Operation and maintenance threshold regulation and control method, device and equipment in cloud native environment, medium and program product

Abstract

The invention provides an operation and maintenance threshold regulation and control method, device and equipment in a cloud native environment, a medium and a program product, and relates to the technical field of intelligent operation and maintenance. The method comprises the following steps: processing a resource index and a service index through a resource channel and a service channel respectively, and fusing features obtained after processing based on a dynamic gate control mechanism to obtain a feature vector; generating a dynamic service dependency relationship graph based on the association relationship among the services; the association relationship comprises a direct calling relationship between services and a cross-cluster multi-hop association relationship; performing state transition analysis according to the service dependency graph and the feature vector, and determining the current state of the system; based on the current state and resource information of the system, the initial solution space is optimized from the three dimensions of performance, cost and stability, and a target regulation and control strategy is generated; and generating a cloud resource regulation and control instruction based on the regulation and control strategy, wherein the cloud resource regulation and control instruction is used for carrying out operation and maintenance threshold regulation and control.

Description

Technical Field

[0001] This disclosure relates to the field of cloud-native intelligent operation and maintenance technology, specifically to a method, device, equipment, medium, and program product for adjusting operation and maintenance thresholds in a cloud-native environment. Background Technology

[0002] With the rapid development of cloud computing technology, cloud-native environments have become the mainstream approach for modern application deployment. Compared with traditional IT architectures, cloud-native environments are characterized by high dynamism, elastic scaling, and microservices.

[0003] Traditional operation and maintenance systems have the following technical shortcomings in cloud-native environments: Insufficient dynamic adaptability to the environment: The elastic scheduling characteristics of container resources (such as dynamic scaling, rolling updates, and service migration) cause traditional fixed threshold monitoring models to frequently fail; Lack of service topology awareness: Under microservice architecture, the call chain changes dynamically and the topology is complex, significantly increasing the difficulty of cross-service indicator correlation analysis; Insufficient consideration of cost constraints: Lack of quantitative constraints on cloud resource costs; Coarse-grained state identification: Insufficient ability to model the entire lifecycle of cloud-native unique states such as elastic scaling and rolling updates. Summary of the Invention

[0004] This disclosure provides a method, apparatus, equipment, media, and program product for adjusting operation and maintenance thresholds in a cloud-native environment.

[0005] According to the first aspect of this disclosure, a method for adjusting operation and maintenance thresholds in a cloud-native environment is provided, comprising: processing resource indicators and service indicators through resource channels and service channels respectively, and fusing the processed features based on a dynamic gate control mechanism to obtain a feature vector; the feature vector is used to characterize the current operating state of the system; generating a dynamic service dependency graph based on the association relationships between various services; the association relationships include direct call relationships between services and multi-hop association relationships across clusters; performing state transition analysis based on the service dependency graph and the feature vector to determine the current state of the system; optimizing the initial solution space from three dimensions—performance, cost, and stability—based on the current state of the system and resource information to generate a target adjustment strategy; the initial solution space contains multiple potential adjustment strategies; and generating cloud resource adjustment instructions based on the adjustment strategies, wherein the cloud resource adjustment instructions are used for adjusting operation and maintenance thresholds.

[0006] According to embodiments of this disclosure, resource indicators and service indicators are processed through resource channels and service channels respectively, and the processed features are fused based on a dynamic gating mechanism to obtain a feature vector. This includes: processing resource indicators with time decay factors based on resource channels to obtain resource features, which include the dynamic change trend of resources; processing service indicators based on gated loop networks and multi-head attention mechanisms based on service channels to obtain service features, where service features include dependencies between services; calculating weight allocation based on the current system state; and fusing the weight information of the resource features and service features to obtain a feature vector.

[0007] According to embodiments of this disclosure, generating a dynamic service dependency graph based on the relationships between various services includes: determining the call relationships and dependency strengths between services based on real-time collected operational data of each service; wherein, services include services in the same service mesh and services in different clusters; determining the position and connection method of each service in the topology graph based on the call relationships and dependency strengths, and constructing an initial dependency graph; representing the data in the initial dependency graph as a sparse matrix and calculating the sparse matrix based on an attention calculation model to obtain a calculation result, the calculation result containing quantitative information on the service association strength; and adjusting the initial dependency graph based on the calculation result of the sparse matrix to obtain a service dependency graph.

[0008] According to embodiments of this disclosure, state transition analysis is performed based on service dependency graphs and feature vectors to determine the current state of the system, including: dividing different states based on historical operating data, historical maintenance information, and technical documents, and setting transition conditions between different states; matching feature vectors with transition conditions, and updating the current state of the system when the feature vectors meet the transition conditions.

[0009] According to embodiments of this disclosure, based on the current system state and resource information, the initial solution space is optimized from three dimensions: performance, cost, and stability, to generate a target control strategy. This includes: obtaining current resource information in real time from a cloud service provider's API and optimizing the initial solution space based on the resource information to narrow its range; the initial solution space includes various resource allocation and configuration schemes; optimizing the filtered initial solution space based on multi-dimensional objectives to obtain an intermediate solution space; and introducing random perturbations into the intermediate solution space based on an annealing perturbation algorithm to obtain the target control strategy.

[0010] According to embodiments of this disclosure, optimizing the initial solution space after screening based on multi-dimensional objectives to obtain an intermediate solution space includes: determining multiple objective functions and assigning corresponding weights to different objective functions based on the current state of the system; based on the initial solution space after screening, finding a resource allocation and configuration scheme that satisfies multiple objective functions by adjusting the position and velocity of particles in the solution space, and determining the resource allocation and configuration scheme as the intermediate solution space, wherein the particles are potential solutions in the solution space.

