Container cloud intelligent console implementation method, system and device and storage medium

Abstract

The invention discloses a container cloud intelligent console implementation method, system and device and a storage medium, belongs to the technical field of cloud native technologies of Internet technologies, and solves the technical problem that a large language model has limitation when being applied to Kubernetes operation and maintenance. The method comprises the following steps: selecting an LLM pre-training large model; the method comprises the following steps: establishing a Kubernetes knowledge base through a data collector; retraining the LLM pre-training model by adopting the large model instruction data set; performing parameter fine tuning on the LLM pre-training model; the method comprises the following steps: deploying an LLM large model service, an AI middle console service related to large model interaction and an intelligent container cloud management console system in a Kubernetes cluster; providing the following seven intelligent functions according to the output of the LLM pre-training model; carrying out production environment feedback and data backflow; updating a knowledge base according to feedback and data backflow, and retraining the LLM pre-training model; and redeploying the retrained LLM pre-training model. Therefore, the operation and maintenance operation becomes more efficient, accurate and intelligent.

Description

Technical Field

[0001] The present invention relates to the field of Internet technology and cloud native technology, and more specifically, to a method, system, device and storage medium for implementing a container cloud intelligent console. Background Art

[0002] With the rapid development of information technology, cloud computing has become a core component of enterprise IT structure. It provides flexible and scalable IT resources through the network, allowing enterprises to quickly adjust computing, storage and network resources according to demand. The development of cloud computing has spawned a series of new technologies, among which containerization technology is particularly prominent. Container is a lightweight form of virtualization that allows applications and their dependencies to be packaged together for rapid deployment and operation in any environment. Due to its efficient resource utilization and portability, container technology has quickly gained favor among developers and enterprises.

[0003] Among the many container orchestration tools, Kubernetes (K8s), as an open source project, has become an industry standard. Kubernetes provides powerful functions to automate the deployment, expansion and management of containers, allowing applications to run flexibly and efficiently in multiple cloud environments. With the popularity of microservice architecture, more and more companies are beginning to use Kubernetes to improve the maintainability and elasticity of applications. However, the complexity of Kubernetes also brings new challenges to developers and operators (DevOps).

[0004] Behind the flexibility and power of Kubernetes lies the complexity of management and operation. Here are some of the main challenges:

[0005] Complex configuration management: Kubernetes relies on YAML files for configuration and resource management. Any error in the YAML file may cause service interruption or failure. Ensuring the accuracy and consistency of the configuration file is a difficult task.

[0006] Difficulty in troubleshooting: In distributed systems, failures may be caused by a variety of reasons, including network problems, insufficient resources, configuration errors, etc. Operation and maintenance personnel need to have rich experience and knowledge to quickly locate problems and develop solutions.

[0007] Insufficient resource optimization: As the scale of applications increases, how to reasonably allocate, manage, and optimize cluster resources to improve performance and reduce costs is an urgent problem to be solved.

[0008] The ever-changing technology stack: The Kubernetes ecosystem changes very quickly, with new tools and technologies emerging in an endless stream. Operations and maintenance personnel need to continue learning and mastering new knowledge to keep up with the pace of technology.

[0009] Complex operational processes: Traditional DevOps processes often have problems with fragmented tool chains and lack a global perspective and optimized decision support system, which makes the continuous integration and continuous deployment (CI / CD) process inefficient and error-prone.

[0010] In recent years, natural language processing (NLP) technology based on deep learning has made breakthrough progress, especially the emergence of large language models (LLM), which has opened a new chapter in artificial intelligence. Models such as OpenAI's GPT series and Google's BERT have demonstrated superb capabilities in language understanding and generation. These models can learn the structure, semantics and contextual relationships of language by training large amounts of text data, and then generate fluent text, perform language translation, answer questions, and conduct conversations.

[0011] These large language models provide new ideas for solving complex cognitive tasks. For example, by interacting with computers through natural language, users can complete tasks in a more intuitive and easy way without having to deeply understand the underlying technical details. This intelligent technical solution has achieved remarkable success in many fields, and some companies have even begun to apply it to scenarios such as automated customer service, content generation, and code review.

[0012] Combining the big language model with Kubernetes operation and maintenance can provide great help to operation and maintenance personnel:

[0013] 1. Intelligent troubleshooting: The large language model can analyze error logs, system status, and historical data, quickly present potential problems and solutions to operation and maintenance personnel, and shorten troubleshooting time.

