MicroShed Testing

microshed-testing

alt text

Get going on true-to-production tests without the hassle

Why use MicroShed Testing?

MicroShed Testing offers a fast and simple way of writing and running true-to-production integration tests for Java microservice applications. MicroShed Testing exercises your containerized application from outside the container so you are testing the exact same image that runs in production.

MicroShed Testing aims to:

  1. be easy to get started with
  2. work with any Java EE, Jakarta EE or MicroProfile runtime
  3. provide true-to-production tests

What it looks like

Assume we have a basic JAX-RS application that can perform create, update, and delete operations on ‘Person’ data objects. It may look something like this:

@Path("/people")
@ApplicationScoped
@Produces(MediaType.APPLICATION_JSON)
@Consumes(MediaType.APPLICATION_JSON)
public class PersonService {

    private final PersonRepo personRepo = // ...

    @GET
    public Collection<Person> getAllPeople() {
        return personRepo.values();
    }

    @GET
    @Path("/{personId}")
    public Person getPerson(@PathParam("personId") long id) {
        Person foundPerson = personRepo.get(id);
        if (foundPerson == null)
            throw new NotFoundException("Person with id " + id + " not found.");
        return foundPerson;
    }
    
    // ...
}

Using MicroShed Testing, we can write an integration test that looks something like this:

import static org.junit.jupiter.api.Assertions.*;
import jakarta.ws.rs.NotFoundException;
import org.junit.jupiter.api.Test;
import org.microshed.testing.jaxrs.RESTClient;
import org.microshed.testing.jupiter.MicroShedTest;
import org.microshed.testing.testcontainers.ApplicationContainer;
import org.testcontainers.junit.jupiter.Container;

@MicroShedTest
public class BasicJAXRSServiceTest {

    // This will search for a Dockerfile in the repository and start up the application
    // in a Docker container, and wait for it to be ready before starting the tests.
    @Container
    public static ApplicationContainer app = new ApplicationContainer()
                    .withAppContextRoot("/myservice");

    // This injects a REST _Client_ proxy of the PersonService shown above
    // This allows us to easily invoke HTTP requests on the running application container
    @RESTClient
    public static PersonService personSvc;

    @Test
    public void testGetPerson() {
        // This invokes an HTTP POST request to the running container, which triggers
        // the PersonService#createPerson endpoint and returns the generated ID
        Long bobId = personSvc.createPerson("Bob", 24);
        
        // Using the generated ID, invoke an HTTP GET request to read the record we just created
        // The JSON response will be automatically converted to a 'Person' object using JSON-B 
        Person bob = personSvc.getPerson(bobId);
        
        assertEquals("Bob", bob.name);
        assertEquals(24, bob.age);
        assertNotNull(bob.id);
    }
    
    @Test
    public void testGetUnknownPerson() {
        // This invokes an HTTP GET request to get a person with ID -1, which does not exist
        // asserts that the application container returns an HTTP 404 (not found) exception
        assertThrows(NotFoundException.class, () -> personSvc.getPerson(-1L));
    }

    // ...
}

If the repository containing the tests does not have a Dockerfile in it, there are a few other options:

  • If the application’s container image is produced by a different repository, a String docker image label can be supplied instead:
    @Container
    public static ApplicationContainer app = new ApplicationContainer("myservice:latest")
                    .withAppContextRoot("/myservice");
  • If a Dockerfile or container image label is not available, it is possible to use vendor-specific adapters that will provide the default logic for building an application container. For example, the microshed-testing-liberty adapter will automatically produce a testable container image roughly equivalent to the following Dockerfile:
FROM openliberty/open-liberty:full-java17-openj9-ubi
COPY src/main/liberty/config /config/
ADD target/$APP_FILE /config/dropins

Quick Start

To get started writing a test with MicroShed Testing, add system-test and junit-jupiter as test-scoped dependencies: microshed-testing-core supports the Javax namespace up to and including version 0.9.2. Starting from version 0.9.3, the Jakarta namespace is supported.

<dependency>
    <groupId>org.microshed</groupId>
    <artifactId>microshed-testing-testcontainers</artifactId>
    <version>0.9.2</version>
    <scope>test</scope>
</dependency>

<!-- Any compatible version of JUnit Jupiter 5.X will work -->
<dependency>
    <groupId>org.junit.jupiter</groupId>
    <artifactId>junit-jupiter</artifactId>
    <version>5.10.1</version>
    <scope>test</scope>
</dependency>

Once you have the above dependencies added, create a new test class with the following items:

  1. Annotate the class with @MicroShedTest
  2. Create a public static ApplicationContainer field
  3. Inject one or more public static JAX-RS resource classes
import org.microshed.testing.jaxrs.RESTClient;
import org.microshed.testing.jupiter.MicroShedTest;
import org.microshed.testing.testcontainers.ApplicationContainer;
import org.testcontainers.junit.jupiter.Container;

@MicroShedTest
public class MyTest {

    @Container
    public static ApplicationContainer app = new ApplicationContainer()
                    .withAppContextRoot("/myservice");
                    
    @RESTClient
    public static MyService mySvc;
    
    // write @Test methods as normal
}

For a more complete introduction, see the Walkthrough page