Operation and maintenance policy determination method and device based on step log, and storage medium

Abstract

The application discloses an operation and maintenance strategy determination method and device based on a step log and a storage medium, and relates to the technical field of data processing. The method is used to solve the problem that traditional operation and maintenance is separated from business and cannot quantitatively perceive business effects. The method is used to solve the problem of complex unknown business performance that has no matching preset rules. The method is used to break through the limitations of fixed rules and artificial experience. The method is used to improve the operation and maintenance capability of an intelligent operation and maintenance system when dealing with complex business problems.

Description

Technical Field

[0001] This application relates to the field of data processing technology, and in particular to a method, device and storage medium for determining operation and maintenance strategies based on step logs. Background Technology

[0002] During the process of enterprise digital transformation, a large number of domain knowledge documents are accumulated. Knowledge application systems are the core carriers for realizing the transformation of knowledge value. In order to ensure the continuous and efficient operation of knowledge application systems and stably support business operations, it is urgent to rely on a reliable intelligent operation and maintenance system to build a solid operational guarantee.

[0003] Currently, common operation and maintenance methods mainly rely on monitoring and alarms with fixed thresholds combined with human experience to handle anomalies, or on using predefined automated scripts to deal with known and well-defined basic faults. However, these methods are difficult to quantify the business performance of knowledge application systems, and cannot provide effective operation and maintenance strategies when complex or undefined business performance problems occur.

[0004] The above content is only used to help understand the technical solution of this application and does not represent an admission that the above content is prior art. Summary of the Invention

[0005] The main purpose of this application is to provide a method for determining operation and maintenance strategies based on step logs, which aims to solve the technical problem of how to improve the operation and maintenance capabilities of intelligent operation and maintenance systems when dealing with complex business problems.

[0006] To achieve the above objectives, this application proposes a method for determining operation and maintenance strategies based on step logs, the method comprising:

[0007] When an operation and maintenance event is triggered, the operation and maintenance type to which the operation and maintenance event belongs is determined based on the business tag of the target step log associated with the operation and maintenance event; If no corresponding operation and maintenance rule is found in the operation and maintenance rule base based on the operation and maintenance type, a scenario context is constructed based on the step logs within a preset time period before the event is triggered. The scenario context and the operation and maintenance objectives of the operation and maintenance event are input into a pre-trained large language model to obtain the operation and maintenance strategy output by the large language model. Based on the aforementioned operation and maintenance strategies and historical operational data prior to the event triggering, simulations are performed in the sandbox to obtain the operation and maintenance indicator data corresponding to each of the aforementioned operation and maintenance strategies. The target operation and maintenance strategy is determined based on the operation and maintenance indicator data.

[0008] In one embodiment, the step of determining the operation and maintenance type to which the operation and maintenance event belongs based on the service tag of the target step log associated with the operation and maintenance event includes: The descriptor of the operation and maintenance event is matched with the business tags of the step logs within a preset time period before the event is triggered, and the target step log associated with the operation and maintenance event is determined from the step logs. Determine the number of tags for each of the business tags corresponding to the target step log; Based on the number of tags, determine the operation and maintenance type to which the operation and maintenance event belongs.

[0009] In one embodiment, the step of constructing the scene context based on the step log within a preset time period before the event is triggered includes: Extract the entity words of the preset type and the indicator data associated with the entity words from the step log; The attribute information of the entity words and the calling relationship between each entity word are determined according to the preset service topology. The attribute information and the indicator data are used as node attributes of the entity words, and the call relationship is used as node edges of the entity words to construct a relationship graph; The relationship graph and the step log are used as the scenario context.

[0010] In one embodiment, the step of inputting the scenario context and the operation and maintenance objectives of the operation and maintenance event into a pre-trained large language model to obtain the operation and maintenance strategy output by the large language model includes: The scenario context and the operation and maintenance objectives are input into the large language model, and a list of diagnostic tasks is generated through the large language model. The execution engine is invoked to execute each task in the diagnostic task list, and the execution results are obtained. The execution result is input into the large language model to obtain the operation and maintenance strategy output by the large language model.

[0011] In one embodiment, the step of performing simulations in a sandbox based on the operation and maintenance strategies and historical operational data prior to the event triggering to obtain the operation and maintenance indicator data corresponding to each operation and maintenance strategy includes: Within the sandbox, a first environment is constructed based on the historical operational data, and a second environment is constructed based on the historical operational data and the operation and maintenance strategy. Select historical requests and execute them in the first environment and the second environment respectively to obtain the first operation and maintenance indicator data corresponding to the first environment and the second operation and maintenance indicator data corresponding to the second environment.

[0012] In one embodiment, the step of determining the target operation and maintenance strategy based on the operation and maintenance indicator data includes: The difference between the first operation and maintenance indicator data and each of the second operation and maintenance indicator data is used as the input parameter of a preset utility function, and the score of the operation and maintenance strategy corresponding to each second environment is determined by the utility function. The target operation and maintenance strategy is determined in the operation and maintenance strategy based on the score.

[0013] In one embodiment, before the step of determining the operation and maintenance type of the operation and maintenance event based on the service tag of the target step log associated with the operation and maintenance event when the operation and maintenance event is triggered, the method further includes: Obtain business indicator data from the service instance on the client, and construct a time series based on the business indicator data; Based on the business indicator values ​​at the same time within a preset period in the time series, determine the predicted indicator value and the confidence interval corresponding to the predicted indicator value at the same time in the future; Obtain the actual indicator value corresponding to the current time. If the actual indicator value is not within the confidence interval corresponding to the current time, trigger an operation and maintenance event; or In response to received operation and maintenance instructions, an operation and maintenance event is triggered.

[0014] In one embodiment, after the step of determining the target operation and maintenance strategy based on the operation and maintenance indicator data, The target operation and maintenance policy, the scenario context, and the operation and maintenance type corresponding to the operation and maintenance event are associated and stored in the operation and maintenance rule base.

[0015] Furthermore, to achieve the above objectives, this application also proposes an operation and maintenance strategy determination device based on step logs. The device includes: a memory, a processor, and a computer program stored in the memory and executable on the processor. The computer program is configured to implement the steps of the operation and maintenance strategy determination method based on step logs as described above.