[0014] 2. Automated configuration recommendation: Through natural language processing, the large language model can automatically generate or recommend configuration files based on user needs, reducing the complexity of manual operations and avoiding human errors.

[0015] 3. Dynamic learning and adaptation: Combining machine learning and user feedback, the system can continuously update the knowledge base to adapt to new operation and maintenance scenarios and technical changes, enhancing system flexibility.

[0016] 4. Optimize decision support: Utilize the reasoning capabilities of large language models to provide data-driven recommendations for complex decisions such as resource allocation and architecture optimization, thereby improving operation and maintenance efficiency.

[0017] 5. User-friendly interface: Interacting with users through natural language reduces the barriers between users and technology, allowing people without a deep technical background to effectively manage Kubernetes clusters.

[0018] Although large language models have shown their potential in many areas, they still have some limitations when applied to Kubernetes operations. For example:

[0019] Insufficient knowledge depth: While large language models are capable of generating natural language text, they may not be able to fully grasp the deep knowledge of the Kubernetes domain, especially when dealing with specific technical problems.

[0020] Model adaptability: Since large language models are mainly based on general corpora, when they are applied to specific fields, they may need to be retrained to improve their performance on specific tasks.

[0021] Data privacy and security: How to ensure data privacy and security when handling sensitive enterprise information remains an important challenge. When users interact with AI systems, how to protect their data from being abused will be a major consideration in the design. Summary of the invention

[0022] The technical problem to be solved by the present invention is to address the above-mentioned deficiencies in the prior art. The first purpose of the present invention is to provide a method for implementing a container cloud intelligent console, which can combine a large language model with container cloud management, and realize end-to-end operation and maintenance automation through an intelligent console, making operation and maintenance operations more efficient, accurate and intelligent.

[0023] The second objective of the present invention is to provide a container cloud intelligent console implementation system.

[0024] A third object of the present invention is to provide a computer device.

[0025] A fourth objective of the present invention is to provide a computer storage medium.

[0026] In order to achieve the above-mentioned objective 1, the present invention provides a method for implementing a container cloud intelligent console, comprising the following steps:

[0027] Step S1. Select the type of LLM big model to be accessed; by adapting a variety of LLM big model interfaces, different LLM big models can be accessed, including open source general big models and industry big models on the Internet; after determining the LLM pre-trained big model to be used, modify the corresponding configuration file to complete the access of the LLM big model;

[0028] Step S2. Establish a Kubernetes knowledge base through a data collector; the data collector automatically collects the large model instruction data set in the Kubernetes cluster, and also collects the knowledge data actively uploaded by the user; the data collector integrates the large model instruction data set and the knowledge data actively uploaded by the user to form a knowledge base;

[0029] Step S3. Based on the LLM pre-trained model, the LLM pre-trained model is retrained using the large model instruction data set to enhance the LLM pre-trained model's ability to grasp, reason, and generate Kubernetes knowledge;

[0030] Step S4. Fine-tune the parameters of the retrained LLM pre-trained model to improve the versatility and correctness of the solution provided by the LLM pre-trained model;

[0031] Step S5. Deploy the LLM large model service of the fine-tuned LLM pre-trained model, the AI ​​middle platform service involving large model interaction, and the intelligent container cloud management console system in the Kubernetes cluster;

[0032] Step S6. The intelligent container cloud management console system interacts with the fine-tuned LLM pre-trained model through the AI ​​middle platform service, and provides the following 7 intelligent functions according to the output of the LLM pre-trained model: intelligent configuration recommendation, intelligent YAML completion, Kubernetes error analysis, container application deployment optimization suggestions, network topology optimization suggestions, pipeline compilation error analysis, alarm analysis and processing methods;

[0033] Step S7. Perform production environment feedback and data reflow;

[0034] Step S8. After automatic or manual screening by the smart container cloud management console system, the returned data is added to the knowledge base;

[0035] Step S9. The knowledge base is updated according to the feedback and data reflux, and the LLM pre-trained model is retrained according to the updated knowledge base;

[0036] Step S10. Redeploy the LLM large model service of the retrained LLM pre-trained model, and continue to iteratively optimize the model performance.

