CN115562978A - Performance test system and method based on service scene - Google Patents
Performance test system and method based on service scene Download PDFInfo
- Publication number
- CN115562978A CN115562978A CN202211175645.XA CN202211175645A CN115562978A CN 115562978 A CN115562978 A CN 115562978A CN 202211175645 A CN202211175645 A CN 202211175645A CN 115562978 A CN115562978 A CN 115562978A
- Authority
- CN
- China
- Prior art keywords
- test
- service
- scene
- performance
- model
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/36—Preventing errors by testing or debugging software
- G06F11/3668—Software testing
- G06F11/3672—Test management
- G06F11/3688—Test management for test execution, e.g. scheduling of test suites
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Computer Hardware Design (AREA)
- Quality & Reliability (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Debugging And Monitoring (AREA)
Abstract
The invention relates to the field of performance test, and provides a performance test system based on a service scene for comprehensively testing the performance of a system, which comprises: a service scene recognition module; a performance target determination module; the system comprises a business scene model establishing module and a test executing module. The performance test method based on the service scene comprises the following steps: step 1, identifying a service scene; step 2, determining an expected index and a performance target according to a service scene, and quantizing the performance target into a performance index; step 3, establishing a service scene model by combining test parameters according to a service scene and performance indexes thereof, wherein the service scene model comprises a single-scene test model, a mixed-scene test model and a service proportioning model; and 4, performing performance test based on the service scene model to obtain a response index. By adopting the mode, the system performance can be conveniently and comprehensively tested.
Description
Technical Field
The invention relates to the field of performance test, in particular to a performance test system and a performance test method based on a service scene.
Background
With the arrival of the 5G era and the everything interconnection era, more and more cloud applications and cloud services are available, and the data volume is exponentially increased. Can cause various industries to start to appear clouds, thereby hastening the birth of various products with individuation. How to deal with the increase of users and flow and simultaneously ensure the stable operation of the application becomes a problem which needs to be solved by each manufacturer. The performance problem of the system also gradually becomes the focus of our attention: the problem of the large data volume 'impact' every day, at what performance level the system can stabilize, and whether the system can withstand 'test' when the business of the industry company is increased, needs to be answered by a complete performance test. In the face of increasingly complex service scenes and different system architectures, different test strategies are of great importance to whether the test results meet the expected targets.
At present, in the prior art, the system performance cannot be comprehensively and accurately effectively tested, the system bottleneck cannot be accurately positioned, an effective means is lacked for comprehensively evaluating the performance of an application software system from local to whole, and an effective model is lacked for ensuring the comprehensive implementation of the performance test. If the test method is not appropriate, the service scene is not designed reasonably enough, so that the comprehensive performance test cannot reflect the actual load pressure model of the application software, and the bottleneck of the positioning system is difficult.
Disclosure of Invention
In order to comprehensively test the system performance, the application provides a performance test system and method based on a service scene.
The technical scheme adopted by the invention for solving the problems is as follows:
a performance test system based on a service scene comprises:
a service scene recognition module: for determining a business scenario;
a performance goal determination module: the system is used for determining the expected indexes and the performance targets corresponding to the service scenes, and quantizing the performance targets into performance indexes;
the business scene model establishing module: the system is used for establishing a service scene model according to a service scene and performance indexes thereof in combination with test parameters; the business scenario model comprises a single scenario test model, a mixed scenario test model and a business matching model, wherein the single scenario test model is used for testing single businesses, the mixed scenario test model is used for testing comprehensive businesses, and the business matching model is used for determining the proportion of each single business in the comprehensive businesses;
a test execution module: and the method is used for performing performance test in combination with the service scene model to obtain the response index.
Further, the system also comprises an optimization judgment module: and the method is used for determining whether the system needs to be optimized according to the response index and the expected index, and if so, giving factors needing to be optimized.
Further, the factors include internal factors and external factors.
Further, the business proportioning model determines the proportion of each business condition actually accessed by the user according to the log analysis of the historical user, thereby determining the proportion of each single business in the integrated business.