[0016] In addition, to achieve the above objectives, this application also proposes a storage medium, which is a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it implements the steps of the operation and maintenance strategy determination method based on step logs as described above.

[0017] This application provides a method for determining operation and maintenance (O&M) strategies based on step logs. When an O&M event is triggered, the method can determine the O&M type of the O&M event based on the business tags of the target step log associated with the O&M event. If no corresponding O&M rule is found in the O&M rule base based on the O&M type, a scenario context is constructed based on the step logs within a preset time period before the event is triggered. The scenario context and the O&M goal of the O&M event are input into a pre-trained large language model to obtain the O&M strategy output by the large language model. Then, based on the O&M strategy and historical operation data before the event is triggered, the O&M indicator data corresponding to each O&M strategy is obtained through deduction in a sandbox. Finally, the target O&M strategy is determined based on the O&M indicator data.

[0018] After an operational event is triggered, the above method first uses business tags in the step logs to deeply associate operations and maintenance with the business scenarios of the knowledge application system, solving the problem that traditional operations and maintenance are detached from business and cannot quantify and perceive business effects. For complex and unknown business performance problems without matching preset rules, the method can construct scenario context through step logs to supplement scenario information, and then use a pre-trained large language model to achieve intelligent strategy reasoning, breaking through the limitations of fixed rules and human experience. Finally, the method uses sandbox simulation combined with historical data to quantify and evaluate the strategy effect, select the optimal solution, and thus improve the operational and maintenance capabilities of the intelligent operations and maintenance system when dealing with complex business problems. Attached Figure Description

[0019] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0020] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1 This is a flowchart illustrating an embodiment of the method for determining operation and maintenance strategies based on step logs in this application. Figure 2 This is a flowchart illustrating Embodiment 2 of the method for determining operation and maintenance strategies based on step logs in this application. Figure 3 This is an intelligent operation and maintenance architecture diagram of the operation and maintenance strategy determination method based on step logs provided in Embodiment 2 of this application; Figure 4 A simplified flowchart illustrating the operation and maintenance strategy determination method based on step logs provided in Embodiment 2 of this application; Figure 5This is a schematic diagram of the device structure of the hardware operating environment involved in the operation and maintenance strategy determination method based on step logs in the embodiments of this application.

[0022] The purpose, features, and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0023] It should be understood that the specific embodiments described herein are only used to explain the technical solutions of this application and are not intended to limit this application.

[0024] To better understand the technical solution of this application, a detailed description will be provided below in conjunction with the accompanying drawings and specific embodiments. It should be noted that all actions involving the acquisition of signals, information, or data in this application are performed in accordance with the relevant data protection laws and regulations of the country where the application is located, and with authorization from the owner of the corresponding device.

[0025] Enterprises accumulate a large amount of domain knowledge documents during their digital transformation, and knowledge application systems are a key support for realizing the transformation of knowledge value. To ensure the continuous and efficient operation of knowledge application systems and their stable support for business operations, they need to have reliable automated operation and maintenance capabilities.

[0026] Currently, common operation and maintenance methods mainly rely on monitoring and alarms with fixed thresholds combined with human experience to handle anomalies, or on using predefined automated scripts to deal with known and well-defined basic faults. However, these methods are difficult to quantify the business performance of knowledge application systems, and cannot provide effective operation and maintenance strategies when complex or undefined business performance problems occur.

[0027] In view of the above problems, this application proposes an operation and maintenance strategy determination method based on step logs. This method first deeply associates operation and maintenance with the business scenarios of knowledge application systems through business tags in the step logs, solving the problem that traditional operation and maintenance is detached from business and cannot quantify and perceive business effects. For complex and unknown business performance problems without matching preset rules, the method can construct scenario context through step logs to supplement scenario information, and then use a pre-trained large language model to realize intelligent strategy reasoning, breaking through the limitations of fixed rules and human experience. Finally, the method uses sandbox inference combined with historical data to quantitatively evaluate the strategy effect and select the optimal solution, thereby improving the operation and maintenance capabilities of the intelligent operation and maintenance system when dealing with complex business problems.

[0028] It should be noted that the executing entity in this embodiment can be a computing service device with data processing, network communication, and program execution functions, such as a tablet computer, personal computer, or mobile phone, or an electronic device or intelligent operation and maintenance system capable of performing the above functions. The following description uses an intelligent operation and maintenance system as an example to illustrate the various embodiments.

[0029] First Embodiment The first embodiment of this application provides a method for determining operation and maintenance strategies based on step logs, referring to... Figure 1 In this embodiment, the method for determining operation and maintenance strategies based on step logs includes steps S10 to S50: Step S10: When an operation and maintenance event is triggered, determine the operation and maintenance type to which the operation and maintenance event belongs based on the business tag of the target step log associated with the operation and maintenance event.

[0030] Operational events refer to events that trigger the intelligent operations and maintenance (O&M) system to initiate processing flows. These events can be triggered automatically by anomaly alarms generated by the dynamic anomaly detection engine within the intelligent O&M system based on real-time detected business indicator data, or they can be manually triggered by O&M personnel or users. Step logs record detailed operational information for each call to service components throughout the knowledge application chain, such as retrieval, reordering, and generation components. This includes, but is not limited to, key fields such as component name, input, output, time consumption, status code, and tracking code. In this embodiment, the knowledge application chain refers to a complete end-to-end knowledge processing and response process. When a user raises a question or request, the intelligent O&M system needs to work collaboratively through multiple steps to ultimately generate an accurate answer or result. This complete process, from "user asking a question" to "system providing the final answer," constitutes the knowledge application chain. Service components are independent modules with specific artificial intelligence or data processing functions that constitute the aforementioned knowledge application chain. Each service component is responsible for a specific task within the knowledge application chain, such as retrieval and sorting.

[0031] Business tags are metadata attached to the step logs and have business semantics. They are used to quickly filter and aggregate massive logs from a business perspective.

