这篇文章主要讲解了“如何实现SpringCloud Gateway请求响应日志”，文中的讲解内容简单清晰，易于学习与理解，下面请大家跟着小编的思路慢慢深入，一起来研究和学习“如何实现SpringCloud Gateway请求响应日志”吧！

专注于为中小企业提供成都网站设计、网站制作服务,电脑端+手机端+微信端的三站合一,更高效的管理,为中小企业宁武免费做网站提供优质的服务。我们立足成都，凝聚了一批互联网行业人才，有力地推动了上1000+企业的稳健成长，帮助中小企业通过网站建设实现规模扩充和转变。

获取输入输出参数

首先我们先定义一个日志体

@Data public class GatewayLog {     /**访问实例*/     private String targetServer;     /**请求路径*/     private String requestPath;     /**请求方法*/     private String requestMethod;     /**协议 */     private String schema;     /**请求体*/     private String requestBody;     /**响应体*/     private String responseData;     /**请求ip*/     private String ip;  /**请求时间*/     private Date requestTime;  /**响应时间*/     private Date responseTime;     /**执行时间*/     private long executeTime; }

【关键】在网关定义日志过滤器，获取输入输出参数

/**  * 日志过滤器，用于记录日志  * @author jianzh6  * @date 2020/3/24 17:17  */ @Slf4j @Component public class AccessLogFilter implements GlobalFilter, Ordered {     @Autowired     private AccessLogService accessLogService;      private final List> messageReaders = HandlerStrategies.withDefaults().messageReaders();      @Override     public int getOrder() {         return -100;     }      @Override     @SuppressWarnings("unchecked")     public Mono filter(ServerWebExchange exchange, GatewayFilterChain chain) {          ServerHttpRequest request = exchange.getRequest();          // 请求路径         String requestPath = request.getPath().pathWithinApplication().value();          Route route = getGatewayRoute(exchange);           String ipAddress = WebUtils.getServerHttpRequestIpAddress(request);          GatewayLog gatewayLog = new GatewayLog();         gatewayLog.setSchema(request.getURI().getScheme());         gatewayLog.setRequestMethod(request.getMethodValue());         gatewayLog.setRequestPath(requestPath);         gatewayLog.setTargetServer(route.getId());         gatewayLog.setRequestTime(new Date());         gatewayLog.setIp(ipAddress);          MediaType mediaType = request.getHeaders().getContentType();          if(MediaType.APPLICATION_FORM_URLENCODED.isCompatibleWith(mediaType) || MediaType.APPLICATION_JSON.isCompatibleWith(mediaType)){             return writeBodyLog(exchange, chain, gatewayLog);         }else{             return writeBasicLog(exchange, chain, gatewayLog);         }     }      private Mono writeBasicLog(ServerWebExchange exchange, GatewayFilterChain chain, GatewayLog accessLog) {         StringBuilder builder = new StringBuilder();         MultiValueMap queryParams = exchange.getRequest().getQueryParams();         for (Map.Entry> entry : queryParams.entrySet()) {             builder.append(entry.getKey()).append("=").append(StringUtils.join(entry.getValue(), ","));         }         accessLog.setRequestBody(builder.toString());          //获取响应体         ServerHttpResponseDecorator decoratedResponse = recordResponseLog(exchange, accessLog);          return chain.filter(exchange.mutate().response(decoratedResponse).build())                 .then(Mono.fromRunnable(() -> {                     // 打印日志                     writeAccessLog(accessLog);                 }));     }       /**      * 解决 request body 只能读取一次问题，      * 参考: org.springframework.cloud.gateway.filter.factory.rewrite.ModifyRequestBodyGatewayFilterFactory      * @param exchange      * @param chain      * @param gatewayLog      * @return      */     @SuppressWarnings("unchecked")     private Mono writeBodyLog(ServerWebExchange exchange, GatewayFilterChain chain, GatewayLog gatewayLog) {         ServerRequest serverRequest = ServerRequest.create(exchange,messageReaders);          Mono modifiedBody = serverRequest.bodyToMono(String.class)                 .flatMap(body ->{                     gatewayLog.setRequestBody(body);                     return Mono.just(body);                 });          // 通过 BodyInserter 插入 body(支持修改body), 避免 request body 只能获取一次         BodyInserter bodyInserter = BodyInserters.fromPublisher(modifiedBody, String.class);         HttpHeaders headers = new HttpHeaders();         headers.putAll(exchange.getRequest().getHeaders());         // the new content type will be computed by bodyInserter         // and then set in the request decorator         headers.remove(HttpHeaders.CONTENT_LENGTH);          CachedBodyOutputMessage outputMessage = new CachedBodyOutputMessage(exchange, headers);          return bodyInserter.insert(outputMessage,new BodyInserterContext())                 .then(Mono.defer(() -> {                     // 重新封装请求                     ServerHttpRequest decoratedRequest = requestDecorate(exchange, headers, outputMessage);                      // 记录响应日志                     ServerHttpResponseDecorator decoratedResponse = recordResponseLog(exchange, gatewayLog);                      // 记录普通的                     return chain.filter(exchange.mutate().request(decoratedRequest).response(decoratedResponse).build())                             .then(Mono.fromRunnable(() -> {                                 // 打印日志                                 writeAccessLog(gatewayLog);                             }));                 }));     }      /**      * 打印日志      * @author javadaily      * @date 2021/3/24 14:53      * @param gatewayLog 网关日志      */     private void writeAccessLog(GatewayLog gatewayLog) {         log.info(gatewayLog.toString());       }        private Route getGatewayRoute(ServerWebExchange exchange) {         return exchange.getAttribute(ServerWebExchangeUtils.GATEWAY_ROUTE_ATTR);     }       /**      * 请求装饰器，重新计算 headers      * @param exchange      * @param headers      * @param outputMessage      * @return      */     private ServerHttpRequestDecorator requestDecorate(ServerWebExchange exchange, HttpHeaders headers,                                                        CachedBodyOutputMessage outputMessage) {         return new ServerHttpRequestDecorator(exchange.getRequest()) {             @Override             public HttpHeaders getHeaders() {                 long contentLength = headers.getContentLength();                 HttpHeaders httpHeaders = new HttpHeaders();                 httpHeaders.putAll(super.getHeaders());                 if (contentLength > 0) {                     httpHeaders.setContentLength(contentLength);                 } else {                     // TODO: this causes a 'HTTP/1.1 411 Length Required' // on                     // httpbin.org                     httpHeaders.set(HttpHeaders.TRANSFER_ENCODING, "chunked");                 }                 return httpHeaders;             }              @Override             public Flux getBody() {                 return outputMessage.getBody();             }         };     }       /**      * 记录响应日志      * 通过 DataBufferFactory 解决响应体分段传输问题。      */     private ServerHttpResponseDecorator recordResponseLog(ServerWebExchange exchange, GatewayLog gatewayLog) {         ServerHttpResponse response = exchange.getResponse();         DataBufferFactory bufferFactory = response.bufferFactory();          return new ServerHttpResponseDecorator(response) {             @Override             public Mono writeWith(Publisher body) {                 if (body instanceof Flux) {                     Date responseTime = new Date();                     gatewayLog.setResponseTime(responseTime);                     // 计算执行时间                     long executeTime = (responseTime.getTime() - gatewayLog.getRequestTime().getTime());                      gatewayLog.setExecuteTime(executeTime);                      // 获取响应类型，如果是 json 就打印                     String originalResponseContentType = exchange.getAttribute(ServerWebExchangeUtils.ORIGINAL_RESPONSE_CONTENT_TYPE_ATTR);                       if (ObjectUtil.equal(this.getStatusCode(), HttpStatus.OK)                             && StringUtil.isNotBlank(originalResponseContentType)                             && originalResponseContentType.contains("application/json")) {                          Flux fluxBody = Flux.from(body);                         return super.writeWith(fluxBody.buffer().map(dataBuffers -> {                              // 合并多个流集合，解决返回体分段传输                             DataBufferFactory dataBufferFactory = new DefaultDataBufferFactory();                             DataBuffer join = dataBufferFactory.join(dataBuffers);                             byte[] content = new byte[join.readableByteCount()];                             join.read(content);                              // 释放掉内存                             DataBufferUtils.release(join);                             String responseResult = new String(content, StandardCharsets.UTF_8);                                gatewayLog.setResponseData(responseResult);                              return bufferFactory.wrap(content);                         }));                     }                 }                 // if body is not a flux. never got there.                 return super.writeWith(body);             }         };     } }

