6、spring cloud zuul使用

作者: KissGoodby | 来源:发表于2018-11-19 18:20 被阅读0次

使用zuul生成关联traceID

这里我们使用zuul的过虑器，完成一个trace日志的功能，创建一个traceID，关联整个链路，打印在日志中。

从一个请求的开始和结束，这整个链路traceID唯一，这在生产开发中也是很常见的功能，不仅仅可以将整个链路的日志关联起来，方便排查问题，还为后期日志的收集奠定基础。

接下来，我们进入代码环节

首先搭建eureka-server，zuul-server，consumer-server，provider-server服务，这在之前都说过，这里不再详细说明，简单把controller代码和配置贴一下
provider-server的controller，仅仅提供了一个服务

@RestController
public class ProviderController {

    private final Logger LOGGER = Logger.getLogger(ProviderController.class);

    /**
     * get方式接口
     * @param request 请求参数
     */
    @RequestMapping(value = "/provider", method = RequestMethod.GET)
    public String provider(@RequestParam String request) {
        LOGGER.info("========================================");
        LOGGER.info("provider service ");
        LOGGER.info("========================================");
        return "provider, " + request;
    }

}

consumer-server的controller

@RestController
public class Controller {

    private final Logger LOGGER = Logger.getLogger(Controller.class);

    @Autowired
    private RestTemplate restTemplate;

    /**
     * GET请求传参数1
     */
    @RequestMapping("/consumer/v1")
    public String consumerV1() {
        return restTemplate.getForEntity("http://zuul-server/api/provider?request={1}", String.class, "test").getBody();
    }

    /**
     * GET请求传参数2
     */
    @RequestMapping("/consumer/v2")
    public String consumerV2() {
        Map<String, String> map = Maps.newHashMap();
        map.put("request", "test");
        return restTemplate.getForEntity("http://provider-server/provider?request={request}", String.class, map).getBody();
    }
}

zuul-server的application配置

#整合eureka
eureka:
  instance:
    prefer-ip-address: true #注册服务的IP
  client:
    serviceUrl:
      defaultZone: http://localhost:8761/eureka
    register-with-eureka: true
    fetch-registry: true


zuul:
  routes:
    consumer-server:  #手动定义路由映射
      path: /a/**
      url: consumer-server
    provider-server:    #手动定义路由映射
      path: /api/**
      url: provider-server

现在分别启动服务，测试调用没有问题，路由信息如下：

{
    "/a/**": "consumer-server",
    "/api/**": "provider-server",
    "/consumer-server/**": "consumer-server",
    "/provider-server/**": "provider-server"
}

这样基础工作完成。

来创建保存traceID的上下文，这里我新创建了一个module，将下面代码保存到新的jar包里面。

/**
 * trace model
 *
 * @author hui.wang
 * @since 19 November 2018
 */
public class Trace {

    private String traceId;
    private Date createTime;

    public String getTraceId() {
        return traceId;
    }

    public void setTraceId(String traceId) {
        this.traceId = traceId;
    }

    public Date getCreateTime() {
        return createTime;
    }

    public void setCreateTime(Date createTime) {
        this.createTime = createTime;
    }
}

/**
 * @author hui.wang
 * @since 19 November 2018
 */
public class TraceContext {

    private static final ThreadLocal<Trace> TRACE_CONTEXT = new ThreadLocal<Trace>();

    public static Trace getTraceContext() {
        return TRACE_CONTEXT.get();
    }

    public static void setTraceContext(Trace traceContext) {
        Assert.notNull(traceContext, "Only non-null traceContext instances are permitted");
        TRACE_CONTEXT.set(traceContext);
    }

    public static void clean() {
        TRACE_CONTEXT.remove();
    }
}

上面创建了trace model和trace context使用threadLocal保存上下文中的trace

接下来我们编写zuul的filter，zuul的filter分为前置过虑器，后置过虑器和路由过虑器，type类型如下：

pre：在请求被路由之前调用。
routing：在路由请求时候被调用。
post：在routing和error过滤器之后被调用。
error：处理请求时发生错误时被调用。
首先写了前置过虑器TrackingFilter.java

