Sentinel源码—7.参数限流和注解的实现一

大纲

1.参数限流的原理和源码

2.@SentinelResource注解的使用和实现

1.参数限流的原理和源码

(1)参数限流规则ParamFlowRule的配置Demo

(2)ParamFlowSlot根据参数限流规则验证请求

(1)参数限流规则ParamFlowRule的配置Demo

一.参数限流的应用场景

二.参数限流规则的属性

三.参数限流规则的配置Demo

一.参数限流的应用场景

传统的流量控制，一般是通过资源维度来限制某接口或方法的调用频率。但有时需要更细粒度地控制不同参数条件下的访问速率，即参数限流。参数限流允许根据不同的参数条件设置不同的流量控制规则，这种方式非常适合处理特定条件下的请求，因为能更加精细地管理流量。

假设有一个在线电影订票系统，某个接口允许用户查询电影的放映时间。但只希望每个用户每10秒只能查询接口1次，以避免过多的查询请求。这时如果直接将接口的QPS限制为5是不能满足要求的，因为需求是每个用户每5分钟只能查询1次，而不是每秒一共只能查询5次，因此参数限流就能派上用场了。

可以设置一个规则，根据用户ID来限制每个用户的查询频率，将限流的维度从资源维度细化到参数维度，从而实现每个用户每10秒只能查询接口1次。比如希望影院工作人员可以每秒查询10次，老板可以每秒查询100次，而购票者则只能每10秒查询一次，其中工作人员的userId值为100和200，老板的userId值为9999，那么可以如下配置：需要注意限流阈值是以秒为单位的，所以需要乘以统计窗口时长10。

二.参数限流规则的属性

public class ParamFlowRule extends AbstractRule {...//The threshold type of flow control (0: thread count, 1: QPS).//流量控制的阈值类型(0表示线程数，1表示QPS)private int grade = RuleConstant.FLOW_GRADE_QPS;//Parameter index.//参数下标private Integer paramIdx;//The threshold count.//阈值private double count;//Original exclusion items of parameters.//针对特定参数的流量控制规则列表private List<ParamFlowItem> paramFlowItemList = new ArrayList<ParamFlowItem>();//Indicating whether the rule is for cluster mode.//是否集群private boolean clusterMode = false;...
}//针对特定参数的流量控制规则
public class ParamFlowItem {private String object;private Integer count;private String classType;...
}