[0011] According to a second aspect of this disclosure, an operation and maintenance threshold control device in a cloud-native environment is provided, comprising: a processing module, used to process resource indicators and service indicators through resource channels and service channels respectively, and to fuse the processed features based on a dynamic gate control mechanism to obtain a feature vector; the feature vector is used to characterize the current operating state of the system; a first generation module, used to generate a dynamic service dependency graph based on the association relationships between various services; the association relationships include direct call relationships between services and multi-hop association relationships across clusters; an analysis module, used to perform state transition analysis based on the service dependency graph and the feature vector to determine the current state of the system; a second generation module, used to optimize the initial solution space from three dimensions—performance, cost, and stability—based on the current state of the system and resource information to generate a target control strategy; the initial solution space contains multiple potential control strategies; and a third generation module, used to generate cloud resource control instructions based on the control strategy, the cloud resource control instructions being used for operation and maintenance threshold control.

[0012] According to a third aspect of this disclosure, an electronic device is provided, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the methods described above.

[0013] According to a fourth aspect of this disclosure, a computer-readable storage medium is provided that stores a computer program or instructions thereon, characterized in that the computer program or instructions, when executed by a processor, implement the steps of the method described above.

[0014] According to a fifth aspect of this disclosure, a computer program product is provided, comprising a computer program or instructions that, when executed by a processor, implement the steps of the method described above. Attached Figure Description

[0015] The above and other objects, features, and advantages of this disclosure will become clearer from the following description of embodiments of the present disclosure with reference to the accompanying drawings, in which:

[0016] Figure 1This diagram illustrates an application scenario of the operation and maintenance threshold control method in a cloud-native environment according to embodiments of the present disclosure.

[0017] Figure 2 A flowchart illustrating an operation and maintenance threshold control method in a cloud-native environment according to an embodiment of the present disclosure is shown schematically.

[0018] Figure 3 The diagram illustrates the principle of processing resource indicators and service indicators through resource channels and service channels respectively according to an embodiment of the present disclosure, and fusing the processed features based on a dynamic gate control mechanism to obtain a feature vector.

[0019] Figure 4 This illustration schematically shows a flowchart of generating a dynamic service dependency graph based on the association relationships between various services according to an embodiment of the present disclosure;

[0020] Figure 5 The schematic diagram illustrates the principle of different system state transitions according to embodiments of the present disclosure;

[0021] Figure 6 This schematically illustrates a flowchart of how, according to an embodiment of the present disclosure, the initial solution space is optimized from three dimensions—performance, cost, and stability—to generate a target control strategy based on the current system state and resource information.

[0022] Figure 7 This schematically illustrates a structural block diagram of an operation and maintenance threshold control device in a cloud-native environment according to an embodiment of the present disclosure;

[0023] Figure 8 A block diagram of an electronic device for an operation and maintenance threshold control method in a cloud-native environment according to an embodiment of the present disclosure is shown schematically. Detailed Implementation

[0024] To make the objectives, technical solutions, and advantages of this disclosure clearer, the following detailed description is provided in conjunction with specific embodiments and accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this disclosure. All other embodiments obtained by those skilled in the art based on the embodiments of this disclosure without inventive effort are within the scope of protection of this disclosure.

[0025] The terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the scope of this disclosure. The terms “comprising,” “including,” etc., as used herein indicate the presence of a feature, step, operation, and / or component, but do not exclude the presence or addition of one or more other features, steps, operations, or components.

[0026] In this disclosure, unless otherwise expressly specified and limited, the terms "installation," "connection," "linking," "fixing," etc., should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral part; they can refer to a mechanical connection, an electrical connection, or a connection that allows communication between them; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication of two components or the interaction between two components. Those skilled in the art can understand the specific meaning of the above terms in this disclosure according to the specific circumstances.

[0027] In the description of this disclosure, it should be understood that the terms "longitudinal", "length", "circumferential", "front", "rear", "left", "right", "top", "bottom", "inner", "outer", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings, and are only for the convenience of describing this disclosure and simplifying the description, and do not indicate or imply that the subsystem or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of this disclosure.

[0028] Throughout the accompanying drawings, identical elements are represented by the same or similar reference numerals. Conventional structures or configurations have been omitted where they may cause confusion in understanding this disclosure. Furthermore, the shapes, dimensions, and positional relationships of the components in the drawings do not reflect actual size, scale, or actual positional relationships. Additionally, any reference symbols enclosed in parentheses should not be construed as limiting.

[0029] Similarly, to simplify this disclosure and aid in understanding one or more of the various aspects of the disclosure, in the above description of exemplary embodiments of the present disclosure, various features of the present disclosure are sometimes grouped together in a single embodiment, figure, or description thereof. The use of terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., refers to a specific feature, structure, material, or characteristic described in connection with that embodiment or example, which is included in at least one embodiment or example of the present disclosure. In this specification, illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0030] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this disclosure, "multiple" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0031] In the technical solution disclosed herein, the collection, storage, use, processing, transmission, provision, disclosure, and application of user personal information all comply with relevant laws and regulations, necessary confidentiality measures have been taken, and they do not violate public order and good morals. In the technical solution disclosed herein, user authorization or consent has been obtained before acquiring or collecting user personal information.

