Cloud platform health check method, device and computer program product

Abstract

Provided in the embodiments of the present disclosure are a cloud platform health check method, a device and a computer program product. The health check method comprises: acquiring current multi-dimensional indicators corresponding to a cloud platform in a current period; determining a combined statistical table corresponding to operation indicators of the cloud platform in a previous period, the combined statistical table at least comprising indicator statistical characteristics obtained by performing statistical processing on historical multi-dimensional indicators of the cloud platform in the previous period, wherein the indicator statistical characteristics at least comprise the average value, the standard deviation, and the number of statistical indicators; on the basis of the combined statistical table, determining a statistical model used for implementing health check operation; and, on the basis of the combined statistical table, the statistical model and the current multi-dimensional indicators, performing health check on the cloud platform, so as to obtain a health check result. The present embodiment can achieve health check operation without acquiring specific historical multi-dimensional indicators, thereby reducing the data volume required for the health check operation, and also improving the quality and efficiency of the check operation.

Description

[0001] Cloud Platform Health Monitoring Method, Device, and Computer Program Product This disclosure claims priority to Chinese patent application No. 202410382515.6, filed with the China Patent Office on March 29, 2024, entitled "Cloud Platform Health Monitoring Method, Device, and Computer Program Product," the entire contents of which are incorporated herein by reference. Technical Field This disclosure relates to the field of cloud platform technology, and more particularly to a cloud platform health monitoring method, device, and computer program product. Background: With the rapid development of cloud technology, cloud platforms are becoming increasingly widely used. Because cloud platforms can securely, economically, and efficiently store large amounts of data, many businesses and individuals rely on cloud platforms to remotely store and access data and provide computing power. Therefore, cloud platforms have become an essential tool for many businesses. When using cloud platforms for data storage and computing, the health of the cloud platforms is crucial. When an enterprise is unable to access its critical data due to cloud platform unavailability, it can result in financial losses, even leading to the collapse of data processing operations and loss of customer trust. Cloud platforms are typically large-scale distributed systems, generating large amounts of temporal multidimensional data. This data can include log data, trace data, and metrics data. For example, a cloud service consists of 1,000 clusters deployed in many different regions around the world, each consisting of 100 commodity machines. Each machine can detect over 10,000 dimensional objects, and data can be detected for 10 metrics every 10 seconds. Each metric requires 8 bytes to store. This generates approximately 70 terabytes (8.64 trillion rows) of temporal multidimensional data per day, accumulating to 260 trillion rows and approximately 2 petabytes (2 petabytes) of data within a month. oAs can be seen from the above, when a cloud platform health check is required, the large amount of metering data not only increases the difficulty of performing real-time and efficient health checks on the cloud service platform, but also reduces the quality and efficiency of health checks on the cloud platform. SUMMARY OF THE INVENTION Embodiments of the present disclosure provide a cloud platform health check method, device, and computer program product that not only reduce the amount of data required for cloud platform health checks but also improve the quality and efficiency of cloud platform health checks. In a first aspect, an embodiment of the present disclosure provides a health monitoring method for a cloud platform, comprising: obtaining current multi-dimensional indicators corresponding to the cloud platform in a current cycle; determining a spliced ​​statistical table corresponding to the operating indicators of the cloud platform in a previous cycle, the spliced ​​statistical table including at least indicator statistical features obtained by statistically processing historical multi-dimensional indicators of the cloud platform in the previous cycle, the indicator statistical features including at least an average value, a standard deviation, and a number of statistical indicators; determining a statistical model for implementing a health monitoring operation based on the spliced ​​statistical table; and performing a health monitoring on the cloud platform based on the spliced ​​statistical table, the statistical model, and the current multi-dimensional indicators to obtain a health monitoring result. In a second aspect, embodiments of the present disclosure provide a cloud platform health monitoring device, comprising: a first acquisition module for acquiring current multi-dimensional indicators corresponding to the cloud platform in a current cycle; a first determination module for determining a concatenated statistical table corresponding to the cloud platform's operating indicators in a previous cycle, wherein the concatenated statistical table includes at least indicator statistical features obtained by statistically processing historical multi-dimensional indicators of the cloud platform in the previous cycle, wherein the indicator statistical features include at least an average value, a standard deviation, and a number of statistical indicators; the first determination module for determining a statistical model for implementing a health monitoring operation based on the concatenated statistical table; and a first processing module for performing a health monitoring operation on the cloud platform based on the concatenated statistical table, the statistical model, and the current multi-dimensional indicators to obtain a health monitoring result. In a third aspect, embodiments of the present disclosure provide an electronic device, comprising: a memory and a processor; wherein the memory is configured to store one or more computer instructions, wherein when executed by the processor, the one or more computer instructions implement the cloud platform health monitoring method of the first aspect. In a fourth aspect, an embodiment of the present disclosure provides a computer storage medium for storing a computer program, wherein the computer program enables a computer to implement the cloud platform health detection method in the first aspect when executed.In a fifth aspect, embodiments of the present disclosure provide a computer program product, comprising: a computer program that, when executed by a processor of an electronic device, causes the processor to perform the steps of the cloud platform health monitoring method of the first aspect. The cloud platform health monitoring method, device, and computer program product provided in this embodiment obtain current multi-dimensional indicators corresponding to the cloud platform in the current cycle; determine a spliced ​​statistical table corresponding to the cloud platform's operating indicators in the previous cycle; and determine a statistical model for implementing a health monitoring operation based on the spliced ​​statistical table. Health monitoring operations are then performed on the cloud platform based on the spliced ​​statistical table, the statistical model, and the current multi-dimensional indicators, thereby quickly and efficiently obtaining health monitoring results. Because the spliced ​​statistical table stores statistical features of indicator data corresponding to different aggregation dimensions in previous historical cycles and the current cycle, re-statistics are performed on a large amount of original fine-grained data, significantly reducing the amount of data required to process health monitoring operations for the cloud platform, improving the quality and efficiency of health monitoring operations for the cloud platform, and further enhancing the practicality of the method. BRIEF DESCRIPTION OF THE DRAWINGS To more clearly illustrate the embodiments of the present disclosure or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present disclosure. Those skilled in the art can derive other drawings based on these drawings without inventive effort. Figure 1 is a schematic diagram of the principle of a health detection method for a cloud platform provided by an embodiment of the present disclosure; Figure 2 is a schematic diagram of the flow of a health detection method for a cloud platform provided by an embodiment of the present disclosure; Figure 3 is a schematic diagram of the flow of splicing multiple aggregate statistical tables to obtain a spliced ​​statistical table provided by an embodiment of the present disclosure; Figure 4 is a schematic diagram of the flow of updating the statistical model based on the spliced ​​statistical table and the current multi-dimensional indicators to obtain an updated statistical model provided by an embodiment of the present disclosure; Figure 5 is a schematic diagram of the first flow of performing a health detection on the cloud platform based on the updated statistical model to obtain a health detection result provided by an embodiment of the present disclosure; Figure 6 is a schematic diagram of the second flow of performing a health detection on the cloud platform based on the updated statistical model to obtain a health detection result provided by an embodiment of the present disclosure; Figure 7 is a schematic diagram of the flow of another health detection method for a cloud platform provided by an embodiment of the present disclosure; Figure 8 is a schematic diagram of the structure of a health detection device for a cloud platform provided by an embodiment of the present disclosure; Figure 9 is a schematic diagram of the structure of an electronic device corresponding to the health detection device for the cloud platform provided by the embodiment shown in Figure 8.To further clarify the objectives, technical solutions, and advantages of the embodiments of the present disclosure, the technical solutions of the embodiments of the present disclosure will be described clearly and completely below in conjunction with the accompanying drawings. It should be understood that the described embodiments represent only a portion of the embodiments of the present disclosure, and are not exhaustive. All other embodiments devised by persons of ordinary skill in the art based on the embodiments of the present disclosure without inventive effort are within the scope of protection of the present disclosure. The terms used in the embodiments of the present disclosure are intended solely to describe specific embodiments and are not intended to limit the present disclosure. The singular forms "a," "an," "the," and "the" used in the embodiments of the present disclosure and the appended claims are intended to include the plural forms, unless the context clearly indicates otherwise. "A plurality" generally includes at least two, but does not exclude the inclusion of at least one. It should be understood that the term "and / or" as used herein is merely a term used to describe an association between associated objects, indicating that three possible relationships exist. For example, "A and / or B" can represent: A alone, A and B simultaneously, or B alone. Furthermore, the character " / " generally indicates an "or" relationship between the associated objects. Depending on the context, the words "if" and "if" as used herein can be interpreted as "at the time of" or "when" or "in response to determining" or "in response to detecting." Similarly, depending on the context, the phrases "if it is determined" or "if (a stated condition or event) is detected" can be interpreted as "when it is determined" or "in response to determining" or "when (a stated condition or event) is detected" or "in response to detecting (a stated condition or event)." It should also be noted that the terms "comprise," "comprising," or any other variations thereof are intended to encompass a non-exclusive inclusion, such that a product or system comprising a list of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such product or system. Without further limitation, the elements defined by the phrase "comprising a..." do not preclude the presence of additional identical elements in the product or system comprising the elements. In addition, the sequence of steps in the following method embodiments is merely an example and not a strict limitation. Term Definitions: Telemetry: refers to the automated measurement and transmission of data from remote or inaccessible sources to a monitoring system in real time. It involves the collection and transmission of measurement, statistical, or diagnostic data from various devices or sensors to a centralized system for analysis, monitoring, and control.Observability engineering refers to the practice of designing, implementing, and managing highly observable systems. It involves ensuring that developers, operators, and other stakeholders have access to the necessary tools, data, and insights to effectively understand and address system behavior. Observability engineering goes beyond traditional instrumentation and logging approaches, emphasizing the need for holistic visibility into a system's internal state. It focuses on capturing relevant telemetry data, such as logs, metrics, traces, and events, and making it accessible. To facilitate understanding the specific implementation of the technical solutions in this embodiment, the following briefly describes the relevant technologies. In recent years, with the rapid development of cloud technology, the application of cloud platforms has become increasingly widespread. Because cloud platforms can securely, cost-effectively, and efficiently store large amounts of data, many businesses and individuals rely on cloud platforms for remote data storage and access, as well as for computing power. Consequently, cloud platforms have become a vital tool for many enterprises. The health of cloud platforms for data storage and computing is crucial. When an enterprise loses access to critical data due to cloud platform unavailability, it can result in financial losses, even disrupted data processing operations, and loss of customer trust. To monitor and maintain the health of cloud platforms, instrumentation and telemetry can be leveraged to collect data about their performance and health, transmitting this data to remote locations for analysis and monitoring. This allows engineers to gain insight into how cloud platforms operate in real time and make adjustments as needed. This approach is also known as observability engineering. In the practice of observability engineering, cloud platforms are typically large-scale distributed systems. Large-scale cloud service platforms typically generate large amounts of temporal, multi-dimensional data. This data can include log data, trace data, and metrics data. Log data, as described above, records events that occur in an application or system and is often used for troubleshooting and debugging. Trace data tracks the path of a request as it moves through an application or system. It is often used to identify and diagnose performance issues, identify bottlenecks, and optimize performance. Metrics provide information about the performance of an application or system over a period of time. Latency, network traffic, error counts, and system load (for example, request counts, CPU utilization, and memory usage) are the most common and useful metrics. After obtaining log data, trace data, and indicator data, they can be organized into time multidimensional data. The time multidimensional data can be viewed as a data table consisting of these three types of data fields. The data table can include the following types of data: (1) time period or timestamp, (2) discrete dimension columns indicating the context of events or statistical information, and (3) numerical measurements.Specifically, the aforementioned time period or timestamp may include: time period start time. Discrete dimension columns may include: user identifier, cluster, server, user storage space name, user data table name, user data partition name, software component identifier, request operation, request parameter string, response return code, total read data size, and average latency. Numerical metrics may include: time period length in seconds, total request count, and total write data size. Because cloud platforms are typically large-scale distributed systems, with the assistance of instrumentation and telemetry technologies, large-scale cloud service platforms typically generate large amounts of temporal multi-dimensional data. For example, if a cloud service consists of 1,000 clusters deployed in many different regions around the world, each cluster consisting of 100 commodity machines, each machine needs to detect over 10,000 dimensional objects, such as the aforementioned combinations of partitions, software components, and request operations. Then, statistical calculations are performed on these objects using 10 metrics every 10 seconds, with each metric requiring 8 bytes to store. In the form of time-dependent multidimensional data, a cloud platform generates approximately 70 terabytes (8.64 trillion rows) of data daily, accumulating to 2 petabytes (260 trillion rows) of data within a month. Such a large amount of metering data not only increases the difficulty of performing real-time and efficient health monitoring of the cloud service platform, but also reduces the quality and efficiency of health monitoring operations. To address the aforementioned technical issues, this embodiment provides a cloud platform health monitoring method, device, and computer program product. Referring to FIG. 1 , the cloud platform health monitoring method provided in this embodiment may be executed by a cloud platform health monitoring device 200. It should be noted that the cloud platform health monitoring device 200 can be implemented as any device capable of providing cloud platform health monitoring services, such as a terminal device, a personal computer, a tablet computer, a local server, or a cloud server. In this case, when the cloud platform health monitoring device 200 is implemented as a cloud server, the cloud platform health monitoring method can be executed in the cloud. Several computing nodes (cloud servers) may be deployed in the cloud, each of which has processing resources such as computing and storage. In the cloud, multiple computing nodes can be organized to provide a service. Of course, a single computing node can also provide one or more services. The cloud can provide this service by providing a service interface, which users call to access the corresponding service.Service interfaces include software development kits (SDKs) and application programming interfaces (APIs). The cloud platform health monitoring device 200 is communicatively connected to the cloud platform 100. The cloud platform 100 can provide cloud services for user applications and generate corresponding operating data based on the provided cloud services, enabling health monitoring operations on the cloud platform to be performed based on the operating data. In the above embodiment, the cloud platform 100 and the cloud platform health monitoring device 200 are connected via a network, which can be either wireless or wired. If the cloud platform 100 is communicatively connected to the cloud platform health monitoring device 200, the mobile network standard can be any of 2G (GSM), 2.5G (GPRS), 3G (WCDMA, TD-SCDMA, CDMA2000, UTMS), 4G (LTE), 4G+ (LTE+), Wi-Fi, 5G, 6G, etc. The cloud platform health monitoring device 200 is a device that can provide health monitoring operations for the cloud platform in a network virtual environment, typically a device that utilizes the network for information planning and health monitoring operations for the cloud platform. In physical implementation, the cloud platform health monitoring device 200 can be any device that can provide computing services and perform health monitoring operations for the cloud platform in response to health monitoring requests from the cloud platform. For example, it can be a cluster server, a conventional server, a cloud server, a cloud host, a virtual center, etc. The cloud platform health monitoring device 200 is primarily composed of a processor, a hard disk, memory, a system bus, etc., similar to a general computer architecture. In an embodiment of the present disclosure, the cloud platform 100 provides multiple cloud services for invocation by multiple users, and different cloud services can implement different data processing operations. When the cloud services of the cloud platform 100 are invoked by one or more users to perform corresponding data processing operations, the cloud platform 100 can generate operating indicators in multiple dimensions. In order to detect the operating status of the cloud platform 100, the operating indicators generated by the cloud platform 100 can be collected according to a preset period, so as to obtain the current multi-dimensional indicators corresponding to the cloud platform 100 in the current period. The obtained current multi-dimensional indicators can then be actively or passively sent to the health detection device 200 of the cloud platform, so that the health detection device 