feign的bean创建过程-底层请求过程-源码走读

feign的bean创建过程，请求过程-源码走读

目标

1. 了解feign的bean的创建过程
2. 了解feign发送请求时的过程

前置知识：了解spring的bean创建过程

feign的版本是spring-cloud-openfeign-core:4.1.2

采用的nacos做注册中心

项目需要配置 spring.cloud.loadbalancer.nacos.enabled=true

bean的创建过程

一般我们会通过注解 @EnableFeignClients(basePackages = {"org.example.**.feign"}) 指定扫描的bean 启动项目之后，就会在对应目录下面扫描出来，feign是采用factoryBean的形式创建bean的，所以断点直接打到FeignClientFactoryBean.getObject

    @Override
	public Object getObject() {
		return getTarget();
	}

	/**
	 * @param <T> the target type of the Feign client
	 * @return a {@link Feign} client created with the specified data and the context
	 * information
	 */
	@SuppressWarnings("unchecked")
	<T> T getTarget() {
        FeignClientFactory feignClientFactory = beanFactory != null ? beanFactory.getBean(FeignClientFactory.class)
                : applicationContext.getBean(FeignClientFactory.class);
        Feign.Builder builder = feign(feignClientFactory);
        if (!StringUtils.hasText(url) && !isUrlAvailableInConfig(contextId)) { // 没有指定url的情况，feign要自己拼

            if (LOG.isInfoEnabled()) {
                LOG.info("For '" + name + "' URL not provided. Will try picking an instance via load-balancing.");
            }
            if (!name.startsWith("http://") && !name.startsWith("https://")) {
                url = "http://" + name;
            }
            else {
                url = name;
            }
            url += cleanPath();
            return (T) loadBalance(builder, feignClientFactory, new HardCodedTarget<>(type, name, url)); // 没有写死url，选择负载均衡，进入这里看一下怎么创建的
        }
        if (StringUtils.hasText(url) && !url.startsWith("http://") && !url.startsWith("https://")) {
            url = "http://" + url;
        }
        Client client = getOptional(feignClientFactory, Client.class);
        if (client != null) {
            if (client instanceof FeignBlockingLoadBalancerClient) {
                // not load balancing because we have a url,
                // but Spring Cloud LoadBalancer is on the classpath, so unwrap
                client = ((FeignBlockingLoadBalancerClient) client).getDelegate();
            }
            if (client instanceof RetryableFeignBlockingLoadBalancerClient) {
                // not load balancing because we have a url,
                // but Spring Cloud LoadBalancer is on the classpath, so unwrap
                client = ((RetryableFeignBlockingLoadBalancerClient) client).getDelegate();
            }
            builder.client(client);
        }

        applyBuildCustomizers(feignClientFactory, builder);

        Targeter targeter = get(feignClientFactory, Targeter.class);
        return targeter.target(this, builder, feignClientFactory, resolveTarget(feignClientFactory, contextId, url));
    }

loadBalance 里面实际创建是最后一步调用target函数，一步一步点进去之后，会发现是采用jdk代理创建的

protected <T> T loadBalance(Feign.Builder builder, FeignClientFactory context, HardCodedTarget<T> target) {
    Client client = getOptional(context, Client.class);
    if (client != null) {
        builder.client(client);
        applyBuildCustomizers(context, builder);
        Targeter targeter = get(context, Targeter.class);
        return targeter.target(this, builder, context, target);
    }
}
public <T> T target(FeignClientFactoryBean factory, Feign.Builder feign, FeignClientFactory context,
        Target.HardCodedTarget<T> target) {
    return feign.target(target);
}

public <T> T target(Target<T> target) {
    return build().newInstance(target); // build 函数内部会注册 ResponseHandler，里面有编码器，解码器，日志，interceptor等东西
}