[0032] The embodiments of this disclosure provide a method for adjusting operation and maintenance thresholds in a cloud-native environment, including: processing resource indicators and service indicators through resource channels and service channels respectively, and fusing the processed features based on a dynamic gate control mechanism to obtain a feature vector; the feature vector is used to characterize the current operating state of the system; generating a dynamic service dependency graph based on the association relationships between various services; the association relationships include direct call relationships between services and multi-hop association relationships across clusters; performing state transition analysis based on the service dependency graph and feature vectors to determine the current state of the system; optimizing the initial solution space from three dimensions—performance, cost, and stability—based on the current state of the system and resource information to generate a target adjustment strategy; the initial solution space contains multiple potential adjustment strategies; and generating cloud resource adjustment instructions based on the adjustment strategies, which are used for adjusting operation and maintenance thresholds.

[0033] Figure 1 The diagram illustrates an application scenario of the operation and maintenance threshold control method in a cloud-native environment according to an embodiment of the present disclosure.

[0034] like Figure 1 As shown, application scenario 100 according to this embodiment may include terminal devices 101, 102, and 103, network 104, and server 105. Network 104 is used as a medium to provide a communication link between terminal devices 101, 102, and 103 and server 105. Network 104 may include various connection types, such as wired or wireless communication links or fiber optic cables, etc.

[0035] Users can use terminal devices 101, 102, and 103 to interact with server 105 via network 104 to receive or send messages, etc. Various communication client applications can be installed on terminal devices 101, 102, and 103, such as physics simulation applications, shopping applications, web browser applications, search applications, instant messaging tools, email clients, social media platform software, etc. (for example only).

[0036] Terminal devices 101, 102, and 103 can be various electronic devices with displays and web browsing capabilities, including but not limited to smartphones, tablets, laptops, and desktop computers.

[0037] Server 105 can be a server that provides various services, such as a backend management server that supports websites browsed by users using terminal devices 101, 102, and 103 (for example only). The backend management server can analyze and process data such as received user requests, and feed back the processing results (such as web pages, information, or data obtained or generated according to user requests) to the terminal devices.

[0038] It should be noted that the operation and maintenance threshold control method in a cloud-native environment provided in this disclosure embodiment can generally be executed by server 105. Correspondingly, the operation and maintenance threshold control device in a cloud-native environment provided in this disclosure embodiment can generally be set in server 105. The operation and maintenance threshold control method in a cloud-native environment provided in this disclosure embodiment can also be executed by a server or server cluster that is different from server 105 and can communicate with terminal devices 101, 102, 103 and / or server 105. Correspondingly, the operation and maintenance threshold control device in a cloud-native environment provided in this disclosure embodiment can also be set in a server or server cluster that is different from server 105 and can communicate with terminal devices 101, 102, 103 and / or server 105.

[0039] It should be understood that Figure 1 The number of terminal devices, networks, and servers shown is merely illustrative. Depending on implementation needs, any number of terminal devices, networks, and servers can be included.

[0040] The following will be based on Figure 1 The described scene, through Figures 2-6 The present disclosure provides a detailed description of the operation and maintenance threshold control method in a cloud-native environment according to embodiments of the present disclosure.

[0041] Figure 2 A flowchart illustrating an operation and maintenance threshold control method in a cloud-native environment according to an embodiment of the present disclosure is shown.

[0042] like Figure 2 As shown, the operation and maintenance threshold control method in the cloud-native environment of this embodiment includes operations S210 to S250.

[0043] In operation S210, resource indicators and service indicators are processed through resource channels and service channels respectively, and the processed features are fused based on the dynamic gate control mechanism to obtain feature vectors; feature vectors are used to characterize the current operating state of the system.

[0044] In some embodiments, resource metrics and service metrics are processed through resource channels and service channels respectively to obtain service characteristics and resource characteristics. Resource metrics may include hardware resource-related data such as CPU utilization, memory usage, and disk I / O; service metrics may include service operation status data such as service response time, throughput, and error rate.

[0045] The processed resource and service characteristics can be fused using a dynamic gating mechanism to obtain a feature vector that comprehensively represents the current operating state of the system. The dynamic gating mechanism can dynamically adjust the weights of different features during the fusion process based on the real-time operating conditions of the system. For example, when the system is under high load, the weights of resource indicators such as CPU utilization and memory usage, which reflect system load characteristics, are increased; when a service failure occurs, the weights of service indicators such as error rate, which reflect service anomalies, are increased.

[0046] When operating S220, a dynamic service dependency graph is generated based on the relationships between various services; the relationships include direct call relationships between services and multi-hop relationships across clusters.

[0047] In some embodiments, the relationships between services can be determined based on their respective operational data. These relationships can include direct call relationships and multi-hop relationships across clusters. Direct call relationships can be obtained through service call logs, API call records, etc., while multi-hop relationships across clusters can be determined by analyzing information such as network communication and data flow between services. A dynamic service dependency graph can be generated based on the collected relationships. This graph uses services as nodes and the relationships between services as edges, with the direction of the edges indicating the direction of the call.

[0048] During operation S230, state transition analysis is performed based on the service dependency graph and feature vectors to determine the current state of the system.

[0049] In some embodiments, state transition analysis can be performed based on service dependency graphs and feature vectors. By calculating the transition probabilities between different states, the most likely current state of the system can be determined. For example, the system states may include: steady-state operation state, elastic scaling state, rolling update state, overload recovery state, and fault propagation state.

