一、Sentinel簡(jiǎn)介

Sentinel 以流量為切入點(diǎn)，從流量控制、熔斷降級(jí)、系統(tǒng)負(fù)載保護(hù)等多個(gè)維度保護(hù)服務(wù)的穩(wěn)定性。

Sentinel 具有以下特征:

?豐富的應(yīng)用場(chǎng)景：Sentinel 承接了阿里巴巴近 10 年的雙十一大促流量的核心場(chǎng)景，例如秒殺（即突發(fā)流量控制在系統(tǒng)容量可以承受的范圍）、消息削峰填谷、集群流量控制、實(shí)時(shí)熔斷下游不可用應(yīng)用等。

?完備的實(shí)時(shí)監(jiān)控：Sentinel 同時(shí)提供實(shí)時(shí)的監(jiān)控功能。您可以在控制臺(tái)中看到接入應(yīng)用的單臺(tái)機(jī)器秒級(jí)數(shù)據(jù)，甚至 500 臺(tái)以下規(guī)模的集群的匯總運(yùn)行情況。

?廣泛的開(kāi)源生態(tài)：Sentinel 提供開(kāi)箱即用的與其它開(kāi)源框架/庫(kù)的整合模塊，例如與 Spring Cloud、Apache Dubbo、gRPC、Quarkus 的整合。您只需要引入相應(yīng)的依賴并進(jìn)行簡(jiǎn)單的配置即可快速地接入 Sentinel。同時(shí) Sentinel 提供 Java/Go/C++ 等多語(yǔ)言的原生實(shí)現(xiàn)。

?完善的 SPI 擴(kuò)展機(jī)制：Sentinel 提供簡(jiǎn)單易用、完善的 SPI 擴(kuò)展接口。您可以通過(guò)實(shí)現(xiàn)擴(kuò)展接口來(lái)快速地定制邏輯。例如定制規(guī)則管理、適配動(dòng)態(tài)數(shù)據(jù)源等。

有關(guān)Sentinel的詳細(xì)介紹以及和Hystrix的區(qū)別可以自行網(wǎng)上檢索，推薦一篇文章：https://mp.weixin.qq.com/s/Q7Xv8cypQFrrOQhbd9BOXw

本次主要使用了Sentinel的降級(jí)、限流、系統(tǒng)負(fù)載保護(hù)功能

二、Sentinel關(guān)鍵技術(shù)源碼解析

無(wú)論是限流、降級(jí)、負(fù)載等控制手段，大致流程如下：

?StatisticSlot 則用于記錄、統(tǒng)計(jì)不同維度的 runtime 指標(biāo)監(jiān)控信息

?責(zé)任鏈依次觸發(fā)后續(xù) slot 的 entry 方法，如 SystemSlot、FlowSlot、DegradeSlot 等的規(guī)則校驗(yàn)；

?當(dāng)后續(xù)的 slot 通過(guò)，沒(méi)有拋出 BlockException 異常，說(shuō)明該資源被成功調(diào)用，則增加執(zhí)行線程數(shù)和通過(guò)的請(qǐng)求數(shù)等信息。

關(guān)于數(shù)據(jù)統(tǒng)計(jì)，主要會(huì)牽扯到 ArrayMetric、BucketLeapArray、MetricBucket、WindowWrap 等類。

項(xiàng)目結(jié)構(gòu)

以下主要分析core包里的內(nèi)容

2.1注解入口

2.1.1 Entry、Context、Node

SphU門(mén)面類的方法出參都是Entry，Entry可以理解為每次進(jìn)入資源的一個(gè)憑證，如果調(diào)用SphO.entry()或者SphU.entry()能獲取Entry對(duì)象，代表獲取了憑證，沒(méi)有被限流，否則拋出一個(gè)BlockException。

Entry中持有本次對(duì)資源調(diào)用的相關(guān)信息：

?createTime：創(chuàng)建該Entry的時(shí)間戳。

?curNode：Entry當(dāng)前是在哪個(gè)節(jié)點(diǎn)。

?orginNode：Entry的調(diào)用源節(jié)點(diǎn)。

?resourceWrapper：Entry關(guān)聯(lián)的資源信息。

?

Entry是一個(gè)抽象類，CtEntry是Entry的實(shí)現(xiàn)，CtEntry持有Context和調(diào)用鏈的信息

Context的源碼注釋如下，

This class holds metadata of current invocation

Node的源碼注釋

Holds real-time statistics for resources

Node中保存了對(duì)資源的實(shí)時(shí)數(shù)據(jù)的統(tǒng)計(jì)，Sentinel中的限流或者降級(jí)等功能就是通過(guò)Node中的數(shù)據(jù)進(jìn)行判斷的。Node是一個(gè)接口，里面定義了各種操作request、exception、rt、qps、thread的方法。

在細(xì)看Node實(shí)現(xiàn)時(shí)，不難發(fā)現(xiàn)LongAddr的使用，關(guān)于LongAddr和DoubleAddr都是java8 java.util.concurrent.atomic里的內(nèi)容，感興趣的小伙伴可以再深入研究一下，這兩個(gè)是高并發(fā)下計(jì)數(shù)功能非常優(yōu)秀的數(shù)據(jù)結(jié)構(gòu)，實(shí)際應(yīng)用場(chǎng)景里需要計(jì)數(shù)時(shí)可以考慮使用。
關(guān)于Node的介紹后續(xù)還會(huì)深入，此處大致先提一下這個(gè)概念。

2.2 初始化

2.2.1 Context初始化

在初始化slot責(zé)任鏈部分前，還執(zhí)行了context的初始化，里面涉及幾個(gè)重要概念，需要解釋一下：

可以發(fā)現(xiàn)在Context初始化的過(guò)程中，會(huì)把EntranceNode加入到Root子節(jié)點(diǎn)中（實(shí)際Root本身是一個(gè)特殊的EntranceNode），并把EntranceNode放到contextNameNodeMap中。

之前簡(jiǎn)單提到過(guò)Node，是用來(lái)統(tǒng)計(jì)數(shù)據(jù)用的，不同Node功能如下：

?Node：用于完成數(shù)據(jù)統(tǒng)計(jì)的接口

?StatisticNode：統(tǒng)計(jì)節(jié)點(diǎn)，是Node接口的實(shí)現(xiàn)類，用于完成數(shù)據(jù)統(tǒng)計(jì)

?EntranceNode：入口節(jié)點(diǎn)，一個(gè)Context會(huì)有一個(gè)入口節(jié)點(diǎn)，用于統(tǒng)計(jì)當(dāng)前Context的總體流量數(shù)據(jù)

?DefaultNode：默認(rèn)節(jié)點(diǎn)，用于統(tǒng)計(jì)一個(gè)資源在當(dāng)前Context中的流量數(shù)據(jù)

?ClusterNode：集群節(jié)點(diǎn)，用于統(tǒng)計(jì)一個(gè)資源在所有Context中的總體流量數(shù)據(jù)

protected static Context trueEnter(String name, String origin) {
        Context context = contextHolder.get();
        if (context == null) {
            Map localCacheNameMap = contextNameNodeMap;
            DefaultNode node = localCacheNameMap.get(name);
            if (node == null) {
                if (localCacheNameMap.size() > Constants.MAX_CONTEXT_NAME_SIZE) {
                    setNullContext();
                    return NULL_CONTEXT;
                } else {
                    LOCK.lock();
                    try {
                        node = contextNameNodeMap.get(name);
                        if (node == null) {
                            if (contextNameNodeMap.size() > Constants.MAX_CONTEXT_NAME_SIZE) {
                                setNullContext();
                                return NULL_CONTEXT;
                            } else {
                                node = new EntranceNode(new StringResourceWrapper(name, EntryType.IN), null);
                                // Add entrance node.
                                Constants.ROOT.addChild(node);

