写个日志请求切面，先后端甩锅更方便

时间 2020-03-21

标签日志请求切面后端方便繁體版

原文原文链接

最近项目进入联调阶段，服务层的接口须要和协议层进行交互，协议层须要将入参[json字符串]组装成服务层所需的json字符串，组装的过程当中很容易出错。入参出错致使接口调试失败问题在联调中出现不少次，所以就想写一个请求日志切面把入参信息打印一下，同时协议层调用服务层接口名称对不上也出现了几回，经过请求日志切面就能够知道上层是否有没有发起调用，方便先后端甩锅还能拿出证据

写在前面

本篇文章是实战性的，对于切面的原理不会讲解，只会简单介绍一下切面的知识点java

切面介绍

面向切面编程是一种编程范式，它做为OOP面向对象编程的一种补充，用于处理系统中分布于各个模块的横切关注点，好比事务管理、权限控制、缓存控制、日志打印等等。
AOP把软件的功能模块分为两个部分：核心关注点和横切关注点。业务处理的主要功能为核心关注点，而非核心、须要拓展的功能为横切关注点。AOP的做用在于分离系统中的各类关注点，将核心关注点和横切关注点进行分离，使用切面有如下好处：编程

集中处理某一关注点/横切逻辑
能够很方便的添加/删除关注点
侵入性少，加强代码可读性及可维护性

所以当想打印请求日志时很容易想到切面，对控制层代码0侵入json

切面的使用【基于注解】

@Aspect => 声明该类为一个注解类

切点注解：后端

@Pointcut => 定义一个切点，能够简化代码

通知注解：缓存

@Before => 在切点以前执行代码
@After => 在切点以后执行代码
@AfterReturning => 切点返回内容后执行代码，能够对切点的返回值进行封装
@AfterThrowing => 切点抛出异常后执行
@Around => 环绕，在切点先后执行代码

动手写一个请求日志切面

使用@Pointcut定义切点并发
```
@Pointcut("execution(* your_package.controller..*(..))")
public void requestServer() {
}
```
@Pointcut定义了一个切点，由于是请求日志切边，所以切点定义的是Controller包下的全部类下的方法。定义切点之后在通知注解中直接使用requestServer方法名就能够了框架

使用@Before再切点前执行dom

@Before("requestServer()")
public void doBefore(JoinPoint joinPoint) {
    ServletRequestAttributes attributes = (ServletRequestAttributes) 
RequestContextHolder.getRequestAttributes();
    HttpServletRequest request = attributes.getRequest();

    LOGGER.info("===============================Start========================");
    LOGGER.info("IP                 : {}", request.getRemoteAddr());
    LOGGER.info("URL                : {}", request.getRequestURL().toString());
    LOGGER.info("HTTP Method        : {}", request.getMethod());
    LOGGER.info("Class Method       : {}.{}", joinPoint.getSignature().getDeclaringTypeName(), joinPoint.getSignature().getName());
}

在进入Controller方法前，打印出调用方IP、请求URL、HTTP请求类型、调用的方法名ide

使用@Around打印进入控制层的入参高并发

@Around("requestServer()")
public Object doAround(ProceedingJoinPoint proceedingJoinPoint) throws Throwable {
    long start = System.currentTimeMillis();
    Object result = proceedingJoinPoint.proceed();
    LOGGER.info("Request Params       : {}", getRequestParams(proceedingJoinPoint));
    LOGGER.info("Result               : {}", result);
    LOGGER.info("Time Cost            : {} ms", System.currentTimeMillis() - start);

    return result;
}

打印了入参、结果以及耗时

getRquestParams方法

private Map<String, Object> getRequestParams(ProceedingJoinPoint proceedingJoinPoint) {
     Map<String, Object> requestParams = new HashMap<>();

      //参数名
     String[] paramNames = ((MethodSignature)proceedingJoinPoint.getSignature()).getParameterNames();
     //参数值
     Object[] paramValues = proceedingJoinPoint.getArgs();

     for (int i = 0; i < paramNames.length; i++) {
         Object value = paramValues[i];