三.参数限流规则的配置Demo

//This demo demonstrates flow control by frequent ("hot spot") parameters.
public class ParamFlowQpsDemo {private static final int PARAM_A = 1;private static final int PARAM_B = 2;private static final int PARAM_C = 3;private static final int PARAM_D = 4;//Here we prepare different parameters to validate flow control by parameters.private static final Integer[] PARAMS = new Integer[] {PARAM_A, PARAM_B, PARAM_C, PARAM_D};private static final String RESOURCE_KEY = "resA";public static void main(String[] args) throws Exception {initParamFlowRules();final int threadCount = 20;ParamFlowQpsRunner<Integer> runner = new ParamFlowQpsRunner<>(PARAMS, RESOURCE_KEY, threadCount, 120);runner.tick();Thread.sleep(1000);runner.simulateTraffic();}private static void initParamFlowRules() {//QPS mode, threshold is 5 for every frequent "hot spot" parameter in index 0 (the first arg).ParamFlowRule rule = new ParamFlowRule(RESOURCE_KEY).setParamIdx(0).setGrade(RuleConstant.FLOW_GRADE_QPS).setCount(5);//We can set threshold count for specific parameter value individually.//Here we add an exception item. That means: //QPS threshold of entries with parameter `PARAM_B` (type: int) in index 0 will be 10, rather than the global threshold (5).ParamFlowItem item = new ParamFlowItem().setObject(String.valueOf(PARAM_B)).setClassType(int.class.getName()).setCount(10);rule.setParamFlowItemList(Collections.singletonList(item));//ParamFlowRuleManager类加载的一个时机是：它的静态方法被调用了//所以下面会先初始化ParamFlowRuleManager，再执行loadRules()方法ParamFlowRuleManager.loadRules(Collections.singletonList(rule));}
}public final class ParamFlowRuleManager {private static final Map<String, List<ParamFlowRule>> PARAM_FLOW_RULES = new ConcurrentHashMap<>();private final static RulePropertyListener PROPERTY_LISTENER = new RulePropertyListener();private static SentinelProperty<List<ParamFlowRule>> currentProperty = new DynamicSentinelProperty<>();static {currentProperty.addListener(PROPERTY_LISTENER);}//Load parameter flow rules. Former rules will be replaced.public static void loadRules(List<ParamFlowRule> rules) {try {//设置规则的值为rulescurrentProperty.updateValue(rules);} catch (Throwable e) {RecordLog.info("[ParamFlowRuleManager] Failed to load rules", e);}}static class RulePropertyListener implements PropertyListener<List<ParamFlowRule>> {@Overridepublic void configUpdate(List<ParamFlowRule> list) {Map<String, List<ParamFlowRule>> rules = aggregateAndPrepareParamRules(list);if (rules != null) {PARAM_FLOW_RULES.clear();PARAM_FLOW_RULES.putAll(rules);}RecordLog.info("[ParamFlowRuleManager] Parameter flow rules received: {}", PARAM_FLOW_RULES);}@Overridepublic void configLoad(List<ParamFlowRule> list) {Map<String, List<ParamFlowRule>> rules = aggregateAndPrepareParamRules(list);if (rules != null) {PARAM_FLOW_RULES.clear();PARAM_FLOW_RULES.putAll(rules);}RecordLog.info("[ParamFlowRuleManager] Parameter flow rules received: {}", PARAM_FLOW_RULES);}...}...
}public class DynamicSentinelProperty<T> implements SentinelProperty<T> {protected Set<PropertyListener<T>> listeners = new CopyOnWriteArraySet<>();private T value = null;public DynamicSentinelProperty() {}//添加监听器到集合@Overridepublic void addListener(PropertyListener<T> listener) {listeners.add(listener);//回调监听器的configLoad()方法初始化规则配置listener.configLoad(value);}//更新值@Overridepublic boolean updateValue(T newValue) {//如果值没变化，直接返回if (isEqual(value, newValue)) {return false;}RecordLog.info("[DynamicSentinelProperty] Config will be updated to: {}", newValue);//如果值发生了变化，则遍历监听器，回调监听器的configUpdate()方法更新对应的值value = newValue;for (PropertyListener<T> listener : listeners) {listener.configUpdate(newValue);}return true;}...
}//A traffic runner to simulate flow for different parameters.
class ParamFlowQpsRunner<T> {private final T[] params;private final String resourceName;private int seconds;private final int threadCount;private final Map<T, AtomicLong> passCountMap = new ConcurrentHashMap<>();private final Map<T, AtomicLong> blockCountMap = new ConcurrentHashMap<>();private volatile boolean stop = false;public ParamFlowQpsRunner(T[] params, String resourceName, int threadCount, int seconds) {this.params = params;this.resourceName = resourceName;this.seconds = seconds;this.threadCount = threadCount;for (T param : params) {passCountMap.putIfAbsent(param, new AtomicLong());blockCountMap.putIfAbsent(param, new AtomicLong());}}public void tick() {Thread timer = new Thread(new TimerTask());timer.setName("sentinel-timer-task");timer.start();}public void simulateTraffic() {for (int i = 0; i < threadCount; i++) {Thread t = new Thread(new RunTask());t.setName("sentinel-simulate-traffic-task-" + i);t.start();}}final class TimerTask implements Runnable {@Overridepublic void run() {long start = System.currentTimeMillis();System.out.println("Begin to run! Go go go!");System.out.println("See corresponding metrics.log for accurate statistic data");Map<T, Long> map = new HashMap<>(params.length);for (T param : params) {map.putIfAbsent(param, 0L);}while (!stop) {sleep(1000);//There may be a mismatch for time window of internal sliding window.//See corresponding `metrics.log` for accurate statistic log.for (T param : params) {System.out.println(String.format("[%d][%d] Parameter flow metrics for resource %s: pass count for param <%s> is %d, block count: %d",seconds, TimeUtil.currentTimeMillis(), resourceName, param,passCountMap.get(param).getAndSet(0), blockCountMap.get(param).getAndSet(0)));}System.out.println("=============================");if (seconds-- <= 0) {stop = true;}}long cost = System.currentTimeMillis() - start;System.out.println("Time cost: " + cost + " ms");System.exit(0);}}final class RunTask implements Runnable {@Overridepublic void run() {while (!stop) {Entry entry = null;T param = generateParam();try {entry = SphU.entry(resourceName, EntryType.IN, 1, param);//Add pass for parameter.passFor(param);} catch (BlockException e) {//block.incrementAndGet();blockFor(param);} catch (Exception ex) {//biz exceptionex.printStackTrace();} finally {//total.incrementAndGet();if (entry != null) {entry.exit(1, param);}}sleep(ThreadLocalRandom.current().nextInt(0, 10));}}}//Pick one of provided parameters randomly.private T generateParam() {int i = ThreadLocalRandom.current().nextInt(0, params.length);return params[i];}private void passFor(T param) {passCountMap.get(param).incrementAndGet();}private void blockFor(T param) {blockCountMap.get(param).incrementAndGet();}private void sleep(int timeMs) {try {TimeUnit.MILLISECONDS.sleep(timeMs);} catch (InterruptedException e) {}}
}