/**
 * zuul 前置过虑器
 * 设置关联ID
 *
 * @author hui.wang
 * @since 19 November 2018
 */
@Component
public class TrackingFilter extends ZuulFilter{

    private final Logger LOGGER = Logger.getLogger(TrackingFilter.class);

    private static final int FILTER_ORDER = 1;
    private static final boolean SHOULD_FILTER = true;
    private static final String FILTER_TYPE = "pre";

    /**
     * filter类型，前置过虑器，后置过虑器和路由过虑器
     */
    @Override
    public String filterType() {
        return FILTER_TYPE;
    }

    /**
     * 返回一个整数值，表示filter执行顺序
     */
    @Override
    public int filterOrder() {
        return FILTER_ORDER;
    }

    /**
     * 返回一个boolean，表示该过虑器是否执行
     */
    @Override
    public boolean shouldFilter() {
        return SHOULD_FILTER;
    }

    /**
     * 每次filter执行的代码
     */
    @Override
    public Object run() {
        if (StringUtils.isEmpty(FilterUtils.getTraceId())) {
            FilterUtils.setTraceId();
        }

        RequestContext requestContext = RequestContext.getCurrentContext();
        String URL = requestContext.getRequest().getRequestURL().toString();
        String traceId = FilterUtils.getTraceId();
        LOGGER.info("======================================");
        LOGGER.info("request url = " + URL + ", traceId = " + traceId);
        LOGGER.info("======================================");
        return null;
    }
}

代码也很简单，先判断header里面有没有trace_id，如果有，不做其他操作，只是打印日志，如果没有生成一个trace_id，放到header里面。这里需要说明一下trace_id是保存到HTTP header里面进行传递的。

这里使用到了FilterUtils工具类，这里我也贴一下代码

public class FilterUtils {

    public static final String TRACE_ID = "trace_id";

    /**
     * 获取随机ID
     */
    public static String generateTraceId() {
        return java.util.UUID.randomUUID().toString();
    }

    /**
     * 获取traceId
     */
    public static String getTraceId() {
        RequestContext requestContext = RequestContext.getCurrentContext();

        if (StringUtils.isNotEmpty(requestContext.getRequest().getHeader(TRACE_ID))) {
            return requestContext.getRequest().getHeader(TRACE_ID);
        }
        return requestContext.getZuulRequestHeaders().get(TRACE_ID);
    }

    /**
     * 设置traceId
     */
    public static void setTraceId() {
        RequestContext requestContext = RequestContext.getCurrentContext();
        requestContext.addZuulRequestHeader(TRACE_ID, generateTraceId());
    }
}

这里已经把前置zuul过滤器已经编写完成，有这个过滤器是不够的，还需要在请求完成后，将trace_id设置到response header里面进行传递。编写后置zuul 过滤器

public class ResponseFilter extends ZuulFilter {

    private static final Logger LOGGER = LoggerFactory.getLogger(ResponseFilter.class);

    private static final int FILTER_ORDER = 1;
    private static final boolean SHOULD_FILTER = true;
    private static final String FILTER_TYPE = "post";

    @Override
    public String filterType() {
        return FILTER_TYPE;
    }

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

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

    @Override
    public Object run() {
        RequestContext requestContext = RequestContext.getCurrentContext();
        String URL = requestContext.getRequest().getRequestURL().toString();
        LOGGER.info("======================================");
        LOGGER.info("response url " + URL + "traceId = " + FilterUtils.getTraceId());
        LOGGER.info("======================================");
        requestContext.getResponse().addHeader(FilterUtils.TRACE_ID, FilterUtils.getTraceId());
        return null;
    }
}

这里代码也是很简单的，将上下文中的trace_id设置到response header里面

zuul filter编写完成后，只是保证trace_id可以在网关路由阶段可以传递，但当请求打到了具体的应用，如果在应用内不作处理，这个trace_id在后面的链路就丢了，因此还要编写HTTP servlet过虑器，传递trace_id

public class TraceFilter implements Filter{

    private static final Logger LOGGER = LoggerFactory.getLogger(TraceFilter.class);

