WO2019019975A1

WO2019019975A1 - Method and device for cloud platform performance testing

Info

Publication number: WO2019019975A1
Application number: PCT/CN2018/096688
Authority: WO
Inventors: 杨蕾; 王黎; 吴波; 谢东
Original assignee: 华为技术有限公司
Priority date: 2017-07-24
Filing date: 2018-07-23
Publication date: 2019-01-31
Also published as: CN107547261B; CN107547261A

Abstract

Provided in the present application are a method and device for cloud platform performance testing. The method comprises: within a preset test run cycle, performing the following operation at each preset stress injection time interval with respect to a running cloud platform: randomly selecting a stress injection path on the basis of a prestored stress type set, a stress application object set, a stress application specification range, a stress application tool set, and a stress level range of said stress application tools, the stress injection path comprising N selected stress types, the stress application objects respectively corresponding to the N stress types and the number of said stress application objects, and the stress application tools and stress levels of the stress application tools respectively corresponding to the N stress types. Stress injection is performed on the basis of the stress injection path selected; the running of the cloud platform is monitored via the cloud platform monitoring system to acquire the result of the test run. Hence, the test result produced is of increased comprehensiveness, thus increasing the accuracy of the test.

Description

Cloud platform performance testing method and device

Technical field

The present application relates to the field of performance testing technologies, and in particular, to a cloud platform performance testing method and apparatus.

Background technique

The cloud platform is a large-scale, distributed management system. Its essence is a service provision model. Through this model, resources of a shared resource pool can be accessed through the network anytime, anywhere, and on-demand. The resources of the shared resource pool include computing resources and network resources. And storage resources, etc., these resources can be dynamically allocated and adjusted, and can be flexibly divided among different users. The cloud platform provides computing, storage, and network support for various applications in the upper layer through the form of services. The services and applications carried on the cloud platform are spread across all walks of life, with rich usage scenarios, large application scales, and complex application models. Cloud platforms generally need to have versatility and large-scale features in order to support various complex usage scenarios. The system function modules are more dependent and complex. Therefore, the performance capability of the cloud platform is an important indicator to measure the quality of the cloud platform, directly affecting the performance of the application on the cloud, and plays an important role in the availability and reliability of the cloud platform. Performance testing is an important test content of the cloud platform. As the scale of applications on the cloud increases, the scenarios become more and more complex. Different application types, different request pressures, and different data volumes have completely impacted the cloud platform. Different.

In the related art, the benchmark load test method is used to perform performance test of the cloud platform, and the test process is: first selecting a benchmark test object and a pressure test performance indicator item, and then selecting a reference pressure test tool and deploying on the selected benchmark test object, According to the pressure model, the selected reference pressure test tool is configured, and finally the test is started. When the pressure model continuously generates pressure, the test result of the reference pressure test tool is obtained, and the selected test object is determined by analyzing and comparing the test results. The performance capability of the pressure measurement performance indicator, for example, the different types of data disk reading and writing capabilities of the cloud platform are tested according to different read and write models for the number of readings per second and the time delay.

It can be seen that the above test method is based on the performance test under the single-point pressure injection model, and the obtained performance test data and evaluation conclusions are too simplistic, and in the cloud platform, a variety of cloud-on-line hybrid online applications and large-scale The pressure of requesting access data is different from the pressure point generated by the cloud platform and the single point pressure point generated by the reference pressure tool, and the resulting performance bottleneck point and system failure rate are also different. Therefore, the obtained test method is obtained. The test results are not comprehensive enough and the accuracy is not high.

Summary of the invention

The application provides a cloud platform performance testing method and device, which can simulate the performance of the cloud platform in a real application scenario, and the test result is more comprehensive and improves the accuracy of the test.

In a first aspect, the application provides a cloud platform performance testing method, including:

During the preset test run period, the following operations are performed on the running cloud platform at each preset pressure injection time interval:

A pressure injection path is randomly selected according to the pre-stored pressure type set, the pressure object set, the pressure specification range, the pressure tool set, and the pressure level range of the pressure tool, and the pressure injection path includes the selected N pressure types and N The pressure type corresponds to the number of pressure objects and the number of pressure objects, and the pressure levels of the pressure tools and pressure tools corresponding to the N pressure types, N≥1, wherein the pressure type is the pressure that causes performance problems in the cloud platform. The type of pressure object is the application system of the cloud platform and the object of the performance bottleneck in the management system. The pressure specification range is the range of the number of pressure objects, and the pressure tool is a tool set or a program set for pressure injection. Pressure injection according to the selected pressure injection path,

During the test operation period, the cloud platform monitoring system monitors the operation of the cloud platform to obtain the test operation result.