// biuld->internalBuild
public Feign internalBuild() {
    final ResponseHandler responseHandler =
        new ResponseHandler(logLevel, logger, decoder, errorDecoder,
            dismiss404, closeAfterDecode, decodeVoid, responseInterceptorChain());
    MethodHandler.Factory<Object> methodHandlerFactory =
        new SynchronousMethodHandler.Factory(client, retryer, requestInterceptors,
            responseHandler, logger, logLevel, propagationPolicy,
            new RequestTemplateFactoryResolver(encoder, queryMapEncoder),
            options);
    return new ReflectiveFeign<>(contract, methodHandlerFactory, invocationHandlerFactory,
        () -> null);
}
// 回到instance

public <T> T newInstance(Target<T> target) {
return newInstance(target, defaultContextSupplier.newContext());
}

public <T> T newInstance(Target<T> target, C requestContext) {
    TargetSpecificationVerifier.verify(target);

    Map<Method, MethodHandler> methodToHandler =
        targetToHandlersByName.apply(target, requestContext);
    InvocationHandler handler = factory.create(target, methodToHandler); // 熟悉的InvocationHandler，jdk动态代理的写法
    T proxy = (T) Proxy.newProxyInstance(target.type().getClassLoader(),
        new Class<?>[] {target.type()}, handler);

    for (MethodHandler methodHandler : methodToHandler.values()) {
        if (methodHandler instanceof DefaultMethodHandler) {
        ((DefaultMethodHandler) methodHandler).bindTo(proxy);
        }
    }

    return proxy;
}

InvocationHandler 是一个内部类实现的 FeignInvocationHandler

static class FeignInvocationHandler implements InvocationHandler {

    private final Target target;
    private final Map<Method, MethodHandler> dispatch;

    FeignInvocationHandler(Target target, Map<Method, MethodHandler> dispatch) {
      this.target = checkNotNull(target, "target");
      this.dispatch = checkNotNull(dispatch, "dispatch for %s", target);
    }

    @Override
    public Object invoke(Object proxy, Method method, Object[] args) throws Throwable {
      if ("equals".equals(method.getName())) {
        try {
          Object otherHandler =
              args.length > 0 && args[0] != null ? Proxy.getInvocationHandler(args[0]) : null;
          return equals(otherHandler);
        } catch (IllegalArgumentException e) {
          return false;
        }
      } else if ("hashCode".equals(method.getName())) {
        return hashCode();
      } else if ("toString".equals(method.getName())) {
        return toString();
      } else if (!dispatch.containsKey(method)) {
        throw new UnsupportedOperationException(
            String.format("Method \"%s\" should not be called", method.getName()));
      }

      return dispatch.get(method).invoke(args); // dispatch是一个map，key是我们的远程调用函数，value是SynchronousMethodHandler，也就是实际执行时候要调用的函数，在第二节执行的时候会展开
    }

    @Override
    public boolean equals(Object obj) {
      if (obj instanceof FeignInvocationHandler) {
        FeignInvocationHandler other = (FeignInvocationHandler) obj;
        return target.equals(other.target);
      }
      return false;
    }

    @Override
    public int hashCode() {
      return target.hashCode();
    }

    @Override
    public String toString() {
      return target.toString();
    }
  }

feign的调用过程

直接从方法入口进入

@FeignClient(value = "database", contextId = "db")
public interface DatabaseFeign {

    @GetMapping(value = Constant.FEIGN_URL_PREFIX + "/dbInfo/{id}")
    Response<DbInfo> getDbInfo(@PathVariable("id") Integer id);
}

@GetMapping("/v1")
public Response<DbInfo> v1() {
    return databaseFeign.getDbInfo(1);
}

第一节提到，代理是jdk动态代理，所以预期应该是invocationHandler里面执行，直接把断点打到FeignInvocationHandler的invoke上面

static class FeignInvocationHandler implements InvocationHandler {

    private final Target target;
    private final Map<Method, MethodHandler> dispatch;

    FeignInvocationHandler(Target target, Map<Method, MethodHandler> dispatch) {
      this.target = checkNotNull(target, "target");
      this.dispatch = checkNotNull(dispatch, "dispatch for %s", target);
    }