The performance test method based on the service scene comprises the following steps:
step 2, determining an expected index and a performance target according to a service scene, and quantizing the performance target into a performance index;
step 3, establishing a service scene model by combining test parameters according to a service scene and performance indexes thereof, wherein the service scene model comprises a single scene test model, a mixed scene test model and a service proportioning model;
and 4, performing performance test based on the service scene model to obtain a response index.
Further, the performance index comprises the overall performance index of the system and the performance index of a single service.
Further, the single scene model includes: single scene benchmark test and single scene concurrent test; the hybrid scenario test model includes: the method comprises a mixed scene benchmark test, a mixed scene concurrent test and a stability test.
Further, the single scene benchmark test specifically comprises the following steps: under the condition that the system has no pressure, a single user iteratively executes a single service for a period of time or times to obtain a response index when the service runs;
the single-scene concurrent test specifically comprises the following steps: based on a single scene benchmark test, user behavior simulation is carried out by increasing the number of users, a single service is operated concurrently, and a response index of the service during operation is obtained;
the mixed scene benchmark test specifically comprises the following steps: under the condition that the system is not under pressure, a single user iteratively executes a plurality of combined services for a period of time or times to obtain a response index when each service runs;
the mixed scene concurrency test comprises the following specific steps: based on a mixed scene benchmark test, user behavior simulation is carried out by increasing the number of users, a plurality of services are operated concurrently according to the proportion set by a service proportioning model on the basis of data volume of a certain scale, and response indexes during service operation are obtained;
the stability test comprises the following specific steps: and based on the concurrent test of the mixed scene, acquiring a certain number of users as the load of the stability test, operating for a period of time, and acquiring the response index of the service during operation.
And further, the method also comprises a step 5 of determining whether the system needs to be optimized according to the response index and the expected index, giving factors needing to be optimized when the system needs to be optimized, and judging whether the system needs to be optimized again after the optimization is adjusted until the requirements are met.
Further, the performance test execution sequence is as follows: the method comprises the following steps of single-scene benchmark testing, single-scene concurrent testing, mixed-scene benchmark testing, mixed-scene concurrent testing and stability testing, wherein the next testing is executed after the current testing meets the requirement.
Compared with the prior art, the invention has the beneficial effects that: the invention can test whether each service of the system meets the performance requirement in a single scene, and the bottleneck of the system is locally analyzed; testing whether the system meets the performance requirement in a mixed scene, and integrally evaluating the performance of the system; the load consistent with the actual use environment of a user can be simulated, the system performance can be tested more accurately and efficiently, the system bottleneck can be positioned accurately, and a basis is provided for system performance tuning; the system can guide the whole process of implementing the performance project, and comprehensively monitors the detailed process of finding system bottleneck, finding out root causes, tuning and verifying.
Drawings
Fig. 1 is a flowchart of a performance testing method based on a service scenario according to an embodiment;
FIG. 2 is a schematic diagram of a business scenario model.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is further described in detail with reference to the following embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and do not limit the invention.
A performance test system based on a service scene comprises:
a service scene recognition module: for determining a business scenario;
a performance target determination module: the system is used for determining an expected index and a performance target corresponding to a service scene, and quantizing the performance target into a performance index;
a business scene model establishing module: the system is used for establishing a service scene model according to a service scene and performance indexes thereof in combination with test parameters; the business scene model comprises a single scene test model, a mixed scene test model and a business matching model, wherein the single scene test model is used for testing single business, the mixed scene test model is used for testing integrated business, and the business matching model is used for determining the matching of the actual access of each business condition of a user according to the log analysis of historical users and determining the proportion of each single business in the integrated business.
A test execution module: and the method is used for performing performance test in combination with the service scene model to obtain the response index.
Further, the system also comprises an optimization judgment module: and the method is used for determining whether the system needs to be optimized according to the response index and the expected index, and if so, giving factors needing to be optimized. The factors include internal factors and external factors.
