在前面寫的一篇文章中，熱心網(wǎng)友【地藏Kelvin】評(píng)論說在多線程中還是有可能會(huì)亂掉夺姑，建議通過MDC打印traceId來個(gè)全鏈路調(diào)用跟蹤。掘金里個(gè)個(gè)都是人才掌猛，說話又好聽盏浙，超喜歡在里面。掘金使我進(jìn)步荔茬，熱心網(wǎng)友總能提出改進(jìn)意見
寫在前面
通過本文將了解到什么是MDC废膘、MDC應(yīng)用中存在的問題、如何解決存在的問題
MDC介紹
簡介：
MDC（Mapped Diagnostic Context慕蔚，映射調(diào)試上下文）是 log4j 丐黄、logback及l(fā)og4j2 提供的一種方便在多線程條件下記錄日志的功能。MDC 可以看成是一個(gè)與當(dāng)前線程綁定的哈希表坊萝，可以往其中添加鍵值對(duì)孵稽。MDC 中包含的內(nèi)容可以被同一線程中執(zhí)行的代碼所訪問。當(dāng)前線程的子線程會(huì)繼承其父線程中的 MDC 的內(nèi)容十偶。當(dāng)需要記錄日志時(shí)菩鲜，只需要從 MDC 中獲取所需的信息即可。MDC 的內(nèi)容則由程序在適當(dāng)?shù)臅r(shí)候保存進(jìn)去惦积。對(duì)于一個(gè) Web 應(yīng)用來說接校，通常是在請求被處理的最開始保存這些數(shù)據(jù)
API說明：
- clear() => 移除所有MDC
- get (String key) => 獲取當(dāng)前線程MDC中指定key的值
- getContext() => 獲取當(dāng)前線程MDC的MDC
- put(String key, Object o) => 往當(dāng)前線程的MDC中存入指定的鍵值對(duì)
- remove(String key) => 刪除當(dāng)前線程MDC中指定的鍵值對(duì)
優(yōu)點(diǎn)：
- 代碼簡潔，日志風(fēng)格統(tǒng)一，不需要在log打印中手動(dòng)拼寫traceId蛛勉，即LOGGER.info("traceId:{} ", traceId)
暫時(shí)只能想到這一點(diǎn)
MDC使用
-
添加攔截器
public class LogInterceptor implements HandlerInterceptor { @Override public boolean preHandle(HttpServletRequest request, HttpServletResponse response, Object handler) throws Exception { //如果有上層調(diào)用就用上層的ID String traceId = request.getHeader(Constants.TRACE_ID); if (traceId == null) { traceId = TraceIdUtil.getTraceId(); } MDC.put(Constants.TRACE_ID, 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 { //調(diào)用結(jié)束后刪除 MDC.remove(Constants.TRACE_ID); } } -
修改日志格式
<property name="pattern">[TRACEID:%X{traceId}] %d{HH:mm:ss.SSS} %-5level %class{-1}.%M()/%L - %msg%xEx%n</property>重點(diǎn)是%X{traceId}鹿寻，traceId和MDC中的鍵名稱一致
簡單使用就這么容易，但是在有些情況下traceId將獲取不到
MDC 存在的問題
- 子線程中打印日志丟失traceId
- HTTP調(diào)用丟失traceId
解決MDC存在的問題
子線程日志打印丟失traceId
子線程在打印日志的過程中traceId將丟失诽凌，解決方式為重寫線程池毡熏，對(duì)于直接new創(chuàng)建線程的情況不考略【實(shí)際應(yīng)用中應(yīng)該避免這種用法】，重寫線程池?zé)o非是對(duì)任務(wù)進(jìn)行一次封裝
-
線程池封裝類：ThreadPoolExecutorMdcWrapper.java
public class ThreadPoolExecutorMdcWrapper extends ThreadPoolExecutor { public ThreadPoolExecutorMdcWrapper(int corePoolSize, int maximumPoolSize, long keepAliveTime, TimeUnit unit, BlockingQueue<Runnable> workQueue) { super(corePoolSize, maximumPoolSize, keepAliveTime, unit, workQueue); } public ThreadPoolExecutorMdcWrapper(int corePoolSize, int maximumPoolSize, long keepAliveTime, TimeUnit unit, BlockingQueue<Runnable> workQueue, ThreadFactory threadFactory) { super(corePoolSize, maximumPoolSize, keepAliveTime, unit, workQueue, threadFactory); } public ThreadPoolExecutorMdcWrapper(int corePoolSize, int maximumPoolSize, long keepAliveTime, TimeUnit unit, BlockingQueue<Runnable> workQueue, RejectedExecutionHandler handler) { super(corePoolSize, maximumPoolSize, keepAliveTime, unit, workQueue, handler); } public ThreadPoolExecutorMdcWrapper(int corePoolSize, int maximumPoolSize, long keepAliveTime, TimeUnit