代码较长建议直接拷贝到编辑器，只要注意下面一个关键点：

getOrder()方法返回的值必须要<-1，「否则标准的NettyWriteResponseFilter将在您的过滤器被调用的机会之前发送响应，即不会执行获取后端响应参数的方法」

通过上面的两步我们已经可以获取到请求的输入输出参数了，在  writeAccessLog()中将其输出到了日志文件，大家可以在Postman发送请求观察日志。

存储日志

如果需要将日志持久化方便后期检索的话可以考虑将日志存储在MongoDB中，实现过程很简单。(安装MongoDB可以参考这篇文章：实战|MongoDB的安装配置)

• 引入MongoDB

     org.springframework.boot     spring-boot-starter-data-mongodb-reactive 

由于gateway是基于webflux，所以我们需要选择reactive版本。

• 在GatewayLog上添加对应的注解

@Data @Document public class GatewayLog {     @Id     private String id;  ... }

• 建立AccessLogRepository

@Repository public interface AccessLogRepository extends ReactiveMongoRepository {    }

• 建立Service

public interface AccessLogService {      /**      * 保存AccessLog      * @param gatewayLog 请求响应日志      * @return 响应日志      */     Mono saveAccessLog(GatewayLog gatewayLog);  }

• 建立实现类

@Service public class AccessLogServiceImpl implements AccessLogService {     @Autowired     private AccessLogRepository accessLogRepository;      @Override     public Mono saveAccessLog(GatewayLog gatewayLog) {         return accessLogRepository.insert(gatewayLog);     } }

• 在Nacos配置中心添加MongoDB对应配置

spring:   data:     mongodb:       host: xxx.xx.x.xx       port: 27017       database: accesslog       username: accesslog       password: xxxxxx

执行请求，打开MongoDB客户端，查看日志结果

感谢各位的阅读，以上就是“如何实现SpringCloud Gateway请求响应日志”的内容了，经过本文的学习后，相信大家对如何实现SpringCloud Gateway请求响应日志这一问题有了更深刻的体会，具体使用情况还需要大家实践验证。这里是创新互联，小编将为大家推送更多相关知识点的文章，欢迎关注！

标题名称：如何实现SpringCloudGateway请求响应日志
标题网址：http://scyanting.com/article/ijcjgg.html