The performance test method based on the service scene comprises the following steps:
step 2, determining an expected index and a performance target according to a service scene, and quantizing the performance target into a performance index; the performance indexes comprise the overall performance index of the system and the performance index of a single service;
step 3, establishing a service scene model by combining test parameters according to a service scene and performance indexes thereof, wherein the service scene model comprises a single-scene test model, a mixed-scene test model and a service proportioning model;
and 4, performing performance test based on the service scene model to obtain a response index.
Specifically, the single scene test model includes: single scene benchmark test and single scene concurrent test; the mixed scene test model comprises the following steps: the method comprises a mixed scene benchmark test, a mixed scene concurrent test and a stability test.
The single-scene benchmark test comprises the following specific steps: under the condition that the system has no pressure, a single user iteratively executes a single service for a period of time or times to obtain a response index when the service runs;
the single-scene concurrent test specifically comprises the following steps: based on a single scene benchmark test, user behavior simulation is carried out by increasing the number of users, a single service is operated concurrently, and a response index when the service runs is obtained;
the mixed scene benchmark test specifically comprises the following steps: under the condition that the system is not under pressure, a single user iteratively executes a plurality of combined services for a period of time or times to obtain a response index when each service runs;
the mixed scene concurrency test comprises the following specific steps: based on a mixed scene benchmark test, user behavior simulation is carried out by increasing the number of users, a plurality of services are operated concurrently according to the proportion set by a service proportioning model on the basis of data volume of a certain scale, and response indexes during service operation are obtained;
the stability test comprises the following specific steps: based on the mixed scene concurrent test, a certain number of users is obtained as the load of the stability test, the operation is carried out for a period of time, and the response index of the operation of the service is obtained.
And further, the method also comprises a step 5 of determining whether the system needs to be optimized according to the response index and the expected index, giving factors needing to be optimized when the system needs to be optimized, and judging whether the system needs to be optimized again after the optimization is adjusted until the requirements are met.
Specifically, the performance test execution sequence is: the method comprises the following steps of single-scene benchmark test, single-scene concurrent test, mixed-scene benchmark test, mixed-scene concurrent test and stability test, wherein the next test is executed after the current test meets the requirement.
Examples
As shown in fig. 1, a service scenario is identified by a service scenario identification module: the service scene describes the specific behavior of the user operation, and mainly comprises the specific operation of the user on the system service function. The service function may be specific behaviors of a user for operation, such as login, query, and browsing, and correspondingly, the service scenario may be user operation descriptions, such as login by the user in a login interface, query by inputting conditions in a search box, and browsing a home page.
And the performance target determining module determines a performance target related to the service scene according to the service scene, quantifies the performance target into a performance index and determines an expected index value.
The performance targets comprise the number of the concurrent users supported by the system, the acceptable response time under the predefined number of the concurrent users, the TPS, the resource utilization rate, the inspection stability, the bottleneck determination and the like. And decomposing and quantizing the performance target into performance indexes, wherein each decomposed index can be measured, such as the number of concurrent users, the response time, the TPS, the resource utilization rate and the like.
After the index item is clear, the required value of the index item, namely the expected index, in the aspect of the service capability or quality provided by the system is determined according to the analysis of the application field, the technical specification, the user scale, the user requirement, the historical data and the like of the system. For example: the index values of the number of concurrent users, the response time, the TPS, the stable operation time and the resource utilization rate are determined as follows:
the number of concurrent users = (total number of users/statistical time) × influence factor. The influence factor can be set according to actual conditions and can be set to be 5. Such as: in a card punching system, the number of card punching persons per day is 5W, the early peak is 8 to 9 points per day, and the late peak is 18 to 19 points per day, so that the number of concurrent users = (50000/(2 × 60 × 60)) × 5 ≈ 35.
The response time depends on the requirements of users, the response time acceptable to the users is different for different service systems, the response time needs to be unified for specific service scenes, and the response time index is usually predefined by adopting a 2/5/8 second principle.
TPS may be predefined according to a 2/8 principle, specifically meaning that 80% of the traffic is completed in 20% of the time, then TPS = traffic x 80%/(8 x 20% × 3600) (8 hours a day).
