A method for monitoring computing resources in a multi-cloud environment

Abstract

The application discloses a kind of multi-cloud environment under computing resource monitoring method, specifically relates to resource monitoring technical field;It is by collecting the running index data of virtual computing resources in multi-cloud environment and physical host resource distribution data, constructs multi-cloud monitoring dataset;According to the resource adjacent topology relationship graph between virtual instance constructed by multi-cloud monitoring dataset, the load interference characteristic data of adjacent virtual instance is extracted;Identify the load fluctuation characteristics and mutation trend of target virtual instance simultaneously, generate the abnormal fluctuation identification data of target virtual instance;Based on load interference characteristic data and abnormal fluctuation identification data, establish the distinguishing model of target virtual instance real load and adjacent virtual instance interference load, adjust abnormal monitoring judgment logic according to the distinguishing result, output interference corrected monitoring state label data, to effectively distinguish real load change and adjacent instance load interference, improve the accuracy of resource monitoring and the reliability of resource scheduling decision.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of resource monitoring, and more particularly to a computing resource monitoring method in a multi-cloud environment. BACKGROUND

[0002] With the development and application of cloud computing technology, enterprises choose to deploy their business systems in a multi-cloud environment constructed by multiple different cloud service providers to obtain higher resource utilization and service reliability. In the multi-cloud environment, users usually obtain virtual computing resources through virtualization technology and perform specific tasks, but due to the great difference in the underlying physical host resource configuration strategies adopted by different cloud service providers, there is a potential possibility of mutual interference between virtual computing resources.

[0003] In the prior art, when monitoring the running state of virtual computing resources in a multi-cloud environment, the load interference influence between virtual computing resources due to sharing physical resources is not considered, and it is difficult to accurately distinguish the real load change in the virtual computing resource running state data and the interference fluctuation caused by adjacent virtual computing resources. SUMMARY

[0004] In order to overcome the above-mentioned defects of the prior art, the embodiments of the present application provide a computing resource monitoring method in a multi-cloud environment to solve the problems raised in the background art.

[0005] To achieve the above-mentioned purpose, the present application provides the following technical solutions:

[0006] A computing resource monitoring method in a multi-cloud environment, comprising the following steps:

[0007] S1: Collecting running index data and physical host resource distribution data of virtual computing resources deployed in a multi-cloud environment, and performing time alignment and normalization processing to generate a standardized multi-cloud monitoring data set;

[0008] S2: Based on the standardized multi-cloud monitoring data set, constructing a resource adjacency topology relationship graph between virtual instances;

[0009] S3: Based on the resource adjacency topology relationship graph, calculating a load similarity index between virtual instances, and extracting load interference feature data of adjacent virtual instances;

[0010] S4: Based on the running index change data extracted by the virtual instance in the multi-cloud monitoring data set, identifying the load fluctuation feature and mutation trend of the target virtual instance, and generating abnormal fluctuation identification data of the target virtual instance;

[0011] S5: Based on the load interference feature data and the abnormal fluctuation identification data, establishing a model for distinguishing the real load of the target virtual instance and the interference load of the adjacent virtual instance.

[0012] S6: According to the distinguishing result of the target virtual instance real load and the adjacent virtual instance interference load, adjusting the abnormal monitoring judgment logic of the multi-cloud computing resource, and outputting the interference corrected monitoring state label data.

[0013] In one preferred embodiment, S1, specifically:

[0014] Collecting running index data of virtual computing resources deployed in multiple different cloud service provider environments;

[0015] Collecting distribution data of physical host resources deployed in multiple different cloud service provider environments;

[0016] Calibrating the running index data of virtual computing resources and the distribution data of physical host resources to a unified time reference;

[0017] Normalizing the calibrated running index data of virtual computing resources and the distribution data of physical host resources to generate a standardized multi-cloud monitoring data set.

[0018] In one preferred embodiment, S2, specifically:

[0019] According to the standardized multi-cloud monitoring data set, grouping virtual instances and generating a mapping table of virtual instances and physical host resources;

[0020] Based on the mapping table, assigning a unique node number to each physical host resource, marking all virtual instances mapped to the same physical host resource node number as homolayer nodes, and marking all virtual instances mapped to different physical host resource node numbers as heterolayer nodes;

[0021] Establishing undirected edges between homolayer nodes according to shared processor core identifiers, shared memory channel identifiers, and shared input / output bus identifiers, and not establishing edges between heterolayer nodes, to generate a resource adjacency topology relationship graph between virtual instances.

[0022] In one preferred embodiment, S3, specifically:

[0023] Based on the resource adjacency topology relationship graph, extracting processor usage rate data, memory usage rate data, and network bandwidth usage rate data of each node of each adjacent virtual instance node pair from the standardized multi-cloud monitoring data set;

[0024] Calculating the similarity of processor usage rate data, the similarity of memory usage rate data, and the similarity of network bandwidth usage rate data between each adjacent virtual instance node pair, and performing weighted summation to obtain a load similarity index between each adjacent virtual instance node pair;

[0025] extracting a pair of adjacent virtual instance nodes with a load similarity index exceeding a preset similarity threshold, and generating load interference feature data of the adjacent virtual instance.

[0026] In one preferred embodiment, S4, specifically:

[0027] Based on the standardized multi-cloud monitoring dataset, the node identification information of the target virtual instance is determined.

[0028] According to the node identification information of the target virtual instance, the disk input / output operation rate data, disk delay data and processor usage rate data of the target virtual instance within the continuous monitoring time window are extracted from the standardized multi-cloud monitoring dataset.