unit, BlockingQueue<Runnable> workQueue, ThreadFactory threadFactory, RejectedExecutionHandler handler) { super(corePoolSize, maximumPoolSize, keepAliveTime, unit, workQueue, threadFactory, handler); } @Override public void execute(Runnable task) { super.execute(ThreadMdcUtil.wrap(task, MDC.getCopyOfContextMap())); } @Override public <T> Future<T> submit(Runnable task, T result) { return super.submit(ThreadMdcUtil.wrap(task, MDC.getCopyOfContextMap()), result); } @Override public <T> Future<T> submit(Callable<T> task) { return super.submit(ThreadMdcUtil.wrap(task, MDC.getCopyOfContextMap())); } @Override public Future<?> submit(Runnable task) { return super.submit(ThreadMdcUtil.wrap(task, MDC.getCopyOfContextMap())); } }說明：
- 繼承ThreadPoolExecutor類侣诵，重新執(zhí)行任務(wù)的方法
- 通過ThreadMdcUtil對(duì)任務(wù)進(jìn)行一次包裝
-
線程traceId封裝工具類：ThreadMdcUtil.java
public class ThreadMdcUtil { public static void setTraceIdIfAbsent() { if (MDC.get(Constants.TRACE_ID) == null) { MDC.put(Constants.TRACE_ID, TraceIdUtil.getTraceId()); } } public static <T> Callable<T> wrap(final Callable<T> callable, final Map<String, String> context) { return () -> { if (context == null) { MDC.clear(); } else { MDC.setContextMap(context); } setTraceIdIfAbsent(); try { return callable.call(); } finally { MDC.clear(); } }; } public static Runnable wrap(final Runnable runnable, final Map<String, String> context) { return () -> { if (context == null) { MDC.clear(); } else { MDC.setContextMap(context); } setTraceIdIfAbsent(); try { runnable.run(); } finally { MDC.clear(); } }; } }說明【以封裝Runnable為例】：
- 判斷當(dāng)前線程對(duì)應(yīng)MDC的Map是否存在痢法，存在則設(shè)置
- 設(shè)置MDC中的traceId值，不存在則新生成杜顺，針對(duì)不是子線程的情況财搁，如果是子線程，MDC中traceId不為null
- 執(zhí)行run方法
代碼等同于以下寫法躬络，會(huì)更直觀
public static Runnable wrap(final Runnable runnable, final Map<String, String> context) { return new Runnable() { @Override public void run() { if (context == null) { MDC.clear(); } else { MDC.setContextMap(context); } setTraceIdIfAbsent(); try { runnable.run(); } finally { MDC.clear(); } } }; }重新返回的是包裝后的Runnable尖奔，在該任務(wù)執(zhí)行之前【runnable.run()】先將主線程的Map設(shè)置到當(dāng)前線程中【 即MDC.setContextMap(context)】，這樣子線程和主線程MDC對(duì)應(yīng)的Map就是一樣的了
HTTP調(diào)用丟失traceId
在使用HTTP調(diào)用第三方服務(wù)接口時(shí)traceId將丟失穷当，需要對(duì)HTTP調(diào)用工具進(jìn)行改造提茁，在發(fā)送時(shí)在request header中添加traceId，在下層被調(diào)用方添加攔截器獲取header中的traceId添加到MDC中
HTTP調(diào)用有多種方式膘滨，比較常見的有HttpClient甘凭、OKHttp、RestTemplate火邓，所以只給出這幾種HTTP調(diào)用的解決方式
HttpClient：
- 實(shí)現(xiàn)HttpClient攔截器
public class HttpClientTraceIdInterceptor implements HttpRequestInterceptor {
@Override
public void process(HttpRequest httpRequest, HttpContext httpContext) throws HttpException, IOException {
String traceId = MDC.get(Constants.TRACE_ID);
//當(dāng)前線程調(diào)用中有traceId丹弱，則將該traceId進(jìn)行透傳
if (traceId != null) {
//添加請求體