200 of the cloud platform can perform health detection operations on the cloud platform 100 based on the current multi-dimensional indicators.The cloud platform health monitoring device 200 is used to obtain current multi-dimensional indicators corresponding to the cloud platform 100 in the current cycle. The above-mentioned current multi-dimensional indicators may include at least one of the following: a time period or timestamp, a discrete dimension column for indicating the context of an event or statistical information, and a digital metric. In order to minimize the amount of data required for the health monitoring operation of the cloud platform, the health monitoring operation of the cloud platform is simplified. After obtaining the current multi-dimensional indicators corresponding to the cloud platform 100 in the current cycle, a spliced ​​statistical table corresponding to the operating indicators of the cloud platform 100 in the previous cycle can be determined. The spliced ​​statistical table includes at least: indicator statistical characteristics obtained by statistically processing the historical multi-dimensional indicators of the cloud platform 100 in the previous cycle. The indicator statistical characteristics include at least: an average value, a standard deviation, and the number of statistical indicators. After obtaining the spliced ​​statistical table, a statistical model for implementing health checks can be determined based on the spliced ​​statistical table. The obtained statistical model can determine the indicator variation patterns or indicator variation patterns of each dimensional indicator, allowing health checks on the cloud platform to be performed based on the indicator variation patterns, indicator variation patterns, and current multi-dimensional indicators. Specifically, health checks on the cloud platform can be performed based on the spliced ​​statistical table, the statistical model, and the current multi-dimensional indicators, thereby quickly and efficiently obtaining health check results. Because the spliced ​​statistical table includes at least the indicator statistical features obtained by statistically processing the cloud platform's historical multi-dimensional indicators in the previous cycle, there is no need to obtain the specific multi-dimensional indicators corresponding to the cloud platform in the previous operating cycle. This significantly reduces the amount of data required for health checks on the cloud platform 100, improves the quality and efficiency of health checks on the cloud platform 100, and further enhances the practicality of the method. The following detailed description of some embodiments of the present disclosure is provided in conjunction with the accompanying drawings. The following embodiments and features may be combined unless they conflict with each other. Furthermore, the sequence of steps in the following method embodiments is provided for illustrative purposes only and is not a strict limitation. FIG2 is a flow chart illustrating a cloud platform health monitoring method provided by an embodiment of the present disclosure. Referring to FIG2 , this embodiment provides a cloud platform health monitoring method. The method may be performed by a cloud platform health monitoring device. It is understood that the cloud platform health monitoring device may be implemented as software or a combination of software and hardware. Specifically, when the cloud platform health monitoring device is implemented as hardware, it includes, but is not limited to, a tablet computer, a personal computer (PC), a server, and the like. When the cloud platform health monitoring device is implemented as software, it may be installed in the electronic devices listed above.Based on the aforementioned cloud platform health monitoring device, the cloud platform health monitoring method in this embodiment may include the following steps: Step S201: Obtaining the current multi-dimensional indicators corresponding to the cloud platform in the current cycle. Step S202: Determining a concatenated statistical table corresponding to the cloud platform's operating indicators in the previous cycle. The concatenated statistical table includes at least indicator statistical features obtained by statistically processing the cloud platform's historical multi-dimensional indicators in the previous cycle. The indicator statistical features include at least the mean, standard deviation, and number of statistical indicators. Step S203: Determining a statistical model for implementing health monitoring operations based on the concatenated statistical table. Step S204: Performing a health monitoring of the cloud platform based on the concatenated statistical table, the statistical model, and the current multi-dimensional indicators to obtain health monitoring results. The specific implementation principles and effects of each of the above steps are described in detail below: Step S201: Obtaining the current multi-dimensional indicators corresponding to the cloud platform in the current cycle. The cloud platform may provide multiple cloud services for implementing different functions. Users can call corresponding cloud services as needed to implement preset functions. To ensure the operational stability of the cloud platform and a positive user experience with the cloud server, the cloud platform can be subjected to real-time or periodic health checks. This allows the cloud platform's health check device (hereinafter referred to as the "health check device") to obtain the cloud platform's current multi-dimensional indicators for the current period. In some instances, these current multi-dimensional indicators can be collected using a preset detection module or telemetry technology. Obtaining the cloud platform's current multi-dimensional indicators for the current period can include: obtaining a preset collection period (e.g., 10s / time, 15s / time, 20s / time, etc.); determining a detection module or telemetry module for performing health checks on the cloud platform. The detection module or telemetry module can be located on the cloud platform; or the detection module or telemetry module can be in communication with the cloud platform. Data can then be checked on the cloud platform based on the detection module or telemetry module to obtain the cloud platform's current multi-dimensional indicators for the current period. The current multi-dimensional metrics corresponding to the cloud platform in the current cycle can include at least one of the following: log data, trace data, and metric data. Log data is a record of events occurring in an application or system and is typically used for troubleshooting and debugging the application or system. Trace data is used to track the path of a request as it moves through the application or system. Trace data is typically used to identify and diagnose performance issues, identify bottlenecks, and optimize performance.Metric data provides performance information about an application or system over a specific time period. Latency, network traffic, error counts, and system load (such as request counts, CPU utilization, and memory utilization) are common and useful metrics. It is understood that the current multi-dimensional metrics may include not only the metric types listed above, but also other types. Those skilled in the art can flexibly adjust and configure the specific types of metrics included in the obtained multi-dimensional metrics based on specific application scenarios or requirements, as long as accurate health monitoring of the cloud platform is ensured. Step S202: Determine a concatenated statistical table corresponding to the cloud platform's operating metrics for the previous cycle. The concatenated statistical table includes at least the metric statistical features obtained by statistically processing the cloud platform's historical multi-dimensional metrics for the previous cycle. The metric statistical features include at least the mean, standard deviation, and number of statistical metrics. To accurately perform health checks on the cloud platform, it is necessary to obtain not only the multi-dimensional operating indicators corresponding to the cloud platform in the current cycle, but also the historical multi-dimensional operating indicators corresponding to the cloud platform in previous cycles. However, due to the enormous amount of data from the cloud platform's historical multi-dimensional operating indicators in previous cycles, directly obtaining all historical multi-dimensional operating indicators from the cloud platform in previous cycles and then combining them with the current multi-dimensional operating indicators to perform health checks on the cloud platform would not only reduce the efficiency of the health check operation, but also require a large number of resources due to the large amount of data. Therefore, to minimize the number of indicators required for cloud platform health checks and ensure the quality and efficiency of cloud platform health checks, the implementation method of this embodiment does not require obtaining the historical multi-dimensional operating indicators of the cloud platform in previous cycles. Instead, it is necessary to determine a concatenated statistical table corresponding to the cloud platform's operating indicators in the previous cycle. This concatenated statistical table includes at least indicator statistical features obtained by statistically processing the historical multi-dimensional indicators of the cloud platform in the previous cycle. The indicator statistical features include at least the mean value, standard deviation, and number of statistical indicators. It is understandable that the indicator statistical characteristics may include not only the above-mentioned average value, standard deviation and number of statistical indicators, but also other statistical characteristics, such as: indicator sum value, indicator maximum value, indicator minimum value, indicator record value, etc. Those skilled in the art may flexibly configure and adjust the indicator statistical characteristics according to specific application scenarios or application requirements, which will not be elaborated here.This embodiment does not limit the specific method for determining the splicing statistics table. In some instances, the cloud platform may have different splicing statistics tables corresponding to different operating cycles. The splicing statistics tables corresponding to each operating cycle may be tables configured using a preset algorithm and stored in a preset area. In this case, determining the splicing statistics table corresponding to the operating indicators of the cloud platform in the previous cycle may include: obtaining a storage area for storing the splicing statistics table. By accessing the storage area, the splicing statistics table corresponding to the operating indicators of the cloud platform in the previous cycle can be determined and obtained, thereby ensuring the accuracy and reliability of the determination of the splicing statistics table. In other instances, the splicing statistical table can be determined not only by accessing a preset region but also by analyzing and processing the multi-dimensional indicators of the cloud platform in the previous cycle. In this case, determining the splicing statistical table corresponding to the cloud platform's operating indicators in the previous cycle may include: obtaining historical multi-dimensional indicators corresponding to the cloud platform in the previous cycle; aggregating the historical multi-dimensional indicators to obtain multiple aggregated statistical tables; and splicing the multiple aggregated statistical tables to obtain a spliced ​​statistical table, wherein the spliced ​​statistical table also includes: preset splicing dimensions and splicing dimension values ​​corresponding to the preset splicing dimensions. Specifically, since the splicing statistical table includes at least indicator statistical features obtained by statistically processing the historical multi-dimensional indicators of the cloud platform in the previous cycle, in order to accurately determine the splicing statistical table, the historical multi-dimensional indicators corresponding to the cloud platform in the previous cycle can be first obtained. The specific method for obtaining the historical multi-dimensional indicators is similar to the specific method for obtaining the current multi-dimensional indicators in the above step and is not further described here. After acquiring historical multi-dimensional indicator data, to reduce data sparsity, the historical multi-dimensional indicators can be aggregated to obtain multiple aggregated statistical tables. In some instances, indicator aggregation can be achieved using a preset fixed dimension. In this case, aggregating the historical multi-dimensional indicators to obtain multiple aggregated statistical tables may include: obtaining a pre-configured target aggregation dimension for aggregating the historical multi-dimensional indicators. The target aggregation dimension may include at least one of the following: a user information dimension, a hierarchical entity dimension, a cloud service entity dimension, a software component dimension, a request dimension, a response dimension, etc. The target aggregation dimension for aggregating the historical multi-dimensional indicators can be determined based on the application scenario or application requirements. Once determined, the target aggregation dimension cannot be arbitrarily changed. Subsequently, the historical multi-dimensional indicators can be aggregated based on at least one aggregation dimension to obtain multiple aggregated statistical tables corresponding to the aggregation dimension. Different aggregation statistical tables may correspond to different aggregation dimensions.In other instances, metric aggregation operations can be implemented not only using fixed dimensions but also based on flexibly adjustable dimensions. In this case, aggregating historical multi-dimensional metrics to obtain multiple aggregated statistical tables may include: obtaining preset aggregation dimensions for implementing a single aggregation operation. The preset aggregation dimensions include at least: time period, return code, and one or two replaceable dimensions. The replaceable dimensions are obtained by traversing and replacing the remaining dimensions of the historical multi-dimensional metrics except for time period and return code; and aggregating the historical multi-dimensional metrics based on the preset aggregation dimensions to obtain multiple aggregated statistical tables. Specifically, to fully perform an aggregation operation based on each aggregation dimension, the preset aggregation dimensions for implementing a single aggregation operation may be first obtained. Since the return code is immediate context information that can indicate whether a request response contains an error, the return code dimension may be added to the preset aggregation dimensions to perform the aggregation operation. In this case, the preset aggregation dimensions include at least: time period, return code, and one or two replaceable dimensions. The aforementioned return codes may include: client return code, authorization return code, meta-process return code, data processing return code, etc. In addition, the replaceable dimensions included in the preset aggregation dimensions may change as the number of aggregation operations changes, and may be obtained by traversing and replacing the remaining dimensions in the historical multi-dimensional indicators except for the time period and the return code. For example 1, the historical multi-dimensional indicators obtained include: time period, region 1, region group 1, cluster 1, client return code, authorization return code, meta-process return code, and data processing return code. The above-mentioned "time period" and "user request return code" can be two aggregation dimensions used for aggregation operations. The other aggregation dimension can be obtained by traversing and replacing the "region, region group, cluster" dimensions. In this way, the following three preset aggregation dimensions can be obtained: preset aggregation dimension 1 is [time period, region, user request return code]; preset aggregation dimension 2 is [time period, region group, user request return code]; preset aggregation dimension 3 is [time period, cluster, user request return code]; As can be seen from the above, the "time period" and "return code" in the above-mentioned preset aggregation dimensions are fixed. The above-mentioned "replaceable dimension" will be traversed and replaced in any one of "region", "region group" and "cluster", so as to obtain different preset aggregation dimensions. After obtaining the above-mentioned preset aggregation dimensions, An aggregation operation may be performed on the historical multi-dimensional indicators based on each preset aggregation dimension, thereby obtaining multiple aggregation statistical tables corresponding to each preset aggregation dimension.In Example 2, the historical multi-dimensional indicators obtained include: time period, client ID, region, region group, cluster, and user request return code. The above-mentioned "time period" and "user request return code" can be two aggregation dimensions used for the aggregation operation. In addition, the two aggregation dimensions can be obtained by traversing and replacing the "client ID1, client ID2" and "region, region group, cluster" dimensions. In this way, the following six preset aggregation dimensions can be obtained: Preset aggregation dimension 1 is [time period, client ID, region, user request return code]; Preset aggregation dimension 2 is [time period, client ID, region group, user request return code]; Preset aggregation dimension 3 is [time period, client ID, cluster, user request return code]; Preset aggregation dimension 4 is [time period, region, region group, user request return code]; Preset aggregation dimension 5 is [time period, region, cluster, user request return code]; Preset aggregation dimension 6 is [time period, region group, cluster, user request return code]; As can be seen above, the "time period" and "return code" in the above-mentioned preset aggregation dimensions are fixed. The above-mentioned "replaceable dimensions" are obtained by traversing and replacing "client ID" and "region, region group, and cluster" in "client ID", "region", "region group", and "cluster", thereby obtaining different preset aggregation dimensions. After obtaining the above-mentioned preset aggregation dimensions, an aggregation operation can be performed on historical multi-dimensional indicators based on each preset aggregation dimension, thereby obtaining multiple aggregation statistical tables corresponding to each preset aggregation dimension. After the above-mentioned aggregation operation on the historical multi-dimensional indicators, multiple aggregate statistical tables can be obtained. Since the number of aggregate statistical tables is closely related to the number of aggregation dimensions and historical multi-dimensional indicators, in order to reduce the amount of data required for cloud platform health detection and simplify data processing operations, after obtaining multiple aggregate statistical tables, the multiple aggregate statistical tables can be spliced ​​to obtain a spliced ​​statistical table. The spliced ​​statistical table also includes: preset splicing dimensions and splicing dimension values ​​corresponding to the preset splicing dimensions. The preset splicing dimensions can be obtained by performing upper processing or rough statistics on replaceable dimensions in multiple aggregate statistical tables. For example, the preset splicing dimensions can be "client", "cluster", "region", "first splicing dimension", etc.In Example 3, when a splicing operation is performed on multiple aggregated statistical tables obtained based on the preset aggregation dimension 1 [time period, region, user request return code] in Example 1 above, the resulting spliced ​​statistical table may include the following dimensions: [time period, dimension name, dimension value, user request return code, indicator statistical characteristics]. When the indicator is a data processing request, the indicator statistical characteristics may include: request count, average request duration, minimum request duration, maximum request duration, standard deviation of request duration, etc.; when a splicing operation is performed on multiple aggregated statistical tables obtained based on the preset aggregation dimension 3 [time period, client ID, cluster, user request return code] in Example 2 above, the resulting spliced ​​statistical table may include the following dimensions: [time period, dimension 1 name, dimension value corresponding to dimension 1, dimension 2 name, dimension value corresponding to dimension 2, user request return code, indicator statistical characteristics]. When the indicator is a data processing request, the indicator statistical characteristics may include: request count, average request duration, minimum request duration, maximum request duration, standard deviation of request duration, etc. In some instances, the concatenation operation can be implemented by vertically stacking aggregate tables based on various preset dimension sets. In this case, multiple aggregate statistical tables are concatenated, with the dimension names from each aggregate statistical table entered into the dimension name fields in the data rows of the concatenated data table, and the dimension values ​​from each