Create an architecture for web test automation
5 likes1,382 views
The document outlines best practices and strategies for automated testing in web applications, including design patterns, testing architecture, and data generation techniques. Key topics include the use of page object models, fluent interfaces, logging strategies, parallel test execution, and data generation methods for both sensitive and non-sensitive information. It emphasizes the importance of clear reporting, exception handling, and the use of tools like Java Faker for effective test management.
Create an architecture for web test automation
- 1. ELIAS NOGUEIRA TATIANE NOGUEIRA @eliasnogueira @tatianeaguirres
- 2. TATIANE NOGUEIRA Consultant Developer @ Thoughtworks @tatianeaguirres linkedin.com/tatianeaguirres ELIAS NOGUEIRA Software Engineer in Test @ Sicredi @eliasnogueira linkedin.com/eliasnogueira
- 3. THE STACK LANGUAGE TESTING INFRADATA GENERATION javafaker LOG & REPORTS ExtentReports
- 4. PAGE OBJECT MODEL • Page Objects • Page Factory • Abstraction • Waiting Strategy PARALLEL EXECUTION • Infrastructure • Containers LOGS AND REPORTS • Exception logs • General reports • Evidence DATA GENERATION • Fake • Static creation • Dynamic creation PIPELINE • Execution strategy BASE ARCHITECTURE • Clean Architecture • Design Patters • Testing Patterns BASIC ITEMS FOR A TEST ARCHITECTURE with focus on web automation
- 5. an abstract class that will take care of commons actions in your automated tests FACTORY the Design Pattern to create, in our case, browser instances LISTENER a non-intrusive way to know what is happening during the test execution BASE TEST BASE ARCHITECTURE to apply DRY and KISS
- 6. Smart use of inheritance • test inherit common test actions One test case per class • provide an easy way to add more tests • ease division of tests in suites BASE TEST CLASS TEST 1 TEST 2 TEST N • browser initialization/close • open/close database, logs … • connect/disconnect servers • login/logout app BASE TEST CLASS
- 7. Apply Factory Design Pattern will help us to create a browser instance and make easy the parallel execution against many environments. BROWSER FACTORY chrome firefox edge FACTORY CLASS safari
- 8. Using TestNG we can use some listeners that allow modifying (or just watch) the test behaviors. Helpful o watch test lifecycle and do something. LISTENERS MY TEST LISTENER • test start • test finish • on test fail • on test skipped • on start • on finish • on success TEST 1 @MyTestListener TEST 2 @MyTestListener TEST N @MyTestListener
- 9. way to create more readability service class LOAD STRATEGY making the code wait for async executions FLUENT INTERFACE create tests in a fluent way PAGE FACTORY PAGE OBJECTS MODEL more maintainability and readability
- 10. TESTSPAGE OBJECTSAPP PAGES PAGE OBJECTS Page Object is a class that serves as an interface to a page of your web page. The class provides methods to do the page actions. Tests will use these methods. PAGE OBJECT FLIGHT SELECTION PAGE OBJECT SEARCH PAGE OBJECT PAYMENT SUCCESSFUL BOOK INVALID DATES SEARCH PAGE FLIGHT SELECTION PAGE PAYMENT PAGE PAYMENT PROBLEM
- 11. LOAD STRATEGY A Load Strategy is responsible for wait for a certain time by any event on the web page, most of the time related to async requests (Ajax). PAUSE IMPLICITLY any type of sleep that will pause the execution you’ll won’t know, in your code witch action will wait EXPLICITLY AJAX LOCATORthe best choice to use with Page Factory strategy with this strategy you can see, in the code, witch element will take time
- 12. FLUENT INTERFACE Creates a method chaining to perform a series of actions to make the code more readable and easy to use. @Test public void testWithoutFluentInterface() { GeneralMenuPage menu = new GeneralMenuPage(); menu.clickinExperience(); menu.clickInOurFleet(); menu.clickInSeatingCharts(); } @Test public void testWithFluentInterface() { GeneralMenuPage menu = new GeneralMenuPage(); menu.clickinExperience().clickInOurFleet().clickInSeatingCharts(); }
- 13. know all the exceptions to solve the problems root-cause GENERAL REPORTS evidence and executive reports EXCEPTION LOGS LOGS AND REPORTS because we need to know about any error
- 14. By using any log strategy, saving a log file, we can understand the common errors occurred during the test execution. These errors can be of: • assertion errors • timeout exceptions • locator exception • an exception on your architecture If you want to analyze test errors across teams a good way is using Elasticsearch with Grafana/Kibana. EXCEPTION LOGS
- 15. Generate xUnit reports to attach on your CI/CD and, rapidly, see the test status. GENERAL REPORTS Create an executive report to provide information and evidence about the test execution. This report may contain screenshots when an error occurs to help to analyze the root cause of a problem.