[0050] When operating S240, based on the current system state and resource information, the initial solution space is optimized from three dimensions: performance, cost, and stability, to generate a target control strategy; the initial solution space contains a variety of potential control strategies.

[0051] In some embodiments, the initial solution space can be optimized from three dimensions—performance, cost, and stability—based on the current system state and resource information. The initial solution space includes various potential adjustment strategies, such as increasing or decreasing resource allocation, adjusting service deployment locations, and optimizing service call parameters. In terms of performance, the resource allocation ratio can be adjusted according to the system's current processing capacity and business needs to improve system processing speed and throughput. For example, if a service's response time is found to be too long, the CPU and memory resources of the node hosting that service can be increased. In terms of cost, the cost of resource usage is considered, and resource usage is reduced as much as possible while meeting performance and stability requirements. For example, lower-cost cloud service instance types can be selected, or the ratio of resource reservation to on-demand usage can be reasonably adjusted. In terms of stability, the current system failure risks are analyzed, and corresponding measures are taken to improve system stability. For example, service redundancy can be increased, and critical services can be backed up and fault-tolerant.

[0052] The optimal solution that satisfies the requirements of performance, cost, and stability in the initial solution space is searched using multi-objective optimization algorithms (such as particle swarm optimization, genetic algorithm, etc.) to generate the target control strategy.

[0053] When operating the S250, cloud resource control instructions are generated based on the control strategy. These instructions are used to control operation and maintenance thresholds.

[0054] In some embodiments, cloud resource control instructions are generated based on target control strategies. These instructions are used to control operational thresholds, such as adjusting thresholds for CPU utilization, memory usage, and service response time. The instructions can be encapsulated in a specific format (e.g., JSON, XML) and include information such as the object being controlled (e.g., a specific service or resource instance), the control operation (e.g., adding, reducing, or modifying), and the control parameters (e.g., the specific value of the threshold). The generated cloud resource control instructions can be sent to the corresponding cloud resource management platform or operation and maintenance system to achieve real-time control of cloud resources.

[0055] This embodiment of the disclosure, by processing resource and service indicators separately and performing feature fusion based on a dynamic gating mechanism, can comprehensively and accurately characterize the current operating state of the system. The dynamic gating mechanism can dynamically adjust feature weights according to the real-time situation of the system, making the feature vector more reflective of the system's key state information and providing a reliable basis for subsequent decision-making. Optimizing the initial solution space from three dimensions—performance, cost, and stability—the generated target control strategy can minimize resource usage costs and improve resource utilization efficiency while meeting system performance and stability requirements, effectively enhancing the accuracy and flexibility of operation and maintenance threshold control.

[0056] Figure 3The diagram illustrates the principle of processing resource indicators and service indicators through resource channels and service channels respectively according to an embodiment of the present disclosure, and fusing the processed features based on a dynamic gate control mechanism to obtain a feature vector.

[0057] like Figure 3 As shown, this embodiment processes resource indicators and service indicators through resource channels and service channels respectively, and fuses the processed features based on a dynamic gate control mechanism to obtain a feature vector. The feature vector includes: processing resource indicators with time decay factors based on resource channels to obtain resource features, which include the dynamic change trend of resources; processing service indicators based on gated loop networks and multi-head attention mechanisms based on service channels to obtain service features, which include the dependencies between services; calculating weight allocation according to the current system state, and fusing the weight information of resource features and service features to obtain a feature vector.

[0058] In some embodiments, the resource channel can employ a Long Short-Term Memory (LSTM) network to process time-series resource indicators with a time decay factor, extract resource features, and capture the dynamic characteristics of the resource indicators. The role of the time decay factor is to make the model place greater emphasis on recent data. By introducing the time decay factor, data from different time points are weighted according to certain rules during the calculation process, giving recent data a greater weight in feature extraction, thereby better capturing the dynamic changes of resource indicators.

[0059] In some embodiments, the service channel combines a Gated Recurrent Unit (GRU) network and a multi-head attention mechanism to extract key features of service metrics to analyze complex dependencies between services. For example, GRUs can be used to process service metrics to obtain preliminary service features that reflect basic trends in service metrics, such as fluctuations in service response time and changes in error rate. After obtaining the preliminary service features, a multi-head attention mechanism is used to focus on different parts of the input data (i.e., the preliminary service features) from different perspectives. By mapping the input data to multiple subspaces and independently calculating attention weights in each space, the attention results from these subspaces are concatenated and linearly transformed to obtain the final attention representation. For example, eight attention heads can be used in parallel to analyze service dependencies across different dimensions, thereby providing a more comprehensive understanding of complex interactions between services. The number of attention heads can be dynamically adjusted according to the cluster size.

[0060] After obtaining resource and service characteristics, a dynamic gating mechanism calculates weight allocation based on the current system state, dynamically adjusting the weights of resource and service characteristics to better adapt to the dynamic changes in the cloud-native environment, achieving adaptive fusion and avoiding model lag issues caused by static weights. The linear combination result can be mapped to the [0,1] interval using the Sigmoid function, achieving adaptive weighted fusion of resource and service characteristics. For example, when the system is in a state of elastic scaling, the weight of resource characteristics can be increased to better reflect the impact of resource changes on the system.

[0061] Figure 4 The flowchart illustrating the generation of a dynamic service dependency graph based on the association between various services is shown in the embodiment of the present disclosure.

[0062] like Figure 4 As shown, the generation of a dynamic service dependency graph based on the association between various services in this embodiment includes operations S410 to S440.