By setting the test operation period and the pressure injection time interval in advance, during the test operation period, a pressure injection path is selected for each cloud injection time interval in the running cloud platform, and pressure injection is performed according to the selected pressure injection path. . The pressure injection path includes the selected N pressure types, the pressure objects corresponding to the N pressure types, the number of pressure objects, the pressure levels of the pressure tools corresponding to the N pressure types, and the pressure level of the pressure tool. During the test operation period, the cloud platform monitoring system monitors the operation of the cloud platform to obtain test operation results. Since the selection of each pressure injection path is randomly selected according to different pressure types, the pressure type corresponding to the pressure object and the number of pressure objects, the pressure type corresponding pressure tool and the pressure level of the pressure tool, Therefore, the hybrid and overlay services in the real application scenario of the cloud platform can be simulated to the greatest extent, and the performance problems that may occur after the long-running of the business in the real application scenario are mined, compared with the single-point based pressure injection model in the related art. The performance test of this application is more comprehensive and improves the accuracy of the test.

In one possible design, a pressure injection path is randomly selected according to a pre-existing pressure type set, a set of pressure objects, a pressure specification range, a pressure tool set, and a pressure level range of the pressure tool, including: from the pressure type set Select N pressure types; determine the pressure objects corresponding to the N pressure types, randomly select the number of pressure objects corresponding to the N pressure types according to the pressure specification range; determine the pressure tools corresponding to the N pressure types respectively According to the pressure level range of the pressure applying tool, the pressure levels of the pressure tools corresponding to the N pressure types are randomly selected.

In a possible design, the pressure injection is performed according to the selected pressure injection path, including: corresponding pressure objects and pressure objects corresponding to the N pressure types, and pressure tools corresponding to the N pressure types respectively. Pressure injection is performed for the pressure levels of the pressure tools corresponding to the N pressure types.

In one possible design, the types of stress include business layer load, management system resources, and user operations;

The pressure object corresponding to the load of the business layer is a virtual machine, and the pressure object corresponding to the management system resource is a management node, and the pressure object corresponding to the user operation is a simulated user front-end interface operation and an open application program interface call.

In one possible design, the business layer load includes computationally intensive services, memory intensive services, storage intensive services, and network intensive services, and the pressure-intensive tools corresponding to computationally intensive services are central processor CPU benchmark tools. The pressure-intensive tool corresponding to the memory-intensive service is the memory benchmark tool set, the pressure-intensive tool corresponding to the storage-intensive service is the network benchmark tool set, and the pressure-intensive tool corresponding to the network-intensive service is the storage reference tool set. The management system resources include: management CPU resources, management memory resources, management storage resources, and management network resources; the pressure tools corresponding to the management CPU resources are the CPU consumption program collection, and the management memory resources corresponding to the pressure tools. For the memory consumption program set, the pressure tool corresponding to the management storage resource is a storage input and output IO consumption program set, and the pressure tool corresponding to the management network resource is a network IO consumption and a network interrupt injection program set. The user operation includes a computing class, a storage class, and a network class provided by the cloud platform, and the user-oriented cloud service is operable by the user; the corresponding pressure tool of the user operation is a program set that simulates the user operation.

In a second aspect, the application provides a cloud platform performance testing device, including:

The pressure injection module is configured to perform the following operations on the running cloud platform at each preset pressure injection time interval during a preset test operation period:

A pressure injection path is randomly selected according to the pre-stored pressure type set, the pressure object set, the pressure specification range, the pressure tool set, and the pressure level range of the pressure tool, and the pressure injection path includes the selected N pressure types and N The pressure type corresponds to the number of pressure objects and the number of pressure objects, and the pressure levels of the pressure tools and pressure tools corresponding to the N pressure types, N≥1, wherein the pressure type is the pressure that causes performance problems in the cloud platform. The type of pressure object is the application system of the cloud platform and the object of the performance bottleneck in the management system. The pressure specification range is the range of the number of pressure objects, and the pressure tool is a tool set or a program set for pressure injection. Pressure injection is performed according to the selected pressure injection path.

