Atlas processing method, task processing method, electronic device, storage medium, and program product

Abstract

Embodiments of the present application provide a graph processing method, a task processing method, an electronic device, a storage medium and a program product. In the graph structure of the unified modeling graph, a new type of operation manual set node is introduced. The definition file of the operation manual set of the operation manual set node describes the operation capability information required for the observable entity to perform task processing, and through the operation manual link information, the association relationship between the operation manual set node and the entity set node or the data set node is established in the unified modeling graph, so that the operation capability is accurately and dynamically bound to the corresponding observable entity or observable data. In this way, the unified modeling graph can realize deep collaboration with the task execution system, provide a closed-loop mechanism from "state perception" to "task execution", support task processing capability for observable entities, and support intelligent active governance needs of cloud computing systems.

Description

Technical Field

[0001] This application relates to the field of cloud computing technology, and in particular to a graph processing method, a task processing method, an electronic device, a storage medium, and a program product. Background Technology

[0002] In the rapid evolution of cloud-native architectures, observability has become a core capability to ensure the stable operation of cloud computing systems. Currently, observability management platforms construct a unified modeling graph of cloud computing systems through unified modeling technology. This graph structure manages observable entities, datasets, and their relationships, providing a structured foundation for comprehensive perception of the system state of cloud computing systems.

[0003] However, the observability value of existing unified modeling graphs remains relatively limited. Their capabilities are primarily confined to data collection and storage of observable entities, making it difficult to fully support collaboration with task execution systems such as automated operation and maintenance, fault self-healing, and elastic scaling. Due to the lack of a closed-loop mechanism from "state awareness" to "task execution," existing unified modeling graphs cannot support task processing capabilities for observable entities, thus failing to meet the intelligent proactive governance needs of cloud computing systems. Summary of the Invention

[0004] This application provides a graph processing method, a task processing method, an electronic device, a storage medium, and a program product, which enable the unified modeling graph to achieve deep collaboration with the task execution system and support task processing capabilities for observable entities.

[0005] This application provides a graph processing method applied to an observability management platform. The observability management platform is used to manage the observability of at least one observable entity in a cloud computing system. The method includes: acquiring an initial unified modeling graph, wherein the initial unified modeling graph organizes multiple regular graph nodes in a graph structure, and regular graph nodes with relationships are connected by edges. The types of regular graph nodes include at least one of the following: entity set nodes and dataset nodes. Entity set nodes represent entity sets describing a class of observable entities, and dataset nodes represent observable data sets of observable entities. In response to configuration operations on a graphical interface, the method configures the definition file and operation manual links for an operation manual set. The information includes: the definition file of the operation manual set, which describes the operational capabilities required to perform tasks on observable entities; and the operation manual link information, which defines the association between the operation manual set and the entity set or observable data set. Based on the definition file of the operation manual set, operation manual set nodes are created in the initial unified modeling graph, and their node attribute information is configured. The node attribute information of the operation manual set nodes includes the definition file of the operation manual set. Based on the operation manual link information of the operation manual set nodes, edges connecting the operation manual set nodes and their associated regular graph nodes are created in the initial unified modeling graph, and the edge attribute information of the edges connected to the operation manual set nodes is configured to obtain the unified modeling graph.

[0006] This application also provides a task processing method applied to an observability management platform. The observability management platform is used to perform observability management on at least one observable entity in a cloud computing system. The method includes: receiving a capability discovery request sent by a task agent system. The capability discovery request is used to request the acquisition of operational capability information required to perform task processing on a target observable entity, and the capability discovery request carries entity context information of the target observable entity; obtaining the definition file of at least one level of operation manual set matching the entity context information from a pre-constructed unified modeling graph. The unified modeling graph is constructed according to a graph processing method, and the scope of operation manual sets at different levels is different; obtaining operational capability information according to the definition file of the operation manual set at least one level. The operational capability information includes a tool list, a skill list, and a knowledge list corresponding to the target observable entity; and returning the operational capability information to the task agent system so that the task agent system can perform task processing on the target observable entity according to the operational capability information.

[0007] This application also provides an electronic device, including: a memory and a processor; the memory for storing a computer program; the processor coupled to the memory for executing the computer program to perform steps in a map processing method or task processing method.

[0008] This application also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, enables the processor to implement the steps in a graph processing method or task processing method.

[0009] This application also provides a computer program product, including a computer program / instructions, which, when executed by a processor, enable the processor to implement the steps in a graph processing method or task processing method.

[0010] The technical solution provided in this application introduces a novel operation manual set node into the graph structure of the unified modeling graph. The operation manual set node describes the operational capability information required for observable entities to perform task processing through its definition file. By linking operation manuals, an association is established between the operation manual set node and entity set nodes or dataset nodes in the unified modeling graph. This allows for precise and dynamic binding of operational capabilities to corresponding observable entities or observable data. In this way, the unified modeling graph can achieve deep collaboration with the task execution system, providing a closed-loop mechanism from "state awareness" to "task execution," supporting task processing capabilities oriented towards observable entities. Based on the hierarchical structure of the unified modeling graph, it can achieve layered organization and on-demand supply of operational capabilities, improving decision-making accuracy and task execution efficiency in complex cloud environments, and supporting the intelligent proactive governance needs of cloud computing systems. Attached Figure Description

[0011] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings: Figure 1 A flowchart of a spectrum processing method provided in an embodiment of this application; Figure 2 A flowchart illustrating a task processing method provided in an embodiment of this application; Figure 3 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0012] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below in conjunction with specific embodiments and corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0013] The following is a description of some of the terms used in this application: Cloud computing service systems refer to software systems that rely on cloud infrastructure to provide users with computing, storage, data, and application capabilities in a platform-based manner. Examples of cloud computing service systems include, but are not limited to, database systems, cloud server systems, and e-commerce service systems.

[0014] Observability refers to the ability to infer the internal state, behavior, and root causes of failures from telemetry data such as metrics, logs, traces, or events generated by a cloud computing service system during runtime.

[0015] An Observability Management Platform (ODP) is an intelligent operation and maintenance infrastructure for cloud computing service systems. It collects, manages, stores, and analyzes multi-source telemetry data such as logs, metrics, and link tracing in a unified manner to build a full-link context association centered on observable entities. On this basis, it provides visualization insights, intelligent root cause diagnosis, and cross-team collaborative handling capabilities, transforming discrete telemetry data into executable decisions that drive system stability assurance and performance optimization.

[0016] An observable entity refers to a resource object in a cloud computing service system that can be detected, analyzed, and manipulated. Resource objects include, but are not limited to, specific hosts, specific containers, specific processes, service instances, microservice instances, or database instances. The entity category to which an observable entity belongs (also referred to as the object category of a resource object) includes, but is not limited to, hosts, containers, processes, services, and databases.

[0017] Entity Set: A collection of metadata shared by observable entities of the same entity class. For example, an entity set may include primary keys, attribute data, state data, etc., of observable entities of the same entity class.