    @Override
    public void init(FilterConfig filterConfig) throws ServletException {

    }

    @Override
    public void doFilter(ServletRequest servletRequest, ServletResponse servletResponse, FilterChain filterChain) throws IOException, ServletException {
        try {
            try {
                HttpServletRequest httpServletRequest = (HttpServletRequest) servletRequest;
                if (StringUtils.isNotEmpty(httpServletRequest.getHeader(FilterUtils.TRACE_ID))) {
                    Trace trace = new Trace();
                    trace.setTraceId(httpServletRequest.getHeader(FilterUtils.TRACE_ID));
                    trace.setCreateTime(new Date());
                    TraceContext.setTraceContext(trace);
                    LOGGER.info("filter set trace success");
                }
            } catch (Exception e) {
                LOGGER.error("filter set trace error", e);
            }
            filterChain.doFilter(servletRequest, servletResponse);
        } finally {
            TraceContext.clean();
        }
    }

    @Override
    public void destroy() {

    }
}

代码也很简单，就是判断header里面有没有trace_id，如果有的话，将trace_id保存到上下文中

上面完成后，只是保证从路由到应用，trace_id不会丢失，但是如果consumer-server调用provider-server的服务时候，如果trace_id没有被传播，此时调用到provider-server的服务时候，trace_id就丢失了，因此要编写spring Interceptor然后将Interceptor整合到RestTemplate中

public class TraceInterceptor implements ClientHttpRequestInterceptor{

    @Override
    public ClientHttpResponse intercept(HttpRequest httpRequest, byte[] bytes, ClientHttpRequestExecution clientHttpRequestExecution) throws IOException {
        if (TraceContext.getTraceContext() != null && StringUtils.isNotEmpty(TraceContext.getTraceContext().getTraceId())) {
            HttpHeaders httpHeaders = httpRequest.getHeaders();
            httpHeaders.set(FilterUtils.TRACE_ID, TraceContext.getTraceContext().getTraceId());
        }
        return clientHttpRequestExecution.execute(httpRequest, bytes);
    }
}

代码也很简单，在上下文中取trace_id然后设置到请求的header里面。

配置filter和Interceptor
在zuul-server的启动类上配置zuul filter

@SpringBootApplication
@EnableZuulProxy
public class ZuulAppllication {

    @Bean
    public TrackingFilter trackingFilter() {
        return new TrackingFilter();
    }

    @Bean
    public ResponseFilter responseFilter() {
        return new ResponseFilter();
    }

    public static void main(String[] args) {
        SpringApplication.run(ZuulAppllication.class, args);
    }
}

在consumer-server的启动类上配置filter和Interceptor

@SpringBootApplication
@EnableDiscoveryClient
public class ConsumerApplication {

    @LoadBalanced
    @Bean
    RestTemplate restTemplate() {
        RestTemplate template = new RestTemplate();
        List interceptors = template.getInterceptors();
        if (Objects.isNull(interceptors)) {
            template.setInterceptors(Lists.newArrayList(new TraceInterceptor()));
        } else {
            interceptors.add(new TraceInterceptor());
            template.setInterceptors(interceptors);
        }
        return template;
    }

    @Bean
    public TraceFilter traceFilter() {
        return new TraceFilter();
    }

    public static void main(String[] args) {
        SpringApplication.run(ConsumerApplication.class, args);
    }
}

在provider-server启动类上配置filter

@SpringBootApplication
@EnableDiscoveryClient
public class ProviderApplicatioin {

    @Bean
    public TraceFilter traceFilter() {
        return new TraceFilter();
    }

    public static void main(String[] args) {
        SpringApplication.run(ProviderApplicatioin.class, args);
    }
}