                                Map newMap = new HashMap(contextNameNodeMap.size() + 1);
                                newMap.putAll(contextNameNodeMap);
                                newMap.put(name, node);
                                contextNameNodeMap = newMap;
                            }
                        }
                    } finally {
                        LOCK.unlock();
                    }
                }
            }
            context = new Context(node, name);
            context.setOrigin(origin);
            contextHolder.set(context);
        }

        return context;
    }

2.2.2 通過(guò)SpiLoader默認(rèn)初始化8個(gè)slot

每個(gè)slot的主要職責(zé)如下：

?NodeSelectorSlot 負(fù)責(zé)收集資源的路徑，并將這些資源的調(diào)用路徑，以樹(shù)狀結(jié)構(gòu)存儲(chǔ)起來(lái)，用于根據(jù)調(diào)用路徑來(lái)限流降級(jí)；

?ClusterBuilderSlot 則用于存儲(chǔ)資源的統(tǒng)計(jì)信息以及調(diào)用者信息，例如該資源的 RT, QPS, thread count 等等，這些信息將用作為多維度限流，降級(jí)的依據(jù)；

?StatisticSlot 則用于記錄、統(tǒng)計(jì)不同緯度的 runtime 指標(biāo)監(jiān)控信息；

?FlowSlot 則用于根據(jù)預(yù)設(shè)的限流規(guī)則以及前面 slot 統(tǒng)計(jì)的狀態(tài)，來(lái)進(jìn)行流量控制；

?AuthoritySlot 則根據(jù)配置的黑白名單和調(diào)用來(lái)源信息，來(lái)做黑白名單控制；

?DegradeSlot 則通過(guò)統(tǒng)計(jì)信息以及預(yù)設(shè)的規(guī)則，來(lái)做熔斷降級(jí)；

?SystemSlot 則通過(guò)系統(tǒng)的狀態(tài)，例如 集群QPS、線程數(shù)、RT、負(fù)載 等，來(lái)控制總的入口流量；

2.3 StatisticSlot

2.3.1 Node

深入看一下Node，因?yàn)榻y(tǒng)計(jì)信息都在里面，后面不論是限流、熔斷、負(fù)載保護(hù)等都是結(jié)合規(guī)則+統(tǒng)計(jì)信息判斷是否要執(zhí)行

從Node的源碼注釋看，它會(huì)持有資源維度的實(shí)時(shí)統(tǒng)計(jì)數(shù)據(jù)，以下是接口里的方法定義，可以看到totalRequest、totalPass、totalSuccess、blockRequest、totalException、passQps等很多request、qps、thread的相關(guān)方法：

/**
 * Holds real-time statistics for resources.
 *
 * @author qinan.qn
 * @author leyou
 * @author Eric Zhao
 */
public interface Node extends OccupySupport, DebugSupport {
    long totalRequest();
    long totalPass();
    long totalSuccess();
    long blockRequest();
    long totalException();
    double passQps();
    double blockQps();
    double totalQps();
    double successQps();
    ……
}

2.3.2 StatisticNode

我們先從最基礎(chǔ)的StatisticNode開(kāi)始看，源碼給出的定位是：

The statistic node keep three kinds of real-time statistics metrics：
metrics in second level ({@code rollingCounterInSecond})
metrics in minute level ({@code rollingCounterInMinute})
thread count

StatisticNode只有四個(gè)屬性，除了之前提到過(guò)的LongAddr類型的curThreadNum外，還有兩個(gè)屬性是Metric對(duì)象，通過(guò)入?yún)⒁呀?jīng)屬性命名可以看出，一個(gè)用于秒級(jí)，一個(gè)用于分鐘級(jí)統(tǒng)計(jì)。接下來(lái)我們就要看看Metric

// StatisticNode持有兩個(gè)Metric，一個(gè)秒級(jí)一個(gè)分鐘級(jí)，由入?yún)⒖芍爰?jí)統(tǒng)計(jì)劃分了兩個(gè)時(shí)間窗口，窗口程度是500ms
private transient volatile Metric rollingCounterInSecond = new ArrayMetric(SampleCountProperty.SAMPLE_COUNT,
    IntervalProperty.INTERVAL);

// 分鐘級(jí)統(tǒng)計(jì)劃分了60個(gè)時(shí)間窗口，窗口長(zhǎng)度是1000ms
private transient Metric rollingCounterInMinute = new ArrayMetric(60, 60 * 1000, false);

/**
 * The counter for thread count.
 */
private LongAdder curThreadNum = new LongAdder();

/**
 * The last timestamp when metrics were fetched.
 */
private long lastFetchTime = -1;

ArrayMetric只有一個(gè)屬性LeapArray，其余都是用于統(tǒng)計(jì)的方法，LeapArray是sentinel中統(tǒng)計(jì)最基本的數(shù)據(jù)結(jié)構(gòu)，這里有必要詳細(xì)看一下，總體就是根據(jù)timeMillis去獲取一個(gè)bucket，分為：沒(méi)有創(chuàng)建、有直接返回、被廢棄后的reset三種場(chǎng)景。

//以分鐘級(jí)的統(tǒng)計(jì)屬性為例，看一下時(shí)間窗口初始化過(guò)程
private transient Metric rollingCounterInMinute = new ArrayMetric(60, 60 * 1000, false);


public LeapArray(int sampleCount, int intervalInMs) {
        AssertUtil.isTrue(sampleCount > 0, "bucket count is invalid: " + sampleCount);
        AssertUtil.isTrue(intervalInMs > 0, "total time interval of the sliding window should be positive");
        AssertUtil.isTrue(intervalInMs % sampleCount == 0, "time span needs to be evenly divided");
        // windowLengthInMs = 60*1000 / 60 = 1000 滑動(dòng)窗口時(shí)間長(zhǎng)度，可見(jiàn)sentinel默認(rèn)將單位時(shí)間分為了60個(gè)滑動(dòng)窗口進(jìn)行數(shù)據(jù)統(tǒng)計(jì)
        this.windowLengthInMs = intervalInMs / sampleCount;
        // 60*1000
        this.intervalInMs = intervalInMs;
        // 60
        this.intervalInSecond = intervalInMs / 1000.0;
        // 60
        this.sampleCount = sampleCount;
        // 數(shù)組長(zhǎng)度60
        this.array = new AtomicReferenceArray(sampleCount);
    }

/**
     * Get bucket item at provided timestamp.
     *
     * @param timeMillis a valid timestamp in milliseconds
     * @return current bucket item at provided timestamp if the time is valid; null if time is invalid
     */
    public WindowWrap currentWindow(long timeMillis) {
        if (timeMillis < 0) {
            return null;
        }
        // 根據(jù)當(dāng)前時(shí)間戳算一個(gè)數(shù)組索引
        int idx = calculateTimeIdx(timeMillis);
        // Calculate current bucket start time.
        // timeMillis % 1000
        long windowStart = calculateWindowStart(timeMillis);