The plateau run time is usually measured in 7 x 24 hours.
Resource utilization, such as CPU utilization less than or equal to 75%, memory utilization less than or equal to 75%, etc.
Because the use frequency and the importance degree of different service functions are different, the performance indexes not only comprise the overall performance indexes of the system, but also comprise the performance indexes of different service functions. Accordingly, the determination of the expected performance metric value also includes the determination of the overall system performance metric value and the various business function performance metric values. For example, the overall performance index of the system is: 1) 20% of users log in, 30% of users browse home pages, and 50% of users are in a query state, and the online throughput rate of 100 users of the system reaches 10 transactions/second; 2) And (3) stable running time of the system: 7 × 24 hours; 3) Resource utilization rate: CPU utilization rate is less than or equal to 75%, memory utilization rate is less than or equal to 75%, and the like. Performance indexes of different service functions are as follows: 1) The single-point operation response time of the main service function of the system is within 2 seconds; 2) The system supports 100 user concurrence, and the response time is within 5 seconds; 3) The resource utilization rate is as follows: CPU utilization rate is less than or equal to 75%, and memory utilization rate is less than or equal to 75%.
The service scene model establishing module establishes service scene models including a single scene test model, a mixed scene test model and a service matching model according to service scenes and performance indexes thereof and by combining factors such as user quantity, active time, access frequency and scene interaction under different service functions, as shown in fig. 2.
The single-scene test model is used for testing whether the single service processing capacity of the system reaches the expectation or not and acquiring performance indexes such as the optimal load, the maximum TPS (transfer protocol security), the average response time and the like which meet the expectation indexes, and aims to find defects and position bottlenecks. And the business proportioning model is used for analyzing and extracting according to the operation log of the daily user, determining the business model actually accessed by the online user, and setting the proportion according to the business model in a mixed scene. Such as: 100 online users exist in a preset time period of the system, wherein 30 users browse the home page, 50 users inquire related contents, 20 users log in, and at the moment, a business matching model of home page browsing, inquiry and login is 3:5:2. the mixed scene model is established based on a business matching model of the actual access condition of the user, accords with the actual business requirement, can simulate the real production environment of the system, and tests whether the comprehensive business processing capacity of the system meets the expected requirement or not, so as to evaluate the performance of the whole system.
The single scene test model comprises a single scene benchmark test and a single scene concurrent test.
The single-scenario benchmark test is used for testing whether a single service meets the performance requirements expected by a system design or a user, and acquiring various response indexes of the single user during operation, and aims to establish a measurable reference standard to provide a comparison reference for other test scenarios or an optimization process, for example, when the software and hardware environment of the system changes, the influence of the change on the performance can be determined by performing the benchmark test again. The test method comprises the following steps: under the condition that the system is not stressed, a single user iteratively executes a single service for a period of time or times, and response indexes of the service during operation are obtained to serve as analysis measurement indexes. For example: performing single scene benchmark test on the login operation, wherein the specific test is as follows: the single user iteratively executes login operation for 5 times, and the average value of the best three times is taken as a response index for analysis; or, the single user executes the login operation for 1 minute, and an average value within 1 minute is acquired as a response index for analysis.
The single scene concurrent test is based on a benchmark test, user behavior simulation is carried out by increasing the number of users, a single service is operated concurrently, response indexes of the service during operation are obtained, performance of the service is observed, and whether concurrency problems exist is verified. The test method comprises the following steps: increasing pressure by increasing the number of users continuously until a certain resource item of the server reaches saturation (for example, the utilization rate of a CPU reaches 90%) or a certain index reaches an expected index, and acquiring the optimal load which can be borne by the service to be tested, wherein under the optimal load, the system response index meets the expected index. In addition, the optimal load can be used as constant concurrent decompression, whether the system has errors under the condition of high load of the server is tested, whether the response index exceeds an expected index, whether the capability of providing continuous service can be provided, and the like, so that the stability of the tested service can be verified. For example: the number of users increases progressively with anticipated number of concurrent users U, constantly pressurizes the system, and until certain index reaches the expectation, the step is as follows:
(1) And simultaneously operating the same service by the user number U, acquiring a response index within preset time, simultaneously setting m =1, and judging whether the response index exceeds an expected index, if so, adjusting and optimizing the system.
