices and containerization has created unprecedented challenges for the non-functional governance of these systems. Non-functional governance encompasses multiple areas, including availability, maintainability, compatibility, and security. This wide scope and varying standards across these areas lead to a lack of standardization and consistency in governance solutions.

[0003] At the same time, the financial sector has extremely high requirements for system stability and requires real-time calculation of a large number of complex and related non-functional indicators, such as hundreds of indicators in the fields of business continuity, error management, and log management. Existing technical solutions have obvious deficiencies in indicator collection, real-time calculation, and efficient processing.

[0004] In addition, the handling of non-functional issues is highly dependent on manual experience, with low levels of automation and intelligence, making it difficult to meet the stringent requirements of financial services for timeliness and accuracy.

[0005] Therefore, how to achieve standardization, intelligence and automation of non-functional governance of financial application systems and improve the real-time calculation and processing efficiency of indicators has become a technical problem that needs to be solved urgently. [Summary of the invention]

[0006] The embodiments of the present application provide a non-functional governance system and method for financial application systems, aiming to solve the technical problem in related technologies that the non-functional governance capabilities and efficiency of financial application systems are difficult to meet actual needs.

[0007] In a first aspect, an embodiment of the present application provides a non-functional governance system for a financial application system, including:

[0008] a knowledge service agent, configured to obtain input information from any task agent in the task agent set, retrieve non-functional governance-related information of the input information, and feed back the non-functional governance-related information to the task agent that provides the input information;

[0009] The task agent set includes:

[0010] An indicator construction agent is used to send system-related data of the financial application system as input information to the knowledge service agent, and determine non-functional governance indicator information of the financial application system based on the feedback results of the knowledge service agent and a preset large language model;

[0011] a diagnostic detection agent, configured to send the non-functional indicator results corresponding to the non-functional governance indicator information as input information to the knowledge service agent, and determine a non-functional governance diagnostic disposal strategy for the financial application system based on the feedback results of the knowledge service agent and the large language model;

[0012] an operation processing agent, configured to send the non-functional governance diagnosis and disposal strategy as input information to the knowledge service agent, and determine non-functional governance task planning information for the financial application system based on feedback results from the knowledge service agent and the large language model;

[0013] The tool calling agent is used to send the non-functional governance task planning information as input information to the knowledge service agent, and determine the non-functional governance tool configuration information of the financial application system based on the feedback results of the knowledge service agent and the large language model, so as to control the actual application process of the non-functional governance tool based on the non-functional governance tool configuration information.

[0014] In one embodiment of the present application, optionally, the knowledge service agent is specifically used to:

[0015] Retrieving a vector search result corresponding to the input information in the financial vector data through a vector index;

[0016] Retrieving a graph search result corresponding to the input information in the financial graph data through a graph index;

[0017] Searching the financial keyword data and the financial summary data for keyword search results corresponding to the input information through a keyword index;

[0018] integrating the vector search results, the graph search results, and the keyword search results into a comprehensive search result according to a predetermined ratio;

[0019] The comprehensive search results are used as the non-functional governance related information and fed back to the task agent that provides the input information.

[0020] In one embodiment of the present application, optionally, the knowledge service agent is specifically used to:

[0021] Before the step of obtaining input information from any task agent in the task agent set, the original financial data is converted into the financial vector data, the financial graph data, the financial keyword data and the financial summary data respectively;

[0022] For any knowledge subject in the original financial data, adding the association relationship between the vector, graph nodes and edges, keywords and summary information corresponding to the knowledge subject to the financial vector data;

[0023] constructing the vector index based on the financial vector data;

[0024] Constructing the graph index based on the financial graph data;

[0025] The keyword index is constructed based on the financial keyword data and the financial summary data.

[0026] In one embodiment of the present application, optionally, the indicator construction agent is specifically used to:

[0027] Obtain system-related data of the financial application system and send the system-related data as input information to the knowledge service agent, and obtain a first search result of the knowledge service agent for the system-related data, and use the first search result as the input of a preset large language model to determine the non-functional governance indicator information of the financial application system through the large language model, wherein the non-functional governance indicator information includes: non-functional governance indicators and calculation strategies of the non-functional governance indicators.

[0028] In one embodiment of the present application, optionally, the diagnostic detection agent is specifically used to:

[0029] Obtain the non-functional governance indicator information from the indicator construction agent, call an external real-time computing tool to process the non-functional governance indicator information to obtain a non-functional indicator result, and send the non-functional indicator result as input information to the knowledge service agent, and obtain a second search result of the knowledge service agent for the non-functional indicator result, and use the second search result as the input of the large language model to determine the non-functional governance diagnosis and disposal strategy of the financial application system through the large language model.

[0030] In one embodiment of the present application, optionally, the operation processing agent is specifically used to:

[0031] Acquire the non-functional governance diagnosis and disposal strategy from the diagnostic detection agent, send the non-functional governance diagnosis and disposal strategy as input information to the knowledge service agent, and obtain the third search result of the knowledge service agent for the non-functional governance diagnosis and disposal strategy, and use the third search result as the input of the large language model to determine the non-functional governance task planning information of the financial application system through the large language model.

[0032] In one embodiment of the present application, optionally, the tool calling agent is specifically used to:

[0033] Obtain the non-functional governance task planning information from the operation processing agent, send the non-functional governance task planning information as input information to the knowledge service agent, and obtain the fourth search result of the knowledge service agent for the non-functional governance task planning information, and use the fourth search result as the input of the large language model to determine the non-functional governance tool configuration information of the financial application system through the large language model, wherein the non-functional governance tool configuration information includes: tool type, tool configuration parameters and tool status monitoring strategy.

[0034] In one embodiment of the present application, optionally, the system further includes:

[0035] The communication bus agent establishes a communication channel for any two agents in the task agent set and the knowledge service agent, wherein:

[0036] If the agent is a strategy planning agent, establishing an A2A protocol communication channel for the agent;

[0037] If the agent is a tool-using agent, an MCP protocol communication channel is established for the agent.

[0038] In a second aspect, an embodiment of the present application provides a non-functional governance method for a financial application system, the method being executed in the non-functional governance system for a financial application system described in any one of the first aspects above, the method comprising:

[0039] Obtain system-related data of financial application systems;

[0040] Searching a preset financial knowledge set based on the system-related data to obtain a first search result, wherein the financial knowledge set includes: financial vector data, financial graph data, financial keyword data, and financial summary data;

[0041] Using the first search result as input to a preset large language model, and determining non-functional governance indicator information of the financial application system through the large language model;

[0042] Calling an external real-time computing tool to process the non-functional governance indicator information to obtain a non-functional indicator result;

[0043] Searching the financial knowledge set based on the non-functional indicator result to obtain a second search result;

[0044] Using the second search result as input to the large language model, determining a non-functional governance diagnosis and disposal strategy for the financial application system through the large language model;

[0045] Searching the financial knowledge set based on the non-functional governance diagnosis and disposal strategy to obtain a third search result;

[0046] Using the third search result as input to the large language model, determining non-functional governance task planning information of the financial application system through the large language model;

[0047] Searching the financial knowledge set based on the non-functional governance task planning information to obtain a fourth search result;

[0048] Using the fourth search result as input to the large language model, determining non-functional governance tool configuration information of the financial application system through the large language model;

[0049] Based on the non-functional governance tool configuration information, the corresponding non-functional governance tool is called to perform non-functional governance on the financial application system.