[0018] An observable dataset (Telemetry Data Set) refers to a collection of multiple observable data sets. Observable data includes, but is not limited to, metrics, logs, traces, and events. Metrics include, but are not limited to, CPU usage, request latency, QPS (Queries Per Second / RPS), and memory usage. Logs include, but are not limited to, error logs, debug logs, and access logs. Trace data is used to trace the complete call path of a single request in a distributed system, helping to identify performance bottlenecks in the call chain. Event data includes, but is not limited to, service launch events, configuration update events, and alarm trigger events. Different observable data categories can correspond to different observable datasets. Observable datasets can be categorized as metric sets, log sets, trace sets, and event sets.

[0019] Unified Modeling (UModel) technology is a data and capability fusion modeling technology for observability management platforms. It takes a graph-based structure as its core and uses nodes and links to uniformly abstract, organize and associate the core elements in the observable domain. For example, it uses a graph structure to uniformly abstract, organize and associate entity sets, observable data sets and their relationships.

[0020] In the rapid evolution of cloud-native architectures, observability has become a core capability to ensure the stable operation of cloud computing systems. Currently, observability management platforms construct a unified modeling graph of cloud computing systems through unified modeling technology. This graph structure manages observable entities, datasets, and their relationships, providing a structured foundation for comprehensive perception of the system state of cloud computing systems.

[0021] However, the observability value of existing unified modeling graphs remains relatively limited. Their capabilities are primarily confined to data collection and storage of observable entities, lacking deep integration with task execution systems such as automated operations and maintenance, fault self-healing, and elastic scaling. Due to the lack of a closed-loop mechanism from "state awareness" to "task execution," existing unified modeling graphs cannot support task processing capabilities for observable entities, making it difficult to support the intelligent proactive governance needs of cloud computing systems.

[0022] The technical solution provided in this application introduces a novel operation manual set node into the graph structure of the unified modeling graph. The operation manual set node describes the operational capability information required for observable entities to perform task processing through its definition file. By linking operation manuals, an association is established between the operation manual set node and entity set nodes or dataset nodes in the unified modeling graph. This allows for precise and dynamic binding of operational capabilities to corresponding observable entities or observable data. In this way, the unified modeling graph can achieve deep collaboration with the task execution system, providing a closed-loop mechanism from "state awareness" to "task execution," supporting task processing capabilities oriented towards observable entities. Based on the hierarchical structure of the unified modeling graph, it can achieve layered organization and on-demand supply of operational capabilities, improving decision-making accuracy and task execution efficiency in complex cloud environments, and supporting the intelligent proactive governance needs of cloud computing systems.

[0023] The technical solutions of this application and how they solve the aforementioned technical problems are described in detail below with specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments. The technical solutions provided by each embodiment of this application are described in detail below with reference to the accompanying drawings.

[0024] Figure 1 This is a flowchart illustrating a graph processing method provided in an embodiment of this application. The method is applied to an observability management platform, which manages the observability of at least one observable entity in a cloud computing system. (See also...) Figure 1 The method may include the following steps: 101. Obtain the initial unified modeling graph, wherein the initial unified modeling graph organizes multiple regular graph nodes in the form of a graph structure. Regular graph nodes with relationships are connected by edges. The types of regular graph nodes include at least one of the following: entity set nodes and dataset nodes. Entity set nodes represent entity sets that describe a class of observable entities, and dataset nodes represent observable data sets of observable entities.

[0025] Specifically, the initial unified modeling graph refers to a unified modeling graph that does not support task processing capabilities for observable entities. The initial unified modeling graph organizes multiple regular graph nodes in the form of a graph structure. Regular graph nodes include, but are not limited to, entity set nodes and dataset nodes. Entity set nodes represent entity sets that describe a class of observable entities, and dataset nodes represent observable data sets of observable entities.

[0026] In the initial unified modeling graph, nodes with relationships in the regular graph are connected by edges, which represent the relationships between nodes. For example, there is a calling relationship between entity set node 1 and entity set node 2, and they are connected by an edge; there is a belonging relationship between entity set node 1 and its corresponding observable data set, and they are connected by an edge.

[0027] 102. In response to configuration operations on the graphical interface, configure the definition file and operation manual link information of the operation manual set. The definition file of the operation manual set is used to describe the operational capability information required to perform task processing on observable entities. The operation manual link information is used to define the association between the operation manual set and the entity set or observable data set.

[0028] Specifically, the observability management platform provides a graphical interface for configuration personnel to define definition files and link information for operation manual sets for observable entities.

[0029] A RunbookSet is an information carrier that encapsulates the operational capabilities required by observable entities to perform tasks. The RunbookSet uses definition files to fully describe the operational capabilities required by observable entities to perform tasks in scenarios such as fault diagnosis, daily inspection, elastic scaling, and security compliance.

[0030] Optionally, the definition file of the operation manual set includes, but is not limited to: toolbox configuration information, skill configuration information, and domain knowledge configuration information.

[0031] Toolkit configuration information describes the configuration of a toolkit. A toolkit is a collection of callable tools. The tools in a toolkit are service modules that provide atomic operation capabilities. Tool types include, but are not limited to: MCP (Model-Context Protocol) tools, script execution tools, database operation tools, and RESTful API tools. A RESTful API is an application programming interface that conforms to the REST (Representational State Transition) architectural style.

[0032] Optionally, toolbox configuration information may include, but is not limited to, runtime configuration information shared by all tools in the toolbox and tool parameter information for each tool in the toolbox.

[0033] Runtime configuration information includes, but is not limited to: toolbox identification information, toolbox type, shared configuration parameters, tags, and extended attributes.

[0034] The toolbox types include, but are not limited to: MCP type, Platform-Oriented Protocol API type, RESTful API type, script type, etc.

[0035] Shared configuration parameters are runtime parameters automatically inherited by all tools within the toolbox. These parameters include, but are not limited to, connection information, authentication credentials, and timeout configuration information. Connection information is the information required by the AIAgent system to access the tool server when invoking the tool. This information includes, but is not limited to, access address, communication protocol, port, and host number. Authentication credentials are the security credentials used by the AIAgent system to prove its identity and obtain operating permissions when connecting to the tool server. The timeout configuration information refers to the maximum waiting time set for each stage of the tool invocation process, used to prevent system resource exhaustion or task stagnation due to network latency, service unresponsiveness, or execution blocking. Timeout configuration information includes, but is not limited to, connection timeout, execution timeout, and read / write timeout.

[0036] The tool's parameter information includes, but is not limited to: input parameters, output parameters, risk level, idempotency flag, and confirmation requirements. Specifically, the risk level identifies the potential impact of the tool's operation; the idempotency flag indicates whether the tool's operation can be safely repeated without side effects; and the confirmation requirements indicate whether secondary manual confirmation is required before the tool is invoked.

[0037] In this embodiment, the skill configuration information is the definition information describing at least one skill. A skill is a reusable instruction module that encapsulates the operation process, professional knowledge and best practices of a specific task, covering capabilities such as script execution, data processing, and external API calls. It aims to enable the task intelligence system or operation and maintenance personnel to execute tasks consistently and efficiently according to preset steps.