[0037] As a further improvement, in step S2, the large model instruction dataset focuses on Kubernetes operation and maintenance, and is a specific solution covering various Kubernetes problems, including Kubernetes official guidance, all YAML data of the current cluster, ETCD data, Kubernetes exception analysis and solutions on the Internet, and compilation and construction error analysis and solutions.

[0038] Furthermore, in step S2, document data uploaded by the user and custom knowledge data added by the user to the knowledge base are obtained; the document data includes word files, pdf files and directly uploaded text data; word files and pdf files will be parsed into text and included in the knowledge base together with the directly uploaded text data.

[0039] Furthermore, in step S6, first select the question that needs to be submitted according to the intelligent container cloud management console system to request the LLM pre-trained model, determine the category to which the question belongs, and then submit it to the corresponding large model Agent to process the intelligent function of the corresponding category.

[0040] Furthermore, in step S6, the seven intelligent functions are specifically:

[0041] Intelligent configuration recommendation: When creating Kubernetes resources in wizard mode, the intelligent container cloud management console system automatically fills in some input boxes and highlights the options that need to be enabled or selected, which can reduce the need for users to frequently copy and paste scenarios and nearly duplicate options;

[0042] Intelligent YAML completion: When writing a Kuberntes YAML file to define a Kubernetes resource object, the intelligent container cloud management console system will automatically complete the YAML at the current cursor. You can use shortcut keys to apply the YAML fragment added by intelligent YAML completion.

[0043] Kubernetes error analysis: When an error occurs in the container cloud console, the intelligent analysis function can be triggered. The intelligent container cloud management console system will combine the current content of the cluster with official documents and solutions on the Internet to describe in detail the possible causes of the error and the corresponding solutions.

[0044] Container application deployment optimization suggestions: It can provide intelligent optimization suggestions for deployed applications, improve performance and resource utilization through some intelligent optimization suggestions, make containerized application deployment more intelligent and efficient, and improve resource utilization and performance; intelligent optimization suggestions include: adjusting the number of deployments, adding health checks, node affinity suggestions, node anti-affinity suggestions, and pod affinity suggestions;

[0045] Network topology optimization suggestions: It can provide optimization suggestions for service discovery methods, select optimization strategies that are suitable for the current architecture and business needs, effectively manage resources and reduce operation and maintenance costs; optimization suggestions include: service type, adjustment of Pod distribution plan;

[0046] Pipeline compilation error analysis: When a compilation error occurs during pipeline execution, the Smart Container Cloud Management Console system will generate a corresponding error log, which records the specific location of the error, error code, and related context information. Through the pipeline compilation error analysis function, the root cause of the compilation error can be accurately determined by analyzing the error log. The root cause includes syntax problems, missing dependencies, and configuration errors. For each error, the Smart Container Cloud Management Console system will provide specific solutions, including code modification operations, dependency update operations, and configuration adjustment operations.

[0047] Alarm analysis: The alarm analysis function can analyze historical alarm data to discover potential problem patterns and trends; and provide regularly generated alarm reports that summarize alarm events and their solutions in the recent period.

[0048] Furthermore, in step S7, feedback and data reflow include two forms:

[0049] The first form is: when you actively select the recommended configuration items and use YAML completion, the adoption status is automatically submitted to the AI ​​middle office service to mark the approved data;

[0050] The second form is: after the LLM pre-trained model submits analysis and provides suggestions, the page of the intelligent container cloud management console system will provide an option to agree or deny the suggestions of the large model below the answer; users can make choices based on whether the analysis is correct and whether the suggestions are reasonable and effective, thereby submitting data to the AI ​​middle-end service; in addition, users can also upload their own answers for data return.

[0051] In order to achieve the above-mentioned second objective, the present invention provides a container cloud intelligent console implementation system, including:

[0052] The selection module is used to select the type of LLM big model to be connected; by adapting a variety of LLM big model interfaces, different LLM big models can be connected, including open source general big models and industry big models on the Internet; after determining the LLM pre-trained big model to be used, modify the corresponding configuration file to complete the access of the LLM big model;

[0053] Create a knowledge base module to build a Kubernetes knowledge base through a data collector. The data collector automatically collects the large model instruction data set in the Kubernetes cluster, and also collects the knowledge data actively uploaded by the user. The data collector integrates the large model instruction data set and the knowledge data actively uploaded by the user to form a knowledge base.

