Added full-link stress testing platform probe management to help businesses achieve 10-fold performance improvement!

The full-link stress test was born in the preparation process of Alibaba’s Double 11. If the Double 11 promotion is the “final exam” of Ali’s business, the full-link stress test is a “mock exam” before the big exam. 11 "Nuclear Weapons" for Stability Assurance. The full-link stress test performs high-simulation simulations of high-traffic business scenarios in the production environment to obtain the most realistic online actual carrying capacity, execute accurate capacity planning, and ensure system availability.

The distributed architecture and rapid business development have brought uncertainty to the business system. Any node in a distributed environment may become a bottleneck/short board/problem. At the same time, system availability faces more severe challenges and uncertainties with the rapid growth of business.

For example, single-link stress testing lacks external interference and various resource competitions, and the results of single-link stress testing are generally optimistic and cannot reflect the real system carrying capacity. Some problems will only be exposed under real heavy traffic, such as network bandwidth, inter-system impact, basic dependencies, and so on. The full-link stress test is not only a stress test, but also a real big promotion rehearsal, a unified acceptance of high-availability solutions such as pre-plan drills, current limiting verification, and destructive drills.

This automatic test platform 1.0.32 update adds the online management function of the probes of the full-link stress test platform. How is it realized? Let's look down together~

1. Added online management of full-link pressure test platform probes

Basic overview: Full-link stress testing is based on actual production business scenarios and system environments, simulating massive user requests and data to conduct stress tests on the entire business chain, trying to find potential system performance bottlenecks and continuously optimize the process.

A probe is a tool embedded in an application that can be used to monitor and diagnose the performance and behavior of the application. The full-link stress testing platform implements this function in a code-free way based on java Agent technology.

2. Added full-link stress testing platform application management

Basic overview: The application management of the full-link stress testing platform supports viewing the tested applications and probes, and setting shadow libraries/shadow tables for applications to achieve data isolation and prevent stress testing data from polluting production data.

 Operation example:

Step 1: Call an interface in the application under test "easydemo-gateway".

picture

 

Note: After this operation, the probe will record all link information requested by the application interface under test.

 

Step 2: Click "Application Management" on the full-link stress testing platform to view the connected tested applications and the number of probes.

picture

Step 3: Click the application "Details".

picture

 

 

Note: The application is responsible for writing to the database in the interface request link, so a new shadow library is created in the application.

 

Step 4: Click Add Shadow Library/Table.

picture

 

Note: The function of the newly added shadow database/table is to write the data of the full-link stress test into the shadow database/table to prevent the stress test data from polluting the real data.

 

Step 5: Click "OK".

picture

 

Step 6: Add a shadow table with the same structure as the official data table in the database of the application under test, and the naming format is "shadow_formal data table name".

picture

 

 

illustrate:

1. The naming format is fixed "shadow_formal environment data table name". If the official environment database is named user, the shadow table is named shadow_user, so that the program will write the data of the full link stress test into the table without Write to the official user table. The name must be consistent and cannot be mistaken.

2. If the new one is a shadow library, this step will add a shadow library that is the same as the official database, and the naming format is "shadow_official database name".

Notice:

If no shadow database/table is created, the data of the full-link stress test will be written into the database table of the formal environment, which will pollute the data of the real environment.

3. Add link management of full-link stress testing platform

Basic overview: A link refers to the entire path of a request in a distributed system, including all service components participating in the request. The link management of the full link stress testing platform supports viewing the application link recorded by the probe, including the topology relationship and service component details of each service component in the link.

 Operation example:

Step 1: Call an interface in the application under test "easydemo-gateway".

picture

Step 2: Click "Link Management > Query" on the full link stress testing platform to view the request link of the interface in step 1.

picture

 

Explanation: Because the interface request calls two services, each application displays a link.

 

Step 3: Click "Details" of the "POST:/api/register" link.

picture

Step 4: You can view the complete link of each service component that the interface request passes through.

picture

Step 5: Click the service component to view the service details.

picture

 

 

 

 

4. New full-link stress test platform stress test task management

 

 

Basic overview: Executing a stress test task can initiate a stress test on the application link recorded by the probe to obtain a test report, helping users test the performance, reliability and stability of the link, identify performance bottlenecks, and optimize system performance.

 

 Operation example:

 

Step 1: Click Pressure Test Task > New Pressure Test Task.

picture

Step 2: Fill in the name of the stress test task, select a resource pool, and click "Select Link".

picture

 

 

Note: The resource pool needs to be configured in the system management in advance. For specific configuration, please refer to the chapter on system management of the full-link stress testing platform.

 

Step 3: Select the link and click "OK".

picture

Step 4: Fill in the pressure configuration as required, and click "OK".

picture

 

 

illustrate:

Concurrent mode: Concurrent access by the target number of users per unit time, used for stress testing system TPS.

QPS mode: A fixed number of requests are initiated per unit time, which is used to detect whether the system business activities can support the number of requests.

Target Users: The number of users per second or per minute.

Stress test duration: The stress test time. It is recommended that the stress test last no less than 2 minutes. After the stress test starts, it can be stopped manually at any time.

Request step rate: Ladder pressurization, according to the number of target users and pressure test duration/request step rate, increase step by step, if no step pressurization is required, set 1.

Continue running time: How long to continue running after reaching the target number.

 

Step 5: Click "More > Start Pressure Test" of the pressure test task.

picture

Step 6: After the startup is successful, it will automatically jump to the report details.

picture

Step 7: Check the shadow database. The data generated by the pressure test will be automatically written into the shadow database without interfering with the official database.

picture

 

 

 

 

5. New full-link stress testing platform system management

 

 

Basic overview: The system configuration supports the configuration of resource pools, and the resources in the resource pools can be used as machines that initiate stress tests to perform stress testing tasks.

 

 Operation example:

 

Step 1: Enter the "feisuanyz-pressure-engine" directory, double-click "startup.bat" to start the full link pressure test engine.

picture

Step 2: If the console log shows the icon information, the startup is successful.

picture

Step 3: Click "System Configuration > Resource Pool Management > Add Resource Pool" on the full-link stress testing platform.

picture

Step 4: Fill in the resource pool name and click "Submit".

picture

Step 5: Click "Resource Management".

picture

Step 6: Click "Add Resource".

picture

Step 7: Fill in the resource name and full link stress test engine service address, and click "Submit" to complete the configuration of the resource pool.

picture

 

The above is the whole content of this update. If you are interested, you can apply for a free trial: http://feisuanyz.mikecrm.com/9dW4GeZ

Graduates of the National People’s University stole the information of all students in the school to build a beauty scoring website, and have been criminally detained. The new Windows version of QQ based on the NT architecture is officially released. The United States will restrict China’s use of Amazon, Microsoft and other cloud services that provide training AI models . Open source projects announced to stop function development LeaferJS , the highest-paid technical position in 2023, released: Visual Studio Code 1.80, an open source and powerful 2D graphics library , supports terminal image functions . The number of Threads registrations has exceeded 30 million. "Change" deepin adopts Asahi Linux to adapt to Apple M1 database ranking in July: Oracle surges, opening up the score again
{{o.name}}
{{m.name}}

Guess you like

Origin my.oschina.net/u/4868096/blog/10087099