Method, device, equipment and medium for generating stateful application instance operation and maintenance components

Abstract

The embodiment of the present invention provides a method, device, equipment and medium for generating a stateful application instance operation and maintenance component, wherein the method comprises: obtaining learning data, the learning data comprising the writing syntax of the Kubernetes Operator, the principle of the application state and the operation and maintenance data of the stateful application instance; using a deep learning algorithm to analyze the learning data to obtain the characteristic data of the Kubernetes Operator object, and generating a Kubernetes Operator software package based on the characteristic data, wherein the characteristic data comprises the abstract characteristics of the full life cycle activities of the Kubernetes Operator object. This solution can be beneficial to improving efficiency and improving the performance or accuracy of the constructed Kubernetes Operator.

Description

Technical Field

[0001] The present invention relates to the field of operation and maintenance technology, and in particular to a method, device, equipment and medium for generating a stateful application instance operation and maintenance component. Background Art

[0002] Kubernetes (k8s for short), as an open source container scheduling platform across host clusters, provides a container-centric infrastructure and has become the de facto standard in the field of container orchestration since 2017.

[0003] like Figure 1 As shown in the figure, the core of Kubernetes project orchestration capability lies in the design and use of its declarative API. By performing CRUD operations on the API group under the REST path, resource information can be obtained and changed. There are many types of resources in Kubernetes, such as pods (the smallest atomic scheduling unit), services, events, etc. Usually, most users only care about resource types such as pods and services that are directly related to the success of service operation. They only pay attention to cluster operation information when an exception occurs in the cluster or service deployment and problem tracking is required.

[0004] Currently, the recording of cluster internal operation information is mainly done through events. Events contain information generated by the core components kubelet and ControllerManager (control manager) during the operation of Kubernetes. Analyzing the information contained in events and the change history of events can effectively track problems and solve them in a targeted manner. Currently, events are mainly viewed in the industry through the kubectl (binary executable file, which is effective by accessing the kube-apiserver) command. Specific commands such as kubectl get events and kubectl describe [resourceName] can be used to view events content. The existing method is command-line level operation, which requires access to the cluster. To obtain accurate events information, it is necessary to set complex parameters for multi-condition matching, and the final query result display is not user-friendly. For some specific needs, such as querying the corresponding pod name changes before and after an application upgrade, multiple queries and manual matching are required, and it is not easy to normalize the operation.

[0005] It is still a difficult task for Kubernetes to manage the high availability, fault protection, and persistence of stateful applications. Kubernetes version 1.5 provides StatefulSet to handle stateful applications, but the implementation and effect are not good. To solve this problem, Kubernetes Operator appeared after Kubernetes 1.9+, which can well solve the high availability, fault protection, and persistence of stateful applications. Figure 2 As shown in the figure, Kubernetes Operator is a framework launched by CoreOS to simplify the management of complex stateful applications. It is a controller that perceives the application state and automatically creates, manages and configures application instances by extending the Kubernetes API. Kubernetes Operator extends new application resources based on CustomerResources Define and ensures that the application is in the expected state through the controller. For example, etcd operator simulates the behavior of managing etcd (a distributed database) cluster through the following three steps:

[0006] 1. Observe the current status of the cluster through the Kubernetes API;

[0007] 2. Analyze the difference between the current state and the desired state;

[0008] 3. Call etcd cluster management API or Kubernetes API to eliminate these differences.

[0009] However, to write a Kubernetes Operator, it is necessary to precipitate the operation and maintenance experience into code to achieve code-based, automated, and intelligent operation and maintenance. In other words, the operation and maintenance personnel need to have rich operation and maintenance experience to build a Kubernetes Operator for an application. These operation and maintenance experiences are the experience of the operation and maintenance personnel, but no one is perfect. The operation and maintenance experience of each operation and maintenance personnel is limited, and the more operation and maintenance experience, the better the Kubernetes Operator can be formed. Therefore, the method of writing an application's Kubernetes Operator based on the experience of the operation and maintenance personnel is inefficient, which also affects the performance or accuracy of the constructed Kubernetes Operator. Summary of the invention

[0010] The embodiment of the present invention provides a method for generating an operation and maintenance component of a stateful application instance to solve the technical problems of low efficiency and poor performance or accuracy of the constructed Kubernetes Operator in the prior art. The method includes:

[0011] Obtain learning data, including the writing syntax of Kubernetes Operator, the principle of application state, and operation and maintenance data of stateful application instances;

[0012] The learning data is analyzed by using a deep learning algorithm to obtain feature data of the Kubernetes Operator object, and a software package of the Kubernetes Operator is generated according to the feature data, wherein the feature data includes abstract features of full life cycle activities of the Kubernetes Operator object.

