Spring zuul 快速入门实践 --服务转发实现解析

　　zuul 作为springCloud 的全家桶组件之一，有着不可或缺的分量。它作为一个普通java API网关，自有网关的好处：

　　　　避免将内部信息暴露给外部；
　　　　统一服务端应用入口；
　　　　为微服务添加额外的安全层；
　　　　支持混合通信协议；
　　　　降低构建微服务的复杂性；
　　　　微服务模拟与虚拟化；

　　zuul 基本上已经被springCloud 处理为一个开箱即用的一个组件了，所以基本上只需要添加相应依赖和一些必要配置，该网关就可以跑起来了。（这表面和nginx反向代理部分功能看起来是差不多的）

　　让我们来快速实践一下吧！

 

一、zuul入坑基本实践步骤

1.1. 引入 pom 依赖

    <parent>
        <groupId>org.springframework.bootgroupId>
        <artifactId>spring-boot-starter-parentartifactId>
        <version>2.0.5.RELEASEversion>
        <relativePath/> 
    parent>
    
    <modelVersion>4.0.0modelVersion>
    <groupId>zuul-testgroupId>
    <artifactId>com.yougeartifactId>
    <version>1.0version>
    
    
    <dependencyManagement>
        <dependencies>
            <dependency>
                <groupId>org.springframework.cloudgroupId>
                <artifactId>spring-cloud-dependenciesartifactId>
                <version>Finchley.RC2version>
                <type>pomtype>
                <scope>importscope>
            dependency>
        dependencies>
    dependencyManagement>
    
    <dependencies>
        
        <dependency>
            <groupId>org.springframework.cloudgroupId>
            <artifactId>spring-cloud-starter-netflix-zuulartifactId>
        dependency>
    dependencies>

　　以上就是我们整个demo的全部maven依赖了，很简洁吧。这也是springboot的初衷，把所有的工作都移到幕后，让业务更简洁。

 

1.2. 编写网关入口类

　　如下为整个网关的入口类，实际上就是两个注解发生了化学反应。@EnableZuulProxy 是本文的主角，它会开启网关相关的服务。

@SpringBootApplication
@EnableZuulProxy
public class MyZuulGateway {
    // 只有一个空的main方法
    public static void main(String[] args) {
        SpringApplication.run(MyZuulGateway.class, args);
    }
}

　　就是这么简单！

 

1.3. 添加测试配置项

　　在application.properties配置文件中添加如下配置，主要使用一个路由配置验证即可！

server.port=9000
spring.application.name=my-zuul-gateway

#本地环境配置zuul转发的规则：
# 忽略所有微服务，只路由指定微服务
# 如下配置为将 /sapi/** 的路径请求，转发到 http://127.0.0.1:8082/fileenc/ 上去。
zuul.ignored-services=*
zuul.routes.fileenc1.url=http://127.0.0.1:8082/fileenc/
zuul.routes.fileenc1.path=/sapi/**

　　如上就可以将网关跑起来了，如果你连后台服务也没有，没关系，自己写一个就好了。

    @GetMapping("hello")
    public Object hello() {
        return "hello, world";
    }

　　

1.4. 测试网关

　　以上就已经将整个网关搞好了，run一下就ok. 测试方式就是直接浏览器里访问下该网关地址就好了:

http://localhost:9000/sapi/test/hello?a=1&b=22 .

　　如果你看到 “hello, world”, 恭喜你，zuul已入坑。

 

二、zuul是如何转发请求的？

　　根据上面的观察，zuul已经基本可以满足我们的开发需求了，后续更多要做的可能就是一些安全相关，业务相关，优化相关的东西了。不过在做这些之前，我们可以先多问一个问题，zuul是如何将请求转发给后台服务的呢？

　　这实际上和zuul的架构相关：

 

　　zuul的中核心概念是：Filter. 运行时逻辑上分为多种类型的Filter，各类型Filter处理时机不同！  PRE：这种过滤器在请求被路由之前调用；ROUTING：这种过滤器将请求路由到微服务；POST：这种过滤器在路由到微服务以后执行；ERROR：在其他阶段发生错误时执行该过滤器；

　　所以，整体上来说，它的转发流程会经过一系列的过滤器，然后再进行实际的转发。

如果只想了解其最终是如何转的可以直奔主题，而如果要添加你的功能，则需要编写一些对应生命周期的过滤器。

　　原本要分析zuul是如何处理请求的，但是实际上，zuul被整合到spring之后，就完全地符合了一个springmvc的编程模型了。所有对该网关的请求会先调用 ZuulController 进行请求的接收，然后到 service处理，再到响应这么一个过程。

　　整个 ZuulController 非常地简单：就是一个请求的委托过程！

// org.springframework.cloud.netflix.zuul.web.ZuulController
public class ZuulController extends ServletWrappingController {

    public ZuulController() {
        setServletClass(ZuulServlet.class);
        setServletName("zuul");
        setSupportedMethods((String[]) null); // Allow all
    }

    @Override
    public ModelAndView handleRequest(HttpServletRequest request, HttpServletResponse response) throws Exception {
        try {
            // We don't care about the other features of the base class, just want to
            // handle the request
            return super.handleRequestInternal(request, response);
        }
        finally {
            // @see com.netflix.zuul.context.ContextLifecycleFilter.doFilter
            RequestContext.getCurrentContext().unset();
        }
    }