The obtaining module is configured to obtain a test running result by monitoring the running of the cloud platform through the cloud platform monitoring system during the test running period.

In a possible design, the pressure injection module is specifically configured to: select N pressure types from the pressure type set; determine the pressure objects corresponding to the N pressure types respectively, and randomly select N pressure types according to the pressure specification range The number of corresponding pressing objects; determining the pressing tools corresponding to the N pressure types respectively, and randomly selecting the pressure levels of the pressing tools corresponding to the N pressure types according to the pressure level range of the pressing tool.

In a possible design, the pressure injection module is specifically configured to: respectively correspond to the pressure object and the pressure object corresponding to the N pressure types, the pressure tools corresponding to the N pressure types, and the N pressure types respectively. The pressure level of the pressure tool is used for pressure injection.

In a third aspect, the application provides a cloud platform performance testing device, including: a memory and a processor;

The memory is used to store program instructions;

The processor is used to call a program instruction in the memory to perform the following method:

During the preset test run period, a pressure injection path is selected for each of the preset pressure injection time intervals for the running cloud platform, and pressure injection is performed according to the selected pressure injection path:

Wherein, selecting a pressure injection path specifically includes:

A pressure injection path is randomly selected according to the pre-stored pressure type set, the pressure object set, the pressure specification range, the pressure tool set, and the pressure level range of the pressure tool, and the pressure injection path includes the selected N pressure types and N The pressure type corresponds to the number of pressure objects and the number of pressure objects, and the pressure levels of the pressure tools and pressure tools corresponding to the N pressure types, N≥1, wherein the pressure type is the pressure that causes performance problems in the cloud platform. The type of pressure object is the application system of the cloud platform and the object of the performance bottleneck in the management system. The pressure specification range is the range of the number of pressure objects, and the pressure tool is a tool set or a program set for pressure injection;

In a fourth aspect, the application provides a readable storage medium, where the execution instructions are stored, and when the at least one processor of the cloud platform performance testing device executes the execution instruction, the cloud platform performance testing device executes the first aspect. And the method of any of the possible aspects of the first aspect.

In a fifth aspect, the present application provides a program product comprising an execution instruction stored in a readable storage medium. At least one processor of the cloud platform performance testing device can read the execution instruction from a readable storage medium, and the at least one processor executes the execution instruction to cause the cloud platform performance testing device to implement the first aspect and any one of the possible aspects of the first aspect The method in .

DRAWINGS

1 is a flowchart of an embodiment of a cloud platform performance testing method provided by the present application;

2 is a flowchart of an embodiment of a cloud platform performance testing method provided by the present application;

3 is a schematic structural diagram of an embodiment of a cloud platform performance testing apparatus according to the present application;

FIG. 4 is a schematic structural diagram of an embodiment of a cloud platform performance testing apparatus according to the present application.

Detailed ways

In order to make the test result of the cloud platform performance test more comprehensive and accurate, the present application simulates the performance of the real application scenario test cloud platform, specifically by setting the test operation cycle and the pressure injection time interval in advance, and in the set operation cycle, the operation is in operation. The cloud platform selects a pressure injection path at each pressure injection time interval, and performs pressure injection according to the selected pressure injection path. The pressure injection path includes the selected N pressure types and N pressure types respectively corresponding to the pressure. The number of objects and pressure objects, and the pressure levels of the pressure tools and pressure tools corresponding to the N pressure types respectively classify the pressures that cause performance problems in the cloud platform. The selection of the pressure injection paths is random, so it can be maximized. To simulate the hybrid and overlay services in the real application scenario of the cloud platform, and to explore the performance problems that may occur after the long-running of the business in the real application scenario, compared with the performance test based on the single-point pressure injection model in the related technology. The test results of this application are more comprehensive and accurate. The cloud platform performance testing method and apparatus provided by the present application are described in detail below with reference to the accompanying drawings.