[0050] In a third aspect, an embodiment of the present application provides a computer device comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are configured to execute the method described in the second aspect above.

[0051] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium storing computer-executable instructions, wherein the computer-executable instructions are used to execute the method described in the second aspect above.

[0052] The above technical solution addresses the technical problem that the non-functional governance capabilities and efficiency of financial application systems in related technologies cannot meet actual needs. By building an architecture system that collaborates with a knowledge service agent and multiple task agents, it achieves intelligent governance throughout the entire process, from indicator construction and problem diagnosis to task planning and tool invocation. The knowledge service agent serves as the core, integrating multi-dimensional financial data and providing intelligent retrieval services. The indicator construction agent automatically generates governance indicators, the diagnostic and detection agent performs real-time problem analysis, the operation processing agent intelligently orchestrates governance tasks, and the tool invocation agent precisely configures execution tools. This overall solution builds a complete non-functional governance closed-loop system through multi-agent collaboration and large language model empowerment, changes the traditional operation and maintenance model that relies on manual experience, and realizes the comprehensive intelligence and automation of non-functional governance of financial systems. It significantly improves governance efficiency and quality through knowledge sharing and intelligent decision-making. Its indicator design is more scientific, problem diagnosis is more accurate, and task execution is more efficient. At the same time, the system has strong adaptive capabilities and can dynamically respond to various complex scenarios, greatly reducing operation and maintenance costs, providing intelligent guarantees for the high availability, high reliability and security of financial business systems, and effectively solving the problems of knowledge dispersion, slow response, and inefficient execution in traditional governance methods.

Brief Description of the Drawings

[0053] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0054] Figure 1 A schematic diagram of a non-functional governance system for a financial application system according to an embodiment of the present application is shown;

[0055] Figure 2 A schematic diagram showing the interaction of intelligent agents in a non-functional governance system for a financial application system according to an embodiment of the present application is shown;

[0056] Figure 3 A system framework diagram of a non-functional governance system for a financial application system according to an embodiment of the present application is shown;

[0057] Figure 4 A flowchart of a non-functional governance method for a financial application system according to an embodiment of the present application is shown;

[0058] Figure 5 A block diagram of a computer device according to an embodiment of the present application is shown;

[0059] Figure 6 A block diagram of a computer device according to another embodiment of the present application is shown. [Specific implementation method]

[0060] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0061] Figure 1 A schematic diagram of a non-functional governance system for a financial application system according to an embodiment of the present application is shown.

[0062] like Figure 1 As shown, the embodiment of the present application provides a non-functional governance system for a financial application system, including:

[0063] The knowledge service agent is used to obtain input information from any task agent in the task agent set, retrieve non-functional governance related information of the input information, and feed back the non-functional governance related information to the task agent that provides the input information.

[0064] Non-functional governance refers to the standardized management and optimized assurance of non-business functional characteristics such as performance, reliability, and security of financial application systems. Its role is to ensure stable system operation, improve service quality, and meet the strict requirements of financial business for high availability and security compliance.

[0065] As the core data processing unit of non-functional governance, the knowledge service agent receives input information from each task agent and retrieves non-functional governance knowledge related to the input information. The input information includes key parameters such as system operation data and indicator calculation results. The non-functional governance related information is the knowledge content such as solutions, historical cases, etc. related to the input problems obtained through multi-dimensional retrieval.

[0066] This knowledge service agent integrates scattered non-functional governance experience into a structured knowledge system by establishing a unified knowledge retrieval and feedback mechanism, realizing centralized management and intelligent sharing of non-functional governance knowledge in the financial field, effectively solving the problems of knowledge dispersion and difficulty in reusing experience in traditional governance methods, providing accurate and reliable knowledge support for each task agent, significantly improving the scientific nature and efficiency of non-functional governance decision-making, and at the same time reducing the need for manual intervention through standardized knowledge services, laying a solid foundation for the automated governance of financial application systems.

[0067] In one possible design, the knowledge service agent is specifically used to: retrieve the vector retrieval results corresponding to the input information in the financial vector data through a vector index; retrieve the graph retrieval results corresponding to the input information in the financial graph data through a graph index; retrieve the keyword retrieval results corresponding to the input information in the financial keyword data and financial summary data through a keyword index; integrate the vector retrieval results, the graph retrieval results and the keyword retrieval results into a comprehensive retrieval result according to a predetermined ratio; and feed back the comprehensive retrieval result as the non-functional governance related information to the task agent that provides the input information.

[0068] The knowledge service agent retrieves financial data from different dimensions using vector indexing, graph indexing, and keyword indexing. It then integrates the search results into a comprehensive search result proportionally and feeds it back to the task agent. Optionally, the ratio is 3:3:4. Vector indexing is used to process similarity searches for high-dimensional data, graph indexing is used to explore relationships between data, and keyword indexing is used to quickly match text content. This multimodal retrieval mechanism comprehensively covers the characteristics of financial data, improving retrieval accuracy and efficiency. Through multi-dimensional data fusion, the task agent is provided with more comprehensive and accurate non-functional governance-related information, thereby enhancing the reliability and intelligence of subsequent decision-making.

[0069] Before obtaining input information from any task agent in the task agent set, the knowledge service agent is also used to: convert the original financial data into the financial vector data, the financial graph data, the financial keyword data and the financial summary data respectively; for any knowledge subject in the original financial data, add the association relationship between the vector, graph nodes and edges, keywords and summary information corresponding to the knowledge subject to the financial vector data; construct the vector index based on the financial vector data; construct the graph index based on the financial graph data; and construct the keyword index based on the financial keyword data and the financial summary data.

[0070] Before the knowledge service agent processes input information, it must convert raw financial data into vector data, graph data, keywords, and summary data, and establish various indexes. Raw financial data includes unstructured information such as system logs and design documents. Converting this data into structured data facilitates subsequent retrieval and analysis. Vector data captures semantic information through embedding models, graph data displays relationships between entities, and keywords and summaries provide a concise summary of the content. Data preprocessing and index construction provide a foundation for efficient retrieval for the knowledge service agent, ensuring data accessibility and consistency during non-functional governance processes, thereby supporting the agent's rapid response and decision-making.

[0071] The task agent set includes: indicator construction agent, diagnosis and detection agent, operation processing agent and tool calling agent.

[0072] Among them, the indicator construction agent is used to send the system-related data of the financial application system as input information to the knowledge service agent, and determine the non-functional governance indicator information of the financial application system based on the feedback results of the knowledge service agent and the preset large language model.

