Cloud platform operation and maintenance instruction operation method and device, equipment and medium

Abstract

The invention relates to the technical field of intelligent decision making, can be applied to business system platforms of financial science and technology, medical health and the like, and discloses a cloud platform operation and maintenance instruction operation method, device, equipment and medium, a natural language instruction input through an instant messaging interface is acquired, and intention analysis is performed on the natural language instruction to obtain an instruction operation intention; acquiring system state data in real time according to the instruction operation intention, and performing plan matching on the instruction operation intention through the system state data to obtain a target plan; performing user operation authority verification on the target plan according to a preset authority hierarchical control mechanism to obtain a verification result; and performing execution confirmation on the target plan according to the verification result. By performing execution confirmation and pre-execution verification on the target plan, forcible examination and approval avoids unnecessary operation caused by mistaken input of an instruction; meanwhile, potential problems in the plan can be found in advance through pre-execution verification, and errors in actual execution are prevented.

Description

Technical Field

[0001] The present invention relates to the field of intelligent decision-making technology, and in particular to a method, device, equipment and medium for executing cloud platform operation and maintenance instructions. Background Art

[0002] Cloud platform operation and maintenance refers to a series of technical and management activities that monitor, manage, optimize, troubleshoot and ensure the security of the entire life cycle of the cloud computing platform. With the deep penetration of cloud computing in various industries, these operation and maintenance technologies and management methods centered on the stable operation of the cloud platform have also been extended and applied to various fields such as finance, medical care, manufacturing, and the Internet. For example, in the medical field, there are business volume fluctuations and innovative scenarios. Cloud platform operation and maintenance can use dynamic resource scheduling technology to automatically expand or shrink computing and storage resources according to real-time business load to avoid resource waste or shortage. For example, the core business in the financial technology field has extremely high requirements for system availability. Millisecond-level delays or interruptions may lead to financial losses, user complaints and even market risks. Cloud platform operation and maintenance ensures the stable operation of the trading system by building a high-availability architecture, combining real-time monitoring, automatic fault switching, full-link stress testing and other means.

[0003] Currently, the cloud platform operation and maintenance instruction execution method adopts the traditional plan library approach, but this approach does not integrate the mandatory approval process and pre-verification mechanism, and cannot meet the security operation and maintenance requirements of the dynamic cloud environment. Summary of the Invention

[0004] The present invention provides a cloud platform operation and maintenance instruction execution method, device, equipment and medium. By performing execution confirmation and pre-execution verification on the target plan, mandatory approval is avoided to avoid unnecessary operations caused by erroneous input of instructions; at the same time, pre-execution verification can discover potential problems in the plan in advance and prevent errors during actual execution.

[0005] In a first aspect, a method for executing cloud platform operation and maintenance instructions is provided, comprising: Obtaining a natural language instruction input through an instant messaging interface, and performing intent analysis on the natural language instruction to obtain an instruction operation intention; Acquire system status data in real time according to the instruction operation intention, match the instruction operation intention with the system status data to obtain a target plan; Performing a user operation authority check on the target plan according to a preset authority hierarchical control mechanism to obtain a verification result; Performing execution confirmation on the target plan according to the verification result, and performing pre-execution verification on the operation instructions in the target plan based on the confirmation result to obtain an execution verification result; The operation instruction is sent to the cloud platform for execution according to the execution verification result to obtain the execution result.

[0006] In a second aspect, a cloud platform operation and maintenance instruction execution device is provided, comprising: An acquisition and analysis module is used to acquire natural language instructions input through the instant messaging interface and perform intent analysis on the natural language instructions to obtain the instruction operation intention; An acquisition and matching module is used to acquire system status data in real time according to the instruction operation intention, and perform plan matching on the instruction operation intention through the system status data to obtain a target plan; A verification module is used to verify the user's operation authority of the target plan according to a preset authority classification control mechanism to obtain a verification result; A confirmation module, configured to confirm the execution of the target plan based on the verification result; A verification module, configured to perform pre-execution verification on the operation instructions in the target plan by confirming the result, and obtain an execution verification result; The execution module is used to send the operation instruction to the cloud platform for execution according to the execution verification result to obtain the execution result.

[0007] In a third aspect, a computer device is provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the above-mentioned cloud platform operation and maintenance instruction execution method are implemented.

[0008] In a fourth aspect, a computer-readable storage medium is provided, which stores a computer program. When the computer program is executed by a processor, the steps of the above-mentioned cloud platform operation and maintenance instruction running method are implemented.

[0009] In the solution implemented by the above-mentioned cloud platform operation and maintenance instruction execution method, device, computer equipment and storage medium, instructions are issued through everyday language to improve the usability of the system; by analyzing natural language instructions through intent, ambiguous expressions in natural language can be filtered out, the user's true purpose can be clarified, and erroneous operations caused by instruction ambiguity can be avoided; system status data is obtained in real time according to the operation intention, and the target plan is matched with the system status data. The system status data is the "reality basis". Matching the plan based on the real-time state can avoid the failure of the plan due to changes in the system state; at the same time, system data is obtained in real time and the plan is quickly matched, reducing the manual judgment link. Especially in dynamically changing scenarios, it can shorten the response time from instruction to plan and improve processing efficiency; The target plan is verified for user operation permissions based on the permission classification control mechanism. Permission classification can limit the operation scope of different users to avoid system failure or information leakage caused by unauthorized operations. At the same time, the permission requirements of different scenarios are different, and the classification mechanism can be configured on demand to meet diverse management needs. The target plan is executed and the execution confirmation mandatory approval process gives users a second chance to judge to avoid unnecessary operations caused by erroneous input of instructions. At the same time, pre-execution verification can discover potential problems in the plan in advance to prevent errors during actual execution. The verified instructions are executed through the cloud platform, which can make use of the distributed architecture and fault-tolerant mechanism of the cloud platform to avoid operation interruption due to local equipment failure and ensure that the instructions are "not lost or deviated." BRIEF DESCRIPTION OF THE DRAWINGS