aggregate statistical table entered into the dimension value fields in the data rows of the concatenated data table. Because the concatenated statistical table includes not only the preset concatenated dimensions and their corresponding concatenated dimension values, but also the indicator statistical features obtained by statistically processing historical multi-dimensional indicators, the concatenated statistical table can be used to replace multiple aggregate statistical tables for cloud platform health monitoring. This reduces the number of aggregate tables, simplifies automated data processing, and improves the efficiency of cloud platform health monitoring operations. Step S203: Determine a statistical model for implementing health monitoring operations based on the concatenated statistical table. Since the spliced ​​statistical table includes the statistical characteristics of indicators obtained by statistically processing the historical multi-dimensional indicators of the cloud platform in the previous cycle, in order to accurately perform health monitoring operations on the cloud platform, after obtaining the spliced ​​statistical table, the spliced ​​statistical table can be analyzed and processed to determine the statistical model used to implement the health monitoring operation. This statistical model is used to identify the regularities or patterns of changes in each historical multi-dimensional indicator.In some instances, the statistical model can be obtained by analyzing and processing the spliced ​​statistical table using a preset algorithm formula. In this case, determining the statistical model for implementing the health check operation based on the spliced ​​statistical table may include: obtaining a preset algorithm formula for analyzing and processing the spliced ​​statistical table; and analyzing and processing the spliced ​​statistical table using the preset algorithm formula, thereby determining the statistical model for implementing the health check operation. In other instances, the statistical model can be determined not only by the preset algorithm formula but also based on a corresponding dimension combination in the spliced ​​statistical table. In this case, determining the statistical model for implementing the health check operation based on the spliced ​​statistical table may include: obtaining at least one parameter dimension combination in the spliced ​​statistical table, the parameter dimension combination including at least one parameter dimension and dimension values ​​corresponding to each parameter dimension; and determining the statistical model for implementing the health check operation based on the at least one parameter dimension combination. For example, for parameter dimension combination 1, when there is a corresponding dependent variable indicator y and a set of (pT) independent variable indicators %, the statistical model 1 determined by the parameter dimension combination 1 in the spliced ​​statistical table may be the following relationship: The model parameters of the statistical model are: %2-^pi is the independent variable composed of pT indicators collected by the cloud platform in different periods. For parameter dimension combination 2, when there is a corresponding dependent variable indicator / ■ and a group of (pT) independent variable indicators m, the statistical model 2 determined by splicing the parameter dimension combination 2 in the statistical table can be the following relationship: f = b 00 + Ru Dish + b 22m2+ •- +such as TpTTHpT, where / ■ is the dependent variable indicator, and the above-mentioned such as, such as, force 22". such as -Qiong-1 are the model parameters of the statistical model. TH], m2...TYlp^ are a set of pT independent variable indicators m collected by the cloud platform in different periods. The above example shows that the statistical models corresponding to different parameter dimension combinations in the spliced ​​statistical table have the same model structure, and the model parameters of the statistical models corresponding to different parameter dimension combinations are different. That is, the number of statistical models determined by a spliced ​​statistical table can be multiple, and the model parameters of the statistical models determined by different dimension value combinations are different, while the model structure can be the same. For example: {cluster dimension value: C1001, server dimension value: SVR-12345678, error return code value: 500}. Statistical model 1 can be calculated and determined based on the corresponding indicator data through the above different dimension value combinations. When the above dimensions take different values, other statistical models are generated based on other dimension value combinations corresponding to the above dimensions and the corresponding indicator data. The structures of statistical model 1 and other models can be identical. However, since the dimensional values ​​corresponding to each dimension are different, the corresponding indicator data is different, and the model parameters of the statistical model determined by the indicator data calculation are different. This enables the statistical model to accurately perform prediction and anomaly detection operations for different parameter-dimensional combinations. Each set of dimensional value combinations can be assigned customized model parameters, further improving the quality and effectiveness of health monitoring operations on the cloud platform. Step S204: Perform a health check on the cloud platform based on the concatenated statistical table, the statistical model, and the current multi-dimensional indicators to obtain a health check result. After obtaining the statistical model, a health check operation can be performed on the cloud platform based on the concatenated statistical table, the statistical model, and the current multi-dimensional indicators to obtain a health check result. In some instances, the health check operation can be implemented by analyzing and processing the current multi-dimensional indicators using the statistical model. In this case, performing a health check on the cloud platform based on the concatenated statistical table, the statistical model, and the current multi-dimensional indicators to obtain a health check result can include: analyzing and processing the current multi-dimensional indicators using the statistical model to obtain the current indicator test result; The cloud platform is tested for health based on the current indicator test results to obtain health test results, thereby effectively implementing the health test operation of the cloud platform.In other instances, not only can the health check operation of the cloud platform be implemented by analyzing and processing the current multi-dimensional indicators of the current period of the spliced ​​statistical table through the statistical model, but the statistical model can also be updated and then the updated statistical model can be used to perform a health check on the cloud platform. In this case, performing a health check on the cloud platform based on the spliced ​​statistical table, the statistical model, and the current multi-dimensional indicators to obtain a health check result may include: updating the statistical model based on the spliced ​​statistical table and the current multi-dimensional indicators to obtain an updated statistical model; and performing a health check on the cloud platform based on the updated statistical model to obtain a health check result. As the cloud platform system changes over time, the corresponding multi-dimensional time series indicators will also change, and the model relationships between the indicators may also change. Therefore, once the model parameters are calculated, they cannot be fixed. To ensure the accuracy of the cloud platform's health check, an online incremental algorithm can be used to update the statistical model to obtain an updated statistical model. Specifically, after obtaining the concatenated statistical table and the current multi-dimensional indicators for the current cycle, updating the statistical model is an iterative calculation process. The current multi-dimensional indicators for the current cycle can be used to update the statistical model updated in the previous cycle to obtain an updated statistical model. The updated statistical model for this cycle overwrites the previous cycle's statistical model, and the model for this cycle will be updated in the next cycle. After obtaining the updated statistical model, the cloud platform health check can be performed using the updated statistical model to obtain a health check result, which can include the first test result or the second test result. If the health check result is the first test result, it indicates that the cloud platform is operating normally; if the health check result is the second test result, it indicates that the cloud platform is operating abnormally. This allows for quick and stable completion of the cloud platform health check operation. Furthermore, after obtaining the cloud platform health check result, if the health check result is the second test result, which indicates abnormal cloud platform operation, the health check result can be saved in a preset dedicated data table for easy access and viewing by users on demand.The cloud platform health monitoring method provided in this embodiment obtains the current multi-dimensional indicators corresponding to the cloud platform in the current cycle; determines a concatenated statistical table corresponding to the cloud platform's operating indicators in the previous cycle; and uses the concatenated statistical table to determine a statistical model for implementing the health monitoring operation. The cloud platform health monitoring operation is then performed based on the concatenated statistical table, the statistical model, and the current multi-dimensional indicators, thereby quickly and efficiently obtaining health monitoring results. Because the concatenated statistical table stores the statistical characteristics of indicator data corresponding to different aggregation dimensions in previous historical cycles and the current cycle, there is no need to re-compile statistics on a large amount of original fine-grained data, thereby significantly reducing the amount of data required to process the cloud platform health monitoring operation. This improves the quality and efficiency of the cloud platform health monitoring operation, further enhancing the practicality of the method. FIG3 is a flow chart illustrating a process for splicing multiple aggregated statistical tables to obtain a spliced ​​statistical table, according to an embodiment of the present disclosure. Based on the above embodiment, and with reference to FIG3 , a splicing operation is provided by performing structural adjustment operations on multiple aggregated statistical tables based on replaceable dimensions in the multiple aggregated statistical tables. In this case, splicing multiple aggregated statistical tables to obtain the spliced ​​statistical table may include the following: Step S301: Obtaining replaceable dimensions from the multiple aggregated statistical tables. Since multiple aggregated statistical tables can be obtained by aggregating historical multi-dimensional indicators based on preset aggregation dimensions, different preset aggregation dimensions may correspond to different aggregated statistical tables. The preset aggregation dimensions may include preset fixed dimensions and replaceable dimensions for implementing the aggregation operation. The preset fixed dimensions may include a "time period" dimension and a "return code" dimension, while the replaceable dimension is used to implement the splicing operation. The "time period" dimension may be time information used to identify a time collection period. The "return code" dimension may include "4XX" or "5XX" values, with "4XX" or "5XX" used to indicate invalid user requests or other system errors. In order to ensure the quality and effect of table splicing processing, replaceable dimensions in multiple aggregate statistical tables can be obtained. The replaceable dimensions can be determined by whether the dimension values ​​of the preset dimensions in the multiple aggregate statistical tables are flexibly changed; for example, when the multiple aggregate statistical tables include aggregate statistical table 1, aggregate statistical table 2, and aggregate statistical table 3, and the "region" in aggregate statistical table 1 can include specific dimension values ​​such as "region 1", "region 2", and "region 3", Bei0 can determine that the "region" dimension in aggregate statistical table 1 can be a replaceable dimension."Client ID" in Aggregate Statistical Table 2 may include specific dimension values ​​such as "Client 1" and "Client 2," and "Cluster" in Aggregate Statistical Table 2 may include specific dimension values ​​such as "Cluster 1," "Cluster 2," and "Cluster 3." Therefore, the "Client" and "Cluster" dimensions in Aggregate Statistical Table 2 can be considered interchangeable dimensions in multiple aggregated statistical tables. If "Region" in Aggregate Statistical Table 3 includes specific dimension values ​​such as "Region Group 1" and "Region Group 2," then the "Region Group" dimension in Aggregate Statistical Table 3 can be determined to be an interchangeable dimension. Step S302: Adjust multiple aggregated statistical tables based on the interchangeable dimensions to obtain an adjusted statistical table with a unified structure. After obtaining the replaceable dimension, multiple aggregate statistical tables can be adjusted based on the replaceable dimension to obtain an adjusted statistical table with a unified structure. The adjustment operation can be performed by adding a dimension name and a dimension value for splicing the multiple aggregate statistical tables. At this time, adjusting the multiple aggregate statistical tables based on the replaceable dimension to obtain an adjusted statistical table with a unified structure may include: adding a dimension name for splicing the multiple aggregate statistical tables in the aggregate statistical table based on the replaceable dimension; determining the dimension value corresponding to the replaceable dimension based on the aggregate statistical table; and adjusting each aggregate statistical table based on the dimension name and dimension value to obtain an adjusted statistical table with a unified structure. Specifically, after obtaining the replaceable dimension, a dimension name for concatenating multiple aggregate statistical tables can be added to the aggregate statistical table based on the replaceable dimension. The dimension name can be obtained by performing upper-level processing on the replaceable dimension. After obtaining the dimension name, the dimension value corresponding to the replaceable dimension can be determined based on the aggregate statistical table. Then, based on the dimension name and dimension value, each aggregate statistical table can be adjusted to obtain an adjusted statistical table with a unified structure. This effectively ensures the accuracy and reliability of obtaining the adjusted statistical tables. For example, aggregate statistical table 1 and aggregate statistical table 2 can include the following dimensions: Dimensions of aggregate statistics table 1 By analyzing and processing the above-mentioned Aggregate Statistics Table 1 and Aggregate Statistics Table 2, it can be determined that one dimension in Aggregate Statistics Table 1 and Aggregate Statistics Table 2 is changing. The corresponding replaceable dimensions of Aggregate Statistics Table 1 and Aggregate Statistics Table 2 are "client ID" and "cluster," respectively. The structures of Aggregate Statistics Table 1 and Aggregate Statistics Table 2, that is, the fields of the two tables, are not completely consistent. Then, based on the above-mentioned replaceable dimensions, the dimension names of the replaceable dimensions used to splice the above-mentioned Aggregate Statistics Table 1 and Aggregate Statistics Table 2 can be determined. Dimension values ​​corresponding to the respective dimension names can also be determined based on the above-mentioned Aggregate Statistics Table 1 and Aggregate Statistics Table 2. By adjusting each aggregate table based on the dimension names and dimension values, a stable adjusted statistical table with a unified structure can be obtained. Specifically, the adjusted statistical table can include the following dimensions: Dimensions of adjusted statistical table 3 Dimensions of Adjusted Statistical Table 4: Adjusted Statistical Table 3 is obtained by adjusting Aggregate Statistical Table 3, and Adjusted Statistical Table 4 is obtained by adjusting Aggregate Statistical Table 4, thereby stably obtaining a spliced ​​statistical table. Step S303: Vertically splice the adjusted statistical tables to obtain a spliced ​​table. After obtaining the adjusted statistical tables, they can be vertically spliced, that is, spliced ​​according to the preset splicing dimensions, thereby stably obtaining a spliced ​​table. For example, Adjusted Statistical Table 1 and Adjusted Statistical Table 2 may include the following data: Dimensions of adjusted statistical table 1 After obtaining the above-mentioned adjusted statistical table, the dimensions of the adjusted statistical table 2 can be spliced ​​according to the dimension name and dimension value, so that the spliced ​​table 1 corresponding to the adjusted statistical table 1 and the adjusted statistical table 2 can be stably obtained. Specifically, the spliced ​​table 1 can include the following dimensions: Dimension Example 2 of Spliced ​​Table 1, Adjusted Statistics Table 3, and Adjusted Statistics Table 4 may include the following data: Dimensions of adjusted statistical table 3 After obtaining the above-mentioned adjusted statistical table, the dimensions of the adjusted statistical table 4 can be spliced ​​according to the dimension name and dimension value, so that the spliced ​​table 2 corresponding to the adjusted statistical table 3 and the adjusted statistical table 4 can be stably obtained. Specifically, the spliced ​​table 2 can include the following dimensions: Step S304 of splicing the dimensions of Table 2: performing data statistical operations on the operating indicators in the multiple aggregated statistical tables to obtain indicator statistical features, which also include: maximum value and minimum value. Since the concatenated table is obtained by concatenating multiple aggregated statistical tables, it is used to replace multiple aggregated statistical tables for cloud platform health monitoring. To ensure the stability and reliability of health monitoring operations, after obtaining multiple aggregated statistical tables, not only do they need to be concatenated, but data statistical operations can also be performed on the operating indicators in the multiple aggregated statistical tables. The aforementioned operating indicators are indicators that can be expressed as numerical values, such as the number of requests or the number of requested bytes. These data statistical operations can include data average calculation, data size comparison, data standard deviation calculation, and so on. These data statistical operations can obtain indicator statistical characteristics. These indicator statistical characteristics can include at least the average value, standard deviation, and the number of statistical indicators. In addition to these indicator statistical characteristics, indicator statistical characteristics can also include maximum and minimum values. These indicator statistical characteristics can directly reflect the indicator attribute characteristics of the operating indicators in the multiple aggregated statistical tables, thereby facilitating health monitoring operations on the cloud platform based on these indicator statistical characteristics. Step S305: Add the indicator statistical characteristics to the concatenated table to obtain the concatenated statistical table. After obtaining the above-mentioned indicator statistical features, the indicator statistical features can be added to the splicing table. In this way, a splicing statistical table including the indicator statistical features can be obtained, ensuring the accuracy and reliability of obtaining the splicing statistical table. Specifically, the splicing statistical table can include the following contents: After obtaining the aforementioned spliced ​​statistical table 1 and determining the indicator statistical characteristics, including the average value, standard deviation, and number of indicators, the aforementioned spliced ​​statistical table 1 is obtained by adding the indicator statistical characteristics to the aforementioned spliced ​​statistical table 1. This effectively enables the stable acquisition of relevant statistical characteristics of historical multi-dimensional indicators through the spliced ​​statistical table 1. In this embodiment, by obtaining replaceable dimensions from multiple aggregated statistical tables, adjusting the multiple aggregated statistical tables based on the replaceable dimensions to obtain an adjusted statistical table with a unified structure, and then vertically splicing the adjusted statistical tables to obtain a spliced ​​table. Data statistical operations are performed on the operating indicators in the multiple aggregated statistical tables to obtain the indicator statistical characteristics, and the indicator statistical characteristics are added to the spliced ​​table. This allows the spliced ​​statistical