FIG. 1 is a flowchart of an embodiment of a cloud platform performance testing method provided by the present application. As shown in FIG. 1 , the method in this embodiment may include:

S101. Perform the following operations on the running cloud platform at each preset pressure injection time interval in a preset test operation period:

A pressure injection path is randomly selected according to the pre-stored pressure type set, the pressure object set, the pressure specification range, the pressure tool set, and the pressure level range of the pressure tool, and the pressure injection path includes the selected N pressure types and N The pressure type corresponds to the number of pressure objects and pressure objects, and the pressure levels of the pressure tools and pressure tools corresponding to the N pressure types are respectively N≥1, and pressure injection is performed according to the selected pressure injection path.

Specifically, for example, the test operation period is 7-10 days, etc., for example, pressure injection is performed for 5 minutes, and the pressure injection time interval is 5 minutes, 10 minutes, 15 minutes, and the like. For each pressure injection, a pressure injection path is selected, and pressure injection is performed according to the selected pressure injection path.

Among them, the pressure type is the type of pressure in the cloud platform that causes performance problems. For example, the pressure on the cloud platform that may cause performance problems is divided into two dimensions. One is the performance bottleneck point that may exist in the system, such as the consumption of system resources. That is, the management system resources; the second is the operation in the user's actual application scenario, such as the execution of the business application (service layer load), the user operation of the cloud platform, and the pressure type is randomly selected. The object of pressure is the application system of the cloud platform and the object that generates performance bottlenecks in the management system, such as a virtual machine, a management node, a simulated user front-end interface (protal) operation, and an open application programming interface (API) call. The pressure specification range is the range of the number of objects to be pressed, or the maximum number of objects to be pressed. When the pressure object is a virtual machine, the corresponding pressure specification range is the range of the number of virtual machines that the system can support. When the pressure object is the management node, the corresponding pressure specification range is the scope of the management node deployed by the system; To simulate user protal operations and open API calls, the corresponding pressure specification range is the maximum number of concurrent operations supported by the cloud platform, which is a random positive integer. The pressure tool is a collection of tools or a collection of programs for pressure injection. The pressure level of the pressure tool can be the maximum value of the pressure level, and the pressure level of the pressure tool is randomly selected and determined.

Table 1 below is a correspondence table of pressure type, pressure object, pressure specification range, pressure level of pressure tool and pressure tool, as shown in Table 1:

Table I

The pressure types can be divided into three categories: business layer load, management system resources, and user operations. The pressure object corresponding to the business layer load is a virtual machine. The pressure object corresponding to the management system resource is the management node. The corresponding pressure object is the simulated user protal operation and open API call.

Among them, the business layer load includes computationally intensive services, memory intensive services, storage intensive services, and network intensive services. The pressure-intensive tool corresponding to the computation-intensive service is a central processing unit (CPU) benchmark tool set, and the pressure-intensive service corresponding to the memory-intensive service is a set of memory benchmark tools, and a load-intensive tool corresponding to the storage-intensive service. For the network benchmark tool set, the pressure tool corresponding to the network-intensive business is a collection of benchmark tools.

The management system resources include: management CPU resources, management memory resources, management storage resources, and management network resources. The pressure tool corresponding to the management CPU resource is a CPU consumption program set, the pressure tool corresponding to the management memory resource is a memory consumption program set, and the pressure storage tool corresponding to the management storage resource is a storage input and output IO consumption program set, and the management layer The pressure tool corresponding to the network resource is a collection of programs for network IO consumption and network interruption.

User operations include user-oriented operations such as computing, storage, and network classes provided by the cloud platform. For example, computing services, network services, and storage services. The user-operated corresponding pressure tool is a collection of programs that simulate user operations, such as creating, deleting, stopping, and starting a collection of operations.

The pressure levels of the above-mentioned pressure applying tools are randomly determined according to the pressure level ranges of the corresponding pressure applying tools.

Specifically, a pressure injection path is randomly selected according to the pre-stored pressure type set, the pressure object set, the pressure specification range, the pressure tool set, and the pressure level range of the pressure tool, which may be:

S101a: Select N pressure types from the pressure type set.

S101b: determining a pressing object corresponding to each of the N pressure types, and randomly selecting the number of pressing objects corresponding to the N pressure types according to the pressure specification range.

S101c: determining pressure tools corresponding to the N pressure types respectively, and randomly selecting pressure levels of the pressure tools corresponding to the N pressure types according to the pressure level range of the pressure application tool.