(2) If not, the number of users is increased progressively by U, namely, (m + 1) xU users operate the same service at the same time, response indexes within preset time are obtained, whether the response indexes exceed expected indexes is judged, and if yes, the optimal load meeting the expected indexes is searched between [ m x U, (m + 1) xU ] by a dichotomy principle; if not, m = m +1, repeating the step (2) until a certain index reaches the expectation.
(3) And (3) applying pressure to the optimal load as constant concurrence, so that the resources of the server are under high load and continuously run for a long time, and testing whether the system has errors or not and whether the system can provide continuous service capability or not, thereby verifying the stability of the tested service. If the system makes a fault, or the response index exceeds the expected index, the system needs to adjust the optimization process.
Taking the registration of one web site as an example, a concurrency test is performed (the expected number of concurrent users is 100). The method comprises the following steps:
(1) The user number 100 is used for simultaneously executing login operation, response indexes within 5 minutes are obtained, meanwhile, m =1 is set, and it is judged that the response indexes do not exceed expected indexes;
(2) The number of users is increased by 100, namely 200 users execute login operation at the same time, response indexes within 5 minutes are obtained, and the response indexes are judged not to exceed expected indexes;
(3) At this time, m = m +1=2, at this time, the number of concurrent users is 300, login is performed simultaneously, a response index within 5 minutes is obtained, at this time, the response index exceeds the expected index, then between [200, 300], an optimal load meeting the expected index is found according to the dichotomy principle, and the optimal load under the current test environment is obtained as 224.
(4) And (2) applying pressure by taking 224 as a constant concurrent way, so that the resources of the server are under high load and continuously run for 2 hours for a long time, the response index does not exceed the expected index, and the detected service is relatively stable.
The mixed scene test model comprises a mixed scene benchmark test, a mixed scene concurrent test and a stability test.
And performing benchmark test, namely testing whether the multiple combined services meet the performance requirements expected by a system design or a user, and acquiring various indexes of a single user during operation, wherein the aims are to quantize indexes such as response time and TPS (TPS), facilitate subsequent comparison and provide reference basis for performance analysis such as multi-user concurrency and stability test. The test method comprises the following steps: the method comprises the following steps: under the condition that the system is not stressed, a single user iteratively executes a plurality of combined services for a period of time or times to obtain a response index when each service runs. For example: and carrying out benchmark test on login, home page browsing and query services. The specific tests are as follows: performing iterative operation login, home page browsing and business query for 5 times by a single user, and taking the average value of the best three times as a response index for analysis; or, the single user operates to log in, browse home pages and inquire services for 1 minute, and the average value within 1 minute is obtained as a response index for analysis.
And (3) performing concurrent test on a mixed scene, performing user behavior simulation by increasing the number of users based on benchmark test, setting a proportion according to a service proportioning model on a certain scale of data volume to concurrently operate a plurality of services, acquiring a response index when the services run, observing performance of the services, and verifying whether concurrency problems exist. The actual production environment can be truly simulated under the scene. The data size of a certain scale refers to the data size of a certain scale already existing in the system, such as 100 ten thousand of data; business operations are performed on a certain scale of data volume, such as operations of querying and browsing one piece of data by a user in 100 ten thousand data volume. The test method comprises the following steps: increasing pressure by continuously increasing the number of users until a certain resource item of a server is saturated (for example, the CPU utilization rate reaches 90%) or a certain index reaches an expectation, and acquiring the optimal load which can be borne by the tested system, wherein under the optimal load, the system response index meets the expectation index; the optimal load can be used as constant concurrent decompression, so that the resources of the server are under high load and continuously run for a long time, whether the system has errors or whether the response index exceeds an expected index or not, whether the capability of providing continuous service can be provided or not can be tested, and the stability of the system can be verified. For example: the number of users increases progressively with the anticipated number of concurrent users U, constantly pressurizes the system, and until a certain index reaches expectation, the step is as follows:
(1) The user number U operates a plurality of combined services according to the service proportioning model setting proportion, response indexes in preset time are obtained, m =1 is set, whether the response indexes exceed expected indexes or not is judged, and if yes, the system needs to be adjusted and optimized.