[0032] Optionally, as an alternative solution for injecting business tags into the step log, when the intelligent operation and maintenance system processes a specific business request, business information generated during the processing of the business request, such as tenant identifier, usage channel, service component information, and whether the service component call was successful, can be written into the log as tags.

[0033] Alternatively, as another option for injecting business tags into the step log, users can specify preset business tags when creating a service instance for a business request.

[0034] Furthermore, step S10 may include steps S11 to S13: Step S11: Match the descriptor of the operation and maintenance event with the business tags of the step logs within a preset time period before the event is triggered, and determine the target step log associated with the operation and maintenance event from the step logs.

[0035] A descriptor for an operations and maintenance (O&M) event refers to a textual summary of the core characteristics of the O&M event. As an optional method for generating descriptors, when an O&M event is triggered, the intelligent O&M system parses the O&M objectives (such as "restore knowledge retrieval success rate") from abnormal alarm information or user-sent O&M commands, the component names of the service components that are abnormal or require optimization, and abnormal business indicator data detected by the dynamic anomaly detection engine. The aforementioned O&M objectives, component names, and abnormal business indicator data are then used as the descriptor for the O&M event.

[0036] Furthermore, the step of matching the descriptor of the operation and maintenance event with the business tags of the step logs within a preset time period before the event is triggered may include: extracting entity words from the descriptor, performing text matching of these entity words with the business tags of all step logs, and taking all successfully matched step logs as the target step logs related to the operation and maintenance event.

[0037] Step S12: Determine the number of tags for each of the business tags corresponding to the target step log.

[0038] Step S13: Determine the operation and maintenance type to which the operation and maintenance event belongs based on the number of tags.

[0039] For example, read the business tags for each target step log, count the occurrences of each type of business tag, and obtain the tag count for each type of business tag. Determine the proportion of each business tag's tag count to the total number of business tags, and select the k business tags with the highest proportion, or sorted by proportion from largest to smallest, as the operation and maintenance type to which the operation and maintenance event belongs.

[0040] Optionally, to improve the processing efficiency and policy determination accuracy of subsequent steps, before step S12, all target step logs can be aggregated based on the tenant identifier. If the proportion of target step logs corresponding to a certain target tenant identifier to the total number of target step logs is greater than or equal to a preset proportion threshold, then the anomaly is determined to be mainly concentrated in that target tenant. Step logs corresponding to other tenant identifiers in the target step logs are deleted, and the step logs corresponding to the target tenant identifier are retained as the target step logs for the operation and maintenance event.

[0041] Furthermore, if the proportion of the number of target step logs corresponding to a certain target tenant identifier to the total number of all target step logs is less than a preset proportion threshold, then when executing the above step S12, the business tags of the target step logs are matched according to the operation and maintenance type tags corresponding to each operation and maintenance type item on the preset checklist. Based on the number of matches for the business tags corresponding to each operation and maintenance type item, the operation and maintenance type to which the operation and maintenance event belongs is determined. Specifically, the operation and maintenance type item with the most matches, or k operation and maintenance type items sorted from largest to smallest match count, can be selected as the operation and maintenance type to which the operation and maintenance event belongs.

[0042] Step S20: If no corresponding operation and maintenance rule is found in the operation and maintenance rule base based on the operation and maintenance type, then construct a scenario context based on the step log within a preset time period before the event is triggered.

[0043] The operations and maintenance (O&M) rule base is a pre-stored, structured collection of knowledge formed from historical O&M cases. Each O&M rule defines a specific O&M type and its corresponding O&M strategy. When the intelligent O&M system detects that the O&M type of the aforementioned O&M event matches a certain O&M rule, it can quickly invoke the existing O&M strategy. If no corresponding O&M rule is found in the O&M rule base for the O&M type, a scenario context can be constructed based on the step logs within a preset time period before the event is triggered, to support the subsequent intelligent generation of O&M strategies.

[0044] Furthermore, step S20 above may include steps S21 to S24: Step S21: Extract the entity words of the preset type and the indicator data associated with the entity words from the step log.

[0045] Predefined entity terms refer to words extracted from the step log that are predefined as key elements or objects of a specific type. These may include component names, configuration items, data sources, database table names, and knowledge base identifiers of service components. The associated metrics data for these entity terms are quantitative values ​​reflecting their running status or performance, linked to the same step log entry. These metrics include time consumption, status codes, and internal parameters, representing the performance of the corresponding entity at a specific call time.

[0046] For example, based on a preset list of entity types, natural language processing techniques, such as named entity recognition or rule matching based on fixed fields, are used to identify and extract entity words belonging to preset types from the text content of the step log, such as component name fields, input fields, output fields, and error message fields. For instance, the entity words "vector retrieval service" (type: service component name) and "v2.1" (type: configuration item - version) are extracted from the log message "failed to call vector retrieval component v2.1". At the same time, the corresponding metric data, such as the time taken for this call: "response_time:150ms" and the status code: status_code:500, are read from the structured fields of the same log.

[0047] Step S22: Determine the attribute information of the entity words and the calling relationship between each entity word according to the preset service topology relationship.

[0048] Service topology relationships can be stored in the intelligent operation and maintenance system in the form of graphs or database tables. This describes the call dependencies between various entities within the system, as well as the associated attribute information of each entity. For example, when designing the system architecture, it is determined that "frontend service A can call backend service B" or "service C depends on database D." Attribute information is the inherent state and configuration metadata of the service components themselves, including but not limited to deployment information such as IP address, hostname, port, and cluster affiliation; version information such as service version number, software package version, and configuration version; and attribution information such as the business domain, responsible person, and team.

[0049] For example, step S22 may specifically include: traversing each entity word of a preset type in the step log, using it as a query identifier, directly initiating a query to the database or graph system of the service topology relationship, searching for the node record or field corresponding to the entity word in the service topology relationship, reading all its associated attribute information, determining the target node or target field corresponding to each entity word in the service topology relationship, and determining the calling relationship between the target nodes or target fields in the service topology relationship, as the calling relationship between the aforementioned entity words.

[0050] Step S23: Use the attribute information and the indicator data as node attributes of the entity words, and use the call relationship as node edges of the entity words to construct a relationship graph.

