A method and system for automatic operation of operation and maintenance tasks

Abstract

The application discloses a kind of operation and maintenance task automatic running method and system;Script operation and maintenance environment is dependent on risk high and complex task scheduling inefficient, the present application obtains low code arrangement feature, Playbook structure feature and host load parameter is merged and fused to obtain automatic operation and maintenance feature, and obtains server certificate execution method, utilizes automatic operation and maintenance feature and server certificate execution method to build certificate execution operation and maintenance automation model.The present application considers low code arrangement feature, Playbook structure feature and host load parameter to build multi-parameter correction neural network model, more Playbook structure feature data is generated by reinforcement learning method, so that certificate execution operation and maintenance automation model is more accurate and intelligent to obtain server certificate intelligent execution method, while task runs in container sandbox to isolate risk, combined with the self-recovery mechanism that drag-and-drop flowchart parameter generates executable script, applied to the intelligent batch execution of multiple server certificate update and avoids fault node.

Description

A method and system for automatically running operation and maintenance tasks Technical Field

[0001] This invention relates to a method for automatically running operation and maintenance tasks in the field of script-based operation and maintenance engineering technology research and development, and particularly to a method and system for automatically running operation and maintenance tasks. Background Technology

[0002] In the early days of rapid development in information technology, traditional operations and maintenance (O&M) relied heavily on manual operations, facing problems such as low efficiency, high error rates, and slow response times. With the increasing complexity of enterprise IT architectures and the widespread adoption of technologies like cloud computing and big data, manual O&M has become insufficient to meet the demands of large-scale management, and automated O&M has gradually become a core industry solution. Its core is to automate operations such as system monitoring, configuration management, and task scheduling through scripting or the use of automation tools. Scripting languages, due to their flexibility, ease of integration, and cross-platform capabilities, have become the cornerstone of automation, while automated O&M platforms further improve O&M efficiency and system reliability through modular design and workflow engines. However, early script-based O&M still suffered from low levels of intelligence, reliance on manual intervention, and cross-platform compatibility challenges, driving its continuous evolution towards intelligence, integration, and security compliance.

[0003] Therefore, when facing the challenges of high-risk dependencies in script-based operations and maintenance environments and inefficient scheduling of complex tasks, significant issues remain regarding how to obtain and merge relevant script operation and maintenance parameters to derive operation and maintenance task-related characteristics for intelligent batch execution of server certificates. This requires constructing an automated certificate execution operation and maintenance model, modifying conventionally used neural network models, and amplifying the most difficult-to-obtain Playbook structural feature data to make the automated model more accurate and intelligent in deriving intelligent server certificate execution methods. Simultaneously, tasks should run in a container sandbox to isolate risks, and a self-healing mechanism combining drag-and-drop flowchart parameters to generate executable scripts should be implemented for intelligent batch execution and avoidance of faulty nodes during multi-server certificate updates. Current technologies still present significant challenges in addressing these issues. Summary of the Invention

[0004] To address the aforementioned technical problems, this invention proposes a method and system for automatically running operation and maintenance tasks.

[0005] In a first aspect of the present invention, a method for automatically running maintenance tasks is provided, the method comprising:

[0006] Collect and acquire low-code orchestration features, and parse natural language work orders into Playbook structure features based on natural language processing technology, synchronously acquire host load parameters, and acquire server certificate execution methods;

[0007] The original automated operation and maintenance features are formed by merging and integrating the low-code orchestration features, the Playbook structure features, and the host load parameters.

[0008] The Playbook structure features are obtained based on reinforcement learning methods. The corresponding low-code orchestration generation features and host load generation parameters are set by the technicians. The low-code orchestration generation features, the Playbook structure features, and the host load generation parameters are merged and fused to form automatic operation and maintenance generation features. The server certificate generation execution method is obtained simultaneously.

[0009] Based on the automated operation and maintenance features, the server certificate execution method, the automated operation and maintenance generation features, and the server certificate generation and execution method, an automated certificate execution operation and maintenance model is constructed.

