AutoTest Assist

This feature is available only in API Connect Enterprise as a Service

The AutoTest assistant exercises your API by sending it a series of randomly-generated requests, guided by the API definition and any custom rules that you provide, and alerts you about any conformance errors and server-side errors that it observes.

You control the auto-tester by creating and running an AutoTest profile, and you can create multiple profiles against the same API to cover different use-cases. A profile specifies how the auto-tester should run, which endpoints it should test, and what kind of data it should send. A default profile is constructed so that you can get immediate feedback by providing no more than a target URL and credentials.

The AutoTest assistant is especially useful while an API is under development as it can quickly identify errors in an implementation without you having to write test cases. It is also valuable later in the cycle for spotting regressions, as the random nature of its requests can provide coverage that your existing test cases do not match. However, it is not meant as a replacement for other testing tools in API Connect, but is rather complementary, and the information it provides when it identifies an error is often a good starting point for creating a test case.

Follow these steps to create and run your first profile:

Consult these additional guides for detailed instructions on how to configure the auto-tester to support your use case:

Known limitations

  • The API definition must conform to OpenAPI 2.0 (Swagger 2.0)
  • The API definition document must have a .json or .yaml extension (you cannot select a document with another extension)
  • If there are errors in the API definition then you will see only a blank screen, and the errors are not reported
  • You must provide either a username and password, or a bearer token, to authenticate against the target API; other authentication methods are not supported
  • You can run only one profile at a time