[0054] A training module is used to retrain the LLM pre-trained model using the large model instruction data set based on the LLM pre-trained model, thereby enhancing the LLM pre-trained model's ability to grasp, reason, and generate Kubernetes knowledge; and fine-tune the parameters of the retrained LLM pre-trained model to improve the versatility and correctness of the solution provided by the LLM pre-trained model;

[0055] The deployment module is used to deploy the LLM large model service of the fine-tuned LLM pre-trained model, the AI ​​middle platform service involving large model interaction, and the intelligent container cloud management console system in the Kubernetes cluster;

[0056] Intelligent function module, used to interact the intelligent container cloud management console system with the fine-tuned LLM pre-trained model through the AI ​​middle platform service, and provide the following 7 intelligent functions according to the output of the LLM pre-trained model: intelligent configuration recommendation, intelligent YAML completion, Kubernetes error analysis, container application deployment optimization suggestions, network topology optimization suggestions, pipeline compilation error analysis, alarm analysis and processing methods;

[0057] A feedback module is used for production environment feedback and data reflow; after automatic or manual screening by the intelligent container cloud management console system, the reflowed data is added to the knowledge base;

[0058] An update module is used to update the knowledge base according to feedback and data reflux, and the LLM pre-trained model is retrained according to the updated knowledge base; the LLM large model service of the retrained LLM pre-trained model is redeployed to continuously iteratively optimize the model performance.

[0059] In order to achieve the third objective above, the present invention provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor implements the above-mentioned container cloud intelligent console implementation method when executing the computer program.

[0060] In order to achieve the above-mentioned fourth objective, the present invention provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the above-mentioned method for implementing a container cloud intelligent console is implemented.

[0061] Beneficial Effects

[0062] Compared with the prior art, the present invention has the following advantages:

[0063] 1. The present invention significantly reduces the working time of operation and maintenance personnel through automated operation and maintenance processes and intelligent troubleshooting, enabling the team to solve problems more quickly, thereby improving overall work efficiency. The analysis capabilities and intelligent recommendation functions of large language models can reduce the incidence of human errors, enhance the accuracy of configuration and operation, and ensure the stability and reliability of the system. The introduction of intelligent monitoring and early warning mechanisms in operation and maintenance management can identify potential security risks in a timely manner and reduce the risk of data leakage and system failures through rapid response.

[0064] 2. The present invention supports intelligent decision-making. Through self-learning and loop feedback mechanisms, the console can continuously adjust itself according to the changing environment, provide data-based intelligent decision-making support, and help the operation and maintenance team make more targeted optimizations and adjustments. Based on intelligent tracking of resource usage, the system dynamically adjusts resource allocation to maximize resource utilization efficiency and reduce the waste of idle resources. This optimization directly leads to cost reduction, so that enterprises can effectively control operation and maintenance costs and improve the overall response speed and market competitiveness of the business.

[0065] 3. The present invention has a user-friendly experience. Users interact with the console through natural language, which lowers the technical threshold, enables non-technical personnel to operate smoothly, and broadens the scope of use.

[0066] 4. The present invention can improve the operation and maintenance efficiency of Kubernetes by building an intelligent container cloud management solution, making the operation and maintenance operations more efficient, accurate and intelligent. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] Figure 1 is a flow chart of the present invention;

[0068] Figure 2 A summary diagram of the functions provided for the Smart Container Cloud Management Console. DETAILED DESCRIPTION

[0069] The present invention will be further described below with reference to the specific embodiments in the accompanying drawings.

[0070] See also Figure 1-2 , a method for implementing a container cloud intelligent console, comprising the following steps S1 to S10:

[0071] Step S1. Select the type of LLM big model to be connected; connect different LLM big models by adapting multiple LLM big model interfaces. Multiple LLM big models include open source general big models and industry big models on the Internet. After determining the LLM pre-trained big model to be used, modify the corresponding configuration file to complete the access of the LLM big model. As the core component of the system, the LLM big model is responsible for processing and analyzing complex data and instructions.

[0072] By adapting to different large language model interfaces, we can enable access to multiple models, provide users with flexible choices, and ensure that the system can be adjusted according to specific needs.