[0038] Optionally, the skill definition information may include, but is not limited to, the following: skill name, skill description, list of pre-authorized tools, skill dependency files, priority information, and compatibility information.

[0039] The skill name serves as the unique identifier for the skill. The skill description explains the skill's function and usage scenarios, helping the task agent system identify when to use the skill. The pre-authorized tool list indicates the list of tools the skill can invoke, establishing a connection between the skill and tools in the toolbox. Skill dependency files are various files required to perform a task using the skill, including but not limited to: the SKILL.md file (the main skill file), reference materials, and auxiliary scripts. The main skill file defines the skill's triggering conditions, task flow, and constraint rules. Priority information indicates the priority of selecting the execution order among multiple candidate skills. Environmental compatibility refers to the environmental requirements for skill usage, including but not limited to: dependent system software packages, runtime versions, and network access permissions.

[0040] In this embodiment, the domain knowledge configuration information is used to configure at least one knowledge item related to task execution; each knowledge item includes, but is not limited to, content type field, content body field, application strategy field, priority field, and tag field, etc. The content type field indicates the content type of the domain knowledge. Content types include, but are not limited to: Markdown (Lightweight Markup Language), HTML (Hyper Text Markup Language), and URL (Uniform Resource Locator). Markdown domain knowledge represents text content written in Markdown syntax; HTML domain knowledge represents structured content in HTML format; and URL domain knowledge represents a Uniform Resource Locator for external domain knowledge.

[0041] The content body field is used to carry specific domain knowledge.

[0042] The application strategy field defines how knowledge items are used during task execution. The field information of the application strategy field includes, but is not limited to: auto, manual, always, and custom. Among them, auto means that the system automatically determines whether to apply based on the context; manual means that it is applied only when the user explicitly specifies it; always means that it is always applied when the basic conditions are met; and custom means that it is applied based on custom rules or logic. Among them, the priority field information is used to determine the application order among multiple matching knowledge items; the tag field information is used for classification, retrieval or association of keywords or metadata, which facilitates knowledge organization, filtering and intelligent matching.

[0043] In this embodiment, the operation manual link information is used to define the association between the operation manual set and the entity set or observable data set.

[0044] Specifically, the user manual link information includes the following fields: source node field, target node field, and link type field. Optionally, the user manual link information may also include: entity filter condition field.

[0045] The field information of the source node field is used to indicate the attribute information of the entity set or observable data set that is associated with the manual set; for example, the field information of the source node field includes the domain, type or name to which the entity set or observable data set belongs; The field information of the target node is used to indicate the attribute information of the corresponding manual set; for example, the field information of the target node includes the domain, type or name of the manual set. The link type field information is used to indicate the link type between the manual set and the entity set or observable data set; The field information of the entity filter condition field is used to indicate the scope of the operation manual set; for example, the field information of the entity filter condition field can be an entity-level filter expression, which limits the scope of the operation manual set.

[0046] In some optional embodiments, the implementation of configuring the definition file of the operation manual set in response to configuration operations on the graphical interface includes: displaying toolbox configuration items, skill configuration items, and domain knowledge configuration items on the first graphical interface; configuring toolbox configuration information in response to configuration operations for toolbox configuration items, the toolbox configuration information including: runtime configuration information shared by all tools in the toolbox and tool parameter information for each tool in the toolbox; configuring skill configuration information in response to configuration operations for skill configuration items, the skill configuration information including skill definition information for at least one skill, the skill definition information for each skill including at least one of the following: skill name, skill description information, pre-authorized tool list, skill dependency file, usage priority information, and environment compatibility information; configuring domain knowledge configuration information in response to configuration operations for domain knowledge configuration items, each knowledge item in the domain knowledge configuration information including at least one of the following: domain knowledge content type, content body, application strategy, selection priority, and tag; and generating the definition file of the operation manual set based on the toolbox configuration information, skill configuration information, and domain knowledge configuration information in response to the configuration completion operation.

[0047] Specifically, the first graphical interface is used to configure the definition files of the operation manual set; the graphical configuration mechanism completes the visualization, modularization and low-code construction of the definition files of the operation manual set, thereby improving the configuration efficiency of the definition files of the operation manual set.

[0048] In some optional embodiments, for each knowledge item, the content body of the domain knowledge is divided into blocks according to the content type of the domain knowledge to obtain multiple knowledge blocks of the domain knowledge, and a full-text index and a vector index are established for each knowledge block; the multiple knowledge blocks of the domain knowledge and the full-text index and vector index of each knowledge block are saved to the index storage system.

[0049] Specifically, the content of domain knowledge is segmented according to its content type. For example, Markdown content can be segmented by heading level or paragraph semantic boundaries; HTML content is extracted based on the DOM (Document Object Model) structure; and URL content is first crawled and parsed before being segmented according to the above rules. After segmentation, two complementary indexes are built for each generated knowledge block: a full-text index to support precise keyword matching and a vector index to support vector retrieval based on semantic similarity. Subsequently, all knowledge blocks and their corresponding full-text and vector indexes are persistently stored in a unified index storage system. The index storage system should be selected based on the supported retrieval types. For example, when keyword retrieval is required, a full-text search engine based on an inverted index can be used; when vector semantic retrieval is required, a vector database supporting the Approximate Nearest Neighbor (ANN) algorithm can be used. In a preferred embodiment, a unified index storage system supporting hybrid retrieval is used to reduce system complexity and improve retrieval efficiency.

[0050] In this embodiment, the content body of each knowledge item is divided into blocks according to its content type, and a full-text index and vector index are built simultaneously for each knowledge block. These structured index data are then uniformly stored in the index storage system, which improves the retrieval efficiency of domain knowledge. This enables the task intelligent agent system to quickly and accurately locate the most relevant knowledge fragments in complex operation and maintenance scenarios, thereby enhancing the context awareness of decision-making and the accuracy of response.

[0051] In some optional embodiments, the configuration of the operation manual link information in response to the configuration operation on the graphical interface is implemented as follows: Source node field configuration items, target node field configuration items, and link type field configuration items are displayed on the second graphical interface; in response to the configuration operation for the source node field configuration item, the field information of the source node field is configured, the field information of the source node field is used to indicate the attribute information of the entity set or observable data set that is associated with the operation manual set; in response to the configuration operation for the target node field configuration item, the field information of the target node field is configured, the field information of the target node field is used to indicate the attribute information of the corresponding operation manual set; in response to the configuration operation for the link type field configuration item, the field information of the link type field is configured, the field information of the link type field is used to indicate the link type between the operation manual set and the entity set or observable data set; in response to the configuration completion operation, the operation manual link information is generated based on the field information of the source node field, the field information of the target node field, and the field information of the link type field.

[0052] Specifically, the second graphical interface is used to configure the operation manual link information; through the graphical configuration mechanism, the operation manual link information is visualized, modularized, and built in a low-code manner, thus improving the configuration efficiency of the operation manual link information.