[0063] When operating S410, the call relationships and dependency strength between services are determined based on the real-time collected runtime data of each service; among them, services include services in the same service mesh and services in different clusters.

[0064] When operating S420, the location and connection method of each service in the topology graph are determined based on the call relationship and dependency strength, and an initial dependency graph is constructed.

[0065] In operation S430, the data in the initial dependency graph is represented as a sparse matrix and the sparse matrix is ​​calculated based on the attention calculation model to obtain the calculation result, which contains quantitative information on the service association strength.

[0066] In operation S440, the initial dependency graph is adjusted based on the calculation results of the sparse matrix to obtain the service dependency graph.

[0067] In some embodiments, a real-time topology association engine can construct service dependencies based on input feature vectors, which contain information about resources, services, and other aspects during service operation. For example, by collecting data during service operation (such as call logs, resource usage data, and business metric data), the direct call relationships and dependency strength between services within the same service mesh can be obtained. Furthermore, by collecting relevant information from different clusters (such as service lists and network configurations within the cluster), the call relationships and dependency strength between services in different clusters can be obtained.

[0068] Based on the obtained call relationships and dependency strengths, the position and connection method of each service in the topology graph are determined, and an initial dependency graph is constructed. In the service topology, the call relationships between most services are relatively sparse, that is, only a few services have direct or indirect call relationships. By representing this with a sparse matrix, the amount of data storage and computation can be reduced.

[0069] The data in the initial dependency graph can be represented as a sparse matrix, and lightweight computational algorithms (such as attention-based computation models) can be used to operate on the sparse matrix to reduce the computational complexity from O(n²) to O(n log n), thereby reducing computational load and improving computational efficiency. Based on the calculation results of the sparse matrix, the initially constructed dependency graph is simplified to remove some redundant edges and nodes, resulting in a service dependency graph.

[0070] This disclosure improves the comprehensiveness and accuracy of service dependency analysis by acquiring services from the same service mesh and different clusters, comprehensively analyzing the call relationships and dependency strength between services, and calculating the sparse matrix based on an attention model to accurately quantify the association strength between services, thereby improving the accuracy of the service dependency graph and enabling more reasonable resource allocation and performance optimization.

[0071] According to one embodiment of this disclosure, state transition analysis is performed based on service dependency graphs and feature vectors to determine the current state of the system, including: dividing different states according to historical operation data, historical maintenance information and technical documents and setting transition conditions between different states; matching feature vectors with transition conditions, and updating the current state of the system when the feature vectors meet the transition conditions.

[0072] Figure 5 The schematic diagram illustrates the principle of different system state transitions according to embodiments of the present disclosure.

[0073] like Figure 5 As shown, based on historical operational data, historical maintenance information, and technical documentation, the system state can be divided into steady-state operation, elastic scaling, rolling update, overload recovery, and fault propagation states, and the transition conditions and rules between different states can be defined. For example, when the system load continues to increase and exceeds a certain threshold, it transitions from the steady-state operation state to the elastic scaling state; when the elastic scaling operation is completed and the system load returns to normal, it transitions back from the elastic scaling state to the steady-state operation state.

[0074] By matching real-time acquired feature vectors with transition conditions and determining whether a system state transition has occurred according to defined state transition rules, the current system state is updated if the feature vectors meet the transition conditions. Analyzing state transitions allows for the prediction of future system state change trends, and the optimization of operational strategies and resource allocation based on the analysis results can improve the timeliness of operational threshold adjustments.

[0075] Continue to combine Figure 5 Furthermore, service dependency graphs can be used for anomaly propagation prediction. For example, when a system anomaly occurs (such as service failure or resource overload), the possible propagation paths can be analyzed based on the service dependency graph to predict anomaly propagation. For instance, if a core service fails, the dependency graph can predict which services it depends on and which functions will be affected. By predicting anomaly propagation, measures can be taken in advance to prevent the spread of failures, reduce the impact of failures on the system, and optimize the system architecture and dependencies, thereby improving the system's fault tolerance and stability.

[0076] Figure 6 The illustration shows a flowchart of how, according to an embodiment of the present disclosure, the initial solution space is optimized from three dimensions—performance, cost, and stability—to generate a target control strategy based on the current system state and resource information.

[0077] like Figure 6 As shown, this embodiment optimizes the initial solution space from three dimensions—performance, cost, and stability—based on the current system state and resource information, and generates a target control strategy including operations S610 to S630.

[0078] When operating the S610, the current resource information is obtained in real time from the cloud service provider's API, and the initial solution space is optimized based on the resource information to narrow the range of the initial solution space, which includes various resource allocation and configuration schemes.

[0079] In some embodiments, current resource information can be obtained from the cloud service provider's API. This current resource information may include resource unit price data, which reflects the current market prices of various cloud resources and serves as the basis for cost optimization.

[0080] Based on the acquired resource information, the initial solution space is optimized. The initial solution space contains various possible resource allocation and configuration schemes. Through cost-aware initialization, some unreasonable schemes with significantly high costs can be eliminated, thereby narrowing the search range and improving the subsequent optimization process.

[0081] When operating the S620, the initial solution space after screening is optimized based on multi-dimensional objectives to obtain an intermediate solution space.

[0082] In some embodiments, the initial solution space can be optimized by determining the objective function, adjusting the elastic weights, and executing the PSO algorithm. For example, multiple objective functions are determined and corresponding weights are assigned to different objective functions based on the current state of the system; based on the filtered initial solution space, resource allocation and configuration schemes that satisfy multiple objective functions are found by adjusting the positions and velocities of particles in the solution space, and these resource allocation and configuration schemes are determined as the intermediate solution space, where particles are potential solutions in the solution space.

[0083] In some embodiments, the objective function may include, for example, a performance objective function, a cost objective function, and a stability objective function. The performance objective function measures system performance, the cost objective function calculates resource consumption costs, and the stability function calculates system jitter metrics.

[0084] The flexible weight adjustment mechanism can dynamically adjust the cost constraint weights according to the business cycle to adapt to different needs during peak and off-peak periods. For example, during peak periods, the cost weight can be appropriately reduced to prioritize system performance and stability in order to meet business needs.

[0085] Based on the initial solution space after filtering, the PSO algorithm is used to search for the optimal solution in the solution space. The PSO algorithm continuously adjusts the position and velocity of particles by simulating the collective behavior of flocks of birds or schools of fish in order to find the optimal resource allocation and configuration scheme (i.e., intermediate solution space) that satisfies multiple objective functions.

[0086] When operating the S630, random perturbations are introduced into the intermediate solution space based on the annealing perturbation algorithm to obtain the target control strategy.

[0087] In some embodiments, after obtaining the intermediate solution space based on the PSO algorithm, random perturbations are introduced through simulated annealing to help the algorithm escape the predicament of local optima and improve its global optimization capabilities. After annealing perturbation optimization, the final optimized solution (i.e., the target control strategy) is obtained, which represents the optimal resource allocation and configuration scheme under the current business scenario and system state. For example, in an IoT scenario involving 20TB of data processing, through annealing perturbation optimization, the system can reduce the data processing latency from 1200ms using traditional methods to 350ms, a latency reduction of 71%.

[0088] The following examples, including specific application scenarios, illustrate how... Figure 2 The technical effects of the operation and maintenance threshold control method shown in the cloud-native environment will be further explained.

[0089] Taking the implementation of high concurrency + cross-cluster calls in the performance control scenario of e-commerce microservice system as an example, the number of microservices in this scenario is 200+, the daily average request volume is 200 million, and during the promotion period, it is elastically scaled up to 5000+ container instances. The service level agreement (SLA) requires: response time <500ms and error rate <0.1%.

[0090] The cloud-native environment operation and maintenance threshold control method provided in this disclosure collects container CPU / memory metrics, service channel processing API response time, and error rate. A microservice call topology is constructed, and the dependency relationship between payment and inventory services is identified. When the system enters an "elastic scaling state," a target control strategy is generated: daily response time thresholds [100ms, 400ms], dynamically adjusted to [200ms, 500ms] during peak sales periods. The comparison results between the e-commerce system controlled according to this disclosure and traditional methods are shown in the table below.

[0091] Table 1 Comparison of E-commerce System Control Indicators

[0092]

[0093] Table 2 Comparison of E-commerce System Response Time

[0094]

[0095] Taking the implementation of edge computing and real-time requirements in the industrial equipment monitoring scenario under a cloud-native environment as an example, this scenario is based on a Kubernetes-managed equipment monitoring cluster, with 1000+ industrial equipment monitored, a data acquisition frequency of 5 seconds / time, and SLA requirements: fault prediction accuracy >95% and response time <2 seconds.

[0096] This disclosure provides a cloud-native environment operation and maintenance threshold control method for equipment monitoring and fault prediction. It involves collecting device sensor data (e.g., deploying a collection service on edge nodes via an Operator); using Flink to process streaming data and aggregate metrics with a 1-second window; constructing device group topology relationships and identifying critical path devices; and automatically adjusting monitoring thresholds based on load, with the threshold range expanded by 20% during peak hours.

[0097] Table 3 Comparison of Industrial Equipment Monitoring Indicators

[0098]

[0099] Table 4 Comparison of Fault Prediction Accuracy for Various Equipment

[0100]

[0101] Taking the implementation of massive devices + streaming processing in the IoT data processing scenario under the cloud-native environment as an example, this scenario is based on the Kubernetes IoT data processing platform, with a daily data processing volume of 20TB, a number of devices of 50,000+, and SLA requirements: data processing latency <500ms, system availability >99.95%.

[0102] Data processing is performed based on the cloud-native environment operation and maintenance threshold control method provided in this disclosure. An Operator is used to manage edge data acquisition services, streaming data processing is implemented using Flink, the window size is dynamically adjusted, the processing threshold is automatically adjusted according to the data volume and system load, and the data processing Pod is automatically scaled based on Prometheus metrics.

[0103] Table 5 Comparison of Performance Indicators

[0104]

[0105] Table 6 Comparison of Data Processing Latency

[0106]

[0107] Table 7. Cost vs. SLA Trade-offs

[0108]

[0109] As can be seen from the above embodiments, the operation and maintenance threshold control method in the cloud-native environment provided in this disclosure has significant advantages in different cloud-native scenarios, including improvements in alarm accuracy, response speed, resource cost optimization, and service stability.

