小蚂蚁 发表于 2021-9-18 19:35:57

Spring-cloud Feign 的深入理解

这篇文章主要介绍了Spring-cloud Feign 的深入理解,小编觉得挺不错的,现在分享给大家,也给大家做个参考。一起跟随小编过来看看吧
feign的调用流程
读取注解信息:enablefeignclients-->feignclientsregistrar-->feignclientfactorybean
feigh流程:reflectivefeign-->contract-->synchronousmethodhandler
相关configuration:feignclientsconfiguration,feignautoconfiguration,defaultfeignloadbalancedconfiguration,feignribbonclientautoconfiguration(ribbon)
在feignclientsregistrar中:


@override
public void registerbeandefinitions(annotationmetadata metadata,
    beandefinitionregistry registry) {
//注册feign配置信息
registerdefaultconfiguration(metadata, registry);
//注册feign client
registerfeignclients(metadata, registry);
}

private void registerfeignclient(beandefinitionregistry registry,
    annotationmetadata annotationmetadata, map<string, object> attributes) {
string classname = annotationmetadata.getclassname();
//准备注入feignclientfactorybean
beandefinitionbuilder definition = beandefinitionbuilder
      .genericbeandefinition(feignclientfactorybean.class);
...
}
查看feignclientfactorybean:


@override
public object getobject() throws exception {
feigncontext context = applicationcontext.getbean(feigncontext.class);
//构建feign.builder
feign.builder builder = feign(context);
//如果注解没有指定url
if (!stringutils.hastext(this.url)) {
    string url;
    if (!this.name.startswith("http")) {
      url = "http://" + this.name;
    }
    else {
      url = this.name;
    }
    url += cleanpath();
    return loadbalance(builder, context, new hardcodedtarget<>(this.type,
      this.name, url));
}
//如果指定了url
if (stringutils.hastext(this.url) && !this.url.startswith("http")) {
    this.url = "http://" + this.url;
}
string url = this.url + cleanpath();
client client = getoptional(context, client.class);
if (client != null) {
    if (client instanceof loadbalancerfeignclient) {
      // 因为指定了url且classpath下有ribbon,获取client的delegate(unwrap)
      // not load balancing because we have a url,
      // but ribbon is on the classpath, so unwrap
      client = ((loadbalancerfeignclient)client).getdelegate();
    }
    builder.client(client);
}
targeter targeter = get(context, targeter.class);
return targeter.target(this, builder, context, new hardcodedtarget<>(
      this.type, this.name, url));
}

protected <t> t loadbalance(feign.builder builder, feigncontext context,
    hardcodedtarget<t> target) {
//获取feign client实例
client client = getoptional(context, client.class);
if (client != null) {
    builder.client(client);
    //defaulttargeter或者hystrixtargeter
    targeter targeter = get(context, targeter.class);
    //调用builder的target,其中就调用了feign的newinstance
    return targeter.target(this, builder, context, target);
}

throw new illegalstateexception(
      "no feign client for loadbalancing defined. did you forget to include spring-cloud-starter-netflix-ribbon?");
}
在feignclientsconfiguration配置了feign.builder,prototype类型:


@bean
@scope("prototype")
@conditionalonmissingbean
public feign.builder feignbuilder(retryer retryer) {
return feign.builder().retryer(retryer);
}
feign的builder.build返回了一个reflectivefeign:


public feign build() {
synchronousmethodhandler.factory synchronousmethodhandlerfactory =
   new synchronousmethodhandler.factory(client, retryer, requestinterceptors, logger,
                      loglevel, decode404);
parsehandlersbyname handlersbyname =
   new parsehandlersbyname(contract, options, encoder, decoder,
               errordecoder, synchronousmethodhandlerfactory);
//reflectivefeign构造参数
//parsehandlersbyname作用是通过传入的target返回代理接口下的方法的各种信息(methodhandler)
//contract:解析接口的方法注解规则,生成methodmetadata
//options:request超时配置
//encoder:请求编码器
//decoder:返回解码器
//errordecoder:错误解码器
//synchronousmethodhandler.factory是构建synchronousmethodhandler的工厂
//client:代表真正执行http的组件
//retryer:该组决定了在http请求失败时是否需要重试
//requestinterceptor:请求前的拦截器
//logger:记录日志组件,包含各个阶段记录日志的方法和留给用户自己实现的log方法
//logger.level:日志级别
//decode404:处理404的策略,返回空还是报错
//synchronousmethodhandlerfactory通过所有的信息去包装一个synchronousmethodhandler,在调用invoke方法的时候执行http
return new reflectivefeign(handlersbyname, invocationhandlerfactory);
}
在调用feign.builder的target的时候,调用了reflectivefeign.newinstance:


/**
* creates an api binding to the {@code target}. as this invokes reflection, care should be taken
* to cache the result.
*/
@suppresswarnings("unchecked")
@override
//接收target参数(包含feign代理接口的类型class,名称,http url)
public <t> t newinstance(target<t> target) {
//首先通过**parsehandlersbyname**解析出接口中包含的方法,包装requesttemplate,组装成<name, methodhandler>
map<string, methodhandler> nametohandler = targettohandlersbyname.apply(target);
map<method, methodhandler> methodtohandler = new linkedhashmap<method, methodhandler>();
//接口default方法list
list<defaultmethodhandler> defaultmethodhandlers = new linkedlist<defaultmethodhandler>();

for (method method : target.type().getmethods()) {
if (method.getdeclaringclass() == object.class) {
   continue;
} else if(util.isdefault(method)) {
   defaultmethodhandler handler = new defaultmethodhandler(method);
   defaultmethodhandlers.add(handler);
   methodtohandler.put(method, handler);
} else {
   methodtohandler.put(method, nametohandler.get(feign.configkey(target.type(), method)));
}
}
//invocationhandlerfactory.default()返回了一个reflectivefeign.feigninvocationhandler对象,通过传入的methodhandler map 调用目标对象的对应方法
invocationhandler handler = factory.create(target, methodtohandler);
//生成jdk代理对象
t proxy = (t) proxy.newproxyinstance(target.type().getclassloader(), new class<?>[]{target.type()}, handler);
//绑定接口的默认方法到代理对象
for(defaultmethodhandler defaultmethodhandler : defaultmethodhandlers) {
defaultmethodhandler.bindto(proxy);
}
return proxy;
}
生成feign代理对象的基本流程图:

当调用接口方法时,实际上就是调用代理对象invoke方法:


@override
public object invoke(object[] argv) throws throwable {
//工厂创建请求模版
requesttemplate template = buildtemplatefromargs.create(argv);
//每次克隆一个新的retryer
retryer retryer = this.retryer.clone();
while (true) {
try {
   //这里调用实际的feign client execute
   return executeanddecode(template);
} catch (retryableexception e) {
   //失败重试
   retryer.continueorpropagate(e);
   if (loglevel != logger.level.none) {
    logger.logretry(metadata.configkey(), loglevel);
   }
   continue;
}
}
}
在defaultfeignloadbalancedconfiguration里实例化了loadbalancerfeignclient


@override
public response execute(request request, request.options options) throws ioexception {
try {
    uri asuri = uri.create(request.url());
    string clientname = asuri.gethost();
    uri uriwithouthost = cleanurl(request.url(), clientname);
    //delegate这里是client.default实例,底层调用的是java.net原生网络访问
    feignloadbalancer.ribbonrequest ribbonrequest = new feignloadbalancer.ribbonrequest(
      this.delegate, request, uriwithouthost);

    iclientconfig requestconfig = getclientconfig(options, clientname);
    //executewithloadbalancer会根据ribbon的负载均衡算法构建url,这里不展开
    return lbclient(clientname).executewithloadbalancer(ribbonrequest,
      requestconfig).toresponse();
}
catch (clientexception e) {
    ioexception io = findioexception(e);
    if (io != null) {
      throw io;
    }
    throw new runtimeexception(e);
}
}
以上就是本文的全部内容,希望对大家的学习有所帮助,也希望大家多多支持CodeAE代码之家。
原文链接:https://segmentfault.com/a/1190000018077675

http://www.zzvips.com/article/176520.html
页: [1]
查看完整版本: Spring-cloud Feign 的深入理解