[0013] The embodiment of the present invention also provides a method for generating an operation and maintenance component of a stateful application instance to solve the technical problems of low efficiency and poor performance or accuracy of the constructed Kubernetes Operator in the prior art. The device includes:

[0014] A data acquisition module is used to acquire learning data, wherein the learning data includes the writing syntax of Kubernetes Operator, the principle of application status, and operation and maintenance data of stateful application instances;

[0015] A component generation module is used to analyze the learning data using a deep learning algorithm to obtain feature data of the Kubernetes Operator object, and generate a software package of the Kubernetes Operator based on the feature data, wherein the feature data includes abstract features of the full life cycle activities of the Kubernetes Operator object.

[0016] An embodiment of the present invention also provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements a method for generating an operation and maintenance component of any of the above-mentioned stateful application instances when executing the computer program, so as to solve the technical problems in the prior art of low efficiency in constructing a Kubernetes Operator and poor performance or accuracy of the constructed Kubernetes Operator.

[0017] An embodiment of the present invention further provides a computer-readable storage medium storing a computer program for executing the above-mentioned method for generating an operation and maintenance component of any stateful application instance, so as to solve the technical problems existing in the prior art of low efficiency in constructing a Kubernetes Operator and poor performance or accuracy of the constructed Kubernetes Operator.

[0018] In an embodiment of the present invention, it is proposed to analyze learning data including the writing syntax of Kubernetes Operator, the principle of application state, and the operation and maintenance data of stateful application instances through a deep learning algorithm to obtain feature data including abstract features of the full life cycle activities of the Kubernetes Operator object, and then generate a software package of the Kubernetes Operator based on the feature data, that is, complete the generation of the operation and maintenance component Kubernetes Operator of the stateful application instance. The method of automatically constructing the Kubernetes Operator by analyzing a large amount of learning data through a deep learning algorithm is realized. Compared with the method in the prior art where operation and maintenance personnel write Kubernetes Operator based on operation and maintenance experience, manual writing of Kubernetes Operator can be avoided, which is conducive to improving efficiency. In addition, by analyzing and learning a large amount of learning data, the influence of gaps or limitations of personal experience on the Kubernetes Operator can be avoided, which is conducive to improving the performance or accuracy of the constructed Kubernetes Operator. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] The drawings described herein are used to provide a further understanding of the present invention, constitute a part of the present application, and do not constitute a limitation of the present invention. In the drawings:

[0020] Figure 1 It is a schematic diagram of Kubernetes in the prior art;

[0021] Figure 2 This is a schematic diagram of a Kubernetes Operator in the prior art;

[0022] Figure 3 It is a flow chart of a method for generating an operation and maintenance component of a stateful application instance provided by an embodiment of the present invention;

[0023] Figure 4 is a model schematic diagram of an unsupervised learning algorithm provided by an embodiment of the present invention;

[0024] Figure 5 is a schematic diagram of a Gaussian mixture model provided by an embodiment of the present invention;

[0025] Figure 6 is a structural block diagram of a device provided by an embodiment of the present invention;

[0026] Figure 7 It is a structural block diagram of a device for generating an operation and maintenance component of a stateful application instance provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0027] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the embodiments and the accompanying drawings. Here, the illustrative embodiments of the present invention and their descriptions are used to explain the present invention, but are not intended to limit the present invention.

[0028] In an embodiment of the present invention, a method for generating an operation and maintenance component of a stateful application instance is provided, such as Figure 1 As shown, the method includes:

[0029] Step 302: Acquire learning data, where the learning data includes the writing syntax of the Kubernetes Operator, the principle of application state, and operation and maintenance data of stateful application instances;

[0030] Step 304: Analyze the learning data using a deep learning algorithm to obtain feature data of the Kubernetes Operator object, and generate a software package for the Kubernetes Operator based on the feature data, wherein the feature data includes abstract features of the full life cycle activities of the Kubernetes Operator object.