Pressure injection according to the selected pressure injection path can be:

The pressure injection is performed according to the pressure targets corresponding to the pressure objects and the pressure objects corresponding to the N pressure types, the pressure tools corresponding to the N pressure types, and the pressure levels of the pressure tools respectively corresponding to the N pressure types.

For example, N is 3, the selected pressure type includes 2 computationally intensive services and 1 computing service, and 2 computationally intensive services correspond to 5 pressure objects respectively, and the pressure object corresponding to one computing service is The pressure tools corresponding to the five computing intensive services are the CPU benchmark tools. The pressure levels are M1 and M2 respectively. The pressure tools corresponding to one computing service are the set of creation, deletion, stop and start operations, and the pressure level. For M3, the pressure-bearing objects corresponding to computation-intensive services and computing services are virtual machine and simulated user protal operations and open API calls. When pressure injection is performed, five virtual machines adopt the CPU benchmark tool set to perform pressure level M1. Injection, pressure injection with pressure level M2 for 5 virtual machines using CPU benchmark tool set, pressure injection with pressure level M3 for 5 simulated user protal operations, open API calls using set of creation, deletion, stop and start operations .

S102. Monitor the running of the cloud platform through the cloud platform monitoring system during the test operation period to obtain a test operation result.

For example, the test running period is 7-10 days, the cloud platform runs normally for 7-10 days, and the normal functional operation is performed during the whole running process. Under the background of the random pressure injection described above, the response time of the cloud platform system is observed through the cloud platform monitoring system. Whether the performance monitoring index items such as resource occupancy exceed the threshold value, and observe whether each function module of the system generates an error log, a suspension, a non-response abnormality, etc., and performs analysis to obtain a test operation result.

The cloud platform performance testing method provided by the embodiment selects a test injection period and a pressure injection time interval in advance, and selects a pressure injection path for each of the pressure injection time intervals in the running cloud platform during the test operation period. Pressure injection is performed according to the selected pressure injection path. The pressure injection path includes the selected N pressure types, the pressure objects corresponding to the N pressure types, the number of pressure objects, the pressure levels of the pressure tools corresponding to the N pressure types, and the pressure level of the pressure tool. During the test operation period, the cloud platform monitoring system monitors the operation of the cloud platform to obtain test operation results. Since the selection of each pressure injection path is randomly selected according to different pressure types, the pressure type corresponding to the pressure object and the number of pressure objects, the pressure type corresponding pressure tool and the pressure level of the pressure tool, Therefore, the hybrid and overlay services in the real application scenario of the cloud platform can be simulated to the greatest extent, and the performance problems that may occur after the long-running of the business in the real application scenario are mined, compared with the single-point based pressure injection model in the related art. The performance test of this application is more comprehensive and improves the accuracy of the test.

The technical solution of the method embodiment shown in FIG. 1 is described in detail below by using a specific embodiment.

2 is a flowchart of an embodiment of a cloud platform performance testing method provided by the present application. As shown in FIG. 2, the method in this embodiment may include:

S201, the performance test begins.

S202: Perform random pressure injection during a preset test operation period.

Specifically, performing random pressure injection includes:

S202a. Select a pressure type 1 from the set of pressure types.

S202b, determining the pressing object X corresponding to the selected pressure type 1, and randomly selecting the number n of pressing objects corresponding to the pressure type 1 according to the pressure specification range [0, m], and the maximum number of pressing objects is m, n ≤ m.

S202c. Determine the pressure application tool R1 corresponding to the pressure type 1 from the pressure tool set R, and select the pressure level L1 according to the pressure level range [0, L] of the pressure application tool R1.

S202d, pressure injection of the pressure level L1 is performed using the pressing tool R1 for the n pressure applying objects X1 to Xn.

If the test run cycle ends, the pressurization is stopped, otherwise random pressure injection continues at each predetermined pressure injection time interval.

S203, the whole process performance indicator observation, that is, the test operation result is obtained by monitoring the operation of the cloud platform through the cloud platform monitoring system during the test operation period.

In this embodiment, the selected pressure type is taken as an example. When the pressure type is multiple, the selection of the pressure injection path and the pressure injection are similar, and are not exemplified herein.

