如何在 Zuul post 过滤器中获取响应主体？

Question

在 post 过滤器中使用 Zuul 作为代理时如何读取响应主体？

我正在尝试这样调用代码：

@Component
public class PostFilter extends ZuulFilter {

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

    @Override
    public String filterType() {
        return "post";
    }

    @Override
    public int filterOrder() {
        return 2000;
    }

    @Override
    public boolean shouldFilter() {
        return true;
    }

    @Override
    public Object run() {
        RequestContext ctx = RequestContext.getCurrentContext();
        ctx.getResponseBody(); // null

        // cant't do this, cause input stream is used later in other filters and I got InputStream Closed exception
        // GZIPInputStream gzipInputStream = new GZIPInputStream(stream);
        return null;
    }

}

Answer 1

如您在此 example 中所见，您有两种方法可用于提取响应正文：

1- ctx.getResponseBody();

2- ctx.getResponseDataStream();

你必须检查哪个不为空并使用那个。

Answer 2

我已经设法克服了这个问题。该解决方案包括 4 个步骤：

将 ctx.getResponseDataStream() 读入 ByteArrayOutputStream
将 OutputStream 复制到 2 个 InputStreams。
将其中之一用于您的自定义目的。
使用第二个重新分配给上下文：context.setResponseBody(inputStream)
- 从点 1 读取流会导致无法再次读取该流，因此您将通过这种方式传递尚未读取的新流

Answer 3

感谢您的建议，这是我使用的有效代码。

try (final InputStream responseDataStream = ctx.getResponseDataStream()) {
   final String responseData = CharStreams.toString(new InputStreamReader(responseDataStream, "UTF-8"));
   ctx.setResponseBody(responseData);
} catch (IOException e) {
   logger.warn("Error reading body",e);
}

Answer 4

小心 filterNumber

Using anything greater than 1000 leads to an "InputStream already closed" error because the response body has already been read and

我用的是数字 10，效果很好

Answer 5

如果有人在压缩答案上苦苦挣扎，这是我使用的解决方案：

// Read the compressed response
RequestContext ctx = RequestContext.getCurrentContext();
InputStream compressedResponseDataStream = ctx.getResponseDataStream();
try {
    // Uncompress and transform the response
    InputStream responseDataStream = new GZIPInputStream(compressedResponseDataStream);
    String responseAsString = StreamUtils.copyToString(responseDataStream, Charset.forName("UTF-8"));
    // Do want you want with your String response
    ...
    // Replace the response with the modified object
    ctx.setResponseBody(responseAsString);
} catch (IOException e) {
    logger.warn("Error reading body", e);
}

Answer 6

None 个答案对我有用。 1) 过滤器的阶数需要低于 1000（发送响应过滤器）

2）代码：

 private String getResponseData(RequestContext ctx) throws IOException {
    String responseData = null;

    final InputStream responseDataStream = ctx.getResponseDataStream();
    ByteArrayOutputStream bos = new ByteArrayOutputStream();
    ByteArrayOutputStream copy = new ByteArrayOutputStream();
    int read = 0;
    byte[] buff = new byte[1024];
    while ((read = responseDataStream.read(buff)) != -1) {
        bos.write(buff, 0, read);
        copy.write(buff, 0, read);
    }
    InputStream isFromFirstData = new ByteArrayInputStream(bos.toByteArray());

    boolean responseGZipped = ctx.getResponseGZipped();
    try {
        InputStream zin = null;
        if (responseGZipped) {
            zin = new GZIPInputStream(isFromFirstData);
        } else {
            zin = responseDataStream;
        }
        responseData = CharStreams.toString(new InputStreamReader(zin, "UTF-8"));
        ctx.setResponseDataStream(new ByteArrayInputStream(copy.toByteArray()));

    } catch (IOException e) {
        logger.warn("Error reading body {}", e.getMessage());
    }

    return responseData;
}

Answer 7

最简单的解决方案是获取一个输入流，用InputStreamReader读取它，最后根据读取的字符串创建一个新的流并将其设置为上下文的响应数据流。另外，不要忘记将过滤器的顺序设置为小于 1000 的数字。

final InputStream responseDataStream = ctx.getResponseDataStream();
String response = CharStreams.toString(new InputStreamReader(responseDataStream, StandardCharsets.UTF_8));
InputStream stream = new ByteArrayInputStream(response.getBytes(StandardCharsets.UTF_8));
ctx.setResponseDataStream(stream);

如何在 Zuul post 过滤器中获取响应主体？

How to get response body in Zuul post filter?

java

proxy

netflix-zuul