[0053] In some optional embodiments, the method further includes: displaying entity filtering condition field configuration items on a second graphical interface; configuring field information of the filtering condition field in response to a configuration operation for the filtering condition field configuration item, the field information of the filtering condition field being used to indicate the scope of the operation manual set; the scope of the operation manual set includes at least one of the following: entity type scope, domain scope, and global scope; wherein, the global scope operation manual set applies to all observable entities of the cloud computing system; the domain scope operation manual set applies to observable entities of the cloud computing system belonging to a specified domain; and the entity type scope operation manual set applies to a specified observable entity of the cloud computing system.

[0054] Specifically, the entity filtering condition field configuration item on the second graphical interface enables configuration personnel to accurately define the scope of the operation manual link information in a visual and low-code manner, thereby realizing fine-grained, context-aware dynamic binding between the operation manual and observable entities, and improving the accuracy and applicability of automated operation and maintenance.

[0055] 103. Based on the definition file of the operation manual set, create operation manual set nodes in the initial unified modeling graph and configure the node attribute information of the operation manual set nodes. The node attribute information of the operation manual set nodes includes the definition file of the operation manual set.

[0056] 104. Based on the operation manual link information of the operation manual set node, create edges connecting the operation manual set node and its associated regular graph nodes in the initial unified modeling graph, and configure the edge attribute information of the edges connected to the operation manual set node to obtain the unified modeling graph.

[0057] Specifically, after creating operation manual set nodes and establishing edges connecting them in the initial unified modeling graph, the node attribute information of the operation manual set nodes and the edge attribute information of the corresponding edges are configured to obtain the unified modeling graph. The node attribute information of the operation manual set nodes is directly derived from the definition file of the operation manual set, and the edge attribute information of the edges connected to the operation manual set nodes is derived from the operation manual link information. Finally, after attribute injection of operation manual set nodes and edges, the initial unified modeling graph is enhanced into a unified modeling graph that integrates operation capability information. This unified modeling graph not only supports efficient graph traversal and semantic query, but also provides task processing capabilities for subsequent task execution.

[0058] The technical solution provided in this application introduces a novel operation manual set node into the graph structure of the unified modeling graph. The operation manual set node describes the operational capability information required for task processing of observable entities through its definition file. By linking operation manuals, an association is established between the operation manual set node and entity set nodes or dataset nodes in the unified modeling graph. This allows for precise and dynamic binding of operational capabilities to corresponding observable entities or observable data. In this way, the unified modeling graph can achieve deep collaboration with the task execution system, providing a closed-loop mechanism from "state awareness" to "task execution," supporting task processing capabilities for observable entities. Based on the hierarchical structure of the unified modeling graph, it enables layered organization and on-demand supply of operational capabilities, improving decision-making accuracy and task execution efficiency in complex cloud environments, and supporting the intelligent proactive governance needs of cloud computing systems.

[0059] In some optional embodiments, after obtaining the unified modeling graph, entity topology change events in the cloud computing system can also be detected. The change type of the entity topology change event includes at least one of the following: entity deletion, entity type change, or entity domain change. In response to detecting any entity topology change event, the affected regular graph node and its associated operation manual set node are located in the unified modeling graph. The edges and corresponding edge attribute information between the affected regular graph node and its associated operation manual set node are adjusted according to the entity topology change event.

[0060] Specifically, entity topology change events in the cloud computing system can be detected periodically or according to event triggering. For example, entity topology change events of the entity deletion type include: an existing observable entity in the cloud computing system is removed from the cloud computing system due to being offline, destroyed, or retired; entity topology change events of the entity type change include: a host is changed from "database server" to "caching node"; entity topology change events of the entity's domain change include: a microservice is migrated from "payment domain" to "user domain".

[0061] Upon detecting any entity topology change event, the graph update process is automatically triggered: First, the affected regular graph nodes are precisely located in the unified modeling graph, and their adjacency relationships are further traversed to identify all associated manual set nodes. Then, based on the specific semantics of the change event, the edges and their attribute information of the affected graph are dynamically adjusted. For example, if an observable entity is deleted, the associated edges between it and the manual set nodes are removed to avoid invalid calls; if the entity type or domain changes, the entity filtering conditions of the original links are reassessed to determine if they are still applicable. If necessary, the link type and priority of the edges are updated, or mismatched associations are directly disconnected, and new associations are established according to a preset strategy.

[0062] In this embodiment, by detecting entity topology change events in the cloud computing system after the unified modeling graph is constructed, the changes in the operation and maintenance environment can be dynamically perceived, and the affected regular graph nodes and their associated operation manual set nodes can be automatically located. Then, the associated edges and their edge attributes between the two can be adjusted in real time. In this way, the unified modeling graph can reflect the latest topology state of the cloud environment in real time, and realize the adaptive maintenance of the association between the operation manual and observable entities.

[0063] In some optional embodiments, after obtaining the unified modeling graph, the task execution effect score is determined based on the execution feedback data of the task agent system in the most recent time period; if the task execution effect score does not meet the task execution effect qualification conditions, the first operation manual set node associated with the task processing performed by the task agent system in the most recent time period is located in the unified modeling graph. Root cause analysis is performed based on the definition file of the first operation manual set node, the entity set information of the entity set node connected to the first operation manual set node, and the observable data set corresponding to the dataset node connected to the first operation manual set node; optimization suggestions for the first operation manual set node are generated based on the root cause analysis results; the optimization suggestions for the first operation manual set node are displayed on the third graphical interface; in response to modification operations initiated based on the optimization suggestions for the first operation manual set node, the node attribute information of the first operation manual set node and the edge attribute information of its associated edges in the unified modeling graph are modified.

[0064] Specifically, the most recent time period is the period closest to the current time. The duration of the most recent time period is not limited; for example, the most recent time period could be the last 10 minutes. Execution feedback data from the task intelligence system within this most recent time period is collected. This feedback data includes, but is not limited to, task success rate, average execution latency, and result accuracy. The task success rate reflects the percentage of successfully completed tasks out of the total number of triggered tasks. For example, if 85 out of 100 calls succeed, the success rate is 85%. Average execution latency reflects the average time taken from the start to the completion of a task. Result accuracy reflects whether the task execution result meets expectations. When determining the task execution performance score based on the execution feedback data from the task intelligence system within the most recent time period, a multi-dimensional weighted fusion approach can be used for quantitative calculation. For example, predefine the weight coefficients for each execution feedback data point, such as task success rate accounting for 50%, average execution latency accounting for 30%, and result accuracy accounting for 20%. Normalize each execution feedback data point to a unified scoring range (e.g., 0–100 points), and then generate the task execution performance score through weighted summation.

[0065] In this embodiment, the task execution performance meets certain criteria. For example, if the task execution performance score is greater than a set score threshold, the criteria are considered met. If the score is less than the set score threshold, the criteria are considered not met. Alternatively, if the score falls within a set score range, the criteria are considered met. If the score does not fall within the set score range, the criteria are considered not met.