[0010] The certificate execution and maintenance automation model is run in a container sandbox to obtain the corresponding intelligent execution method for server certificates in the new script maintenance environment. Risks are isolated based on the intelligent execution method for server certificates, and the model is applied to intelligently batch execution and avoid faulty nodes when updating multiple server certificates.

[0011] Furthermore, the low-code orchestration features include drag-and-drop flowchart parameters, flowchart orchestration parameters, and data structure parameters. The drag-and-drop flowchart parameters include flowchart text feature values, colors, and data source calculations.

[0012] Furthermore, the drag-and-drop flowchart parameters are calculated using flowchart text feature values, colors, and data sources.

[0013] Furthermore, the step of parsing the natural language work order into Playbook structural features based on natural language processing technology involves parsing the text in the natural language work order into descriptions and processing them accordingly to set description feature values.

[0014] Furthermore, the host load parameters include memory utilization, read / write latency, and CPU utilization.

[0015] Furthermore, the reinforcement learning method utilizes a generative adversarial network model and adjusts it according to the drag-and-drop flowchart parameters. Specifically, it uses the drag-and-drop flowchart parameters to modify the generator of the generative adversarial network model. The generator uses a deep neural network, and its activation function is modified using the drag-and-drop flowchart parameters.

[0016] Furthermore, the original automated operation and maintenance features are formed by merging and fusing the low-code orchestration features, the Playbook structure features, and the host load parameters, and the automated operation and maintenance generation features are formed by merging and fusing the low-code orchestration generation features, the generated Playbook structure features, and the host load generation parameters, both of which are obtained through data dimension linking.

[0017] Furthermore, a unified training dataset is constructed based on the automatic operation and maintenance features, the server certificate execution method, the automatic operation and maintenance generation features, and the server certificate generation and execution method. The certificate execution operation and maintenance automation model is constructed using the unified training dataset, and the certificate execution operation and maintenance automation model is a neural network model with an improved activation function.

[0018] In the activation function, changes in the activation function can effectively show the gradient changes of the model under the application scenario, thus affecting the classification accuracy of the neural network model. By using the host load parameter coefficients unique to this scenario to correct the model, a more accurate intelligent execution method for server certificates can be obtained.

[0019] It also provides a system for automatically running operation and maintenance tasks. This system implements a method for automatically running operation and maintenance tasks, including a low-code orchestration feature acquisition module, a natural language work order parsing module, a host load parameter acquisition module, a certificate execution operation and maintenance automation model construction module, and a server certificate intelligent execution module.

[0020] The low-code orchestration feature acquisition module is used to collect and acquire low-code orchestration features.

[0021] The natural language work order parsing module has a built-in natural language processing model, which is used to parse natural language work orders into Playbook structural features based on natural language processing technology;

[0022] The host load parameter acquisition module is used to acquire host load parameters and acquire server certificate execution methods.

[0023] The certificate execution and maintenance automation model construction module is used to obtain the Playbook structure features based on reinforcement learning methods, and according to the low-code orchestration generation features and corresponding host load generation parameters set by technical personnel, the low-code orchestration generation features, the Playbook structure features, and the host load generation parameters are merged and fused to form automatic operation and maintenance generation features, and the server certificate generation execution method is obtained simultaneously.

[0024] Based on the automated operation and maintenance features, the server certificate execution method, the automated operation and maintenance generation features, and the server certificate generation and execution method, an automated certificate execution operation and maintenance model is constructed.

[0025] The server certificate intelligent execution module is used to run the certificate execution operation and maintenance automation model in a container sandbox, obtain the server certificate intelligent execution method corresponding to the new script operation and maintenance environment, isolate risks according to the server certificate intelligent execution method, and apply it to intelligent batch execution and avoid fault nodes when updating multiple server certificates.

[0026] Furthermore, the reinforcement learning method utilizes a generative adversarial network model and adjusts it according to the drag-and-drop flowchart parameters. Specifically, it uses the drag-and-drop flowchart parameters to modify the generator of the generative adversarial network model. The generator uses a deep neural network, and its activation function is modified using the drag-and-drop flowchart parameters.