         //若是是文件对象
         if (value instanceof MultipartFile) {
             MultipartFile file = (MultipartFile) value;
             value = file.getOriginalFilename();  //获取文件名
         }

         requestParams.put(paramNames[i], value);
     }

     return requestParams;
 }

经过 @PathVariable以及@RequestParam注解传递的参数没法打印出参数名，所以须要手动拼接一下参数名，同时对文件对象进行了特殊处理，只需获取文件名便可

@After方法调用后执行

@After("requestServer()")
public void doAfter(JoinPoint joinPoint) {
    LOGGER.info("===============================End========================");
}

没有业务逻辑只是打印了End

完整切面代码

@Component
@Aspect
public class RequestLogAspect {
    private final static Logger LOGGER = LoggerFactory.getLogger(RequestLogAspect.class);

    @Pointcut("execution(* your_package.controller..*(..))")
    public void requestServer() {
    }

    @Before("requestServer()")
    public void doBefore(JoinPoint joinPoint) {
        ServletRequestAttributes attributes = (ServletRequestAttributes) 
RequestContextHolder.getRequestAttributes();
        HttpServletRequest request = attributes.getRequest();

        LOGGER.info("===============================Start========================");
        LOGGER.info("IP                 : {}", request.getRemoteAddr());
        LOGGER.info("URL                : {}", request.getRequestURL().toString());
        LOGGER.info("HTTP Method        : {}", request.getMethod());
        LOGGER.info("Class Method       : {}.{}", joinPoint.getSignature().getDeclaringTypeName(), 
 joinPoint.getSignature().getName());
    }

@Around("requestServer()")
    public Object doAround(ProceedingJoinPoint proceedingJoinPoint) throws Throwable {
        long start = System.currentTimeMillis();
        Object result = proceedingJoinPoint.proceed();
        LOGGER.info("Request Params     : {}", getRequestParams(proceedingJoinPoint));
        LOGGER.info("Result               : {}", result);
        LOGGER.info("Time Cost            : {} ms", System.currentTimeMillis() - start);

        return result;
    }

    @After("requestServer()")
    public void doAfter(JoinPoint joinPoint) {
        LOGGER.info("===============================End========================");
    }

    /**
     * 获取入参
     * @param proceedingJoinPoint
     *
     * @return
     * */
    private Map<String, Object> getRequestParams(ProceedingJoinPoint proceedingJoinPoint) {
        Map<String, Object> requestParams = new HashMap<>();

        //参数名
        String[] paramNames = 
((MethodSignature)proceedingJoinPoint.getSignature()).getParameterNames();
        //参数值
        Object[] paramValues = proceedingJoinPoint.getArgs();

        for (int i = 0; i < paramNames.length; i++) {
            Object value = paramValues[i];

            //若是是文件对象
            if (value instanceof MultipartFile) {
                MultipartFile file = (MultipartFile) value;
                value = file.getOriginalFilename();  //获取文件名
            }