(2) If not, the number of users is increased by U, namely (m + 1) multiplied by U users operate a plurality of combined services according to the set proportion of the service proportioning model, response indexes in preset time are obtained, whether the target indexes are exceeded or not is judged, and if yes, the best load meeting the expected indexes is searched between [ m multiplied by U, (m + 1) multiplied by U ] by the dichotomy principle;
(3) If not, m = m +1, repeating the step (2) until a certain index reaches the expectation.
(4) And (3) applying pressure to the optimal load as constant concurrence, so that the resources of the server are under high load and continuously run for a long time, and testing whether the system has errors or not and whether the system can provide continuous service capability or not, thereby verifying the stability of the system. If the system makes a fault, or the response index exceeds the expected index, the system needs to adjust the optimization process.
And stability testing, namely obtaining a more reasonable user number as a load of the stability testing based on a mixed scene concurrent test, running for a relatively long time, obtaining the performance of the stability testing, and verifying the capability of the server for continuously providing stable service. The test method comprises the following steps: firstly, determining the load of a stability test; based on a mixed scene concurrency test, searching an inflection point by observing a TPS trend graph; when the number of concurrent users increases to a certain number, the response time increases, the TPS is reduced, the current concurrent user is the inflection point in the performance test, and the inflection point in the performance test can be used as the load pressure in the stability test. The method comprises the steps of applying pressure to a system in a constant concurrent mode, enabling resources of a server to be in a limit state and continuously run for a long time (for example, 7 multiplied by 24 hours), monitoring resource consumption of the server, database processing capacity and the like, and testing whether hidden problems of long response time, memory leakage, insufficient disk space and the like occur to the system under the condition of high load, so that the stability of the system is verified.
And the test execution module performs performance test on the tested system according to the service scene model to obtain a plurality of response indexes. Wherein the response index comprises response time, TPS, resource utilization rate, smooth operation time and the like.
The optimization judgment module determines whether the system needs to be optimized according to the response index and the expected index, gives factors needing to be optimized when the system needs to be optimized, and judges again after the optimization is adjusted until the requirements are met: and judging whether the response index exceeds an expected index, if so, adjusting the system. For example: when 500 users log in concurrently, the CPU utilization rate reaches 99%. In the example, the CPU utilization rate reaches 99 percent and exceeds the expected index (the CPU is less than or equal to 75 percent), and the system needs to be adjusted and optimized. And if not, performing the performance test of the next scene until the performance test of the current round is finished. For example, the single-scene concurrent test can be performed only if the response index in the single-scene benchmark test meets the expected index; the mixed scene test can be carried out only when the response index in the single scene test model meets the expected index; otherwise, the system is verified after being adjusted and optimized. And determining factors needing to be adjusted by the system according to the service scene, the response index and the expected index, and verifying after adjusting and optimizing. The factors for the adjustment include internal factors and external factors. Internal factors refer to performance problems caused by the software itself, such as codes, algorithms, SQL, etc.; external factors refer to performance issues due to software and hardware configuration, such as bandwidth, I/O, databases, middleware, operating systems, and the like. For example: when 500 users log in concurrently, the CPU utilization rate reaches 99%. In this example, when a user logs in, the user needs to interact with the database, and as the efficiency of the SQL statements is low, 500 performance problems of long time consumption when the user logs in at the same time, high CPU utilization rate and the like are caused; adjustments need to be made from internal factors due to performance issues caused by the software itself.