[0073] The indicator-building agent leverages system-related data and feedback from the knowledge-serving agent to generate non-functional governance indicator information through a large language model. System-related data includes, but is not limited to, performance logs and configuration information. The large language model understands complex context and generates structured output of non-functional governance indicator information, encompassing both indicator definitions and calculation strategies. This automated generation of non-functional governance indicators reduces manual intervention, ensures scientific and consistent indicator design, and improves the efficiency and coverage of indicator planning, laying the foundation for subsequent diagnosis and resolution.

[0074] Specifically, the indicator construction agent obtains the system-related data of the financial application system and sends the system-related data as input information to the knowledge service agent, and obtains the first search result of the knowledge service agent for the system-related data, and uses the first search result as the input of the preset large language model, and determines the non-functional governance indicator information of the financial application system through the large language model, wherein the non-functional governance indicator information includes: non-functional governance indicators and the calculation strategy of the non-functional governance indicators.

[0075] The indicator construction agent collects system-related data, such as operational logs and configuration parameters, from financial application systems. This raw data is then passed as input to the knowledge service agent for knowledge retrieval. The agent then obtains a first search result containing knowledge content such as historical cases and industry standards. This result is then analyzed and processed by a large language model, ultimately generating non-functional governance indicator information, including specific monitoring indicators and their calculation methods. The system-related data provides the foundational input for governance, while the knowledge service agent's search results supplement industry experience and knowledge. The large language model enables intelligent analysis of knowledge and indicator generation. This process automates and intelligently generates non-functional governance indicators, transforming the traditional indicator design process, which relied on manual experience, into an automated process based on big data and artificial intelligence. This not only significantly improves the efficiency and coverage of indicator design, but also dynamically adapts to the specific needs of different financial systems. Furthermore, the generated indicators and their calculation strategies are more scientific and targeted, providing a precise and reliable quantitative basis for subsequent non-functional governance, effectively addressing the challenges of traditional indicator design, such as high subjectivity, limited coverage, and delayed updates.

[0076] The diagnostic detection agent is used to send the non-functional indicator results corresponding to the non-functional governance indicator information as input information to the knowledge service agent, and determine the non-functional governance diagnostic disposal strategy of the financial application system based on the feedback results of the knowledge service agent and the large language model.

[0077] The diagnostic detection agent receives non-functional governance indicator information generated by the indicator construction agent, including quantitative data such as system performance and reliability. It then uses these indicator results as diagnostic basis and inputs them into the knowledge service agent for knowledge retrieval. It obtains feedback results containing diagnostic knowledge such as historical failure cases and best practices. Combined with the intelligent analysis capabilities of the large language model, it ultimately forms a non-functional governance diagnostic and disposal strategy that includes problem location and solutions. This process enables the automatic generation of intelligent diagnosis and disposal strategies for non-functional problems in the financial system, transforming the traditional analysis process that relies on manual experience into an automated process based on real-time data and artificial intelligence. This not only significantly improves the accuracy of problem diagnosis and disposal efficiency, but also dynamically adapts to various complex system scenarios. The generated diagnostic and disposal strategies are more scientific and operational, effectively solving the problems of traditional diagnostic methods such as slow response, narrow coverage, and reliance on personal experience, providing intelligent protection for the stable operation of the financial system.

[0078] Specifically, the diagnostic detection agent obtains the non-functional governance indicator information from the indicator construction agent, calls an external real-time computing tool to process the non-functional governance indicator information to obtain non-functional indicator results, and sends the non-functional indicator results as input information to the knowledge service agent, and obtains the second search result of the knowledge service agent for the non-functional indicator results, and uses the second search result as the input of the large language model, and determines the non-functional governance diagnosis and disposal strategy of the financial application system through the large language model.

[0079] In other words, the diagnostic and detection agent processes non-functional governance indicator information by invoking an external real-time computing tool, generates indicator results, and then combines the feedback from the knowledge service agent and the large language model to generate diagnostic and remediation strategies. External real-time computing tools can efficiently process streaming data, ensuring the timeliness of indicator calculations. This enables real-time detection and intelligent diagnosis of non-functional issues, quickly locating system anomalies and providing remediation recommendations, thereby shortening response time and improving system stability and reliability.

[0080] In non-functional governance systems in the financial sector, external real-time computing tools such as Apache Flink can be used to process high-throughput, low-latency streaming data, Spark Streaming is suitable for building scalable real-time analytics pipelines, and Kafka Streams provides lightweight processing capabilities for event-driven applications. These external real-time computing tools include but are not limited to the examples mentioned above. They utilize in-memory computing and distributed architectures to ensure fast computing power, providing real-time data support for diagnostic and detection agents.

[0081] The operation processing agent is used to send the non-functional governance diagnosis and disposal strategy as input information to the knowledge service agent, and determine the non-functional governance task planning information of the financial application system based on the feedback results of the knowledge service agent and the large language model.

[0082] The Operation Processing Agent receives the non-functional governance diagnosis and resolution strategy from the Diagnosis and Detection Agent. This strategy includes system anomaly analysis and suggested solutions. The Operation Processing Agent sends this strategy as a query request to the Knowledge Service Agent to obtain feedback such as historical O&M cases and best practices. The Operation Processing Agent then conducts in-depth analysis of the diagnosis strategy and knowledge feedback using a large language model, ultimately outputting non-functional governance task planning information with details such as task priority, execution order, and resource allocation. The diagnosis and resolution strategy provides the initial direction for problem resolution, the feedback from the Knowledge Service Agent supplements practical O&M experience, and the large language model enables strategy optimization and task orchestration. This process transforms the diagnosis results into an executable task sequence and achieves optimal allocation of O&M resources through intelligent planning. This addresses the inefficiency of task scheduling in traditional O&M, which relies on manual experience, ensuring the orderly and efficient implementation of non-functional governance work. Furthermore, by dynamically adjusting task planning to adapt to changes in system operating status, it significantly improves the O&M quality and responsiveness of the financial system.

[0083] Specifically, the operation processing agent obtains the non-functional governance diagnosis and disposal strategy from the diagnosis and detection agent, sends the non-functional governance diagnosis and disposal strategy as input information to the knowledge service agent, and obtains the third search result of the knowledge service agent for the non-functional governance diagnosis and disposal strategy, and uses the third search result as the input of the large language model to determine the non-functional governance task planning information of the financial application system through the large language model.

[0084] By receiving the non-functional governance diagnosis and disposal strategy provided by the diagnostic detection agent, which includes system anomaly analysis and remediation suggestions, and sending this strategy as a query condition to the knowledge service agent for knowledge retrieval, the system obtains a third-party search result containing historical operation and maintenance records and industry best practices. This result is then input into the large language model for intelligent analysis, ultimately generating non-functional governance task planning information that includes task execution sequence, resource allocation plan, and expected effect evaluation. The diagnostic disposal strategy clarifies the direction for addressing system issues, the third-party search results provided by the knowledge service agent supplement actual operation and maintenance experience, and the large language model implements strategy optimization and task orchestration. This process automatically converts diagnostic conclusions into implementable operation and maintenance plans. Through intelligent task planning, it solves the inefficiency caused by the reliance on manual experience in traditional operation and maintenance, ensures the scientific nature and enforceability of governance measures, and significantly improves the accuracy and responsiveness of financial system operation and maintenance by dynamically adjusting task priorities and resource allocation, providing intelligent support for the stable operation of the system.