[0073] Step S2. Establish a Kubernetes knowledge base through a data collector. The data collector will automatically collect large model instruction data sets in the Kubernetes cluster, and will also collect knowledge data actively uploaded by users. The data collector integrates the large model instruction data sets and the knowledge data actively uploaded by users to form a knowledge base. The data collector is used to collect and integrate multi-source data, including but not limited to: collecting Kubernetes official guidance, all YAML data of the current cluster, ETCD data, Kubernetes exception analysis and solutions on the Internet, compilation and construction error analysis and solutions, and knowledge data actively uploaded by users. These data will be integrated and a knowledge base will be established for the system to analyze and make decisions.

[0074] Furthermore, the large model instruction dataset focuses on Kubernetes operation and maintenance, and is a specific solution covering various Kubernetes problems, including Kubernetes official guidance, all YAML data of the current cluster, ETCD data, Kubernetes exception analysis and solutions on the Internet, and compilation and construction error analysis and solutions.

[0075] Furthermore, the document data uploaded by the user and the custom knowledge data added by the user to the knowledge base are obtained; the document data includes word files, pdf files and text data uploaded directly; the word files and pdf files will be parsed into text and included in the knowledge base together with the text data uploaded directly.

[0076] By implementing data collection and building a rich knowledge base, including user-defined data, we ensure that the model has targeted learning and reasoning capabilities.

[0077] Step S3. Based on the LLM pre-trained model, the LLM pre-trained model is retrained using a large model instruction data set to enhance the LLM pre-trained model's ability to grasp, reason, and generate Kubernetes knowledge.

[0078] Step S4. Fine-tune the parameters of the retrained LLM pre-trained model to improve the versatility and correctness of the solution provided by the LLM pre-trained model.

[0079] Utilize user feedback to achieve dynamic knowledge iteration and model retraining to ensure the availability and accuracy of the system.

[0080] Step S5. Deploy the LLM large model service of the fine-tuned LLM pre-trained model, the AI ​​middle-end service involving large model interaction, and the intelligent container cloud management console system in the Kubernetes cluster.

[0081] AI middle platform service: responsible for interacting with the large language model, generating feedback data and screening it to optimize the system's decision-making process. The intelligent container cloud management console system provides a user interface, is responsible for interacting with users, and provides intelligent container cloud platform solutions to help users efficiently manage and maintain the container cloud platform.

[0082] Step S6. The intelligent container cloud management console system interacts with the fine-tuned LLM pre-trained model through the AI ​​middleware service, and provides the following seven intelligent functions based on the output of the LLM pre-trained model: intelligent configuration recommendations, intelligent YAML completion, Kubernetes error analysis, container application deployment optimization suggestions, network topology optimization suggestions, pipeline compilation error analysis, alarm analysis, and processing methods.

[0083] Furthermore, first select the questions that need to be submitted according to the intelligent container cloud management console system to request the LLM pre-trained model, determine the category to which the question belongs, and then submit it to the corresponding large model Agent to process the intelligent function of the corresponding category.

[0084] The 7 intelligent functions are as follows:

[0085] Intelligent configuration recommendations: When creating Kubernetes resources using the wizard mode, the intelligent container cloud management console system automatically fills in some input boxes and highlights the options that need to be enabled or selected, which can reduce the need for users to frequently copy and paste scenarios and nearly duplicate options.

[0086] Intelligent YAML completion: When writing a Kuberntes YAML file to define a Kubernetes resource object, the intelligent container cloud management console system will automatically complete the YAML at the current cursor. You can use shortcut keys to apply the YAML fragment added by intelligent YAML completion.

[0087] Kubernetes error analysis: In the case of an error event in the container cloud console, the intelligent analysis function can be triggered. The intelligent container cloud management console system will combine the current content of the cluster with official documents and solutions on the Internet to describe in detail the possible causes of the error and the corresponding solutions.

[0088] Container application deployment optimization suggestions: It can provide intelligent optimization suggestions for deployed applications, improve performance and resource utilization through some intelligent optimization suggestions, make containerized application deployment more intelligent and efficient, and improve resource utilization and performance; intelligent optimization suggestions include: adjusting the number of deployments, adding health checks, node affinity suggestions, node anti-affinity suggestions, and pod affinity suggestions.