table to be accurately and stably obtained. In this way, when performing health monitoring operations on the cloud platform based on the spliced ​​statistical table, the number of tables involved in data processing can be reduced, data processing can be simplified, and the quality and efficiency of the health monitoring operations can be effectively improved. Figure 4 is a flowchart of an embodiment of the present disclosure for updating a statistical model based on a spliced ​​statistical table and current multi-dimensional indicators to obtain an updated statistical model. Based on the above embodiment, with reference to Figure 4 , this embodiment provides a technical solution for determining updated model parameters based on the spliced ​​statistical table and current multi-dimensional indicators, and determining an updated statistical model based on the updated model parameters. In this case, updating the statistical model based on the spliced ​​statistical table and current multi-dimensional indicators to obtain the updated statistical model may include: Step S401: Obtaining a dependent variable for performing health detection on a cloud platform, where the dependent variable corresponds to at least one independent variable. When performing a health check on the cloud platform, anomaly detection can be performed using a single metric or multiple metric. When performing anomaly detection using multiple metric, a dependent variable for performing health check on the cloud platform can be obtained. The dependent variable corresponds to at least one independent variable, that is, a preset association relationship exists between the dependent variable and the at least one independent variable. For example, when the dependent variable is request latency, the independent variables may be the total request volume, the read traffic size, and the write traffic size. The greater the total request volume, the greater the request latency; the smaller the total request volume, the smaller the request latency; the greater the read traffic size, the greater the request latency; the smaller the read traffic size, the smaller the request latency; the greater the write traffic size, the greater the request latency; and the smaller the write traffic size, the smaller the request latency.Furthermore, this embodiment does not limit the specific method for obtaining dependent variables. In some instances, dependent variables may be obtained through human-computer interaction. In this case, obtaining dependent variables for performing health checks on the cloud platform may include: displaying a human-computer interaction interface and obtaining an execution operation input by a user in the human-computer interaction interface; and obtaining dependent variables for performing health checks on the cloud platform based on the execution operation. In other instances, dependent variables may not only be obtained through human-computer interaction, but also be determined based on the application scenario corresponding to the cloud platform. In this case, obtaining dependent variables for performing health checks on the cloud platform may include: obtaining a mapping relationship between pre-configured application identifiers and dependent variables required for performing health checks on the cloud platform in different application scenarios; determining the current application scenario of the cloud platform, where the current application scenario may be determined based on information such as the request type, data type, and user type; and then determining dependent variables for performing health checks on the cloud platform based on the mapping relationship and the current application scenario. It should be noted that dependent variables are not limited to being acquired through the implementation methods listed above. Those skilled in the art may adjust or select the dependent variable acquisition method based on specific application scenarios or application requirements, as long as the accuracy and reliability of the dependent variable acquisition can be guaranteed. This description will not be repeated here. Step S402: Based on the splicing statistics table and the current multi-dimensional indicators, determine the first correlation between the dependent variable and the independent variable, and the second correlation between the independent variables. To accurately implement a health check operation for the cloud platform, after obtaining a dependent variable for health checking the cloud platform, and the dependent variable corresponds to at least one independent variable, the spliced ​​statistical table and the current multidimensional index may be analyzed and processed to determine a first correlation relationship between the dependent variable and the independent variable and a second correlation relationship between the independent variable and the independent variable. In some instances, the first correlation and the second correlation may be determined by analyzing and processing the spliced ​​statistical table and the current multidimensional index. In this case, determining the first correlation between the dependent variable and the independent variable and the second correlation between the independent variable and the independent variable based on the spliced ​​statistical table and the current multidimensional index includes: inputting the spliced ​​statistical table, the current multidimensional index, the dependent variable, and the at least one independent variable corresponding to the dependent variable into a statistical model for analysis and processing, thereby stably determining the first correlation between the dependent variable and the independent variable and the second correlation between the independent variable and the independent variable.In other examples, the first degree of correlation between the dependent variable and the independent variable and the second degree of correlation between the independent variable and the independent variable can be obtained not only synchronously but also asynchronously. For example, the first degree of correlation can be obtained first and then the second degree of correlation; or the second degree of correlation can be obtained first and then the first degree of correlation, and so on. That is, the operation of obtaining the first degree of correlation is independent of the operation of obtaining the second degree of correlation. Determining the first degree of correlation between the dependent variable and the independent variable based on the splicing statistical table and the current multi-dimensional index may include: determining the inter-variable product between the dependent variable and the independent variable in the current cycle based on the current multi-dimensional index; determining the first historical degree of correlation between the dependent variable and the independent variable in the previous cycle based on the splicing statistical table; and determining the first degree of correlation between the dependent variable and the independent variable based on the inter-variable product and the first historical degree of correlation. Specifically, since the current multi-dimensional indicator is indicator data collected and obtained by the cloud platform in the current period, after obtaining the current multi-dimensional indicator, the current multi-dimensional indicator can be analyzed and processed to determine the inter-variable product between the dependent variable and the independent variable of the cloud platform in the current period. For example, when the dependent variable in the current period is y(t) and the independent variables are y(t) and y(t), the inter-variable product between the dependent variable and the independent variable in the current period is y(t)*y(t). In order to accurately determine the first correlation between the dependent variable and the independent variable, it is necessary not only to determine the inter-variable product of the dependent variable and the independent variable in the current period but also to determine the historical product of the dependent variable and the independent variable in the previous period. Specifically, the splicing statistical table can be analyzed and processed to determine the first historical correlation Sk(t-1) between the dependent variable and the independent variable in the previous period, where the first historical correlation s is t. k The specific method of determining (t-1) is similar to the specific method of determining the product between variables mentioned above. For details, please refer to the above statement and will not be repeated here. k (t) * y(l) and the first historical correlation s k After (t - 1), you can multiply the variables by % k (t) * y(l) and the first historical correlation s k (t-1) is analyzed and processed, so that the first correlation degree S between the dependent variable and the independent variable can be determined k (t), in some instances, the first correlation Sk(. can be directly obtained by multiplying the product x between variables k (t) * y(l) and the first historical correlation s k (t - 1) is accumulated, that is, s k (t) = xk (t) * y(t) + s k (t - l) o In other examples, the first correlation can be obtained by performing a weighted summation of the inter-variable product and the first historical correlation. In this case, determining the first correlation between the dependent variable and the independent variable based on the inter-variable product and the first historical correlation may include: determining a weight coefficient corresponding to the first historical correlation, where the weight coefficient is greater than 0 and less than or equal to 1; determining a product value between the weight coefficient and the first historical correlation; and determining the sum of the inter-variable product and the product value as the first correlation. Specifically, for the inter-variable product x fc (t) * y(t) and the first historical correlation s k (t - 1), due to the sum of products between variables. ) * y(i) is obtained based on the analysis and processing of the cloud platform's indicator data in the current cycle, which can reflect the latest operating characteristics of the cloud platform. The first historical correlation Sk(l - 1) is obtained based on the analysis and processing of the cloud platform's historical indicator data in the previous cycle, which is used to reflect the operating characteristics of the cloud platform at the historical moment. When performing health checks on the cloud platform, the sum of products between variables corresponding to the cloud platform's indicator data in the current cycle. ) * y(i) has a higher importance or attention than the first historical correlation s corresponding to the cloud platform's indicator data in the previous cycle. k (t - 1) importance or attention, therefore, the product between variables can be kThe weight of (t) * y(t) can be set to 1, and the weight coefficient of the first historical correlation can be set to a. The first historical correlation corresponds to a weight coefficient a used to identify the importance of the first historical correlation, and the weight coefficient a can be a value greater than 0 and less than or equal to 1. In some instances, when it is necessary to determine the first correlation between a dependent variable and an independent variable, and the weight coefficients are stored in a preset area or a preset device, the weight coefficient a corresponding to the first historical correlation can be determined by accessing the preset area or the preset device. Alternatively, the weight coefficient corresponding to the first historical correlation can be determined through human-computer interaction. In this case, determining the weight coefficient corresponding to the first historical correlation can include: displaying a human-computer interaction interface, obtaining an execution operation input by a user in the human-computer interaction interface, and determining the weight coefficient corresponding to the first historical correlation based on the execution operation. After determining the weight coefficient a corresponding to the first historical correlation, the product value a*Sk(t-1) between the weight coefficient a and the first historical correlation Sk(t-1) can be determined, and then the product between the variables can be summed. ) * y(l) and the product value a * ) is determined. Specifically, the product %i(t) * x7(t) between the independent variable (1) and the independent variable % / (*) can be directly determined as the intra-variable product, or the sum of the product %i(t) * Xj (t) between the independent variable (1) and the independent variable % / (*) and the preset deviation E can be determined as the intra-variable product. In order to accurately determine the second correlation between the independent variables, it is necessary not only to determine the intra-variable product of the independent variables in the current cycle, but also to determine the historical correlation between the independent variables in the previous cycle. Specifically, the splicing statistical table can be analyzed and processed to determine the second historical correlation between the independent variables in the previous cycle. The specific method for determining the second historical correlation is similar to the specific method for determining the intra-variable product described above. The cumulative first 0) * Xj (t) is obtained by analyzing and processing the indicator data of the cloud platform in the current cycle, which can reflect the cloud - 1) importance or attention, therefore, the product r within the variable can be £; The weight of (t - 1) can be set to 1 OIn this case, determining the second degree of association between the independent variables based on the intra-variable product and the second historical degree of association may include: determining a weight coefficient corresponding to the second historical degree of association, where the weight coefficient is greater than 0 and less than or equal to 1; determining a product value between the weight coefficient and the second historical degree of association; and determining the sum of the intra-variable product and the product value as the second degree of association. The second historical degree of association r £; (t-1) corresponds to a weight coefficient a for identifying the importance of the second historical correlation. The weight coefficient a can be a value greater than 0 and less than or equal to 1. When the second correlation between independent variables needs to be determined, and the weight coefficient is stored in a preset area or a preset device, the weight coefficient a corresponding to the second historical correlation can be determined by accessing the preset area or the preset device. Alternatively, when the second correlation between independent variables needs to be determined, the weight coefficient corresponding to the second historical correlation can be determined through human-computer interaction. In this case, determining the weight coefficient corresponding to the second historical correlation can include: displaying a human-computer interaction interface, obtaining an execution operation input by a user in the human-computer interaction interface, and determining the weight coefficient corresponding to the second historical correlation based on the execution operation. ar £; (t=3), a is a weight greater than 0 and less than or equal to 1 Fixed r £; (t - 3) = x k (t - 3)y(t - 3) + ar £;(t - 4), etc., thereby effectively implementing the use of an online incremental algorithm to determine the historical correlation. The advantages of the online incremental algorithm are similar to those of the online incremental algorithm in the above embodiment. For details, please refer to the above description and will not be repeated here. Furthermore, for the weight coefficient corresponding to the first historical correlation, to improve the accuracy and reliability of determining the first correlation, the weight coefficient may decrease exponentially with the number of intervals between the historical cycle and the current cycle. In this case, determining the weight coefficient corresponding to the first historical correlation may include: obtaining an initial weight corresponding to the initial cycle, where the initial weight is greater than 0 and less than 1; determining the number of period intervals between the previous cycle and the initial cycle; and determining the weight coefficient corresponding to the historical correlation based on the number of period intervals and the initial weight. Determining the weight coefficient corresponding to the historical correlation based on the number of period intervals and the initial weight may include: using the initial weight as a base and the number of period intervals as an exponential coefficient of the initial weight to obtain a racing function weight corresponding to the initial weight; and determining the racing function weight as the weight coefficient corresponding to the historical correlation. Specifically, the weight coefficient a can be a weight coefficient greater than 0 and less than or equal to 1, which is used to identify the importance of samples used to calculate the correlation between variables. Samples collected in the current calculation period t have a weight of 1, samples collected in the previous calculation period t-1 have a weight of 1, samples collected in the previous calculation period t-2 have a weight of 2, and so on. The weight of samples collected in the first time period farthest from the current moment is a weight of 1. If a is greater than 0 and less than 1, the sample weight a decays exponentially with the age of the sample. The older the sample, the smaller the weight a and the smaller its contribution to the correlation between variables. If a is equal to 1, all samples have a weight of 1, meaning that samples collected in all time periods are equally important. The specific value of the weight coefficient a can be determined based on the specific application scenario or application requirements. For example, the user can set the sample weight coefficient to drop to 1 / 2 every 3 days. Then the weight of the sample 6 days ago is reduced to 1 / 4, the weight of the sample 9 days ago is reduced to 1 / 8, and so on. Assuming that the data collection cycle is 10 seconds, there are 8640 data collection cycles in one day and 25920 data collection cycles in 3 days. Then a 25920 =1 / 2, we can calculate a= 2-<1 / 25920), so if the application scenario requires the sample weight to be reduced to 1 / 2 every n days, and there are m data collection cycles per day, then calculate a= 2-<1 / (nm)) o Setting the weight coefficients to decay over time not only helps the statistical model capture the latest relationships between indicators, but also helps the model capture the latest relationships between indicators. Specifically, in the early stages of a cloud platform, the corresponding traffic is very low. As the cloud platform's data processing progresses, the platform's traffic increases, and the statistical models at different processing stages may differ. If the weight coefficients do not decay, the importance of the operating indicators is dominated by historical data indicators, and new operating indicators have limited impact on the statistical model. However, the decay of the weight coefficients can suppress the importance of old data samples, allowing the latest operating indicators to play a more important role. Therefore, the decay of the weight coefficients can help the statistical model truly reflect the underlying patterns between these operating indicators. Another benefit of the weight decay process is that it enables incremental updates of the statistical model. The updated statistical model can then perform health checks on the cloud platform, making health checks more efficient in terms of time and storage space. Step S403: Based on the first and second correlations, the updated model parameters are determined. After obtaining the first and second degrees of association, the first and second degrees of association can be analyzed and processed to determine updated model parameters. In some instances, the updated model parameters can be obtained by analyzing and processing the first and second degrees of association using a preset machine learning model. In this case, determining the updated model parameters based on the first and second degrees of association may include: obtaining a preset machine learning model for analyzing and processing the first and second degrees of association; and analyzing and processing the first and second degrees of association using the preset machine learning model, thereby obtaining the updated model parameters. In other instances, the parameters can be obtained by directly analyzing and processing the first and second degrees of association. In this case, determining the updated model parameters based on the first and second degrees of association may include: determining first matrix information based on the first degree of association; determining second matrix information based on the second degree of association; and determining the updated model parameters based on the first and second matrix information. Specifically, after obtaining the first degree of association, the first degree of association may be analyzed and processed to obtain first matrix information. In some instances, the first degree of association corresponding to different operating cycles of the cloud platform may be determined as a one-dimensional matrix.