[0110] This disclosure employs a dual-mode, adaptive approach to perceive container scheduling and service topology changes in the cloud-native environment in real time, improving the stability and reliability of operational threshold control. It also flexibly adjusts operational thresholds by monitoring service resource usage and performance metrics in real time, ensuring smooth service operation under different conditions. Furthermore, when determining target threshold strategies, it comprehensively considers multiple dimensions such as performance, cost, and stability, improving the comprehensiveness and accuracy of cloud resource control strategy determination. This method boasts a high level of intelligence and automation, automatically adapting to dynamic changes in the cloud-native environment. It exhibits strong versatility and scalability, adapting to various cloud-native scenarios without frequent manual intervention, thus improving operational and management efficiency. While enhancing operational efficiency and service quality, it significantly optimizes cloud resource costs, providing a strong guarantee for the stable and efficient operation of cloud-native applications.

[0111] Figure 7 The diagram illustrates a structural block diagram of an operation and maintenance threshold control device in a cloud-native environment according to an embodiment of the present disclosure.

[0112] like Figure 7 As shown, the cloud-native environment operation and maintenance threshold control device 700 of this embodiment includes a processing module 710, a first generation module 720, an analysis module 730, a second generation module 740 and a third generation module 750.

[0113] The processing module 710 processes resource indicators and service indicators through resource channels and service channels respectively, and fuses the processed features based on a dynamic gate control mechanism to obtain a feature vector; the feature vector is used to characterize the current operating state of the system. In one embodiment, the processing module 710 can be used to execute the operation S210 described above, which will not be repeated here.

[0114] The first generation module 720 is used to generate a dynamic service dependency graph based on the relationships between various services; the relationships include direct call relationships between services and multi-hop relationships across clusters. In one embodiment, the first generation module 720 can be used to perform the operation S220 described above, which will not be repeated here.

[0115] The analysis module 730 is used to perform state transition analysis based on the service dependency graph and feature vectors to determine the current state of the system. In one embodiment, the analysis module 730 can be used to perform the operation S230 described above, which will not be repeated here.

[0116] The second generation module 740 is used to optimize the initial solution space from three dimensions—performance, cost, and stability—based on the current system state and resource information, and generate a target control strategy; the initial solution space contains a variety of potential control strategies. In one embodiment, the second generation module 740 can be used to execute the operation S240 described above, which will not be repeated here.

[0117] The third generation module 750 is used to generate cloud resource control instructions based on the control strategy. These cloud resource control instructions are used to control operation and maintenance thresholds. In one embodiment, the third generation module 750 can be used to execute the operation S250 described above, which will not be repeated here.

[0118] According to embodiments of this disclosure, any plurality of modules among processing module 710, first generation module 720, analysis module 730, second generation module 740, and third generation module 750 may be combined into one module, or any one of these modules may be split into multiple modules. Alternatively, at least part of the functionality of one or more of these modules may be combined with at least part of the functionality of other modules and implemented in one module. According to embodiments of this disclosure, at least one of processing module 710, first generation module 720, analysis module 730, second generation module 740, and third generation module 750 may be at least partially implemented as hardware circuitry, such as a field-programmable gate array (FPGA), a programmable logic array (PLA), a system-on-a-chip, a system-on-a-substrate, a system-on-package, an application-specific integrated circuit (ASIC), or implemented in hardware or firmware by any other reasonable means of integrating or packaging circuitry, or implemented in any one of software, hardware, and firmware methods, or in a suitable combination of any of these methods. Alternatively, at least one of the processing module 710, the first generation module 720, the analysis module 730, the second generation module 740, and the third generation module 750 may be implemented at least partially as a computer program module, which can perform corresponding functions when the computer program module is run.

[0119] Figure 8 A block diagram of an electronic device for an operation and maintenance threshold control method in a cloud-native environment according to an embodiment of the present disclosure is shown schematically.

[0120] like Figure 8 As shown, an electronic device 800 according to an embodiment of this disclosure includes a processor 801, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 802 or a program loaded from a storage portion 808 into a random access memory (RAM) 803. The processor 801 may include, for example, a general-purpose microprocessor (e.g., a CPU), an instruction set processor and / or an associated chipset and / or a special-purpose microprocessor (e.g., an application-specific integrated circuit (ASIC)), etc. The processor 801 may also include onboard memory for caching purposes. The processor 801 may include a single processing unit or multiple processing units for performing different actions of the method flow according to an embodiment of this disclosure.

[0121] RAM 803 stores various programs and data required for the operation of electronic device 800. Processor 801, ROM 802, and RAM 803 are interconnected via bus 804. Processor 801 performs various operations of the method flow according to embodiments of the present disclosure by executing programs in ROM 802 and / or RAM 803. It should be noted that the program may also be stored in one or more memories other than ROM 802 and RAM 803. Processor 801 may also perform various operations of the method flow according to embodiments of the present disclosure by executing programs stored in one or more memories.

[0122] According to embodiments of this disclosure, the electronic device 800 may further include an input / output (I / O) interface 805, which is also connected to a bus 804. The electronic device 800 may also include one or more of the following components connected to the input / output (I / O) interface 805: an input section 806 including a keyboard, mouse, etc.; an output section 807 including a cathode ray tube (CRT), liquid crystal display (LCD), etc., and a speaker, etc.; a storage section 808 including a hard disk, etc.; and a communication section 809 including a network interface card such as a LAN card, modem, etc. The communication section 809 performs communication processing via a network such as the Internet. A drive 810 is also connected to the input / output (I / O) interface 805 as needed. A removable medium 811, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., is installed on the drive 810 as needed so that computer programs read from it can be installed into the storage section 808 as needed.

[0123] This disclosure also provides a computer-readable storage medium, which may be included in the device / apparatus / system described in the above embodiments; or it may exist independently and not assembled into the device / apparatus / system. The computer-readable storage medium carries one or more programs that, when executed, implement the method according to the embodiments of this disclosure.