[0089] Network topology optimization suggestions: It can provide optimization suggestions for service discovery methods, select optimization strategies suitable for the current architecture and business needs, effectively manage resources and reduce operation and maintenance costs; optimization suggestions include: service type, adjustment of Pod distribution plan.

[0090] Pipeline compilation error analysis: When a compilation error occurs during pipeline execution, the Smart Container Cloud Management Console system will generate a corresponding error log, which records the specific location of the error, error code, and related context information. Through the pipeline compilation error analysis function, you can accurately determine the root cause of the compilation error by analyzing the error log. The root cause includes syntax problems, missing dependencies, and configuration errors. For each error, the Smart Container Cloud Management Console system will provide specific solutions. Solutions include code modification operations, dependency update operations, and configuration adjustment operations.

[0091] Alarm analysis: The alarm analysis function can analyze historical alarm data to discover potential problem patterns and trends. It also provides regularly generated alarm reports that summarize alarm events and their solutions in the recent period.

[0092] Through intelligent decision support, it provides a variety of intelligent functions, including configuration recommendation, fault analysis, resource optimization, etc., to enhance the work efficiency of operation and maintenance personnel.

[0093] Step S7: Provide production environment feedback and data reflow. Feedback and data reflow include two forms:

[0094] The first form is: when you actively select recommended configuration items and use YAML completion, the adoption status is automatically submitted to the AI ​​middleware service to mark the approved data.

[0095] The second form is: after the LLM pre-trained model submits analysis and provides suggestions, the page of the intelligent container cloud management console system will provide an option to choose to agree or deny the suggestions of the large model below the answer; users can make choices based on whether the analysis is correct and whether the suggestions are reasonable and effective, and then submit data to the AI ​​middle platform service. In addition, users can also upload their own answers for data return.

[0096] Step S8. After automatic or manual screening by the intelligent container cloud management console system, the returned data is added to the knowledge base.

[0097] Step S9: The knowledge base is updated according to the feedback and data reflux, and the LLM pre-trained model is retrained according to the updated knowledge base.

[0098] Step S10. Redeploy the LLM large model service of the retrained LLM pre-trained model and continuously iterate and optimize the model performance. The self-learning and loop feedback mechanism enables the console to continuously adjust itself according to the changing environment, provide data-based intelligent decision support, and help the operation and maintenance team make more targeted optimizations and adjustments.

[0099] A container cloud intelligent console implementation system, comprising:

[0100] The selection module is used to select the type of LLM big model to be connected; by adapting a variety of LLM big model interfaces, different LLM big models can be connected, including open source general big models and industry big models on the Internet; after determining the LLM pre-trained big model to be used, modify the corresponding configuration file to complete the access of the LLM big model;

[0101] Create a knowledge base module to build a Kubernetes knowledge base through a data collector. The data collector automatically collects the large model instruction data set in the Kubernetes cluster, and also collects the knowledge data actively uploaded by the user. The data collector integrates the large model instruction data set and the knowledge data actively uploaded by the user to form a knowledge base.

[0102] The training module is used to retrain the LLM pre-trained model using a large model instruction data set based on the LLM pre-trained model, thereby enhancing the LLM pre-trained model's ability to grasp, reason, and generate Kubernetes knowledge; fine-tune the parameters of the retrained LLM pre-trained model to improve the versatility and correctness of the solution provided by the LLM pre-trained model;

[0103] The deployment module is used to deploy the LLM large model service of the fine-tuned LLM pre-trained model, the AI ​​middle platform service involving large model interaction, and the intelligent container cloud management console system in the Kubernetes cluster;

[0104] Intelligent function module, used to interact the intelligent container cloud management console system with the fine-tuned LLM pre-trained model through the AI ​​middle platform service, and provide the following 7 intelligent functions according to the output of the LLM pre-trained model: intelligent configuration recommendation, intelligent YAML completion, Kubernetes error analysis, container application deployment optimization suggestions, network topology optimization suggestions, pipeline compilation error analysis, alarm analysis and processing methods;

[0105] The feedback module is used for production environment feedback and data reflow. After automatic or manual screening by the intelligent container cloud management console system, the reflowed data is added to the knowledge base.

[0106] The update module is used to update the knowledge base based on feedback and data reflux. The LLM pre-trained model is retrained based on the updated knowledge base. The LLM large model service of the retrained LLM pre-trained model is redeployed to continuously iterate and optimize the model performance.