[0066] In this embodiment, if the task execution performance score does not meet the criteria for achieving the target, the first operation manual set node associated with the task agent system's task processing within the most recent time period is located in the unified modeling graph. Root cause analysis is then performed by integrating the definition file of the operation manual set of the first operation manual set node, the entity set information of the entity set nodes connected to the first operation manual set node, and the observable data set corresponding to the dataset nodes connected to the first operation manual set node. This accurately identifies the root causes of the poor task performance, such as: flaws in skill process design, incorrect tool parameter configuration, insufficient domain knowledge coverage, mismatch between entity filtering conditions and actual entities, or missing environmental compatibility declarations. The results of this root cause analysis serve as the basis for generating subsequent optimization suggestions, supporting the continuous evolution and adaptive optimization of the operation manual set.

[0067] In this embodiment, the third graphical interface is used for users to confirm and modify operation manual set nodes; optimization suggestions for the first operation manual set nodes are automatically generated based on the root cause analysis results and displayed intuitively on the third graphical interface, realizing a complete closed-loop optimization process from "effect evaluation - problem diagnosis - improvement suggestions - manual confirmation - map update"; thereby ensuring that the operation manual is always consistent with the actual operating environment and task effect, enhancing the long-term autonomy and reliability of the task intelligent agent system.

[0068] Figure 2 This is a flowchart illustrating a task processing method provided in an embodiment of this application. The method is applied to an observability management platform, which is used to manage the observability of at least one observable entity in a cloud computing system. (See also...) Figure 2 The method may include the following steps: 201. Receive the capability discovery request sent by the task intelligent agent system. The capability discovery request is used to request the information on the operational capabilities required to perform task processing on the target observable entity, and the capability discovery request carries the entity context information of the target observable entity.

[0069] Specifically, a task-oriented intelligent agent system can be an intelligent agent system capable of performing various tasks, including but not limited to: troubleshooting tasks, inspection and analysis tasks, process approval tasks, and work order processing tasks.

[0070] In practical applications, the task intelligence system can initiate capability discovery requests on demand to request the observability management platform to obtain the operational capability information required to perform task processing for the target observable entity. The target observable entity may be, for example, a cloud host, container, or database instance. The operational capability information may include, but is not limited to, a tool list, a skill list, and a knowledge list.

[0071] In practical applications, capability discovery requests must include entity context information of the target observable entity. Entity context information typically includes: entity unique identifier (entity_id), entity type (entity_type), domain, operating environment, current operating status, or observation metrics.

[0072] 202. Obtain the definition file of at least one level of operation manual set that matches the entity context information from the pre-built unified modeling graph. The scope of operation manual sets at different levels is different.

[0073] 203. Obtain operational capability information based on the definition file of at least one level of the operation manual set. The operational capability information includes a list of tools, a list of skills, and a list of knowledge corresponding to the target observable entity.

[0074] 204. Return the operational capability information to the task agent system so that the task agent system can perform task processing on the target observable entity based on the operational capability information.

[0075] Specifically, a new type of operation manual set node has been introduced into the pre-built unified modeling graph. The operation manual set node describes the operation capability information required to perform task processing on observable entities through the definition file of the operation manual set node. Through operation manual link information, the relationship between operation manual set node and entity set node or dataset node is established in the unified modeling graph. The unified modeling graph supports task processing capabilities for observable entities.

[0076] Upon receiving a capability discovery request, the observability management platform searches for definition files of operation manual sets that match the entity's context information within a pre-built unified modeling graph. Based on the definition files of operation manual sets at one or more levels obtained from the unified modeling graph, it further extracts specific operational capability information for the target observable entity. This specific operational capability information includes, but is not limited to, a list of tools, skills, and knowledge that can be used to perform the task. The observability management platform then returns the compiled operational capability information to the task agent system that initiated the capability discovery request. In this way, the task agent can formulate a specific task plan and execute corresponding operations to complete the task processing based on this operational capability information.

[0077] For example, a task-oriented intelligent agent system (such as Operations Robot A) detects that the CPU usage of an instance named mysql-prod-01 exceeds the threshold for 5 minutes, triggering an anomaly diagnosis task. The task-oriented intelligent agent system sends a capability discovery request to the observability management platform, carrying the entity context information as follows: { "entity_id":"mysql-prod-01" / / Unique identifier for the entity "entity_type": "Pod" / / Entity type "domain":"k8s" / / The domain to which the entity belongs “cloud_provider”:“Aaaaa” / / The cloud computing system where the entity resides "observed_symptoms":["high_cpu","slow_query"] / / Entity characteristics } Among them, the entity features in the entity context information convey the task intent, namely, "diagnose high CPU and slow query problems".

[0078] The task intelligence system receives operational capability information returned by the observability management platform. The list of tools in the operational capability information includes, but is not limited to: slow query log tools, resource usage viewing tools, and slow log summarization tools. The list of skills in the operational capability information includes, but is not limited to: slow query analysis (analyze_slow_queries) skills and lock contention detection (detect_lock_contention) skills. The knowledge list includes: "High CPU Troubleshooting Guide", "Best Practices for Database Performance Tuning", etc.

[0079] When the task intelligence system performs task processing on the target observable entity based on operational capability information, the task intelligence system parses the various contents of the operational capability information and performs capability matching and priority ranking in conjunction with the current task intent (such as "diagnosing high CPU and slow queries"). For example, it will first call resource viewing tools to confirm whether the CPU, memory, and other resources of mysql-prod-01 are continuously at high levels; if a performance bottleneck is confirmed, it will further call slow query log tools or slow log aggregation tools to extract high-frequency or high-time-consuming SQL statements.

[0080] Subsequently, the task agent activates the corresponding skills to perform in-depth analysis. For example, it calls the analyze_slow_queries skill to perform pattern recognition on the extracted slow SQL statements to determine whether there are problems such as missing indexes, full table scans, or complex subqueries. At the same time, it calls the detect_lock_contention skill to check whether there are transaction blocking or lock waiting phenomena, so as to eliminate the misjudgment of "seemingly slow queries are actually lock contention".

[0081] Throughout the task processing, the task intelligence system dynamically references domain knowledge from its knowledge list to improve the accuracy and interpretability of its decisions. Ultimately, the system integrates tool outputs, skill analysis results, and domain knowledge to generate a structured task processing report and automatically performs corrective actions (such as creating missing indexes or terminating abnormal sessions) or pushes actionable suggestions to operations personnel.

[0082] In some optional embodiments, the implementation of obtaining the definition files of at least one level of the operation manual set that matches the entity context information in the pre-built unified modeling graph is as follows: the definition files of at least one level of the operation manual set that matches the entity context information are obtained in the unified modeling graph in hierarchical order; wherein, the hierarchical order is the entity type layer corresponding to the entity type scope, the domain layer corresponding to the domain scope, and the global layer corresponding to the global scope; or, the hierarchical order is the global layer, the domain layer, and the entity type layer.

[0083] Specifically, the hierarchical structure of the unified modeling graph is used to achieve hierarchical management and inheritance of capabilities. The operational capability information of an observable entity can be traversed sequentially according to the hierarchical order of entity type layer, domain layer, and global layer, or the hierarchical order of global layer, domain layer, and entity type layer, etc. This reuses the natural hierarchy of the graph structure, transforming the capability management problem into a graph traversal problem, and ensuring that the hierarchical relationship of capabilities and the organizational relationship of entities maintain structural consistency.