[0085] The tool calling agent is used to send the non-functional governance task planning information as input information to the knowledge service agent, and determine the non-functional governance tool configuration information of the financial application system based on the feedback results of the knowledge service agent and the large language model, so as to control the actual application process of the non-functional governance tool based on the non-functional governance tool configuration information.

[0086] The Tool Invocation Agent receives the non-functional governance task planning information generated by the Operation Processing Agent. This information contains specific O&M tasks and execution requirements. The Tool Invocation Agent sends this information to the Knowledge Service Agent for tool matching and retrieval, obtaining feedback including applicable tool types and usage instructions. The Tool Invocation Agent then uses a large language model to analyze the task requirements and tool characteristics, ultimately generating non-functional governance tool configuration information including tool selection, parameter settings, and monitoring plans. The task planning information clarifies O&M requirements, the Knowledge Service Agent's feedback provides tool knowledge base support, and the Large Language Model implements intelligent tool matching and configuration optimization. This process enables intelligent invocation and precise configuration of O&M tools, resolving the traditional O&M issue of tool selection relying on manual experience and inefficient configuration. Automated tool matching and parameter setting ensure the efficient execution of governance measures, while real-time monitoring of tool operational status significantly improves the automation level and execution efficiency of financial system O&M, providing reliable assurance for system stability.

[0087] Specifically, the tool calling agent obtains the non-functional governance task planning information from the operation processing agent, sends the non-functional governance task planning information as input information to the knowledge service agent, and obtains the fourth search result of the knowledge service agent for the non-functional governance task planning information, and uses the fourth search result as the input of the large language model to determine the non-functional governance tool configuration information of the financial application system through the large language model, wherein the non-functional governance tool configuration information includes: tool type, tool configuration parameters and tool status monitoring strategy.

[0088] The tool-invoking agent receives non-functional governance task planning information from the operation processing agent, which contains specific O&M tasks and execution requirements. It then sends this information to the knowledge service agent for a tool matching query. This results in a fourth search result containing applicable tools and their usage instructions. This result is then fed into a large language model for in-depth analysis, ultimately generating non-functional governance tool configuration information that includes specific tool selection, parameter setting details, and an operation monitoring plan. The task planning information clarifies the specific O&M execution requirements. The knowledge service agent's fourth search result provides a rich tool knowledge base, and the large language model implements intelligent tool matching and parameter optimization. This process enables the precise invocation and automated configuration of financial system O&M tools. Through an intelligent tool selection mechanism and parameter optimization algorithm, it addresses the inaccurate tool matching and inefficient configuration issues inherent in traditional O&M. This ensures that each governance task utilizes the most appropriate tool and automatically sets optimal parameters. Furthermore, a built-in status monitoring strategy tracks tool performance in real time, significantly improving the execution accuracy and automation level of financial system O&M work, providing intelligent assurance for system stability and significantly reducing the need for manual intervention.

[0089] The non-functional governance system for financial application systems provided in the embodiment of the present application realizes intelligent governance of the entire process from indicator construction, problem diagnosis to task planning and tool calling by constructing an architecture system in which a knowledge service agent and multiple task agents work together. Its knowledge service agent serves as the core hub, integrating multi-dimensional financial data and providing intelligent retrieval services, the indicator construction agent automatically generates governance indicators, the diagnosis and detection agent realizes real-time problem analysis, the operation processing agent intelligently arranges governance tasks, and the tool calling agent accurately configures execution tools. This overall solution builds a complete non-functional governance closed-loop system through multi-agent collaboration and large language model empowerment, changes the traditional operation and maintenance mode that relies on manual experience, realizes the comprehensive intelligence and automation of non-functional governance of financial systems, and significantly improves governance efficiency and quality through knowledge sharing and intelligent decision-making. Its indicator design is more scientific, problem diagnosis is more accurate, and task execution is more efficient. At the same time, the system has strong adaptive capabilities and can dynamically respond to various complex scenarios, significantly reducing operation and maintenance costs, providing intelligent guarantees for the high availability, high reliability and security of financial business systems, and effectively solving the problems of knowledge dispersion, slow response, and inefficient execution in traditional governance methods.

[0090] In addition, the system also includes: a communication bus agent, which establishes a communication channel for any two agents in the task agent set and the knowledge service agent.

[0091] If the agent is a strategy planning agent, an A2A protocol (inter-agent communication protocol) communication channel is established for the agent.

[0092] If the agent is a tool-using agent, an MCP protocol (Model Context Protocol) communication channel is established for the agent.

[0093] Optionally, if the agent is both a strategy planning agent and a tool usage agent, both an A2A protocol communication channel and an MCP protocol communication channel can be established for it.

[0094] The communication bus agent, serving as the system's communication hub, intelligently identifies agent roles and establishes differentiated communication channels. The strategy planning agent uses the A2A protocol for efficient transmission of strategy information, while the tool usage agent employs the MCP protocol to ensure reliable tool invocation. The dual-role agent supports both protocol channels simultaneously. The A2A protocol optimizes the efficiency of strategy information exchange, while the MCP protocol ensures the stability of tool operation. The communication bus agent automatically adapts to different protocols, enabling seamless collaboration between heterogeneous agents.

[0095] As a result, a flexible and efficient intelligent agent communication network was built, which solved the interaction barriers caused by the lack of unified communication protocols in multi-agent systems, maintained the professionalism of the communication of each intelligent agent and achieved overall coordination. Through automatic protocol matching and channel isolation management, the system communication efficiency and reliability were significantly improved, providing stable and reliable underlying communication support for complex non-functional governance tasks, ensuring that various intelligent agents can adopt the optimal communication method according to their functional characteristics, and ultimately achieve efficient and coordinated operation of the entire governance system.

[0096] In one possible design, the system adopts a multi-agent architecture based on AIAgents (artificial intelligence agents), dividing agents into two categories: basic agents and task agents. Basic agents include knowledge service agents and communication bus agents: knowledge service agents build a professional knowledge service system by integrating business system operation data and historical non-functional governance field data, providing intelligent support such as strategy generation, task orchestration, data analysis, and tool invocation for upper-level task agents; communication bus agents are responsible for interaction and communication services among all agents. Task agents include indicator construction agents, diagnostic detection agents, operation processing agents, and tool execution agents, which are responsible for core functions such as non-functional indicator planning, real-time indicator calculation, operation task orchestration, and system operation changes.

[0097] This architectural design offers multiple advantages. The Knowledge Service Agent, through a standardized knowledge system, effectively addresses the diverse concepts and inconsistent standards inherent in traditional non-functional governance. The Indicator Construction Agent significantly enhances indicator data collection and computation capabilities. The Diagnostic Detection Agent, Operational Processing Agent, and Tool Execution Agent collectively enhance the system's automation and intelligence. To ensure stable system operation, manual experience can be refined into Standard Operating Procedures (SOPs) and pre-installed into the Task Agent. This effectively addresses issues such as dead loops and insufficient responses in complex task planning, as well as tool call anomalies caused by interface adaptation or network conditions.

