1. 什么是Hystrix

?

?

Hystrix是Netflix的一個(gè)開源框架，地址如下：https://github.com/Netflix/Hystrix

?

中文名為“豪豬”，即平時(shí)很溫順，在感受到危險(xiǎn)的時(shí)候，用刺保護(hù)自己；在危險(xiǎn)過去后，還是一個(gè)溫順的肉球。

?

所以，整個(gè)框架的核心業(yè)務(wù)也就是這2點(diǎn)：

?

何時(shí)需要保護(hù)

如何保護(hù)

?

2. 何時(shí)需要保護(hù)

?

對于一個(gè)系統(tǒng)而言，它往往承擔(dān)著2層角色，服務(wù)提供者與服務(wù)消費(fèi)者。對于服務(wù)消費(fèi)者而言最大的痛苦就是如何“明哲保身”，做過網(wǎng)關(guān)項(xiàng)目的同學(xué)肯定感同身受

上面是一個(gè)常見的系統(tǒng)依賴關(guān)系，底層的依賴往往很多，通信協(xié)議包括 socket、HTTP、Dubbo、WebService等等。當(dāng)通信層發(fā)生網(wǎng)絡(luò)抖動(dòng)以及所依賴的系統(tǒng)發(fā)生業(yè)務(wù)響應(yīng)異常時(shí)，我們業(yè)務(wù)本身所提供的服務(wù)能力也直接會(huì)受到影響。

?

這種效果傳遞下去就很有可能造成雪崩效應(yīng)，即整個(gè)業(yè)務(wù)聯(lián)調(diào)發(fā)生異常，比如業(yè)務(wù)整體超時(shí)，或者訂單數(shù)據(jù)不一致。

?

那么核心問題就來了，如何檢測業(yè)務(wù)處于異常狀態(tài)？

?

成功率！成功率直接反映了業(yè)務(wù)的數(shù)據(jù)流轉(zhuǎn)狀態(tài)，是最直接的業(yè)務(wù)表現(xiàn)。

?

當(dāng)然，也可以根據(jù)超時(shí)時(shí)間做判斷，比如 Sentinel 的實(shí)現(xiàn)。其實(shí)這里概念上可以做一個(gè)轉(zhuǎn)化，用時(shí)間做超時(shí)控制，超時(shí)=失敗，這依然是一個(gè)成功率的概念。

?

3. 如何保護(hù)　

?

如同豪豬一樣，“刺”就是他的保護(hù)工具，所有的攻擊都會(huì)被刺無情的懟回去。

?

在 Hystrix 的實(shí)現(xiàn)中，這就出現(xiàn)了“熔斷器”的概念，即當(dāng)前的系統(tǒng)是否處于需要保護(hù)的狀態(tài)。

?

當(dāng)熔斷器處于開啟的狀態(tài)時(shí)，所有的請求都不會(huì)真正的走之前的業(yè)務(wù)邏輯，而是直接返回一個(gè)約定的信息，即 FallBack。通過這種快速失敗原則保護(hù)我們的系統(tǒng)。?

?

但是，系統(tǒng)不應(yīng)該永遠(yuǎn)處于“有刺”的狀態(tài)，當(dāng)危險(xiǎn)過后需要恢復(fù)正常。

?

于是對熔斷器的核心操作就是如下幾個(gè)功能：

?

如果成功率過低，就打開熔斷器，阻止正常業(yè)務(wù)

隨著時(shí)間的流動(dòng)，熔斷器處于半打開狀態(tài)，嘗試性放入一筆請求

　　熔斷器的核心 API 如下圖：?

?

?

4. 限流、熔斷、隔離、降級(jí)

?

這四個(gè)概念是我們談起微服務(wù)會(huì)經(jīng)常談到的概念，這里我們討論的是 Hystrix 的實(shí)現(xiàn)方式。

?

限流

?

• 這里的限流與 Guava 的 RateLimiter 的限流差異比較大，一個(gè)是為了“保護(hù)自我”，一個(gè)是“保護(hù)下游”

• 當(dāng)對服務(wù)進(jìn)行限流時(shí)，超過的流量將直接 Fallback，即熔斷。而 RateLimiter 關(guān)心的其實(shí)是“流量整形”，將不規(guī)整流量在一定速度內(nèi)規(guī)整

?

熔斷

?

• 當(dāng)我的應(yīng)用無法提供服務(wù)時(shí)，我要對上游請求熔斷，避免上游把我壓垮

• 當(dāng)我的下游依賴成功率過低時(shí)，我要對下游請求熔斷，避免下游把我拖垮

?

降級(jí)

?

• 降級(jí)與熔斷緊密相關(guān)，熔斷后業(yè)務(wù)如何表現(xiàn)，約定一個(gè)快速失敗的 Fallback，即為服務(wù)降級(jí)

?