[0107] A computer device includes a memory and a processor, wherein the memory stores a computer program, and the processor implements the above-mentioned container cloud intelligent console implementation method when executing the computer program.

[0108] A computer-readable storage medium stores a computer program, which implements the above-mentioned container cloud intelligent console implementation method when the computer program is executed by a processor.

[0109] The above are only preferred embodiments of the present invention. It should be pointed out that for those skilled in the art, several modifications and improvements can be made without departing from the structure of the present invention, which will not affect the effect of the implementation of the present invention and the practicality of the patent.

Claims

1. A method for implementing a container cloud intelligent console, characterized in that: The following steps are involved: Step S1. Select the type of LLM model to be accessed; By adapting to a variety of LLM big model interfaces, different LLM big models can be accessed, including open source general big models and industry big models on the Internet; after determining the LLM pre-trained big model to be used, modify the corresponding configuration file to complete the access of the LLM big model; Step S2. Establish a Kubernetes knowledge base through a data collector; The data collector will automatically collect large model instruction data sets in the Kubernetes cluster, and will also collect knowledge data actively uploaded by users; The data collector integrates the large model instruction data set and the knowledge data actively uploaded by users to form a knowledge base; Step S3. Based on the LLM pre-trained model, the LLM pre-trained model is retrained using the large model instruction data set to enhance the LLM pre-trained model's ability to grasp, reason, and generate Kubernetes knowledge; Step S4. Fine-tune the parameters of the retrained LLM pre-trained model to improve the versatility and correctness of the solution provided by the LLM pre-trained model; Step S5. Deploy the LLM large model service of the fine-tuned LLM pre-trained model, the AI ​​middle platform service involving large model interaction, and the intelligent container cloud management console system in the Kubernetes cluster; Step S6. The intelligent container cloud management console system interacts with the fine-tuned LLM pre-trained model through the AI ​​middle platform service, and provides the following 7 intelligent functions according to the output of the LLM pre-trained model: intelligent configuration recommendation, intelligent YAML completion, Kubernetes error analysis, container application deployment optimization suggestions, network topology optimization suggestions, pipeline compilation error analysis, alarm analysis and processing methods; Step S7. Perform production environment feedback and data reflow; Step S8. After automatic or manual screening by the smart container cloud management console system, the returned data is added to the knowledge base; Step S9. The knowledge base is updated according to the feedback and data reflux, and the LLM pre-trained model is retrained according to the updated knowledge base; Step S10. Redeploy the LLM large model service of the retrained LLM pre-trained model, and continue to iteratively optimize the model performance.

2. A method for implementing a container cloud intelligent console according to claim 1, characterized in that: In step S2, the large model instruction dataset focuses on Kubernetes operation and maintenance, and is a specific solution covering various Kubernetes problems, including Kubernetes official guidance, all YAML data of the current cluster, ETCD data, Kubernetes exception analysis and solutions on the Internet, and compilation and construction error analysis and solutions.

3. A method for implementing a container cloud intelligent console according to claim 1, characterized in that: In step S2, the document data uploaded by the user and the custom knowledge data added by the user to the knowledge base are obtained; the document data includes word files, pdf files and text data uploaded directly; the word files and pdf files will be parsed into text and included in the knowledge base together with the text data uploaded directly.

4. A method for implementing a container cloud intelligent console according to claim 1, characterized in that: In step S6, first select the question that needs to be submitted according to the intelligent container cloud management console system to request the LLM pre-training model, determine the category to which the question belongs, and then submit it to the corresponding large model Agent to process the intelligent function of the corresponding category.

