Testing
Testing offers a good way to ensure camel routes behave as expected over time. Before going deeper in the subject, it is strongly advised to read First Steps and Quarkus testing.
When it comes to testing a route in the context of Quarkus, the paved road is to write local integration tests.
This way of doing offers the advantage of running both in JVM and native mode.
The flip side is that the standard Camel testing approach with camel-test
and CamelTestSupport
is not supported.
Let’s enumerate below some points of interest.
A test running in JVM mode
The key point is to use the @QuarkusTest
annotation that will bootstrap Quarkus and start Camel routes before the @Test
logic is executed,
like in the example beneath:
import io.quarkus.test.junit.QuarkusTest;
import org.junit.jupiter.api.Test;
@QuarkusTest
class MyTest {
@Test
public void test() {
// Use any suitable code that send test data to the route and then assert outcomes
...
}
}
An example implementation can be found here.
A test running in native mode
As long as all extensions your application depends on are supported in native mode,
you should definitely test that your application really works in native mode.
The test logic defined in JVM mode can then be reused in native mode thanks to inheriting from the respective JVM mode test.
@NativeImageTest
annotation is there to instruct the Quarkus JUnit extension to compile the application under test to native image
and start it before running the tests.
import io.quarkus.test.junit.NativeImageTest;
@NativeImageTest
class MyIT extends MyTest {
...
}
An implementation of a native test may help to capture more details here.
@QuarkusTest
vs. @NativeImageTest
JVM mode tests annotated with @QuarkusTest
are executed in the same JVM as the application under test.
Thanks to that, @Inject
-ing beans from the application into the test code is possible.
You can also define new beans or even override the beans from the application using @javax.enterprise.inject.Alternative
and @javax.annotation.Priority
.
However all these tricks won’t work in native mode tests annotated with @NativeImageTest
because those are executed in a JVM hosted in a process separate from the running native application.
If you ask why, the answer is actually in the previous sentence: a native executable does not need a JVM to run; it even cannot be run by a JVM, because it is native code, not bytecode.
On the other hand, there is no point in compiling tests to native code. So they are run using a traditional JVM.
An important consequence of this setup is that all communication between tests and the application must go over network (HTTP/REST, or any other protocol your application speaks) or through watching filesystem (log files, etc.) or any kind of interprocess communication.
Testing with external services
Testcontainers
Sometimes your application needs to access some external service, such as messaging broker, database, etc. If a container image is available for the service of interest, Testcontainers come in handy for starting and configuring the services during testing.
For the application to work properly it is often essential to pass the connection configuration data
(host, port, user, password, etc. of the remote service) to the application before it starts.
In Quarkus ecosystem, QuarkusTestResourceLifecycleManager
serves this purpose.
You can start one or more Testcontainers in its start()
method
and you can return the connection configuration from the method in form of a Map
.
The entries of this map are then passed to the application either via command line (-Dkey=value
) in native mode
or through a special MicroProfile configuration provider in JVM mode.
Note that these settings have a higher precedence than the settings in application.properties
file.
import java.util.Map;
import java.util.HashMap;
import io.quarkus.test.common.QuarkusTestResourceLifecycleManager;
import org.testcontainers.containers.GenericContainer;
import org.testcontainers.containers.wait.strategy.Wait;
public class MyTestResource implements QuarkusTestResourceLifecycleManager {
private GenericContainer myContainer;
@Override
public Map<String, String> start() {
// Start the needed container(s)
myContainer = new GenericContainer(...)
.withExposedPorts(1234)
.waitingFor(Wait.forListeningPort());
myContainer.start();
// Pass the configuration to the application under test
return new HashMap<>() {{
put("my-container.host", container.getContainerIpAddress());
put("my-container.port", "" + container.getMappedPort(1234));
}};
}
@Override
public void stop() {
// Stop the needed container(s)
myContainer.stop();
...
}
The defined test resource needs then to be referenced from the test classes with @QuarkusTestResource
as shown below:
import io.quarkus.test.common.QuarkusTestResource;
import io.quarkus.test.junit.QuarkusTest;
@QuarkusTest
@QuarkusTestResource(MyTestResource.class)
class MyTest {
...
}
Please refer to Camel Quarkus source tree for a complete example.
WireMock
It is sometimes useful to stub HTTP interactions with third party services & APIs so that tests do not have to connect to live endpoints, as this can incur costs and the service may not always be 100% available or reliable.
An excellent tool for mocking & recording HTTP interactions is WireMock. It is used extensively throughout the Camel Quarkus test suite for various component extensions. Here follows a typical workflow for setting up WireMock.
First set up the WireMock server. Note that it is important to configure the Camel component under test to pass any HTTP interactions through the WireMock proxy. This is usually achieved by configuring a component property that determines the API endpoint URL. Sometimes things are less straightforward and some extra work is required to configure the API client library, as was the case for Twilio.
import static com.github.tomakehurst.wiremock.client.WireMock.aResponse;
import static com.github.tomakehurst.wiremock.client.WireMock.get;
import static com.github.tomakehurst.wiremock.client.WireMock.urlEqualTo;
import static com.github.tomakehurst.wiremock.core.WireMockConfiguration.wireMockConfig;
import java.util.HashMap;
import java.util.Map;
import com.github.tomakehurst.wiremock.WireMockServer;
import io.quarkus.test.common.QuarkusTestResourceLifecycleManager;
public class WireMockTestResource implements QuarkusTestResourceLifecycleManager {
private WireMockServer server;
@Override
public Map<String, String> start() {
// Setup & start the server
server = new WireMockServer(
wireMockConfig().dynamicPort()
);
server.start();
// Stub a HTTP endpoint. Note that WireMock also supports a record and playback mode
// http://wiremock.org/docs/record-playback/
server.stubFor(
get(urlEqualTo("/api/greeting"))
.willReturn(aResponse()
.withHeader("Content-Type", "application/json")
.withBody("{\"message\": \"Hello World\"}")));
// Ensure the camel component API client passes requests through the WireMock proxy
Map<String, String> conf = new HashMap<>();
conf.put("camel.component.foo.server-url", server.baseUrl());
return conf;
}
@Override
public void stop() {
if (server != null) {
server.stop();
}
}
}
Finally, ensure your test class has the @QuarkusTestResource
annotation with the appropriate test resource class specified as the value. The WireMock server will be started before all tests are
executed and will be shut down when all tests are finished.
import io.quarkus.test.common.QuarkusTestResource;
import io.quarkus.test.junit.QuarkusTest;
@QuarkusTest
@QuarkusTestResource(WireMockTestResource.class)
class MyTest {
...
}
More examples of WireMock usage can be found in the Camel Quarkus integration test source tree such as Geocoder.