        /*
         * Get bucket item at given time from the array.
         *
         * (1) Bucket is absent, then just create a new bucket and CAS update to circular array.
         * (2) Bucket is up-to-date, then just return the bucket.
         * (3) Bucket is deprecated, then reset current bucket.
         */
        while (true) {
            WindowWrap old = array.get(idx);
            if (old == null) {
                /*
                 *     B0       B1      B2    NULL      B4
                 * ||_______|_______|_______|_______|_______||___
                 * 200     400     600     800     1000    1200  timestamp
                 *                             ^
                 *                          time=888
                 *            bucket is empty, so create new and update
                 *
                 * If the old bucket is absent, then we create a new bucket at {@code windowStart},
                 * then try to update circular array via a CAS operation. Only one thread can
                 * succeed to update, while other threads yield its time slice.
                 */
                // newEmptyBucket 方法重寫(xiě)，秒級(jí)和分鐘級(jí)統(tǒng)計(jì)對(duì)象實(shí)現(xiàn)不同
                WindowWrap window = new WindowWrap(windowLengthInMs, windowStart, newEmptyBucket(timeMillis));
                if (array.compareAndSet(idx, null, window)) {
                    // Successfully updated, return the created bucket.
                    return window;
                } else {
                    // Contention failed, the thread will yield its time slice to wait for bucket available.
                    Thread.yield();
                }
            } else if (windowStart == old.windowStart()) {
                /*
                 *     B0       B1      B2     B3      B4
                 * ||_______|_______|_______|_______|_______||___
                 * 200     400     600     800     1000    1200  timestamp
                 *                             ^
                 *                          time=888
                 *            startTime of Bucket 3: 800, so it's up-to-date
                 *
                 * If current {@code windowStart} is equal to the start timestamp of old bucket,
                 * that means the time is within the bucket, so directly return the bucket.
                 */
                return old;
            } else if (windowStart > old.windowStart()) {
                /*
                 *   (old)
                 *             B0       B1      B2    NULL      B4
                 * |_______||_______|_______|_______|_______|_______||___
                 * ...    1200     1400    1600    1800    2000    2200  timestamp
                 *                              ^
                 *                           time=1676
                 *          startTime of Bucket 2: 400, deprecated, should be reset
                 *
                 * If the start timestamp of old bucket is behind provided time, that means
                 * the bucket is deprecated. We have to reset the bucket to current {@code windowStart}.
                 * Note that the reset and clean-up operations are hard to be atomic,
                 * so we need a update lock to guarantee the correctness of bucket update.
                 *
                 * The update lock is conditional (tiny scope) and will take effect only when
                 * bucket is deprecated, so in most cases it won't lead to performance loss.
                 */
                if (updateLock.tryLock()) {
                    try {
                        // Successfully get the update lock, now we reset the bucket.
                        return resetWindowTo(old, windowStart);
                    } finally {
                        updateLock.unlock();
                    }
                } else {
                    // Contention failed, the thread will yield its time slice to wait for bucket available.
                    Thread.yield();
                }
            } else if (windowStart < old.windowStart()) {
                // Should not go through here, as the provided time is already behind.
                return new WindowWrap(windowLengthInMs, windowStart, newEmptyBucket(timeMillis));
            }
        }
    }
// 持有一個(gè)時(shí)間窗口對(duì)象的數(shù)據(jù)，會(huì)根據(jù)當(dāng)前時(shí)間戳除以時(shí)間窗口長(zhǎng)度然后散列到數(shù)組中
private int calculateTimeIdx(/*@Valid*/ long timeMillis) {
        long timeId = timeMillis / windowLengthInMs;
        // Calculate current index so we can map the timestamp to the leap array.
        return (int)(timeId % array.length());
    }

WindowWrap持有了windowLengthInMs, windowStart和LeapArray（分鐘統(tǒng)計(jì)實(shí)現(xiàn)是BucketLeapArray，秒級(jí)統(tǒng)計(jì)實(shí)現(xiàn)是OccupiableBucketLeapArray），對(duì)于分鐘級(jí)別的統(tǒng)計(jì)，MetricBucket維護(hù)了一個(gè)longAddr數(shù)組和一個(gè)配置的minRT

/**
 * The fundamental data structure for metric statistics in a time span.
 *
 * @author jialiang.linjl
 * @author Eric Zhao
 * @see LeapArray
 */
public class BucketLeapArray extends LeapArray {

    public BucketLeapArray(int sampleCount, int intervalInMs) {
        super(sampleCount, intervalInMs);
    }

    @Override
    public MetricBucket newEmptyBucket(long time) {
        return new MetricBucket();
    }

    @Override
    protected WindowWrap resetWindowTo(WindowWrap w, long startTime) {
        // Update the start time and reset value.
        w.resetTo(startTime);
        w.value().reset();
        return w;
    }
}

對(duì)于秒級(jí)統(tǒng)計(jì)，QPS=20場(chǎng)景下，如何準(zhǔn)確統(tǒng)計(jì)的問(wèn)題，此處用到了另外一個(gè)LeapArry實(shí)現(xiàn)FutureBucketLeapArray，至于秒級(jí)統(tǒng)計(jì)如何保證沒(méi)有統(tǒng)計(jì)誤差，讀者可以再研究一下FutureBucketLeapArray的上下文就好。

2.4 FlowSlot

2.4.1 常見(jiàn)限流算法

介紹sentinel限流實(shí)現(xiàn)前，先介紹一下常見(jiàn)限流算法，基本分為三種：計(jì)數(shù)器、漏斗、令牌桶。

計(jì)數(shù)器算法

顧名思義，計(jì)數(shù)器算法就是統(tǒng)計(jì)某個(gè)時(shí)間段內(nèi)的請(qǐng)求，每單位時(shí)間加1，然后與配置的限流值（最大QPS）進(jìn)行比較，如果超出則觸發(fā)限流。但是這種算法不能做到“平滑限流”，以1s為單位時(shí)間，100QPS為限流值為例，如下圖，會(huì)出現(xiàn)某時(shí)段超出限流值的情況

因此在單純計(jì)數(shù)器算法上，又出現(xiàn)了滑動(dòng)窗口計(jì)數(shù)器算法，我們將統(tǒng)計(jì)時(shí)間細(xì)分，比如將1s統(tǒng)計(jì)時(shí)長(zhǎng)分為5個(gè)時(shí)間窗口，通過(guò)滾動(dòng)統(tǒng)計(jì)所有時(shí)間窗口的QPS作為系統(tǒng)實(shí)際的QPS的方式，就能解決上述臨界統(tǒng)計(jì)問(wèn)題，后續(xù)我們看sentinel源碼時(shí)也能看到類似操作。

漏斗算法

不論流量有多大都會(huì)先到漏桶中，然后以均勻的速度流出。如何在代碼中實(shí)現(xiàn)這個(gè)勻速呢？比如我們想讓勻速為100q/s，那么我們可以得到每流出一個(gè)流量需要消耗10ms，類似一個(gè)隊(duì)列，每隔10ms從隊(duì)列頭部取出流量進(jìn)行放行，而我們的隊(duì)列也就是漏桶，當(dāng)流量大于隊(duì)列的長(zhǎng)度的時(shí)候，我們就可以拒絕超出的部分。

漏斗算法同樣的也有一定的缺點(diǎn)：無(wú)法應(yīng)對(duì)突發(fā)流量。比如一瞬間來(lái)了100個(gè)請(qǐng)求，在漏桶算法中只能一個(gè)一個(gè)的過(guò)去，當(dāng)最后一個(gè)請(qǐng)求流出的時(shí)候時(shí)間已經(jīng)過(guò)了一秒了，所以漏斗算法比較適合請(qǐng)求到達(dá)比較均勻，需要嚴(yán)格控制請(qǐng)求速率的場(chǎng)景。

令牌桶算法

令牌桶算法和漏斗算法比較類似，區(qū)別是令牌桶存放的是令牌數(shù)量不是請(qǐng)求數(shù)量，令牌桶可以根據(jù)自身需求多樣性得管理令牌的生產(chǎn)和消耗，可以解決突發(fā)流量的問(wèn)題。