[0031] Depend on Figure 3 As can be seen from the process shown, in the embodiment of the present invention, it is proposed to analyze learning data including the writing syntax of Kubernetes Operator, the principle of application state, and the operation and maintenance data of stateful application instances through machine learning and deep learning algorithms to obtain feature data including abstract features of the full life cycle activities of Kubernetes Operator objects, and then generate a software package of Kubernetes Operator based on the feature data, that is, complete the generation of the operation and maintenance component Kubernetes Operator of the stateful application instance. It is realized that the Kubernetes Operator is automatically constructed by analyzing a large amount of learning data through machine learning and deep learning algorithms. Compared with the method in the prior art where operation and maintenance personnel write Kubernetes Operator based on operation and maintenance experience, manual writing of Kubernetes Operator can be avoided, which is conducive to improving efficiency. In addition, by analyzing and learning a large amount of learning data, the influence of personal experience gaps or limitations on Kubernetes Operator can be avoided, which is conducive to improving the performance or accuracy of the constructed Kubernetes Operator.

[0032] During the specific implementation, the inventor of the present application learned that machine learning is a branch of artificial intelligence. The research history of artificial intelligence has a natural and clear thread from focusing on "reasoning", to focusing on "knowledge", and then to focusing on "learning". Obviously, machine learning is a way to realize artificial intelligence, that is, to solve problems in artificial intelligence by means of machine learning. Machine learning has developed into a multi-disciplinary cross-disciplinary subject in the past 30 years, involving probability theory, statistics, approximation theory, convex analysis, computational complexity theory and other disciplines. The theory of machine learning mainly designs and analyzes some algorithms that allow computers to "learn" automatically. Machine learning algorithms are a type of algorithm that automatically analyzes and obtains rules from data and uses rules to predict unknown data. Because a large number of statistical theories are involved in learning algorithms, machine learning is particularly closely related to inferential statistics, and is also called statistical learning theory. In terms of algorithm design, machine learning theory focuses on achievable and effective learning algorithms. Many inference problems are difficult to follow without procedures, so part of machine learning research is to develop easy-to-handle approximate algorithms.

[0033] Deep learning is an algorithm in machine learning that is based on learning to represent data. Observations (such as an image) can be represented in a variety of ways, such as a vector of intensity values ​​for each pixel, or more abstractly represented as a series of edges, regions of a specific shape, etc. Using certain specific representation methods makes it easier to learn tasks from examples (for example, face recognition or facial expression recognition). The benefit of deep learning is that it replaces manual feature acquisition with efficient algorithms for unsupervised or semi-supervised feature learning and hierarchical feature extraction.

[0034] The goal of representation learning is to find better representations and create better models to learn these representations from large-scale unlabeled data. Representation learning comes from neuroscience and is loosely based on information processing and understanding of communication patterns in similar nervous systems, such as neural coding, which attempts to define the relationship between the response of pulling neurons and the relationship between the electrical activity of neurons in the brain.

[0035] To date, several deep learning frameworks, such as deep neural networks, convolutional neural networks, deep belief networks, and recurrent neural networks have been applied in computer vision, speech recognition, natural language processing, audio recognition, and bioinformatics, and have achieved excellent results.

[0036] In specific implementation, the inventor of the present application proposed to analyze the learning data through a deep learning algorithm to obtain the characteristic data of the Kubernetes Operator object. Specifically, it was proposed to analyze the learning data through an unsupervised learning algorithm to obtain the characteristic data of the Kubernetes Operator object.

[0037] Specifically, Figure 4 As shown in the figure, the unsupervised learning algorithm autoencoder uses the back-propagation principle to make the model output target value equal to the operation and maintenance data input value, where x represents the input data, for example, it can be the above-mentioned learning data; Encoder g represents the encoding function, that is, the encoding function is g(·); z represents the abstract feature after the input feature is encoded by encoder g; Decoder f represents the decoding function, that is, the decoding function is f(·); x′ represents the output, x′ is the feature after the abstract feature z is decoded by Decoder f, and the optimization objective function is f,g=argmin f,g L(x,f(g(x))) is used to calculate the minimum value between the predicted value and the actual value. Since AutoEncoder is based on DNN and has an activation function, it can perform nonlinear transformations to reduce the dimension of the valuable data in the above learning data through nonlinear transformations, obtain valuable data, and then obtain the feature data of the Kubernetes Operator object. The feature data of the Kubernetes Operator object can be the abstract features of the full life cycle activities of the Kubernetes Operator object.