[0010] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments of the present invention. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0011] Figure 1 This is a schematic diagram of an application environment of a cloud platform operation and maintenance instruction execution method according to an embodiment of the present invention; Figure 2 This is a flow chart of a method for executing cloud platform operation and maintenance instructions in one embodiment of the present invention; Figure 3 This is a structural diagram of a cloud platform operation and maintenance instruction execution device in one embodiment of the present invention; Figure 4 is a structural diagram of a computer device according to an embodiment of the present invention; Figure 5 It is another structural schematic diagram of a computer device in one embodiment of the present invention. DETAILED DESCRIPTION

[0012] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0013] The embodiment of the present invention provides a cloud platform operation and maintenance instruction execution method, which can be applied to Figure 1 In an application environment, the client communicates with the server through a network. The server can obtain natural language instructions input through an instant messaging interface, and perform intent analysis on the natural language instructions to obtain instruction operation intentions; obtain system status data in real time according to the instruction operation intentions, match the instruction operation intentions with plans through the system status data, and obtain a target plan; verify the user operation authority of the target plan according to a preset authority hierarchy control mechanism to obtain a verification result; confirm the execution of the target plan according to the verification result, and perform pre-execution verification on the operation instructions in the target plan through the confirmation result to obtain an execution verification result; send the operation instructions to the cloud platform for execution according to the execution verification result to obtain an execution result, and feed the execution result back to the client. The present invention provides a cloud platform operation and maintenance instruction running device. For the execution result business, by performing execution confirmation and pre-execution verification on the target plan, forced approval is performed to avoid unnecessary operations caused by erroneous input of instructions; at the same time, pre-execution verification can discover potential problems in the plan in advance to prevent errors during actual execution. Among them, the client can be, but is not limited to, various personal computers, laptops, smart phones, tablet computers and portable wearable devices. The server side can be implemented by an independent server or a server cluster composed of multiple servers. The present invention will be described in detail below through specific embodiments.

[0014] See also Figure 2 As shown, Figure 2 A flowchart of a method for executing cloud platform operation and maintenance instructions provided by an embodiment of the present invention includes the following steps: S1. Obtain a natural language instruction input through an instant messaging interface, and perform intent analysis on the natural language instruction to obtain an instruction operation intention.

[0015] In an embodiment of the present invention, the acquisition refers to the process of collecting and capturing natural language instructions input by the user from the instant messaging interface, and the intent analysis refers to the process of in-depth understanding and parsing of the acquired natural language instructions to determine the operations the user wants to perform or the needs expressed.

[0016] Specifically, through instant messaging tools (such as chat windows, intelligent customer service interfaces), we receive commands entered by users in everyday language, such as "check my physical examination report" and "help me redeem funds", and then use technical means to analyze the core needs behind the language, clarify the specific types of operations that users want to complete, such as "query medical data" and "financial product transactions", and provide a basis for the subsequent accurate execution of commands.

[0017] For example, when the system receives an alert from the APM monitoring platform 111: "The P99 response time of the application 'OrderSvc' exceeds the threshold of 5 seconds", the AI ​​decision engine is activated, analyzes the alert content, and determines that the intention is to "solve the high latency problem of 'OrderSvc'".

[0018] In specific medical and health scenarios, patients use natural language to initiate requests to medical platforms or smart assistants. The system quickly identifies core needs such as "consulting about the condition" and "outpatient appointments" through intent analysis, and directly connects to the electronic medical record system and registration platform to reduce manual transfer links. This is especially convenient for elderly patients or users who are not familiar with complex operating interfaces to efficiently obtain medical services.

[0019] In the financial technology scenario, users input natural language commands such as "How do I pay my credit card bill?" and "Recommend a stable financial product" through the chat interface. After the system parses the intentions such as "bill inquiry and repayment" and "product recommendation", it directly links the account system and financial product library to push repayment paths or match risk-adapted products, lowering the threshold for using financial services and improving user operation convenience.

[0020] In the embodiment of the present invention, performing intent analysis on the natural language instruction to obtain the instruction operation intention includes: Cleaning the natural language instruction to obtain a cleaned text to be analyzed; Extracting key information from the cleaned text to be analyzed to obtain key information of the text; Match the text key information with the intent category through the preset domain intent library to obtain the information intent category; The information intention category is ambiguously corrected using the preset scene context to obtain the instruction operation intention.

[0021] In an embodiment of the present invention, the cleaning process refers to preprocessing the acquired natural language instructions, removing irrelevant, erroneous or interfering information therein, and converting them into a more standardized text form that is more suitable for subsequent analysis. The key information extraction refers to identifying and extracting core information closely related to the user's intention from the cleaned text to be analyzed. The intent category matching refers to comparing and matching the extracted text key information with various intents in the intent library through a preset domain intent library, so as to determine the intent category to which the user instruction belongs. The ambiguity correction refers to the process of using preset scene context information to judge and correct ambiguities that may arise during the intent type matching process.

[0022] Specifically, the original natural language instructions obtained are cleaned and standardized, including removing redundant punctuation, correcting typos, and unifying sentence formats, such as converting spoken expressions into written expressions to generate structured text to be analyzed; through semantic analysis, the core elements in the text to be analyzed are identified, including operation entities, such as "cloud server", "database", "medical archives", action instructions, such as "expansion", "query", "backup", and "freeze", and the semantic association relationship between entities and actions is marked, such as "expand the cloud server", to generate a key information set containing entity-action associations, namely text key information.

[0023] Furthermore, a key information set containing entity-action associations is input into a preset domain intent library. By comparing the entity-action associations with typical intent templates preset in the intent library, the intent category of the instruction is preliminarily determined and a preliminary intent label is generated.

[0024] Furthermore, the preliminary intent label is combined with the user's historical interaction records and the current scenario context (such as user identity permissions, current system status) to perform ambiguity correction. For example, when there are polysemous entities in the instruction (such as "backup" can refer to data backup or device backup), the unique reference is determined through context association, and the precise instruction operation intention is finally output.

[0025] Specifically, an independent conversation context is maintained for each user or each independent operation and maintenance task. This context records the user's identity information, historical interaction records, current task status (for example, whether it is waiting for user confirmation or waiting for supervisor approval), and alternative plans recommended by the AI ​​decision engine. Message distribution and routing are used as the core message distribution center. After receiving the user's original message, the conversation engine performs preliminary analysis. For messages containing natural language instructions, it forwards them to the AI ​​decision engine for intent recognition. For simple confirmation replies (such as the user enters "execute" or clicks the confirmation button), it directly drives the subsequent process based on the conversation context.