2.4.2 單機(jī)限流模式

接下來(lái)我們看一下Sentinel中的限流實(shí)現(xiàn)，相比上述基本限流算法，Sentinel限流的第一個(gè)特性就是引入“資源”的概念，可以細(xì)粒度多樣性的支持特定資源、關(guān)聯(lián)資源、指定鏈路的限流。

FlowSlot的主要邏輯都在FlowRuleChecker里，介紹之前，我們先看一下Sentinel關(guān)于規(guī)則的模型描述，下圖分別是限流、訪問(wèn)控制規(guī)則、系統(tǒng)保護(hù)規(guī)則（Linux負(fù)載）、降級(jí)規(guī)則

    /**
     * 流量控制兩種模式 
     *   0: thread count（當(dāng)調(diào)用該api的線程數(shù)達(dá)到閾值的時(shí)候，進(jìn)行限流）
     *   1: QPS（當(dāng)調(diào)用該api的QPS達(dá)到閾值的時(shí)候，進(jìn)行限流）
     */
    private int grade = RuleConstant.FLOW_GRADE_QPS;

    /**
     * 流量控制閾值，值含義與grade有關(guān)
     */
    private double count;

    /**
     * 調(diào)用關(guān)系限流策略（可以支持關(guān)聯(lián)資源或指定鏈路的多樣性限流需求）
     *  直接（api 達(dá)到限流條件時(shí)，直接限流）
     *  關(guān)聯(lián)（當(dāng)關(guān)聯(lián)的資源達(dá)到限流閾值時(shí)，就限流自己）
     *  鏈路（只記錄指定鏈路上的流量）
     * {@link RuleConstant#STRATEGY_DIRECT} for direct flow control (by origin);
     * {@link RuleConstant#STRATEGY_RELATE} for relevant flow control (with relevant resource);
     * {@link RuleConstant#STRATEGY_CHAIN} for chain flow control (by entrance resource).
     */
    private int strategy = RuleConstant.STRATEGY_DIRECT;

    /**
     * Reference resource in flow control with relevant resource or context.
     */
    private String refResource;

    /**
     * 流控效果：
     * 0. default(reject directly),直接拒絕，拋異常FlowException
     * 1. warm up, 慢啟動(dòng)模式（根據(jù)coldFactor（冷加載因子，默認(rèn)3）的值，從閾值/coldFactor，經(jīng)過(guò)預(yù)熱時(shí)長(zhǎng)，才達(dá)到設(shè)置的QPS閾值）
     * 2. rate limiter  排隊(duì)等待
     * 3. warm up + rate limiter
     */
    private int controlBehavior = RuleConstant.CONTROL_BEHAVIOR_DEFAULT;

    private int warmUpPeriodSec = 10;

    /**
     * Max queueing time in rate limiter behavior.
     */
    private int maxQueueingTimeMs = 500;

    /**
    *  是否集群限流，默認(rèn)為否
    */
    private boolean clusterMode;
    /**
     * Flow rule config for cluster mode.
     */
    private ClusterFlowConfig clusterConfig;

    /**
     * The traffic shaping (throttling) controller.
     */
    private TrafficShapingController controller;

接著我們繼續(xù)分析FlowRuleChecker

canPassCheck第一步會(huì)好看limitApp，這個(gè)是結(jié)合訪問(wèn)授權(quán)限制規(guī)則使用的，默認(rèn)是所有。

private static boolean passLocalCheck(FlowRule rule, Context context, DefaultNode node, int acquireCount,
                                          boolean prioritized) {
        // 根據(jù)策略選擇Node來(lái)進(jìn)行統(tǒng)計(jì)（可以是本身Node、關(guān)聯(lián)的Node、指定的鏈路）
        Node selectedNode = selectNodeByRequesterAndStrategy(rule, context, node);
        if (selectedNode == null) {
            return true;
        }

        return rule.getRater().canPass(selectedNode, acquireCount, prioritized);
    }


static Node selectNodeByRequesterAndStrategy(/*@NonNull*/ FlowRule rule, Context context, DefaultNode node) {
        // limitApp是訪問(wèn)控制使用的，默認(rèn)是default，不限制來(lái)源
        String limitApp = rule.getLimitApp();
        // 拿到限流策略
        int strategy = rule.getStrategy();
        String origin = context.getOrigin();
        // 基于調(diào)用來(lái)源做鑒權(quán)
        if (limitApp.equals(origin) && filterOrigin(origin)) {
            if (strategy == RuleConstant.STRATEGY_DIRECT) {
                // Matches limit origin, return origin statistic node.
                return context.getOriginNode();
            }
            // 
            return selectReferenceNode(rule, context, node);
        } else if (RuleConstant.LIMIT_APP_DEFAULT.equals(limitApp)) {
            if (strategy == RuleConstant.STRATEGY_DIRECT) {
                // Return the cluster node.
                return node.getClusterNode();
            }

            return selectReferenceNode(rule, context, node);
        } else if (RuleConstant.LIMIT_APP_OTHER.equals(limitApp)
            && FlowRuleManager.isOtherOrigin(origin, rule.getResource())) {
            if (strategy == RuleConstant.STRATEGY_DIRECT) {
                return context.getOriginNode();
            }

            return selectReferenceNode(rule, context, node);
        }

        return null;
    }

static Node selectReferenceNode(FlowRule rule, Context context, DefaultNode node) {
        String refResource = rule.getRefResource();
        int strategy = rule.getStrategy();

        if (StringUtil.isEmpty(refResource)) {
            return null;
        }

        if (strategy == RuleConstant.STRATEGY_RELATE) {
            return ClusterBuilderSlot.getClusterNode(refResource);
        }

        if (strategy == RuleConstant.STRATEGY_CHAIN) {
            if (!refResource.equals(context.getName())) {
                return null;
            }
            return node;
        }
        // No node.
        return null;
    }

// 此代碼是load限流規(guī)則時(shí)根據(jù)規(guī)則初始化流量整形控制器的邏輯，rule.getRater()返回TrafficShapingController
private static TrafficShapingController generateRater(/*@Valid*/ FlowRule rule) {
        if (rule.getGrade() == RuleConstant.FLOW_GRADE_QPS) {
            switch (rule.getControlBehavior()) {
                // 預(yù)熱模式返回WarmUpController
                case RuleConstant.CONTROL_BEHAVIOR_WARM_UP:
                    return new WarmUpController(rule.getCount(), rule.getWarmUpPeriodSec(),
                            ColdFactorProperty.coldFactor);
                // 排隊(duì)模式返回ThrottlingController
                case RuleConstant.CONTROL_BEHAVIOR_RATE_LIMITER:
                    return new ThrottlingController(rule.getMaxQueueingTimeMs(), rule.getCount());
                // 預(yù)熱+排隊(duì)模式返回WarmUpRateLimiterController
                case RuleConstant.CONTROL_BEHAVIOR_WARM_UP_RATE_LIMITER:
                    return new WarmUpRateLimiterController(rule.getCount(), rule.getWarmUpPeriodSec(),
                            rule.getMaxQueueingTimeMs(), ColdFactorProperty.coldFactor);
                case RuleConstant.CONTROL_BEHAVIOR_DEFAULT:
                default:
                    // Default mode or unknown mode: default traffic shaping controller (fast-reject).
            }
        }
        // 默認(rèn)是DefaultController
        return new DefaultController(rule.getCount(), rule.getGrade());
    }