    @Override
    public Object invoke(Object proxy, Method method, Object[] args) throws Throwable {
      if ("equals".equals(method.getName())) {
        try {
          Object otherHandler =
              args.length > 0 && args[0] != null ? Proxy.getInvocationHandler(args[0]) : null;
          return equals(otherHandler);
        } catch (IllegalArgumentException e) {
          return false;
        }
      } else if ("hashCode".equals(method.getName())) {
        return hashCode();
      } else if ("toString".equals(method.getName())) {
        return toString();
      } else if (!dispatch.containsKey(method)) {
        throw new UnsupportedOperationException(
            String.format("Method \"%s\" should not be called", method.getName()));
      }

      return dispatch.get(method).invoke(args); // 正常肯定走到这里
    }

    @Override
    public boolean equals(Object obj) {
      if (obj instanceof FeignInvocationHandler) {
        FeignInvocationHandler other = (FeignInvocationHandler) obj;
        return target.equals(other.target);
      }
      return false;
    }

    @Override
    public int hashCode() {
      return target.hashCode();
    }

    @Override
    public String toString() {
      return target.toString();
    }
  }

dispatch 是一个map，里面存储的是方法对应的SynchronousMethodHandler，都是和http请求相关的属性

final class SynchronousMethodHandler implements MethodHandler {

  private final MethodMetadata metadata;
  private final Target<?> target;
  private final Client client;
  private final Retryer retryer;
  private final List<RequestInterceptor> requestInterceptors;
  private final Logger logger;
  private final Logger.Level logLevel;
  private final RequestTemplate.Factory buildTemplateFromArgs;
  private final Options options;
  private final ExceptionPropagationPolicy propagationPolicy;
  private final ResponseHandler responseHandler;
 // ...
  @Override
  public Object invoke(Object[] argv) throws Throwable { // 上面的代码，接下来进入invoke
    RequestTemplate template = buildTemplateFromArgs.create(argv); // 把参数填到url路径里面
    Options options = findOptions(argv); // 1. 参数，包含超时时间等信息，可以写在函数参数里面
    Retryer retryer = this.retryer.clone();
    while (true) {
      try {
        return executeAndDecode(template, options); // 3.
      } catch (RetryableException e) {
        try { // 2. 失败自动重试，重试取决于retryer，如果超过重试次数就直接抛异常了，不会重新循环，如果不配置是不重试的，会抛出异常
          retryer.continueOrPropagate(e);
        } catch (RetryableException th) {
          Throwable cause = th.getCause();
          if (propagationPolicy == UNWRAP && cause != null) {
            throw cause;
          } else {
            throw th;
          }
        }
        if (logLevel != Logger.Level.NONE) {
          logger.logRetry(metadata.configKey(), logLevel);
        }
        continue;
      }
    }
  }
}

1中 options可以这么实现

@GetMapping(value = Constant.FEIGN_URL_PREFIX + "/dbInfo/{id}")
Response<DbInfo> getDbInfo(@PathVariable("id") Integer id, Request.Options options);

2 中的重试规则，默认不重试

Retryer NEVER_RETRY = new Retryer() {
    public void continueOrPropagate(RetryableException e) {
        throw e;
    }

    public Retryer clone() {
        return this;
    }
};

抛出 RetryableException 有两个位置 ioExecption 和 返回的resp header是否有 Retry-After 字段

private CompletableFuture<Object> executeAndDecode(RequestTemplate template, Options options) {
    Request request = targetRequest(template);

    if (logLevel != Logger.Level.NONE) {
      logger.logRequest(metadata.configKey(), logLevel, request);
    }

    long start = System.nanoTime();
    return client.execute(request, options, Optional.ofNullable(requestContext))
        .thenApply(response ->
        // TODO: remove in Feign 12
        ensureRequestIsSet(response, template, request))
        .exceptionally(throwable -> {
          CompletionException completionException = throwable instanceof CompletionException
              ? (CompletionException) throwable
              : new CompletionException(throwable);
          if (completionException.getCause() instanceof IOException) {
            IOException ioException = (IOException) completionException.getCause();
            if (logLevel != Logger.Level.NONE) {
              logger.logIOException(metadata.configKey(), logLevel, ioException,
                  elapsedTime(start));
            }

            throw errorExecuting(request, ioException); // 这里会包装成RetryableException再抛出
          } else {
            throw completionException;
          }
        })
        .thenCompose(response -> handleResponse(response, elapsedTime(start)));
  }