}
    // org.springframework.web.servlet.mvc.ServletWrappingController#handleRequestInternal
    /**
     * Invoke the wrapped Servlet instance.
     * @see javax.servlet.Servlet#service(javax.servlet.ServletRequest, javax.servlet.ServletResponse)
     */
    @Override
    protected ModelAndView handleRequestInternal(HttpServletRequest request, HttpServletResponse response)
            throws Exception {

        Assert.state(this.servletInstance != null, "No Servlet instance");
        // 该 servletInstance 是 ZuulServlet, 整个zuul的实现框架由其控制
        this.servletInstance.service(request, response);
        return null;
    }
    // com.netflix.zuul.http.ZuulServlet#service
    @Override
    public void service(javax.servlet.ServletRequest servletRequest, javax.servlet.ServletResponse servletResponse) throws ServletException, IOException {
        try {
            // 初始化请求，由 zuulRunner 处理
            init((HttpServletRequest) servletRequest, (HttpServletResponse) servletResponse);

            // Marks this request as having passed through the "Zuul engine", as opposed to servlets
            // explicitly bound in web.xml, for which requests will not have the same data attached
            // setZuulEngineRan 会旋转一个标识： "zuulEngineRan", true
            RequestContext context = RequestContext.getCurrentContext();
            context.setZuulEngineRan();

            try {
                // 前置过滤器
                preRoute();
            } catch (ZuulException e) {
                error(e);
                // 异常时直接调用后置路由完成请求
                postRoute();
                return;
            }
            try {
                // 正常的路由请求处理
                route();
            } catch (ZuulException e) {
                error(e);
                postRoute();
                return;
            }
            try {
                // 正常地后置路由处理
                postRoute();
            } catch (ZuulException e) {
                error(e);
                return;
            }

        } catch (Throwable e) {
            error(new ZuulException(e, 500, "UNHANDLED_EXCEPTION_" + e.getClass().getName()));
        } finally {
            // 重置上下文，以备下次使用
            RequestContext.getCurrentContext().unset();
        }
    }

　　以上就是整个zuul对于普通请求的处理框架部分了。逻辑还是比较清晰的，简单的，前置+转发+后置处理。我们就几个重点部分说明一下：

 

2.1. 请求初始化

　　该部分主要是将外部请求，接入到 zuul 的处理流程上，当然下面的实现主要是使用了 ThreadLocal 实现了上下文的衔接。

    // com.netflix.zuul.http.ZuulServlet#init
    /**
     * initializes request
     *
     * @param servletRequest
     * @param servletResponse
     */
    void init(HttpServletRequest servletRequest, HttpServletResponse servletResponse) {
        zuulRunner.init(servletRequest, servletResponse);
    }
    // com.netflix.zuul.ZuulRunner#init
    /**
     * sets HttpServlet request and HttpResponse
     *
     * @param servletRequest
     * @param servletResponse
     */
    public void init(HttpServletRequest servletRequest, HttpServletResponse servletResponse) {
        // RequestContext 使用 ThreadLocal 进行保存，且保证有值
        // 且 RequestContext 继承了 ConcurrentHashMap, 保证了操作的线程安全
        RequestContext ctx = RequestContext.getCurrentContext();
        if (bufferRequests) {
            ctx.setRequest(new HttpServletRequestWrapper(servletRequest));
        } else {
            ctx.setRequest(servletRequest);
        }

        ctx.setResponse(new HttpServletResponseWrapper(servletResponse));
    }

　　以上就是一个 zuul 请求的初始化了，简单地说就是设置好请求上下文，备用。

 