[0051] Step S24: Use the relationship graph and the step log as the scenario context.

[0052] As an optional approach to constructing the scenario context, each entity word is treated as a node in the relationship graph. The attribute information of the entity words determined from the service topology and their associated indicator data are used as node attributes in the relationship graph. The call relationships between the entity words determined from the service topology are used as node edges in the relationship graph, thus constructing the aforementioned relationship graph. The obtained relationship graph and the original step log are then packaged into a preset format to form the scenario context.

[0053] As an alternative approach to constructing a scenario context, the step logs within a preset time period prior to the event trigger can be quantitatively analyzed and aggregated based on business tags to generate the scenario context. Specifically, the step logs within the preset time period prior to the event trigger can be categorized according to business tags, and the distribution ratio of step logs corresponding to each tag category can be calculated, along with the request volume, success rate, and average latency recorded in the step logs for each time period, as well as the failure rate and success rate of each service component. A statistical report is generated based on the above information, and this statistical report, along with the original step logs, serves as the scenario context.

[0054] Step S30: Input the scenario context and the operation and maintenance objectives of the operation and maintenance event into the pre-trained large language model to obtain the operation and maintenance strategy output by the large language model.

[0055] For example, the aforementioned scenario context and operational goals are encapsulated into prompts according to a preset template. The structure of the prompts typically includes role settings, operational goals, scenario context content, and format instructions for the model's output. Then, the encapsulated prompts are submitted to a pre-trained large language model via API calls or other means. The large language model infers from the input prompts based on its trained model parameters and outputs operational strategies.

[0056] The aforementioned large language model can directly use pre-trained general-purpose large language models, such as OpenAI's GPT series, Anthropic's Claude series, and Google's Gemini series. Alternatively, it can adopt open-source or privately deployable large models, such as Meta's Llama series and Tsinghua University's ChatGLM series. Based on the model weights provided by these models, the model weights are updated by training the model using historical operation and maintenance cases to obtain the aforementioned large language model.

[0057] Furthermore, step S30 above may include steps S31 to S33: Step S31: Input the scenario context and the operation and maintenance goal into the large language model, and generate a list of diagnostic tasks through the large language model.

[0058] For example, a preset diagnostic task list template is input as a prompt, and the large language model is required to use the aforementioned diagnostic task list template as its output format. The large language model interface is called, the aforementioned prompt is sent, and after receiving the natural language response from the large language model, the response content is parsed, and the diagnostic task list conforming to the output format is extracted.

[0059] Step S32: Invoke the execution engine to execute each task in the diagnostic task list and obtain the execution result.

[0060] The diagnostic task list generated in step S31 is transmitted to the execution engine, and a task execution instruction is issued to the execution engine, determining the execution time and data collection scope. Upon receiving the diagnostic task list and execution instruction, the execution engine first parses the core requirements, execution objects, and data collection scope of each diagnostic task, and then starts execution sequentially according to task priority. For each diagnostic task, the execution engine calls the corresponding functional modules of the intelligent operation and maintenance system. For example, for a service component running status diagnostic task, it calls the intelligent operation and maintenance system's collection probes or SDK to collect business indicator data, status codes, and time consumption data of the corresponding service component; for a log anomaly investigation task, it calls the log parsing module to extract key field information from the target step logs; for a service dependency verification task, it calls the service topology relation library to extract the call dependency data of the corresponding entity words. During task execution, the execution engine records the execution status and results of each task in real time.

[0061] Step S33: Input the execution result into the large language model to obtain the operation and maintenance strategy output by the large language model.

[0062] The execution results of each task in the above diagnostic task list are input into the large language model again. The large language model determines and outputs the operation and maintenance strategy based on the above execution results and the previous prompt words.

[0063] Step S40: Based on the operation and maintenance strategy and historical operation data before the event is triggered, a simulation is performed in the sandbox to obtain the operation and maintenance indicator data corresponding to each operation and maintenance strategy.

[0064] Historical operational data prior to an event trigger refers to traffic data records generated by the intelligent operations and maintenance (O&M) system in a real production environment within a period prior to the triggering of an O&M event. This includes real user requests, corresponding step logs, and the business context at that time. A sandbox is a simulation testing environment isolated from the production environment, used to test the effectiveness of O&M strategies. O&M metric data refers to the business metric results obtained after simulating the execution of the aforementioned O&M strategies in the sandbox environment using technical means and leveraging historical operational data to drive the test. These O&M metric data are used to quantitatively evaluate the effectiveness of each O&M strategy and may include technical performance metrics such as average response time, error rate, and throughput, as well as business performance metrics such as retrieval success rate and answer accuracy.

[0065] Furthermore, step S40 above may include steps S41 to S42: Step S41: In the sandbox, a first environment is constructed based on the historical operation data, and a second environment is constructed based on the historical operation data and the operation and maintenance strategy.

[0066] For example, the steps of constructing the first environment based on historical operational data may specifically include: firstly, analyzing the historical operational data used for simulation, and extracting a snapshot of the system state at the corresponding point in time from the historical operational data, including: determining the version number of each service component when processing historical requests; obtaining the versions or content of various configuration files and parameter configurations that were in effect when processing historical requests, such as model parameters, timeout thresholds, business rules, etc.; and determining the versions or data content of data sources such as the knowledge base, vector database, and business database at that time. In an isolated sandbox, based on the above historical operational data, deploying the images or code of each service component of the corresponding version, pulling the configuration files and parameter configurations of the corresponding version from the configuration repository and applying them, rolling back or switching the data source to the specified version, and obtaining a sandbox environment instance consistent with the production environment state before the operational event was triggered, i.e., the first environment.

[0067] For example, the steps for constructing a second environment based on historical operational data and operation and maintenance policies include: first, using the first environment as a template, creating a new, identical sandbox environment instance; then, reading the specific operation instructions defined in the operation and maintenance policy; and executing these operation instructions within the policy to obtain the second environment. These operation instructions include updating a parameter to a new value recommended by the operation and maintenance policy, rolling back or upgrading the version of a specific service component to the version specified by the policy, enabling, disabling, or modifying a business rule or traffic scheduling rule, and adjusting the CPU and memory quotas allocated to a service component, etc.

