Tests

Search…
Tests
You add tests when you want Garden to run your test suites for you. A simple configuration looks like this:
1
# garden.yml
2
kind: Module
3
tests:
4
  - name: unit
5
    args: [npm, run, test:unit]
6
  - name: integ
7
    args: [npm, run, test:integ]
8
    dependencies:
9
      - backend
Copied!
Note that not all modules types support tests.
How it Works
Tests belong to modules and each module can have several tests. Because tests are a part of the Stack Graph and dependency aware, you can easily run integration tests that require other parts of your stack to be running.
Garden caches test results and only re-runs the test if the module the test belongs to, or upstream dependents, have changed. For remote environments, the test results are stored at the cluster level so that the entire team can share the cached results.
You use the command and args directives to specify how the test is run. If the execution exits with 0, the test is considered to have passed, otherwise failed.
If you have expensive tests that you don't want to run on every watch event when in watch mode, you can use the --skip-tests flag or, alternatively, specify what tests to run with the --test-names flag.
You can run a test manually with the garden run test <test-name> command. This will run the test regardless of whether or not the result is cached.
You can view test results from the dashboard or by running garden get test-result <module-name> <test-name>.
Tests in the Stack Graph
Tests correspond to a test action in the Stack Graph.
Tests implicitly depend on the build step of their parent module.
Tests can depend on services and tasks.
Currently, nothing else can depend on tests.
Examples
For full test configuration by module type, please take a look at our reference docs.
Integration Testing
Below is an example of a frontend module that has a unit test and an integ test that depends on a backend module. The integ test checks whether the frontend gets the correct response from the backend. The example is based on our vote example project.
Here's the configuration for frontend module:
1
# garden.yml
2
kind: Module
3
type: container
4
name: frontend
5
tests:
6
  - name: unit
7
    args: [npm, run, test:unit]
8
  - name: integ
9
    args: [npm, run, test:integ]
10
    timeout: 60
11
    dependencies:
12
      - backend
Copied!
The integ test looks like this:
1
// tests/integ/test.js
2
describe('POST /vote', () => {
3
  it('respond with message from hello-function', async () => {
4
    const result = await axios.post('<http://backend/vote/>', `vote=a`);
5
    expect(result.status).to.eql(200);
6
  });
7
});
Copied!
Now when you're in watch mode and make a change to the frontend, Garden will re-run both the unit and integ tests for you.
If you make a change to the backend backend module, Garden will first re-build and re-deploy the backend, and then run the integ test defined for the frontend.
Advanced
Test Artifacts
Many module types, including container, exec and helm, allow you to extract artifacts after tests have been run. This can be handy when you'd like to view reports or logs, or if you'd like a script (via a local exec module, for instance) to validate the output from a test.
By convention, artifacts you'd like to copy can be specified using the artifacts field on test configurations. For example, for the container module, you can do something like this:
1
kind: Module
2
type: container
3
name: my-container
4
...
5
tests:
6
  - name: my-test
7
    command: [some, command]
8
    artifacts:
9
      - source: /report/*
10
        target: my-test-report
Copied!
After running my-test, you can find the contents of the report directory in the test's container, locally under .garden/artifacts/my-test-report.
Please look at individual module type references to see how to configure each module type's tests to extract artifacts after running them.
Disabling Tests
Module types that allow you to configure tests generally also allow you to disable tests by setting disabled: true in the test configuration. You can also disable them conditionally using template strings. For example, to disable a container module test for a specific environment, you could do something like this:
1
kind: Module
2
type: container
3
...
4
tests:
5
  - name: e2e
6
    disabled: ${environment.name == "prod"}
7
    ...
Copied!
Tests are also implicitly disabled when the parent module is disabled.
Kubernetes Provider
Tests are executed in their own Pod inside the project namespace. The Pod is removed once the test has finished running.
Tests results are stored as ConfigMaps in the garden-system--metadata namespace with the format test-result--<hash> and shared across the team.
To clear cached test results, you currently have to delete the ConfigMaps manually with kubectl. Here's an example:
1
kubectl delete -n garden-system--metadata $(kubectl get configmap -n garden-system--metadata -o name | grep test-result)
Copied!
Exec Modules
The exec module type runs tests locally in your shell. By default, the exec module type executes tests in the Garden build directory (under .garden/build/<module-name>). By setting local: true, the tests are executed in the module source directory instead.
Kubernetes and Helm Modules
Because a Kubernetes or Helm module can contain any number of Kubernetes resources, a serviceResource needs to be specified to determine the pod spec for the test pod. You can see the whole pod spec used in the reference docs for kubernetes and helm modules. Please note that the startupProbe, livenessProbe and readinessProbe are stripped from your pod spec. Health checks for your application might fail when the container is used for testing because the main process usually running in that container is replaced by the test command.
Next Steps
In the next section, we'll see how Garden can execute tasks for your. For example populating a database after it has been deployed.
Using Garden - Previous
Services
Next - Using Garden
Tasks
Last modified 26d ago
Copy link
Contents
How it Works
Tests in the Stack Graph
Kubernetes and Helm Modules
Next Steps