Continuous API testing within your CI pipeline

You probably want to run your tests automatically on a Continuous Integration (CI) environment. Good news, our CLI test runner integrates well with most CI solutions, all you need is the latest Java JDK and Maven installed on your CI machine.

Check in our CI/CD solution list to see if yours integrates well with Restlet Client or influence your choice of a solution:

CI/CD solution Maven integration level Comments
Jenkins ★★★ Natively supported, optional Maven integration plugin
Travis CI ★★★ Natively supported
CircleCI ★★★ Natively supported
Codeship ★★★ Natively supported
Team Foundation Server ★★★ Natively supported
Teamcity ★★★ Natively supported
Bamboo ★★★ Natively supported
Bitbucket Pipelines ★★★ Natively supported
Semafore CI ★★★ Natively supported
Shippable ★★★ Natively supported
Wercker ★★☆ Via a Docker image
Drone IO ★★☆ Via a Docker image
CodeFresh ★★☆ Via a Docker image
Gitlab CI ★☆☆ Via low-level configuration
GoCD ★☆☆ Via low-level configuration
AppVeyor ☆☆☆ Not supported
PHP CI ☆☆☆ Not supported

Other CI/CD solution may work, the list above is only the list we have in our radar. If you have another working solution please contact us so we can update this page.

CI workflow

Your API tests should follow the following flow:

Test flow

You should build your API from the sources, then run unit tests, deploy it and finally run integration tests against it using Restlet Client.

Technically, there are two main ways to do so which we will explain and comment below. If you want more precise instructions on how to configure your CI builds, have a look at our tutorials on Travis CI and CircleCI . They should get you started in no time.

Basic workflow: test your main branch once a day

Basic workflow

If you choose to test your API on a static environment, you can run the test on demand or schedule your job to run nightly. This solution is easy to implement and does the job but note that:

  • only one test can run at a time
  • the data in your API gets "polluted" when running a test and you must clean it before running another test or they won't run in the same conditions
  • anybody with access to that environment can request the API while the test is running, which can result in hard-to-debug test failures

If you are testing on a staging environment, you can create a CRON task that runs once a night and make it:

  • pull the API code and test file from GitHub
  • build and test your API code
  • deploy it to your environment
  • run your API tests against it with our maven plugin
  • get your test reports

Once the job exists, you should be able to run it on-demand too.

Advanced workflow: check your API is OK on each commit

Advanced workflow

A more advanced solution is to deploy your API in a container for each job run. That allows you to:

  • test multiple branches at the same time
  • work in a compartmentalized environment, only the test can mess with it
  • test every commit

If you are testing your API in containers, you can create a task that runs each time the source control is updated and make it:

  • build and unit tests the API code
  • deploy it in a container
  • run your API tests against it with our maven plugin
  • get your test reports