2.2. 前置处理过滤器

　　前置处理过滤器主要用于标记一些请求类型，权限验证，安全过滤等等。是不可或缺一环。具体实现自行处理！我们来看一个整体的通用流程：

    // com.netflix.zuul.http.ZuulServlet#preRoute
    /**
     * executes "pre" filters
     *
     * @throws ZuulException
     */
    void preRoute() throws ZuulException {
        zuulRunner.preRoute();
    }
    // com.netflix.zuul.ZuulRunner#preRoute
    /**
     * executes "pre" filterType  ZuulFilters
     *
     * @throws ZuulException
     */
    public void preRoute() throws ZuulException {
        // FilterProcessor 是个单例
        FilterProcessor.getInstance().preRoute();
    }
    // com.netflix.zuul.FilterProcessor#preRoute
    /**
     * runs all "pre" filters. These filters are run before routing to the orgin.
     *
     * @throws ZuulException
     */
    public void preRoute() throws ZuulException {
        try {
            // 调用Type 为 pre 的过滤器
            runFilters("pre");
        } catch (ZuulException e) {
            throw e;
        } catch (Throwable e) {
            throw new ZuulException(e, 500, "UNCAUGHT_EXCEPTION_IN_PRE_FILTER_" + e.getClass().getName());
        }
    }
    // com.netflix.zuul.FilterProcessor#runFilters
    /**
     * runs all filters of the filterType sType/ Use this method within filters to run custom filters by type
     *
     * @param sType the filterType.
     * @return
     * @throws Throwable throws up an arbitrary exception
     */
    public Object runFilters(String sType) throws Throwable {
        if (RequestContext.getCurrentContext().debugRouting()) {
            Debug.addRoutingDebug("Invoking {" + sType + "} type filters");
        }
        boolean bResult = false;
        // 通过 FilterLoader 的单例，获取所有注册为 sType 的过滤器
        // 存放 Filters 的容器自然也是线程安全的，为 ConcurrentHashMap
        // - org.springframework.cloud.netflix.zuul.filters.pre.ServletDetectionFilter
        // - org.springframework.cloud.netflix.zuul.filters.pre.Servlet30WrapperFilter
        // - org.springframework.cloud.netflix.zuul.filters.pre.FormBodyWrapperFilter
        // - org.springframework.cloud.netflix.zuul.filters.pre.DebugFilter
        // - org.springframework.cloud.netflix.zuul.filters.pre.PreDecorationFilter
        List list = FilterLoader.getInstance().getFiltersByType(sType);
        if (list != null) {
            for (int i = 0; i < list.size(); i++) {
                ZuulFilter zuulFilter = list.get(i);
                // 依次处理每个 filter
                Object result = processZuulFilter(zuulFilter);
                if (result != null && result instanceof Boolean) {
                    bResult |= ((Boolean) result);
                }
            }
        }
        return bResult;
    }
    // 获取相应的 filters
    // com.netflix.zuul.FilterLoader#getFiltersByType
    /**
     * Returns a list of filters by the filterType specified
     *
     * @param filterType
     * @return a List
     */
    public List getFiltersByType(String filterType) {

        List list = hashFiltersByType.get(filterType);
        if (list != null) return list;

        list = new ArrayList();

        Collection filters = filterRegistry.getAllFilters();
        for (Iterator iterator = filters.iterator(); iterator.hasNext(); ) {
            ZuulFilter filter = iterator.next();
            if (filter.filterType().equals(filterType)) {
                list.add(filter);
            }
        }
        Collections.sort(list); // sort by priority

        hashFiltersByType.putIfAbsent(filterType, list);
        return list;
    }

    // com.netflix.zuul.FilterProcessor#processZuulFilter
    /**
     * Processes an individual ZuulFilter. This method adds Debug information. Any uncaught Thowables are caught by this method and converted to a ZuulException with a 500 status code.
     *
     * @param filter
     * @return the return value for that filter
     * @throws ZuulException
     */
    public Object processZuulFilter(ZuulFilter filter) throws ZuulException {

        RequestContext ctx = RequestContext.getCurrentContext();
        boolean bDebug = ctx.debugRouting();
        final String metricPrefix = "zuul.filter-";
        long execTime = 0;
        String filterName = "";
        try {
            long ltime = System.currentTimeMillis();
            filterName = filter.getClass().getSimpleName();
            
            RequestContext copy = null;
            Object o = null;
            Throwable t = null;

            if (bDebug) {
                Debug.addRoutingDebug("Filter " + filter.filterType() + " " + filter.filterOrder() + " " + filterName);
                copy = ctx.copy();
            }
            // 调用各filter的 runFilter() 方法，触发filter作用
            // 如果filter被禁用，则不会调用 zuul.ServletDetectionFilter.pre.disable=true, 代表禁用 pre
            // 具体实现逻辑由各 filter 决定 
            ZuulFilterResult result = filter.runFilter();
            ExecutionStatus s = result.getStatus();
            execTime = System.currentTimeMillis() - ltime;

            switch (s) {
                case FAILED:
                    t = result.getException();
                    ctx.addFilterExecutionSummary(filterName, ExecutionStatus.FAILED.name(), execTime);
                    break;
                case SUCCESS:
                    o = result.getResult();
                    // 使用 StringBuilder 记录请求处理日志
                    ctx.addFilterExecutionSummary(filterName, ExecutionStatus.SUCCESS.name(), execTime);
                    if (bDebug) {
                        Debug.addRoutingDebug("Filter {" + filterName + " TYPE:" + filter.filterType() + " ORDER:" + filter.filterOrder() + "} Execution time = " + execTime + "ms");
                        Debug.compareContextState(filterName, copy);
                    }
                    break;
                default:
                    break;
            }
            // 只要发生异常，则抛出
            if (t != null) throw t;
            // 请求计数器增加
            usageNotifier.notify(filter, s);
            return o;