[0029] The change rates of the disk input / output operation rate data, disk delay data and processor usage rate data within the continuous monitoring time window are calculated respectively, and the fluctuation features corresponding to the change rates are identified.

[0030] The mutation trend of each fluctuation feature is detected, and abnormal fluctuation identification data of the target virtual instance is generated.

[0031] In one preferred embodiment, S5, specifically:

[0032] According to the load interference feature data of the adjacent virtual instance, the interference intensity level of the pair of adjacent virtual instance nodes is determined.

[0033] According to the abnormal fluctuation identification data of the target virtual instance, the abnormal fluctuation level of the target virtual instance is determined.

[0034] Based on the interference intensity level and the abnormal fluctuation level, a distinguishing model of the real load of the target virtual instance and the interference load of the adjacent virtual instance is established.

[0035] Based on the distinguishing model, each abnormal fluctuation feature in the abnormal fluctuation identification data of the target virtual instance is classified and determined, and it is determined that each abnormal fluctuation feature belongs to the real load change of the target virtual instance or belongs to the interference load change of the adjacent virtual instance.

[0036] In one preferred embodiment, S6, specifically:

[0037] When the target virtual instance load fluctuation category is determined as the interference load change of the adjacent virtual instance, the abnormal fluctuation identification data of the target virtual instance in the standardized multi-cloud monitoring dataset is marked and corrected to a non-abnormal monitoring state label.

[0038] When the target virtual instance load fluctuation category is determined as the real load change of the target virtual instance, the abnormal fluctuation identification data of the target virtual instance in the standardized multi-cloud monitoring data set is marked and determined as an abnormal monitoring state label;

[0039] The standardized multi-cloud monitoring data set is updated based on the non-abnormal monitoring state label and the abnormal monitoring state label, and the corrected target virtual instance monitoring state label data is output.

[0040] The technical effects and advantages of the computing resource monitoring method in a multi-cloud environment are as follows:

[0041] Through the construction of the standardized multi-cloud monitoring data set, unified management of cross-cloud environment data is realized, and the monitoring deviation caused by clock difference and inconsistent dimensions is eliminated; by constructing the resource adjacency topology relationship graph, the physical resource sharing relationship between virtual instances can be accurately described; by calculating the load similarity index and extracting the load interference feature data, the monitoring abnormality caused by the load fluctuation of adjacent virtual instances can be identified; by identifying the load fluctuation characteristics and mutation trend of the target virtual instance, the real load change and interference fluctuation can be effectively separated; by introducing the distinction model based on the interference intensity level and the abnormal fluctuation level, the separation determination of the real load change and the adjacent interference load is realized; according to the distinction result, the abnormal monitoring judgment logic of the multi-cloud computing resource is dynamically adjusted, the interference corrected monitoring state label data is output, the precision and stability of resource monitoring are significantly improved, and reliable data support is provided for multi-cloud resource scheduling and capacity planning. BRIEF DESCRIPTION OF DRAWINGS

[0042] Figure 1 A schematic diagram of the computing resource monitoring method in a multi-cloud environment is provided. DETAILED DESCRIPTION

[0043] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0044] EMBODIMENT

[0045] Figure 1 A computing resource monitoring method in a multi-cloud environment is provided, which includes the following steps:

[0046] S1: Collecting the running index data and physical host resource distribution data of the virtual computing resources deployed in the multi-cloud environment, and performing time alignment and normalization processing to generate a standardized multi-cloud monitoring data set;

[0047] S2: Construct a resource adjacency topology graph between virtual instances based on the standardized multi-cloud monitoring dataset;

[0048] S3: Calculate the load similarity index between virtual instances based on the resource adjacency topology graph, and extract the load interference feature data of adjacent virtual instances;

[0049] S4: Based on the running index change data extracted by the virtual instance in the multi-cloud monitoring dataset, identify the load fluctuation characteristics and mutation trend of the target virtual instance, and generate abnormal fluctuation identification data of the target virtual instance;

[0050] S5: Based on the load interference feature data and abnormal fluctuation identification data, establish a distinction model of the target virtual instance real load and adjacent virtual instance interference load;

[0051] S6: According to the distinction result of the target virtual instance real load and adjacent virtual instance interference load, adjust the abnormal monitoring judgment logic of multi-cloud computing resources, and output the interference corrected monitoring state label data.

[0052] S1: Collect running index data and physical host resource distribution data of virtual computing resources deployed in a multi-cloud environment, and perform time alignment and normalization processing to generate a standardized multi-cloud monitoring dataset, including:

[0053] Collect running index data of virtual computing resources deployed in multiple different cloud service provider environments;

[0054] Multi-cloud environment refers to the deployment of computing resources in two or more different cloud computing environments provided by different cloud service providers, and the cloud computing environment is maintained and managed by different cloud service providers; Virtual computing resources refer to virtual servers, virtual network storage devices, and virtual containers established through virtualization technology in a cloud computing environment. Each virtual device can independently execute user-assigned computing tasks; Running index data is data obtained by quantitatively measuring the running state of virtual computing resources, such as processor usage percentage, memory usage percentage, network traffic throughput speed, disk read / write operation frequency, and delay time. Real-time continuous acquisition of each running index data is performed to reflect the current running state of the virtual server.