[0002] ST data column elements, thereby obtaining the first matrix information S will be '2. Similarly, after obtaining the second correlation degree s'k. ri,i ■■- rThe elements in i,p' can be used to obtain the second matrix information & - nj ....

[0003] - 7 P,I ■■- r After obtaining the first matrix information and the second matrix information, the first matrix information and the second matrix information can be analyzed and processed to determine the updated model parameters. The updated model parameters can be obtained by directly processing the first matrix information and the second matrix information. In this case, determining the updated model parameters based on the first matrix information and the second matrix information may include: obtaining the inverse matrix of the second matrix information; multiplying the inverse matrix by the first matrix information to obtain a product matrix; and determining the updated model parameters based on the product matrix. For example, when the first matrix information is S(t) and the second matrix information is R(t), in order to accurately determine the updated model parameters, the inverse matrix R(t) of the second matrix information can be first obtained, and then the inverse matrix R(t) can be multiplied by the first matrix information. The updated model parameters can be determined based on the product matrix. Specifically, the elements % and ^{\text ... Then, based on the initialized first matrix information and the second matrix information, Model update operations can ensure the quality and effectiveness of model update operations. After waiting for p+10 acquisition cycles to continuously update the parameters of the first statistical model, the first model is used to infer the value of the dependent variable indicator. Only then can it be used to calculate the value of the dependent variable of the second statistical model. Only then does the calculation and update of the parameters of the second statistical model begin. Then, after waiting for another p+10 acquisition cycles to continuously update the parameters of the second statistical model, both statistical models are used for anomaly detection. This effectively ensures that the statistical models use a sufficient number of data samples to reduce noise interference, ensuring the robustness and reliability of the statistical models. In this embodiment, the dependent variable used for cloud platform health monitoring is obtained. Based on the concatenated statistical table and the current multi-dimensional indicators, a first correlation between the dependent variable and the independent variable, and a second correlation between the independent variables, are determined. Updated model parameters are then determined based on the first and second correlations. Consequently, an updated statistical model can be determined based on the updated model parameters. The entire calculation is an online incremental process. Each data collection cycle updates the statistical model parameters using only the data collected in the current cycle, rather than requiring a batch calculation of data from all or a large number of previous collection cycles. This results in highly efficient calculations. Furthermore, the health monitoring model assigns a weight to samples collected in each collection cycle. Newer data has a higher weight, while older data has a lower weight. Data weight decays exponentially with age, allowing the statistical model to keep pace with system changes. If all data has the same weight, then after the model has been running for a long time, the impact of recent samples on the statistical model will be minimal. However, if the cloud platform changes, the corresponding statistical model will also change. However, since the statistical model uses a large amount of data, the proportion of the most recently collected samples in the overall data is relatively small, making it difficult for the statistical model to keep up with the changes in the cloud platform. The online incremental statistical model in this embodiment uses an exponential decay method to avoid this situation. This effectively ensures the efficiency, accuracy, and reliability of determining the updated statistical model. This facilitates health checks on the cloud platform based on the updated statistical model, further ensuring the quality and effectiveness of health checks.FIG5 is a flowchart illustrating a process for performing health checks on a cloud platform based on an updated statistical model and obtaining health check results, provided by an embodiment of the present disclosure. Based on the above embodiment, and referring to FIG5 , when the health check operation of the cloud platform is implemented using multiple metrics, this embodiment provides a method for performing health checks on the cloud platform based on the updated statistical model and obtaining health check results. Specifically, the method may include: Step S501: Obtaining an actual detection value of a target operating indicator used to implement the health check operation. The target operating indicator is determined by operating indicators from at least one dimension. To implement health checks on the cloud platform using multiple metrics, the actual detection value of the target operating indicator used to implement the health check operation may be obtained. The target operating indicator may not only be determined by operating indicators from at least one dimension, but also may correspond to different types of target operating indicators for different application scenarios. For example, the target operating indicator may include at least one of the following: request latency information, data processing accuracy, network throughput, resource utilization, etc. Of course, those skilled in the art may flexibly configure or adjust the specific type of the target operating indicator based on specific application scenarios or requirements, and this will not be further described here. Furthermore, this embodiment does not limit the specific method for obtaining the actual detection value of the target operating indicator. In some instances, the actual detection value of the target operating indicator can be obtained by a preset detection module provided in the cloud platform, or the actual detection value can be a value obtained through pre-detection and stored in a preset area or device. In this case, by actively or passively obtaining the actual detection value of the target operating indicator used to perform health monitoring operations, the flexibility and reliability of obtaining the actual detection value are effectively ensured. Step S502: Based on the updated statistical model and the current multi-dimensional indicators, a predicted indicator value corresponding to the target operating indicator is obtained. Since the updated statistical model can identify patterns or regularities in operating indicator changes, in order to accurately perform health monitoring operations on the cloud platform, after obtaining the updated statistical model and the current multi-dimensional indicators, the updated statistical model can be used to analyze and process the current multi-dimensional indicators to obtain the predicted indicator value corresponding to the target operating indicator. Specifically, the current multi-dimensional indicators can be input into the updated statistical model to obtain the predicted indicator value output by the updated statistical model. Step S503: Based on the splicing statistics table, determine the standard deviation of the cloud platform corresponding to the target operating indicator in the previous cycle.The spliced ​​statistical table includes statistical features such as "standard deviation" corresponding to each operating cycle. After determining the target operating indicator of the cloud platform in the previous cycle, the standard deviation corresponding to the target operating indicator can be directly extracted from the spliced ​​statistical table based on the target operating indicator of the cloud platform in the previous cycle. This allows for stable and reliable determination of the standard deviation corresponding to the target operating indicator. Step S504: Perform a health check on the cloud platform based on the actual detection value, the predicted indicator value, and the standard deviation to obtain a health check result. In some instances, the health check operation can be implemented using a pre-trained machine learning model. In this case, performing a health check on the cloud platform based on the actual detection value, the predicted indicator value, and the standard deviation to obtain the health check result can include: obtaining the pre-trained machine learning model; inputting the actual detection value, the predicted indicator value, and the standard deviation into the machine learning model; and obtaining the health check result output by the machine learning model. This effectively ensures the accuracy and reliability of the health check result. In other instances, cloud platform health monitoring operations can be implemented not only through machine learning models but also through preset algorithms. Specifically, performing a health check on the cloud platform based on actual detection values, predicted indicator values, and standard deviations to obtain health monitoring results may include: determining a model prediction error corresponding to an updated statistical model based on the actual detection values ​​and predicted indicator values; determining a preset detection coefficient for performing health monitoring on the cloud platform; and performing a health check on the cloud platform based on the model prediction error, the preset detection coefficient, and the standard deviation to obtain health monitoring results. Specifically, after obtaining the actual detection values ​​and predicted indicator values, the actual detection values ​​and predicted indicator values ​​may be analyzed and processed to determine the model prediction error corresponding to the updated statistical model. In some instances, the model prediction error may be the difference between the actual detection value and the predicted indicator value. For example, when the actual detection value is % and the predicted indicator value is m, the model prediction error may be |y. t- y ^ | , the difference can be a value greater than or equal to 0. In order to implement a health check operation for the cloud platform, in addition to determining the model prediction error corresponding to the updated statistical model, it is also necessary to determine a preset detection coefficient for performing health checks on the cloud platform. In some instances, the preset detection coefficient can be obtained through human-computer interaction. In this case, determining the preset detection coefficient for performing health checks on the cloud platform may include: displaying a human-computer interaction interface; obtaining an execution operation input by a user in the human-computer interaction interface; and determining the preset detection coefficient for performing health checks on the cloud platform based on the execution operation. The preset detection coefficient is a hyperparameter and can be set to 3, 4, 5, etc. Those skilled in the art can adjust or configure the preset detection coefficient according to specific application requirements. In other instances, the preset detection coefficient can be obtained not only through human-computer interaction but also by accessing a preset area or device. The preset area or device may store preconfigured or prestored preset detection coefficients. In this case, the preset detection coefficients can be actively or passively obtained by accessing the preset area or device, thus ensuring the accuracy and reliability of the preset detection coefficients. After obtaining the model prediction error, the preset detection coefficient, and the standard deviation, the model prediction error, the preset detection coefficient, and the standard deviation can be analyzed and processed to perform a health check on the cloud platform and obtain a health check result. In some instances, the health check on the cloud platform can be performed using a pretrained machine learning model. In this case, performing a health check on the cloud platform based on the model prediction error, the preset detection coefficient, and the standard deviation to obtain the health check result can include: obtaining a machine learning model pre-trained for health checking; and inputting the model prediction error, the preset detection coefficient, and the standard deviation into the machine learning model to obtain the health check result output by the machine learning model. In other instances, the health check operation of the cloud platform can not only be implemented through a pre-trained machine learning model, but can also be obtained by analyzing and processing the model's preset error, preset detection coefficient, and standard deviation through a preset algorithm or preset rules. In this case, performing a health check on the cloud platform based on the model prediction error, preset detection coefficient, and standard deviation, and obtaining the health check result may include: determining the product value of the preset detection coefficient and the standard deviation; and performing a health check on the cloud platform based on the model prediction error and the product value to obtain the health check result.Specifically, after obtaining the preset detection coefficient and standard deviation, the preset detection technology and the standard deviation can be multiplied to obtain a product value of the preset detection coefficient and the standard deviation. The product value can then be determined as a limit for performing health checks on the cloud platform. Health checks on the cloud platform can then be performed based on the model prediction error and the product value, thereby stably obtaining health check results. In some instances, performing health checks on the cloud platform based on the model prediction error and the product value to obtain health check results can include: analyzing and comparing the model prediction error with the product value; if the model prediction error is greater than the product value, determining the health check result as a first detection result indicating abnormal operation of the cloud platform; and if the model prediction error is less than or equal to the product value, determining the health check result as a second detection result indicating normal operation of the cloud platform. For example, when the model prediction error is y - (b. p + bg + b2x2+ ••• + bp_iXp_D |、 Preset detection system The product value k* of the difference (such as + b 1X1 + b2x2+ - + bp-iXp-i), and then the model prediction error and the product value can be The platform performs stable health checks. In this embodiment, by obtaining the actual detection value of the target operating indicator used to implement the health check, then obtaining the indicator prediction value corresponding to the target operating indicator based on the updated statistical model and the current multi-dimensional indicator, determining the standard deviation of the cloud platform corresponding to the target operating indicator in the previous cycle based on the spliced ​​statistical table, and performing a health check on the cloud platform based on the actual detection value, the indicator prediction value, and the standard deviation to obtain a health check result, thereby effectively achieving stable health checks on the cloud platform. Figure 6 is a second flow diagram of performing health checks on the cloud platform based on the updated statistical model and obtaining health check results, according to an embodiment of the present disclosure. Based on the above embodiment, with reference to Figure 6, when the health check operation of the cloud platform is implemented using a single metric, this embodiment provides a method for performing health checks on the cloud platform based on the updated statistical model and obtaining health check results. Specifically, the method may include: Step S601: Obtaining the actual detection value of the single-dimensional detection indicator used to perform health checks on the cloud platform. To implement health monitoring operations for the cloud platform using a single metric, the actual detection value of the single-dimensional detection metric used to implement the health monitoring operation can be obtained. This embodiment does not limit the specific method for obtaining the actual detection value of the single-dimensional detection metric. In some instances, the actual detection value of the single-dimensional detection metric can be obtained by a preset detection module provided in the cloud platform, or the actual detection value can be a value obtained through pre-detection and stored in a preset area or device. In this case, actively or passively obtaining the actual detection value of the single-dimensional detection metric used to implement the health monitoring operation effectively ensures flexible and reliable acquisition of the actual detection value. Step S602: Based on the updated statistical model and the current multi-dimensional metric, determine the average value and standard deviation of the metric corresponding to the single-dimensional detection metric of the cloud platform in the current period. Since the updated statistical model can identify patterns or regularities in changes in operating indicators, in order to accurately perform health checks on the cloud platform, after obtaining the updated statistical model and the current multi-dimensional indicators, the updated statistical model can be used to analyze and process the current multi-dimensional indicators. This can determine the average and standard deviation of the indicators corresponding to the single-dimensional detection indicators of the cloud platform in the current cycle.In some instances, determining the average value of the indicator corresponding to the single-dimensional detection indicator of the cloud platform in the current cycle based on the updated statistical model and the current multi-dimensional indicator may include: determining the indicator detection sum value of the single-dimensional detection indicator in the current cycle and the previous cycle based on the current multi-dimensional indicator and the splicing statistical table; determining the weight sum value corresponding to the single-dimensional detection indicator in the current cycle and the previous cycle; and determining the indicator average value as the ratio of the indicator detection sum value to the weight sum value. After obtaining the current multi-dimensional indicator and the splicing statistical table, the current multi-dimensional indicator and the splicing statistical table may be analyzed and processed to determine the single-dimensional detection indicator of the cloud platform in the previous cycle. If the single-dimensional detection indicator of the cloud platform in the previous cycle is aS(t-1)> the current multi-dimensional indicator, then the indicator detection sum value of the single-dimensional detection indicator of the cloud platform in the current cycle and the previous cycle may be determined. This indicator sum value may be S(t) = x. t + aS(t - l) o In order to accurately determine the average value of the indicator, it is necessary not only to determine the indicator detection sum value of the single-dimensional detection indicator of the cloud platform in the current cycle and the previous cycle, but also to determine the weight sum value corresponding to the single-dimensional detection indicator in the current cycle and the previous cycle. Since the current multi-dimensional indicator corresponding to the current cycle can identify the operating status of the cloud platform in the current cycle, the obtained weight information of the single-dimensional detection indicator of the cloud platform in the current cycle can be set to 1. The weight information corresponding to the single-dimensional detection indicator of the cloud platform in the previous cycle can be determined by multiplying the preset coefficient corresponding to the previous cycle and the weight itself. For example, the preset coefficient corresponding to the cloud platform in the previous cycle is a, and the weight