下面是我用来在RestTemplate中记录整个http请求/响应而不丢失响应体信息的解决方案。Spring引导版本为<version>2.7.5</version>

1.创建LoggingInterceptor类

@Component
@Slf4j
public class LoggingInterceptor implements ClientHttpRequestInterceptor {

    @Override
    public ClientHttpResponse intercept(final HttpRequest request, final byte[] body, final ClientHttpRequestExecution execution) throws IOException {
        traceRequest(request, body);
        ClientHttpResponse response = execution.execute(request, body);
        response = traceResponse(response);
        return response;
    }

    private void traceRequest(HttpRequest request, byte[] body) throws IOException {
        if (!log.isDebugEnabled()) {
            return;
        }
        log.debug("=========================== Request Begin ===========================");
        log.debug("URI          : " + request.getURI());
        log.debug("Method       : " + request.getMethod());
        log.debug("Headers      : " + request.getHeaders());
        log.debug("Body : " + new String(body, "utf-8"));
        log.debug("============================ Request End ============================");

    }

    private ClientHttpResponse traceResponse(ClientHttpResponse response) throws IOException {
        if (!log.isDebugEnabled()) {
            return response;
        }
        ClientHttpResponse newCopiedResponse = new BufferingClientHttpResponseWrapper(response);
        StringBuilder inputStringBuilder = new StringBuilder();
        BufferedReader bufferedReader = new BufferedReader(new InputStreamReader(newCopiedResponse.getBody(), "UTF-8"));
        String line = bufferedReader.readLine();
        while (line != null) {
            inputStringBuilder.append(line);
            line = bufferedReader.readLine();
        }
        log.debug("=========================== Response Begin ===========================");
        log.debug("Status code   : {}", response.getStatusCode());
        log.debug("Status text   : {}", response.getStatusText());
        log.debug("Headers       : {}", response.getHeaders());
        log.debug("Response Body : {}", inputStringBuilder.toString());
        log.debug("============================ Response End ============================");
        return newCopiedResponse;

    }

    /**
     * Wrapper around ClientHttpResponse, buffers the body so it can be read repeatedly (for logging & consuming the result).
     */
    private static class BufferingClientHttpResponseWrapper implements ClientHttpResponse {

        private final ClientHttpResponse response;
        private byte[] body;

        public BufferingClientHttpResponseWrapper(ClientHttpResponse response) {
            this.response = response;
        }

        @Override
        public InputStream getBody() throws IOException {
            if (body == null) {
                body = StreamUtils.copyToByteArray(response.getBody());
            }
            return new ByteArrayInputStream(body);
        }

        @Override
        public HttpStatus getStatusCode() throws IOException {
            return this.response.getStatusCode();
        }

        @Override
        public int getRawStatusCode() throws IOException {
            return this.response.getRawStatusCode();
        }

        @Override
        public String getStatusText() throws IOException {
            return this.response.getStatusText();
        }

        @Override
        public HttpHeaders getHeaders() {
            return this.response.getHeaders();
        }

        @Override
        public void close() {
            this.response.close();
        }
    }

}

2.将其附加到RestTemplate bean

@Configuration
public class RestTemplateConfig {

    @Bean
    public RestTemplate createRestTemplate(LoggingInterceptor loggingInterceptor) {

        RestTemplate restTemplate = new RestTemplate();
        restTemplate.setInterceptors(Collections.singletonList(loggingInterceptor));

        return restTemplate;
    }
}

3.在应用程序中应用适当的级别

logging:
  level:
    com:
      test: DEBUG

这些答案都不能完全解决问题。Mjj1409实现了大部分功能，但方便地避免了记录响应的问题，因为这需要更多的工作。Paul Sabou提供了一个看似现实的解决方案，但没有提供足够的细节来实际实现(对我来说根本没用)。Sofiene得到了日志记录，但遇到了一个严重的问题:响应不再可读，因为输入流已经被消耗了!

我建议使用BufferingClientHttpResponseWrapper来包装响应对象，以允许多次读取响应体:

public class LoggingRequestInterceptor implements ClientHttpRequestInterceptor {

    private static final Logger logger = LoggerFactory.getLogger(LoggingRequestInterceptor.class);

    @Override
    public ClientHttpResponse intercept(final HttpRequest request, final byte[] body,
            final ClientHttpRequestExecution execution) throws IOException {
        ClientHttpResponse response = execution.execute(request, body);

        response = log(request, body, response);

        return response;
    }

    private ClientHttpResponse log(final HttpRequest request, final byte[] body, final ClientHttpResponse response) {
        final ClientHttpResponse responseCopy = new BufferingClientHttpResponseWrapper(response);
        logger.debug("Method: ", request.getMethod().toString());
        logger.debug("URI: ", , request.getURI().toString());
        logger.debug("Request Body: " + new String(body));
        logger.debug("Response body: " + IOUtils.toString(responseCopy.getBody()));
        return responseCopy;
    }