[0124] According to embodiments of this disclosure, the computer-readable storage medium can be a non-volatile computer-readable storage medium. Examples include, but are not limited to: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this disclosure, the computer-readable storage medium can be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0125] The embodiments of this disclosure have been described above. However, these embodiments are for illustrative purposes only and are not intended to limit the scope of this disclosure. Although various embodiments have been described above, this does not mean that the measures in the various embodiments cannot be used advantageously in combination. Without departing from the scope of this disclosure, those skilled in the art can make various substitutions and modifications, all of which should be included within the protection scope of this disclosure.

Claims

1. A method for adjusting operation and maintenance thresholds in a cloud-native environment, characterized in that, The method includes: Resource indicators and service indicators are processed through resource channels and service channels respectively, and the processed features are fused based on a dynamic gate control mechanism to obtain a feature vector; the feature vector is used to characterize the current operating state of the system. A dynamic service dependency graph is generated based on the relationships between various services; the relationships include direct call relationships between services and multi-hop relationships across clusters. Based on the service dependency graph and the feature vector, state transition analysis is performed to determine the current state of the system; Based on the current system state and resource information, the initial solution space is optimized from three dimensions: performance, cost, and stability, to generate a target control strategy; the initial solution space contains a variety of potential control strategies. Cloud resource control instructions are generated based on the control strategy, and these instructions are used to control operation and maintenance thresholds.

2. The method according to claim 1, characterized in that, The resource indicators and service indicators are processed through resource channels and service channels respectively, and the processed features are fused based on a dynamic gate control mechanism to obtain a feature vector, including: Resource characteristics are obtained by processing resource indicators with time decay factors based on resource channels, and the resource characteristics include the dynamic change trend of resources. Service metrics are processed using a gate loop network and a multi-head attention mechanism based on service channels to obtain service features, wherein the service features include dependencies between services. The weight allocation is calculated based on the current system state, and the weight information of resource features and service features is fused to obtain the feature vector.

3. The method according to claim 1, characterized in that, The generation of a dynamic service dependency graph based on the relationships between various services includes: The call relationships and dependency strength between services are determined based on the real-time collected operational data of each service; wherein, the services include services in the same service mesh and services in different clusters; Based on the call relationships and dependency strength, the position and connection method of each service in the topology graph are determined, and an initial dependency graph is constructed. The data in the initial dependency graph is represented as a sparse matrix and the sparse matrix is ​​calculated based on the attention calculation model to obtain the calculation result, which includes quantitative information on the service association strength. The initial dependency graph is adjusted based on the calculation results of the sparse matrix to obtain a service dependency graph.

4. The method according to claim 1, characterized in that, The step of performing state transition analysis based on the service dependency graph and the feature vector to determine the current system state includes: Based on historical operational data, historical maintenance information, and technical documents, different states are defined, and transition conditions between different states are set. The feature vector is matched with the transition condition, and the current state of the system is updated if the feature vector satisfies the transition condition.

5. The method according to claim 1, characterized in that, Based on the current system state and resource information, the initial solution space is optimized from three dimensions: performance, cost, and stability, to generate a target control strategy, including: The system obtains current resource information in real time from the cloud service provider's API and optimizes the initial solution space based on the resource information to narrow the scope of the initial solution space, which includes various resource allocation and configuration schemes. The initial solution space after screening is optimized based on multi-dimensional objectives to obtain an intermediate solution space; The target control strategy is obtained by introducing random perturbations into the intermediate solution space based on the annealing perturbation algorithm.

6. The method according to claim 5, characterized in that, The optimization of the initial solution space after screening based on multi-dimensional objectives to obtain an intermediate solution space includes: Determine multiple objective functions and assign corresponding weights to different objective functions based on the current state of the system; Based on the initial solution space after screening, by adjusting the position and velocity of particles in the solution space, a resource allocation and configuration scheme that satisfies multiple objective functions is found. The resource allocation and configuration scheme is determined as the intermediate solution space, wherein the particles are potential solutions in the solution space.

7. A cloud-native environment operation and maintenance threshold control device, characterized in that, The device includes: The processing module is used to process resource indicators and service indicators through resource channels and service channels respectively, and to fuse the processed features based on a dynamic gate control mechanism to obtain a feature vector; the feature vector is used to characterize the current operating state of the system. The first generation module is used to generate a dynamic service dependency graph based on the relationships between various services; the relationships include direct call relationships between services and multi-hop relationships across clusters. The analysis module is used to perform state transition analysis based on the service dependency graph and the feature vector to determine the current state of the system. The second generation module is used to optimize the initial solution space from three dimensions—performance, cost, and stability—based on the current system state and resource information, and generate a target control strategy; the initial solution space contains a variety of potential control strategies. The third generation module is used to generate cloud resource control instructions based on the control strategy. The cloud resource control instructions are used to control operation and maintenance thresholds.

8. An electronic device, comprising: One or more processors; Memory, used to store one or more computer programs. The characteristic feature is that the one or more processors execute the one or more computer programs to implement the steps of the method according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program or instructions stored thereon, characterized in that, When the computer program or instructions are executed by a processor, they implement the steps of the method according to any one of claims 1 to 6.

10. A computer program product, comprising a computer program or instructions, characterized in that, When the computer program or instructions are executed by a processor, they implement the steps of the method according to any one of claims 1 to 6.

Images