[0027] Therefore, the beneficial effects of this invention are that it obtains automatic operation and maintenance features by acquiring low-code orchestration features, Playbook structure features, and host load parameters, and obtains server certificate execution methods. It then constructs an automated certificate execution operation and maintenance model using these features and methods. This invention considers low-code orchestration features, Playbook structure features, and host load parameters to construct a multi-parameter corrective neural network model. Through reinforcement learning, it generates more Playbook structure feature data, enabling the automated certificate execution operation and maintenance model to more accurately and intelligently derive intelligent server certificate execution methods. Simultaneously, the task runs in a container sandbox to isolate risks. Combined with a drag-and-drop flowchart parameter-generated executable script self-healing mechanism, it is applied to intelligently batch execution and avoid faulty nodes during multi-server certificate updates.

[0028] Further embodiments and improvements of the present invention will be described in conjunction with the accompanying drawings and specific examples. Attached Figure Description

[0029] Figure 1 is a flowchart of a method for automatically running maintenance tasks according to the present invention;

[0030] Figure 2 is a schematic diagram of a method for automatically running operation and maintenance tasks according to the present invention;

[0031] Figure 3 is an example diagram of low-code applications used in the embodiments of the present invention;

[0032] Figure 4 is a schematic diagram of the activation function of the neural network model in the embodiment of the present invention;

[0033] Figure 5 is a schematic diagram of the electronic device structure for implementing the method of the present invention in an embodiment of the present invention. Detailed Implementation

[0034] The invention will now be further described in conjunction with the accompanying drawings and specific embodiments. In this invention, the reinforcement learning model is an improved generative adversarial network model, and the neural network model used for classification is also an improved model for scenario applications.

[0035] As shown in Figure 2, the system of this invention belongs to the field of script operation and maintenance engineering technology research and development, and belongs to the research and development of script operation and maintenance engineering technology other than marine engineering equipment, new materials, biotechnology, new energy, new energy vehicles, energy conservation, environmental protection, etc. Therefore, it belongs to new technology and innovation and entrepreneurship services.

[0036] The core task of generative artificial intelligence is to generate new data that conforms to certain rules or patterns. Neural networks, with their powerful learning and expressive capabilities, play a crucial role in GAI.

[0037] Generative Adversarial Networks (GANs) are a typical example of GAN-based neural network models. A GAN consists of two networks: a generator and a discriminator, which compete against and learn from each other. The generator is responsible for generating fake data, while the discriminator is responsible for judging the authenticity of the data. During training, the generator continuously attempts to generate more realistic images to deceive the discriminator, while the discriminator continuously improves its judgment to identify the generator's deceptions. This adversarial learning process allows the generator to gradually learn to generate high-quality data. The generator itself is constructed from a neural network, which requires the use of activation functions. The specific steps involved are readily available to those skilled in the art and will not be elaborated upon here.

[0038] In a first aspect of the present invention, a method for automatically running maintenance tasks is provided, the method comprising:

[0039] Collect and acquire low-code orchestration features, and parse natural language work orders into Playbook structure features based on natural language processing technology, synchronously acquire host load parameters, and acquire server certificate execution methods;

[0040] The original automated operation and maintenance features are formed by merging and integrating the low-code orchestration features, the Playbook structure features, and the host load parameters.

[0041] The Playbook structure features are obtained based on reinforcement learning methods. The corresponding low-code orchestration generation features and host load generation parameters are set by the technicians. The low-code orchestration generation features, the Playbook structure features, and the host load generation parameters are merged and fused to form automatic operation and maintenance generation features. The server certificate generation execution method is obtained simultaneously.

[0042] Based on the automated operation and maintenance features, the server certificate execution method, the automated operation and maintenance generation features, and the server certificate generation and execution method, an automated certificate execution operation and maintenance model is constructed.

[0043] The certificate execution and maintenance automation model is run in a container sandbox to obtain the corresponding intelligent execution method for server certificates in the new script maintenance environment. Risks are isolated based on the intelligent execution method for server certificates, and the model is applied to intelligently batch execution and avoid faulty nodes when updating multiple server certificates.