[0098] Regarding the communication mechanism, a communication bus agent was designed. This component integrates the model context protocol and the inter-agent communication protocol, providing protocol conversion, mapping, and assembly capabilities. By decoupling communication interaction tasks from each agent and centralizing them, this not only improves communication efficiency but also ensures system reliability and scalability. This design allows each agent to focus on its core functionality while ensuring efficient and coordinated operation of the entire system.

[0099] Figure 2 A schematic diagram of agent interaction in a non-functional governance system for a financial application system according to an embodiment of the present application is shown.

[0100] like Figure 2As shown in the figure, in the architectural design of the prototype intelligent agent and its functional modules, the prototype intelligent agent contains three core functional modules, namely SOP retrieval, planning decomposition and action execution. SOP retrieval adopts the ReAct mechanism, planning decomposition includes three steps: Observe, Think and Act, and action execution covers three operations: Watch, Run and Publish.

[0101] The system also includes five major areas of intelligent agents. The operation processing intelligent agent is responsible for timeout monitoring, bandwidth control, and log cleaning. The change maintenance intelligent agent realizes flexible start and stop functions including grayscale release and traffic allocation. The diagnosis and detection intelligent agent performs data collection, indicator calculation and anomaly detection. The security control intelligent agent manages identity identification, security audit and data security. The compatibility adaptation intelligent agent handles compatibility issues of operating systems, middleware and mobile terminals. The platform intelligent agent includes the API governance intelligent agent responsible for API registration, API statistics and API scheduling. The process scheduling intelligent agent manages task planning, SOP retrieval and execution monitoring. The entire architecture demonstrates the core capability framework of the prototype intelligent agent and the specific division of responsibilities of each intelligent agent, forming a complete intelligent agent collaboration system.

[0102] In other words, based on their different functional positioning, the system has designed two basic types of agents: the Process Scheduling Agent is responsible for global task planning and process guidance, and the API Governance Agent implements unified management and control of all API interfaces. The system also deploys five specialized domain agents: the Operation Processing Agent focuses on executing system-level operations, the Change Maintenance Agent manages application deployment and traffic scheduling, the Diagnosis and Detection Agent implements data collection and anomaly detection, the Security Management Agent provides comprehensive security assurance, and the Compatibility and Adaptation Agent handles system compatibility issues. All domain agents obtain decision data through the API Governance Agent and receive task scheduling from the Process Scheduling Agent.

[0103] In terms of communication mechanism, the system adopts a collaborative model that combines message bus and SOP. As the core communication hub, the message bus significantly improves the scalability and flexibility of the system by decoupling the intelligent agent components. Its built-in message batch processing mechanism (including the entire process of message collection, merging, sending and result processing) effectively reduces system overhead. At the same time, instance-level flow control and circuit breaker mechanisms are introduced to ensure the high availability of communication links. The system also builds a professional SOP knowledge base, in which the SOP class describes the task process in the form of a state diagram, and the SOP object contains state transfer logic and intelligent agent routing strategy. Among them, the technical implementation includes: the State class encapsulates the intelligent agent state information, the sop._transit() function handles state migration, and the sop._route() function determines the execution subject. The two are uniformly scheduled through the sop.next() function to drive the efficient execution of the entire governance process in the main loop. This design not only ensures the standardization of system operation, but also maintains sufficient flexibility to adapt to various complex non-functional governance scenarios.

[0104] Figure 3 A system framework diagram of a non-functional governance system for a financial application system according to an embodiment of the present application is shown.

[0105] like Figure 3 As shown in the figure, the architecture of the non-functional governance system based on AIAgents is presented. The system consists of two major components: a toolbox and an agent. The toolbox provides four basic capability modules: real-time calculation, operation and maintenance changes, data monitoring, and system diagnosis. The core of the system is composed of six types of agents: the indicator construction agent is responsible for indicator services, indicator adjustments, indicator identification, indicator classification, and indicator construction SOPs; the diagnosis and detection agent has the functions of real-time calculation, threshold setting, diagnosis strategy, disposal strategy, and diagnosis and detection SOPs; the operation processing agent performs task planning, task maintenance, operation management, progress interaction, and operation processing SOPs; the tool execution agent manages tool catalogs, progress interaction, tool matching, scheduling control, and tool execution SOPs. All task agents are built based on the Agent Framework and integrated with LLM capabilities. A message queue is set up at the bottom of the system to realize asynchronous communication.

[0106] The basic agent layer comprises knowledge service agents and communication bus agents. The knowledge service agent provides knowledge construction, management, retrieval, and updating, while the communication bus agent is responsible for protocol conversion, assembly, management, and adaptation. These two basic agents, also based on the Agent Framework and LLM technologies, jointly support the collaborative operation of upper-layer agents, forming a complete non-functional governance solution.

[0107] This system adopts a layered architecture design to achieve the complete construction of the intelligent system. The specific implementation process includes three key stages.

[0108] First, the intelligent agent framework construction stage.

[0109] Based on the underlying API, the Agent Runner base class was expanded, core methods such as AgentState management, Memory storage, and runStep were restructured, and the inform interaction method was added to enable communication between agents. Furthermore, the AgentWorker engine was expanded, the Large Language Model (LLM) was configured, and the Tools calling mechanism was optimized, ultimately forming a complete Agent Framework agent development framework.

[0110] Second, the basic intelligent body construction stage.

[0111] Build a knowledge service agent and a communication bus agent. As the core knowledge hub, the knowledge service agent uses specialized data extraction devices to convert system operation data and design documents into structured knowledge, including vector data, graph data, and keyword summary information, and establishes a multi-dimensional indexing system. Its knowledge retrieval process uses a hybrid retrieval mechanism, combining vector indexing, graph indexing, and keyword indexing. It integrates search results in an optimized ratio of 3:3:4, ultimately generating accurate knowledge responses through a large language model.

[0112] The communication bus agent adopts a modular design, which includes three functional modules: agent role identification, protocol processing and communication interaction.

[0113] The agent role identification module is responsible for distinguishing and managing different types of agent roles. The system currently defines two basic roles: strategic planning agents and tool use agents. During initialization, each agent declares its role through the role field of the framework's inform method. The indicator construction agent and the operation processing agent are registered as strategic planning agents, the tool calling agent is registered as a tool use agent, and the diagnostic detection agent has dual roles. The knowledge service agent is also classified as a strategic planning agent.

[0114] The protocol handling module implements intelligent communication protocol management. It automatically selects the appropriate communication protocol based on the agent's role: Strategy planning agents establish communication channels using the A2A protocol, tool usage agents use the MCP protocol, and dual-role agents establish channels with both protocols simultaneously. The module also features protocol adaptation capabilities, intelligently integrating legacy communication protocols into existing systems and enabling automatic protocol conversion. All communication channels utilize a point-to-point, independent design and asynchronous communication via message queues to ensure efficient system communication.