[0084] In some optional embodiments, the method of obtaining operational capability information based on the definition file of at least one level of the operation manual set includes: obtaining operational capability information based on the definition file of at least one level of the operation manual set, wherein the operational capability information includes a toolbox list, a skill list, and a knowledge list corresponding to the target observable entity; obtaining toolbox configuration information of at least one level from the definition file of at least one level of the operation manual set, and determining the toolbox list corresponding to the target observable entity based on the toolbox configuration information of at least one level of the operation manual set; obtaining skill configuration information of at least one level from the definition file of at least one level of the operation manual set, and determining the skill list corresponding to the target observable entity based on the skill configuration information of at least one level of the operation manual set; obtaining domain knowledge configuration information of at least one level from the definition file of at least one level of the operation manual set; and determining the knowledge list corresponding to the target observable entity based on the domain knowledge configuration information of at least one level of the operation manual set.

[0085] Specifically, when determining the toolbox list corresponding to the target observable entity, the corresponding toolbox configuration information is parsed from the definition files of the operation manual sets of one or more matched levels (such as the entity type layer, domain layer, and global layer). The toolbox configuration information at each level is merged and deduplicated to finally generate a tool list suitable for the target observable entity, ensuring that the task agent system can call the operation tools most appropriate to the current entity context. For the same tool, tools from the entity type layer are prioritized for addition to the tool list; if no tool exists at the entity type layer, tools from the domain layer are prioritized; if no tool exists at the domain layer, tools from the global layer are added.

[0086] When determining the skill list corresponding to the target observable entity, the corresponding skill configuration information is parsed from the definition files of one or more matched layers (such as entity type layer, domain layer, and global layer). The skill configuration information at each layer is merged and deduplicated to finally generate a skill list suitable for the target observable entity. For the same skill, skills from the entity type layer are prioritized for addition to the skill list; if no entity type layer skill exists, then domain layer skills are prioritized; if no domain layer skill exists, then global layer skills are added.

[0087] When determining the knowledge list corresponding to the target observable entity, the corresponding domain knowledge configuration information is parsed from the definition files of the operation manual sets of one or more matched levels (such as the entity type layer, domain layer, and global layer). This information points to structured or unstructured knowledge resources, such as troubleshooting guides, best practice documents, security compliance baselines, and failure mode libraries. Different levels of knowledge have different coverage scopes. The entity type layer provides dedicated documents such as the "High CPU Troubleshooting Manual," the domain layer provides the "Cloud-Native Database Operation and Maintenance White Paper," and the global layer contains basic materials such as "General Methods for Performance Analysis." This domain knowledge configuration information is then integrated into a unified knowledge list.

[0088] In this embodiment, by extracting toolbox configuration information, skill configuration information, and domain knowledge configuration information from the definition files of the multi-level operation manual set, the precise and dynamic construction of the target observable entity's operational capabilities is achieved.

[0089] In some optional embodiments, if the knowledge blocks of domain knowledge in at least one level of domain knowledge configuration information are stored in an index storage system, the implementation method for determining the knowledge list corresponding to the target observable entity based on the domain knowledge configuration information of at least one level is as follows: based on the task question keywords associated with the task processing performed by the task intelligent agent system, keyword retrieval and vector semantic retrieval are performed in parallel in the index storage to obtain a first retrieval result and a second retrieval result; the first retrieval result includes at least one first candidate knowledge block, and the second retrieval result includes at least one second candidate knowledge block; the first retrieval result and the second retrieval result are fused using a reciprocal ranking fusion method to obtain a third retrieval result, and the third retrieval result includes at least one third candidate knowledge block; at least one third candidate knowledge block is sorted according to the distance between the operation manual set node to which the third candidate knowledge block belongs and the entity set node to which the target observable entity belongs, and the level of the operation manual set node to which the third candidate knowledge block belongs; the knowledge list corresponding to the target observable entity is determined according to the top-ranked third candidate knowledge block.

[0090] Specifically, task question keywords are keywords extracted from task question information by the task-oriented intelligent agent system when executing the task. Based on these keywords, two types of searches are initiated in parallel within the index storage system: firstly, precise or fuzzy keyword matching is performed based on the inverted index to obtain a first set of candidate knowledge blocks; secondly, semantic similarity retrieval is performed based on the vector representation of the task question keywords to obtain a second set of semantically related candidate knowledge blocks. Next, the Reciprocal Rank Fusion (RRF) algorithm is used to fuse the two sets of candidate results, generating a comprehensively ranked third list of candidate knowledge blocks. Finally, combining the hierarchical structure information of the unified modeling graph, the third candidate knowledge blocks in the third candidate knowledge block list are finely reordered, and the top-ranked knowledge blocks are selected to form the knowledge list corresponding to the target observable entity.

[0091] The distance between nodes is measured by the number of intermediate nodes along the shortest path between two nodes in the unified modeling graph. The more intermediate nodes, the greater the distance between nodes. When refining the reordering of each third candidate knowledge block, the greater the distance between nodes, the later the ranking; the smaller the distance between nodes, the earlier the ranking. The earlier the level, the later the ranking; the later the level, the earlier the ranking. The levels, from front to back, are: global level, domain level, entity type level, etc.

[0092] In this embodiment, not only are the advantages of keyword retrieval and vector semantic retrieval combined, but topological distance and hierarchical depth from the unified modeling graph are also introduced as priors for structure-aware ranking. This enables the generated knowledge list to achieve high synergy in three dimensions: semantic relevance, task adaptability, and domain structure, thereby significantly enhancing the understanding ability, reasoning efficiency, and decision-making accuracy of the task-oriented intelligent agent system in complex scenarios.

[0093] The technical solution provided in this application introduces a novel operation manual set node into the graph structure of the unified modeling graph. The operation manual set node describes the operational capability information required for observable entities to perform task processing through its definition file. By linking operation manuals, the unified modeling graph establishes associations between operation manual set nodes and entity set nodes or dataset nodes. This allows for precise and dynamic binding of operational capabilities to corresponding observable entities or observable data. In this way, the unified modeling graph can achieve deep collaboration with the task execution system, providing a closed-loop mechanism from "state awareness" to "task execution," supporting task processing capabilities for observable entities. Based on the hierarchical structure of the unified modeling graph, it enables layered organization and on-demand supply of operational capabilities, improving decision-making accuracy and task execution efficiency in complex cloud environments, and supporting the intelligent proactive governance needs of cloud computing systems. Furthermore, this application embodiment achieves dynamic, context-aware operational capability supply for task-oriented intelligent agent systems through a capability discovery mechanism in the task processing method. Specifically, upon receiving a capability discovery request carrying contextual information about a target observable entity, the observability management platform can accurately match operational capability information (including tools, skills, and knowledge) applicable to the entity's current state and task scenario based on the hierarchically organized operation manual set nodes in the unified modeling graph, and return this information to the task agent system in real time. This mechanism not only avoids the redundancy or missing issues caused by traditional static capability configuration but also significantly improves the response speed and execution accuracy of the task agent in complex and ever-changing cloud environments. Simultaneously, relying on the hierarchical scope design of the operation manual set in the graph, the observability management platform can automatically filter operational capabilities with appropriate granularity based on the entity's context, achieving "on-demand supply and precise binding." This effectively supports closed-loop intelligent governance from state awareness to task execution, enhancing the proactive decision-making and autonomous capabilities of cloud computing systems in scenarios such as fault self-healing, resource scheduling, and security response.