Claims (10)
1. A performance test system based on a service scene is characterized by comprising:
a service scene recognition module: for determining a business scenario;
a performance goal determination module: the system is used for determining the expected indexes and the performance targets corresponding to the service scenes, and quantizing the performance targets into performance indexes;
a business scene model establishing module: the system is used for establishing a service scene model according to a service scene and performance indexes thereof in combination with test parameters; the business scenario model comprises a single scenario test model, a mixed scenario test model and a business matching model, wherein the single scenario test model is used for testing single businesses, the mixed scenario test model is used for testing comprehensive businesses, and the business matching model is used for determining the proportion of each single business in the comprehensive businesses;
the test execution module: and the method is used for performing performance test in combination with the service scene model to obtain the response index.
2. The performance testing system based on service scenario of claim 1, further comprising an optimization judgment module: and determining whether the system needs to be optimized according to the response index and the expected index, and if so, giving factors needing to be optimized.
3. The business scenario-based performance testing system of claim 2, wherein the factors include internal factors and external factors.
4. A performance testing system based on service scenarios according to any of claims 1-3, wherein the service proportioning model determines the proportion of each service instance actually visited by the user based on log analysis of historical users, thereby determining the proportion of each individual service in the integrated service.
5. A performance test method based on a service scenario, which is applied to the performance test system based on the service scenario in any one of claims 1 to 4, is characterized by comprising the following steps:
step 1, identifying a service scene;
step 2, determining an expected index and a performance target according to a service scene, and quantizing the performance target into a performance index;
step 3, establishing a service scene model by combining test parameters according to a service scene and performance indexes thereof, wherein the service scene model comprises a single-scene test model, a mixed-scene test model and a service proportioning model;
and 4, performing performance test based on the service scene model to obtain a response index.
6. The performance testing method based on the service scenario as claimed in claim 5, wherein the performance index includes a system overall performance index and a single service performance index.
7. The performance testing method based on service scenario of claim 5, wherein the single scenario model comprises: single scene benchmark test and single scene concurrent test; the mixed scene test model comprises the following steps: the method comprises a mixed scene benchmark test, a mixed scene concurrent test and a stability test.
8. The performance testing method based on service scenario of claim 7,
the single-scene benchmark test comprises the following specific steps: under the condition that the system is not under pressure, a single user iteratively executes a single service for a period of time or times to obtain a response index when the service runs;
the single-scene concurrent test specifically comprises the following steps: based on a single scene benchmark test, user behavior simulation is carried out by increasing the number of users, a single service is operated concurrently, and a response index when the service runs is obtained;
the mixed scene benchmark test comprises the following specific steps: under the condition that the system has no pressure, a single user iteratively executes a plurality of combined services for a period of time or times to obtain a response index when each service runs;
the concurrent test of the mixed scene comprises the following specific steps: on the basis of a mixed scene benchmark test, user behavior simulation is carried out by increasing the number of users, on a certain scale of data volume, a plurality of services are operated concurrently according to the proportion set by a service proportioning model, and response indexes during service operation are obtained;
the stability test comprises the following specific steps: based on the mixed scene concurrent test, a certain number of users is obtained as the load of the stability test, the operation is carried out for a period of time, and the response index of the operation of the service is obtained.
9. The performance testing method based on the service scenario as claimed in claim 7 or 8, further comprising step 5, determining whether the system needs to be optimized according to the response index and the expected index, when the system needs to be optimized, giving a factor needing to be optimized, and judging whether the system needs to be optimized again after adjusting and optimizing until the requirement is met.