[0038] In specific implementation, in order to realize that the exception handling mechanism can be automatically executed when the Kubernetes Operator runs abnormally, in this embodiment, in the process of obtaining the feature data of the Kubernetes Operator object, the feature data of the Kubernetes Operator object obtained by analyzing the learning data using the unsupervised learning algorithm can also be input into a Gaussian mixture model, and the Gaussian mixture model outputs the probability of each feature in the feature data; then, the abstract feature indicating the abnormal operation of the Kubernetes Operator is determined according to the size of the probability; and then the exception handling abstract feature corresponding to the abstract feature indicating the abnormal operation of the Kubernetes Operator is determined from the feature data. Taking the KubernetesOperator object as an example of a stateful application instance, the Kubernetes Operator constructed based on the feature data can operate and maintain the activities of the entire life cycle of the stateful application instance, and even if a certain operation of the Kubernetes Operator is abnormal, the exception handling mechanism can be automatically executed.

[0039] In specific implementation, the full life cycle activities of a Kubernetes Operator object may include creation, scaling, log processing, monitoring indicators, and exception handling.

[0040] In the specific implementation, all feature data of the Kubernetes Operator object output by the unsupervised learning algorithm model AutoEncoder is used as the input of the GMM (Gaussian Mixture-based Model) model, such as Figure 5 As shown, Figure 5 The original unclustered data in the figure on the left is the original sample without labeling. Figure 5 The Clustereddata in the figure on the right is the clustered sample, that is, the output probability of the Gaussian mixture model. The probability can determine whether a certain operation of the Kubernetes Operator is abnormal. For example, if one of the output probabilities is distributed in the preset area, it can be considered that the Kubernetes Operator is running abnormally, and then the abnormality handling mechanism can be automatically executed. Among them, GMM, as a common clustering algorithm, can be used to separate the foreground and background in the scene, or called background subtraction. So what is background subtraction? In the monitoring system, static cameras are usually used to capture the scene, so one of the more challenging steps is how to detect sudden intruders in the scene. Traditional applications will assume that there are no such intruders in the scene, but in the actual monitoring scene, this situation actually exists, so we cannot ignore it. For the detection of intruders, if we know the statistical model of a scene, those target pixels that do not obey the distribution of the model can be regarded as the background. Such a process is called background subtraction.

[0041] Specifically, the principle of the GMM Gaussian probability model is as follows:

[0042] The probability distribution model of the Gaussian probability model is as follows:

[0043]

[0044] Among them, α k is the coefficient, α k ≥0, φ(y|θ k ) is the Gaussian distribution density, is called the kth sub-model, y is the probability of Gaussian distribution, μ k is the expectation of k Gaussian distributions, σ k is the standard deviation of k Gaussian distributions, θ k is the expectation, variance, and probability of each sub-model occurring in the mixed model.

[0045] In this embodiment, a computer device is provided, such as Figure 6As shown, it includes a memory 602, a processor 604, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the method for generating the operation and maintenance component of any of the above-mentioned stateful application instances is implemented.

[0046] Specifically, the computer device may be a computer terminal, a server or a similar computing device.

[0047] In this embodiment, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores a computer program for executing the method for generating an operation and maintenance component of any of the above-mentioned stateful application instances.

[0048] Specifically, computer-readable storage media include permanent and non-permanent, removable and non-removable media that can be used to store information by any method or technology. Information can be computer-readable instructions, data structures, program modules or other data. Examples of computer-readable storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer-readable storage media does not include temporary computer-readable media (transitory media), such as modulated data signals and carrier waves.