        } catch (Throwable e) {
            if (bDebug) {
                Debug.addRoutingDebug("Running Filter failed " + filterName + " type:" + filter.filterType() + " order:" + filter.filterOrder() + " " + e.getMessage());
            }
            usageNotifier.notify(filter, ExecutionStatus.FAILED);
            if (e instanceof ZuulException) {
                throw (ZuulException) e;
            } else {
                ZuulException ex = new ZuulException(e, "Filter threw Exception", 500, filter.filterType() + ":" + filterName);
                ctx.addFilterExecutionSummary(filterName, ExecutionStatus.FAILED.name(), execTime);
                throw ex;
            }
        }
    }
    // com.netflix.zuul.ZuulFilter#runFilter
    /**
     * runFilter checks !isFilterDisabled() and shouldFilter(). The run() method is invoked if both are true.
     *
     * @return the return from ZuulFilterResult
     */
    public ZuulFilterResult runFilter() {
        ZuulFilterResult zr = new ZuulFilterResult();
        // 如果被禁用则不会触发真正地调用
        if (!isFilterDisabled()) {
            // shouldFilter() 由各filter决定，返回true时执行filter
            if (shouldFilter()) {
                Tracer t = TracerFactory.instance().startMicroTracer("ZUUL::" + this.getClass().getSimpleName());
                try {
                    Object res = run();
                    zr = new ZuulFilterResult(res, ExecutionStatus.SUCCESS);
                } catch (Throwable e) {
                    t.setName("ZUUL::" + this.getClass().getSimpleName() + " failed");
                    zr = new ZuulFilterResult(ExecutionStatus.FAILED);
                    zr.setException(e);
                } finally {
                    t.stopAndLog();
                }
            } else {
                // 打上跳过标识
                zr = new ZuulFilterResult(ExecutionStatus.SKIPPED);
            }
        }
        return zr;
    }
    // run样例： org.springframework.cloud.netflix.zuul.filters.pre.ServletDetectionFilter#run
    @Override
    public Object run() {
        RequestContext ctx = RequestContext.getCurrentContext();
        HttpServletRequest request = ctx.getRequest();
        if (!(request instanceof HttpServletRequestWrapper) 
                && isDispatcherServletRequest(request)) {
            ctx.set(IS_DISPATCHER_SERVLET_REQUEST_KEY, true);
        } else {
            ctx.set(IS_DISPATCHER_SERVLET_REQUEST_KEY, false);
        }

        return null;
    }

　　如上，就是一个preFilter的处理流程了：

　　　　1. 从 FilterLoader 中获取所有 pre 类型的filter;
　　　　2. 依次调用各filter的runFilter()方法，触发filter;
　　　　3. 调用前先调用 shouldFilter() 进行判断该filter对于此次请求是否有用, 各filter实现可以从上下文中取得相应的信息，各自判定;
　　　　4. 计数器加1;
　　　　5. 默认就会有多个filter可调用, 不够满足业务场景再自行添加，各业务执行方法为 run();

 

2.3. 正常路由处理

　　zuul 的本职工作，是对路径的转发路由（正向代理 or 反向代理），如下处理：

    // com.netflix.zuul.http.ZuulServlet#route
    /**
     * executes "route" filters
     *
     * @throws ZuulException
     */
    void route() throws ZuulException {
        zuulRunner.route();
    }
    // com.netflix.zuul.ZuulRunner#route
    /**
     * executes "route" filterType  ZuulFilters
     *
     * @throws ZuulException
     */
    public void route() throws ZuulException {
        FilterProcessor.getInstance().route();
    }
    // com.netflix.zuul.FilterProcessor#route
    /**
     * Runs all "route" filters. These filters route calls to an origin.
     *
     * @throws ZuulException if an exception occurs.
     */
    public void route() throws ZuulException {
        try {
            // 同样，获取filter类型为 route 的 filters, 进行调用处理即可
            // - org.springframework.cloud.netflix.zuul.filters.route.RibbonRoutingFilter
            // - org.springframework.cloud.netflix.zuul.filters.route.SimpleHostRoutingFilter
            // - org.springframework.cloud.netflix.zuul.filters.route.SendForwardFilter
            runFilters("route");
        } catch (ZuulException e) {
            throw e;
        } catch (Throwable e) {
            throw new ZuulException(e, 500, "UNCAUGHT_EXCEPTION_IN_ROUTE_FILTER_" + e.getClass().getName());
        }
    }
    // 其中，Ribbon 的处理需要有 ribbon 组件的引入和配置
    // org.springframework.cloud.netflix.zuul.filters.route.RibbonRoutingFilter#shouldFilter
    @Override
    public boolean shouldFilter() {
        RequestContext ctx = RequestContext.getCurrentContext();
        // 判断是否有 serviceId, 且 sendZuulResponse=true 才会进行 ribbon 处理
        return (ctx.getRouteHost() == null && ctx.get(SERVICE_ID_KEY) != null
                && ctx.sendZuulResponse());
    }
    
    以下是普通路由转发的实现,只要配置了相应的路由信息，则会进行相关转发：
    // org.springframework.cloud.netflix.zuul.filters.route.SimpleHostRoutingFilter#shouldFilter
    @Override
    public boolean shouldFilter() {
        return RequestContext.getCurrentContext().getRouteHost() != null
                && RequestContext.getCurrentContext().sendZuulResponse();
    }

    @Override
    public Object run() {
        RequestContext context = RequestContext.getCurrentContext();
        // step1. 构建http请求头信息
        HttpServletRequest request = context.getRequest();
        MultiValueMap headers = this.helper
                .buildZuulRequestHeaders(request);
        // step2. 构建 params 信息, 如: a=111&&b=222
        MultiValueMap params = this.helper
                .buildZuulRequestQueryParams(request);
        // 获取请求类型, GET,POST,PUT,DELETE
        String verb = getVerb(request);
        // step3. 构建请求体信息，如文件
        InputStream requestEntity = getRequestBody(request);
        // 如果没有 Content-Length 字段，则设置 chunkedRequestBody:true
        if (getContentLength(request) < 0) {
            context.setChunkedRequestBody();
        }
        // step4. 构建要转发的uri地址信息
        String uri = this.helper.buildZuulRequestURI(request);
        this.helper.addIgnoredHeaders();

        try {
            // step5. 请求转发出去，等待响应
            // 具体如何转发请求，是在 forward 中处理的
            CloseableHttpResponse response = forward(this.httpClient, verb, uri, request,
                    headers, params, requestEntity);
            // 将结果放到上下文中，以备后续filter处理
            setResponse(response);
        }
        catch (Exception ex) {
            throw new ZuulRuntimeException(ex);
        }
        return null;
    }