[0055] Collect distribution data of physical host resources deployed in multiple different cloud service provider environments;

[0056] The distribution data of physical host resources is the resource configuration and allocation of physical servers, physical storage devices, and network interface devices that actually carry and run virtual computing resources in a cloud computing environment. For example, the total number of central processor cores and the current occupation of each core, the capacity and actual used capacity of physical memory modules, the current actual available network bandwidth and bandwidth allocation of each network interface device, and the total capacity, remaining available capacity, and used capacity of storage devices. The distribution data of physical host resources also needs to be collected in real time. For example, continuing the above example of the two cloud service provider environments deployed by the enterprise, for each physical server in the two cloud service provider environments of the enterprise, the total number of processor cores, the used processor cores, the total memory capacity, the used memory capacity, the total disk capacity, the remaining capacity, and the available bandwidth and used bandwidth of the network interface of each physical server are collected one by one, so as to obtain complete and accurate physical resource distribution.

[0057] The running index data of virtual computing resources and the distribution data of physical host resources are calibrated with a unified time reference.

[0058] The unified time reference calibration refers to the unified time alignment processing of the running index data of virtual computing resources and the distribution data of physical host resources with the same standard clock frequency as the reference. Specifically, first, determine the unified clock standard, for example, take the standard clock signal of coordinated universal time or universal time protocol as the time reference. Record the original time stamp corresponding to each data collection for each running index data of virtual computing resources and each distribution data of physical host resources. Then, calculate the time difference between each data original time stamp and the unified standard clock signal, and use the time difference as the offset to adjust the time stamp of each running index data and physical host resource distribution data, so that all data have a unified time reference.

[0059] The calibrated running index data of virtual computing resources and the distribution data of physical host resources are normalized to generate a standardized multi-cloud monitoring data set.

[0060] The normalization processing refers to the uniform conversion of the scales of all running index data and resource distribution data after time reference calibration, so that each data is converted into the same value range. Specifically, first, the value range of each type of data is determined, for example, the maximum and minimum possible values of the processor usage rate data, the maximum and minimum possible values of the memory usage rate data, the maximum and minimum possible values of the disk read / write speed data, etc. are determined. The original value of each data is subtracted by the minimum possible value in the value range of the data, and divided by the difference between the maximum possible value and the minimum possible value, to obtain a uniform standardized data value, so that the standardized data value is between zero and one. The normalization calculation is performed on all running index data and physical host resource distribution data, so as to obtain a standardized multi-cloud monitoring data set with uniform scale.

[0061] S2: Based on the standardized multi-cloud monitoring data set, a resource adjacency topology relationship graph between virtual instances is constructed, including:

[0062] According to the standardized multi-cloud monitoring data set, the virtual instances are grouped and a mapping table of virtual instances and physical host resources is generated;

[0063] The virtual instance is a virtual device such as a virtual server, a virtual network storage device or a virtual container that can independently complete a computing task through virtualization technology. The physical host resource is a physical server, a storage device and a network interface device that actually deploys the virtual instance. According to the standardized multi-cloud monitoring data set, the virtual instances are grouped according to the distribution of the physical host resources. It is assumed that the enterprise deploys ten virtual servers in the first cloud computing service provider environment and fifteen virtual servers in the second cloud computing service provider environment. Each virtual server runs on different physical host resources in the respective cloud computing service provider environment. The physical host resource location corresponding to each virtual server is determined through the physical host resource distribution data in the standardized multi-cloud monitoring data set, forming a one-to-one grouping relationship. Each virtual server is mapped to the physical server that carries it. For example, the ten virtual servers in the first cloud computing service provider environment are mapped to four different physical servers, and the fifteen virtual servers in the second cloud computing service provider environment are mapped to five different physical servers. A mapping relationship table is generated, for example:

[0064]

[0065] Based on the mapping table, a unique node number is allocated to each physical host resource. All virtual instances mapped to the same physical host resource node number are marked as same-layer nodes, and all virtual instances mapped to different physical host resource node numbers are marked as different-layer nodes.

[0066] Based on the mapping table of virtual instances and physical host resources, each physical host resource is sequentially assigned a node number with uniqueness and no repetition. The node number is a mark for representing different physical host resources, and through the node number, different physical host resources can be distinguished. For example, taking four physical servers in the first cloud computing service provider environment and five physical servers in the second cloud computing service provider environment as an example, each physical server sequentially obtains a node number different from each other, wherein the first physical server obtains the first node number, the second physical server obtains the second node number, and so on, until the last physical server obtains the last node number. According to the node number of the physical host resource mapped by each virtual instance, all virtual instances are divided into two types of nodes: if multiple virtual instances are mapped to the same node number, the virtual instance is defined as a same-layer node, and the same-layer node refers to a virtual instance set running on the same physical server; if multiple virtual instances are mapped to different node numbers, they are defined as different-layer nodes, and the different-layer node refers to a virtual instance set not running on the same physical server.

[0067] A non-directed edge is established between the same-layer nodes according to the shared processor core identifier, the shared memory channel identifier and the shared input / output bus identifier, and no edge is established between the different-layer nodes, thereby generating a resource adjacency topology relationship graph between the virtual instances.

[0068] The physical resource sharing relationship between the virtual instances is expressed in the form of a topology graph. The nodes in the topology graph are the same layer or different layer virtual instances marked, and each node represents a virtual instance. The edges in the topology graph represent the resource sharing relationship between two virtual instances. The undirected edge does not contain directionality and only represents the association relationship between the nodes with resource sharing. The basis for establishing an undirected edge between nodes is the shared processor core identifier, the shared memory channel identifier, and the shared input / output bus identifier, which respectively represent the shared resource markers of the processor, the memory, and the input / output device. If multiple virtual instances use the same processor core, the same memory channel, or the same input / output bus, an undirected edge is established between the same layer nodes representing the virtual instances. For example, taking four physical servers in the first cloud computing service provider environment and five physical servers in the second cloud computing service provider environment as an example, assuming that a certain physical server in the first cloud computing service provider environment contains four processor cores, each core has an independent identifier, and two virtual servers simultaneously occupy a processor core, i.e., occupy the same identifier of the processor core, an undirected edge is established between the nodes representing the two virtual servers. Similarly, if another three virtual servers occupy the same memory channel or input / output bus, corresponding undirected edges are established between the nodes. In contrast, no edge is established between different layer nodes mapped to different node numbers, indicating that there is no resource sharing relationship between the virtual instances. Through the above method, the real physical resource sharing relationship between the virtual instances can be expressed, thereby generating a resource adjacency topology relationship graph between the virtual instances.

[0069] S3: Based on the resource adjacency topology relationship graph, the load similarity index between the virtual instances is calculated, and the load interference feature data of adjacent virtual instances is extracted, including:

[0070] Based on the resource adjacency topology relationship graph, the processor usage rate data, the memory usage rate data, and the network bandwidth usage rate data of each node of each adjacent virtual instance node pair are extracted from the standardized multi-cloud monitoring data set;

[0071] The resource adjacency topology relationship diagram is a topology structure diagram generated according to the grouping mapping relationship of virtual instances and physical host resources, each node represents a virtual instance, and an undirected edge represents the physical resource relationship shared between two virtual instances; a node pair refers to two virtual instance nodes that have a sharing relationship and are connected by an undirected edge, and is referred to as an adjacent virtual instance node pair; for example, an enterprise has two physical servers in a first cloud computing service provider environment, each of which runs two virtual servers, so there are a total of four virtual servers, forming a plurality of same-layer node pairs; for each adjacent virtual instance node pair, running index data is extracted from the standardized multi-cloud monitoring data set, including processor usage rate data, memory usage rate data, and network bandwidth usage rate data of each virtual instance; the processor usage rate data refers to the proportion of virtual instance current occupation of processor computing resources, the memory usage rate data refers to the proportion of virtual instance occupation of memory space resources, and the network bandwidth usage rate data refers to the proportion of virtual instance current use of network resources.

[0072] The similarity of the processor usage rate data, the similarity of the memory usage rate data, and the similarity of the network bandwidth usage rate data between each adjacent virtual instance node pair are calculated, and a weighted sum is performed to obtain a load similarity index between each adjacent virtual instance node pair;

[0073] The similarity is used to quantify the consistency between the running index data change trend and the use characteristics of the adjacent virtual instance node pair; the similarity of the processor usage rate data refers to the similarity between the change trends of the processor usage rate data of the two virtual instances at the same time; the similarity of the memory usage rate data refers to the similarity of the memory resource usage trends of the two virtual instances; the similarity of the network bandwidth usage rate data refers to the similarity between the network bandwidth usage trends of the two virtual instances; the calculation method is as follows: first, taking the processor usage rate data as an example, the square value of the difference of the processor usage rate data between the two virtual instances at each continuous monitoring time point is calculated, and the square values of all the differences are accumulated to obtain the total square difference value; then, the total square difference value of the processor usage rate data is divided by the square value of the total change range of the corresponding processor usage rate data to obtain the difference proportion value of the processor usage rate data; then, the difference proportion value is subtracted from the numerical value one to obtain the similarity of the processor usage rate data, and the closer the similarity is to one, the more similar the processor usage trends of the two virtual instances are, and the closer the similarity is to zero, the more dissimilar they are; similarly, the similarity of the memory usage rate data and the similarity of the network bandwidth usage rate data are calculated; the three similarities are respectively assigned different weight values, and the determination of the weight values depends on the importance of the similarity in load analysis, and the similarity of the processor usage rate data is usually assigned a higher weight, and the similarities of the memory usage rate data and the network bandwidth usage rate data are respectively assigned lower weights; the similarity of the processor usage rate data is multiplied by the corresponding weight value to obtain the weight similarity of the processor usage rate data; the weight similarity of the memory usage rate data and the weight similarity of the network bandwidth usage rate data are respectively calculated; finally, the three weight similarities are added to obtain the load similarity index.

[0074] Extracting the adjacent virtual instance node pairs with the load similarity index exceeding the preset similarity threshold value to generate the load interference characteristic data of the adjacent virtual instances.

[0075] The preset similarity threshold value is a critical value for judging the size of the load similarity index, and the setting of the critical value depends on specific business needs and experience accumulation; if the load similarity index of the adjacent virtual instance node pair exceeds the preset similarity threshold value, it indicates that there is a correlation and similarity between the running index trends of the virtual instances, and there may be mutual influence and interference between the loads, so as to be marked as having the load interference characteristic; all node pairs with the load similarity index exceeding the preset similarity threshold value are extracted to form a load interference characteristic data set.

[0076] S4: Based on the running index change data extracted from the multi-cloud monitoring data set of the virtual instances, the load fluctuation characteristic and the mutation trend of the target virtual instance are identified, and the abnormal fluctuation identification data of the target virtual instance is generated, including:

[0077] Determine the node identification information of the target virtual instance based on the standardized multi-cloud monitoring dataset;

[0078] The target virtual instance is a specific virtual computing resource currently focused on in monitoring analysis, and the node identification information is characteristic data or attribute information that can uniquely mark and distinguish the location of the target virtual instance in the standardized multi-cloud monitoring dataset. For example, an enterprise deploys multiple virtual servers in a first cloud computing service provider environment, each virtual server has a unique identifier, the identifier of virtual server one is a specific string code combination, and the identifier of virtual server two is another different string code combination. Through the standardized multi-cloud monitoring dataset, according to monitoring requirements and task requirements, the specific target virtual server currently required for monitoring analysis, i.e., the unique node identification information of the target virtual instance, is determined. For example, the current monitoring task needs to monitor a specific virtual server of the enterprise in the first cloud computing service provider environment, and the unique identifier code of the specific virtual server is used to mark the specific virtual server as the target virtual instance, so that the monitoring analysis has a determined object.

[0079] According to the node identification information of the target virtual instance, extract the disk input / output operation rate data, disk latency data, and processor usage rate data of the target virtual instance in the continuous monitoring time window from the standardized multi-cloud monitoring dataset;

[0080] The continuous monitoring time window refers to a continuous and fixed length time period, and the data related to the running state of the target virtual instance is continuously and real-time collected in the time period. The disk input / output operation rate data refers to the number of disk read / write operations per unit time of the target virtual instance in the continuous monitoring time window. The disk latency data refers to the average waiting time for completing disk read / write requests of the target virtual instance in the continuous monitoring time window. The processor usage rate data refers to the proportion of processor computing resources occupied by the target virtual instance in the continuous monitoring time window. Through the standardized multi-cloud monitoring dataset, according to the node identification information, the disk input / output operation rate data, disk latency data, and processor usage rate data of the target virtual instance in each continuous monitoring time window are sequentially extracted from the standardized multi-cloud monitoring dataset. For example, the current continuous monitoring time window is a fixed length continuous time interval, and the disk input / output operation rate data of the target virtual instance in the time interval records the number of disk operations per unit time, the disk latency data records the average waiting time for disk operation requests, and the processor usage rate data records the proportion of processor resources used by the target virtual instance per unit time to the entire processor resource capacity.

[0081] Calculate the change rate of the disk input / output operation rate data, the disk latency data, and the processor usage rate data in the continuous monitoring time window, and identify the fluctuation characteristics corresponding to the change rate;

[0082] The change rate is a quantitative description of the change trend and amplitude of each running index data of the target virtual instance in the continuous monitoring time window. Taking the disk input / output operation rate data as an example, the difference between the disk input / output operation rate data in each continuous monitoring time window at the continuous adjacent time points is calculated, and the relative change rate of the disk input / output operation rate data is obtained by dividing the difference between the adjacent two time points by the disk input / output operation rate corresponding to the previous time point, which represents the change amplitude and trend of the disk operation rate of the target virtual instance. The relative change rates of the disk delay data and the processor usage rate data are calculated in the same way. The trend characteristics of each type of data change rate are identified, specifically, the rules, peak values and fluctuation characteristics of each data change rate in the continuous monitoring time window are analyzed. For example, taking the disk delay data of the target virtual instance as an example, first, the difference between the disk delay data at every two continuous adjacent time points in the continuous monitoring time window is calculated, and then the proportion of the difference value to the disk delay data at the previous time point is calculated to obtain the disk delay data change rate, and the fluctuation trend of the disk delay data change rate in the entire continuous monitoring time window is observed to determine whether there is a fluctuation peak, a change period or an abnormal change phenomenon, so as to determine the fluctuation characteristics of the disk delay data change rate. Similarly, the processor usage rate data and the disk input / output operation rate data are also calculated and analyzed according to the above method to obtain the corresponding change rate fluctuation characteristics, thereby completing the identification of the fluctuation characteristics of different index data of the target virtual instance.

[0083] For each fluctuation characteristic, a mutation trend detection is performed to generate abnormal fluctuation identification data of the target virtual instance.

[0084] The mutation trend detection is to determine whether there is an abnormal data segment in the change rate of each running index data of the target virtual instance, which is different from the general trend and changes sharply in a short time. First, the overall change rule of each fluctuation characteristic in the continuous monitoring time window is analyzed. The average change trend of each data change rate in the continuous monitoring time window is calculated. Then, according to the difference value between the change trend and the actual fluctuation change rate, when the change trend difference value exceeds the normal range or exceeds a certain critical standard, it is considered that an abnormal mutation trend occurs, and the time period and the corresponding index of the mutation trend are recorded and marked. For example, taking the processor usage rate data of the target virtual instance as an example, first, the average change trend of the processor usage rate data change rate in the continuous monitoring time window is calculated. If the processor usage rate data change rate exceeds the reasonable range of the average trend in a continuous short time period, the processor usage rate data in the short time period is marked as mutation trend data, which is considered as abnormal fluctuation. Similarly, the mutation trend detection is also performed on the disk input / output operation rate data and the disk delay data. Finally, all the abnormal mutation trend data detected in each index data of the target virtual instance are uniformly marked, summarized and arranged to form the abnormal fluctuation recognition data.

[0085] S5: Based on the load interference feature data and the abnormal fluctuation recognition data, a distinguishing model of the real load of the target virtual instance and the interference load of the adjacent virtual instance is established, including:

[0086] According to the load interference feature data of the adjacent virtual instance, the interference intensity level of the adjacent virtual instance node pair is determined.

[0087] The load interference feature data refers to data representing the mutual interference load condition between virtual instances, which is extracted by calculating the load similarity index between each pair of adjacent virtual instance nodes and according to a preset similarity threshold; the pair of adjacent virtual instance nodes refers to two virtual instances connected with each other through the identifier of a shared physical resource in a resource adjacency topology relationship graph; the interference intensity level is a grading standard for describing the degree of mutual interference influence between the pair of adjacent virtual instance nodes due to the shared physical resource; the interference intensity level is divided into multiple levels, such as a mild interference level, a moderate interference level, and a severe interference level; the interference intensity level is determined based on the load similarity index of the pair of adjacent virtual instance nodes and the preset similarity threshold; when the difference between the load similarity index of the pair of adjacent virtual instance nodes and the preset similarity threshold accounts for between zero percent and ten percent of the preset similarity threshold, the interference intensity level is determined as the mild interference level; when the difference between the load similarity index of the pair of adjacent virtual instance nodes and the preset similarity threshold accounts for between ten percent and thirty percent of the preset similarity threshold, the interference intensity level is determined as the moderate interference level; when the difference between the load similarity index of the pair of adjacent virtual instance nodes and the preset similarity threshold exceeds thirty percent of the preset similarity threshold, the interference intensity level is determined as the severe interference level.

[0088] determining the abnormal fluctuation level of the target virtual instance according to the abnormal fluctuation identification data of the target virtual instance;

[0089] The abnormal fluctuation identification data includes abnormal mutation trend identification data of the disk input / output operation rate, disk delay, and processor usage rate data during the running of the target virtual instance;

[0090] The determination method of the abnormal fluctuation level is:

[0091] statistically analyzing the abnormal fluctuation identification data of the target virtual instance, wherein the abnormal fluctuation identification data includes the number of mutations and the corresponding amplitude value of each mutation of the processor usage rate, disk input / output operation rate, disk delay, and other running indicators of the target virtual instance within a continuous monitoring time window;

[0092] The abnormal fluctuation frequency is defined as the proportion of the number of abnormal mutation trends of the target virtual instance within a continuous monitoring time window to the number of all monitoring points within the monitoring time window;

[0093] The abnormal fluctuation amplitude is defined as the change ratio of the change value of the running indicator when the abnormal mutation trend occurs to the normal running average value of the corresponding indicator;

[0094] The division method of the abnormal fluctuation level is:

[0095] When the abnormal fluctuation frequency value does not exceed 5% of the total number of monitoring points within the continuous monitoring time window, and the ratio of each abnormal fluctuation amplitude value to the normal operating average value of the corresponding indicator does not exceed 10%, the abnormal fluctuation level of the target virtual instance within this monitoring window is determined as low-level abnormal fluctuation.

[0096] When the abnormal fluctuation frequency value exceeds 5% but does not exceed 15% of the total number of monitoring points within the continuous monitoring time window, and the abnormal fluctuation amplitude value of each time exceeds 10% but does not exceed 30% of the average value of the corresponding indicator under normal operation, the abnormal fluctuation level of the target virtual instance within this monitoring window is determined to be medium-level abnormal fluctuation.

[0097] When the abnormal fluctuation frequency exceeds 15% of the total number of monitoring points within the continuous monitoring time window, and the abnormal fluctuation amplitude of each instance exceeds 30% of the average normal operating value of the corresponding indicator, the abnormal fluctuation level of the target virtual instance within this monitoring window is determined to be a high-level abnormal fluctuation.

[0098] Based on the interference intensity level and abnormal fluctuation level, a model is established to distinguish between the actual load of the target virtual instance and the interference load of adjacent virtual instances.

[0099] Real load refers to load changes directly caused by the target virtual instance's own computational tasks or internal operations. Interference load refers to load interference caused by the target virtual instance sharing the same physical resources with adjacent virtual instances. The distinction model is a rule system for judging and distinguishing the sources of load changes. The establishment process is as follows: Based on the interference intensity level of adjacent virtual instance node pairs, different levels of interference are assigned different impact weights; based on the abnormal fluctuation level of the target virtual instance, different levels of abnormal fluctuation are assigned sensitivity weight values; distinction rules are established based on the two types of weight values. For example: when the abnormal fluctuation is high-level and the interference level is severe, the abnormal fluctuation of the target virtual instance is mainly caused by interference from adjacent instances; if the abnormal fluctuation is high-level and the interference level is not severe, or if the abnormal fluctuation is low-level and the interference level is mild, it is mainly caused by real load; if the abnormal fluctuation is medium-level and the interference level is moderate, real load and interference load coexist. The above rule system is the distinction model.

[0100] Based on the differentiation model, each abnormal fluctuation feature in the abnormal fluctuation identification data of the target virtual instance is classified and determined to determine whether each abnormal fluctuation feature belongs to the real load change of the target virtual instance or the interference load change of adjacent virtual instances.

[0101] Extract each abnormal fluctuation feature in the abnormal fluctuation identification data of the target virtual instance, and determine the interference intensity level corresponding to the occurrence time of each abnormal fluctuation feature respectively; apply the distinguishing model rule for judgment, if the corresponding time of the abnormal fluctuation feature is a heavy interference level or a moderate interference level, it is determined that the abnormal fluctuation feature is caused by interference load; if the corresponding time of the abnormal fluctuation feature is a light interference level, it is determined to be caused by real load change; for example, ten abnormal fluctuation features are recorded in the target virtual instance in the enterprise second cloud computing service provider environment within the monitoring time window, by comparing the interference intensity levels of adjacent instance node pairs, six abnormal fluctuation features correspond to moderate or heavy interference levels, which are determined to be interference load; the remaining four abnormal fluctuation features have low interference levels, which are determined to be real load. Through the above determination method, the classification of the abnormal fluctuation identification data of the target virtual instance is completed.