[0094] In the embodiments of this application, "at least one" refers to one or more, and "more than one" refers to two or more. "And / or" describes the access relationship between associated 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, where A and B can be singular or plural. In the textual description of this application, the character " / " generally indicates that the preceding and following associated objects have an "or" relationship. Furthermore, in the embodiments of this application, "first," "second," "third," etc., are only used to distinguish the content of different objects and have no other special meaning.

[0095] It should be noted that, in the cases involving user information in the embodiments of this application, the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in the embodiments of this application are all information and data authorized by the user or fully authorized by all parties. Furthermore, the collection, use, and processing of related data must comply with the relevant laws, regulations, and standards of the relevant countries and regions, and corresponding operation entry points are provided for users to choose to authorize or refuse. In addition, the various models involved in this application (including but not limited to language models or large models) comply with relevant laws and standards.

[0096] Figure 3 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Figure 3 As shown, the electronic device includes: a memory 31 and a processor 32; Memory 31 is used to store computer programs and can be configured to store various other data to support operation on the computing platform. Examples of this data include instructions for any application or method operating on the computing platform, data structures, contact data, phone book data, messages, pictures, videos, etc.

[0097] The processor 32, coupled to the memory 31, is used to execute the computer program in the memory 31 for: performing steps in a map processing method or a task processing method.

[0098] Optional, such as Figure 3 As shown, the electronic device also includes other components such as a communication component 33, a display 34, a power supply component 35, and an audio component 36. Figure 3 The diagram only shows some components and does not mean that the electronic device includes only these components. Figure 3 The components shown. Additionally... Figure 3 The components within the dashed box are optional, not mandatory, and their specific requirements depend on the product form of the electronic device. The electronic device in this embodiment can be a desktop computer, laptop computer, smartphone, or IoT (Internet of Things) device, or a server-side device such as a conventional server, cloud server, or server array. If the electronic device in this embodiment is a desktop computer, laptop computer, or smartphone, it may include... Figure 3 The components within the dashed box; if the electronic device in this embodiment is implemented as a conventional server, cloud server, or server array, etc., it may be omitted. Figure 3 The component within the dashed box.

[0099] The aforementioned memory can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random-Access Memory (SRAM), Electrically Erasable Programmable Read Only Memory (EEPROM), Erasable Programmable Read Only Memory (EPROM), Programmable Read-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0100] The aforementioned communication component is configured to facilitate wired or wireless communication between the device containing the communication component and other devices. The device containing the communication component can access wireless networks based on communication standards, such as 2G (2nd Generation), 3G (3rd Generation), 4G (4th Generation) / LTE (long Term Evolution), 5G (5th Generation), or combinations thereof. In one exemplary embodiment, the communication component receives broadcast signals or broadcast-related information from an external broadcast management system via a broadcast channel.

[0101] The aforementioned display includes a screen, which may include a Liquid Crystal Display (LCD) and a Touch Panel (TP). If the screen includes a Touch Panel, the screen can be implemented as a touchscreen to receive input signals from the user. The Touch Panel includes one or more touch sensors to sense touches, swipes, and gestures on the Touch Panel. The touch sensors can sense not only the boundaries of touch or swipe actions but also the duration and pressure associated with the touch or swipe operation.

[0102] The aforementioned power supply components provide power to various components within the device in which they reside. These power supply components may include a power management system, one or more power sources, and other components associated with generating, managing, and distributing power to the device in which they reside.

[0103] The aforementioned audio component can be configured to output and / or input audio signals. For example, the audio component includes a microphone (MIC) configured to receive external audio signals when the device containing the audio component is in an operating mode, such as call mode, recording mode, or voice recognition mode. The received audio signals can be further stored in memory or transmitted via a communication component. In some embodiments, the audio component also includes a speaker for outputting audio signals.

[0104] Accordingly, embodiments of this application also provide a computer-readable storage medium storing a computer program, which, when executed by a processor, enables the processor to implement the steps in the above-described method embodiments. The computer-readable storage medium may be volatile, non-volatile, or a combination thereof, and may be removable or non-removable. Examples of computer-readable storage media include, but are not limited to, phase-change random access memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), flash memory or other memory technologies, CD-ROM, digital video disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transfer medium.

[0105] Accordingly, this application also provides a computer program product, which includes a computer program or instructions. When the computer program or instructions are executed by a processor, the processor is able to implement the steps in the above method embodiments. It should be understood that each step or combination of steps in the above method flow can be implemented by the computer program or instructions. In addition, these computer programs or instructions can be applied to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device, so that the processor of the general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing device can be implemented as a means to implement the corresponding functions in the above method embodiments.

[0106] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0107] The above are merely embodiments of this application and are not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A spectrum processing method, characterized in that, An observability management platform is applied to manage the observability of at least one observable entity in a cloud computing system. The method includes: An initial unified modeling graph is obtained, wherein the initial unified modeling graph organizes multiple regular graph nodes in the form of a graph structure, and regular graph nodes with relationships are connected by edges. The types of regular graph nodes include at least one of the following: entity set nodes and dataset nodes. The entity set nodes represent entity sets describing a class of observable entities, and the dataset nodes represent observable data sets of observable entities. In response to configuration operations on the graphical interface, the definition file of the configuration manual set and the manual link information are configured. The definition file of the configuration manual set is used to describe the operational capability information required to perform task processing on observable entities, and the manual link information is used to define the association between the configuration manual set and the entity set or observable data set. According to the definition file of the operation manual set, an operation manual set node is created in the initial unified modeling graph, and the node attribute information of the operation manual set node is configured. The node attribute information of the operation manual set node includes the definition file of the operation manual set. Based on the operation manual link information of the operation manual set node, an edge is created in the initial unified modeling graph to connect the operation manual set node with its associated regular graph node, and the edge attribute information of the edge connected to the operation manual set node is configured to obtain the unified modeling graph.

2. The method according to claim 1, characterized in that, In response to configuration operations on the graphical interface, the definition files for the configuration operation manual set include: The toolbox configuration items, skill configuration items, and domain knowledge configuration items are displayed on the first graphical interface; In response to a configuration operation for the toolbox configuration item, configure toolbox configuration information, which includes: runtime configuration information shared by all tools in the toolbox and tool parameter information for each tool in the toolbox; In response to the configuration operation for the skill configuration item, skill configuration information is configured. The skill configuration information includes skill definition information for at least one skill. The skill definition information for each skill includes at least one of the following: skill name, skill description information, pre-authorized tool list, skill dependency file, usage priority information, and environment compatibility information. In response to a configuration operation for a domain knowledge configuration item, domain knowledge configuration information is configured, wherein each knowledge item in the domain knowledge configuration information includes at least one of the following: content type, content body, application strategy, selection priority, and tag of the domain knowledge; In response to the configuration completion operation, a definition file for the operation manual set is generated based on the toolbox configuration information, the skill configuration information, and the domain knowledge configuration information.