[0115] The communication interaction module, as the final execution layer, completes the actual information exchange through the standard communication methods provided by the Agent Framework. This module encapsulates the underlying communication details and provides a unified execution interface for upper-layer role identification and protocol handling, achieving seamless integration of the entire communication process.

[0116] The system also incorporates a comprehensive task agent architecture, encompassing four core agents: indicator generation, diagnostic testing, operational processing, and tool invocation. All task agents are developed based on the unified AgentFramework framework, with the sopInvoke method specifically extended to integrate various Standard Operating Procedures (SOPs). These SOPs, refined through human experience, provide reliable exception handling for the system.

[0117] First, we built an indicator construction agent. Leveraging the intelligent service mechanism of the underlying LLM and referencing existing indicator design patterns, we targeted financial application systems requiring real-time non-functional governance. This agent converted operational and design data from the financial application system into knowledge data through a knowledge service agent. The indicator construction agent interacted with the communication bus agent, which then conducted knowledge retrieval and interaction via the A2A protocol. This knowledge data was then fed back layer by layer to the indicator construction agent. The relevant knowledge data then entered the underlying LLM model of the indicator construction agent, which then automatically developed the system's indicator list and calculation methods. This information was then shared with the diagnosis and detection agent via the communication bus agent.

[0118] Second, a diagnostic and detection agent is constructed. This agent receives the list of non-functional governance indicators and invokes an external real-time computing tool based on the MCP protocol. The relevant indicator calculation tasks and methods are then dispatched to the tool. The real-time computing tool ensures timely calculations by combining streaming and in-memory computing. This tool calculates non-functional indicator results in real time based on system operation data, indicator calculation methods, and indicator calculation tasks. Based on the A2A protocol, the communication bus agent establishes an access relationship with the knowledge service agent, constructing indicator result data into questions. Knowledge related to governance strategies such as diagnosis and disposal is retrieved. Once the relevant knowledge is returned, the LLM of the operation processing agent is invoked to generate result data, i.e., a complete diagnostic and disposal strategy. The relevant strategy data is transmitted to the operation processing agent via the A2A protocol channel with the communication bus agent.

[0119] Third, the Operation Processing Agent is constructed. This agent receives the non-functional governance diagnosis and disposal strategies. Based on the A2A protocol, it establishes an access relationship with the Knowledge Service Agent through the Communication Bus Agent. It constructs the policy data into questions and retrieves relevant knowledge for governance task planning. Once the relevant knowledge is returned, it calls the Operation Processing Agent's LLM to generate result data, namely, a complete task plan and task orchestration. This task plan and orchestration data is then passed to the Tool Execution Agent via the A2A protocol channel with the Communication Bus Agent.

[0120] Fourth, build a tool execution agent. This receives task planning and task orchestration data from non-functional governance, establishes an access relationship with the knowledge service agent through the communication bus agent based on the A2A protocol, constructs task data into questions, retrieves tool-related knowledge, and matches tasks with tools based on the returned results. Tool invocation, parameter transfer, and status monitoring are completed between tools through the MCP collaborative channel.

[0121] In summary, the system proposed in the present invention realizes the intelligent upgrade of the non-functional governance capabilities of financial application systems. Expert experience is converted into a structured knowledge base through the knowledge service agent to ensure the standardization and consistency of the governance plan; its communication bus agent integrates the MCP and A2A dual-protocol communication mechanisms to achieve efficient and reliable interaction between agents; the four core task agents cooperate with the standard operating procedure SOP to build a complete intelligent collaborative governance system. The system realizes the automation of the entire process from intelligent indicator construction, intelligent strategy generation to intelligent task planning and intelligent tool calling, which changes the traditional governance model that relies on manual experience, improves the governance efficiency of the financial system in terms of performance, reliability, security, etc., and provides intelligent guarantees for the stable operation of financial business.

[0122] The technical solution of this application is further described in detail below through examples in specific scenarios.

[0123] In the actual scenario of non-functional governance of business system connection status, the indicator-building agent, based on the actual system operation, establishes a connection with the knowledge service agent through the communication bus agent to obtain underlying knowledge support. The indicator-building agent constructs key indicators such as connection establishment time, monthly connection stability service efficiency ratio, and quarterly connection stability service efficiency ratio. The monthly connection stability service efficiency ratio = actual stable service time (minutes) / monthly required service time (minutes), and the connection establishment time (seconds) = connection success timestamp - connection initiation timestamp.

[0124] Next, the diagnostic and detection agent calculates the indicator results in real time based on the indicator items provided by the indicator construction agent. The diagnostic and detection agent interacts with the knowledge service agent via the communication bus agent to obtain underlying knowledge support. The diagnostic and detection agent diagnoses the system's operating status based on the real-time indicator results from the historical knowledge base and retrieves and generates corresponding action strategies for identified connection anomalies.

[0125] Furthermore, the Operation Processing Agent, based on the handling strategy provided by the Diagnosis and Detection Agent, communicates with the Knowledge Service Agent via the Communication Bus Agent to obtain underlying knowledge support. The Operation Processing Agent converts the handling strategy for connection anomalies into specific handling steps and task orchestration plans.

[0126] Ultimately, the tool execution agent, based on the task orchestration plan generated by the operations processing agent, precisely calls the appropriate operations and maintenance tools to perform specific operations. This ensures that every governance measure is accurately implemented.

[0127] In the actual scenario of non-functional management of CPU operating efficiency,

[0128] The indicator-building agent, based on the actual system operation, establishes a connection with the knowledge-service agent through the communication bus agent to obtain underlying knowledge support. The indicator-building agent constructs key indicators such as CPU utilization, CPU full load rate, and CPU operation curve. The CPU full load rate is calculated by dividing the daily percentage of time when the CPU utilization is 100% by the total daily CPU operation time.

[0129] Next, the diagnostic and detection agent calculates the indicator results in real time based on the indicator items provided by the indicator construction agent. The diagnostic and detection agent interacts with the knowledge service agent via the communication bus agent to obtain underlying knowledge support. The diagnostic and detection agent diagnoses the CPU operating status based on the real-time indicator results from the historical knowledge base. For identified CPU anomalies, it retrieves and generates corresponding action strategies, including specific measures such as increasing CPU resource supply and shutting down ineffective services.

[0130] Furthermore, the Operation Processing Agent, based on the handling strategy provided by the Diagnosis and Detection Agent, communicates with the Knowledge Service Agent via the Communication Bus Agent to obtain underlying knowledge support. The Operation Processing Agent converts the handling strategy for CPU anomalies into specific handling steps and task scheduling solutions.

[0131] Ultimately, the Tool Execution Agent, based on the task orchestration plan generated by the Operation Processing Agent, precisely invokes the appropriate O&M tools to execute specific actions. The specific measures implemented by the Tool Execution Agent include: first, expanding CPU resource supply; second, migrating services to new CPU resources; and third, decommissioning existing services that are consuming CPU resources. The Tool Execution Agent ensures that each governance measure is accurately implemented.