- 16. pass the responsibility of non-sensitive data generation to a framework STATIC/DYNAMIC GENERATION create the sensitive data and put under your control FAKES DATA GENERATION solve one of the biggest problems
- 17. Ability to create an approach to generate non-sensitive data for your test without the necessity to manually change the test data in each execution. There’re a lot of tools to create these type of data. FAKE GENERATION Example with javafaker Faker faker = new Faker(new Locale("pt-BR")); faker.name().fullName(); faker.address().fullAddress(); faker.internet().emailAddress(); faker.business().creditCardNumber(); faker.date().birthday();
- 18. When the data cause different behaviors in your application. STATIC / DYNAMIC GENERATION A Static approach can be implemented with any kind of solution, like: • Files • CSV | JSON | TXT | YML • Database • Mock A Dynamic approach can be created according to your context. Used for remove the maintenance of test data • Queries in a database • Consume data from a static poll
- 19. run many tests at the same time in a chosen target GRID AND AUTO-SCALE using the proper containers, we can speed up the test execution PARALLELISM PARALLEL EXECUTION to speed up your test execution
- 20. maven-surefire- plugin Have an ability to control how many threads we need inside the pom.xml Junit 4 Has an experimental class called ParallelComputer TestNG Control the parallelism thought the suites in any level of tests (class, methods, etc..) Parallelism, under test, is the ability to perform the same test in different conditions (browser, devices, etc...) or different tests at the same time. PARALLELISM
- 21. GRID SCHEMA Node Windows Node MacOSX Node Linux Test Script Hub send capabilities understands the capabilities and send to proper node
- 22. WAYS TO CREATE A GRID LOCAL Uses machines inside an infrastructure. Can be a bare-metal desktop or a virtual machine CLOUD Uses a cloud infrastructure platform to create virtual machines CONTAINERS Uses containers (locally or cloud-based) to create the infrastructure and support orchestration
- 23. CONTAINERS TO AUTO-SCALE • has containers for each aspect of the grid • selenium-hub • selenium-node-chrome • selenium-node-firefox • auto-scale based on hardware utilization or with some in-house solution • Uses a custom container elgalu/selenium that provides: • live Preview with VNC • video recording • dashboard • automatic auto-scale containers based on the number of tests SELENIUM
- 24. create a pipeline for any type of test execution DIVIDE ALL TYPES OF EXECUTION PIPELINE make the execution process clear
- 25. FUNCTIONAL TESTACCEPTANCE TESTSMOKE TEST DIVIDE ALL TYPES OF EXECUTION WEB PART IN THE PIPELINE Most important tests in a business perspective Most used user scenarios Assure that critical functionalities works each build your determination e.g.: release Your determination e.g.: release
- 26. THANK YOU! TATIANE NOGUEIRA Consultant Developer @ Thoughtworks @tatianeaguirres linkedin.com/tatianeaguirres ELIAS NOGUEIRA Software Engineer in Test @ Sicredi @eliasnogueira linkedin.com/eliasnogueira https://github.com/eliasnogueira/public-speaking