[0026] In an embodiment of the present invention, the core elements are focused on from the cleaned text, irrelevant content is stripped off, and natural language is converted into structured key information to improve the efficiency and accuracy of subsequent intent recognition; the text key information is matched through the domain intent library, and based on the preset domain rules, the key information is bound to the standardized intent category to achieve the initial conversion of natural language to business intent, avoiding generalization or deviation of intent recognition; for polysemous or ambiguous expressions of a word, the real-time scene context is combined to eliminate ambiguity, ensure that the final intent is highly matched with the actual business scenario, and improve the accuracy of instruction understanding.

[0027] In an embodiment of the present invention, intent analysis can be used to quickly locate the core needs of users, avoid inefficient communication caused by manual screening or fuzzy matching, directly connect to the corresponding business system, and shorten the link from instruction input to service execution; at the same time, the frequency of manual intervention is reduced, allowing human resources to focus on solving complex problems and improving the scalability of the overall service.

[0028] S2. Acquire system status data in real time according to the instruction operation intention, perform plan matching on the instruction operation intention through the system status data, and obtain a target plan.

[0029] In an embodiment of the present invention, the acquisition refers to the real-time collection and extraction of system status data necessary for the execution of the intention from system-related modules or data storage based on the determined instruction operation intention, and the plan matching refers to the process of searching for the best-matching plan in a preset plan library based on the acquired system status data and the determined instruction operation intention.

[0030] Specifically, based on the user's command operation intention identified in the early stage, the key status data of the current system is captured in real time, such as the equipment load and patient vital signs of the medical system, and the account risk level and transaction frequency of the financial system. These status data are then accurately matched with the preset standardized plan library to finally determine the execution plan that best suits the current scenario. Among them, real-time system status data includes log data, application performance indicator data, and infrastructure topology data.

[0031] For example, the AI ​​decision engine queries the log platform for error logs through MCP, and queries the host monitoring system for container CPU and memory usage. The AI ​​decision engine integrates alarms and contextual data to generate a fault summary, "High latency was detected in 'OrderSvc', accompanied by a 'database connection timeout error' and 95% CPU usage." The engine semantically matches this summary with the plan library and determines that "Plan ID Name: Application CPU Bottleneck Emergency Expansion Plan" is the best target plan.

[0032] In medical and health scenarios, when the system recognizes the intention of "first aid instructions for patients with acute myocardial infarction", it obtains the patient's current heart rate, blood pressure and other vital signs data in real time, as well as system data such as the availability of rescue equipment and the on-duty status of nearby medical staff. It quickly matches the "Standardized Process for First Aid in Myocardial Infarction" plan, automatically triggering operations such as equipment scheduling, priority issuance of checklists, and emergency calls for specialists, reducing human decision-making delays and buying time for rescue.

[0033] In financial scenarios, when the intention of "large-scale abnormal transactions in different locations" is identified, the system obtains real-time status data such as the account's historical transaction region, recent login device, risk rating, etc., matches the "high-risk transaction interception plan", and automatically triggers identity secondary verification, transaction suspension notification, manual risk control review and other processes to accurately distinguish normal transactions from fraud risks, thereby ensuring the security of funds and reducing interference with users' normal operations.

[0034] In the embodiment of the present invention, the method of matching the instruction operation intention with the system state data to obtain a target plan includes: Extracting core indicators from the system status data and marking the correlation between the core indicators and the instruction operation intention; Generate a data feature set based on the core indicators and the association relationship; Matching the trigger conditions of the data feature set with a preset plan library to obtain a candidate plan set; The candidate plan set is optimally screened according to the priority of the instruction operation intention to obtain a target plan.

[0035] In an embodiment of the present invention, the extraction refers to screening out the most critical and representative indicator information for judging the execution of the instruction operation intention from the collected system status data; the labeling refers to clarifying the logical connection and influence mode between the core indicators and the instruction operation intention; the generation refers to the process of integrating and structuring the extracted core indicators and the labeled association relationships; the trigger condition matching refers to comparing and matching the generated data feature set with the trigger conditions set for each plan in the preset plan library, finding all plans that meet the trigger conditions, and forming a candidate plan set; the optimal screening refers to evaluating and comparing each plan in the candidate plan set according to the priority of the instruction operation intention, and selecting the plan that best suits the current situation and can best realize the instruction operation intention as the target plan.

[0036] Specifically, according to the instruction operation intention (such as "cloud server expansion" and "database fault recovery"), the data interaction interface of the underlying system is called through the Model Context Protocol (MCP) module, and the real-time system status data related to the intention (such as the current server load, number of database connections, disk usage, remaining resources, etc.) is collected according to the unified specifications defined by MCP (such as data format and interaction frequency) to generate a structured system status data set.

[0037] Furthermore, through semantic analysis, the core indicators in the system status data set are identified (such as "load rate > 80%" and "remaining memory < 2GB"), and the association between the indicators and the instruction operation intentions is marked (such as "expansion intention is associated with load rate and remaining memory indicators"), generating a feature set containing the indicator-intent association; the feature set is input into the preset plan library, which stores the trigger conditions corresponding to various operation intentions, such as "when the load rate is > 80% and the remaining memory is < 2GB, match the 'cloud server temporary expansion plan'", and by comparing the consistency between the feature set and the plan trigger conditions, the candidate plan set that meets the current system status is screened out; finally, based on the priority of the instruction operation intention and the real-time changes in the system status, the optimal solution is determined from the candidate plan set, and finally the only target plan is obtained.

[0038] Specifically, upon receiving user intent or warning information, the AI ​​engine uses it as a query to search the descriptive text vector database of the plan library, recalling the most relevant plans. The large language model then sorts them based on real-time context data to select the optimal plan. The plan library stores a series of structured plans, each of which is versioned and supports grayscale release and rollback.

[0039] In an embodiment of the present invention, core indicators are extracted and correlation relationships are marked, key indicators are focused on from massive system status data, and their correlation with operational intent is clarified, providing accurate data anchor points for subsequent plan matching and avoiding interference from invalid data; a data feature set is generated based on core indicators and correlation relationships, and the scattered indicators and correlation relationships are structured and integrated into a feature set, making the mapping between system status and operational intent clearer, providing standardized input for plan library matching, and improving matching efficiency; among multiple candidate plans, the optimal plan is accurately locked based on the priority of operational intent, ensuring that the final selected plan is highly consistent with the core goals, and improving the pertinence and effectiveness of decision-making.