itself is the weight itself w(t-1), where the preset coefficient can decrease exponentially with the increase in the number of time intervals between the previous cycle and the current cycle; after obtaining the preset coefficient a and the weight itself W0-1), the weight information corresponding to the single-dimensional detection indicator of the cloud platform in the previous cycle can be determined to be -1, thereby value, determined as the indicator average value 140), that is, U0) = 4;, thereby stably obtaining the indicator average value 140). Similarly, determining the standard deviation corresponding to the single-dimensional detection indicator of the cloud platform in the current cycle may include: obtaining the indicator square value corresponding to the single-dimensional detection indicator based on the current multi-dimensional indicator; determining the historical indicator square value corresponding to the single-dimensional detection indicator of the cloud platform in the previous cycle based on the splicing statistical table; determining the weight coefficient corresponding to the historical indicator square value; and analyzing and processing the indicator average value, the weight coefficient, the indicator square value, and the historical indicator square value based on the updated statistical model to obtain the standard deviation. Specifically, when the single-dimensional detection indicator is Q, in order to accurately obtain the standard deviation, the square value of the indicator corresponding to the single-dimensional detection indicator can be first obtained based on the current multi-dimensional indicator being detected. After obtaining the splicing statistical table, the splicing statistical table can be analyzed and processed to determine the historical indicator square value Q(t-1) corresponding to the single-dimensional detection indicator of the cloud platform in the previous cycle. Since the importance of the data in the previous cycle is reduced compared to the health detection operation of the cloud platform in the current cycle, in order to accurately implement the health detection of the cloud platform, it is necessary not only to determine the historical indicator square value corresponding to the single-dimensional detection indicator of the cloud platform in the previous cycle, but also to determine the weight coefficient a corresponding to the historical indicator square value. The weight coefficient a can be obtained through human-computer interaction or preset configuration. After obtaining the square value of the indicator, the weight coefficient, and the historical square value of the indicator, the updated statistical model can be used to analyze and process the indicator average value, the weight coefficient, the square value of the indicator, and the historical square value of the indicator to obtain the standard deviation. In some examples, analyzing and processing the indicator average value, the weight coefficient, the square value of the indicator, and the historical square value of the indicator based on the updated statistical model to obtain the standard deviation may include: obtaining the cloud platform in the current cycle and the previous cycle with a single dimension; Then, based on the square value of the indicator, the weight coefficient, and the square value of the historical indicator, determine the weighted square sum of the indicator detection corresponding to the single-dimensional detection indicator in the current cycle and the previous cycle of the cloud platform (Q0) = filter + aQ(t-1); Based on the square value of the indicator detection, the weighted sum value, and the average value of the indicator, After obtaining the variance information, the standard deviation for implementing the health detection operation can be obtained by performing a square root processing on the variance information (J<70), specifically, <J(t) = J<j20), which effectively ensures the accuracy and reliability of determining the standard deviation. In the above process of calculating the statistical characteristics such as the mean value and standard deviation of the detection indicators, the sample weight will also be an online incremental calculation process with exponential decay according to the sample age. The calculation process is very efficient and fast, and it can adaptively adapt to the characteristic changes of the indicators. Step S603: Perform a health check on the cloud platform based on the actual detection value, the indicator mean value and the standard deviation to obtain a health detection result. After obtaining the actual detection value, the indicator mean value and the standard deviation, the actual detection value, the indicator mean value and the standard deviation can be analyzed and processed to implement the health detection operation of the cloud platform and obtain a health detection result. In some instances, The product value ko is analyzed and compared. When |y - it| > ka, the health check result can be determined as a first detection result indicating abnormal cloud platform operation. When \y - u\ < ka, the health check result can be determined as a second detection result indicating normal cloud platform operation. This effectively implements stable health check operations on the cloud platform. In this embodiment, the actual detection value of a single-dimensional detection indicator used for health check of the cloud platform is obtained, and the average value and standard deviation of the indicator corresponding to the single-dimensional detection indicator of the cloud platform in the current cycle are determined based on the updated statistical model and the current multi-dimensional indicators. The health check of the cloud platform is then performed based on the actual detection value, the average value, and the standard deviation of the indicator, thereby effectively implementing stable health check operations on the cloud platform. FIG7 is a flow chart illustrating another cloud platform health check method provided by an embodiment of the present disclosure. Based on any of the above embodiments, and referring to FIG7 , after obtaining a health check result, to improve the quality and effectiveness of the health check operation, this embodiment provides a technical solution for optimizing and adjusting a preset detection coefficient used to implement the health check operation. In this case, the method in this embodiment may further include: Step S701: Obtaining result feedback information corresponding to the health check result. Step S702: Adjusting the preset detection coefficient based on the result feedback information to obtain an adjusted detection coefficient. After obtaining the health check result, a user may provide feedback on the health check result. For example, if the health check result is inaccurate, the user may provide negative result feedback information regarding the health check result; if the health check result is accurate, the user may provide positive result feedback information regarding the health check result. After obtaining the result feedback information, the preset detection coefficient can be adjusted based on the result feedback information to obtain the adjusted detection coefficient. In some instances, the adjustment operation of the preset detection coefficient can be implemented by a pre-trained machine learning model, or the adjustment operation of the preset detection coefficient can be implemented by a preset rule. In this case, adjusting the preset detection coefficient based on the result feedback information to obtain the adjusted detection coefficient may include: determining a detection false alarm rate and a detection missed alarm rate corresponding to the preset detection coefficient based on the result feedback information; determining a false alarm rate weight corresponding to the detection false alarm rate and a missed alarm rate weight corresponding to the detection missed alarm rate; determining a detection false alarm value based on the detection false alarm rate and the false alarm rate weight; determining a detection missed alarm value based on the missed alarm rate weight and the detection missed alarm rate; and adjusting the preset detection coefficient based on the detection false alarm value and the missed alarm value to obtain the adjusted detection coefficient.Specifically, user feedback information regarding health check results may include a false positive rate and a false negative rate. The false positive rate indicates the probability of errors in the health check results. For example, if the cloud platform is operating normally but the health check result indicates that the cloud platform is operating abnormally, a false positive has occurred. If the cloud platform is operating abnormally but the health check result indicates that the cloud platform is operating normally, a false positive has occurred. The false negative rate indicates the probability of false negatives in the health check results. For example, if node A on the cloud platform is operating abnormally but nodes B, C, and D in the health check result indicate that they are operating abnormally, a false negative has occurred. If node A on the cloud platform is operating normally but nodes B, D, and E in the health check result indicate that they are operating normally, a false negative has occurred. Since missed detection and false alarm detection may occur simultaneously during the health check operation of the cloud platform, in order to accurately adjust the preset detection coefficient, after obtaining the false alarm rate and the missed detection rate, the false alarm rate weight corresponding to the false alarm rate and the missed detection rate weight corresponding to the missed detection rate can be determined. The false alarm rate and the false alarm rate weight can then be analyzed and processed to determine the false alarm value. In some examples, the false alarm value can be determined by multiplying the false alarm rate by the false alarm rate weight. For example, the false alarm value = false alarm rate * false alarm rate weight, or the false alarm value = false alarm rate * false alarm rate weight + preset offset. Similarly, the detection omission rate and the omission rate weight can also be analyzed and processed to determine the detection omission value. In some instances, the detection omission value can be determined by multiplying the detection omission rate and the omission rate weight, for example: detection omission value = detection omission rate * omission rate weight, or detection omission value = detection omission rate * omission rate weight + preset offset.After obtaining the detection false alarm value and the detection omission value, the detection false alarm value and the detection omission value can be analyzed and processed to adjust the preset detection coefficient and obtain the adjusted detection coefficient. In some instances, the adjusted detection coefficient can be implemented by a pre-trained machine learning model, or the adjusted detection coefficient can be obtained by analyzing and processing the detection false alarm value and the detection omission value using a preset algorithm. In this case, adjusting the preset detection coefficient based on the detection false alarm value and the detection omission value to obtain the adjusted detection coefficient may include: when the detection false alarm value is greater than the detection omission value, increasing the preset detection coefficient to obtain the adjusted detection coefficient; when the detection false alarm value is less than the detection omission value, decreasing the preset detection coefficient to obtain the adjusted detection coefficient; and when the detection false alarm value is equal to the detection omission value, maintaining the preset detection coefficient unchanged. Specifically, after obtaining the detection false alarm value and the detection missed alarm value, the detection false alarm value and the detection missed alarm value can be analyzed and compared. When the detection false alarm value is greater than the detection missed alarm value, it means that the detection false alarm situation is greater than the detection missed alarm situation in the health detection operation of the cloud platform at this time. In order to reduce the detection false alarm situation, the preset detection coefficient can be increased, so that the adjusted detection coefficient can be obtained. In some instances, increasing the preset detection coefficient to obtain the adjusted detection coefficient may include: obtaining amplitude information for adjusting the preset detection coefficient; and increasing the preset detection coefficient based on the amplitude information, so as to obtain the adjusted detection coefficient. Correspondingly, when the false positive detection value is less than the false negative detection value, it indicates that the number of false positive detections in the cloud platform's health check operation is currently lower than the number of false negative detections. To reduce the number of false negative detections, the preset detection coefficient can be adjusted downward to obtain an adjusted detection coefficient. In some instances, adjusting the preset detection coefficient downward to obtain the adjusted detection coefficient can include: obtaining amplitude information for adjusting the preset detection coefficient; and adjusting the preset detection coefficient downward based on the amplitude information to obtain the adjusted detection coefficient. Correspondingly, when the false positive detection value is equal to or substantially equal to the false negative detection value, it indicates that the number of false positive detections and false negative detections in the cloud platform's health check operation is at the same or nearly the same level. Furthermore, it indicates that the preset detection coefficient used to implement the cloud platform's health check operation is relatively reasonable. Therefore, the preset detection coefficient can be maintained unchanged, thereby completing the adjustment of the preset detection coefficient.In other examples, the preset detection coefficient can be adjusted not only based on the false alarm rate and missed alarm rate included in the result feedback information, but also based on the false alarm rate or missed alarm rate included in the result feedback information. In this case, adjusting the preset detection coefficient based on the result feedback information to obtain the adjusted detection coefficient may include: determining a false alarm rate corresponding to the preset detection coefficient based on the result feedback information; determining a false alarm rate weight corresponding to the false alarm rate; and determining a false alarm value based on the false alarm rate weight and the false alarm rate. If the false alarm value is greater than a preset threshold, it indicates that a high number of false alarms are occurring during health checks on the cloud platform. To reduce false alarms, the preset detection coefficient may be increased, thereby obtaining an adjusted detection coefficient. If the false alarm value is less than the preset threshold, it indicates that a low number of false alarms are occurring during health checks on the cloud platform. To improve the accuracy of health checks on the cloud platform, the preset detection coefficient may be decreased, thereby obtaining an adjusted detection coefficient. When the false positive detection value is equal to the preset threshold, it indicates that the false positive detection rate in the health check operation of the cloud platform is relatively moderate, and the preset detection coefficient can be maintained unchanged. Alternatively, adjusting the preset detection coefficient based on the result feedback information to obtain the adjusted detection coefficient may include: determining a false negative detection rate corresponding to the preset detection coefficient based on the result feedback information; determining a false negative detection rate weight corresponding to the false negative detection rate; and determining a false negative detection value based on the false negative detection rate weight and the false negative detection rate. When the false negative detection value is greater than the preset threshold, it indicates that the false negative detection rate in the health check operation of the cloud platform is relatively high. To reduce the false negative detection rate, the preset detection coefficient may be reduced, thereby obtaining the adjusted detection coefficient. When the detection omission value is less than the preset threshold, it indicates that the cloud platform health check operation has relatively few omissions. To improve the accuracy of the cloud platform health check, the preset detection coefficient can be increased to obtain an adjusted detection coefficient. When the detection omission value is equal to the preset threshold, it indicates that the cloud platform health check operation has a moderate omission rate. In this case, the preset detection coefficient can be maintained unchanged. It is understood that those skilled in the art can flexibly configure or select an implementation method for adjusting the preset detection coefficient based on specific application scenarios or application requirements. As long as the preset detection coefficient can be accurately adjusted, this will not be further described here.In this embodiment, by obtaining result feedback information corresponding to the health detection result, the preset detection coefficient can be adjusted based on the result feedback information to obtain the adjusted detection coefficient. This effectively implements the adjustment of the preset detection coefficient. When performing a health detection operation on the cloud platform based on the adjusted detection coefficient, the quality and effectiveness of the health detection operation on the cloud platform can be improved. Figure 8 is a schematic structural diagram of a health detection device for a cloud platform provided by an embodiment of the present disclosure. Referring to Figure 8, this embodiment provides a health detection device for a cloud platform, which is used to perform the health detection method for the cloud platform shown in Figure 2. In this case, the health detection device for the cloud platform in this embodiment may include: a first acquisition module 11 for obtaining current multi-dimensional indicators corresponding to the cloud platform in the current cycle; a first determination module 12 for determining a spliced ​​statistical table corresponding to the operating indicators of the cloud platform in the previous cycle. The spliced ​​statistical table includes at least indicator statistical features obtained by statistically processing the historical multi-dimensional indicators of the cloud platform in the previous cycle. The indicator statistical features include at least the average value, standard deviation, and number of statistical indicators. The first determination module 12 is configured to determine a statistical model for implementing a health check operation based on the spliced ​​statistical table. The first processing module 13 is configured to perform a health check on the cloud platform based on the spliced ​​statistical table, the statistical model, and the current multi-dimensional indicators to obtain a health check result. In some examples, when the first determination module 12 determines the spliced ​​statistical table corresponding to the cloud platform's operating indicators in the previous cycle, the first determination module 12 is configured to: obtain historical multi-dimensional indicators corresponding to the cloud platform in the previous cycle; aggregate the historical multi-dimensional indicators to obtain multiple aggregated statistical tables; and splice the multiple aggregated statistical tables to obtain a spliced ​​statistical table, where the spliced ​​statistical table also includes: preset splicing dimensions and splicing dimension values ​​corresponding to the preset splicing dimensions. In some instances, when the first determining module 12 aggregates the historical multi-dimensional indicators to obtain multiple aggregated statistical tables, the first determining module 12 is configured to: obtain preset aggregation dimensions