            requestParams.put(paramNames[i], value);
        }

        return requestParams;
    }
}
```

高并发下请求日志切面

写完之后对本身的代码很满意，可是想着可能还有完善的地方就和朋友交流了一下。emmmm

果真还有继续优化的地方
每一个信息都打印一行，在高并发请求下确实会出现请求之间打印日志串行的问题，由于测试阶段请求数量较少没有出现串行的状况，果真生产环境才是第一发展力，可以遇到更多bug，写更健壮的代码
解决日志串行的问题只要将多行打印信息合并为一行就能够了，所以构造一个对象

RequestInfo.java

@Data
public class RequestInfo {
    private String ip;
    private String url;
    private String httpMethod;
    private String classMethod;
    private Object requestParams;
    private Object result;
    private Long timeCost;
}

环绕通知方法体

@Around("requestServer()")
public Object doAround(ProceedingJoinPoint proceedingJoinPoint) throws Throwable {
    long start = System.currentTimeMillis();
    ServletRequestAttributes attributes = (ServletRequestAttributes) RequestContextHolder.getRequestAttributes();
    HttpServletRequest request = attributes.getRequest();
    Object result = proceedingJoinPoint.proceed();
    RequestInfo requestInfo = new RequestInfo();
            requestInfo.setIp(request.getRemoteAddr());
    requestInfo.setUrl(request.getRequestURL().toString());
    requestInfo.setHttpMethod(request.getMethod());
    requestInfo.setClassMethod(String.format("%s.%s", proceedingJoinPoint.getSignature().getDeclaringTypeName(),
            proceedingJoinPoint.getSignature().getName()));
    requestInfo.setRequestParams(getRequestParamsByProceedingJoinPoint(proceedingJoinPoint));
    requestInfo.setResult(result);
    requestInfo.setTimeCost(System.currentTimeMillis() - start);
    LOGGER.info("Request Info      : {}", JSON.toJSONString(requestInfo));

    return result;
}

将url、http request这些信息组装成RequestInfo对象，再序列化打印对象<br/> 打印序列化对象结果而不是直接打印对象是由于序列化有更直观、更清晰，同时能够借助在线解析工具对结果进行解析

是否是还不错
在解决高并发下请求串行问题的同时添加了对异常请求信息的打印，经过使用 @AfterThrowing注解对抛出异常的方法进行处理

RequestErrorInfo.java

@Data
public class RequestErrorInfo {
    private String ip;
    private String url;
    private String httpMethod;
    private String classMethod;
    private Object requestParams;
    private RuntimeException exception;
}

异常通知环绕体

@AfterThrowing(pointcut = "requestServer()", throwing = "e")
public void doAfterThrow(JoinPoint joinPoint, RuntimeException e) {
    ServletRequestAttributes attributes = (ServletRequestAttributes) RequestContextHolder.getRequestAttributes();
    HttpServletRequest request = attributes.getRequest();
    RequestErrorInfo requestErrorInfo = new RequestErrorInfo();
    requestErrorInfo.setIp(request.getRemoteAddr());
    requestErrorInfo.setUrl(request.getRequestURL().toString());
    requestErrorInfo.setHttpMethod(request.getMethod());
    requestErrorInfo.setClassMethod(String.format("%s.%s", joinPoint.getSignature().getDeclaringTypeName(),
            joinPoint.getSignature().getName()));
    requestErrorInfo.setRequestParams(getRequestParamsByJoinPoint(joinPoint));
    requestErrorInfo.setException(e);
    LOGGER.info("Error Request Info      : {}", JSON.toJSONString(requestErrorInfo));
}

对于异常，耗时是没有意义的，所以不统计耗时，而是添加了异常的打印

最后放一下完整日志请求切面代码：

@Component
@Aspect
public class RequestLogAspect {
    private final static Logger LOGGER = LoggerFactory.getLogger(RequestLogAspect.class);

    @Pointcut("execution(* your_package.controller..*(..))")
    public void requestServer() {
    }

    @Around("requestServer()")
    public Object doAround(ProceedingJoinPoint proceedingJoinPoint) throws Throwable {
        long start = System.currentTimeMillis();
        ServletRequestAttributes attributes = (ServletRequestAttributes) RequestContextHolder.getRequestAttributes();
        HttpServletRequest request = attributes.getRequest();
        Object result = proceedingJoinPoint.proceed();
        RequestInfo requestInfo = new RequestInfo();
                requestInfo.setIp(request.getRemoteAddr());
        requestInfo.setUrl(request.getRequestURL().toString());
        requestInfo.setHttpMethod(request.getMethod());
        requestInfo.setClassMethod(String.format("%s.%s", proceedingJoinPoint.getSignature().getDeclaringTypeName(),
                proceedingJoinPoint.getSignature().getName()));
        requestInfo.setRequestParams(getRequestParamsByProceedingJoinPoint(proceedingJoinPoint));
        requestInfo.setResult(result);
        requestInfo.setTimeCost(System.currentTimeMillis() - start);
        LOGGER.info("Request Info      : {}", JSON.toJSONString(requestInfo));