[0132] The technical solution of this application realizes the intelligent management of the connection status of the business system and the CPU operating efficiency through the collaboration of multiple agents. In the connection management scenario, the system builds a connection timeliness indicator, diagnoses anomalies in real time and generates a disposal plan; in the CPU management scenario, it establishes a CPU load indicator, intelligently identifies anomalies and performs resource allocation. Both cases follow the closed-loop process of "indicator construction-real-time diagnosis-task scheduling-precise execution", providing decision support through the knowledge service agent, ensuring smooth interaction through the communication bus agent, and the four major task agents working together to realize the automation of the entire process from problem discovery to disposal, significantly improving the system stability and operation and maintenance efficiency.

[0133] Of course, the actual scenarios in which the technical solution of this application can be applied include but are not limited to the above examples, and can also be actual scenarios that any financial application system may face, such as transaction concurrency management, interface timeout management, message queue backlog management, cache hit rate management, service avalanche protection management, database deadlock management, file transfer stability management, third-party service call management, system resource leakage management, security vulnerability protection management, etc.

[0134] Figure 4 A flowchart of a non-functional governance method for a financial application system according to an embodiment of the present application is shown.

[0135] like Figure 4 As shown, the method includes:

[0136] Step 402: Acquire system-related data of the financial application system.

[0137] Step 404: Search a preset financial knowledge set based on the system-related data to obtain a first search result, wherein the financial knowledge set includes: financial vector data, financial graph data, financial keyword data, and financial summary data.

[0138] Step 406: Using the first search result as input to a preset large language model, the non-functional governance indicator information of the financial application system is determined by the large language model.

[0139] Step 408: Call an external real-time computing tool to process the non-functional governance indicator information to obtain a non-functional indicator result.

[0140] Step 410: Search the financial knowledge set based on the non-functional indicator result to obtain a second search result.

[0141] Step 412: Using the second search result as input to the large language model, a non-functional governance diagnosis and disposal strategy for the financial application system is determined by the large language model.

[0142] Step 414 : Search the financial knowledge set based on the non-functional governance diagnosis and disposal strategy to obtain a third search result.

[0143] Step 416: Using the third search result as input to the large language model, the non-functional governance task planning information of the financial application system is determined by the large language model.

[0144] Step 418: Search the financial knowledge set based on the non-functional governance task planning information to obtain a fourth search result.

[0145] Step 420: Using the fourth search result as input to the large language model, the non-functional governance tool configuration information of the financial application system is determined by the large language model.

[0146] Step 422: Based on the non-functional governance tool configuration information, call the corresponding non-functional governance tool to perform non-functional governance on the financial application system.

[0147] This method uses any one of the solutions in the above embodiments, and therefore has all the above technical effects, which will not be repeated here.

[0148] In addition, in one embodiment, the present application provides a computer device, which may be a server, and its internal structure diagram may be as follows: Figure 5 As shown. The computer device includes a processor, a memory, a network interface and a database connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile and / or volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external client via a network connection. When the computer program is executed by the processor, it can implement the information processing method described in any of the above embodiments.

[0149] In one embodiment, the present application further provides a computer device, which may be a client, and its internal structure diagram may be as follows: Figure 6As shown. The computer device includes a processor, memory, network interface, display screen, and input device connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external server via a network connection. When executed by the processor, the computer program can implement the information processing method described in any of the above embodiments.

[0150] Any of the aforementioned computer devices in the embodiments of the present application may exist in various forms, including but not limited to:

[0151] (1) Mobile communication devices: These devices are characterized by their mobile communication capabilities and are primarily designed to provide voice and data communications. These terminals include smartphones (e.g., iPhones), multimedia phones, feature phones, and low-end phones.

[0152] (2) Ultra-mobile personal computer devices: These devices fall under the category of personal computers, have computing and processing capabilities, and generally also have mobile Internet access. These terminals include PDAs, MIDs, and UMPCs, such as the iPad.

[0153] (3) Portable entertainment devices: These devices can display and play multimedia content. These devices include audio and video players (such as iPods), handheld game consoles, e-books, as well as smart toys, wearable devices, and portable car navigation devices.

[0154] (4) Server: A device that provides computing services. The server consists of a processor, hard disk, memory, system bus, etc. The server is similar to a general computer architecture, but because it needs to provide highly reliable services, it has higher requirements in terms of processing power, stability, reliability, security, scalability, and manageability.

[0155] (5) Other electronic devices with data interaction functions.

[0156] In addition, an embodiment of the present application provides a computer-readable storage medium storing computer-executable instructions, wherein the computer-executable instructions are used to perform the following steps:

[0157] Obtain system-related data of financial application systems;

[0158] Searching a preset financial knowledge set based on the system-related data to obtain a first search result, wherein the financial knowledge set includes: financial vector data, financial graph data, financial keyword data, and financial summary data;

[0159] Using the first search result as input to a preset large language model, and determining non-functional governance indicator information of the financial application system through the large language model;

[0160] Calling an external real-time computing tool to process the non-functional governance indicator information to obtain a non-functional indicator result;

[0161] Searching the financial knowledge set based on the non-functional indicator result to obtain a second search result;

[0162] Using the second search result as input to the large language model, determining a non-functional governance diagnosis and disposal strategy for the financial application system through the large language model;

[0163] Searching the financial knowledge set based on the non-functional governance diagnosis and disposal strategy to obtain a third search result;

[0164] Using the third search result as input to the large language model, determining non-functional governance task planning information of the financial application system through the large language model;

[0165] Searching the financial knowledge set based on the non-functional governance task planning information to obtain a fourth search result;

[0166] Using the fourth search result as input to the large language model, determining non-functional governance tool configuration information of the financial application system through the large language model;

[0167] Based on the non-functional governance tool configuration information, the corresponding non-functional governance tool is called to perform non-functional governance on the financial application system.

[0168] It should be noted that the above functions or steps that can be implemented by the computer-readable storage medium or computer device can refer to the relevant description in the aforementioned method embodiment. To avoid repetition, they will not be described one by one here.

[0169] The above describes in detail the technical solution of the present application in combination with the accompanying drawings. Through the technical solution of the present application, a complete non-functional governance closed-loop system is constructed, which changes the traditional operation and maintenance mode that relies on manual experience, realizes the comprehensive intelligence and automation of non-functional governance of the financial system, and significantly improves the governance efficiency and quality through knowledge sharing and intelligent decision-making. Its indicator design is more scientific, problem diagnosis is more accurate, and task execution is more efficient. At the same time, the system has strong adaptive capabilities and can dynamically respond to various complex scenarios, greatly reducing operation and maintenance costs, providing intelligent guarantees for the high availability, high reliability and security of financial business systems, and effectively solving the problems of knowledge dispersion, slow response, and inefficient execution in traditional governance methods.

[0170] The word "if," as used herein, may be interpreted as "at the time of" or "when" or "in response to determining" or "in response to detecting," depending on the context. Similarly, the phrases "if it is determined" or "if (stated condition or event) is detected" may be interpreted as "when it is determined" or "in response to the determination" or "when detecting (stated condition or event)" or "in response to detecting (stated condition or event)," depending on the context.