3. The method according to claim 2, characterized in that, Also includes: The content body of the domain knowledge is divided into blocks according to the content type of the domain knowledge to obtain multiple knowledge blocks of the domain knowledge, and a full-text index and vector index are established for each knowledge block; Multiple knowledge blocks of the domain knowledge, along with the full-text index and vector index of each knowledge block, are saved to the index storage system.

4. The method according to claim 1, characterized in that, In response to configuration operations on the graphical interface, configuration manual link information is provided, including: The source node field configuration items, target node field configuration items, and link type field configuration items are displayed on the second graphical interface; In response to a configuration operation for the source node field configuration item, the field information of the source node field is configured. The field information of the source node field is used to indicate the attribute information of the entity set or observable data set that is associated with the operation manual set. In response to a configuration operation for the target node field configuration item, the field information of the target node field is configured, and the field information of the target node field is used to indicate the attribute information of the corresponding operation manual set; In response to a configuration operation for the link type field configuration item, the field information of the link type field is configured, wherein the field information of the link type field is used to indicate the link type between the operation manual set and the entity set or observable data set; In response to the configuration completion operation, the operation manual link information is generated based on the field information of the source node field, the field information of the target node field, and the field information of the link type field.

5. The method according to claim 4, characterized in that, Also includes: The entity filtering condition field configuration items are displayed on the second graphical interface; In response to a configuration operation for the filter condition field configuration item, the field information of the filter condition field is configured, and the field information of the filter condition field is used to indicate the scope of the operation manual set; The scope of the operation manual set includes at least one of the following: entity type scope, domain scope, and global scope; The operation manual set for the global scope applies to all observable entities in the cloud computing system; the operation manual set for the domain scope applies to observable entities in the cloud computing system belonging to a specified domain; and the operation manual set for the entity type scope applies to specified observable entities in the cloud computing system.

6. The method according to any one of claims 1 to 5, characterized in that, After obtaining the unified modeling map, it also includes: Detect entity topology change events in the cloud computing system. The change type of the entity topology change event includes at least one of the following: entity deletion, entity type change, or entity domain change. In response to the detection of any entity topology change event, locate the affected regular graph node and its associated operation manual set node in the unified modeling graph; Based on the entity topology change event, adjust the edges and corresponding edge attribute information between the affected regular graph nodes and their associated operation manual set nodes.

7. The method according to any one of claims 1 to 5, characterized in that, After obtaining the unified modeling map, the following is also included: The task execution performance score is determined based on the execution feedback data of the task intelligence system in the most recent time period; If the task execution effect score does not meet the task execution effect qualification conditions, then locate the first operation manual set node associated with the task intelligent agent system performing task processing in the most recent time period in the unified modeling graph; Root cause analysis is performed based on the definition file of the operation manual set of the first operation manual set node, the entity set information of the entity set node connected to the first operation manual set node, and the observable data set corresponding to the dataset node connected to the first operation manual set node. Optimization suggestions for the nodes of the first operation manual set are generated based on the root cause analysis results; The optimization suggestions for the nodes of the first operation manual set are displayed on the third graphical interface; In response to a modification operation initiated based on the optimization suggestions of the first operation manual set nodes, the node attribute information of the first operation manual set nodes and the edge attribute information of their associated edges in the unified modeling graph are modified.

8. A task processing method, characterized in that, An observability management platform is applied to manage the observability of at least one observable entity in a cloud computing system. The method includes: The system receives a capability discovery request sent by the task intelligent agent system. The capability discovery request is used to request the acquisition of the operational capability information required to perform task processing on the target observable entity, and the capability discovery request carries the entity context information of the target observable entity. In a pre-built unified modeling graph, a definition file for at least one level of operation manual set that matches the entity context information is obtained. The unified modeling graph is constructed according to the method described in any one of claims 1-7, and the scope of operation manual sets at different levels is different. The operational capability information is obtained from the definition file of at least one level of the operation manual set, and the operational capability information includes a tool list, a skill list and a knowledge list corresponding to the target observable entity; The operational capability information is returned to the task agent system so that the task agent system can perform task processing on the target observable entity based on the operational capability information.

9. The method according to claim 8, characterized in that, Obtain the definition file of at least one level of the operation manual set that matches the entity context information from the pre-built unified modeling graph, including: In a hierarchical order, obtain the definition files of at least one level of the operation manual set that matches the context information of the entity in the unified modeling graph; The hierarchical order is as follows: the entity type layer corresponding to the entity type scope, the domain layer corresponding to the domain scope, and the global layer corresponding to the global scope; or, the hierarchical order is as follows: global layer, domain layer, and entity type layer.

10. The method according to claim 8, characterized in that, Obtaining the operational capability information based on the definition file of at least one level of the operation manual set includes: Obtain toolbox configuration information for at least one level from the definition file of the operation manual set at at least one level, and determine the toolbox list corresponding to the target observable entity based on the toolbox configuration information for at least one level. Obtain skill configuration information for at least one level from the definition file of the operation manual set at at least one level, and determine the skill list corresponding to the target observable entity based on the skill configuration information for at least one level. Obtain domain knowledge configuration information for at least one level from the definition file of the operation manual set at at least one level; and determine the knowledge list corresponding to the target observable entity based on the domain knowledge configuration information for at least one level.

11. The method according to claim 10, characterized in that, If the knowledge blocks of the domain knowledge in the domain knowledge configuration information at at least one level are stored in the index storage system, then the knowledge list corresponding to the target observable entity is determined based on the domain knowledge configuration information at at least one level, including: Based on the task problem keywords associated with the task processing performed by the task intelligent agent system, keyword retrieval and vector semantic retrieval are performed in parallel in the index storage to obtain a first retrieval result and a second retrieval result; the first retrieval result includes at least one first candidate knowledge block, and the second retrieval result includes at least one second candidate knowledge block; The first search result and the second search result are merged using a reciprocal ranking fusion method to obtain a third search result, which includes at least one third candidate knowledge block; Based on the distance between the operation manual set node to which the third candidate knowledge block belongs and the entity set node to which the target observable entity belongs, and the level of the operation manual set node to which the third candidate knowledge block belongs, at least one third candidate knowledge block is sorted. The knowledge list corresponding to the target observable entity is determined based on the third candidate knowledge block ranked first.

12. An electronic device, characterized in that, include: Memory and processor; The memory is used to store computer programs; The processor is coupled to the memory for executing the computer program to perform the steps of the method according to any one of claims 1-11.

13. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it causes the processor to perform the steps of the method according to any one of claims 1-11.

14. A computer program product, characterized in that, Includes a computer program / instruction that, when executed by a processor, causes the processor to perform the steps of the method according to any one of claims 1-11.

Images