}

这将不会消耗InputStream，因为响应体被加载到内存中，可以多次读取。如果你的类路径上没有BufferingClientHttpResponseWrapper，你可以在这里找到简单的实现:

https://github.com/spring-projects/spring-android/blob/master/spring-android-rest-template/src/main/java/org/springframework/http/client/BufferingClientHttpResponseWrapper.java

设置RestTemplate:

LoggingRequestInterceptor loggingInterceptor = new LoggingRequestInterceptor();
restTemplate.getInterceptors().add(loggingInterceptor);

除了上面的讨论，这只代表了快乐的场景。如果出现错误，您可能无法记录响应。

在这种情况下加上上面所有的情况，您必须重写DefaultResponseErrorHandler并像下面那样设置它

restTemplate.setErrorHandler(new DefaultResponseErrorHandlerImpl());

So many responses here require coding changes and customized classes and it really is not necessary. Gte a debugging proxy such as fiddler and set your java environment to use the proxy on the command line (-Dhttp.proxyHost and -Dhttp.proxyPort) then run fiddler and you can see the requests and responses in their entirety. Also comes with many ancillary advantages such as the ability to tinker with the results and responses before and after they are sent to run experiments before committing to modification of the server.

可能出现的最后一个问题是，如果必须使用HTTPS，则需要从fiddler导出SSL证书并将其导入java密钥存储库(cacerts)提示:默认的java密钥存储库密码通常是“changeit”。

正如@MilacH指出的，在实现中有一个错误。如果返回statusCode > 400，则从拦截器抛出IOException，因为没有调用errorHandler。可以忽略异常，然后在处理程序方法中再次捕获异常。

package net.sprd.fulfillment.common;

import org.slf4j.Logger;
import org.slf4j.LoggerFactory;
import org.springframework.http.HttpRequest;
import org.springframework.http.client.ClientHttpRequestExecution;
import org.springframework.http.client.ClientHttpRequestInterceptor;
import org.springframework.http.client.ClientHttpResponse;

import java.io.BufferedReader;
import java.io.IOException;
import java.io.InputStreamReader;

import static java.nio.charset.StandardCharsets.UTF_8;

public class LoggingRequestInterceptor implements ClientHttpRequestInterceptor {

    final static Logger log = LoggerFactory.getLogger(LoggingRequestInterceptor.class);

    @SuppressWarnings("HardcodedLineSeparator")
    public static final char LINE_BREAK = '\n';

    @Override
    public ClientHttpResponse intercept(HttpRequest request, byte[] body, ClientHttpRequestExecution execution) throws IOException {
        try {
            traceRequest(request, body);
        } catch (Exception e) {
            log.warn("Exception in LoggingRequestInterceptor while tracing request", e);
        }

        ClientHttpResponse response = execution.execute(request, body);

        try {
            traceResponse(response);
        } catch (IOException e) {
            // ignore the exception here, as it will be handled by the error handler of the restTemplate
            log.warn("Exception in LoggingRequestInterceptor", e);
        }
        return response;
    }

    private void traceRequest(HttpRequest request, byte[] body) {
        log.info("===========================request begin================================================");
        log.info("URI         : {}", request.getURI());
        log.info("Method      : {}", request.getMethod());
        log.info("Headers     : {}", request.getHeaders());
        log.info("Request body: {}", new String(body, UTF_8));
        log.info("==========================request end================================================");
    }

    private void traceResponse(ClientHttpResponse response) throws IOException {
        StringBuilder inputStringBuilder = new StringBuilder();
        try (BufferedReader bufferedReader = new BufferedReader(new InputStreamReader(response.getBody(), UTF_8))) {
            String line = bufferedReader.readLine();
            while (line != null) {
                inputStringBuilder.append(line);
                inputStringBuilder.append(LINE_BREAK);
                line = bufferedReader.readLine();
            }
        }

        log.info("============================response begin==========================================");
        log.info("Status code  : {}", response.getStatusCode());
        log.info("Status text  : {}", response.getStatusText());
        log.info("Headers      : {}", response.getHeaders());
        log.info("Response body: {}", inputStringBuilder);
        log.info("=======================response end=================================================");
    }

}

Related to the response using ClientHttpInterceptor, I found a way of keeping the whole response without Buffering factories. Just store the response body input stream inside byte array using some utils method that will copy that array from body, but important, surround this method with try catch because it will break if response is empty (that is the cause of Resource Access Exception) and in catch just create empty byte array, and than just create anonymous inner class of ClientHttpResponse using that array and other parameters from the original response. Than you can return that new ClientHttpResponse object to the rest template execution chain and you can log response using body byte array that is previously stored. That way you will avoid consuming InputStream in the actual response and you can use Rest Template response as it is. Note, this may be dangerous if your's response is too big