这样代码基本完成，简单测试一下，访问一下http://localhost:8822/a/consumer/v1
zuul-server的日志：

2018-11-19 18:18:57.670  INFO 46324 --- [nio-8822-exec-9] c.h.w.s.c.t.a.zuulFilter.TrackingFilter  : ======================================
2018-11-19 18:18:57.670  INFO 46324 --- [nio-8822-exec-9] c.h.w.s.c.t.a.zuulFilter.TrackingFilter  : request url = http://localhost:8822/a/consumer/v1, traceId = 4d25f02d-0a7a-4002-8694-32493f7daa1b
2018-11-19 18:18:57.670  INFO 46324 --- [nio-8822-exec-9] c.h.w.s.c.t.a.zuulFilter.TrackingFilter  : ======================================
2018-11-19 18:18:57.683  INFO 46324 --- [io-8822-exec-10] c.h.w.s.c.t.a.zuulFilter.TrackingFilter  : ======================================
2018-11-19 18:18:57.683  INFO 46324 --- [io-8822-exec-10] c.h.w.s.c.t.a.zuulFilter.TrackingFilter  : request url = http://10.1.32.187:8822/api/provider, traceId = 4d25f02d-0a7a-4002-8694-32493f7daa1b
2018-11-19 18:18:57.683  INFO 46324 --- [io-8822-exec-10] c.h.w.s.c.t.a.zuulFilter.TrackingFilter  : ======================================
2018-11-19 18:18:57.700  INFO 46324 --- [io-8822-exec-10] c.h.w.s.c.t.a.zuulFilter.ResponseFilter  : ======================================
2018-11-19 18:18:57.701  INFO 46324 --- [io-8822-exec-10] c.h.w.s.c.t.a.zuulFilter.ResponseFilter  : response url http://10.1.32.187:8822/api/providertraceId = 4d25f02d-0a7a-4002-8694-32493f7daa1b
2018-11-19 18:18:57.701  INFO 46324 --- [io-8822-exec-10] c.h.w.s.c.t.a.zuulFilter.ResponseFilter  : ======================================
2018-11-19 18:18:57.705  INFO 46324 --- [nio-8822-exec-9] c.h.w.s.c.t.a.zuulFilter.ResponseFilter  : ======================================
2018-11-19 18:18:57.705  INFO 46324 --- [nio-8822-exec-9] c.h.w.s.c.t.a.zuulFilter.ResponseFilter  : response url http://localhost:8822/a/consumer/v1traceId = 4d25f02d-0a7a-4002-8694-32493f7daa1b
2018-11-19 18:18:57.705  INFO 46324 --- [nio-8822-exec-9] c.h.w.s.c.t.a.zuulFilter.ResponseFilter  : ======================================

从zuul-server的日志可以看出整个链路请求过程
consumer-server端的日志为：

2018-11-19 18:18:57.676  INFO 46320 --- [nio-8111-exec-3] c.h.w.s.c.t.a.servletFIlter.TraceFilter  : filter set trace success
2018-11-19 18:18:57.679  INFO 46320 --- [nio-8111-exec-3] c.h.w.s.cloud.controller.Controller      : =================================
2018-11-19 18:18:57.679  INFO 46320 --- [nio-8111-exec-3] c.h.w.s.cloud.controller.Controller      : traceId = 4d25f02d-0a7a-4002-8694-32493f7daa1b
2018-11-19 18:18:57.679  INFO 46320 --- [nio-8111-exec-3] c.h.w.s.cloud.controller.Controller      : =================================

provider-server端的日志为：

2018-11-19 18:18:57.690  INFO 46315 --- [nio-8081-exec-3] c.h.w.s.c.t.a.servletFIlter.TraceFilter  : filter set trace success
2018-11-19 18:18:57.698  INFO 46315 --- [nio-8081-exec-3] c.h.w.s.c.controller.ProviderController  : ========================================
2018-11-19 18:18:57.698  INFO 46315 --- [nio-8081-exec-3] c.h.w.s.c.controller.ProviderController  : provider service 被调用,traceId = 4d25f02d-0a7a-4002-8694-32493f7daa1b
2018-11-19 18:18:57.698  INFO 46315 --- [nio-8081-exec-3] c.h.w.s.c.controller.ProviderController  : ========================================

6、spring cloud zuul使用

使用zuul生成关联traceID

相关文章

网友评论

延伸阅读

深度阅读

栏目导航

热点阅读