[0102] S6: According to the distinguishing results of the real load of the target virtual instance and the interference load of the adjacent virtual instance, adjust the abnormal monitoring judgment logic of the multi-cloud computing resource, and output the interference corrected monitoring state label data, including:

[0103] When the load fluctuation category of the target virtual instance is determined to be the interference load change of the adjacent virtual instance, the abnormal fluctuation identification data of the target virtual instance in the standardized multi-cloud monitoring data set is marked and corrected to a non-abnormal monitoring state label;

[0104] The running index data state label corresponding to the abnormal fluctuation feature in the abnormal fluctuation identification data of the target virtual instance is modified from the initial state to the non-abnormal monitoring state label; the initial state is the label state of the abnormal fluctuation data of the default monitoring system, and the non-abnormal monitoring state label is the label state of the abnormal fluctuation feature caused by the load interference of the adjacent virtual instance, so it is marked as a non-abnormal monitoring state; for example, the target virtual instance of the enterprise in the multi-cloud environment records five abnormal mutation trends of processor utilization within a continuous monitoring time window, and the five abnormal mutation trends correspond to the interference intensity level of the adjacent virtual instance, which is determined by the distinguishing model, i.e. the five abnormal mutation trends are all caused by the load interference of the adjacent virtual instance; therefore, the processor utilization data label state corresponding to the five abnormal mutation trends is modified from the default initial abnormal state to the non-abnormal monitoring state label, indicating that the abnormal change of the monitoring index of the target virtual instance comes from the interference of the adjacent virtual instance of the shared physical resource, avoiding the misjudgment of the running state of the target virtual instance by the monitoring system and the operation and maintenance personnel.

[0105] When the load fluctuation category of the target virtual instance is determined to be the real load change of the target virtual instance, the abnormal fluctuation identification data of the target virtual instance in the standardized multi-cloud monitoring data set is marked and determined as an abnormal monitoring state label;

[0106] The running index data corresponding to the abnormal fluctuation feature recorded in the target virtual instance abnormal fluctuation identification data is modified from the initial state to an abnormal monitoring state label; the abnormal monitoring state label indicates that the abnormal fluctuation feature is a real abnormality caused by the real load change of the target virtual instance itself; for example, the target virtual instance of an enterprise in a multi-cloud environment records three abnormal mutation trends of disk input / output operation rate in a continuous monitoring time window, and the adjacent virtual instance interference intensity level corresponding to the three abnormal mutation trends is determined by the distinguishing model to be a mild interference level, which cannot cause interference to the target virtual instance, so the three abnormal mutation trends are confirmed to be caused by the real load change of the target virtual instance itself; the three abnormal mutation trends recorded by the disk input / output operation rate data of the target virtual instance are marked from the default initial state to the abnormal monitoring state label, which reflects the real source of the abnormal fluctuation feature, facilitates accurate judgment of the running status of the target virtual instance, and timely monitoring, early warning and resource scheduling measures.

[0107] The standardized multi-cloud monitoring data set is updated based on the non-abnormal monitoring state label and the abnormal monitoring state label, and the corrected target virtual instance monitoring state label data is output;

[0108] The running index data marked with the non-abnormal monitoring state label is uniformly corrected to the non-abnormal monitoring state classification in the standardized multi-cloud monitoring data set, which represents that the target virtual instance monitoring data has an abnormal mutation trend caused by interference from adjacent virtual instances sharing physical resources, so that resource emergency scheduling and load adjustment strategies are not needed; the running index data marked with the abnormal monitoring state label is uniformly corrected to the abnormal monitoring state classification in the standardized multi-cloud monitoring data set, which represents that the abnormal change of the target virtual instance monitoring data is caused by its own real load change, and resource reallocation or early warning intervention may be needed; through classification correction, the abnormal index data of the target virtual instance in the standardized multi-cloud monitoring data set is accurately divided into the corresponding monitoring state classification, thereby forming monitoring state label data that has been corrected and can accurately express the actual running state of the target virtual instance; for example, the target virtual instance deployed by an enterprise in the environment of a second cloud computing service provider records ten abnormal mutation trends of running index in a continuous monitoring time window, and after classification and determination by the distinguishing model, six of them belong to the non-abnormal monitoring state and the remaining four belong to the abnormal monitoring state; after updating the above classification results in the standardized multi-cloud monitoring data set, the corrected monitoring state label data for the target virtual instance in the current monitoring window is output; the monitoring state label data includes non-abnormal monitoring state data and abnormal monitoring state data, so that the monitoring and management personnel can distinguish the actual running state of the virtual instance and make accurate decisions and optimization adjustments.

[0109] The above-described embodiments can be implemented in whole or in part by software, hardware, firmware, or any combination thereof. When implemented by software, the above-described embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the processes or functions described in the embodiments of the present application are wholly or partially generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another computer-readable storage medium, for example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center through a wired (for example, infrared, wireless, microwave, etc.) manner. The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server, data center, etc. containing one or more available medium collections. The available medium can be a magnetic medium (for example, a floppy disk, a hard disk, a magnetic tape), an optical medium (for example, a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state disk.

[0110] Those of ordinary skill in the art can realize that the modules and algorithm steps of the examples described in conjunction with the embodiments disclosed herein can be implemented by electronic hardware or a combination of computer software and electronic hardware. Whether the functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0111] Those of ordinary skill in the art can clearly understand that, for the convenience and brevity of the description, the specific working processes of the above-described system, device, and module can refer to the corresponding processes in the foregoing method embodiments, which will not be described here.

[0112] In several embodiments provided in the present application, it should be understood that the disclosed system, device, and method can be implemented in other ways. For example, the above-described device embodiments are only schematic, for example, the division of the modules is only a logical function division, and actual implementation can have another division manner, for example, multiple modules or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the shown or discussed ones can be indirect coupling or communication connection through some interfaces, devices, or modules, which can be electrical, mechanical, or other forms.

[0113] The modules described as separate components may or may not be physically separate, and the components shown as modules may or may not be physical modules, and may be located in one place or distributed on multiple network modules. Part or all of the modules can be selected according to actual needs to achieve the purpose of the embodiment.