5. A method for implementing a container cloud intelligent console according to claim 1, characterized in that: In step S6, the seven intelligent functions are specifically: Intelligent configuration recommendation: When creating Kubernetes resources in wizard mode, the intelligent container cloud management console system automatically fills in some input boxes and highlights the options that need to be enabled or selected, which can reduce the need for users to frequently copy and paste scenarios and nearly duplicate options; Intelligent YAML completion: When writing a Kuberntes YAML file to define a Kubernetes resource object, the intelligent container cloud management console system will automatically complete the YAML at the current cursor. You can use shortcut keys to apply the YAML fragment added by intelligent YAML completion. Kubernetes error analysis: When an error occurs in the container cloud console, the intelligent analysis function can be triggered. The intelligent container cloud management console system will combine the current content of the cluster with official documents and solutions on the Internet to describe in detail the possible causes of the error and the corresponding solutions. Container application deployment optimization suggestions: It can provide intelligent optimization suggestions for deployed applications, improve performance and resource utilization through some intelligent optimization suggestions, make containerized application deployment more intelligent and efficient, and improve resource utilization and performance; Intelligent optimization suggestions include: adjusting the number of deployments, adding health checks, node affinity suggestions, node anti-affinity suggestions, and pod affinity suggestions; Network topology optimization suggestions: It can provide optimization suggestions for service discovery methods, select optimization strategies that are suitable for the current architecture and business needs, effectively manage resources and reduce operation and maintenance costs; optimization suggestions include: service type, adjustment of Pod distribution plan; Pipeline compilation error analysis: When a compilation error occurs during pipeline execution, the Smart Container Cloud Management Console system will generate a corresponding error log, which records the specific location of the error, error code, and related context information. Through the pipeline compilation error analysis function, the root cause of the compilation error can be accurately determined by analyzing the error log. The root cause includes syntax problems, missing dependencies, and configuration errors. For each error, the Smart Container Cloud Management Console system will provide specific solutions, including code modification operations, dependency update operations, and configuration adjustment operations. Alarm analysis: The alarm analysis function can analyze historical alarm data to discover potential problem patterns and trends; and provide regularly generated alarm reports that summarize alarm events and their solutions in the recent period.

6. A method for implementing a container cloud intelligent console according to claim 1, characterized in that: In step S7, feedback and data reflow include two forms: The first form is: when you actively select the recommended configuration items and use YAML completion, the adoption status is automatically submitted to the AI ​​middle office service to mark the approved data; The second form is: after the LLM pre-trained model submits analysis and provides suggestions, the page of the intelligent container cloud management console system will provide an option to choose to agree or deny the suggestions of the large model below the answer; Users can submit data to the AI ​​middle-office service by making choices based on whether the analysis is correct and whether the suggestions are reasonable and effective; in addition, users can also upload their own answers for data return.

7. A container cloud intelligent console implementation system, characterized in that: include: Selection module, used to select the type of LLM model to be connected; By adapting to a variety of LLM big model interfaces, different LLM big models can be accessed, including open source general big models and industry big models on the Internet; after determining the LLM pre-trained big model to be used, modify the corresponding configuration file to complete the access of the LLM big model; Create a knowledge base module to build a Kubernetes knowledge base through the data collector; The data collector will automatically collect large model instruction data sets in the Kubernetes cluster, and will also collect knowledge data actively uploaded by users; The data collector integrates the large model instruction data set and the knowledge data actively uploaded by users to form a knowledge base; A training module is used to retrain the LLM pre-trained model using the large model instruction data set based on the LLM pre-trained model, so as to enhance the LLM pre-trained model's ability to grasp, reason and generate Kubernetes knowledge; Fine-tune the parameters of the retrained LLM pre-trained model to improve the versatility and correctness of the solution provided by the LLM pre-trained model; The deployment module is used to deploy the LLM large model service of the fine-tuned LLM pre-trained model, the AI ​​middle platform service involving large model interaction, and the intelligent container cloud management console system in the Kubernetes cluster; Intelligent function module, used to interact the intelligent container cloud management console system with the fine-tuned LLM pre-trained model through the AI ​​middle platform service, and provide the following 7 intelligent functions according to the output of the LLM pre-trained model: intelligent configuration recommendation, intelligent YAML completion, Kubernetes error analysis, container application deployment optimization suggestions, network topology optimization suggestions, pipeline compilation error analysis, alarm analysis and processing methods; Feedback module, used for production environment feedback and data reflow; After automatic or manual screening by the intelligent container cloud management console system, the returned data is added to the knowledge base; An updating module, used for updating the knowledge base according to feedback and data reflux, and the LLM pre-trained model is retrained according to the updated knowledge base; Redeploy the LLM large model service of the retrained LLM pre-trained model and continuously iterate and optimize the model performance.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the method for implementing a container cloud intelligent console as described in any one of claims 1 to 6 is implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, a method for implementing a container cloud intelligent console as described in any one of claims 1 to 6 is implemented.

Images