[0044] In this embodiment, a method for automatically running operation and maintenance tasks is shown in Figure 1.

[0045] Furthermore, the low-code orchestration features include drag-and-drop flowchart parameters, flowchart orchestration parameters, and data structure parameters. The drag-and-drop flowchart parameters include flowchart text feature values, colors, and data source calculations.

[0046] In this embodiment, low-code orchestration features include core architecture parameters, UI interface orchestration parameters, and integrated extension parameters. However, during model data processing, only three types of parameters that can be represented in data form—drag-and-drop flowchart parameters, process orchestration parameters, and data structure parameters—are selected as low-code orchestration features to adapt to subsequent reinforcement learning and neural network model training and use.

[0047] Figure 3 shows the application diagram related to the low-code orchestration feature in this embodiment.

[0048] This embodiment therefore only provides a detailed description of drag-and-drop flowchart parameters, flowchart orchestration parameters, and data structure parameters. Drag-and-drop flowchart parameters are a type of UI interface orchestration parameter, possessing variable attributes, including text, color, and data source. Flowchart orchestration parameters include the supported node types, connection conditions, and the product of variable parameters for the entire script. Data structure parameters primarily consist of data table structures, which are divided into field names, types, and indexes, obtained by multiplying the set values ​​corresponding to the field names, types, and indexes. Therefore, in this embodiment, the low-code orchestration features are represented by three numerical values ​​characterizing the generation of the server certificate execution method.

[0049] Furthermore, the parameters of the drag-and-drop flowchart are calculated using flowchart text feature values, colors, and data sources:

[0050]

[0051] In the formula, This represents the total number of drag-and-drop flowchart parameters during low-code orchestration, where i represents the i-th drag-and-drop flowchart during low-code orchestration, and n represents the total number of drag-and-drop flowcharts. This represents the text feature value of the i-th drag-and-drop flowchart. Indicates the color complexity of the i-th drag-and-drop flowchart. The complexity of the data source for the i-th drag-and-drop flowchart is represented by the complexity of the color, which is defined by the amount of memory occupied by the color in the computer, and the complexity of the data source is represented by the reciprocal of the time taken for data transfer during script execution.

[0052] The more complex the parameters of the drag-and-drop flowchart, the more complex the operation and maintenance method. Therefore, in this embodiment, the drag-and-drop flowchart parameters are specifically calculated based on the flowchart text feature value, color, and data source complexity to characterize the automatic operation and maintenance task.

[0053] Furthermore, the step of parsing the natural language work order into Playbook structural features based on natural language processing technology involves parsing the text in the natural language work order into descriptions and processing them accordingly to set description feature values.

[0054] In this embodiment, the text in the natural language work order is parsed into descriptions and processed accordingly, and set as description feature values, as shown in the table below:

[0055] ;

[0056] The feature values ​​are set by those skilled in the art according to the needs of subsequent model training and processing. Therefore, a Playbook structural feature in this embodiment can be (234, 133, 165).

[0057] Furthermore, the host load parameters include memory utilization, read / write latency, and CPU utilization;

[0058] Due to the correlation between operation and maintenance tasks and host load, this application fully considers the most important host load parameters, such as memory usage, read / write latency, and CPU usage, to form the relevant parameters for building the model, so that the intelligent execution method of server certificates obtained by the model is more objective and realistic.

[0059] Furthermore, the reinforcement learning method utilizes a generative adversarial network (GAN) model, adjusting it based on the drag-and-drop flowchart parameters. Specifically, it uses the drag-and-drop flowchart parameters to modify the generator of the GAN model. The generator employs a deep neural network, and its activation function is modified using the drag-and-drop flowchart parameters. The modified activation function is as follows:

[0060]

[0061] In the formula, For the activation function value, Let m be the total drag-and-drop flowchart parameter for low-code orchestration of the j-th data in the training set, and m be the total amount of data in the training set. The weighted sum of the input Playbook structural features is used to modify the activation function by dragging and dropping flowchart parameters. This makes the Playbook structural features generated by the generator more suitable for generating automatic operation and maintenance features in this scenario. At the same time, since the automatic operation and maintenance features generated by the generator have a small difference from the actual data, the discriminator is selected in the conventional way in this embodiment. By optimizing the execution parameters through generated data, the original time consumption of script operation and maintenance is reduced, and the efficiency of building the subsequent multi-parameter correction neural network model is improved.