[0114] In addition, the functional modules in each embodiment of the present application can be integrated into one processing module, or each module can exist physically alone, or two or more modules can be integrated into one module.

[0115] The functions, if realized in the form of software function modules and sold or used as independent products, can be stored in a computer readable storage medium. Based on this understanding, the technical solutions of the present application essentially or the part that contributes to the prior art or part of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes a number of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and various program code storage media.

[0116] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art can easily think of changes or replacements within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

[0117] Finally, the above is only the preferred embodiment of the present application, and is not used to limit the present application. Any modification, equivalent replacement, improvement, etc. within the spirit and principles of the present application should be included in the protection scope of the present application.

Claims

1. A method for monitoring computing resources in a multi-cloud environment, characterized in that, The method comprises the following steps: S1: Collecting running index data of virtual computing resources deployed in a multi-cloud environment and physical host resource distribution data, and performing time alignment and normalization processing to generate a standardized multi-cloud monitoring data set; S2: Based on the standardized multi-cloud monitoring data set, a resource adjacency topology relationship graph between virtual instances is constructed; S3: Based on the resource adjacency topology relationship graph, the load similarity index between virtual instances is calculated, and the load interference feature data of adjacent virtual instances is extracted; S4: Based on the running index change data extracted from the multi-cloud monitoring data set, the load fluctuation characteristics and mutation trend of the target virtual instance are identified, and the abnormal fluctuation identification data of the target virtual instance is generated; S5: Based on the load interference feature data and the abnormal fluctuation identification data, a model for distinguishing the real load of the target virtual instance from the interference load of adjacent virtual instances is established; S6: According to the distinction result of the real load of the target virtual instance and the interference load of adjacent virtual instances, the abnormal monitoring judgment logic of the multi-cloud computing resource is adjusted, and the interference corrected monitoring state label data is output.

2. The method of claim 1, wherein, S1, specifically: Collecting running index data of virtual computing resources deployed in multiple different cloud service provider environments; Collecting distribution data of physical host resources deployed in multiple different cloud service provider environments; Calibrating the time base of the virtual computing resource running index data and the physical host resource distribution data uniformly; The calibrated virtual computing resource running index data and physical host resource distribution data are normalized to generate a standardized multi-cloud monitoring data set.

3. The method of claim 2, wherein, S2, specifically: According to the standardized multi-cloud monitoring data set, the virtual instances are grouped and a mapping table of virtual instances and physical host resources is generated; Based on the mapping table, each physical host resource is assigned a unique node number, and all virtual instances mapped to the same physical host resource node number are marked as same-layer nodes, and all virtual instances mapped to different physical host resource node numbers are marked as different-layer nodes; In the same layer nodes, a non-directed edge is established according to the shared processor core identifier, shared memory channel identifier and shared input / output bus identifier, and no edge is established between different layer nodes, generating a resource adjacency topology relationship graph between virtual instances.

4. The method of claim 3, wherein, S3, specifically: Based on the resource adjacency topology relationship graph, the processor usage rate data, memory usage rate data and network bandwidth usage rate data of each node of each adjacent virtual instance pair are extracted from the standardized multi-cloud monitoring data set; The similarity of processor usage rate data, the similarity of memory usage rate data and the similarity of network bandwidth usage rate data between each adjacent virtual instance pair are calculated and weighted summed to obtain the load similarity index between each adjacent virtual instance pair; Adjacent virtual instance node pairs with a load similarity index exceeding a preset similarity threshold are extracted to generate load interference feature data of adjacent virtual instances.

5. The method of claim 4, wherein, S4, specifically: Based on the standardized multi-cloud monitoring data set, the node identification information of the target virtual instance is determined; According to the node identification information of the target virtual instance, disk input / output operation rate data, disk delay data and processor usage rate data of the target virtual instance in a continuous monitoring time window are extracted from the standardized multi-cloud monitoring data set; The change rates of the disk input / output operation rate data, the disk delay data and the processor usage rate data in the continuous monitoring time window are calculated respectively, and the fluctuation characteristics corresponding to the change rates are identified; The mutation trend of each fluctuation characteristic is detected, and abnormal fluctuation identification data of the target virtual instance is generated.

6. The method of claim 5, wherein, S5, specifically: According to the load interference feature data of the adjacent virtual instance, the interference intensity level of the adjacent virtual instance node pair is determined; According to the abnormal fluctuation identification data of the target virtual instance, the abnormal fluctuation level of the target virtual instance is determined; Based on the interference intensity level and the abnormal fluctuation level, a distinguishing model of the real load of the target virtual instance and the interference load of the adjacent virtual instance is established; Based on the distinguishing model, each abnormal fluctuation characteristic in the abnormal fluctuation identification data of the target virtual instance is classified and determined, and it is determined that each abnormal fluctuation characteristic belongs to the real load change of the target virtual instance or belongs to the interference load change of the adjacent virtual instance.

7. The method of claim 6, wherein, S6, specifically: When the target virtual instance load fluctuation category is determined as the adjacent virtual instance interference load change, the abnormal fluctuation identification data of the target virtual instance in the standardized multi-cloud monitoring data set is marked and corrected to a non-abnormal monitoring state label; When the target virtual instance load fluctuation category is determined as the real load change of the target virtual instance, the abnormal fluctuation identification data of the target virtual instance in the standardized multi-cloud monitoring data set is marked and determined as an abnormal monitoring state label; Based on the non-abnormal monitoring state label and the abnormal monitoring state label, the standardized multi-cloud monitoring data set is updated, and the corrected target virtual instance monitoring state label data is output.

Images