[0040] In the embodiment of the present invention, the plan is matched with real-time system status data to avoid "one-size-fits-all" execution that is divorced from the actual scenario, ensure that the target plan is highly adapted to the current environment, and reduce the risk of operational errors. At the same time, there is no need to manually judge the details of the scenario one by one. Through automatic matching of the system's real-time data with the plan library, the optimal execution plan can be quickly located.

[0041] S3. Perform user operation authority verification on the target plan according to a preset authority classification control mechanism to obtain a verification result.

[0042] In the embodiment of the present invention, the user operation authority verification refers to the process of checking whether the current user has the corresponding authority for the operations involved in the target plan according to the preset authority classification control mechanism.

[0043] Specifically, after determining the target plan, the system will verify whether the user who initiated the instruction has the operational authority to execute the plan based on the system's preset authority classification rules (for example, determining whether the user has the authority to call high-risk medical equipment or execute large financial transactions), and ultimately obtain a "pass" or "fail" verification result.

[0044] Specifically, the authority classification rules are integrated by the financial institution's internal identity authentication system, which defines a detailed authority matrix. For example, ordinary users can only execute plans that affect CPU usage less than 50% or do not involve data changes. Plans for high-risk operations (such as service restart and data rollback) require secondary approval and confirmation by the operation and maintenance supervisor or department head.

[0045] In specific scenarios of medical health, when the target plan involves highly sensitive operations such as "retrieval of patient private medical records" and "prescribing narcotic drugs", the system verifies the user's identity through a permission classification mechanism. For example, only the attending physician is allowed to view the complete medical records of the patients in the department, and interns need authorization from their superiors to prescribe special drugs, preventing medical data leakage or illegal operations, while complying with privacy protection requirements such as the "Basic Specifications for Medical Record Writing".

[0046] In a financial scenario, if the target plan is to "approve corporate loans worth tens of millions of yuan," the system verifies whether the initiating user is a risk control officer or senior executive with authority to approve the corresponding quota. It strictly divides business boundaries through hierarchical authority to prevent low-authority personnel from accessing high-risk businesses, while also improving the standardization and traceability of the approval process.

[0047] In an embodiment of the present invention, the user operation authority verification of the target plan according to the preset authority hierarchical control mechanism to obtain the verification result includes: Extracting the risk level and operational characteristics of the target plan; Obtaining the identity role of the user who currently initiates the operation based on the risk level and the operation characteristics; Using a preset permission classification control mechanism, the risk level and the operation characteristics are matched with the user identity role to obtain a matching result; Conduct hierarchical approval and judgment on the target plan to obtain the approval result; A verification result is generated according to the matching result and the approval result.

[0048] In an embodiment of the present invention, the extraction refers to finding and determining the two key information, risk level and operation characteristics, of the target plan from the target plan; the acquisition refers to finding and determining the identity role corresponding to the user who currently initiates the operation based on the risk level and operation characteristics of the extracted target plan; the basic authority matching refers to using a pre-set authority hierarchical control mechanism to compare the risk level and operation characteristics of the extracted target plan with the acquired identity role of the user who currently initiates the operation; the hierarchical approval judgment refers to judging whether the target plan needs to be approved and which levels of approval it needs to go through according to pre-set hierarchical approval rules and processes; the generation refers to generating the final verification result based on certain logical rules and business requirements by comprehensively combining the matching results obtained from the basic authority matching and the approval results obtained from the hierarchical approval judgment.

[0049] Specifically, core information is parsed from the target plan, including the plan's risk level (such as L1-low risk, L2-normal, L3-high risk) and specific operation type (such as whether it involves data changes, resource start-up and shutdown, configuration modification, etc.), to generate plan attribute information containing risk level and operation characteristics; the plan attribute information is integrated with the financial institution's internal identity authentication system (such as LDAP / AD) to obtain the current user identity initiating the operation (such as user name, work number) and associated roles (such as ordinary user, operation and maintenance supervisor, department head), and generate a user identity-role mapping table.

[0050] Furthermore, based on the preset permission classification control mechanism, the permission matrix is ​​called (defining the risk level, operation type and approval requirements that different roles can operate), and the user role is matched with the plan risk level and operation characteristics to determine the user's basic operation permissions for the current plan, such as whether ordinary users are allowed to perform L2 non-data change operations.

[0051] Furthermore, the target plan is triggered to undergo hierarchical approval verification to determine whether the target plan is a high-risk operation or a low-risk operation. The approval result is finally obtained based on the judgment result, and the basic authority matching result and the hierarchical approval result are combined to output the final verification result.

[0052] In detail, the authority classification control strategy makes authority judgments based on the user role and the risk level of the target plan. For plans preset to a high-risk level, the authority verification step includes triggering a multi-level approval process; for example, the security control module is used to verify user authority. Assuming that this operation requires supervisor approval (risk level L3-HighRisk), the system will automatically send an approval request to the designated supervisor, and the supervisor clicks "Approve" in the IM.

[0053] In the embodiment of the present invention, the step of performing hierarchical approval and determination on the target plan to obtain an approval result includes: Determining whether the target plan is a high-risk operation; If the target plan is a high-risk operation, the target plan is subject to approval by the corresponding role to obtain an approval result; If the target plan is not a high-risk operation, it is confirmed that the target plan does not require hierarchical approval.

[0054] In an embodiment of the present invention, the corresponding role approval refers to determining the roles with corresponding approval authority based on key factors such as the risk level and operational characteristics of the target plan, combined with pre-set role permissions and approval process rules, and these roles review and judge the target plan according to the established process, and finally reach an approval result on whether to approve the execution of the plan.

[0055] Specifically, if the target plan belongs to a high-risk operation (such as risk level L3, involving data rollback), the secondary approval process is triggered according to the permission matrix rules, and the operation request submitted by the user is sent to the corresponding approval role (such as the operation and maintenance supervisor or department head) to obtain the approval / rejection result; if it is a low-risk or ordinary operation (such as L2 non-data change), it will directly pass the basic permission check.