[0049] Based on the same inventive concept, an embodiment of the present invention also provides a device for generating an operation and maintenance component of a stateful application instance, as described in the following embodiments. Since the principle of solving the problem by the device for generating the operation and maintenance component of a stateful application instance is similar to the method for generating the operation and maintenance component of a stateful application instance, the implementation of the device for generating the operation and maintenance component of a stateful application instance can refer to the implementation of the method for generating the operation and maintenance component of a stateful application instance, and the repeated parts will not be repeated. As used below, the term "unit" or "module" can be a combination of software and / or hardware that implements a predetermined function. Although the device described in the following embodiments is preferably implemented in software, the implementation of hardware, or a combination of software and hardware, is also possible and conceived.

[0050] Figure 7 is a structural block diagram of a device for generating an operation and maintenance component of a stateful application instance according to an embodiment of the present invention, such as Figure 7 As shown, the device comprises:

[0051] A data acquisition module 702 is used to acquire learning data, where the learning data includes the writing syntax of the Kubernetes Operator, the principle of application status, and operation and maintenance data of stateful application instances;

[0052] The component generation module 704 is used to analyze the learning data using a deep learning algorithm to obtain feature data of the Kubernetes Operator object, and generate a software package of the Kubernetes Operator based on the feature data, wherein the feature data includes abstract features of the full life cycle activities of the Kubernetes Operator object.

[0053] In one embodiment, the component generation module includes:

[0054] The learning unit is used to analyze the learning data by using an unsupervised learning algorithm to obtain feature data of the KubernetesOperator object.

[0055] In one embodiment, the learning unit is further used to input feature data of the Kubernetes Operator object obtained by analyzing the learning data using the unsupervised learning algorithm into a Gaussian probability model, and the Gaussian probability model outputs the probability of each feature in the feature data; determine an abstract feature indicating abnormal operation of the Kubernetes Operator according to the size of the probability; and determine an exception handling abstract feature corresponding to the abstract feature indicating abnormal operation of the Kubernetes Operator from the feature data.

[0056] In one embodiment, the full life cycle activities of the Kubernetes Operator object include creation, scaling, log processing, monitoring indicators, and exception handling.

[0057] The embodiment of the present invention achieves the following technical effects: it proposes to analyze learning data including the writing syntax of Kubernetes Operator, the principle of application state, and the operation and maintenance data of stateful application instances through a deep learning algorithm to obtain feature data including abstract features of the full life cycle activities of the Kubernetes Operator object, and then generate a software package of the Kubernetes Operator based on the feature data, that is, complete the generation of the operation and maintenance component Kubernetes Operator of the stateful application instance. It realizes the automatic construction of Kubernetes Operator by analyzing a large amount of learning data through a deep learning algorithm. Compared with the method in the prior art where operation and maintenance personnel write Kubernetes Operator based on operation and maintenance experience, manual writing of Kubernetes Operator can be avoided, which is conducive to improving efficiency. In addition, by analyzing and learning a large amount of learning data, the influence of personal experience gaps or limitations on Kubernetes Operator can be avoided, which is conducive to improving the performance or accuracy of the constructed Kubernetes Operator.

[0058] Obviously, those skilled in the art should understand that the modules or steps of the above-mentioned embodiments of the present invention can be implemented by a general computing device, they can be concentrated on a single computing device, or distributed on a network composed of multiple computing devices, and optionally, they can be implemented by a program code executable by a computing device, so that they can be stored in a storage device and executed by the computing device, and in some cases, the steps shown or described can be executed in a different order from that here, or they can be made into individual integrated circuit modules, or multiple modules or steps therein can be made into a single integrated circuit module for implementation. In this way, the embodiments of the present invention are not limited to any specific combination of hardware and software.