    // step1. 构建http请求头信息
    // org.springframework.cloud.netflix.zuul.filters.ProxyRequestHelper#buildZuulRequestHeaders
    public MultiValueMap buildZuulRequestHeaders(
            HttpServletRequest request) {
        RequestContext context = RequestContext.getCurrentContext();
        MultiValueMap headers = new HttpHeaders();
        Enumeration headerNames = request.getHeaderNames();
        // 获取所有的 header 信息，还原到 headers 中
        if (headerNames != null) {
            while (headerNames.hasMoreElements()) {
                String name = headerNames.nextElement();
                // 排除一些特别的的头信息
                if (isIncludedHeader(name)) {
                    Enumeration values = request.getHeaders(name);
                    while (values.hasMoreElements()) {
                        String value = values.nextElement();
                        headers.add(name, value);
                    }
                }
            }
        }
        // 添加本次路由转发新增的头信息
        Map zuulRequestHeaders = context.getZuulRequestHeaders();
        for (String header : zuulRequestHeaders.keySet()) {
            headers.set(header, zuulRequestHeaders.get(header));
        }
        headers.set(HttpHeaders.ACCEPT_ENCODING, "gzip");
        return headers;
    }
    
    // step2. 构建 params 信息, 如: a=111&&b=222
    // org.springframework.cloud.netflix.zuul.filters.ProxyRequestHelper#buildZuulRequestQueryParams
    public MultiValueMap buildZuulRequestQueryParams(
            HttpServletRequest request) {
        // 解析 getQueryString 中的 a=111&b=222... 信息
        Map> map = HTTPRequestUtils.getInstance().getQueryParams();
        MultiValueMap params = new LinkedMultiValueMap<>();
        if (map == null) {
            return params;
        }
        for (String key : map.keySet()) {
            for (String value : map.get(key)) {
                params.add(key, value);
            }
        }
        return params;
    }
    // 解析请求url中的k=v&k2=v2 为 map 格式
    // com.netflix.zuul.util.HTTPRequestUtils#getQueryParams
    /**
     * returns query params as a Map with String keys and Lists of Strings as values
     * @return
     */
    public Map> getQueryParams() {

        Map> qp = RequestContext.getCurrentContext().getRequestQueryParams();
        if (qp != null) return qp;

        HttpServletRequest request = RequestContext.getCurrentContext().getRequest();

        qp = new LinkedHashMap>();

        if (request.getQueryString() == null) return null;
        StringTokenizer st = new StringTokenizer(request.getQueryString(), "&");
        int i;

        while (st.hasMoreTokens()) {
            String s = st.nextToken();
            i = s.indexOf("=");
            if (i > 0 && s.length() >= i + 1) {
                String name = s.substring(0, i);
                String value = s.substring(i + 1);

                try {
                    name = URLDecoder.decode(name, "UTF-8");
                } catch (Exception e) {
                }
                try {
                    value = URLDecoder.decode(value, "UTF-8");
                } catch (Exception e) {
                }

                List valueList = qp.get(name);
                if (valueList == null) {
                    valueList = new LinkedList();
                    qp.put(name, valueList);
                }

                valueList.add(value);
            }
            else if (i == -1)
            {
                String name=s;
                String value="";
                try {
                    name = URLDecoder.decode(name, "UTF-8");
                } catch (Exception e) {
                }
               
                List valueList = qp.get(name);
                if (valueList == null) {
                    valueList = new LinkedList();
                    qp.put(name, valueList);
                }

                valueList.add(value);
                
            }
        }

        RequestContext.getCurrentContext().setRequestQueryParams(qp);
        return qp;
    }

    
    // step3. 构建请求体信息，如文件
    // org.springframework.cloud.netflix.zuul.filters.route.SimpleHostRoutingFilter#getRequestBody
    protected InputStream getRequestBody(HttpServletRequest request) {
        InputStream requestEntity = null;
        try {
            // 先向 requestEntity 中获取输入流，如果没有则向 servlet 中获取
            requestEntity = (InputStream) RequestContext.getCurrentContext().get(REQUEST_ENTITY_KEY);
            if (requestEntity == null) {
                // 向 HttpServletRequest 中获取原始的输入流
                requestEntity = request.getInputStream();
            }
        }
        catch (IOException ex) {
            log.error("error during getRequestBody", ex);
        }
        return requestEntity;
    }
    