[0062] Figure 4 shows the basic principle diagram of the activation function in this embodiment.

[0063] Furthermore, the original automated operation and maintenance features are formed by merging and fusing the low-code orchestration features, the Playbook structure features, and the host load parameters, and the automated operation and maintenance generation features are formed by merging and fusing the low-code orchestration generation features, the generated Playbook structure features, and the host load generation parameters, both of which are obtained through data dimension linking.

[0064] In this embodiment, an original automated operation and maintenance feature is (low-code orchestration feature, Playbook structure feature, host load parameters), and an automated operation and maintenance generation feature is (low-code orchestration generation feature, generated Playbook structure feature, host load generation parameters).

[0065] Furthermore, a unified training dataset is constructed based on the automated operation and maintenance features, the server certificate execution method, the automated operation and maintenance generation features, and the server certificate generation and execution method. The certificate execution operation and maintenance automation model is constructed using this unified training dataset. The certificate execution operation and maintenance automation model is a neural network model with an improved activation function, and the activation function it adopts is:

[0066]

[0067] In the formula, X represents the activation function value, where X is the feature value of the original automated operation and maintenance features or the features generated by automated operation and maintenance through a linear transformation using weights and biases. The total number of features generated by automated operations and maintenance is added to the original automated operations and maintenance features used to train the model. The host load parameter coefficient for the kth training set data is calculated by taking the reciprocal of the product of memory utilization, read / write latency, and CPU utilization.

[0068] In the activation function, changes in the activation function can effectively show the gradient changes of the model under the application scenario, thus affecting the classification accuracy of the neural network model. By using the host load parameter coefficients unique to this scenario to correct the model, a more accurate intelligent execution method for server certificates can be obtained.

[0069] It also provides a system for automatically running operation and maintenance tasks. This system implements a method for automatically running operation and maintenance tasks, including a low-code orchestration feature acquisition module, a natural language work order parsing module, a host load parameter acquisition module, a certificate execution operation and maintenance automation model construction module, and a server certificate intelligent execution module.

[0070] The low-code orchestration feature acquisition module is used to collect and acquire low-code orchestration features.

[0071] The natural language work order parsing module has a built-in natural language processing model, which is used to parse natural language work orders into Playbook structural features based on natural language processing technology;

[0072] The host load parameter acquisition module is used to acquire host load parameters and acquire server certificate execution methods.

[0073] The certificate execution and maintenance automation model construction module is used to obtain the Playbook structure features based on reinforcement learning methods, and according to the low-code orchestration generation features and corresponding host load generation parameters set by technical personnel, the low-code orchestration generation features, the Playbook structure features, and the host load generation parameters are merged and fused to form automatic operation and maintenance generation features, and the server certificate generation execution method is obtained simultaneously.

[0074] Based on the automated operation and maintenance features, the server certificate execution method, the automated operation and maintenance generation features, and the server certificate generation and execution method, an automated certificate execution operation and maintenance model is constructed.

[0075] The server certificate intelligent execution module is used to run the certificate execution operation and maintenance automation model in a container sandbox, obtain the server certificate intelligent execution method corresponding to the new script operation and maintenance environment, isolate risks according to the server certificate intelligent execution method, and apply it to intelligent batch execution and avoid fault nodes when updating multiple server certificates.

[0076] Furthermore, the reinforcement learning method utilizes a generative adversarial network (GAN) model, adjusting it based on the drag-and-drop flowchart parameters. Specifically, it uses the drag-and-drop flowchart parameters to modify the generator of the GAN model. The generator employs a deep neural network, and its activation function is modified using the drag-and-drop flowchart parameters. The modified activation function is as follows:

[0077]

[0078] In the formula, For the activation function value, Let m be the total drag-and-drop flowchart parameter for low-code orchestration of the j-th data in the training set, and m be the total amount of data in the training set. The weighted sum of the input Playbook structural features is used to modify the activation function by dragging and dropping flowchart parameters. This makes the Playbook structural features generated by the generator more suitable for generating automatic operation and maintenance features in this scenario. At the same time, since the automatic operation and maintenance features generated by the generator have a small difference from the actual data, the discriminator is selected in the conventional way in this embodiment. By optimizing the execution parameters through generated data, the original time consumption of script operation and maintenance is reduced, and the efficiency of building the subsequent multi-parameter correction neural network model is improved.

[0079] Therefore, the beneficial effects of the present invention are as follows:

[0080] This invention obtains automated operation and maintenance (O&M) features by merging low-code orchestration features, Playbook structure features, and host load parameters, and acquires server certificate execution methods. It then constructs an automated certificate execution O&M model using these features and methods. This invention considers low-code orchestration features, Playbook structure features, and host load parameters to construct a multi-parameter corrective neural network model. Reinforcement learning is used to generate more Playbook structure feature data, enabling the automated certificate execution O&M model to more accurately and intelligently derive intelligent server certificate execution methods. Simultaneously, the task runs in a container sandbox to isolate risks. Combined with a drag-and-drop flowchart parameter-based self-healing mechanism to generate executable scripts, this is applied to intelligently batch execution and avoid faulty nodes during multi-server certificate updates.

[0081] Of course, it is understood that each embodiment of the present invention can achieve one of the effects individually, and the combination of multiple embodiments of the present invention can achieve all the above effects. However, it is not required that each embodiment of the present invention achieve all the above advantages and effects, because each embodiment of the present invention can constitute a separate technical solution and make one or more contributions to the prior art. Each embodiment of the present invention is not affected, and the solution can still be implemented even if one embodiment is deleted.

[0082] For any module structures not specifically defined in this invention, the existing technical descriptions shall prevail. The prior art mentioned in the foregoing background and specific embodiments sections can be considered part of this invention and used to understand the meaning of certain technical features or parameters.

Claims

1. A method for automatically running operation and maintenance tasks, characterized in that, The method includes: collecting and acquiring low-code orchestration features, parsing natural language work orders into Playbook structure features using natural language processing technology, synchronously acquiring host load parameters, and acquiring server certificates to execute methods; merging and integrating the low-code orchestration features, the Playbook structure features, and the host load parameters to form original automated operation and maintenance features; the low-code orchestration features include drag-and-drop flowchart parameters, process orchestration parameters, and data structure parameters, wherein the drag-and-drop flowchart parameters are calculated using flowchart text feature values, colors, and data sources. In the formula, This represents the total number of drag-and-drop flowchart parameters during low-code orchestration, where i represents the i-th drag-and-drop flowchart during low-code orchestration, and n represents the total number of drag-and-drop flowcharts. This represents the text feature value of the i-th drag-and-drop flowchart. Indicates the color complexity of the i-th drag-and-drop flowchart. This represents the data source complexity of the i-th drag-and-drop flowchart. The Playbook structure features are obtained using reinforcement learning, and corresponding low-code orchestration generation features and host load generation parameters are set by technical personnel. These low-code orchestration features, Playbook structure features, and host load generation parameters are then merged to form automatic operation and maintenance generation features, and a server certificate generation execution method is obtained simultaneously. The reinforcement learning method utilizes a generative adversarial network (GAN) model, adjusted based on the drag-and-drop flowchart parameters. Specifically, the generator of the GAN model is modified using the drag-and-drop flowchart parameters. The generator employs a deep neural network, and its activation function is modified using the drag-and-drop flowchart parameters. The modified activation function is: In the formula, For the activation function value, Let m be the total drag-and-drop flowchart parameter for low-code orchestration of the j-th data in the training set, and m be the total amount of data in the training set. The input Playbook structure features are weighted and summed; a certificate execution operation and maintenance automation model is constructed based on the automatic operation and maintenance features, the server certificate execution method, the automatic operation and maintenance generation features, and the server certificate generation and execution method; the certificate execution operation and maintenance automation model is run in a container sandbox to obtain the server certificate intelligent execution method corresponding to the new script operation and maintenance environment; risks are isolated according to the server certificate intelligent execution method, and it is applied to intelligent batch execution and avoidance of fault nodes when updating multiple server certificates.