这是ErrorDecoder的代码

public Exception decode(String methodKey, Response response) {
    FeignException exception = errorStatus(methodKey, response, maxBodyBytesLength,
        maxBodyCharsLength);
    Long retryAfter = retryAfterDecoder.apply(firstOrNull(response.headers(), RETRY_AFTER)); // RETRY_AFTER 是http头里面的字符串 Retry-After，有这个字符串就抛出 RetryableException
    if (retryAfter != null) {
    return new RetryableException(
        response.status(),
        exception.getMessage(),
        response.request().httpMethod(),
        exception,
        retryAfter,
        response.request());
    }
    return exception;
}

来看3中executeAndDecode请求是如何执行的

executeAndDecode 执行请求

Object executeAndDecode(RequestTemplate template, Options options) throws Throwable {
    Request request = targetRequest(template); // 主要是遍历RequestInterceptor，装填interceptor，可以实现这个接口，对请求头加一些操作

    if (logLevel != Logger.Level.NONE) {
      logger.logRequest(metadata.configKey(), logLevel, request);
    }

    Response response;
    long start = System.nanoTime();
    try {
      response = client.execute(request, options); // 这里就开始执行请求了
      // ensure the request is set. TODO: remove in Feign 12
      response = response.toBuilder()
          .request(request)
          .requestTemplate(template)
          .build();
    } catch (IOException e) {
      if (logLevel != Logger.Level.NONE) {
        logger.logIOException(metadata.configKey(), logLevel, e, elapsedTime(start));
      }
      throw errorExecuting(request, e);
    }

    long elapsedTime = TimeUnit.NANOSECONDS.toMillis(System.nanoTime() - start);
    return responseHandler.handleResponse(
        metadata.configKey(), response, metadata.returnType(), elapsedTime);
  }

execute的过程

public Response execute(Request request, Request.Options options) throws IOException {
    final URI originalUri = URI.create(request.url());
    String serviceId = originalUri.getHost();
    Assert.state(serviceId != null, "Request URI does not contain a valid hostname: " + originalUri);
    final LoadBalancedRetryPolicy retryPolicy = loadBalancedRetryFactory.createRetryPolicy(serviceId,
            loadBalancerClient); // 这里创建的是 BlockingLoadBalancedRetryPolicy，这里面代码不多，简单来说就是，默认get肯定重试，或者如果配置文件指定哪些请求可以重试。哪些状态码和异常可以重试。重试次数通过配置文件的次数限制。
    RetryTemplate retryTemplate = buildRetryTemplate(serviceId, request, retryPolicy);
    return retryTemplate.execute(context -> {
        Request feignRequest = null;
        ServiceInstance retrievedServiceInstance = null;
        Set<LoadBalancerLifecycle> supportedLifecycleProcessors = LoadBalancerLifecycleValidator
                .getSupportedLifecycleProcessors(
                        loadBalancerClientFactory.getInstances(serviceId, LoadBalancerLifecycle.class),
                        RetryableRequestContext.class, ResponseData.class, ServiceInstance.class); // 返回空
        String hint = getHint(serviceId);
        DefaultRequest<RetryableRequestContext> lbRequest = new DefaultRequest<>(
                new RetryableRequestContext(null, buildRequestData(request), hint));
        // On retries the policy will choose the server and set it in the context
        // and extract the server and update the request being made
        if (context instanceof LoadBalancedRetryContext lbContext) {
            retrievedServiceInstance = lbContext.getServiceInstance();
            if (retrievedServiceInstance == null) {
                if (LOG.isDebugEnabled()) {
                    LOG.debug("Service instance retrieved from LoadBalancedRetryContext: was null. "
                            + "Reattempting service instance selection");
                }
                ServiceInstance previousServiceInstance = lbContext.getPreviousServiceInstance(); // 这里没有值
                lbRequest.getContext().setPreviousServiceInstance(previousServiceInstance);
                supportedLifecycleProcessors.forEach(lifecycle -> lifecycle.onStart(lbRequest)); // supportedLifecycleProcessors 是空的
                retrievedServiceInstance = loadBalancerClient.choose(serviceId, lbRequest); // 最终在这里获取实例
                if (LOG.isDebugEnabled()) {
                    LOG.debug(String.format("Selected service instance: %s", retrievedServiceInstance));
                }
                lbContext.setServiceInstance(retrievedServiceInstance);
            }