10. The performance testing method based on service scenarios as claimed in claim 9, wherein the performance testing execution sequence is: the method comprises the following steps of single-scene benchmark test, single-scene concurrent test, mixed-scene benchmark test, mixed-scene concurrent test and stability test, wherein the next test is executed after the current test meets the requirement.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211175645.XA CN115562978A (en) | 2022-09-26 | 2022-09-26 | Performance test system and method based on service scene |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211175645.XA CN115562978A (en) | 2022-09-26 | 2022-09-26 | Performance test system and method based on service scene |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115562978A true CN115562978A (en) | 2023-01-03 |
Family
ID=84743443
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211175645.XA Pending CN115562978A (en) | 2022-09-26 | 2022-09-26 | Performance test system and method based on service scene |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115562978A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116680165A (en) * | 2023-04-25 | 2023-09-01 | 厦门国际银行股份有限公司 | Interface performance testing method, device and equipment |
| CN116737554A (en) * | 2023-05-30 | 2023-09-12 | 福芯高照(上海)科技有限公司 | Intelligent analysis processing system and method based on big data |
| CN116991693A (en) * | 2023-09-27 | 2023-11-03 | 宁波银行股份有限公司 | Test method, device, equipment and storage medium |
| CN117520133A (en) * | 2024-01-03 | 2024-02-06 | 梅州客商银行股份有限公司 | Method and device for non-functional test of full-stack domestic database of bank core system |
-
2022
- 2022-09-26 CN CN202211175645.XA patent/CN115562978A/en active Pending
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116680165A (en) * | 2023-04-25 | 2023-09-01 | 厦门国际银行股份有限公司 | Interface performance testing method, device and equipment |
| CN116680165B (en) * | 2023-04-25 | 2024-01-23 | 厦门国际银行股份有限公司 | Interface performance testing method, device and equipment |
| CN116737554A (en) * | 2023-05-30 | 2023-09-12 | 福芯高照(上海)科技有限公司 | Intelligent analysis processing system and method based on big data |
| CN116737554B (en) * | 2023-05-30 | 2023-12-22 | 内蒙古蒙嘟嘟科技服务有限公司 | Intelligent analysis processing system and method based on big data |
| CN116991693A (en) * | 2023-09-27 | 2023-11-03 | 宁波银行股份有限公司 | Test method, device, equipment and storage medium |
| CN116991693B (en) * | 2023-09-27 | 2023-12-26 | 宁波银行股份有限公司 | Test method, device, equipment and storage medium |
| CN117520133A (en) * | 2024-01-03 | 2024-02-06 | 梅州客商银行股份有限公司 | Method and device for non-functional test of full-stack domestic database of bank core system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN115562978A (en) | Performance test system and method based on service scene | |
| CN108446210B (en) | System performance measurement method, storage medium and server | |
| CN106803799B (en) | Performance test method and device | |
| CN111459698A (en) | Database cluster fault self-healing method and device | |
| CN110377519B (en) | Performance capacity test method, device and equipment of big data system and storage medium | |
| CN111582488A (en) | Event deduction method and device | |
| CN115237804A (en) | Performance bottleneck assessment method, performance bottleneck assessment device, electronic equipment, medium and program product | |
| US10884909B1 (en) | Code release with automatic callback | |
| US20120323926A1 (en) | Efficient Optimization over Uncertain Data | |
| CN109101403A (en) | A kind of pair of mobile terminal generates the method and system that SQL is monitored in real time | |
| CN113360353A (en) | Test server and cloud platform | |
| RU2532714C2 (en) | Method of acquiring data when evaluating network resources and apparatus therefor | |
| CN111177640A (en) | Data center operation and maintenance work performance evaluation system | |
| CN115952098A (en) | Performance test tuning scheme recommendation method and system | |
| CN112860531B (en) | Block chain wide consensus performance evaluation method based on deep heterogeneous graph neural network | |
| CN114881503A (en) | Scoring determination method, device, equipment and storage medium | |
| CN113610225A (en) | Quality evaluation model training method and device, electronic equipment and storage medium | |
| Jehangiri et al. | Distributed predictive performance anomaly detection for virtualised platforms | |
| CN113742243B (en) | Application evaluation method, device, electronic equipment and computer readable medium | |
| US9921900B1 (en) | Methods and apparatus for system monitoring | |
| CN116401113B (en) | Environment verification method, device and medium for heterogeneous many-core architecture acceleration card | |
| Lei et al. | Data mining and statistical analysis on smart city services based on 5G network | |
| CN117131791B (en) | Model evaluation method, system and storage medium based on big data platform | |
| CN111324542B (en) | Web application regression test case selection system, method and equipment | |
| CN113656299B (en) | Method and device for determining limit QPS, electronic equipment and readable storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |