Access the UrlConnectionHttpClient Use and configure the UrlConnectionHttpClient Proxy configuration example

Configure the URLConnection-based HTTP client

The AWS SDK for Java 2.x offers a lighter-weight UrlConnectionHttpClient HTTP client in comparison to the default Apache5HttpClient. The UrlConnectionHttpClient is based on Java's URLConnection.

The UrlConnectionHttpClient loads more quickly than the Apache-based HTTP client, but has fewer features. Because it loads more quickly, it is a good solution for Java AWS Lambda functions.

The UrlConnectionHttpClient has several configurable options that you can access.

Note

The UrlConnectionHttpClient does not support the HTTP PATCH method.

A handful of AWS API operations require PATCH requests. Those operation names usually start with Update*. The following are several examples.

Several Update* operations in the AWS Security Hub CSPM API and also the BatchUpdateFindings operation
All Amazon API Gateway API Update* operations

If you might use the UrlConnectionHttpClient, first refer to the API Reference for the AWS service that you're using. Check to see if the operations you need use the PATCH operation.

Note

The UrlConnectionHttpClient does not support a configurable idle connection timeout.

Connection pooling is managed internally by the JVM. To configure the idle timeout, set the JVM system property http.keepAlive.time.server to the desired value in seconds:


$ java -Dhttp.keepAlive.time.server=60 -cp ... App

This is a JVM-global setting that affects all HttpURLConnection usage in the process, not just SDK clients. If you need per-client control over idle connection timeout, use the Apache5HttpClient with the connectionMaxIdleTime setting.

Access the `UrlConnectionHttpClient`

To configure and use the UrlConnectionHttpClient, you declare a dependency on the url-connection-client Maven artifact in your pom.xml file.

Unlike the Apache5HttpClient, the UrlConnectionHttpClient is not automatically added to your project, so declare it explicitly.

The following example of a pom.xml file shows the dependencies required to use and configure the HTTP client.


<dependencyManagement>
   <dependencies>
        <dependency>
            <groupId>software.amazon.awssdk</groupId>
            <artifactId>bom</artifactId>
            <version>2.57.3</version>
            <type>pom</type>
            <scope>import</scope>
        </dependency>
    </dependencies>
</dependencyManagement>

<!-- other dependencies such as s3 or dynamodb -->

<dependencies>
    <dependency>
        <groupId>software.amazon.awssdk</groupId>
        <artifactId>url-connection-client</artifactId>
    </dependency>
</dependencies>

Replace 2.57.3 with the version of the Java SDK that you want to use. For information about the latest version, see the Maven Central website.

Each service artifact, such as s3, also adds the apache5-client and netty-nio-client artifacts to your project by default. If you use only the UrlConnectionHttpClient, you can exclude those artifacts from each service dependency to reduce the number of jars in your application.

The following Maven and Gradle examples exclude both artifacts from the s3 dependency and add the url-connection-client dependency:

Note

Before version 2.46.0, the default synchronous HTTP client artifact was apache-client. If you use an earlier version, exclude apache-client instead of apache5-client.

The UrlConnectionHttpClient supports only synchronous service clients. If you exclude netty-nio-client and also use asynchronous service clients, such as S3AsyncClient, add another asynchronous HTTP client.

Use and configure the `UrlConnectionHttpClient`

You can configure an instance of UrlConnectionHttpClient along with building a service client, or you can configure a single instance to share across multiple service clients.

With either approach, you use the UrlConnectionHttpClient.Builder to configure the properties for the URLConnection-based HTTP client.

Best practice: dedicate an `UrlConnectionHttpClient` instance to a service client

If you need to configure an instance of the UrlConnectionHttpClient, we recommend that you build the dedicated UrlConnectionHttpClient instance. You can do so by using the httpClientBuilder method of the service client's builder. This way, the lifecycle of the HTTP client is managed by the SDK, which helps avoid potential memory leaks if the UrlConnectionHttpClient instance is not closed down when it's no longer needed.

The following example creates an S3Client and configures the embedded instance of UrlConnectionHttpClient with socketTimeout and proxyConfiguration values. The proxyConfiguration method takes a Java lambda expression of type Consumer<ProxyConfiguration.Builder>.

Imports


import software.amazon.awssdk.http.SdkHttpClient;
import software.amazon.awssdk.http.urlconnection.UrlConnectionHttpClient;
import java.net.URI;
import java.time.Duration;

Code


// Singleton: Use the s3Client for all requests.
S3Client s3Client = 
    S3Client.builder()
            .httpClientBuilder(UrlConnectionHttpClient.builder()
                    .socketTimeout(Duration.ofMinutes(5))
                    .proxyConfiguration(proxy -> proxy.endpoint(URI.create("http://proxy.mydomain.net:8888"))))
            .credentialsProvider(EnvironmentVariableCredentialsProvider.create())
            .build();

// Perform work with the s3Client.

s3Client.close();   // Requests completed: Close the s3client.

Alternative approach: share an `UrlConnectionHttpClient` instance

To help keep resource and memory usage lower for your application, you can configure an UrlConnectionHttpClient and share it across multiple service clients. The HTTP connection pool will be shared, which lowers resource usage.

Note

When an UrlConnectionHttpClient instance is shared, you must close it when it is ready to be disposed. The SDK will not close the instance when the service client is closed.

The following example configures an URLConnection-based HTTP client that is used by two service clients. The configured UrlConnectionHttpClient instance is passed to the httpClient method of each builder. When the service clients and the HTTP client are no longer needed, the code explicitly closes them. The code closes the HTTP client last.

Imports


import software.amazon.awssdk.auth.credentials.EnvironmentVariableCredentialsProvider;
import software.amazon.awssdk.awscore.defaultsmode.DefaultsMode;
import software.amazon.awssdk.http.SdkHttpClient;
import software.amazon.awssdk.http.urlconnection.ProxyConfiguration;
import software.amazon.awssdk.http.urlconnection.UrlConnectionHttpClient;
import software.amazon.awssdk.services.dynamodb.DynamoDbClient;
import software.amazon.awssdk.services.s3.S3Client;
import java.net.URI;
import java.time.Duration;

Code


SdkHttpClient urlHttpClient = UrlConnectionHttpClient.create();

// Singletons: Use the s3Client and dynamoDbClient for all requests.
S3Client s3Client = 
    S3Client.builder()
            .httpClient(urlHttpClient)
            .defaultsMode(DefaultsMode.IN_REGION)
            .credentialsProvider(EnvironmentVariableCredentialsProvider.create())
            .build();

DynamoDbClient dynamoDbClient = 
    DynamoDbClient.builder()
                  .httpClient(urlHttpClient)
                  .defaultsMode(DefaultsMode.IN_REGION)
                  .credentialsProvider(EnvironmentVariableCredentialsProvider.create())
                  .build();

// Perform work with the s3Client and dynamoDbClient.

// Requests completed: Close all service clients.
s3Client.close();
dynamoDbClient.close();
urlHttpClient.close();

If you don't specify an HTTP client for a service client, the SDK chooses one for you from the HTTP clients on your classpath. When there's more than one, the SDK uses the one with the highest priority. Apache5HttpClient has the highest priority, so the SDK uses it for any service client that doesn't specify an HTTP client, even if another service client uses the UrlConnectionHttpClient.

In the following example, the dynamoDbClient uses the UrlConnectionHttpClient. The s3Client doesn't specify an HTTP client, so the SDK uses Apache5HttpClient:


// The dynamoDbClient uses the UrlConnectionHttpClient.
DynamoDbClient dynamoDbClient = DynamoDbClient.builder()
        .httpClientBuilder(UrlConnectionHttpClient.builder())
        .build();

// The s3Client doesn't specify an HTTP client, so the SDK uses the Apache5HttpClient.
S3Client s3Client = S3Client.create();

// Perform work with the s3Client and dynamoDbClient.

dynamoDbClient.close();
s3Client.close();

To choose the HTTP client for a service client, specify it explicitly:


DynamoDbClient dynamoDbClient = DynamoDbClient.builder()
        .httpClientBuilder(UrlConnectionHttpClient.builder())
        .build();

S3Client s3Client = S3Client.builder()
        .httpClientBuilder(Apache5HttpClient.builder())    // Explicitly use the Apache5HttpClient.
        .build();

// Perform work with the s3Client and dynamoDbClient.

dynamoDbClient.close();
s3Client.close();

Note

To explicitly use the Apache5HttpClient, add a dependency on the apache5-client artifact. For more information about adding the dependency, see Configuration needed.

Before version 2.22.0, the SDK threw an SdkClientException with the message that multiple HTTP implementations were found on the classpath. To avoid the exception, specify the HTTP client for each service client.

Proxy configuration example

The following code snippet uses the proxy configuration builder for the URL connection HTTP client.


SdkHttpClient urlHttpClient = UrlConnectionHttpClient.builder()
                .proxyConfiguration(ProxyConfiguration.builder()
                        .endpoint(URI.create("http://example.com:1234"))
                        .username("username")
                        .password("password")
                        .addNonProxyHost("localhost")
                        .addNonProxyHost("host.example.com")
                        .build())
                .build();

The equivalent Java system properties for the proxy configuration are shown in the following command line snippet.


$ java -Dhttp.proxyHost=example.com -Dhttp.proxyPort=1234 -Dhttp.proxyUser=username \
-Dhttp.proxyPassword=password -Dhttp.nonProxyHosts=localhost|host.example.com -cp ... App

The equivalent setup that uses environment variables is:


// Set the following environment variables.
// $ export HTTP_PROXY="http://username:password@example.com:1234"
// $ export NO_PROXY="localhost|host.example.com"

// Set the 'useSystemPropertyValues' to false on the proxy configuration.
SdkHttpClient apacheHttpClient = UrlConnectionHttpClient.builder()
                .proxyConfiguration(ProxyConfiguration.builder()
                        .useSystemPropertyValues(Boolean.FALSE)
                        .build())
                .build();

// Run the application.
// $ java -cp ... App

Note

The URLConnection-based HTTP client does not currently support HTTPS proxy system properties or the HTTPS_PROXY environment variable.

Warning Javascript is disabled or is unavailable in your browser.

To use the Amazon Web Services Documentation, Javascript must be enabled. Please refer to your browser's Help pages for instructions.

Document Conventions

Configure the Apache-based HTTP client

Configure the Netty-based HTTP client

Configure the URLConnection-based HTTP client

Note

Note

Access the UrlConnectionHttpClient

Note

Use and configure the UrlConnectionHttpClient

Best practice: dedicate an UrlConnectionHttpClient instance to a service client

Alternative approach: share an UrlConnectionHttpClient instance

Note

Note

Proxy configuration example

Note

Access the `UrlConnectionHttpClient`

Use and configure the `UrlConnectionHttpClient`

Best practice: dedicate an `UrlConnectionHttpClient` instance to a service client

Alternative approach: share an `UrlConnectionHttpClient` instance