The application may divide the functional modules of the cloud platform performance testing device according to the above method example. For example, each functional module may be divided according to each function, or two or more functions may be integrated into one processing module. The above integrated modules can be implemented in the form of hardware or in the form of software functional modules. It should be noted that the division of the modules in the embodiments of the present application is schematic, and is only a logical function division. In actual implementation, there may be another division manner.

FIG. 3 is a schematic structural diagram of an embodiment of a cloud platform performance testing apparatus according to the present application. As shown in FIG. 3, the apparatus includes: a memory 11, a processor 12, an interface circuit 13, and a bus 14, wherein

The memory 11, the processor 12, and the interface circuit 13 are connected by the bus 14 and complete communication with each other. The processor 12 receives or transmits information, such as control information, data, etc., through the interface circuit 13.

The memory 11 stores a set of program codes, and the processor 12 calls the program code stored in the memory 11 to perform the following operations:

A pressure injection path is randomly selected according to the pre-stored pressure type set, the pressure object set, the pressure specification range, the pressure tool set, and the pressure level range of the pressure tool, and the pressure injection path includes the selected N pressure types and N The pressure type corresponds to the number of pressure objects and the number of pressure objects, and the pressure levels of the pressure tools and pressure tools corresponding to the N pressure types are respectively N≥1.

Among them, the pressure type is the type of pressure in the cloud platform that causes performance problems. For example, the pressure on the cloud platform that may cause performance problems is divided into two dimensions. One is the performance bottleneck point that may exist in the system, such as the consumption of system resources. That is, the management system resources; the second is the operation in the user's actual application scenario, such as the execution of the business application (service layer load), the user operation of the cloud platform, and the pressure type is randomly selected. The object of pressure is the application system of the cloud platform and the objects that cause performance bottlenecks in the management system, such as virtual machines, management nodes, simulated user front-end interface (protal) operations, and open application interface (open API) calls. The pressure specification range is the range of the number of objects to be pressed, or the maximum number of objects to be pressed. When the pressure object is a virtual machine or a management node, the corresponding pressure specification range is the range of the number of virtual machines or management nodes; when the pressure object is the simulated user protal operation and the open API call, the corresponding pressure specification range is cloud. The maximum number of concurrent operations for a user supported by the platform. The pressure tool is a collection of tools or a collection of programs for pressure injection. The pressure level of the pressure tool can be the maximum value of the pressure level, and the pressure level of the pressure tool is randomly selected and determined.

Pressure injection is performed according to the selected pressure injection path.

4 is a schematic structural diagram of an embodiment of a cloud platform performance testing device according to the present application. As shown in FIG. 4, the device in this embodiment may include: a pressure injection module 21 and an acquisition module 22, where

The pressure injection module 21 is configured to perform the following operations on the running cloud platform at each preset pressure injection time interval during a preset test operation period:

The obtaining module 22 is configured to obtain a test running result by monitoring the running of the cloud platform by the cloud platform monitoring system during the test running period.

Further, the pressure injection module 21 is specifically configured to: select N pressure types from the pressure type set, determine the pressure objects corresponding to the N pressure types, and randomly select the pressure corresponding to the N pressure types according to the pressure specification range. The number of objects is determined, and the pressure tools corresponding to the N pressure types are respectively determined, and the pressure levels of the pressure tools corresponding to the N pressure types are randomly selected according to the pressure level range of the pressure application tool.

Further, the pressure injection module 21 is specifically configured to: respectively, according to the pressure types of the N pressure types, the number of pressure objects, the pressure tools corresponding to the N pressure types, and the pressure tools corresponding to the N pressure types respectively. The pressure level is pressure injected.

Optionally, the pressure type includes a service layer load, a management system resource, and a user operation, and the pressure object corresponding to the service layer load is a virtual machine, and the pressure object corresponding to the management system resource is a management node, and the user operation corresponding to the application The object is the mock user front-end interface operation and the open application interface call.

Optionally, the business layer load includes computationally intensive services, memory intensive services, storage intensive services, and network intensive services. The pressure-intensive tool corresponding to the computation-intensive service is a central processing unit (CPU) benchmark tool set, and the pressure-intensive service corresponding to the memory-intensive service is a set of memory benchmark tools, and a load-intensive tool corresponding to the storage-intensive service. For the network benchmark tool set, the pressure tool corresponding to the network-intensive business is a collection of benchmark tools.