[0056] Specifically, if a user inputs: "Delete the production database 'OrderSvc'", the AI ​​decision engine parses the high-risk intention of "deleting the database" and matches it in the plan library. However, due to security policies, no such plan exists in the library, and the system fails to match the plan. The security control module then determines the intention as a high-risk violation, immediately blocks the process, and provides feedback to the user through the conversation engine: "The operation is rejected. The requested operation 'delete the production database' is a high-risk instruction and no authorized plan is matched. The behavior has been recorded and notified to the security administrator." Subsequently, the secondary approval process is triggered according to the permission matrix rules, and this operation request is sent to the corresponding approval role (such as the operation and maintenance supervisor or department head) to obtain the approval / rejection result.

[0057] Finally, based on the basic permission matching results and the hierarchical approval results, if the permissions match and are approved (such as the L2 plan meets the permissions of ordinary users), the verification result will be "passed"; if the permissions are insufficient or the approval is rejected, the verification result will be "failed" and the reason will be fed back, such as "the current user does not have high-risk operation permissions."

[0058] In an embodiment of the present invention, the risk level and operational characteristics of the target plan are extracted, the core attributes of the plan are accurately located, and a quantitative basis is provided for subsequent authority matching and approval judgment, thereby avoiding management and control omissions caused by risk ambiguity; based on preset rules, automated authority verification of "risk-operation-role" is implemented to ensure that low-risk operations are efficiently passed and high-risk operations are strictly restricted, thereby ensuring compliance and improving the execution efficiency of low-risk processes; authority compliance and approval compliance are integrated to form a final verification conclusion of "double compliance", ensuring that only plans with "authority matching and approval" can enter the execution link, thereby blocking unauthorized operations and non-compliant processes from the source.

[0059] In the embodiment of the present invention, permission verification is used to ensure that only users with corresponding qualifications can execute the target plan, thereby blocking data leakage, financial loss or medical errors caused by unauthorized operations at the source, and avoiding security risks caused by permission confusion; at the same time, permission verification is implemented through system automation rules, reducing the arbitrariness of human judgment and ensuring that the execution of the target plan strictly follows the preset permission framework.

[0060] S4. Perform execution confirmation on the target plan according to the verification result, and perform pre-execution verification on the operation instructions in the target plan based on the confirmation result to obtain an execution verification result.

[0061] In an embodiment of the present invention, the execution confirmation refers to the final review and determination process of whether the target plan meets the execution conditions and whether the execution process can be formally started based on the verification results generated by the previous process. The pre-execution verification refers to the simulation or small-scale actual testing of the operating instructions in the target plan after the execution confirmation is passed, so as to verify whether these instructions can operate normally as expected and achieve the expected results in the actual execution environment.

[0062] Specifically, after the permission verification is passed, the execution permission for the target plan is first obtained through manual or system confirmation (such as the operation and maintenance personnel reply "execute" or the system automatically confirms compliance). Based on this confirmation result, the specific operation instructions in the plan (such as expansion instructions, data query instructions) are placed in a simulation environment (sandbox) for pre-execution testing. Finally, it is determined whether the instructions can be executed normally and whether there are any potential risks, forming an "execution verification result" to provide security pre-guarantee for the actual execution in the subsequent production environment.

[0063] In an embodiment of the present invention, performing pre-execution verification on the operation instructions in the target plan through the confirmation result to obtain the execution verification result includes: Determining whether the confirmation result is execution; If the confirmation result is not to be executed, then it is confirmed that the execution of the target plan has failed; If the confirmation result is execution, the operation instructions in the target plan are simulated and recorded in a preset sandbox environment to obtain a pre-execution process record; The normal execution of the operation instruction is judged through the pre-execution process record to obtain an execution verification result.

[0064] In an embodiment of the present invention, the simulation record refers to the process of using a preset sandbox environment to simulate the operation instructions in the target plan in an approximate real scenario, and recording in detail and completely various relevant information and behaviors of the operation instructions during the simulated execution process. The normal execution judgment refers to the process of evaluating and judging whether the operation instructions in the target plan can be smoothly executed under normal conditions as expected based on the pre-execution process record obtained from the simulation record.

[0065] Specifically, based on the result of the permission check, the system sends a confirmation message containing the target plan information to the operation and maintenance personnel through the conversation engine (such as "[Alarm] 'OrderSvc' has high latency due to CPU bottleneck. It is recommended to execute 'Plan No. 007: Emergency Expansion'. Do you want to execute it?"); receives the operation and maintenance personnel's reply (such as "Execute" or "Cancel"), and generates a clear execution confirmation result.

[0066] Furthermore, if the execution is confirmed, the specific operation instructions are parsed and extracted from the target plan (such as "Plan 007") to clarify the content of the operation to be executed; then the extracted operation instructions are input into the sandbox environment (a simulation environment independent of the production environment), the execution of the instructions is simulated and the execution process is monitored, including the correctness of the instruction syntax, the matching degree of resource dependencies (such as whether there is an "OrderSvc" deployment), and the simulation of system status changes after execution (such as the simulation of CPU load changes after expansion), to generate a pre-execution process record.

[0067] Furthermore, based on the sandbox pre-execution process records, it is determined whether the instruction can be executed normally (such as no syntax errors, no resource conflicts, no potential risks, and the simulation effect meets expectations), and the final execution verification result is output; if the sandbox verification passes (such as the instruction successfully simulates the expansion and there are no abnormalities), the result is "verification passed"; if there are syntax errors, non-existent resources and other problems, the result is "verification failed" and the specific reason is fed back.

[0068] In detail, the pre-execution verification process is as follows: the security control module starts a temporary, network-isolated containerized sandbox environment, which runs a lightweight image that simulates the target cloud platform API service endpoint; instruction injection and simulated execution: the plan instructions are injected into the sandbox for execution; behavior detection: the sandbox performs syntax checking and high-risk keyword scanning on the instructions; status assertion: after the simulation execution, the system verifies whether the status of the simulated deployment object in the sandbox has changed as expected, and confirms that there are no changes in the production environment; verification pass / fail processing: if the verification passes, the process continues. If it fails, for example, if a high-risk keyword is detected, the verification fails, the system immediately terminates the operation, and notifies the user and the security audit system through the conversation engine: such as "Plan '007' was intercepted during sandbox verification because it contains potentially risky operations."

[0069] In an embodiment of the present invention, simulating the execution of instructions in an isolated sandbox environment can not only fully record every detail of the operation without affecting the production environment, but also provide real process data for subsequent analysis, while avoiding interference of simulated execution on actual business; by analyzing the key nodes in the records, it is possible to accurately judge whether the instructions meet expectations, identify potential risks in advance, and provide a pre-condition for formal execution that "the instructions are valid and the risks are controllable", thereby reducing the probability of actual execution failure.

[0070] In an embodiment of the present invention, the execution confirmation link is verified twice manually or by the system to avoid unnecessary operations caused by passing the permission check but misjudging the scenario; the pre-execution verification tests the feasibility of the instruction in a simulation environment, discovers instruction logic errors, parameter conflicts and other problems in advance, and reduces operational errors during formal execution from the source.

[0071] S5. Send the operation instruction to the cloud platform for execution according to the execution verification result to obtain the execution result.

[0072] In an embodiment of the present invention, the sending of execution to the cloud platform means that based on the execution verification results, the operation instructions in the target plan that have been confirmed to be feasible through simulation records and normal execution judgment are transmitted to the cloud platform through specific technical means and management processes and triggered to run on the actual resources of the cloud environment.

[0073] Specifically, after completing the pre-execution verification and confirming that the results are risk-free (such as the instruction logic is correct and the resource matching is normal), the operation instructions in the target plan will be formally issued to the cloud platform, which will actually execute the instructions, and finally obtain the actual results of the instruction execution, and feedback the execution results of the operation instructions to the user.

[0074] In the embodiment of the present invention, the step of sending the operation instruction to the cloud platform for execution based on the execution verification result to obtain the execution result includes: Performing a production environment execution judgment on the execution verification result to obtain a judgment result; Extracting complete information of the operation instruction according to the judgment result, and encapsulating the complete information according to a preset cloud platform interface specification to obtain an encapsulated execution request; The packaged execution request is sent to the cloud execution environment through the preset cloud platform resource scheduling interface, and feedback is received on the sent request to obtain the execution result.

[0075] In an embodiment of the present invention, the production environment execution judgment refers to the process of comprehensively evaluating whether the operation instruction is suitable for execution in a real production environment based on the execution verification result; the extraction refers to the process of accurately obtaining the complete information of the operation instruction from the target plan based on the judgment result of the production environment execution judgment; the encapsulation refers to the process of organizing and formatting the extracted complete information of the operation instruction according to the preset cloud platform interface specification to form a packaged execution request that can be recognized and processed by the cloud platform; the sending refers to the process of transmitting the packaged execution request to the cloud execution environment through the preset cloud platform resource scheduling interface; the receiving feedback refers to the process of receiving and processing the return result after the cloud platform processes the packaged execution request.

[0076] Specifically, the result output by the pre-execution verification phase (such as "verification passed" or "verification failed") is judged to clarify whether the operation instruction meets the conditions for execution in the production environment; if the result is "verification passed", the complete information of the instruction (such as instruction content, target resource identifier, execution parameters) is extracted; if the result is "verification failed", the process is terminated and the reason is fed back (such as "sandbox verification found instruction syntax error").

[0077] Furthermore, the confirmed operation instruction information will be encapsulated into an execution request according to the interface specifications of the cloud platform, clarifying the execution object of the instruction (such as the target cloud server, database instance), operation type (such as expansion, restart) and execution priority (such as emergency operation marked "high priority"); the generated execution request will be sent to the cloud execution environment through the resource scheduling interface of the cloud platform, and at the same time, real-time monitoring of the execution status will be triggered (such as recording the request sending time and the assigned execution node ID).

[0078] Furthermore, after the cloud platform executes the instruction, it returns the real-time execution status (such as "executing", "execution successful", "execution failed") and detailed logs (such as the new configuration parameters of the server after expansion, and execution time) through the scheduling interface; if the execution fails, the failure reason is returned synchronously; the cloud platform feedback information is integrated and organized into a structured execution result. If the instruction is executed successfully, "execution successful" and key results are output, such as "Plan 007 expansion is completed, and the server CPU utilization rate drops from 95% to 40%"; if the execution fails, "execution failed" is output and the failure reason and suggestions are associated, such as "Due to resource locking failure, it is recommended to unlock and re-execute", and the results are synchronized to the session engine and operation log.

[0079] For example, after the system monitors the completion of an operation and confirms through MCP that the number of OrderSvc replicas has increased to four and the P99 response time has returned to normal, it provides feedback to the user via IM: "Plan 007 was successfully executed, OrderSvc has been expanded to four replicas, and performance has been restored." In an embodiment of the present invention, instructions that have "passed pre-execution verification and are suitable for the production environment" are screened out through final judgment, thereby avoiding execution risks caused by differences between verification results and the production environment, ensuring that instructions entering the next link have practical feasibility, and reducing invalid execution costs; through standardized packaging, operation instructions are converted into a format recognizable by the cloud platform, eliminating information ambiguity, ensuring the compatibility of instructions with the cloud platform interface, improving the accuracy of request issuance, and reducing the probability of execution failure due to format problems.

[0080] In an embodiment of the present invention, issuing instructions based on verification results can ensure that the instructions executed by the cloud platform are completely matched with the target plan, reducing the problem of unsatisfactory execution results due to instruction deviations, and at the same time relying on the computing power and distributed capabilities of the cloud platform to improve the efficiency and stability of instruction execution.

[0081] It can be seen that in the above scheme, for the execution result business, the natural language instructions input through the instant messaging interface are obtained, and the intent analysis of the natural language instructions is performed to obtain the instruction operation intention; the system status data is obtained in real time according to the instruction operation intention, and the instruction operation intention is matched with the plan through the system status data to obtain the target plan; the user operation authority of the target plan is verified according to the preset authority classification control mechanism to obtain the verification result; the execution confirmation of the target plan is performed according to the verification result, and the operation instructions in the target plan are pre-executed and verified through the confirmation result to obtain the execution verification result; the operation instructions are sent to the cloud platform for execution according to the execution verification result to obtain the execution result, and by performing execution confirmation and pre-execution verification on the target plan, forced approval is performed to avoid unnecessary operations caused by erroneous input of instructions; at the same time, pre-execution verification can discover potential problems in the plan in advance to prevent errors during actual execution.