    
    // step4. 构建要转发的uri地址信息
    // org.springframework.cloud.netflix.zuul.filters.ProxyRequestHelper#buildZuulRequestURI
    public String buildZuulRequestURI(HttpServletRequest request) {
        RequestContext context = RequestContext.getCurrentContext();
        // 原始请求 uri
        String uri = request.getRequestURI();
        // 路由转换之后的请求 uri
        String contextURI = (String) context.get(REQUEST_URI_KEY);
        if (contextURI != null) {
            try {
                // 防止乱码，urlencode 一下
                uri = UriUtils.encodePath(contextURI, characterEncoding(request));
            }
            catch (Exception e) {
                log.debug(
                        "unable to encode uri path from context, falling back to uri from request",
                        e);
            }
        }
        return uri;
    }
    
    // step5. 请求转发出去，等待响应
    // 具体如何转发请求，是在 forward 中处理的
    // org.springframework.cloud.netflix.zuul.filters.route.SimpleHostRoutingFilter#forward
    private CloseableHttpResponse forward(CloseableHttpClient httpclient, String verb,
            String uri, HttpServletRequest request, MultiValueMap headers,
            MultiValueMap params, InputStream requestEntity)
            throws Exception {
        Map info = this.helper.debug(verb, uri, headers, params,
                requestEntity);
        // 配置的路由地址前缀
        URL host = RequestContext.getCurrentContext().getRouteHost();
        HttpHost httpHost = getHttpHost(host);
        // 取出uri
        uri = StringUtils.cleanPath((host.getPath() + uri).replaceAll("/{2,}", "/"));
        long contentLength = getContentLength(request);

        ContentType contentType = null;

        if (request.getContentType() != null) {
            contentType = ContentType.parse(request.getContentType());
        }
        // 使用InputStreamEntity封装inputStream请求，该inputStream是从socket接入后的原始输入流
        // 后续 httpclient 进行数据读取时，将由其进行提供相应读数据方法
        InputStreamEntity entity = new InputStreamEntity(requestEntity, contentLength,
                contentType);
        // 构建本次要请求的数据，关键
        HttpRequest httpRequest = buildHttpRequest(verb, uri, entity, headers, params,
                request);
        try {
            log.debug(httpHost.getHostName() + " " + httpHost.getPort() + " "
                    + httpHost.getSchemeName());
            // 提交给 httpclient 组件执行 http 请求，并返回结果
            CloseableHttpResponse zuulResponse = forwardRequest(httpclient, httpHost,
                    httpRequest);
            this.helper.appendDebug(info, zuulResponse.getStatusLine().getStatusCode(),
                    revertHeaders(zuulResponse.getAllHeaders()));
            return zuulResponse;
        }
        finally {
            // When HttpClient instance is no longer needed,
            // shut down the connection manager to ensure
            // immediate deallocation of all system resources
            // httpclient.getConnectionManager().shutdown();
        }
    }
    // org.springframework.cloud.netflix.zuul.filters.route.SimpleHostRoutingFilter#buildHttpRequest
    protected HttpRequest buildHttpRequest(String verb, String uri,
            InputStreamEntity entity, MultiValueMap headers,
            MultiValueMap params, HttpServletRequest request) {
        HttpRequest httpRequest;
        String uriWithQueryString = uri + (this.forceOriginalQueryStringEncoding
                ? getEncodedQueryString(request) : this.helper.getQueryString(params));
        // 根据原始请求的不同类型，做相应类型的转发
        // 以下请求处理，都包含了对 文件流一类请求的逻辑
        switch (verb.toUpperCase()) {
        case "POST":
            HttpPost httpPost = new HttpPost(uriWithQueryString);
            httpRequest = httpPost;
            httpPost.setEntity(entity);
            break;
        case "PUT":
            HttpPut httpPut = new HttpPut(uriWithQueryString);
            httpRequest = httpPut;
            httpPut.setEntity(entity);
            break;
        case "PATCH":
            HttpPatch httpPatch = new HttpPatch(uriWithQueryString);
            httpRequest = httpPatch;
            httpPatch.setEntity(entity);
            break;
        case "DELETE":
            BasicHttpEntityEnclosingRequest entityRequest = new BasicHttpEntityEnclosingRequest(
                    verb, uriWithQueryString);
            httpRequest = entityRequest;
            // DELETE 时会做两步操作
            entityRequest.setEntity(entity);
            break;
        default:
            // 除以上几种情况，都使用 BasicHttpRequest 进行处理即可
            httpRequest = new BasicHttpRequest(verb, uriWithQueryString);
            log.debug(uriWithQueryString);
        }
        // 统一都设置请求头，将map转换为 BasicHeader
        httpRequest.setHeaders(convertHeaders(headers));
        return httpRequest;
    }
    // org.springframework.cloud.netflix.zuul.filters.route.SimpleHostRoutingFilter#forwardRequest
    private CloseableHttpResponse forwardRequest(CloseableHttpClient httpclient,
            HttpHost httpHost, HttpRequest httpRequest) throws IOException {
        return httpclient.execute(httpHost, httpRequest);
    }

　　可见整个真正的转发流程，主要分几步：

　　　　1. 解析http请求头信息，并添加自己部分的头信息;
　　　　2. 解析并保留请求参数信息, 如: a=111&&b=222;
　　　　3. 获取原始的inputStream信息，如文件;
　　　　4. 根据路由配置，构建要转发的uri地址信息;
　　　　5. 使用httpclient组件，将请求转发出去，并等待响应，设置到 response中;

　　实际上，真正的转发仍然是依次做好相应判断，然后还原成对应的请求，再转发后后端服务中。

　　以上，就是一个普通的服务转发实现了。并没有太多的技巧，而是最基础的步骤：接收请求，解析参数，重新构建请求，请求后端，获得结果。

 

2.4. 后置过滤器

　　后置处理器可以做一些请求完服务端之后，对客户端的响应数据，包括正常数据流的输出，错误信息的返回等。如 SendResponseFilter, SendErrorFilter...

