Introduction

Testing becomes crucial as the complexity and the adoption of a solution grows. In Kubecon NA 2020, Jian from Airbnb talked about

Why is functional parallelisation needed?

For example, we have a banking solution for which we want to make two-factor authentication mandatory for all users at the time of sign-in. Now testing this functionality across multiple browsers sequentially would not only take a lot of time but also resources. Here the input would be a parameterized user for which the two-factor authentication is enabled and happens successfully in a minimal amount of time.

With functional parallelisation, we can test this functionality with various inputs across multiple browsers at the same time. Suppose the test fails for any of the browsers, we would get to know in a lot less time and work on identifying the root cause along with its solution.

Enabling functional parallelisation also has some challenges, let us understand these.

Challenges with functional parallelisation

• Test Isolation: Each test may have its own set of dependencies, configurations, and compatibility considerations. Setting up and configuring multiple test environments for different tests can be time-consuming and complex, especially when ensuring consistency across environments.




• Recording Results: In the previous example, we would require a system that can record different input cases and the combination browser and the actual results. These recorded results would then be used to compare with the expected result to determine the performance of functionality.




• Resource Allocation: Identifying which functionality test would require more resources and then distributing the resources accordingly could be a challenging task. Suppose for a banking solution, the sign-in and transaction functionalities have to be tested. Now these tests take time even though run in parallel so we need to determine the resource utilization limit for each.

Seeing the challenges associated with load and functional parallelisation, we need a testing solution that automates testing with parameterized virtual users or other inputs across multiple browsers or operating systems simultaneously. Also, the solution should allow dynamic resource allocation based on test requirements and optimize resource utilization. For systems under test, the testing solution should provide real-time visualization of tests and aggregated results to perform result analysis.

Parallel Testing with Testkube

Testkube recently , and can be created using both extensive Wizards and samples for different scenarios which can then be further enhanced in any way required to fulfill your testing requirements using the .

Workflows for Load Parallelisation

As described above, running a load test from multiple nodes in parallel is often required to simulate large amounts of user traffic. The parallel keyword in Test Workflows allows you to parallelize any load-testing tool or script for this purpose, for example:

• K6 - parallel can be used to distribute and parameterise k6 scripts far beyond what can be achieved with the k6-operator -

  
  
  

  Examples for distributed load-testing with several of these tools are available in the

  
  
  

  As above, examples for distributed functional testing with these tools are available in the

  
  
  

  Read more about Composite Testing in .

  
  
  

  
  
  
  Troubleshooting, Artifacts and Reporting
  

  
  
  

  Testkube automatically captures the log output of any testing tool it runs on all nodes the tools are running on, to help you ensure that tests were executed as desired and debug issues with your executions. Furthermore, Testkube can be configured to capture any artifacts produced by your tests, including aggregate reports, videos, JUnit-reports, etc. When captured, Testkube will make these available via the Dashboard.

  
  
  

  to generate massive load, or the parallelisation of functionali/E2E tests to increase test coverage and improve test execution times.
  
  The recently introduced Test Workflows engine in Testkube allows for parallelisation of any testing tool, both load and functional, with additional support for sharding, parameterisation and management of dependent services to execute your tests.

  
  
  

  Head over to testkube.io/get-started to learn more and give Testkube a try using either our demo environment or your existing tests running in your own infrastructure.

  Vollständiger Original-Bericht

  Ausführliche Details, Code-Beispiele & Hersteller-Stellungnahme auf dev.to.

  ↗ Original-Artikel auf dev.to lesen