[0068] Step S42: Select historical requests and execute the historical requests in the first environment and the second environment respectively to obtain the first operation and maintenance indicator data corresponding to the first environment and the second operation and maintenance indicator data corresponding to the second environment.

[0069] From the historical operational data used to build the environment, select historical requests covering different business scenarios, request types, and load periods. Re-execute these historical requests sequentially or in parallel in the first environment, detecting and recording the processing results of each historical request, and statistically analyzing first-level operational metrics, such as average response time, success rate, and time distribution of each service component. Then, re-execute the same historical requests in the second environment, similarly statistically analyzing second-level operational metrics.

[0070] Step S50: Determine the target operation and maintenance strategy based on the operation and maintenance indicator data.

[0071] After obtaining the operation and maintenance indicator data corresponding to each operation and maintenance strategy through simulation in the sandbox, the operation and maintenance indicator data corresponding to each operation and maintenance strategy can be compared. The operation and maintenance indicator data can be weighted and summed with the corresponding preset weights to determine the score corresponding to each operation and maintenance strategy. The operation and maintenance strategy with the highest score is determined as the target operation and maintenance strategy.

[0072] Optionally, the scores corresponding to the operation and maintenance strategies can be sorted in descending order, and the top k operation and maintenance strategies can be output as a combination of target operation and maintenance strategies for users to choose from.

[0073] Furthermore, step S50 above may include steps S51 to S52: Step S51: The difference between the first operation and maintenance indicator data and each of the second operation and maintenance indicator data is used as the input parameter of a preset utility function, and the score of the operation and maintenance strategy corresponding to each second environment is determined through the utility function.

[0074] Preferably, the utility function can be expressed as U = a × S_e + b × S_r + c × S_c. Where U represents the score of the operation and maintenance strategy; S_e represents the expected effect score of the operation and maintenance strategy, used to quantify the degree to which the strategy improves the operation and maintenance objectives and related positive indicators; S_r represents the risk score of the operation and maintenance strategy, used to quantify the probability and severity of the potential negative impacts of the strategy; S_c represents the execution cost score of the operation and maintenance strategy, used to quantify the resource costs required to implement the strategy; a, b, and c are weighting coefficients, representing the relative importance of the expected effect score, risk score, and cost score in the final decision-making process, respectively.

[0075] Further, the expression for S_e above can be S_e=Σ(w_i×Normalize(ΔM_i)). Here, ΔM_i is the change value of the i-th operational indicator data in the simulation. ΔM_i=M_i_post-M_i_pre. Here, M_i_pre is the first operational indicator data in the first environment, and M_i_post is the corresponding second operational indicator data in the second environment. For positive indicators such as success rate and accuracy, ΔM_i is positive, indicating improvement. For negative indicators such as error rate and average latency, ΔM_i is negative, indicating improvement. Negative or absolute values ​​can be used in the calculation to unify the direction. Normalize(ΔM_i) is a normalization function that maps the change values ​​of indicators with different dimensions to the [0, 1] score interval. w_i is the weight of the i-th operational indicator data.

[0076] The expression for S_r above can be S_r = Σ(p_j × s_j). Here, p_j represents the probability of the j-th identified risk item occurring, such as triggering a new error type, causing performance degradation, or generating data inconsistency. This probability can be estimated based on historical data, the complexity of configuration changes, or the frequency of anomalous events in model simulations. s_j represents the severity score of the j-th risk item if it occurs.

[0077] The expression for S_c above can be S_c = w_time × C_time + w_human × C_human + w_resource × C_resource. Here, w_time represents the time cost, which is the total time required from the start of implementing the operation and maintenance strategy to obtaining the execution result; C_human represents the human resource cost, which is the number of person-hours required for different functional personnel such as development, operation and maintenance, and testing to implement the operation and maintenance strategy; and C_resource represents the computing / resource cost, which is the additional CPU, memory, storage, or API call cost consumed in implementing the operation and maintenance strategy. w_time, w_human, and w_resource represent the weights of time, human resource, and resource costs in the total cost calculation, respectively.

[0078] Step S52: Determine the target operation and maintenance strategy in the operation and maintenance strategy based on the score.

[0079] After calculating the score for each operation and maintenance strategy, the operation and maintenance strategy with the highest score is selected as the target operation and maintenance strategy, or the operation and maintenance strategies are sorted in descending order of their scores, and the top k operation and maintenance strategies are selected as the target operation and maintenance strategies.

[0080] Furthermore, after determining the target operation and maintenance strategy based on the operation and maintenance indicator data, the target operation and maintenance strategy, scenario context, and operation and maintenance type corresponding to the operation and maintenance event are associated and transformed into reusable scenario-based rules, which are stored in the operation and maintenance rule library so that when similar operation and maintenance events are encountered in the future, the target operation and maintenance strategy can be directly matched and reused from the operation and maintenance rule library.

[0081] In this embodiment, the intelligent operation and maintenance system quickly determines the operation and maintenance type based on the business tags in the step logs associated with the operation and maintenance events. If a historical experience rule base cannot be matched, a scenario context containing information such as a relationship graph is automatically constructed, and a large language model is driven to generate candidate operation and maintenance strategies. Subsequently, the strategies are deduced and verified based on historical operational data in a secure sandbox environment, and their expected effects, risks, and costs are quantified. Finally, each strategy is comprehensively scored through a preset utility function, and the optimal strategy is selected as the target operation and maintenance strategy. This embodiment can transform the traditional passive response-based operation and maintenance that relies on human experience into a proactive decision-making closed loop driven by data and intelligence. It not only improves the accuracy of problem location and the adaptability of strategy generation, but also achieves risk prevention through sandbox deduction. By accumulating verified and effective strategies into the rule base, the intelligent operation and maintenance system has the ability to continuously self-optimize, thereby significantly reducing operation and maintenance costs and ensuring the high quality and stability of knowledge services.