The user operation includes operations of the computing class, the storage class, and the network class user-oriented cloud service provided by the cloud platform for the user to perform. The user-applied pressure tool is a collection of programs that simulate user operations.

The device in this embodiment may be used to implement the technical solution of the method embodiment shown in FIG. 1 , and the implementation principle is similar, and details are not described herein again.

The cloud platform performance testing device provided by the embodiment selects a test injection cycle and a pressure injection time interval in advance, and selects a pressure injection path for each cloud injection time interval in the running cloud platform during the test operation cycle. Pressure injection is performed according to the selected pressure injection path. The pressure injection path includes the selected N pressure types, the pressure objects corresponding to the N pressure types, the number of pressure objects, the pressure levels of the pressure tools corresponding to the N pressure types, and the pressure level of the pressure tool. During the test operation period, the cloud platform monitoring system monitors the operation of the cloud platform to obtain test operation results. Since the selection of each pressure injection path is randomly selected according to different pressure types, the pressure type corresponding to the pressure object and the number of pressure objects, the pressure type corresponding pressure tool and the pressure level of the pressure tool, Therefore, the hybrid and overlay services in the real application scenario of the cloud platform can be simulated to the greatest extent, and the performance problems that may occur after the long-running of the business in the real application scenario are mined, compared with the single-point based pressure injection model in the related art. The performance test of this application is more comprehensive and improves the accuracy of the test.

The application further provides a readable storage medium, where the execution instructions are stored, and when at least one processor of the cloud platform performance testing device executes the execution instruction, the cloud platform performance testing device executes the foregoing FIG. Cloud platform performance testing method.

The application also provides a program product comprising an execution instruction stored in a readable storage medium. The at least one processor of the cloud platform performance testing device can read the execution instruction from the readable storage medium, and the at least one processor executes the execution instruction to cause the cloud platform performance testing device to implement the cloud platform performance testing method shown in FIG. 1 above.

One of ordinary skill in the art will appreciate that all or part of the steps to implement the various method embodiments described above may be accomplished by hardware associated with the program instructions. The aforementioned program can be stored in a computer readable storage medium. The program, when executed, performs the steps including the foregoing method embodiments; and the foregoing storage medium includes various media that can store program codes, such as a ROM, a RAM, a magnetic disk, or an optical disk.

Claims

A cloud platform performance testing method, comprising:

During the preset test run period, the following operations are performed on the running cloud platform at each preset pressure injection time interval:

A pressure injection path is randomly selected according to the pre-stored pressure type set, the pressure object set, the pressure specification range, the pressure tool set, and the pressure level range of the pressure tool, the pressure injection path including the selected N pressure types, a pressure level corresponding to the pressing object and the pressing object corresponding to the N pressure types, and a pressure level of the pressing tool and the pressing tool respectively corresponding to the N pressure types, N≥1, wherein the pressure type is a type of pressure in a cloud platform that causes a performance problem, the pressure object is an application system of the cloud platform and an object of a performance bottleneck in the management system, and the pressure specification range is a range of the number of the pressure object, the pressure A collection of tools or programs that are used for pressure injection;

Pressure injection according to the selected pressure injection path;

During the test operation period, the cloud platform monitoring system monitors the operation of the cloud platform to obtain a test operation result.
The method according to claim 1, wherein said randomly selecting a pressure injection path according to a pre-stored pressure type set, a pressure object set, a pressure specification range, a pressure tool set, and a pressure level range of the pressure tool ,include:

Selecting the N pressure types from the set of pressure types;

Determining the pressing objects corresponding to the N pressure types respectively, and randomly selecting the number of pressing objects corresponding to the N pressure types according to the pressure specification range;

Determining the pressing tools corresponding to the N pressure types respectively, and randomly selecting the pressure levels of the pressing tools corresponding to the N pressure types according to the pressure level range of the pressing tool.
The method according to claim 1 or 2, wherein said performing pressure injection according to the selected pressure injection path comprises:

And pressing according to the pressure levels of the pressing objects corresponding to the N pressure types, the pressing objects corresponding to the N pressure types, and the pressure tools corresponding to the N pressure types respectively injection.
The method according to any one of claims 1 to 3, wherein the type of stress comprises a service layer load, a management system resource, and a user operation;