Sentinel單機(jī)限流算法

上面我們看到根據(jù)限流規(guī)則controlBehavior屬性（流控效果），會(huì)初始化以下實(shí)現(xiàn)：

?DefaultController：是一個(gè)非常典型的滑動(dòng)窗口計(jì)數(shù)器算法實(shí)現(xiàn)，將當(dāng)前統(tǒng)計(jì)的qps和請(qǐng)求進(jìn)來(lái)的qps進(jìn)行求和，小于限流值則通過(guò)，大于則計(jì)算一個(gè)等待時(shí)間，稍后再試

?ThrottlingController：是漏斗算法的實(shí)現(xiàn)，實(shí)現(xiàn)思路已經(jīng)在源碼片段中加了備注

?WarmUpController：實(shí)現(xiàn)參考了Guava的帶預(yù)熱的RateLimiter，區(qū)別是Guava側(cè)重于請(qǐng)求間隔，類似前面提到的令牌桶，而Sentinel更關(guān)注于請(qǐng)求數(shù)，和令牌桶算法有點(diǎn)類似

?WarmUpRateLimiterController：低水位使用預(yù)熱算法，高水位使用滑動(dòng)窗口計(jì)數(shù)器算法排隊(duì)。

DefaultController

    @Override
    public boolean canPass(Node node, int acquireCount, boolean prioritized) {
        int curCount = avgUsedTokens(node);
        if (curCount + acquireCount > count) {
            if (prioritized && grade == RuleConstant.FLOW_GRADE_QPS) {
                long currentTime;
                long waitInMs;
                currentTime = TimeUtil.currentTimeMillis();
                waitInMs = node.tryOccupyNext(currentTime, acquireCount, count);
                if (waitInMs < OccupyTimeoutProperty.getOccupyTimeout()) {
                    node.addWaitingRequest(currentTime + waitInMs, acquireCount);
                    node.addOccupiedPass(acquireCount);
                    sleep(waitInMs);

                    // PriorityWaitException indicates that the request will pass after waiting for {@link @waitInMs}.
                    throw new PriorityWaitException(waitInMs);
                }
            }
            return false;
        }
        return true;
    }

ThrottlingController

 public ThrottlingController(int queueingTimeoutMs, double maxCountPerStat) {
        this(queueingTimeoutMs, maxCountPerStat, 1000);
    }

    public ThrottlingController(int queueingTimeoutMs, double maxCountPerStat, int statDurationMs) {
        AssertUtil.assertTrue(statDurationMs > 0, "statDurationMs should be positive");
        AssertUtil.assertTrue(maxCountPerStat >= 0, "maxCountPerStat should be >= 0");
        AssertUtil.assertTrue(queueingTimeoutMs >= 0, "queueingTimeoutMs should be >= 0");
        this.maxQueueingTimeMs = queueingTimeoutMs;
        this.count = maxCountPerStat;
        this.statDurationMs = statDurationMs;
        // Use nanoSeconds when durationMs%count != 0 or count/durationMs> 1 (to be accurate)
        // 可見(jiàn)配置限流值count大于1000時(shí)useNanoSeconds會(huì)是true否則是false
        if (maxCountPerStat > 0) {
            this.useNanoSeconds = statDurationMs % Math.round(maxCountPerStat) != 0 || maxCountPerStat / statDurationMs > 1;
        } else {
            this.useNanoSeconds = false;
        }
    }

    @Override
    public boolean canPass(Node node, int acquireCount) {
        return canPass(node, acquireCount, false);
    }

    private boolean checkPassUsingNanoSeconds(int acquireCount, double maxCountPerStat) {
        final long maxQueueingTimeNs = maxQueueingTimeMs * MS_TO_NS_OFFSET;
        long currentTime = System.nanoTime();
        // Calculate the interval between every two requests.
        final long costTimeNs = Math.round(1.0d * MS_TO_NS_OFFSET * statDurationMs * acquireCount / maxCountPerStat);

        // Expected pass time of this request.
        long expectedTime = costTimeNs + latestPassedTime.get();

        if (expectedTime <= currentTime) {
            // Contention may exist here, but it's okay.
            latestPassedTime.set(currentTime);
            return true;
        } else {
            final long curNanos = System.nanoTime();
            // Calculate the time to wait.
            long waitTime = costTimeNs + latestPassedTime.get() - curNanos;
            if (waitTime > maxQueueingTimeNs) {
                return false;
            }

            long oldTime = latestPassedTime.addAndGet(costTimeNs);
            waitTime = oldTime - curNanos;
            if (waitTime > maxQueueingTimeNs) {
                latestPassedTime.addAndGet(-costTimeNs);
                return false;
            }
            // in race condition waitTime may <= 0
            if (waitTime > 0) {
                sleepNanos(waitTime);
            }
            return true;
        }
    }
    
    // 漏斗算法具體實(shí)現(xiàn)
    private boolean checkPassUsingCachedMs(int acquireCount, double maxCountPerStat) {
        long currentTime = TimeUtil.currentTimeMillis();
        // 計(jì)算兩次請(qǐng)求的間隔（分為秒級(jí)和納秒級(jí)）
        long costTime = Math.round(1.0d * statDurationMs * acquireCount / maxCountPerStat);

        // 請(qǐng)求的期望的時(shí)間
        long expectedTime = costTime + latestPassedTime.get();

        if (expectedTime <= currentTime) {
            // latestPassedTime是AtomicLong類型，支持volatile語(yǔ)義
            latestPassedTime.set(currentTime);
            return true;
        } else {
            // 計(jì)算等待時(shí)間
            long waitTime = costTime + latestPassedTime.get() - TimeUtil.currentTimeMillis();
            // 如果大于最大排隊(duì)時(shí)間，則觸發(fā)限流
            if (waitTime > maxQueueingTimeMs) {
                return false;
            }
            
            long oldTime = latestPassedTime.addAndGet(costTime);
            waitTime = oldTime - TimeUtil.currentTimeMillis();
            if (waitTime > maxQueueingTimeMs) {
                latestPassedTime.addAndGet(-costTime);
                return false;
            }
            // in race condition waitTime may <= 0
            if (waitTime > 0) {
                sleepMs(waitTime);
            }
            return true;
        }
    }

    @Override
    public boolean canPass(Node node, int acquireCount, boolean prioritized) {
        // Pass when acquire count is less or equal than 0.
        if (acquireCount <= 0) {
            return true;
        }
        // Reject when count is less or equal than 0.
        // Otherwise, the costTime will be max of long and waitTime will overflow in some cases.
        if (count <= 0) {
            return false;
        }
        if (useNanoSeconds) {
            return checkPassUsingNanoSeconds(acquireCount, this.count);
        } else {
            return checkPassUsingCachedMs(acquireCount, this.count);
        }
    }

    private void sleepMs(long ms) {
        try {
            Thread.sleep(ms);
        } catch (InterruptedException e) {
        }
    }

    private void sleepNanos(long ns) {
        LockSupport.parkNanos(ns);
    }



long costTime = Math.round(1.0d * statDurationMs * acquireCount / maxCountPerStat);

由上述計(jì)算兩次請(qǐng)求間隔的公式我們可以發(fā)現(xiàn)，當(dāng)maxCountPerStat（規(guī)則配置的限流值QPS）超過(guò)1000后，就無(wú)法準(zhǔn)確計(jì)算出勻速排隊(duì)模式下的請(qǐng)求間隔時(shí)長(zhǎng)，因此對(duì)應(yīng)前面介紹的，當(dāng)規(guī)則配置限流值超過(guò)1000QPS后，會(huì)采用checkPassUsingNanoSeconds，小于1000QPS會(huì)采用checkPassUsingCachedMs，對(duì)比一下checkPassUsingNanoSeconds和checkPassUsingCachedMs，可以發(fā)現(xiàn)主體思路沒(méi)變，只是統(tǒng)計(jì)維度從毫秒換算成了納秒，因此只看checkPassUsingCachedMs實(shí)現(xiàn)就可以