[0082] Second Embodiment Based on the first embodiment described above, in the operation and maintenance strategy determination method based on step logs provided in this embodiment, refer to Figure 2 Before step S10 above, the method for determining operation and maintenance strategies based on step logs further includes steps S60 to S90: Step S60: Obtain business indicator data from the service instance of the client, and construct a time series based on the business indicator data.

[0083] A service instance refers to a specific process or container that is actually deployed and running in the client and can independently provide services. It is the runtime entity of each intelligent component that constitutes the entire knowledge application chain.

[0084] For example, the intelligent operation and maintenance system continuously collects business indicator data, including response time, error rate, retrieval rejection rate, and answer accuracy, through a collection probe or SDK deployed on client service instances at a fixed sampling period. Each time a data point is collected, it is arranged in chronological order to construct a time series.

[0085] Step S70: Based on the business indicator values ​​at the same time within a preset period in the time series, determine the predicted indicator value at the same time in the future and the confidence interval corresponding to the predicted indicator value.

[0086] For example, the intelligent operation and maintenance system analyzes the business indicator values ​​at the same time each day, such as 10:00 AM, within the aforementioned time series over several preset periods, such as the past 7 calendar days. By performing statistical analysis on these historical business indicator values ​​at the same time, such as calculating the mean and standard deviation, or using more complex time series forecasting models, the system calculates the predicted indicator values ​​for the same time in the future. Simultaneously, based on the fluctuation range of historical business indicator values, the system calculates the confidence interval for the predicted indicator value to represent the possible fluctuation range of the predicted indicator value.

[0087] As an alternative approach to determining the predictive index value for the same future moment, the STL (Seasonal-Trend Decomposition using Loess) algorithm is applied to decompose the aforementioned time series into trend components, periodic components, and residual components. From the periodic components, the periodic component value corresponding to each complete historical period, such as 3:00 PM each day in the past 28 days, is extracted to form a subsequence S_historical. The exponential decay function is applied to the period index corresponding to each data point in the subsequence S_historical to calculate the decay weight of each period. The formula for calculating the decay weight W_time_i is W_time_i = λ^i, where λ is the decay factor (0 < λ < 1), and i is the period index; the further back in time the period, the smaller its decay weight. Next, taking the same future moment, such as 3:00 PM, as the center, a time period is extended forward and backward (e.g., 1 hour before and after, totaling 13 data points), forming a theoretical target window Window_target. Since this is a prediction of the future, the average value of the periodic components of windows at the same time in neighboring periods can be used to approximate the shape of this target window. Specifically, the dynamic time warping distance between each historical window (Window_historical_i) and the theoretical target window (Window_target) is calculated. A smaller dynamic time warping distance indicates more similar fluctuation patterns of the data points within the two windows. Then, the dynamic time warping distance is converted into a similarity weight W_shape_i = exp(-γ × D_i). Here, γ is a scaling parameter, γ > 0, used to control the sensitivity of the distance to the weight's influence. Afterward, based on the aforementioned decay weight, similarity weight, and the specific periodic component value C_i of each period in the subsequence S_historical at the same time, the predicted index value for a future time period is determined. Specifically, for each historical period i, its comprehensive weight W_i = W_time_i is calculated. W_shape_i. Normalize all the comprehensive weights so that Σ(W_i)=1 to obtain the final normalized weight W_i_norm, and then calculate the predicted index value Pred_A=Σ(W_i_norm×C_i) at the same time in the future.

[0088] Based on the above optional schemes, the steps for calculating the confidence interval corresponding to the predicted index value at the same future time may include: extracting the residual sequence from the residual components and calculating the standard deviation σ_R of the residual sequence; setting a preset confidence level and its corresponding Z-score. For example, a confidence level of 95% corresponds to a Z-score of 1.96. Calculating the confidence interval CI_A = [Pred_A - Z × σ_R, Pred_A + Z × σ_R] based on the above standard deviation and predicted index value, where Z represents the Z-score corresponding to the preset confidence level.

[0089] Step S80: Obtain the actual indicator value corresponding to the current time. If the actual indicator value is not within the confidence interval corresponding to the current time, then trigger an operation and maintenance event.

[0090] The collected real indicator value at the current moment is compared with the corresponding confidence interval. If the real indicator value is not within this confidence interval, that is, less than the lower limit of the confidence interval or greater than the upper limit of the confidence interval, it means that the real indicator value is not a normal fluctuation with a high probability. Therefore, it is judged as abnormal and an operation and maintenance event is triggered.

[0091] As another optional approach to triggering maintenance events, step S90 may be included before step S10: triggering a maintenance event in response to a received maintenance instruction. This maintenance instruction includes the maintenance objective, the service component or functional module requiring maintenance, and a corresponding event description.

[0092] For example, to help understand the implementation flow of the operation and maintenance strategy determination method based on step logs obtained by combining this embodiment with the above embodiment one, please refer to... Figure 3 , Figure 3 A smart operations and maintenance architecture diagram is provided for a method of determining operations and maintenance strategies based on step logs, specifically: The intelligent operation and maintenance system includes a log collection layer, an intelligent analysis layer, a decision engine layer, and a continuous optimization layer.

[0093] The log collection layer uniformly specifies the standard for step log data that all service components must output, including component name, input / output, time consumption, status code, key internal parameters, and trace codes. The log collection layer uses collection probes or SDKs to collect step logs and business indicator data generated by the client during the execution of services by each service component in the knowledge application process in real time.

[0094] The intelligent analytics layer centrally processes massive amounts of step logs. The log hub stores all step logs and provides query capabilities. The dynamic anomaly detection engine monitors in real time whether the current real-time indicator value is within the corresponding confidence interval. The tag correlation analyzer supports rapid filtering, aggregation, statistics, and analysis of step logs by multi-dimensional business tags such as scenario and tenant.

[0095] The decision engine layer includes a rule engine and an intelligent agent engine. The rule engine matches the type of operation and maintenance (O&M) event against an O&M rule base. The intelligent agent engine constructs a scenario context based on step logs within a preset time period before the event is triggered, and inputs the scenario context and the O&M event's O&M objectives into a pre-trained large language model to obtain the O&M strategy output by the large language model.