        return result;
    }


    @AfterThrowing(pointcut = "requestServer()", throwing = "e")
    public void doAfterThrow(JoinPoint joinPoint, RuntimeException e) {
        ServletRequestAttributes attributes = (ServletRequestAttributes) RequestContextHolder.getRequestAttributes();
        HttpServletRequest request = attributes.getRequest();
        RequestErrorInfo requestErrorInfo = new RequestErrorInfo();
        requestErrorInfo.setIp(request.getRemoteAddr());
        requestErrorInfo.setUrl(request.getRequestURL().toString());
        requestErrorInfo.setHttpMethod(request.getMethod());
        requestErrorInfo.setClassMethod(String.format("%s.%s", joinPoint.getSignature().getDeclaringTypeName(),
                joinPoint.getSignature().getName()));
        requestErrorInfo.setRequestParams(getRequestParamsByJoinPoint(joinPoint));
        requestErrorInfo.setException(e);
        LOGGER.info("Error Request Info      : {}", JSON.toJSONString(requestErrorInfo));
    }

    /**
     * 获取入参
     * @param proceedingJoinPoint
     *
     * @return
     * */
    private Map<String, Object> getRequestParamsByProceedingJoinPoint(ProceedingJoinPoint proceedingJoinPoint) {
        //参数名
        String[] paramNames = ((MethodSignature)proceedingJoinPoint.getSignature()).getParameterNames();
        //参数值
        Object[] paramValues = proceedingJoinPoint.getArgs();

        return buildRequestParam(paramNames, paramValues);
    }

    private Map<String, Object> getRequestParamsByJoinPoint(JoinPoint joinPoint) {
        //参数名
        String[] paramNames = ((MethodSignature)joinPoint.getSignature()).getParameterNames();
        //参数值
        Object[] paramValues = joinPoint.getArgs();

        return buildRequestParam(paramNames, paramValues);
    }

    private Map<String, Object> buildRequestParam(String[] paramNames, Object[] paramValues) {
        Map<String, Object> requestParams = new HashMap<>();
        for (int i = 0; i < paramNames.length; i++) {
            Object value = paramValues[i];

            //若是是文件对象
            if (value instanceof MultipartFile) {
                MultipartFile file = (MultipartFile) value;
                value = file.getOriginalFilename();  //获取文件名
            }

            requestParams.put(paramNames[i], value);
        }

        return requestParams;
    }

    @Data
    public class RequestInfo {
        private String ip;
        private String url;
        private String httpMethod;
        private String classMethod;
        private Object requestParams;
        private Object result;
        private Long timeCost;
    }

    @Data
    public class RequestErrorInfo {
        private String ip;
        private String url;
        private String httpMethod;
        private String classMethod;
        private Object requestParams;
        private RuntimeException exception;
    }
}

赶忙给大家的应用加上吧【若是没加的话】，没有日志的话，总怀疑上层出错，可是却拿不出证据

==================== 如下内容更新于2019/3/14 ==============

关于traceId 跟踪定位【地藏Kelvin的评论】，能够根据traceId跟踪整条调用链，以log4j2为例介绍如何加入traceId

添加拦截器

public class LogInterceptor implements HandlerInterceptor {
    private final static String TRACE_ID = "traceId";

    @Override
    public boolean preHandle(HttpServletRequest request, HttpServletResponse response, Object handler) throws Exception {
        String traceId = java.util.UUID.randomUUID().toString().replaceAll("-", "").toUpperCase();
        ThreadContext.put("traceId", traceId);

        return true;
    }

    @Override
    public void postHandle(HttpServletRequest request, HttpServletResponse response, Object handler, ModelAndView modelAndView)
            throws Exception {
    }

    @Override
    public void afterCompletion(HttpServletRequest request, HttpServletResponse response, Object handler, Exception ex)
            throws Exception {        
        ThreadContext. remove(TRACE_ID);
    }
}

在调用前经过ThreadContext加入traceId，调用完成后移除

修改日志配置文件

在原来的日志格式中
添加traceId的占位符

<property name="pattern">[TRACEID:%X{traceId}] %d{HH:mm:ss.SSS} %-5level %class{-1}.%M()/%L - %msg%xEx%n</property>

执行效果

日志跟踪更方便

DMC是配置logback和log4j使用的，使用方式和ThreadContext差很少，将ThreadContext.put替换为MDC.put便可，同时修改日志配置文件。
推荐使用log4j2，为何推荐使用log4j2能够看下这篇文章：日志框架，选择Logback Or Log4j2？

==================== 如下内容更新于2019/3/16 ==============

log4j2也是能够配合MDC一块儿使用的

MDC是slf4j包下的，其具体使用哪一个日志框架与咱们的依赖有关