WarmUpController

@Override
    public boolean canPass(Node node, int acquireCount, boolean prioritized) {
        long passQps = (long) node.passQps();

        long previousQps = (long) node.previousPassQps();
        syncToken(previousQps);

        // 開(kāi)始計(jì)算它的斜率
        // 如果進(jìn)入了警戒線，開(kāi)始調(diào)整他的qps
        long restToken = storedTokens.get();
        if (restToken >= warningToken) {
            long aboveToken = restToken - warningToken;
            // 消耗的速度要比warning快，但是要比慢
            // current interval = restToken*slope+1/count
            double warningQps = Math.nextUp(1.0 / (aboveToken * slope + 1.0 / count));
            if (passQps + acquireCount <= warningQps) {
                return true;
            }
        } else {
            if (passQps + acquireCount <= count) {
                return true;
            }
        }

        return false;
    }

protected void syncToken(long passQps) {
        long currentTime = TimeUtil.currentTimeMillis();
        currentTime = currentTime - currentTime % 1000;
        long oldLastFillTime = lastFilledTime.get();
        if (currentTime <= oldLastFillTime) {
            return;
        }

        long oldValue = storedTokens.get();
        long newValue = coolDownTokens(currentTime, passQps);

        if (storedTokens.compareAndSet(oldValue, newValue)) {
            long currentValue = storedTokens.addAndGet(0 - passQps);
            if (currentValue < 0) {
                storedTokens.set(0L);
            }
            lastFilledTime.set(currentTime);
        }

    }

private long coolDownTokens(long currentTime, long passQps) {
        long oldValue = storedTokens.get();
        long newValue = oldValue;

        // 添加令牌的判斷前提條件:
        // 當(dāng)令牌的消耗程度遠(yuǎn)遠(yuǎn)低于警戒線的時(shí)候
        if (oldValue < warningToken) {
            newValue = (long)(oldValue + (currentTime - lastFilledTime.get()) * count / 1000);
        } else if (oldValue > warningToken) {
            if (passQps < (int)count / coldFactor) {
                newValue = (long)(oldValue + (currentTime - lastFilledTime.get()) * count / 1000);
            }
        }
        return Math.min(newValue, maxToken);
    }

2.4.3 集群限流

passClusterCheck方法(因?yàn)閏lusterService找不到會(huì)降級(jí)到非集群限流)

private static boolean passClusterCheck(FlowRule rule, Context context, DefaultNode node, int acquireCount,
                                            boolean prioritized) {
        try {
            // 獲取當(dāng)前節(jié)點(diǎn)是Token Client還是Token Server
            TokenService clusterService = pickClusterService();
            if (clusterService == null) {
                return fallbackToLocalOrPass(rule, context, node, acquireCount, prioritized);
            }
            long flowId = rule.getClusterConfig().getFlowId();
            // 根據(jù)獲取的flowId通過(guò)TokenService進(jìn)行申請(qǐng)token。從上面可知，它可能是TokenClient調(diào)用的，也可能是ToeknServer調(diào)用的。分別對(duì)應(yīng)的類是DefaultClusterTokenClient和DefaultTokenService
            TokenResult result = clusterService.requestToken(flowId, acquireCount, prioritized);
            return applyTokenResult(result, rule, context, node, acquireCount, prioritized);
            // If client is absent, then fallback to local mode.
        } catch (Throwable ex) {
            RecordLog.warn("[FlowRuleChecker] Request cluster token unexpected failed", ex);
        }
        // Fallback to local flow control when token client or server for this rule is not available.
        // If fallback is not enabled, then directly pass.
        return fallbackToLocalOrPass(rule, context, node, acquireCount, prioritized);
    }

//獲取當(dāng)前節(jié)點(diǎn)是Token Client還是Token Server。
//1） 如果當(dāng)前節(jié)點(diǎn)的角色是Client，返回的TokenService為DefaultClusterTokenClient；
//2）如果當(dāng)前節(jié)點(diǎn)的角色是Server，則默認(rèn)返回的TokenService為DefaultTokenService。
private static TokenService pickClusterService() {
        if (ClusterStateManager.isClient()) {
            return TokenClientProvider.getClient();
        }
        if (ClusterStateManager.isServer()) {
            return EmbeddedClusterTokenServerProvider.getServer();
        }
        return null;
    }

集群限流模式

Sentinel 集群限流服務(wù)端有兩種啟動(dòng)方式：

?嵌入模式（Embedded）適合應(yīng)用級(jí)別的限流，部署簡(jiǎn)單，但對(duì)應(yīng)用性能有影響

?獨(dú)立模式（Alone）適合全局限流，需要獨(dú)立部署

考慮到文章篇幅，集群限流有機(jī)會(huì)再展開(kāi)詳細(xì)介紹。

集群限流模式降級(jí)

private static boolean passClusterCheck(FlowRule rule, Context context, DefaultNode node, int acquireCount,
                                            boolean prioritized) {
        try {
            TokenService clusterService = pickClusterService();
            if (clusterService == null) {
                return fallbackToLocalOrPass(rule, context, node, acquireCount, prioritized);
            }
            long flowId = rule.getClusterConfig().getFlowId();
            TokenResult result = clusterService.requestToken(flowId, acquireCount, prioritized);
            return applyTokenResult(result, rule, context, node, acquireCount, prioritized);
            // If client is absent, then fallback to local mode.
        } catch (Throwable ex) {
            RecordLog.warn("[FlowRuleChecker] Request cluster token unexpected failed", ex);
        }
        // Fallback to local flow control when token client or server for this rule is not available.
        // If fallback is not enabled, then directly pass.
        // 可以看到如果集群限流有異常，會(huì)降級(jí)到單機(jī)限流模式，如果配置不允許降級(jí)，那么直接會(huì)跳過(guò)此次校驗(yàn)
        return fallbackToLocalOrPass(rule, context, node, acquireCount, prioritized);
    }

2.5 DegradeSlot

CircuitBreaker

大神對(duì)斷路器的解釋：https://martinfowler.com/bliki/CircuitBreaker.html

首先就看到了根據(jù)資源名稱獲取斷路器列表，Sentinel的斷路器有兩個(gè)實(shí)現(xiàn)：RT模式使用ResponseTimeCircuitBreaker、異常模式使用ExceptionCircuitBreaker

public interface CircuitBreaker {

    /**
     * Get the associated circuit breaking rule.
     *
     * @return associated circuit breaking rule
     */
    DegradeRule getRule();

    /**
     * Acquires permission of an invocation only if it is available at the time of invoking.
     *
     * @param context context of current invocation
     * @return {@code true} if permission was acquired and {@code false} otherwise
     */
    boolean tryPass(Context context);

    /**
     * Get current state of the circuit breaker.
     *
     * @return current state of the circuit breaker
     */
    State currentState();

    /**
     * Record a completed request with the context and handle state transformation of the circuit breaker.
     * Called when a passed invocation finished.
     *
     * @param context context of current invocation
     */
    void onRequestComplete(Context context);

    /**
     * Circuit breaker state.
     */
    enum State {
        /**
         * In {@code OPEN} state, all requests will be rejected until the next recovery time point.
         */
        OPEN,
        /**
         * In {@code HALF_OPEN} state, the circuit breaker will allow a "probe" invocation.
         * If the invocation is abnormal according to the strategy (e.g. it's slow), the circuit breaker
         * will re-transform to the {@code OPEN} state and wait for the next recovery time point;
         * otherwise the resource will be regarded as "recovered" and the circuit breaker
         * will cease cutting off requests and transform to {@code CLOSED} state.
         */
        HALF_OPEN,
        /**
         * In {@code CLOSED} state, all requests are permitted. When current metric value exceeds the threshold,
         * the circuit breaker will transform to {@code OPEN} state.
         */
        CLOSED
    }
}