[0096] The continuous optimization layer includes a strategy execution and evaluation module, a knowledge asset accumulation module, and a baseline self-update module. The strategy execution and evaluation module uses the operation and maintenance (O&M) strategies output by the large language model and historical operational data before the event trigger to deduce the corresponding O&M indicator data for each O&M strategy in a sandbox, and then determines the target O&M strategy based on the O&M indicator data. The knowledge asset accumulation module associates the target O&M strategy, scenario context, and O&M type corresponding to the O&M event, transforming it into reusable scenario-based rules, and stores them in the O&M rule base. The baseline self-update module constructs a time series in real time based on business indicator data, and determines the predicted indicator value and its corresponding confidence interval for the same future time based on the business indicator values ​​at the same time within a preset period in the time series.

[0097] Furthermore, referring to Figure 4 , Figure 4 A simplified flowchart of the operation and maintenance strategy determination method based on step logs in this application is provided.

[0098] Specifically, an operational event is first triggered, which can be automatically alerted by the dynamic anomaly detection engine or manually initiated by human commands. Then, the intelligent operations and maintenance system first attempts to match the operational event type against the operations and maintenance rule base. If historical operational cases of the same type exist, historical experience rules can be directly reused to output a matching target operations and maintenance strategy. If no match is found, a scenario context is constructed and an operations and maintenance strategy is generated based on it. The strategy is then deduced, and a score is determined based on the deduction results. The target operations and maintenance strategy is then determined based on the score. If the score is lower than a preset score threshold, the operations and maintenance strategy is regenerated.

[0099] It should be noted that the above examples are only for understanding this application and do not constitute a limitation on the method for determining operation and maintenance strategies based on step logs in this application. Any simple modifications based on this technical concept are within the protection scope of this application.

[0100] This application provides a step log-based operation and maintenance strategy determination device, which includes: 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 executed by the at least one processor to enable the at least one processor to execute the step log-based operation and maintenance strategy determination method in the above embodiment 1.

[0101] The following is for reference. Figure 5 The diagram illustrates a structural schematic of a device suitable for implementing the operation and maintenance strategy determination based on step logs in the embodiments of this application. The operation and maintenance strategy determination device based on step logs in the embodiments of this application may include, but is not limited to, mobile terminals such as laptops and tablets, and fixed terminals such as desktop computers. Figure 5 The illustrated operation and maintenance strategy based on step logs for determining the device is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of this application.

[0102] like Figure 5 As shown, the operation and maintenance strategy determination device based on the step log may include a processing unit 1001 (e.g., a central processing unit, a graphics processing unit, etc.), which can perform various appropriate actions and processes according to a program stored in read-only memory (ROM) 1002 or a program loaded from storage device 1003 into random access memory (RAM) 1004. The random access memory 1004 also stores various programs and data required for the operation of the operation and maintenance strategy determination device based on the step log. The processing unit 1001, ROM 1002, and RAM 1004 are interconnected via bus 1005. Input / output (I / O) interface 1006 is also connected to the bus. Typically, the following systems can be connected to I / O interface 1006: input devices 1007 including, for example, touchscreens, touchpads, keyboards, mice, image sensors, microphones, accelerometers, gyroscopes, etc.; output devices 1008 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; storage devices 1003 including, for example, magnetic tapes, hard disks, etc.; and communication devices 1009. Communication device 1009 allows the step-log-based maintenance strategy determination device to communicate wirelessly or wiredly with other devices to exchange data. Although a step-log-based maintenance strategy determination device with various systems is shown in the figure, it should be understood that it is not required to implement or have all the systems shown. More or fewer systems can be implemented alternatively.

[0103] Specifically, according to the embodiments disclosed in this application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments disclosed in this application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device, or installed from storage device 1003, or installed from read-only memory 1002. When the computer program is executed by processing device 1001, it performs the functions defined in the methods of the embodiments disclosed in this application.

[0104] The operation and maintenance strategy determination device based on step logs provided in this application, employing the operation and maintenance strategy determination method based on step logs in the above embodiments, can solve the technical problem of how to improve the operation and maintenance capabilities of intelligent operation and maintenance systems when dealing with complex business issues. Compared with the prior art, the beneficial effects of the operation and maintenance strategy determination device based on step logs provided in this application are the same as the beneficial effects of the operation and maintenance strategy determination method based on step logs provided in the above embodiments, and other technical features in this operation and maintenance strategy determination device are the same as those disclosed in the method of the previous embodiment, and will not be repeated here.

[0105] It should be understood that the various parts disclosed in this application can be implemented using hardware, software, firmware, or a combination thereof. In the description of the above embodiments, specific features, structures, materials, or characteristics can be combined in any suitable manner in one or more embodiments or examples.

[0106] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0107] This application provides a computer-readable storage medium having computer-readable program instructions (i.e., a computer program) stored thereon, which are used to execute the operation and maintenance strategy determination method based on step logs in the above embodiments.

[0108] The computer-readable storage medium provided in this application may be, for example, a USB flash drive, but is not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to: electrical connections having one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), or flash memory, optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, system, or device. The program code contained on the computer-readable storage medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, radio frequency (RF), etc., or any suitable combination thereof.

[0109] The aforementioned computer-readable storage medium may be included in the operation and maintenance policy determination device based on the step log; or it may exist independently and not be assembled into the operation and maintenance policy determination device based on the step log.

[0110] The aforementioned computer-readable storage medium carries one or more programs that, when executed by a step-log-based operation and maintenance policy determination device, enable the step-log-based operation and maintenance policy determination device to write computer program code for performing the operations of this application in one or more programming languages ​​or a combination thereof. These programming languages ​​include object-oriented programming languages ​​such as Java, Smalltalk, and C++, as well as conventional procedural programming languages ​​such as C or similar languages. The program code can be executed entirely on the user's computer, partially on the user's computer, or as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0111] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, may be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0112] The modules described in the embodiments of this application can be implemented in software or hardware. The names of the modules do not necessarily limit the functionality of the unit itself.