    // com.netflix.zuul.http.ZuulServlet#postRoute
    /**
     * executes "post" ZuulFilters
     *
     * @throws ZuulException
     */
    void postRoute() throws ZuulException {
        zuulRunner.postRoute();
    }
    
    // com.netflix.zuul.ZuulRunner#postRoute
    /**
     * executes "post" filterType  ZuulFilters
     *
     * @throws ZuulException
     */
    public void postRoute() throws ZuulException {
        FilterProcessor.getInstance().postRoute();
    }

    // com.netflix.zuul.FilterProcessor#postRoute
    /**
     * runs "post" filters which are called after "route" filters. ZuulExceptions from ZuulFilters are thrown.
     * Any other Throwables are caught and a ZuulException is thrown out with a 500 status code
     *
     * @throws ZuulException
     */
    public void postRoute() throws ZuulException {
        try {
            // 获取类型为 post 的 filter, 调用
            // 默认为: org.springframework.cloud.netflix.zuul.filters.post.SendResponseFilter
            runFilters("post");
        } catch (ZuulException e) {
            throw e;
        } catch (Throwable e) {
            throw new ZuulException(e, 500, "UNCAUGHT_EXCEPTION_IN_POST_FILTER_" + e.getClass().getName());
        }
    }
    // org.springframework.cloud.netflix.zuul.filters.post.SendResponseFilter#shouldFilter
    @Override
    public boolean shouldFilter() {
        // 有响应的数据，就可以进行处理
        RequestContext context = RequestContext.getCurrentContext();
        return context.getThrowable() == null
                && (!context.getZuulResponseHeaders().isEmpty()
                    || context.getResponseDataStream() != null
                    || context.getResponseBody() != null);
    }
    // org.springframework.cloud.netflix.zuul.filters.post.SendResponseFilter#run
    @Override
    public Object run() {
        try {
            // 添加header信息
            addResponseHeaders();
            // 输出数据流到请求端
            writeResponse();
        }
        catch (Exception ex) {
            ReflectionUtils.rethrowRuntimeException(ex);
        }
        return null;
    }
    // org.springframework.cloud.netflix.zuul.filters.post.SendResponseFilter#addResponseHeaders
    private void addResponseHeaders() {
        RequestContext context = RequestContext.getCurrentContext();
        HttpServletResponse servletResponse = context.getResponse();
        if (this.zuulProperties.isIncludeDebugHeader()) {
            @SuppressWarnings("unchecked")
            List rd = (List) context.get(ROUTING_DEBUG_KEY);
            if (rd != null) {
                StringBuilder debugHeader = new StringBuilder();
                for (String it : rd) {
                    debugHeader.append("[[[" + it + "]]]");
                }
                servletResponse.addHeader(X_ZUUL_DEBUG_HEADER, debugHeader.toString());
            }
        }
        // 向 response 中添加header
        List> zuulResponseHeaders = context.getZuulResponseHeaders();
        if (zuulResponseHeaders != null) {
            for (Pair it : zuulResponseHeaders) {
                servletResponse.addHeader(it.first(), it.second());
            }
        }
        if (includeContentLengthHeader(context)) {
            Long contentLength = context.getOriginContentLength();
            if(useServlet31) {
                servletResponse.setContentLengthLong(contentLength);
            } else {
                //Try and set some kind of content length if we can safely convert the Long to an int
                if (isLongSafe(contentLength)) {
                    servletResponse.setContentLength(contentLength.intValue());
                }
            }
        }
    }
    // org.springframework.cloud.netflix.zuul.filters.post.SendResponseFilter#writeResponse()
    private void writeResponse() throws Exception {
        RequestContext context = RequestContext.getCurrentContext();
        // there is no body to send
        if (context.getResponseBody() == null
                && context.getResponseDataStream() == null) {
            return;
        }
        HttpServletResponse servletResponse = context.getResponse();
        if (servletResponse.getCharacterEncoding() == null) { // only set if not set
            servletResponse.setCharacterEncoding("UTF-8");
        }
        
