Java Microservices Testing — A Minimalist Approach
In this tutorial, we'll learn about using HttpClient for REST API microservices testing with Java.
Join the DZone community and get the full member experience.
Join For FreeTo prepare integration tests of microservice's REST API you really don't need any additional tools other than the microservice itself and HttpUrlConnection class included in JDK.
In the matter of fact, you can even test such API without writing any code, using cURL. And this is the big advantage of the REST API, despite frequent criticism.
The article shows how to design test cases for microservice REST API using the minimum external dependencies. The REST API of the Cricket Microservices Framework (CMSF) will serve as an example.
Although it is possible and in the case of a quick functionality check - invaluable, using the simplest methods is not convenient when you need automation. For this reason, to create our tests we will use JUnit framework and the wrapper class for HttpUrlConnection.
Another useful trick to facilitate running a test session is to use an embedded microservice. It is worth considering whether this is possible because, thanks to this, we do not have to run a separate test environment.
HttpClient
HttpClient along with Request and StandardResult classes which are part of Cricket Microservices Framework provides several convenient methods to handle the HTTP request with Java. They were included in the CMSF for two main reasons: to avoid external dependencies and to ensure Android compatibility of the client classes.
The code below is the streamlined version of the original HttpClient (without SSL connection and self certificate handling).
package org.cricketmsf.out.http;
import java.io.BufferedReader;
import java.io.IOException;
import java.io.InputStreamReader;
import java.io.OutputStream;
import java.io.OutputStreamWriter;
import java.net.HttpURLConnection;
import java.net.URL;
import org.cricketmsf.in.http.Result;
import org.cricketmsf.in.http.StandardResult;
public class HttpClient {
private int timeout = 0;
public Result send(Request request) {
StandardResult result = new StandardResult();
if (request == null) {
result.setCode(HttpURLConnection.HTTP_INTERNAL_ERROR);
result.setMessage("Request is null");
}
String requestData = "";
if (request.data != null) {
if (requestData instanceof String) {
requestData = (String) request.data;
} else {
requestData = request.data.toString();
}
}
result.setCode(HttpURLConnection.HTTP_INTERNAL_ERROR);
try {
URL urlObj = new URL(request.getUrl());
HttpURLConnection con;
con = (HttpURLConnection) urlObj.openConnection();
con.setReadTimeout(timeout);
con.setConnectTimeout(timeout);
con.setRequestMethod(request.method);
request.properties.keySet().forEach((key) -> {
con.setRequestProperty(key, request.properties.get(key));
});
if (requestData.length() > 0 || "POST".equals(request.method) || "PUT".equals(request.method) || "DELETE".equals(request.method)) {
con.setDoOutput(true);
OutputStream os = con.getOutputStream();
OutputStreamWriter out = new OutputStreamWriter(os);
try {
out.write(requestData);
out.flush();
out.close();
os.close();
} catch (IOException e) {
result.setCode(HttpURLConnection.HTTP_INTERNAL_ERROR);
result.setMessage(e.getMessage());
}
}
con.connect();
result.setCode(con.getResponseCode());
StringBuilder response;
String inputLine;
try {
try (BufferedReader in = new BufferedReader(new InputStreamReader(con.getInputStream()))) {
response = new StringBuilder();
while ((inputLine = in.readLine()) != null) {
response.append(inputLine);
}
}
result.setPayload(response.toString().getBytes());
} catch (IOException e) {
try (BufferedReader in = new BufferedReader(new InputStreamReader(con.getErrorStream()))) {
response = new StringBuilder();
while ((inputLine = in.readLine()) != null) {
response.append(inputLine);
}
}
result.setMessage(response.toString());
}
} catch (IOException e) {
result.setCode(HttpURLConnection.HTTP_INTERNAL_ERROR);
result.setMessage(e.getMessage());
}
return result;
}
}
Test Cases
With the HTTP client in place, we can design required test cases for our REST API. As an example, let's test Cricket's user API.
It would be very convenient if we did not have to run the tested page in a separate thread and start the tests with one click. For services built using CMSF, this is possible thanks to the implemented hexagonal architecture. We can prepare the appropriate configuration of adapters and get a self-contained service. We can treat this configuration as a reference for further development of the website.
Let's create methods to run before and after calling tests, to start the service and generate a session token, and after performing all tests, call the proper shutdown.
/**
* Run the microservice on localhost
*/
@BeforeClass
public static void setup() {
System.out.println("@setup");
String[] args = {"-r", "-s", "Microsite"};
service = Runner.getService(args);
while (!service.isInitialized()) {
try {
Thread.sleep(100);
} catch (InterruptedException ex) {
Thread.currentThread().interrupt();
}
}
System.out.println("service is running");
sessionToken = getSessionToken();
}
@AfterClass
public static void shutdown() {
System.out.println("@shutdown");
service.shutdown();
}
/**
* Log in to the service as user "demo" and get session token
*/
private static String getSessionToken() {
// Given
String login = "demo";
String password = "cricket";
String credentials = Base64.getEncoder()
.encodeToString((login + ":" + password).getBytes());
HttpClient client = new HttpClient();
Request req = new Request()
.setMethod("POST")
.setProperty("Accept", "text/plain")
.setProperty("Authentication", "Basic " + credentials)
.setData("p=ignotethis") /*data must be added to POST or PUT requests */
.setUrl("http://localhost:8080/api/auth");
// When
StandardResult response = (StandardResult) client.send(req);
// Then
Assert.assertEquals(200, response.getCode());
String token = "";
try {
token = new String(response.getPayload(), "UTF-8");
} catch (UnsupportedEncodingException ex) {
Assert.fail(ex.getMessage());
}
Assert.assertFalse(token == null || token.isEmpty());
return token;
}
Then, test methods can use a running service to send requests and check responses.
@Test
public void readingPersonalDataOK() {
// Given
HttpClient client = new HttpClient();
Request req = new Request()
.setMethod("GET")
.setProperty("Accept", "apNplication/json")
.setProperty("Authentication", sessionToken)
.setUrl("http://localhost:8080/api/user/")
.setQuery("demo");
// When
StandardResult response = (StandardResult) client.send(req);
// Then
Assert.assertEquals(200, response.getCode());
String data = null;
try {
data = new String(response.getPayload(), "UTF-8");
} catch (UnsupportedEncodingException ex) {
Assert.fail(ex.getMessage());
}
Assert.assertFalse(data == null || data.isEmpty());
}
@Test
public void readingOtherUserPersonalDataNOK() {
//Given
HttpClient client = new HttpClient();
Request req = new Request()
.setMethod("GET")
.setProperty("Accept", "application/json")
.setProperty("Authentication", sessionToken)
.setUrl("http://localhost:8080/api/user/")
.setQuery("admin");
// When
StandardResult response = (StandardResult) client.send(req);
// Then
// The session token has been generated for user "demo", so accessing
// "admin" account should not be possible.
Assert.assertEquals(403, response.getCode());
}
And that's all. In a few simple steps, we have built a set of tests, working based on simple tools and at the same time, friendly for testing automation.
You can find the complete code in the Cricket Microservices Framework repository.
Conclusion
Testing REST API with Java is not as complicated as it is said. There are many tools that facilitate this task (several referenced above), so you can choose the one that suits you. However, it is always worth knowing what is going on underneath and not being completely dependent on the tool.
Published at DZone with permission of Grzegorz Skorupa. See the original article here.
Opinions expressed by DZone contributors are their own.
Trending
-
Replacing Apache Hive, Elasticsearch, and PostgreSQL With Apache Doris
-
Never Use Credentials in a CI/CD Pipeline Again
-
Is Podman a Drop-in Replacement for Docker?
-
4 Expert Tips for High Availability and Disaster Recovery of Your Cloud Deployment
Comments