Performance Testing Guidelines

Use the following guidelines to construct effective performance tests for the Archer Platform:

• When adding new content to an application, consider the anticipated number of records that will be added each year in your production environment. Performance tests can generate more activity in an application over several minutes than would typically occur in a year of use in production.
• Consider how a real user will work when using an application.
  • Users do not typically log in, perform a single transaction, and then immediately log out. Tests should simulate multiple actions during each login session, over a set period of time.
  • All users in a production environment do not typically log in simultaneously. Tests should simulate multiple user logins spread across a realistic period of time.
  • Users typically require time to think between the steps of an operation. Tests should simulate realistic delays between steps in each operation, such as between a lookup action and the selection of a returned value.
  • Users typically do not perform a Save operation after modifying each individual field in an application or a questionnaire. Tests should simulate Save operations spread across a realistic period of time.
  • Consider the flow of the application when designing a load test. For example, executing a Save and Close operation for a content record will return the user to the search page from which the record was opened or added, and execute the search again. Tests should simulate whether users would typically execute another search from the page, or continue to work with the current search results.
  • Consider whether references to related records in an application will be spread across a range of records, or whether they will all reference the same record. Tests should simulate references to related records as they will likely exist in your production deployment.