隔離

?

• 業(yè)務(wù)之間不可互相影響，不同業(yè)務(wù)需要有獨(dú)立的運(yùn)行空間

• 最徹底的，可以采用物理隔離，不同的機(jī)器部

• 次之，采用進(jìn)程隔離，一個(gè)機(jī)器多個(gè) Tomcat

• 次之，請求隔離

• 由于 Hystrix 框架所屬的層級(jí)為代碼層，所以實(shí)現(xiàn)的是請求隔離，線程池或信號(hào)量

?

5. 源碼分析

?

?

先上一個(gè) Hystrix 的業(yè)務(wù)流程圖

?

?

可以看到 Hystrix 的請求都要經(jīng)過 HystrixCommand 的包裝，其核心邏輯在 AbstractComman.java 類中。

?

下面的源碼是基于 RxJava 的，看之前最好先了解下 RxJava 的常見用法與邏輯，否則看起來會(huì)很迷惑。

?

簡單的說，RxJava 就是基于回調(diào)的函數(shù)式編程。通俗的說，就等同于策略模式的匿名內(nèi)部類實(shí)現(xiàn)。

?

5.1 熔斷器

?

首先看信號(hào)量是如何影響我們請求的：

?

private Observable<R> applyHystrixSemantics(final AbstractCommand<R> _cmd) {// 自定義擴(kuò)展executionHook.onStart(_cmd);//判斷熔斷器是否允許請求過來if (circuitBreaker.attemptExecution()) {//獲得分組信號(hào)量，如果沒有采用信號(hào)量分組，返回默認(rèn)通過的信號(hào)量實(shí)現(xiàn)final TryableSemaphore executionSemaphore = getExecutionSemaphore();final AtomicBoolean semaphoreHasBeenReleased = new AtomicBoolean(false);//調(diào)用終止的回調(diào)函數(shù)final Action0 singleSemaphoreRelease = new Action0() {@Overridepublic void call() {if (semaphoreHasBeenReleased.compareAndSet(false, true)) {executionSemaphore.release();}}};//調(diào)用異常的回調(diào)函數(shù)final Action1<Throwable> markExceptionThrown = new Action1<Throwable>() {@Overridepublic void call(Throwable t) {eventNotifier.markEvent(HystrixEventType.EXCEPTION_THROWN, commandKey);}};//根據(jù)信號(hào)量嘗試競爭信號(hào)量if (executionSemaphore.tryAcquire()) {try {//競爭成功，注冊執(zhí)行參數(shù)executionResult = executionResult.setInvocationStartTime(System.currentTimeMillis());return executeCommandAndObserve(_cmd).doOnError(markExceptionThrown).doOnTerminate(singleSemaphoreRelease).doOnUnsubscribe(singleSemaphoreRelease);} catch (RuntimeException e) {return Observable.error(e);}} else {//競爭失敗，進(jìn)入fallbackreturn handleSemaphoreRejectionViaFallback();}} else {//熔斷器已打開，進(jìn)入fallbackreturn handleShortCircuitViaFallback();}}

?

什么時(shí)候熔斷器可以放請求進(jìn)來：

?

@Overridepublic boolean attemptExecution() {//動(dòng)態(tài)屬性判斷，熔斷器是否強(qiáng)制開著，如果強(qiáng)制開著，就不允許請求if (properties.circuitBreakerForceOpen().get()) {return false;}//如果強(qiáng)制關(guān)閉，就允許請求if (properties.circuitBreakerForceClosed().get()) {return true;}//如果當(dāng)前是關(guān)閉，就允許請求if (circuitOpened.get() == -1) {return true;} else {//如果當(dāng)前開著，就看是否已經(jīng)過了"滑動(dòng)窗口"，過了就可以請求，不過就不可以if (isAfterSleepWindow()) {//only the first request after sleep window should execute//if the executing command succeeds, the status will transition to CLOSED//if the executing command fails, the status will transition to OPEN//if the executing command gets unsubscribed, the status will transition to OPEN//這里使用CAS的方式，只有一個(gè)請求能過來，即"半關(guān)閉"狀態(tài)if (status.compareAndSet(Status.OPEN, Status.HALF_OPEN)) {return true;} else {return false;}} else {return false;}}}}

?

這里有個(gè)重要概念就是"滑動(dòng)窗口"：

?

?

private boolean isAfterSleepWindow() {final long circuitOpenTime = circuitOpened.get();final long currentTime = System.currentTimeMillis();final long sleepWindowTime = properties.circuitBreakerSleepWindowInMilliseconds().get();//滑動(dòng)窗口的判斷就是看看熔斷器打開的時(shí)間與現(xiàn)在相比是否超過了配置的滑動(dòng)窗口return currentTime > circuitOpenTime + sleepWindowTime;}

?