2. The method for automatically running operation and maintenance tasks as described in claim 1, characterized in that: The process of parsing a natural language work order into Playbook structural features using natural language processing technology involves parsing the text in the natural language work order into descriptions and processing them accordingly to set description feature values.

3. The method for automatically running operation and maintenance tasks as described in claim 2, characterized in that: The host load parameters include memory usage, read / write latency, and CPU usage.

4. The method for automatically running operation and maintenance tasks as described in claim 3, characterized in that: The original automated operation and maintenance features are formed by merging and fusing the low-code orchestration features, the Playbook structure features, and the host load parameters. The automated operation and maintenance generated features are formed by merging and fusing the low-code orchestration generated features, the generated Playbook structure features, and the host load generated parameters. Both are obtained through data dimension linking.

5. The method for automatically running operation and maintenance tasks as described in claim 4, characterized in that: A unified training dataset is constructed based on the automated operation and maintenance features, the server certificate execution method, the automated operation and maintenance generation features, and the server certificate generation and execution method. The certificate execution operation and maintenance automation model is constructed using this unified training dataset. The certificate execution operation and maintenance automation model is a neural network model with an improved activation function, and the activation function it adopts is: In the formula, X represents the activation function value, where X is the feature value of the original automated operation and maintenance features or the features generated by automated operation and maintenance through a linear transformation using weights and biases. The total number of features generated by automated operations and maintenance is added to the original automated operations and maintenance features used to train the model. The host load parameter coefficient for the k-th training set data is calculated by taking the reciprocal of the product of memory usage, read / write latency, and CPU usage. In the activation function, the change of the activation function can well show the gradient change of the model under the application scenario, thus affecting the classification accuracy of the neural network model. By using the host load parameter coefficient unique to this scenario to correct the model, a more accurate intelligent execution method for server certificates can be obtained.

6. A system for automatically running operation and maintenance tasks, the system implementing the method as described in claim 5, comprising a low-code orchestration feature acquisition module, a natural language work order parsing module, a host load parameter acquisition module, a certificate execution operation and maintenance automation model construction module, and a server certificate intelligent execution module, characterized in that: The low-code orchestration feature acquisition module is used to collect and acquire low-code orchestration features. The natural language work order parsing module has a built-in natural language processing model, which is used to parse natural language work orders into Playbook structural features based on natural language processing technology; The host load parameter acquisition module is used to acquire host load parameters and server certificate execution methods. The certificate execution and maintenance automation model construction module is used to obtain Playbook structure features based on reinforcement learning methods, and according to the low-code orchestration generation features and corresponding host load generation parameters set by technical personnel, merge and integrate the low-code orchestration generation features, the Playbook structure features, and the host load generation parameters to form automatic operation and maintenance generation features, and simultaneously acquire server certificate generation and execution methods. A certificate execution and maintenance automation model is constructed based on the automatic operation and maintenance features and the server certificate execution methods. The server certificate intelligent execution module is used to run the certificate execution and maintenance automation model in a container sandbox, acquire the corresponding server certificate intelligent execution methods in the new script operation and maintenance environment, isolate risks based on the server certificate intelligent execution methods, and apply intelligent batch execution and avoidance of fault nodes when updating multiple server certificates.

7. A system for automatically running maintenance tasks as described in claim 6, characterized in that: The reinforcement learning method utilizes a generative adversarial network (GAN) model, which is adjusted based on the drag-and-drop flowchart parameters. Specifically, the generator of the GAN model is modified using the drag-and-drop flowchart parameters. The generator employs a deep neural network, and its activation function is modified using the drag-and-drop flowchart parameters. The modified activation function is as follows: In the formula, For the activation function value, Let m be the total drag-and-drop flowchart parameter for low-code orchestration of the j-th data in the training set, and m be the total amount of data in the training set. The sum of the weighted values ​​of the input Playbook structural features.

Images