以ExceptionCircuitBreaker為例看一下具體實(shí)現(xiàn)

public class ExceptionCircuitBreaker extends AbstractCircuitBreaker {
    
    // 異常模式有兩種，異常率和異常數(shù)
    private final int strategy;
    // 最小請(qǐng)求數(shù)
    private final int minRequestAmount;
    // 閾值
    private final double threshold;
    
    // LeapArray是sentinel統(tǒng)計(jì)數(shù)據(jù)非常重要的一個(gè)結(jié)構(gòu)，主要封裝了時(shí)間窗口相關(guān)的操作
    private final LeapArray stat;

    public ExceptionCircuitBreaker(DegradeRule rule) {
        this(rule, new SimpleErrorCounterLeapArray(1, rule.getStatIntervalMs()));
    }

    ExceptionCircuitBreaker(DegradeRule rule, LeapArray stat) {
        super(rule);
        this.strategy = rule.getGrade();
        boolean modeOk = strategy == DEGRADE_GRADE_EXCEPTION_RATIO || strategy == DEGRADE_GRADE_EXCEPTION_COUNT;
        AssertUtil.isTrue(modeOk, "rule strategy should be error-ratio or error-count");
        AssertUtil.notNull(stat, "stat cannot be null");
        this.minRequestAmount = rule.getMinRequestAmount();
        this.threshold = rule.getCount();
        this.stat = stat;
    }

    @Override
    protected void resetStat() {
        // Reset current bucket (bucket count = 1).
        stat.currentWindow().value().reset();
    }

    
    @Override
    public void onRequestComplete(Context context) {
        Entry entry = context.getCurEntry();
        if (entry == null) {
            return;
        }
        Throwable error = entry.getError();
        SimpleErrorCounter counter = stat.currentWindow().value();
        if (error != null) {
            counter.getErrorCount().add(1);
        }
        counter.getTotalCount().add(1);

        handleStateChangeWhenThresholdExceeded(error);
    }

    private void handleStateChangeWhenThresholdExceeded(Throwable error) {
        if (currentState.get() == State.OPEN) {
            return;
        }
        
        if (currentState.get() == State.HALF_OPEN) {
            // In detecting request
            if (error == null) {
                fromHalfOpenToClose();
            } else {
                fromHalfOpenToOpen(1.0d);
            }
            return;
        }
        
        List counters = stat.values();
        long errCount = 0;
        long totalCount = 0;
        for (SimpleErrorCounter counter : counters) {
            
 += counter.errorCount.sum();
            totalCount += counter.totalCount.sum();
        }
        if (totalCount < minRequestAmount) {
            return;
        }
        double curCount = errCount;
        if (strategy == DEGRADE_GRADE_EXCEPTION_RATIO) {
            // Use errorRatio
            curCount = errCount * 1.0d / totalCount;
        }
        if (curCount > threshold) {
            transformToOpen(curCount);
        }
    }

    static class SimpleErrorCounter {
        private LongAdder errorCount;
        private LongAdder totalCount;

        public SimpleErrorCounter() {
            this.errorCount = new LongAdder();
            this.totalCount = new LongAdder();
        }

        public LongAdder getErrorCount() {
            return errorCount;
        }

        public LongAdder getTotalCount() {
            return totalCount;
        }

        public SimpleErrorCounter reset() {
            errorCount.reset();
            totalCount.reset();
            return this;
        }

        @Override
        public String toString() {
            return "SimpleErrorCounter{" +
                "errorCount=" + errorCount +
                ", totalCount=" + totalCount +
                '}';
        }
    }

    static class SimpleErrorCounterLeapArray extends LeapArray {

        public SimpleErrorCounterLeapArray(int sampleCount, int intervalInMs) {
            super(sampleCount, intervalInMs);
        }

        @Override
        public SimpleErrorCounter newEmptyBucket(long timeMillis) {
            return new SimpleErrorCounter();
        }

        @Override
        protected WindowWrap resetWindowTo(WindowWrap w, long startTime) {
            // Update the start time and reset value.
            w.resetTo(startTime);
            w.value().reset();
            return w;
        }
    }
}

2.6 SystemSlot

校驗(yàn)邏輯主要集中在com.alibaba.csp.sentinel.slots.system.SystemRuleManager#checkSystem，以下是片段，可以看到，作為負(fù)載保護(hù)規(guī)則校驗(yàn)，實(shí)現(xiàn)了集群的QPS、線程、RT（響應(yīng)時(shí)間）、系統(tǒng)負(fù)載的控制，除系統(tǒng)負(fù)載以外，其余統(tǒng)計(jì)都是依賴StatisticSlot實(shí)現(xiàn)，系統(tǒng)負(fù)載是通過(guò)SystemRuleManager定時(shí)調(diào)度SystemStatusListener，通過(guò)OperatingSystemMXBean去獲取

/**
     * Apply {@link SystemRule} to the resource. Only inbound traffic will be checked.
     *
     * @param resourceWrapper the resource.
     * @throws BlockException when any system rule's threshold is exceeded.
     */
    public static void checkSystem(ResourceWrapper resourceWrapper, int count) throws BlockException {
        if (resourceWrapper == null) {
            return;
        }
        // Ensure the checking switch is on.
        if (!checkSystemStatus.get()) {
            return;
        }

        // for inbound traffic only
        if (resourceWrapper.getEntryType() != EntryType.IN) {
            return;
        }

        // total qps 此處是拿到某個(gè)資源在集群中的QPS總和，相關(guān)概念可以會(huì)看初始化關(guān)于Node的介紹
        double currentQps = Constants.ENTRY_NODE.passQps();
        if (currentQps + count > qps) {
            throw new SystemBlockException(resourceWrapper.getName(), "qps");
        }

        // total thread 
        int currentThread = Constants.ENTRY_NODE.curThreadNum();
        if (currentThread > maxThread) {
            throw new SystemBlockException(resourceWrapper.getName(), "thread");
        }

        double rt = Constants.ENTRY_NODE.avgRt();
        if (rt > maxRt) {
            throw new SystemBlockException(resourceWrapper.getName(), "rt");
        }

        // load. BBR algorithm.
        if (highestSystemLoadIsSet && getCurrentSystemAvgLoad() > highestSystemLoad) {
            if (!checkBbr(currentThread)) {
                throw new SystemBlockException(resourceWrapper.getName(), "load");
            }
        }

        // cpu usage
        if (highestCpuUsageIsSet && getCurrentCpuUsage() > highestCpuUsage) {
            throw new SystemBlockException(resourceWrapper.getName(), "cpu");
        }
    }

    private static boolean checkBbr(int currentThread) {
        if (currentThread > 1 &&
            currentThread > Constants.ENTRY_NODE.maxSuccessQps() * Constants.ENTRY_NODE.minRt() / 1000) {
            return false;
        }
        return true;
    }

    public static double getCurrentSystemAvgLoad() {
        return statusListener.getSystemAverageLoad();
    }

    public static double getCurrentCpuUsage() {
        return statusListener.getCpuUsage();
    }


public class SystemStatusListener implements Runnable {

    volatile double currentLoad = -1;
    volatile double currentCpuUsage = -1;

    volatile String reason = StringUtil.EMPTY;

    volatile long processCpuTime = 0;
    volatile long processUpTime = 0;

    public double getSystemAverageLoad() {
        return currentLoad;
    }

    public double getCpuUsage() {
        return currentCpuUsage;
    }

    @Override
    public void run() {
        try {
            OperatingSystemMXBean osBean = ManagementFactory.getPlatformMXBean(OperatingSystemMXBean.class);
            currentLoad = osBean.getSystemLoadAverage();