5.2 隔離

?

如果將業(yè)務(wù)請求進(jìn)行隔離？

?

private Observable<R> executeCommandWithSpecifiedIsolation(final AbstractCommand<R> _cmd) {//判斷隔離策略是什么，是線程池隔離還是信號(hào)量隔離　　　　if (properties.executionIsolationStrategy().get() == ExecutionIsolationStrategy.THREAD) {// mark that we are executing in a thread (even if we end up being rejected we still were a THREAD execution and not SEMAPHORE)//線程池隔離的運(yùn)行邏輯如下return Observable.defer(new Func0<Observable<R>>() {@Overridepublic Observable<R> call() {executionResult = executionResult.setExecutionOccurred();if (!commandState.compareAndSet(CommandState.OBSERVABLE_CHAIN_CREATED, CommandState.USER_CODE_EXECUTED)) {return Observable.error(new IllegalStateException("execution attempted while in state : " + commandState.get().name()));}//按照配置生成監(jiān)控?cái)?shù)據(jù)metrics.markCommandStart(commandKey, threadPoolKey, ExecutionIsolationStrategy.THREAD);if (isCommandTimedOut.get() == TimedOutStatus.TIMED_OUT) {// the command timed out in the wrapping thread so we will return immediately// and not increment any of the counters below or other such logicreturn Observable.error(new RuntimeException("timed out before executing run()"));}if (threadState.compareAndSet(ThreadState.NOT_USING_THREAD, ThreadState.STARTED)) {//we have not been unsubscribed, so should proceedHystrixCounters.incrementGlobalConcurrentThreads();threadPool.markThreadExecution();// store the command that is being runendCurrentThreadExecutingCommand = Hystrix.startCurrentThreadExecutingCommand(getCommandKey());executionResult = executionResult.setExecutedInThread();/*** If any of these hooks throw an exception, then it appears as if the actual execution threw an error*/try {//執(zhí)行擴(kuò)展點(diǎn)邏輯executionHook.onThreadStart(_cmd);executionHook.onRunStart(_cmd);executionHook.onExecutionStart(_cmd);return getUserExecutionObservable(_cmd);} catch (Throwable ex) {return Observable.error(ex);}} else {//command has already been unsubscribed, so return immediatelyreturn Observable.empty();}}//注冊各種場景的回調(diào)函數(shù)}).doOnTerminate(new Action0() {@Overridepublic void call() {if (threadState.compareAndSet(ThreadState.STARTED, ThreadState.TERMINAL)) {handleThreadEnd(_cmd);}if (threadState.compareAndSet(ThreadState.NOT_USING_THREAD, ThreadState.TERMINAL)) {//if it was never started and received terminal, then no need to clean up (I don't think this is possible)}//if it was unsubscribed, then other cleanup handled it}}).doOnUnsubscribe(new Action0() {@Overridepublic void call() {if (threadState.compareAndSet(ThreadState.STARTED, ThreadState.UNSUBSCRIBED)) {handleThreadEnd(_cmd);}if (threadState.compareAndSet(ThreadState.NOT_USING_THREAD, ThreadState.UNSUBSCRIBED)) {//if it was never started and was cancelled, then no need to clean up}//if it was terminal, then other cleanup handled it}//將邏輯放在線程池的調(diào)度器上執(zhí)行，即將上述邏輯放入線程池中}).subscribeOn(threadPool.getScheduler(new Func0<Boolean>() {@Overridepublic Boolean call() {return properties.executionIsolationThreadInterruptOnTimeout().get() && _cmd.isCommandTimedOut.get() == TimedOutStatus.TIMED_OUT;}}));} else {//走到這里就是信號(hào)量隔離，在當(dāng)前線程中執(zhí)行，沒有調(diào)度器return Observable.defer(new Func0<Observable<R>>() {@Overridepublic Observable<R> call() {executionResult = executionResult.setExecutionOccurred();if (!commandState.compareAndSet(CommandState.OBSERVABLE_CHAIN_CREATED, CommandState.USER_CODE_EXECUTED)) {return Observable.error(new IllegalStateException("execution attempted while in state : " + commandState.get().name()));}metrics.markCommandStart(commandKey, threadPoolKey, ExecutionIsolationStrategy.SEMAPHORE);// semaphore isolated// store the command that is being runendCurrentThreadExecutingCommand = Hystrix.startCurrentThreadExecutingCommand(getCommandKey());try {executionHook.onRunStart(_cmd);executionHook.onExecutionStart(_cmd);return getUserExecutionObservable(_cmd); ?//the getUserExecutionObservable method already wraps sync exceptions, so this shouldn't throw} catch (Throwable ex) {//If the above hooks throw, then use that as the result of the run methodreturn Observable.error(ex);}}});}}

?