httpRequest.addHeader(Constants.TRACE_ID, traceId);
}
}
}
? 實(shí)現(xiàn)HttpRequestInterceptor接口并重寫process方法
? 如果調(diào)用線程中含有traceId，則需要將獲取到的traceId通過request中的header向下透傳下去
-
為HttpClient添加攔截器
private static CloseableHttpClient httpClient = HttpClientBuilder.create() .addInterceptorFirst(new HttpClientTraceIdInterceptor()) .build();通過addInterceptorFirst方法為HttpClient添加攔截器
OKHttp：
-
實(shí)現(xiàn)OKHttp攔截器
public class OkHttpTraceIdInterceptor implements Interceptor { @Override public Response intercept(Chain chain) throws IOException { String traceId = MDC.get(Constants.TRACE_ID); Request request = null; if (traceId != null) { //添加請求體 request = chain.request().newBuilder().addHeader(Constants.TRACE_ID, traceId).build(); } Response originResponse = chain.proceed(request); return originResponse; } }實(shí)現(xiàn)Interceptor攔截器铲咨，重寫interceptor方法躲胳，實(shí)現(xiàn)邏輯和HttpClient差不多，如果能夠獲取到當(dāng)前線程的traceId則向下透傳
-
為OkHttp添加攔截器
private static OkHttpClient client = new OkHttpClient.Builder() .addNetworkInterceptor(new OkHttpTraceIdInterceptor()) .build();調(diào)用addNetworkInterceptor方法添加攔截器
RestTemplate：
-
實(shí)現(xiàn)RestTemplate攔截器
public class RestTemplateTraceIdInterceptor implements ClientHttpRequestInterceptor { @Override public ClientHttpResponse intercept(HttpRequest httpRequest, byte[] bytes, ClientHttpRequestExecution clientHttpRequestExecution) throws IOException { String traceId = MDC.get(Constants.TRACE_ID); if (traceId != null) { httpRequest.getHeaders().add(Constants.TRACE_ID, traceId); } return clientHttpRequestExecution.execute(httpRequest, bytes); } }實(shí)現(xiàn)ClientHttpRequestInterceptor接口纤勒，并重寫intercept方法坯苹，其余邏輯都是一樣的不重復(fù)說明
-
為RestTemplate添加攔截器
restTemplate.setInterceptors(Arrays.asList(new RestTemplateTraceIdInterceptor()));調(diào)用setInterceptors方法添加攔截器
第三方服務(wù)攔截器：
HTTP調(diào)用第三方服務(wù)接口全流程traceId需要第三方服務(wù)配合，第三方服務(wù)需要添加攔截器拿到request header中的traceId并添加到MDC中
public class LogInterceptor implements HandlerInterceptor {
@Override
public boolean preHandle(HttpServletRequest request, HttpServletResponse response, Object handler) throws Exception {
//如果有上層調(diào)用就用上層的ID
String traceId = request.getHeader(Constants.TRACE_ID);
if (traceId == null) {
traceId = TraceIdUtils.getTraceId();
}
MDC.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 {
MDC.remove(Constants.TRACE_ID);
}
}
說明：
- 先從request header中獲取traceId
- 從request header中獲取不到traceId則說明不是第三方調(diào)用摇天，直接生成一個(gè)新的traceId
- 將生成的traceId存入MDC中
除了需要添加攔截器之外粹湃，還需要在日志格式中添加traceId的打印，如下：
<property name="pattern">[TRACEID:%X{traceId}] %d{HH:mm:ss.SSS} %-5level %class{-1}.%M()/%L - %msg%xEx%n</property>
需要添加%X{traceId}
最后附：項(xiàng)目代碼泉坐，歡迎fork與star为鳄，漲點(diǎn)小星星，卑微乞討
往期文章推薦
1.寫個(gè)日志請求切面腕让，前后端甩鍋更方便
2.為什么阿里巴巴要禁用Executors創(chuàng)建線程池孤钦？
參考文章：
1.在Java項(xiàng)目中使用traceId跟蹤請求全流程日志
2.MDC介紹 -- 一種多線程下日志管理實(shí)踐方式