[0082] It should be understood that the size of the serial numbers of the steps in the above embodiments does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0083] In one embodiment, a cloud platform operation and maintenance instruction execution device is provided, and the cloud platform operation and maintenance instruction execution device corresponds one-to-one to a cloud platform operation and maintenance instruction execution method in the above embodiment. Figure 3 As shown, the cloud platform operation and maintenance instruction execution device includes an acquisition and analysis module 101, an acquisition and matching module 102, a verification module 103, a confirmation module 104, a verification module 105, and an execution module 106. The functional modules are described in detail as follows: The acquisition and analysis module 101 is used to acquire natural language instructions input through the instant messaging interface and perform intent analysis on the natural language instructions to obtain the instruction operation intention; An acquisition and matching module 102 is configured to acquire system status data in real time according to the instruction operation intention, and perform a plan matching on the instruction operation intention through the system status data to obtain a target plan; Verification module 103, configured to verify user operation authority of the target plan according to a preset authority hierarchical control mechanism and obtain a verification result; A confirmation module 104 is configured to confirm the execution of the target plan based on the verification result; The verification module 105 is used to perform pre-execution verification on the operation instructions in the target plan based on the confirmation result to obtain an execution verification result; The execution module 106 is used to send the operation instruction to the cloud platform for execution according to the execution verification result to obtain an execution result.

[0084] In one embodiment, the acquisition and analysis module 101, when performing intent analysis on the natural language instruction and obtaining the instruction operation intent, is configured to: Cleaning the natural language instruction to obtain a cleaned text to be analyzed; Extracting key information from the cleaned text to be analyzed to obtain key information of the text; Match the text key information with the intent category through the preset domain intent library to obtain the information intent category; The information intention category is ambiguously corrected using the preset scene context to obtain the instruction operation intention.

[0085] In one embodiment, the acquisition and matching module 102 is configured to, when performing a plan matching on the system state data and the instruction operation intention to obtain a target plan,: Extracting core indicators from the system status data and marking the correlation between the core indicators and the instruction operation intention; Generate a data feature set based on the core indicators and the association relationship; Matching the trigger conditions of the data feature set with a preset plan library to obtain a candidate plan set; The candidate plan set is optimally screened according to the priority of the instruction operation intention to obtain a target plan.

[0086] In one embodiment, the verification module 103 performs user operation authority verification on the target plan according to a preset authority hierarchical control mechanism and obtains the verification result, which is used to: Extracting the risk level and operational characteristics of the target plan; Obtaining the identity role of the user who currently initiates the operation based on the risk level and the operation characteristics; Using a preset permission classification control mechanism, the risk level and the operation characteristics are matched with the user identity role to obtain a matching result; Conduct hierarchical approval and judgment on the target plan to obtain the approval result; A verification result is generated according to the matching result and the approval result.

[0087] When the target plan is subject to graded approval and judgment and the approval result is obtained, it is used to: Determining whether the target plan is a high-risk operation; If the target plan is a high-risk operation, the target plan is subject to approval by the corresponding role to obtain an approval result; If the target plan is not a high-risk operation, it is confirmed that the target plan does not require hierarchical approval.

[0088] In one embodiment, when the verification module 105 performs pre-execution verification on the operation instructions in the target plan based on the confirmation result and obtains the execution verification result, it is configured to: Determining whether the confirmation result is execution; If the confirmation result is not to be executed, then it is confirmed that the execution of the target plan has failed; If the confirmation result is execution, the operation instructions in the target plan are simulated and recorded in a preset sandbox environment to obtain a pre-execution process record; The normal execution of the operation instruction is judged through the pre-execution process record to obtain an execution verification result.

[0089] In one embodiment, the execution module 106 sends the operation instruction to the cloud platform for execution according to the execution verification result, and when obtaining the execution result, is used to: Performing a production environment execution judgment on the execution verification result to obtain a judgment result; Extracting complete information of the operation instruction according to the judgment result, and encapsulating the complete information according to a preset cloud platform interface specification to obtain an encapsulated execution request; The packaged execution request is sent to the cloud execution environment through the preset cloud platform resource scheduling interface, and feedback is received on the sent request to obtain the execution result.

[0090] The present invention provides a cloud platform operation and maintenance instruction running device, which obtains natural language instructions input through an instant messaging interface for execution result business, and performs intent analysis on the natural language instructions to obtain instruction operation intention; obtains system status data in real time according to the instruction operation intention, matches the instruction operation intention with a plan through the system status data, and obtains a target plan; verifies the user operation authority of the target plan according to a preset authority classification control mechanism to obtain a verification result; confirms the execution of the target plan according to the verification result, and performs pre-execution verification on the operation instructions in the target plan through the confirmation result to obtain an execution verification result; sends the operation instruction to the cloud platform for execution according to the execution verification result to obtain an execution result, and by performing execution confirmation and pre-execution verification on the target plan, forced approval avoids unnecessary operations caused by erroneous input of instructions; at the same time, pre-execution verification can discover potential problems in the plan in advance to prevent errors during actual execution.

[0091] For the specific definition of a cloud platform operation and maintenance instruction execution device, please refer to the definition of a cloud platform operation and maintenance instruction execution method above, which will not be repeated here. The various modules in the above-mentioned cloud platform operation and maintenance instruction execution device can be implemented in whole or in part by software, hardware, and a combination thereof. The above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0092] In one embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as follows: Figure 4 As shown. The computer device includes a processor, memory, network interface and database connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes non-volatile and / or volatile storage media and internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external client via a network connection. When the computer program is executed by the processor, it implements the functions or steps on the server side of a cloud platform operation and maintenance instruction operation method.

[0093] In one embodiment, a computer device is provided. The computer device may be a client, and its internal structure diagram may be as follows: Figure 5 As shown. The computer device includes a processor, memory, network interface, display screen, and input device connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external server via a network connection. When the computer program is executed by the processor, it implements the functions or steps on the client side of a cloud platform operation and maintenance instruction execution method.

[0094] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the following steps are performed: Obtaining a natural language instruction input through an instant messaging interface, and performing intent analysis on the natural language instruction to obtain an instruction operation intention; Acquire system status data in real time according to the instruction operation intention, match the instruction operation intention with the system status data to obtain a target plan; Performing a user operation authority check on the target plan according to a preset authority hierarchical control mechanism to obtain a verification result; Performing execution confirmation on the target plan according to the verification result, and performing pre-execution verification on the operation instructions in the target plan based on the confirmation result to obtain an execution verification result; The operation instruction is sent to the cloud platform for execution according to the execution verification result to obtain the execution result.