[0059] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the embodiments of the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for generating an operation and maintenance component of a stateful application instance, characterized in that: include: Obtain learning data, including the writing syntax of Kubernetes Operator, the principle of application state, and operation and maintenance data of stateful application instances; Analyze the learning data using a deep learning algorithm to obtain feature data of the Kubernetes Operator object, and generate a software package of the Kubernetes Operator according to the feature data, wherein the feature data includes abstract features of full life cycle activities of the Kubernetes Operator object; The learning data is analyzed using a deep learning algorithm to obtain feature data of the Kubernetes Operator object, including: An unsupervised learning algorithm is used to analyze the learning data to obtain characteristic data of the Kubernetes Operator object. The learning data is analyzed using an unsupervised learning algorithm to obtain feature data of the Kubernetes Operator object, including: Inputting feature data of the Kubernetes Operator object obtained by analyzing the learning data using the unsupervised learning algorithm into a Gaussian mixture model, and the Gaussian mixture model outputs the probability of each feature in the feature data; The unsupervised learning algorithm uses the back propagation principle, and the model output target value is equal to the operation and maintenance data input value. The input data is encoded to obtain abstract features, and the output data is obtained through the decoding function. The optimization function is f,g = argmin f,g L(x,f(g(x))); where x represents input data, which is learning data; g represents encoding function; f represents decoding function, which is an optimization function used to calculate the minimum value between the predicted value and the actual value; through nonlinear transformation, the valuable data in the learning data is reduced in dimension through nonlinear transformation to obtain valuable data and obtain the feature data of the Kubernetes Operator object; For the Gaussian mixture model, the probability distribution relationship is: Among them, α k is the coefficient, α k ≥0, φ(y|θ k ) is the Gaussian distribution density, is the kth sub-model, y is the probability of Gaussian distribution, μ k is the expectation of k Gaussian distributions, σ k is the standard deviation of k Gaussian distributions, θ k is the expectation, variance, and probability of each sub-model occurring in the mixed model; Determine the abstract features that indicate abnormal operation of the Kubernetes Operator based on the probability. An exception handling abstract feature corresponding to an abstract feature indicating an abnormal operation of the Kubernetes Operator is determined from the feature data.

2. The method for generating an operation and maintenance component of a stateful application instance according to claim 1, characterized in that: The full life cycle activities of the Kubernetes Operator object include creation, scaling, log processing, monitoring indicators, and exception handling.

3. A device for generating an operation and maintenance component of a stateful application instance, characterized in that: include: A data acquisition module is used to acquire learning data, wherein the learning data includes the writing syntax of Kubernetes Operator, the principle of application status, and operation and maintenance data of stateful application instances; A component generation module, configured to analyze the learning data using a deep learning algorithm to obtain feature data of a Kubernetes Operator object, and generate a software package of the Kubernetes Operator based on the feature data, wherein the feature data includes abstract features of full life cycle activities of the Kubernetes Operator object; The component generation module includes: A learning unit, configured to analyze the learning data using an unsupervised learning algorithm to obtain feature data of a Kubernetes Operator object; The learning unit is further used to input feature data of the Kubernetes Operator object obtained by analyzing the learning data using the unsupervised learning algorithm into a Gaussian probability model, and the Gaussian probability model outputs the probability of each feature in the feature data; determine an abstract feature indicating abnormal operation of the Kubernetes Operator according to the magnitude of the probability; and determine an exception handling abstract feature corresponding to the abstract feature indicating abnormal operation of the Kubernetes Operator from the feature data; The unsupervised learning algorithm uses the back propagation principle, and the model output target value is equal to the operation and maintenance data input value. The input data is encoded to obtain abstract features, and the output data is obtained through the decoding function. The optimization function is f,g = argmin f,g L(x,f(g(x))); where x represents input data, which is learning data; g represents encoding function; f represents decoding function, which is an optimization function used to calculate the minimum value between the predicted value and the actual value; through nonlinear transformation, the valuable data in the learning data is reduced in dimension through nonlinear transformation to obtain valuable data and obtain the feature data of the Kubernetes Operator object; For the Gaussian mixture model, the probability distribution relationship is: Among them, α k is the coefficient, α k ≥0, φ(y|θ k ) is the Gaussian distribution density, is the kth sub-model, y is the probability of Gaussian distribution, μ k is the expectation of k Gaussian distributions, σ k is the standard deviation of k Gaussian distributions, θ k is the expectation, variance, and probability of each sub-model occurring in the mixed model.

4. The device for generating an operation and maintenance component of a stateful application instance according to claim 3, characterized in that: The full life cycle activities of the Kubernetes Operator object include creation, scaling, log processing, monitoring indicators, and exception handling.

5. A computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the method for generating an operation and maintenance component of a stateful application instance according to any one of claims 1 to 2 is implemented.

6. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program for executing the method for generating an operation and maintenance component of a stateful application instance according to any one of claims 1 to 2.

Images