[0113] The readable storage medium provided in this application is a computer-readable storage medium that stores computer-readable program instructions (i.e., a computer program) for executing the above-described method for determining operation and maintenance strategies based on step logs. This addresses the technical problem of how to improve the operation and maintenance capabilities of intelligent operation and maintenance systems when dealing with complex business issues. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided in this application are the same as those of the operation and maintenance strategy determination method based on step logs provided in the above embodiments, and will not be repeated here.

[0114] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the operation and maintenance strategy determination method based on step logs as described above.

[0115] The computer program product provided in this application can solve the technical problem of how to improve the operation and maintenance capabilities of intelligent operation and maintenance systems when dealing with complex business problems. Compared with the prior art, the beneficial effects of the computer program product provided in this application are the same as the beneficial effects of the operation and maintenance strategy determination method based on step logs provided in the above embodiments, and will not be repeated here.

[0116] The above description is only a part of the embodiments of this application and does not limit the patent scope of this application. All equivalent structural transformations made under the technical concept of this application and using the contents of the specification and drawings of this application, or direct / indirect applications in other related technical fields, are included in the patent protection scope of this application.

Claims

1. A method for determining operation and maintenance strategies based on step logs, characterized in that, The method for determining operation and maintenance strategies based on step logs includes the following steps: When an operation and maintenance event is triggered, the operation and maintenance type to which the operation and maintenance event belongs is determined based on the business tag of the target step log associated with the operation and maintenance event; If no corresponding operation and maintenance rule is found in the operation and maintenance rule base based on the operation and maintenance type, a scenario context is constructed based on the step logs within a preset time period before the event is triggered. The scenario context and the operation and maintenance objectives of the operation and maintenance event are input into a pre-trained large language model to obtain the operation and maintenance strategy output by the large language model. Based on the aforementioned operation and maintenance strategies and historical operational data prior to the event triggering, simulations are performed in the sandbox to obtain the operation and maintenance indicator data corresponding to each of the aforementioned operation and maintenance strategies. The target operation and maintenance strategy is determined based on the operation and maintenance indicator data.

2. The method for determining operation and maintenance strategies based on step logs as described in claim 1, characterized in that, The step of determining the operation and maintenance type of the operation and maintenance event based on the service tag of the target step log associated with the operation and maintenance event includes: The descriptor of the operation and maintenance event is matched with the business tags of the step logs within a preset time period before the event is triggered, and the target step log associated with the operation and maintenance event is determined from the step logs. Determine the number of tags for each of the business tags corresponding to the target step log; Based on the number of tags, determine the operation and maintenance type to which the operation and maintenance event belongs.

3. The method for determining operation and maintenance strategies based on step logs as described in claim 1, characterized in that, The step of constructing the scene context based on the step log within a preset time period before the event is triggered includes: Extract the entity words of the preset type and the indicator data associated with the entity words from the step log; The attribute information of the entity words and the calling relationship between each entity word are determined according to the preset service topology. The attribute information and the indicator data are used as node attributes of the entity words, and the call relationship is used as node edges of the entity words to construct a relationship graph; The relationship graph and the step log are used as the scenario context.

4. The method for determining operation and maintenance strategies based on step logs as described in claim 1, characterized in that, The step of inputting the scenario context and the operation and maintenance objectives of the operation and maintenance event into a pre-trained large language model to obtain the operation and maintenance strategy output by the large language model includes: The scenario context and the operation and maintenance objectives are input into the large language model, and a list of diagnostic tasks is generated through the large language model. The execution engine is invoked to execute each task in the diagnostic task list, and the execution results are obtained. The execution result is input into the large language model to obtain the operation and maintenance strategy output by the large language model.

5. The method for determining operation and maintenance strategies based on step logs as described in claim 1, characterized in that, The step of performing simulations in a sandbox based on the operation and maintenance strategies and historical operational data prior to the event triggering to obtain the operation and maintenance indicator data corresponding to each operation and maintenance strategy includes: Within the sandbox, a first environment is constructed based on the historical operational data, and a second environment is constructed based on the historical operational data and the operation and maintenance strategy. Select historical requests and execute them in the first environment and the second environment respectively to obtain the first operation and maintenance indicator data corresponding to the first environment and the second operation and maintenance indicator data corresponding to the second environment.

6. The method for determining operation and maintenance strategies based on step logs as described in claim 5, characterized in that, The step of determining the target operation and maintenance strategy based on the operation and maintenance indicator data includes: The difference between the first operation and maintenance indicator data and each of the second operation and maintenance indicator data is used as the input parameter of a preset utility function, and the score of the operation and maintenance strategy corresponding to each second environment is determined by the utility function. The target operation and maintenance strategy is determined in the operation and maintenance strategy based on the score.

7. The method for determining operation and maintenance strategies based on step logs as described in any one of claims 1 to 6, characterized in that, Before determining the operation and maintenance type of the operation and maintenance event based on the business tag of the target step log associated with the operation and maintenance event when the operation and maintenance event is triggered, the method further includes: Obtain business indicator data from the service instance on the client, and construct a time series based on the business indicator data; Based on the business indicator values ​​at the same time within a preset period in the time series, determine the predicted indicator value and the confidence interval corresponding to the predicted indicator value at the same time in the future; Obtain the actual indicator value corresponding to the current time. If the actual indicator value is not within the confidence interval corresponding to the current time, trigger an operation and maintenance event; or In response to received operation and maintenance instructions, an operation and maintenance event is triggered.

8. The method for determining operation and maintenance strategies based on step logs as described in claim 1, characterized in that, After the step of determining the target operation and maintenance strategy based on the operation and maintenance indicator data... The target operation and maintenance policy, the scenario context, and the operation and maintenance type corresponding to the operation and maintenance event are associated and stored in the operation and maintenance rule base.

9. A method for determining equipment based on step logs for operation and maintenance strategies, characterized in that, The device includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the steps of the operation and maintenance strategy determination method based on step logs as described in any one of claims 1 to 8.

10. A storage medium, characterized in that, The storage medium is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, it implements the steps of the operation and maintenance strategy determination method based on step logs as described in any one of claims 1 to 8.

Images