[0095] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the following steps are implemented: Obtaining a natural language instruction input through an instant messaging interface, and performing intent analysis on the natural language instruction to obtain an instruction operation intention; Acquire system status data in real time according to the instruction operation intention, match the instruction operation intention with the system status data to obtain a target plan; Performing a user operation authority check on the target plan according to a preset authority hierarchical control mechanism to obtain a verification result; Performing execution confirmation on the target plan according to the verification result, and performing pre-execution verification on the operation instructions in the target plan based on the confirmation result to obtain an execution verification result; The operation instruction is sent to the cloud platform for execution according to the execution verification result to obtain the execution result.

[0096] It should be noted that the above functions or steps that can be implemented by the computer-readable storage medium or computer device can be found in the relevant descriptions of the server side and the client side in the aforementioned method embodiment. To avoid repetition, they will not be described one by one here.

[0097] Those skilled in the art will understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in many forms such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), Synchronous Link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0098] Those skilled in the art will clearly understand that for the sake of convenience and brevity of description, only the division of the above-mentioned functional units and modules is used as an example. In actual applications, the above-mentioned functions can be distributed and completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0099] The above-described embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. If software tools or components other than those of the company appear in the application embodiments, they are merely used for illustration and do not represent actual use. Although the present invention has been described in detail with reference to the above-mentioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the above-mentioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included in the scope of protection of the present invention.

Claims

1. A cloud platform operation and maintenance instruction execution method, characterized in that: include: Obtaining a natural language instruction input through an instant messaging interface, and performing intent analysis on the natural language instruction to obtain an instruction operation intention; Acquire system status data in real time according to the instruction operation intention, match the instruction operation intention with the system status data to obtain a target plan; Performing a user operation authority check on the target plan according to a preset authority hierarchical control mechanism to obtain a verification result; Performing execution confirmation on the target plan according to the verification result, and performing pre-execution verification on the operation instructions in the target plan based on the confirmation result to obtain an execution verification result; The operation instruction is sent to the cloud platform for execution according to the execution verification result to obtain the execution result.

2. The cloud platform operation and maintenance instruction execution method according to claim 1, characterized in that: The performing intent analysis on the natural language instruction to obtain the instruction operation intention includes: Cleaning the natural language instruction to obtain a cleaned text to be analyzed; Extracting key information from the cleaned text to be analyzed to obtain key information of the text; Match the text key information with the intent category through the preset domain intent library to obtain the information intent category; The information intention category is ambiguously corrected using the preset scene context to obtain the instruction operation intention.

3. The cloud platform operation and maintenance instruction execution method according to claim 1, characterized in that: The method of matching the command operation intention with the system state data to obtain a target plan includes: Extracting core indicators from the system status data and marking the correlation between the core indicators and the instruction operation intention; Generate a data feature set based on the core indicators and the association relationship; Matching the trigger conditions of the data feature set with a preset plan library to obtain a candidate plan set; The candidate plan set is optimally screened according to the priority of the instruction operation intention to obtain a target plan.

4. The cloud platform operation and maintenance instruction execution method according to claim 1, characterized in that: The user operation authority verification of the target plan is performed according to the preset authority hierarchical control mechanism to obtain the verification result, including: Extracting the risk level and operational characteristics of the target plan; Obtaining the identity role of the user who currently initiates the operation based on the risk level and the operation characteristics; Using a preset permission classification control mechanism, the risk level and the operation characteristics are matched with the user identity role to obtain a matching result; Conduct hierarchical approval and judgment on the target plan to obtain the approval result; A verification result is generated according to the matching result and the approval result.

5. The cloud platform operation and maintenance instruction execution method according to claim 4, characterized in that: The step of performing hierarchical approval and judgment on the target plan to obtain the approval result includes: Determining whether the target plan is a high-risk operation; If the target plan is a high-risk operation, the target plan is subject to approval by the corresponding role to obtain an approval result; If the target plan is not a high-risk operation, it is confirmed that the target plan does not require hierarchical approval.

6. The cloud platform operation and maintenance instruction execution method according to claim 1, characterized in that: The pre-execution verification of the operation instructions in the target plan is performed based on the confirmation result to obtain the execution verification result, including: Determining whether the confirmation result is execution; If the confirmation result is not to be executed, then it is confirmed that the execution of the target plan has failed; If the confirmation result is execution, the operation instructions in the target plan are simulated and recorded in a preset sandbox environment to obtain a pre-execution process record; The normal execution of the operation instruction is judged through the pre-execution process record to obtain an execution verification result.

7. The cloud platform operation and maintenance instruction execution method according to claim 1, characterized in that: The step of sending the operation instruction to the cloud platform for execution according to the execution verification result to obtain the execution result includes: Performing a production environment execution judgment on the execution verification result to obtain a judgment result; Extracting complete information of the operation instruction according to the judgment result, and encapsulating the complete information according to a preset cloud platform interface specification to obtain an encapsulated execution request; The packaged execution request is sent to the cloud execution environment through the preset cloud platform resource scheduling interface, and feedback is received on the sent request to obtain the execution result.

8. A cloud platform operation and maintenance instruction execution device, characterized in that: include: An acquisition and analysis module is used to acquire natural language instructions input through the instant messaging interface and perform intent analysis on the natural language instructions to obtain the instruction operation intention; An acquisition and matching module is used to acquire system status data in real time according to the instruction operation intention, and perform plan matching on the instruction operation intention through the system status data to obtain a target plan; A verification module is used to verify the user's operation authority of the target plan according to a preset authority classification control mechanism to obtain a verification result; A confirmation module, configured to confirm the execution of the target plan based on the verification result; A verification module, configured to perform pre-execution verification on the operation instructions in the target plan by confirming the result, and obtain an execution verification result; The execution module is used to send the operation instruction to the cloud platform for execution according to the execution verification result to obtain the execution result.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the cloud platform operation and maintenance instruction running method according to any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the cloud platform operation and maintenance instruction running method according to any one of claims 1 to 7 is implemented.

Images