            if (retrievedServiceInstance == null) {
                if (LOG.isWarnEnabled()) {
                    LOG.warn("Service instance was not resolved, executing the original request");
                }
                org.springframework.cloud.client.loadbalancer.Response<ServiceInstance> lbResponse = new DefaultResponse(
                        retrievedServiceInstance);
                supportedLifecycleProcessors.forEach(lifecycle -> lifecycle
                        .onComplete(new CompletionContext<ResponseData, ServiceInstance, RetryableRequestContext>(
                                CompletionContext.Status.DISCARD, lbRequest, lbResponse)));
                feignRequest = request;
            }
            else {
                if (LOG.isDebugEnabled()) {
                    LOG.debug(String.format("Using service instance from LoadBalancedRetryContext: %s",
                            retrievedServiceInstance));
                }
                String reconstructedUrl = loadBalancerClient.reconstructURI(retrievedServiceInstance, originalUri)
                        .toString();
                feignRequest = buildRequest(request, reconstructedUrl, retrievedServiceInstance);
            }
        }
        org.springframework.cloud.client.loadbalancer.Response<ServiceInstance> lbResponse = new DefaultResponse(
                retrievedServiceInstance);
        LoadBalancerProperties loadBalancerProperties = loadBalancerClientFactory.getProperties(serviceId);
        Response response = LoadBalancerUtils.executeWithLoadBalancerLifecycleProcessing(delegate, options,
                feignRequest, lbRequest, lbResponse, supportedLifecycleProcessors,
                retrievedServiceInstance != null);
        int responseStatus = response.status();
        if (retryPolicy != null && retryPolicy.retryableStatusCode(responseStatus)) {
            if (LOG.isDebugEnabled()) {
                LOG.debug(String.format("Retrying on status code: %d", responseStatus));
            }
            byte[] byteArray = response.body() == null ? new byte[] {}
                    : StreamUtils.copyToByteArray(response.body().asInputStream());
            response.close();
            throw new LoadBalancerResponseStatusCodeException(serviceId, response, byteArray,
                    URI.create(request.url()));
        }
        return response;
    }, new LoadBalancedRecoveryCallback<Response, Response>() {
        @Override
        protected Response createResponse(Response response, URI uri) {
            return response;
        }
    });
}

loadBalancer.choose的调用

public <T> ServiceInstance choose(String serviceId, Request<T> request) {
    ReactiveLoadBalancer<ServiceInstance> loadBalancer = loadBalancerClientFactory.getInstance(serviceId); // 这里面会返回NacosLoadBalancer
    if (loadBalancer == null) {
        return null;
    }
    Response<ServiceInstance> loadBalancerResponse = Mono.from(loadBalancer.choose(request)).block(); // block表示同步的，同步的从loadBalancer，当前是Nacos，里面获取一个可用的实例，看一下choose
    if (loadBalancerResponse == null) {
        return null;
    }
    return loadBalancerResponse.getServer();
}
public Mono<Response<ServiceInstance>> choose(Request request) {
    ServiceInstanceListSupplier supplier = serviceInstanceListSupplierProvider
            .getIfAvailable(NoopServiceInstanceListSupplier::new);
    return supplier.get(request).next().map(this::getInstanceResponse);
}

getInstanceResponse会从多个节点里面选一个出来

private Response<ServiceInstance> getInstanceResponse(
        List<ServiceInstance> serviceInstances) {
    if (serviceInstances.isEmpty()) {
        log.warn("No servers available for service: " + this.serviceId);
        return new EmptyResponse();
    }

    try {
        String clusterName = this.nacosDiscoveryProperties.getClusterName();

        List<ServiceInstance> instancesToChoose = serviceInstances; // 这里拿到的是所有的实例
        if (StringUtils.isNotBlank(clusterName)) { // 通过配置 clusterName，会优先调用clusterName一样的集群里面的机器
            List<ServiceInstance> sameClusterInstances = serviceInstances.stream()
                    .filter(serviceInstance -> {
                        String cluster = serviceInstance.getMetadata()
                                .get("nacos.cluster");
                        return StringUtils.equals(cluster, clusterName);
                    }).collect(Collectors.toList());
            if (!CollectionUtils.isEmpty(sameClusterInstances)) {
                instancesToChoose = sameClusterInstances;
            }
        }
        else { // 跨集群调用，打个日志
            log.warn(
                    "A cross-cluster call occurs，name = {}, clusterName = {}, instance = {}",
                    serviceId, clusterName, serviceInstances);
        }
        instancesToChoose = this.filterInstanceByIpType(instancesToChoose);

        ServiceInstance instance = NacosBalancer
                .getHostByRandomWeight3(instancesToChoose);  // 通过这个算法，选出一个可用的节点。首先判断节点是否健康，健康才能选，然后从健康的里面，根据权重，随机选一个

        return new DefaultResponse(instance);
    }
    catch (Exception e) {
        log.warn("NacosLoadBalancer error", e);
        return null;
    }
}

然后我们退回发送请求的地方

public Response execute(Request request, Request.Options options) throws IOException {
    final URI originalUri = URI.create(request.url());
    String serviceId = originalUri.getHost();
    Assert.state(serviceId != null, "Request URI does not contain a valid hostname: " + originalUri);
    final LoadBalancedRetryPolicy retryPolicy = loadBalancedRetryFactory.createRetryPolicy(serviceId,
            loadBalancerClient);
    RetryTemplate retryTemplate = buildRetryTemplate(serviceId, request, retryPolicy);
    return retryTemplate.execute(context -> {
        Request feignRequest = null;
        ServiceInstance retrievedServiceInstance = null;
        Set<LoadBalancerLifecycle> supportedLifecycleProcessors = LoadBalancerLifecycleValidator
                .getSupportedLifecycleProcessors(
                        loadBalancerClientFactory.getInstances(serviceId, LoadBalancerLifecycle.class),
                        RetryableRequestContext.class, ResponseData.class, ServiceInstance.class);
        String hint = getHint(serviceId);
        DefaultRequest<RetryableRequestContext> lbRequest = new DefaultRequest<>(
                new RetryableRequestContext(null, buildRequestData(request), hint));
        // On retries the policy will choose the server and set it in the context
        // and extract the server and update the request being made
        if (context instanceof LoadBalancedRetryContext lbContext) {
            retrievedServiceInstance = lbContext.getServiceInstance();
            if (retrievedServiceInstance == null) {
                if (LOG.isDebugEnabled()) {
                    LOG.debug("Service instance retrieved from LoadBalancedRetryContext: was null. "
                            + "Reattempting service instance selection");
                }
                ServiceInstance previousServiceInstance = lbContext.getPreviousServiceInstance();
                lbRequest.getContext().setPreviousServiceInstance(previousServiceInstance);
                supportedLifecycleProcessors.forEach(lifecycle -> lifecycle.onStart(lbRequest));
                retrievedServiceInstance = loadBalancerClient.choose(serviceId, lbRequest); // 从这里退出来的，现在拿到实例了，里面有ip和端口，就能发请求了
                if (LOG.isDebugEnabled()) {
                    LOG.debug(String.format("Selected service instance: %s", retrievedServiceInstance));
                }
                lbContext.setServiceInstance(retrievedServiceInstance);
            }

            if (retrievedServiceInstance == null) {
                if (LOG.isWarnEnabled()) {
                    LOG.warn("Service instance was not resolved, executing the original request");
                }
                org.springframework.cloud.client.loadbalancer.Response<ServiceInstance> lbResponse = new DefaultResponse(
                        retrievedServiceInstance);
                supportedLifecycleProcessors.forEach(lifecycle -> lifecycle
                        .onComplete(new CompletionContext<ResponseData, ServiceInstance, RetryableRequestContext>(
                                CompletionContext.Status.DISCARD, lbRequest, lbResponse)));
                feignRequest = request;
            }
            else {
                if (LOG.isDebugEnabled()) {
                    LOG.debug(String.format("Using service instance from LoadBalancedRetryContext: %s",
                            retrievedServiceInstance));
                }
                String reconstructedUrl = loadBalancerClient.reconstructURI(retrievedServiceInstance, originalUri)
                        .toString(); // 重新编写url，把ip和端口加进去
                feignRequest = buildRequest(request, reconstructedUrl, retrievedServiceInstance); // 根据url构建request
            }
        }
        org.springframework.cloud.client.loadbalancer.Response<ServiceInstance> lbResponse = new DefaultResponse(
                retrievedServiceInstance);
        LoadBalancerProperties loadBalancerProperties = loadBalancerClientFactory.getProperties(serviceId);
        Response response = LoadBalancerUtils.executeWithLoadBalancerLifecycleProcessing(delegate, options,
                feignRequest, lbRequest, lbResponse, supportedLifecycleProcessors,
                retrievedServiceInstance != null); // 实际去发送请求
        int responseStatus = response.status();
        if (retryPolicy != null && retryPolicy.retryableStatusCode(responseStatus)) {
            if (LOG.isDebugEnabled()) {
                LOG.debug(String.format("Retrying on status code: %d", responseStatus));
            }
            byte[] byteArray = response.body() == null ? new byte[] {}
                    : StreamUtils.copyToByteArray(response.body().asInputStream());
            response.close();
            throw new LoadBalancerResponseStatusCodeException(serviceId, response, byteArray,
                    URI.create(request.url()));
        }
            return response;
		}, new LoadBalancedRecoveryCallback<Response, Response>() {
			@Override
			protected Response createResponse(Response response, URI uri) {
				return response;
			}
		});
}

最后就是发送请求了

static Response executeWithLoadBalancerLifecycleProcessing(Client feignClient, Request.Options options,
        Request feignRequest, org.springframework.cloud.client.loadbalancer.Request lbRequest,
        org.springframework.cloud.client.loadbalancer.Response<ServiceInstance> lbResponse,
        Set<LoadBalancerLifecycle> supportedLifecycleProcessors, boolean loadBalanced) throws IOException {
    supportedLifecycleProcessors.forEach(lifecycle -> lifecycle.onStartRequest(lbRequest, lbResponse));
    try {
        Response response = feignClient.execute(feignRequest, options); // 发送请求
        if (loadBalanced) {
            supportedLifecycleProcessors.forEach(
                    lifecycle -> lifecycle.onComplete(new CompletionContext<>(CompletionContext.Status.SUCCESS,
                            lbRequest, lbResponse, buildResponseData(response))));
        }
        return response;
    }
    catch (Exception exception) {
        if (loadBalanced) {
            supportedLifecycleProcessors.forEach(lifecycle -> lifecycle.onComplete(
                    new CompletionContext<>(CompletionContext.Status.FAILED, exception, lbRequest, lbResponse)));
        }
        throw exception;
    }
}
public Response execute(Request request, Options options) throws IOException {
    HttpURLConnection connection = convertAndSend(request, options);
    return convertResponse(connection, request);
}
HttpURLConnection convertAndSend(Request request, Options options) throws IOException {
    final URL url = new URL(request.url());
    final HttpURLConnection connection = this.getConnection(url);
   //......
}

最后的convertAndSend就不仔细看了，目前代码可知，使用的是java.net的发送方式。如果用httpclient，feign会自动装配，只需要引入依赖即可。

总结and面试

首先是feign要初始化bean，EnableFeignClients要写上feign在哪些包里面，扫描出来。 feign的bean是用FeignClientFactoryBean实现的，听名字就知道这实现了FactoryBean接口，里面getObject实现的很复杂。里面需要设置请求url，设置负载均衡的bean等信息。然后会用jdk代理的方式，把这个feign的bean执行代理。实现一个FeignInvocationHandler，重写invoke方法，把远程调用这个过程通过代理实现。 然后在请求的时候，因为要走代理，代理里面实现了如下几个内容。一个是参数信息，比如请求超时时间，这个可以通过函数参数传进来，然后这些参数会封装到请求里面。代理还帮忙做了一些重试，默认是不重试的。 发送请求的位置在executeAndDecode函数。我们可以实现 RequestInterceptor 接口，发送请求的时候，会执行这个interceptor，对header做一些处理。然后因为一般微服务都是不直接写ip的，要向配置中心获取实例地址。假设配置中心是nacos，向nacos请求获得所有的实例，然后nacos会有算法，从健康的实例里面选出一个实例，这样就有ip端口了，就可以得到完整的url。然后在编码一下请求体，就可以发送请求了。