        OutputStream outStream = servletResponse.getOutputStream();
        InputStream is = null;
        try {
            if (context.getResponseBody() != null) {
                String body = context.getResponseBody();
                is = new ByteArrayInputStream(
                                body.getBytes(servletResponse.getCharacterEncoding()));
            }
            else {
                is = context.getResponseDataStream();
                if (is!=null && context.getResponseGZipped()) {
                    // if origin response is gzipped, and client has not requested gzip,
                    // decompress stream before sending to client
                    // else, stream gzip directly to client
                    if (isGzipRequested(context)) {
                        servletResponse.setHeader(ZuulHeaders.CONTENT_ENCODING, "gzip");
                    }
                    else {
                        is = handleGzipStream(is);
                    }
                }
            }
            
            if (is!=null) {
                writeResponse(is, outStream);
            }
        }
        finally {
            /**
            * We must ensure that the InputStream provided by our upstream pooling mechanism is ALWAYS closed
            * even in the case of wrapped streams, which are supplied by pooled sources such as Apache's
            * PoolingHttpClientConnectionManager. In that particular case, the underlying HTTP connection will
            * be returned back to the connection pool iif either close() is explicitly called, a read
            * error occurs, or the end of the underlying stream is reached. If, however a write error occurs, we will
            * end up leaking a connection from the pool without an explicit close()
            *
            * @author Johannes Edmeier
            */
            if (is != null) {
                try {
                    is.close();
                }
                catch (Exception ex) {
                    log.warn("Error while closing upstream input stream", ex);
                }
            }

            try {
                Object zuulResponse = context.get("zuulResponse");
                if (zuulResponse instanceof Closeable) {
                    ((Closeable) zuulResponse).close();
                }
                outStream.flush();
                // The container will close the stream for us
            }
            catch (IOException ex) {
                log.warn("Error while sending response to client: " + ex.getMessage());
            }
        }
    }
    // org.springframework.cloud.netflix.zuul.filters.post.SendResponseFilter#writeResponse
    private void writeResponse(InputStream zin, OutputStream out) throws Exception {
        // 默认大小 8192
        byte[] bytes = buffers.get();
        int bytesRead = -1;
        // 依次向 outputStream 中写入字节流
        while ((bytesRead = zin.read(bytes)) != -1) {
            out.write(bytes, 0, bytesRead);
        }
    }

　　同样，对客户端的输出，就是这么简单：解析出header信息，将response write() 到客户端的socket中。即完成任务。

　　以上，我们主要看了几个非常普通的filter的处理过程，理解了下 zuul 的运行流程，当然主要的目的分析zuul是如何转发请求的。基本上上面所有的filter都会继承 ZuulFilter 的抽象，它提供两个重要的统一的方法：isFilterDisabled() 和 shouldFilter() 方法用于控制过虑器是否启用或者是否应该使用，并统一了返回结果。

 

 　　zuul 整体实现也是非常简单明了，基于模板方法模式 和 责任链模式 和 单例模式，基本搞定。只是更多的花需要应用自己去玩了。

 

三、自行实现一个业务filter

　　要想做到通用的框架，这点事情是必须要做的。当然，还必须要足够简单，如下：一个注解加一个继承实现即可！

// 一个注解，@Component, 成功 spring bean 组件
// 一个继承，ZuulFilter, 以使 zuul 框架可以按照规范进行filter 的接入
@Component
public class MyOneFilter extends ZuulFilter {

    private final UrlPathHelper urlPathHelper = new UrlPathHelper();

    @Autowired
    private ZuulProperties zuulProperties;

    @Autowired
    private RouteLocator routeLocator;

    public MyOneFilter() {
    }

    public MyOneFilter(ZuulProperties zuulProperties,
                       RouteLocator routeLocator) {
        this.routeLocator = routeLocator;
        this.zuulProperties = zuulProperties;
    }

    @Override
    public String filterType() {
        // 自定义过滤器的类型，知道为什么不用枚举类吗？嘿嘿
        return PRE_TYPE;
    }

    @Override
    public int filterOrder() {
        // 定义过滤器的出场顺序，越小越牛
        return 1;
    }

    @Override
    public boolean shouldFilter() {
        // 是否可以启用当前filter, 按你的业务规则来说了算
        return true;
    }

    @Override
    public Object run() {
        // 如果满足了过滤条件，你想怎么做都行，RequestContext中有你想要的一切
        RequestContext ctx = RequestContext.getCurrentContext();
        Route route = routeLocator.getMatchingRoute(
                urlPathHelper.getPathWithinApplication(ctx.getRequest()));
        System.out.println("in my one filter");
        return null;
    }

}

　　至于其他配置项什么的，自行查看官网即可！ https://www.springcloud.cc/spring-cloud-greenwich.html#_router_and_filter_zuul

 

四、几点思考

      zuul 既然作为cloud的通用网关，必然会承受着比其他应用更大的流量，同时也要担起着比其他应用更高的QOC。可谓责任重大！

      然而纵观前面的实现，并没有什么牛逼的技术。相反，看到更多是为了业务的需要，需要进行反复的数据拷贝。

      很显然，网关类的服务，是非常典型的IO密集型应用，但似乎并没有看到它在这方面的努力(默认web服务器是tomcat，这就是其上限，如果换成netty又当如何)。也许，它还得需要前置网关，负载均衡，流量分发，才能够发挥其应有的作用。（把它当作普通应用就没事了，虽然它也在做负载均衡流量分发）

      对于大文件的上传，它是通过先将文件流存储到本地临时文件，再上传后端服务中，这个过程必然会导致响应缓慢以及应对异常能力的变弱。  而且，普通网关请求 zuul 中对于会从servlet中获取输入流，并转化为byte数组，也就是说他会保持全量上传数据，这对于超大文件来说，肯定是不可取的。所以，官网上也特别说你得如何小心处理大文件的上传！