5.3 核心運(yùn)行流程

?

private Observable<R> executeCommandAndObserve(final AbstractCommand<R> _cmd) {final HystrixRequestContext currentRequestContext = HystrixRequestContext.getContextForCurrentThread();//執(zhí)行發(fā)生的回調(diào)final Action1<R> markEmits = new Action1<R>() {@Overridepublic void call(R r) {if (shouldOutputOnNextEvents()) {executionResult = executionResult.addEvent(HystrixEventType.EMIT);eventNotifier.markEvent(HystrixEventType.EMIT, commandKey);}if (commandIsScalar()) {long latency = System.currentTimeMillis() - executionResult.getStartTimestamp();eventNotifier.markEvent(HystrixEventType.SUCCESS, commandKey);executionResult = executionResult.addEvent((int) latency, HystrixEventType.SUCCESS);eventNotifier.markCommandExecution(getCommandKey(), properties.executionIsolationStrategy().get(), (int) latency, executionResult.getOrderedList());circuitBreaker.markSuccess();}}};//執(zhí)行成功的回調(diào)，標(biāo)記下狀態(tài)，熔斷器根據(jù)這個(gè)狀態(tài)維護(hù)熔斷邏輯final Action0 markOnCompleted = new Action0() {@Overridepublic void call() {if (!commandIsScalar()) {long latency = System.currentTimeMillis() - executionResult.getStartTimestamp();eventNotifier.markEvent(HystrixEventType.SUCCESS, commandKey);executionResult = executionResult.addEvent((int) latency, HystrixEventType.SUCCESS);eventNotifier.markCommandExecution(getCommandKey(), properties.executionIsolationStrategy().get(), (int) latency, executionResult.getOrderedList());circuitBreaker.markSuccess();}}};//執(zhí)行失敗的回調(diào)final Func1<Throwable, Observable<R>> handleFallback = new Func1<Throwable, Observable<R>>() {@Overridepublic Observable<R> call(Throwable t) {circuitBreaker.markNonSuccess();Exception e = getExceptionFromThrowable(t);executionResult = executionResult.setExecutionException(e);//各種回調(diào)進(jìn)行各種fallbackif (e instanceof RejectedExecutionException) {return handleThreadPoolRejectionViaFallback(e);} else if (t instanceof HystrixTimeoutException) {return handleTimeoutViaFallback();} else if (t instanceof HystrixBadRequestException) {return handleBadRequestByEmittingError(e);} else {/** Treat HystrixBadRequestException from ExecutionHook like a plain HystrixBadRequestException.*/if (e instanceof HystrixBadRequestException) {eventNotifier.markEvent(HystrixEventType.BAD_REQUEST, commandKey);return Observable.error(e);}return handleFailureViaFallback(e);}}};final Action1<Notification<? super R>> setRequestContext = new Action1<Notification<? super R>>() {@Overridepublic void call(Notification<? super R> rNotification) {setRequestContextIfNeeded(currentRequestContext);}};Observable<R> execution;if (properties.executionTimeoutEnabled().get()) {execution = executeCommandWithSpecifiedIsolation(_cmd).lift(new HystrixObservableTimeoutOperator<R>(_cmd));} else {execution = executeCommandWithSpecifiedIsolation(_cmd);}//注冊各種回調(diào)函數(shù)return execution.doOnNext(markEmits).doOnCompleted(markOnCompleted).onErrorResumeNext(handleFallback).doOnEach(setRequestContext);}

?

6. 小結(jié)

?

• Hystrix 是基于單機(jī)應(yīng)用的熔斷限流框架

• 根據(jù)熔斷器的滑動(dòng)窗口判斷當(dāng)前請求是否可以執(zhí)行

• 線程競爭實(shí)現(xiàn)“半關(guān)閉”狀態(tài)，拿一個(gè)請求試試是否可以關(guān)閉熔斷器

• 線程池隔離將請求丟到線程池中運(yùn)行，限流依靠線程池拒絕策略

• 信號(hào)量隔離在當(dāng)前線程中運(yùn)行，限流依靠并發(fā)請求數(shù)

• 當(dāng)信號(hào)量競爭失敗/線程池隊(duì)列滿，就進(jìn)入限流模式，執(zhí)行 Fallback

• 當(dāng)熔斷器開啟，就熔斷請求，執(zhí)行 Fallback　

• 整個(gè)框架采用的 RxJava 的編程模式，回調(diào)函數(shù)滿天飛　

總結(jié)

以上是生活随笔為你收集整理的关于“豪猪”，你理解的透彻吗？【Hystrix是个什么玩意儿】的全部內(nèi)容，希望文章能夠幫你解決所遇到的問題。

如果覺得生活随笔網(wǎng)站內(nèi)容還不錯(cuò)，歡迎將生活随笔推薦給好友。