[0171] The terms used in the embodiments of the present application are for the purpose of describing specific embodiments only and are not intended to limit the present application. The singular forms "a", "an", "the" and "the" used in the embodiments of the present application and the appended claims are also intended to include plural forms unless the context clearly indicates otherwise.

[0172] In the several embodiments provided in this application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are merely schematic. For example, the division of the units is merely a logical function division. In actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed may be through some interface, indirect coupling or communication connection of the device or unit, which may be electrical, mechanical or other forms.

[0173] In addition, the functional units in the various embodiments of the present application may be integrated into a single processing unit, or each unit may exist physically separately, or two or more units may be integrated into a single unit. The aforementioned integrated units may be implemented in the form of hardware or in the form of hardware plus software functional units.

[0174] Those skilled in the art will appreciate that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0175] The embodiments described above are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention, and should all be included in the scope of protection of the present invention.

Claims

1. A non-functional management system for financial application systems, characterized in that: include: a knowledge service agent, configured to obtain input information from any task agent in the task agent set, retrieve non-functional governance-related information of the input information, and feed back the non-functional governance-related information to the task agent that provides the input information; The task agent set includes: An indicator construction agent is used to send system-related data of the financial application system as input information to the knowledge service agent, and determine non-functional governance indicator information of the financial application system based on the feedback results of the knowledge service agent and a preset large language model; a diagnostic detection agent, configured to send the non-functional indicator results corresponding to the non-functional governance indicator information as input information to the knowledge service agent, and determine a non-functional governance diagnostic disposal strategy for the financial application system based on the feedback results of the knowledge service agent and the large language model; an operation processing agent, configured to send the non-functional governance diagnosis and disposal strategy as input information to the knowledge service agent, and determine non-functional governance task planning information for the financial application system based on feedback results from the knowledge service agent and the large language model; The tool calling agent is used to send the non-functional governance task planning information as input information to the knowledge service agent, and determine the non-functional governance tool configuration information of the financial application system based on the feedback results of the knowledge service agent and the large language model, so as to control the actual application process of the non-functional governance tool based on the non-functional governance tool configuration information.

2. The method according to claim 1, characterized in that The knowledge service agent is specifically used for: Retrieving a vector search result corresponding to the input information in the financial vector data through a vector index; Retrieving a graph search result corresponding to the input information in the financial graph data through a graph index; Searching the financial keyword data and the financial summary data for keyword search results corresponding to the input information through a keyword index; integrating the vector search results, the graph search results, and the keyword search results into a comprehensive search result according to a predetermined ratio; The comprehensive search results are used as the non-functional governance related information and fed back to the task agent that provides the input information.

3. The method according to claim 2, characterized in that The knowledge service agent is specifically used for: Before the step of obtaining input information from any task agent in the task agent set, the original financial data is converted into the financial vector data, the financial graph data, the financial keyword data and the financial summary data respectively; For any knowledge subject in the original financial data, adding the association relationship between the vector, graph nodes and edges, keywords and summary information corresponding to the knowledge subject to the financial vector data; constructing the vector index based on the financial vector data; Constructing the graph index based on the financial graph data; The keyword index is constructed based on the financial keyword data and the financial summary data.

4. The method according to claim 2, characterized in that The indicator-based intelligent agent is specifically used to: Obtain system-related data of the financial application system and send the system-related data as input information to the knowledge service agent, and obtain a first search result of the knowledge service agent for the system-related data, and use the first search result as the input of a preset large language model to determine the non-functional governance indicator information of the financial application system through the large language model, wherein the non-functional governance indicator information includes: non-functional governance indicators and calculation strategies of the non-functional governance indicators.

5. The method according to claim 4, characterized in that The diagnostic detection agent is specifically used for: Obtain the non-functional governance indicator information from the indicator construction agent, call an external real-time computing tool to process the non-functional governance indicator information to obtain a non-functional indicator result, and send the non-functional indicator result as input information to the knowledge service agent, and obtain a second search result of the knowledge service agent for the non-functional indicator result, and use the second search result as the input of the large language model to determine the non-functional governance diagnosis and disposal strategy of the financial application system through the large language model.

6. The method according to claim 5, characterized in that The operation processing agent is specifically used to: Acquire the non-functional governance diagnosis and disposal strategy from the diagnostic detection agent, send the non-functional governance diagnosis and disposal strategy as input information to the knowledge service agent, and obtain the third search result of the knowledge service agent for the non-functional governance diagnosis and disposal strategy, and use the third search result as the input of the large language model to determine the non-functional governance task planning information of the financial application system through the large language model.

7. The method according to claim 6, characterized in that The tool calling agent is specifically used to: Obtain the non-functional governance task planning information from the operation processing agent, send the non-functional governance task planning information as input information to the knowledge service agent, and obtain the fourth search result of the knowledge service agent for the non-functional governance task planning information, and use the fourth search result as the input of the large language model to determine the non-functional governance tool configuration information of the financial application system through the large language model, wherein the non-functional governance tool configuration information includes: tool type, tool configuration parameters and tool status monitoring strategy.

8. The method according to any one of claims 1 to 7, characterized in that Also includes: The communication bus agent establishes a communication channel for any two agents in the task agent set and the knowledge service agent, wherein: If the agent is a strategy planning agent, establishing an A2A protocol communication channel for the agent; If the agent is a tool-using agent, an MCP protocol communication channel is established for the agent.

9. A non-functional governance method for a financial application system, characterized in that: The method implemented in the non-functional governance system for a financial application system according to any one of claims 1 to 8 comprises: Obtain system-related data of financial application systems; Searching a preset financial knowledge set based on the system-related data to obtain a first search result, wherein the financial knowledge set includes: financial vector data, financial graph data, financial keyword data, and financial summary data; Using the first search result as input to a preset large language model, and determining non-functional governance indicator information of the financial application system through the large language model; Calling an external real-time computing tool to process the non-functional governance indicator information to obtain a non-functional indicator result; Searching the financial knowledge set based on the non-functional indicator result to obtain a second search result; Using the second search result as input to the large language model, determining a non-functional governance diagnosis and disposal strategy for the financial application system through the large language model; Searching the financial knowledge set based on the non-functional governance diagnosis and disposal strategy to obtain a third search result; Using the third search result as input to the large language model, determining non-functional governance task planning information of the financial application system through the large language model; Searching the financial knowledge set based on the non-functional governance task planning information to obtain a fourth search result; Using the fourth search result as input to the large language model, determining non-functional governance tool configuration information of the financial application system through the large language model; Based on the non-functional governance tool configuration information, the corresponding non-functional governance tool is called to perform non-functional governance on the financial application system.

10. A computer device, characterized in that: include: at least one processor; and, a memory communicatively coupled to the at least one processor; The memory stores instructions that can be executed by the at least one processor, and the instructions are configured to enable the processor to execute the method of claim 9.

Images