The pressure object corresponding to the service layer load is a virtual machine, and the pressure object corresponding to the management system resource is a management node, and the pressure object corresponding to the user operation is an analog user front-end interface operation and an open application interface call. .
The method of claim 4 wherein:

The service layer load includes computationally intensive services, memory intensive services, storage intensive services, and network intensive services;

The pressure-intensive tool corresponding to the computation-intensive service is a central processor CPU reference tool set, and the pressure-intensive service corresponding pressure-applying tool is a memory reference tool set, and the storage-intensive service corresponding pressure-applying tool is a network a set of benchmark tools, the pressure tool corresponding to the network-intensive service is a set of storage reference tools;

The management system resources include: management CPU resources, management memory resources, management storage resources, and management network resources;

The pressure tool corresponding to the management CPU resource is a CPU consumption program set, the pressure tool corresponding to the management memory resource is a memory consumption program set, and the pressure management tool corresponding to the management storage resource is a storage input and output IO. The collection of the program, the pressure tool corresponding to the management network resource is a network IO consumption, a network interrupt injection program set;

The user operation includes operations of the computing class, the storage class, and the network class user-oriented cloud service provided by the cloud platform for the user to perform;

The pressure tool corresponding to the user operation is a program set that simulates a user operation.
A cloud platform performance testing device, comprising:

The pressure injection module is configured to perform the following operations on the running cloud platform at each preset pressure injection time interval during a preset test operation period:

A pressure injection path is randomly selected according to the pre-stored pressure type set, the pressure object set, the pressure specification range, the pressure tool set, and the pressure level range of the pressure tool, the pressure injection path including the selected N pressure types, a pressure level corresponding to the pressing object and the pressing object corresponding to the N pressure types, and a pressure level of the pressing tool and the pressing tool respectively corresponding to the N pressure types, N≥1, wherein the pressure type is a type of pressure in a cloud platform that causes a performance problem, the pressure object is an application system of the cloud platform and an object of a performance bottleneck in the management system, and the pressure specification range is a range of the number of the pressure object, the pressure A collection of tools or programs that are used for pressure injection;

Pressure injection according to the selected pressure injection path;

And an obtaining module, configured to monitor, by the cloud platform monitoring system, the operation of the cloud platform to obtain a test running result during the testing operation period.
The device according to claim 6, wherein the pressure injection module is specifically configured to:

Selecting the N pressure types from the set of pressure types;

Determining the pressing objects corresponding to the N pressure types respectively, and randomly selecting the number of pressing objects corresponding to the N pressure types according to the pressure specification range;

Determining the pressing tools corresponding to the N pressure types respectively, and randomly selecting the pressure levels of the pressing tools corresponding to the N pressure types according to the pressure level range of the pressing tool.
The device according to claim 6 or 7, wherein the pressure injection module is specifically configured to:

And pressing according to the pressure levels of the pressing objects corresponding to the N pressure types, the pressing objects corresponding to the N pressure types, and the pressure tools corresponding to the N pressure types respectively injection.
The apparatus according to any one of claims 6-8, wherein the type of stress comprises a service layer load, a management system resource, and a user operation;

The pressure object corresponding to the service layer load is a virtual machine, and the pressure object corresponding to the management system resource is a management node, and the pressure object corresponding to the user operation is an analog user front-end interface operation and an open application interface call. .
The device of claim 9 wherein:

The service layer load includes computationally intensive services, memory intensive services, storage intensive services, and network intensive services;

The pressure-intensive tool corresponding to the computation-intensive service is a central processor CPU reference tool set, and the pressure-intensive service corresponding pressure-applying tool is a memory reference tool set, and the storage-intensive service corresponding pressure-applying tool is a network a set of benchmark tools, the pressure tool corresponding to the network-intensive service is a set of storage reference tools;

The management system resources include: management CPU resources, management memory resources, management storage resources, and management network resources;

The pressure tool corresponding to the management CPU resource is a CPU consumption program set, the pressure tool corresponding to the management memory resource is a memory consumption program set, and the pressure management tool corresponding to the management storage resource is a storage input and output IO. The collection of the program, the pressure tool corresponding to the management network resource is a network IO consumption, a network interrupt injection program set;

The user operation includes operations of the computing class, the storage class, and the network class user-oriented cloud service provided by the cloud platform for the user to perform;

The pressure tool corresponding to the user operation is a program set that simulates a user operation.