            /*
             * Java Doc copied from {@link OperatingSystemMXBean#getSystemCpuLoad()}:
             * Returns the "recent cpu usage" for the whole system. This value is a double in the [0.0,1.0] interval.
             * A value of 0.0 means that all CPUs were idle during the recent period of time observed, while a value
             * of 1.0 means that all CPUs were actively running 100% of the time during the recent period being
             * observed. All values between 0.0 and 1.0 are possible depending of the activities going on in the
             * system. If the system recent cpu usage is not available, the method returns a negative value.
             */
            double systemCpuUsage = osBean.getSystemCpuLoad();

            // calculate process cpu usage to support application running in container environment
            RuntimeMXBean runtimeBean = ManagementFactory.getPlatformMXBean(RuntimeMXBean.class);
            long newProcessCpuTime = osBean.getProcessCpuTime();
            long newProcessUpTime = runtimeBean.getUptime();
            int cpuCores = osBean.getAvailableProcessors();
            long processCpuTimeDiffInMs = TimeUnit.NANOSECONDS
                    .toMillis(newProcessCpuTime - processCpuTime);
            long processUpTimeDiffInMs = newProcessUpTime - processUpTime;
            double processCpuUsage = (double) processCpuTimeDiffInMs / processUpTimeDiffInMs / cpuCores;
            processCpuTime = newProcessCpuTime;
            processUpTime = newProcessUpTime;

            currentCpuUsage = Math.max(processCpuUsage, systemCpuUsage);

            if (currentLoad > SystemRuleManager.getSystemLoadThreshold()) {
                writeSystemStatusLog();
            }
        } catch (Throwable e) {
            RecordLog.warn("[SystemStatusListener] Failed to get system metrics from JMX", e);
        }
    }

    private void writeSystemStatusLog() {
        StringBuilder sb = new StringBuilder();
        sb.append("Load exceeds the threshold: ");
        sb.append("load:").append(String.format("%.4f", currentLoad)).append("; ");
        sb.append("cpuUsage:").append(String.format("%.4f", currentCpuUsage)).append("; ");
        sb.append("qps:").append(String.format("%.4f", Constants.ENTRY_NODE.passQps())).append("; ");
        sb.append("rt:").append(String.format("%.4f", Constants.ENTRY_NODE.avgRt())).append("; ");
        sb.append("thread:").append(Constants.ENTRY_NODE.curThreadNum()).append("; ");
        sb.append("success:").append(String.format("%.4f", Constants.ENTRY_NODE.successQps())).append("; ");
        sb.append("minRt:").append(String.format("%.2f", Constants.ENTRY_NODE.minRt())).append("; ");
        sb.append("maxSuccess:").append(String.format("%.2f", Constants.ENTRY_NODE.maxSuccessQps())).append("; ");
        RecordLog.info(sb.toString());
    }
}

三、京東版最佳實(shí)踐

3.1 使用方式

Sentinel使用方式本身非常簡(jiǎn)單，就是一個(gè)注解，但是要考慮規(guī)則加載和規(guī)則持久化的方式，現(xiàn)有的方式有：

?使用Sentinel-dashboard功能：使用面板接入需要維護(hù)一個(gè)配置規(guī)則的管理端，考慮到偏后端的系統(tǒng)需要額外維護(hù)一個(gè)面板成本較大，如果是像RPC框架這種本身有管理端的接入可以考慮次方案。

?中間件（如：zookepper、nacos、eureka、redis等）：Sentinel源碼extension包里提供了類似的實(shí)現(xiàn)，如下圖

結(jié)合京東實(shí)際，我實(shí)現(xiàn)了一個(gè)規(guī)則熱部署的Sentinel組件，實(shí)現(xiàn)方式類似zookeeper的方式，將規(guī)則記錄到ducc的一個(gè)key上，在spring容器啟動(dòng)時(shí)做第一次規(guī)則加載和監(jiān)聽(tīng)器注冊(cè)，組件也做一了一些規(guī)則讀取，校驗(yàn)、實(shí)例化不同規(guī)則對(duì)象的工作

插件使用方式：注解+配置

第一步 引入組件

    com.jd.ldop.tools
    sentinel-tools
    1.0.0-SNAPSHOT

第二步 初始化sentinelProcess

支持ducc、本地文件讀取、直接寫(xiě)入三種方式規(guī)則寫(xiě)入方式

目前支持限流規(guī)則、熔斷降級(jí)規(guī)則兩種模式，系統(tǒng)負(fù)載保護(hù)模式待開(kāi)發(fā)和驗(yàn)證

ducc上配置如下：

第三步 定義資源和關(guān)聯(lián)類型

通過(guò)@SentinelResource可以直接在任意位置定義資源名以及對(duì)應(yīng)的熔斷降級(jí)或者限流方式、回調(diào)方法等，同時(shí)也可以指定關(guān)聯(lián)類型，支持直接、關(guān)聯(lián)、指定鏈路三種

    @Override
    @SentinelResource(value = "modifyGetWaybillState", fallback = "executeDegrade")
    public ExecutionResult> execute(@NotNull Model imodel) {
        // 業(yè)務(wù)邏輯處理
    }

    public ExecutionResult> executeDegrade(@NotNull Model imodel) {
        // 降級(jí)業(yè)務(wù)邏輯處理
    }

3.2 應(yīng)用場(chǎng)景

組件支持任意的業(yè)務(wù)降級(jí)、限流、負(fù)載保護(hù)

四、Sentinel壓測(cè)數(shù)據(jù)

4.1 壓測(cè)目標(biāo)

調(diào)用量：1.2W/m

應(yīng)用機(jī)器內(nèi)存穩(wěn)定在50%以內(nèi)

機(jī)器規(guī)格： 8C16G50G磁盤(pán)*2

Sentinel降級(jí)規(guī)則：

count=350-------慢調(diào)用臨界閾值350ms

timeWindow=180------熔斷時(shí)間窗口180s

grade=0-----降級(jí)模式 慢調(diào)用

statIntervalMs=60000------統(tǒng)計(jì)時(shí)長(zhǎng)1min

4.2 壓測(cè)結(jié)果

應(yīng)用機(jī)器監(jiān)控：

壓測(cè)分為了兩個(gè)階段，分別是組件開(kāi)啟和組件關(guān)閉兩次，前半部分是組件開(kāi)啟的情況，后半部分是組件關(guān)閉的情況

應(yīng)用進(jìn)程內(nèi)存分析,和sentinel有關(guān)的前三對(duì)象是

com.alibaba.csp.sentinel.node.metric.MetricNode

com.alibaba.csp.sentinel.CtEntry

com.alibaba.csp.sentinel.context.Context

4.3 壓測(cè)結(jié)論

使Sentinel組件實(shí)現(xiàn)系統(tǒng)服務(wù)自動(dòng)降級(jí)或限流，由于sentinel會(huì)按照滑動(dòng)窗口周期性統(tǒng)計(jì)數(shù)據(jù)，因此會(huì)占用一定的機(jī)器內(nèi)存，使用時(shí)應(yīng)設(shè)置合理的規(guī)則，如：合理的統(tǒng)計(jì)時(shí)長(zhǎng)、避免過(guò)多的Sentinel資源創(chuàng)建等。

總體來(lái)說(shuō)，使用sentinel組件對(duì)應(yīng)用cpu和內(nèi)存影響不大。

審核編輯 黃宇