for implementing an aggregation operation, where the preset aggregation dimensions include at least a time period, a return code, and one or two replaceable dimensions, where the replaceable dimensions are obtained by traversing and replacing the remaining dimensions of the historical multi-dimensional indicators except the time period and the return code; and aggregate the historical multi-dimensional indicators based on the preset aggregation dimensions to obtain multiple aggregated statistical tables.In some instances, when the first determining module 12 performs splicing processing on multiple aggregate statistical tables to obtain a spliced ​​statistical table, the first determining module 12 is configured to: obtain replaceable dimensions from the multiple aggregate statistical tables; adjust the multiple aggregate statistical tables based on the replaceable dimensions to obtain a structurally unified adjusted statistical table; vertically splice the adjusted statistical tables to obtain a spliced ​​table; perform data statistics on the operating indicators in the multiple aggregate statistical tables to obtain indicator statistical features, which may include a maximum value and a minimum value; and add the indicator statistical features to the spliced ​​table to obtain the spliced ​​statistical table. In some instances, when the first determining module 12 adjusts the multiple aggregate statistical tables based on the replaceable dimensions to obtain a structurally unified adjusted statistical table, the first determining module 12 is configured to: add a dimension name for splicing the multiple aggregate statistical tables to the aggregate statistical table based on the replaceable dimensions; determine a dimension value corresponding to the replaceable dimension based on the aggregate statistical table; and adjust the aggregate statistical tables based on the dimension name and dimension value to obtain a structurally unified adjusted statistical table. In some instances, when the first processing module 13 performs a health check on the cloud platform based on the spliced ​​statistical table, the statistical model, and the current multi-dimensional indicators and obtains a health check result, the first processing module 13 is configured to: update the statistical model based on the spliced ​​statistical table and the current multi-dimensional indicators to obtain an updated statistical model; and perform a health check on the cloud platform based on the updated statistical model to obtain a health check result. In some instances, the statistical models corresponding to different parameter dimension combinations in the spliced ​​statistical table have the same model structure, and the statistical models corresponding to different parameter dimension combinations have different model parameters. In some instances, when the first processing module 13 updates the statistical model based on the spliced ​​statistical table and the current multidimensional indicators to obtain an updated statistical model, the first processing module 13 is configured to perform the following steps: obtaining a dependent variable for health detection of the cloud platform, where the dependent variable corresponds to at least one independent variable; determining a first correlation between the dependent variable and the independent variable, and a second correlation between the independent variables, based on the spliced ​​statistical table and the current multidimensional indicators; determining updated model parameters based on the first correlation and the second correlation; and determining an updated statistical model based on the updated model parameters.In some examples, when the first processing module 13 determines the first correlation between the dependent variable and the independent variable based on the splicing statistical table and the current multidimensional index, the first processing module 13 is configured to: calculate the inter-variable product between the dependent variable and the independent variable in the current period based on the current multidimensional index; determine the first historical correlation between the dependent variable and the independent variable in the previous period based on the splicing statistical table; and determine the first correlation between the dependent variable and the independent variable based on the inter-variable product and the first historical correlation. In some examples, when the first processing module 13 determines the first correlation between the dependent variable and the independent variable based on the inter-variable product and the first historical correlation, the first processing module 13 is configured to: determine a weight coefficient corresponding to the first historical correlation, where the weight coefficient is greater than 0 and less than or equal to 1; determine a product value between the weight coefficient and the first historical correlation; and determine the sum of the inter-variable product and the product value as the first correlation. In some examples, when the first processing module 13 determines the second correlation between the independent variables based on the splicing statistical table and the current multidimensional index, the first processing module 13 is configured to: perform the intra-variable product between the independent variables in the current period based on the current multidimensional index; determine the second historical correlation between the two independent variables in the previous period based on the splicing statistical table; and determine the second correlation between the two independent variables based on the intra-variable product and the second historical correlation. In some examples, when the first processing module 13 determines the second correlation between the two independent variables based on the intra-variable product and the second historical correlation, the first processing module 13 is configured to: determine a weight coefficient corresponding to the second historical correlation, where the weight coefficient is greater than 0 and less than or equal to 1; determine the product value of the weight coefficient and the second historical correlation; and determine the sum of the intra-variable product and the product value as the second correlation. In some instances, when the first processing module 13 determines updated model parameters based on the first degree of association and the second degree of association, the first processing module 13 is configured to: determine first matrix information based on the first degree of association; determine second matrix information based on the second degree of association; and determine updated model parameters based on the first matrix information and the second matrix information. In some instances, when the first processing module 13 determines updated model parameters based on the first matrix information and the second matrix information, the first processing module 13 is configured to: obtain an inverse matrix of the second matrix information; multiply the inverse matrix by the first matrix information to obtain a product matrix; and determine updated model parameters based on the product matrix.In some instances, when the first processing module 13 performs a health check on the cloud platform based on the updated statistical model and obtains a health check result, the first processing module 13 is configured to: obtain an actual detection value of a target operating indicator used to implement the health check operation, where the target operating indicator is determined by operating indicators in at least one dimension; obtain an indicator prediction value corresponding to the target operating indicator based on the updated statistical model and the current multi-dimensional indicators; determine the standard deviation of the cloud platform corresponding to the target operating indicator in the previous cycle based on the spliced ​​statistical table; and perform a health check on the cloud platform based on the actual detection value, the indicator prediction value, and the standard deviation to obtain a health check result. In some instances, when the first processing module 13 performs a health check on the cloud platform based on the actual detection value, the indicator prediction value, and the standard deviation to obtain a health check result, the first processing module 13 is configured to: determine a model prediction error corresponding to the updated statistical model based on the actual detection value and the indicator prediction value; determine a preset detection coefficient for performing the health check on the cloud platform; and perform a health check on the cloud platform based on the model prediction error, the preset detection coefficient, and the standard deviation to obtain a health check result. In some examples, when the first processing module 13 performs a health check on the cloud platform based on the model prediction error, the preset detection coefficient, and the standard deviation to obtain a health check result, the first processing module 13 is configured to: determine a product value of the preset detection coefficient and the standard deviation; and perform a health check on the cloud platform based on the model prediction error and the product value to obtain a health check result. In some examples, when the first processing module 13 performs a health check on the cloud platform based on the model prediction error and the product value to obtain a health check result, the first processing module 13 is configured to: determine the health check result as a first check result indicating abnormal operation of the cloud platform if the model prediction error is greater than the product value; and determine the health check result as a second check result indicating normal operation of the cloud platform if the model prediction error is less than or equal to the product value. In some instances, when the first processing module 13 performs a health check on the cloud platform based on the updated statistical model and obtains a health check result, the first processing module 13 is configured to: obtain an actual detection value of a single-dimensional detection indicator used to perform health check on the cloud platform; determine, based on the updated statistical model and the current multi-dimensional indicators, an indicator average value and a standard deviation corresponding to the single-dimensional detection indicator of the cloud platform in the current cycle; and perform a health check on the cloud platform based on the actual detection value, the indicator average value, and the standard deviation to obtain a health check result.In some instances, when the first processing module 13 performs a health check on the cloud platform based on the actual detection value, the indicator average, and the standard deviation to obtain a health check result, the first processing module 13 is configured to: obtain a preset detection coefficient corresponding to the standard deviation; determine the product value of the indicator deviation between the actual detection value and the indicator average, the preset detection coefficient, and the standard deviation; and perform a health check on the cloud platform based on the indicator deviation and the product value to obtain a health check result. In some instances, when the first processing module 13 performs a health check on the cloud platform based on the indicator deviation and the product value to obtain a health check result, the first processing module 13 is configured to: determine, if the indicator deviation is greater than the product value, that the health check result is a first detection result indicating abnormal operation of the cloud platform; and determine, if the indicator deviation is less than or equal to the product value, that the health check result is a second detection result indicating normal operation of the cloud platform. In some examples, after obtaining a health check result, the first acquisition module 11 and the first processing module 13 in this embodiment are respectively configured to perform the following steps: the first acquisition module 11 is configured to obtain result feedback information corresponding to the health check result; the first processing module 13 is configured to adjust a preset detection coefficient based on the result feedback information to obtain an adjusted detection coefficient. In some examples, when the first processing module 13 adjusts the preset detection coefficient based on the result feedback information to obtain the adjusted detection coefficient, the first processing module 13 is configured to: determine a detection false alarm rate and a detection missed alarm rate corresponding to the preset detection coefficient based on the result feedback information; determine a false alarm rate weight corresponding to the detection false alarm rate and a missed alarm rate weight corresponding to the detection missed alarm rate; determine a detection false alarm value based on the detection false alarm rate and the false alarm rate weight; determine a detection missed alarm value based on the missed alarm rate weight and the detection missed alarm rate; and adjust the preset detection coefficient based on the detection false alarm value and the missed alarm value to obtain the adjusted detection coefficient. In some examples, when the first processing module 13 adjusts the preset detection coefficient based on the false positive detection value and the missed detection value to obtain the adjusted detection coefficient, the first processing module 13 is configured to: increase the preset detection coefficient to obtain the adjusted detection coefficient when the false positive detection value is greater than the missed detection value; decrease the preset detection coefficient to obtain the adjusted detection coefficient when the false positive detection value is less than the missed detection value; and maintain the preset detection coefficient unchanged when the false positive detection value is equal to the missed detection value. The apparatus shown in FIG8 can execute the method of the embodiments shown in FIG1-FIG7. For portions not described in detail in this embodiment, reference is made to the relevant description of the embodiments shown in FIG1-FIG7. The implementation process and technical effects of this technical solution are described in the description of the embodiments shown in FIG1-FIG7 and will not be repeated here.In one possible design, the structure of the cloud platform health monitoring device shown in FIG8 can be implemented as an electronic device, which can be a controller, a personal computer, a server, or other device. As shown in FIG9 , the electronic device may include a first processor 21 and a first memory 22. The first memory 22 is used to store a program that enables the electronic device to execute the cloud platform health monitoring method provided in the embodiments shown in FIG1-7 . The first processor 21 is configured to execute the program stored in the first memory 22. The program includes one or more computer instructions, wherein when executed by the first processor 21, the one or more computer instructions can implement the following steps: obtaining the current multi-dimensional indicators corresponding to the cloud platform in the current cycle; determining a spliced ​​statistical table corresponding to the operating indicators of the cloud platform in the previous cycle, wherein the spliced ​​statistical table includes at least: indicator statistical characteristics obtained by statistically processing the historical multi-dimensional indicators of the cloud platform in the previous cycle, wherein the indicator statistical characteristics include at least: mean value, standard deviation, and number of statistical indicators; determining a statistical model for implementing a health check operation based on the spliced ​​statistical table; and performing a health check on the cloud platform based on the spliced ​​statistical table, the statistical model, and the current multi-dimensional indicators to obtain a health check result. Furthermore, the first processor 21 is further configured to execute all or part of the steps in the embodiments shown in Figures 1-7 above. The electronic device may also include a first communication interface 23 for communicating with other devices or a communication network. Furthermore, embodiments of the present disclosure provide a computer storage medium for storing computer software instructions used by the electronic device, which includes a program for executing the cloud platform health check method in the embodiments shown in Figures 1-7 above. Furthermore, embodiments of the present disclosure provide a computer program product comprising: a computer-readable storage medium storing computer instructions. When the computer instructions are executed by one or more processors, the one or more processors are caused to perform the steps of the cloud platform health detection method in the method embodiments shown in Figures 1 to 7 . The apparatus embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, i.e., they may be located in one location or distributed across multiple network elements. Some or all of the modules may be selected based on actual needs to achieve the objectives of the solutions of this embodiment. Persons of ordinary skill in the art can understand and implement the present invention without inventive effort.Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented by adding the necessary general-purpose hardware platform, or of course, by a combination of hardware and software. Based on this understanding, the essence of the above technical solution, or the portion that contributes to the prior art, can be embodied in the form of a computer product. The present disclosure can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code. The present disclosure is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present disclosure. It should be understood that each process and / or block in the flowcharts and / or block diagrams, as well as combinations of processes and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable device to generate a machine, such that the instructions executed by the processor of the computer or other programmable device generate means for implementing the functions specified in one or more processes in the flowcharts and / or one or more blocks in the block diagrams. These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable device to operate in a specific manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means that implement the functions specified in one or more flows in a flowchart and / or one or more blocks in a block diagram. These computer program instructions may also be loaded onto a computer or other programmable device, causing the computer or other programmable device to execute a series of operational steps to produce a computer-implemented process, such that the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in one or more flows in a flowchart and / or one or more blocks in a block diagram. In a typical configuration, a computing device includes one or more processors (CPUs), input / output interfaces, network interfaces, and memory. The memory may include non-permanent storage in a computer-readable medium, random access memory (RAM), and / or non-volatile memory, such as read-only memory (ROM) or flash memory. OMemory is an example of computer-readable media. Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can implement information storage using any method or technology. Information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase-change RAM (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 technologies, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transmitting medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves. Finally, it should be noted that the above embodiments are merely illustrative of the technical solutions of the present disclosure and are not intended to limit the present disclosure. Although the present disclosure has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they may modify the technical solutions described in the aforementioned embodiments or replace some of the technical features therein with equivalents. However, such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present disclosure.

Claims

Claims 1. A cloud platform health detection method, comprising: Obtain the current multi-dimensional indicators corresponding to the cloud platform in the current cycle; Determining a concatenated statistical table corresponding to operating indicators of the cloud platform in a previous cycle, the concatenated statistical table including at least indicator statistical characteristics obtained by statistically processing historical multi-dimensional indicators of the cloud platform in the previous cycle, the indicator statistical characteristics including at least an average value, a standard deviation, and a number of statistical indicators; determining a statistical model for implementing a health check operation based on the concatenated statistical table; and performing a health check on the cloud platform based on the concatenated statistical table, the statistical model, and the current multi-dimensional indicators to obtain a health check result.

2. The method according to claim 1, wherein determining a splicing statistical table corresponding to an operating indicator of the cloud platform in a previous cycle comprises: Acquiring historical multi-dimensional indicators corresponding to the cloud platform in a previous period; aggregating the historical multi-dimensional indicators to obtain multiple aggregated statistical tables; A splicing process is performed on the multiple aggregation statistical tables to obtain a splicing statistical table, where the splicing statistical table further includes: a preset splicing dimension and a splicing dimension value corresponding to the preset splicing dimension.

3. The method according to claim 2, aggregating the historical multi-dimensional indicators to obtain a plurality of aggregated statistical tables, comprising: Obtaining preset aggregation dimensions for implementing an aggregation operation, where the preset aggregation dimensions include at least a time period, a return code, and one or two replaceable dimensions, where the replaceable dimensions are obtained by traversing and replacing the remaining dimensions of the historical multi-dimensional indicators except the time period and the return code; and aggregating the historical multi-dimensional indicators based on the preset aggregation dimensions to obtain multiple aggregation statistical tables.

4. The method according to claim 2 or 3, wherein the step of concatenating multiple aggregation statistical tables to obtain a concatenated statistical table comprises: Get replaceable dimensions in multiple aggregate statistics tables; Adjusting multiple aggregate statistical tables based on the replaceable dimensions to obtain adjusted statistical tables with a unified structure; Vertically splicing the adjusted statistical tables to obtain the spliced ​​table; Performing data statistical operations on the operating indicators in the multiple aggregate statistical tables to obtain indicator statistical features, wherein the indicator statistical features further include: maximum value and minimum value; The indicator statistical features are added to the splicing table to obtain the splicing statistical table.

5. The method according to claim 4, wherein the step of adjusting a plurality of aggregate statistical tables based on the replaceable dimension to obtain an adjusted statistical table with a unified structure comprises: Based on the replaceable dimension, a dimension name for concatenating multiple aggregation statistics tables is added to the aggregation statistics table; Based on the aggregate statistical table, determining a dimension value corresponding to the replaceable dimension; and adjusting each aggregate statistical table based on the dimension name and the dimension value to obtain an adjusted statistical table with a unified structure.

6. The method according to any one of claims 1 to 5, performing a health check on the cloud platform based on the splicing statistical table, the statistical model, and the current multi-dimensional indicators to obtain a health check result, comprising: updating the statistical model based on the splicing statistical table and the current multi-dimensional indicators to obtain an updated statistical model; Perform a health check on the cloud platform based on the updated statistical model to obtain a health check result.

7. The method according to claim 6, wherein the step of updating the statistical model based on the splicing statistical table and the current multi-dimensional index to obtain an updated statistical model comprises: Obtaining a dependent variable for performing health detection on the cloud platform, wherein the dependent variable corresponds to at least one independent variable; Based on the splicing statistical table and the current multidimensional index, determine a first degree of correlation between the dependent variable and the independent variable, and a second degree of correlation between the independent variables; based on the first degree of correlation and the second degree of correlation, determine updated model parameters; based on the updated model parameters, determine the updated statistical model.

8. The method according to claim 7, wherein determining the first correlation between the dependent variable and the independent variable based on the splicing statistical table and the current multi-dimensional index comprises: Based on the current multi-dimensional indicator, determining the inter-variable product between the dependent variable and the independent variable in the current cycle; based on the splicing statistical table, determining the first historical correlation between the dependent variable and the independent variable in the previous cycle; A first degree of association between the dependent variable and the independent variable is determined based on the inter-variable product and the first historical degree of association.

9. The method according to claim 8, wherein determining the first correlation between the dependent variable and the independent variable based on the inter-variable product and the first historical correlation comprises: Determining a weight coefficient corresponding to the first historical association degree, where the weight coefficient is greater than 0 and less than or equal to 1; determining a product value between the weight coefficient and the first historical association degree; The sum of the product between the variables and the product value is determined as the first correlation degree.

10. The method according to any one of claims 7 to 9, wherein determining a second correlation between independent variables based on the splicing statistical table and the current multi-dimensional index comprises: Based on the current multi-dimensional indicator, the intra-variable product between the independent variables in the current period; Based on the splicing statistical table, a second historical correlation between the two independent variables in the previous period is determined; based on the intra-variable product and the second historical correlation, a second correlation between the two independent variables is determined.

11. The method according to claim 10, wherein determining the second correlation between two independent variables based on the intra-variable product and the second historical correlation comprises: Determining a weight coefficient corresponding to the second historical association degree, where the weight coefficient is greater than 0 and less than or equal to 1; determining a product value between the weight coefficient and the second historical association degree; The sum of the intra-variable product and the product value is determined as the second degree of association.

12. The method according to any one of claims 7 to 11, wherein determining updated model parameters based on the first degree of association and the second degree of association comprises: Determining first matrix information based on the first correlation degree; Determining second matrix information based on the second correlation degree; The updated model parameters are determined based on the first matrix information and the second matrix information.

13. The method according to claim 12, wherein determining the updated model parameters based on the first matrix information and the second matrix information comprises: 36 Obtaining an inverse matrix of the second matrix information; Multiplying the inverse matrix by the first matrix information to obtain a product matrix; Based on the product matrix, the updated model parameters are determined.

14. The method according to any one of claims 6 to 13, wherein the step of performing a health check on the cloud platform based on the updated statistical model to obtain a health check result comprises: Obtaining an actual detection value of a target operating indicator for implementing a health detection operation, wherein the target operating indicator is determined by an operating indicator of at least one dimension; Obtaining an indicator prediction value corresponding to the target operating indicator based on the updated statistical model and the current multi-dimensional indicator; determining a standard deviation of the cloud platform corresponding to the target operating indicator in a previous cycle based on the spliced ​​statistical table; Perform a health check on the cloud platform based on the actual detection value, the indicator prediction value, and the standard deviation to obtain a health check result.

15. The method according to claim 14, wherein the health check of the cloud platform is performed based on the actual detection value, the indicator prediction value, and the standard deviation to obtain a health check result, comprising: Determining a model prediction error corresponding to the updated statistical model based on the actual detection value and the indicator prediction value; Determining a preset detection coefficient for performing health detection on the cloud platform; A health check is performed on the cloud platform based on the model prediction error, the preset detection coefficient, and the standard deviation to obtain a health check result.

16. The method according to claim 15, wherein the performing health detection on the cloud platform based on the model prediction error, the preset detection coefficient, and the standard deviation to obtain the health detection result comprises: Determining a product value of the preset detection coefficient and the standard deviation; Perform a health check on the cloud platform based on the model prediction error and the product value to obtain a health check result.

17. The method according to claim 16, wherein the step of performing a health check on the cloud platform based on the model prediction error and the product value to obtain a health check result comprises: When the model prediction error is greater than the product value, determining that the health detection result is a first detection result for identifying abnormal operation of the cloud platform; When the model prediction error is less than or equal to the product value, the health detection result is determined to be a second detection result for identifying that the cloud platform operates normally.

18. The method according to claim 6, wherein the step of performing a health check on the cloud platform based on the updated statistical model to obtain a health check result comprises: Obtaining actual detection values ​​of single-dimensional detection indicators used to perform health detection on the cloud platform; Based on the updated statistical model and the current multi-dimensional indicator, determine the indicator average and standard deviation of the cloud platform corresponding to the single-dimensional detection indicator in the current cycle; perform a health check on the cloud platform based on the actual detection value, the indicator average and the standard deviation to obtain a health check result.

19. The method according to claim 18, wherein the performing health check on the cloud platform based on the actual detection value, the indicator average value, and the standard deviation to obtain a health check result comprises: Obtaining a preset detection coefficient corresponding to the standard deviation; Determine an indicator deviation between the actual detection value and the indicator average value, and a product value of the preset detection coefficient and the standard deviation; Perform a health check on the cloud platform based on the indicator deviation and the product value to obtain a health check result.

20. The method according to claim 19, wherein the performing a health check on the cloud platform based on the indicator deviation and the product value to obtain a health check result comprises: When the indicator deviation is greater than the product value, determining that the health detection result is a first detection result for identifying abnormal operation of the cloud platform; When the indicator deviation is less than or equal to the product value, the health detection result is determined to be a second detection result for indicating that the cloud platform operates normally.

21. The method according to claim 15 or 19, after obtaining the health test result, the method further comprises: Obtaining result feedback information corresponding to the health test result; The preset detection coefficient is adjusted based on the result feedback information to obtain an adjusted detection coefficient.

22. The method according to claim 21, wherein adjusting the preset detection coefficient based on the result feedback information to obtain the adjusted detection coefficient comprises: Based on the result feedback information, determining a detection false alarm rate and a detection missed alarm rate corresponding to the preset detection coefficient; Determining a false alarm rate weight corresponding to the detection false alarm rate and a missed alarm rate weight corresponding to the detection missed alarm rate; determining the detection false alarm value based on the detection false alarm rate and the false alarm rate weight; Determining the detection omission value based on the omission rate weight and the detection omission rate; The preset detection coefficient is adjusted based on the detection false alarm value and the detection omission value to obtain an adjusted detection coefficient.

23. The method according to claim 22, wherein adjusting the preset detection coefficient based on the false positive detection value and the missed detection detection value to obtain the adjusted detection coefficient comprises: When the detection false alarm value is greater than the detection omission value, the preset detection coefficient is increased to obtain an adjusted detection coefficient; When the detection false alarm value is less than the detection omission value, the preset detection coefficient is adjusted to be smaller to obtain an adjusted detection coefficient; When the detection false alarm value is equal to the detection omission value, the preset detection coefficient is kept unchanged.

24. An electronic device, comprising: A memory and a processor; wherein the memory is used to store one or more computer instructions, wherein the one or more computer instructions, when executed by the processor, implement the method of any one of claims 1 to 23.

25. A computer program product comprising: A computer program, which, when executed by a processor of an electronic device, causes the processor to perform the steps of the method of any one of claims 1 to 23.

26. A computer storage medium for storing a computer program, wherein the computer program enables a computer to implement the method of any one of claims 1 to 23 when executed.

Images