(2)ParamFlowSlot根据参数限流规则验证请求

一.ParamFlowSlot的entry()方法的逻辑

二.不同限流类型 + 阈值类型 + 流控效果的处理

三.流控效果为排队等待和直接拒绝的实现

四.参数限流是如何进行数据统计

五.参数限流验证请求的流程图总结

一.ParamFlowSlot的entry()方法的逻辑

ParamFlowSlot的entry()方法主要干了三件事：参数验证、获取当前资源的全部参数限流规则、循环每一个参数限流规则并判断此次请求是否被允许通过(如果不允许则直接抛出异常)。其中对每一条获取到的参数限流规则，都会通过ParamFlowChecker的passCheck()方法进行判断。

@Spi(order = -3000)
public class ParamFlowSlot extends AbstractLinkedProcessorSlot<DefaultNode> {@Overridepublic void entry(Context context, ResourceWrapper resourceWrapper, DefaultNode node, int count,boolean prioritized, Object... args) throws Throwable {//1.如果没配置参数限流规则，直接触发下一个Slotif (!ParamFlowRuleManager.hasRules(resourceWrapper.getName())) {fireEntry(context, resourceWrapper, node, count, prioritized, args);return;}//2.如果配置了参数限流规则，则调用ParamFlowSlot的checkFlow()方法，该方法执行完成后再触发下一个SlotcheckFlow(resourceWrapper, count, args);fireEntry(context, resourceWrapper, node, count, prioritized, args);}@Overridepublic void exit(Context context, ResourceWrapper resourceWrapper, int count, Object... args) {fireExit(context, resourceWrapper, count, args);}void applyRealParamIdx(/*@NonNull*/ ParamFlowRule rule, int length) {int paramIdx = rule.getParamIdx();if (paramIdx < 0) {if (-paramIdx <= length) {rule.setParamIdx(length + paramIdx);} else {//Illegal index, give it a illegal positive value, latter rule checking will pass.rule.setParamIdx(-paramIdx);}}}void checkFlow(ResourceWrapper resourceWrapper, int count, Object... args) throws BlockException {//1.如果没传递参数，则直接放行，代表不做参数限流逻辑if (args == null) {return;}//2.如果没给resourceWrapper这个资源配置参数限流规则，则直接放行if (!ParamFlowRuleManager.hasRules(resourceWrapper.getName())) {return;}//3.获取此资源的全部参数限流规则，规则可能会有很多个，所以是个ListList<ParamFlowRule> rules = ParamFlowRuleManager.getRulesOfResource(resourceWrapper.getName());//4.遍历获取到的参数限流规则for (ParamFlowRule rule : rules) {//进行参数验证applyRealParamIdx(rule, args.length);//Initialize the parameter metrics.ParameterMetricStorage.initParamMetricsFor(resourceWrapper, rule);//进行验证的核心方法：检查当前规则是否允许通过此请求，如果不允许，则抛出ParamFlowException异常if (!ParamFlowChecker.passCheck(resourceWrapper, rule, count, args)) {String triggeredParam = "";if (args.length > rule.getParamIdx()) {Object value = args[rule.getParamIdx()];//Assign actual value with the result of paramFlowKey methodif (value instanceof ParamFlowArgument) {value = ((ParamFlowArgument) value).paramFlowKey();}triggeredParam = String.valueOf(value);}throw new ParamFlowException(resourceWrapper.getName(), triggeredParam, rule);}}}
}

二.不同限流类型 + 阈值类型 + 流控效果的处理

在ParamFlowChecker的passCheck()方法中，参数值验证通过之后，会判断限流类型。如果是集群限流，则执行ParamFlowChecker的passClusterCheck()方法。如果是单机限流，则执行ParamFlowChecker的passLocalCheck()方法。

在ParamFlowChecker的passLocalCheck()方法中，则会根据不同的参数类型调用ParamFlowChecker的passSingleValueCheck()方法。根据该方法可以知道，参数限流支持两种阈值类型：一种是QPS，另一种是线程数。而QPS类型还支持两种流控效果，分别是排队等待和直接拒绝，但不支持Warm Up。

//Rule checker for parameter flow control.
public final class ParamFlowChecker {public static boolean passCheck(ResourceWrapper resourceWrapper, /*@Valid*/ ParamFlowRule rule, /*@Valid*/ int count, Object... args) {if (args == null) {return true;}//1.判断参数索引是否合法，这个就是配置参数限流时设置的下标，从0开始，也就是对应args里的下标//比如0就代表args数组里的第一个参数，如果参数不合法直接放行，相当于参数限流没生效  int paramIdx = rule.getParamIdx();if (args.length <= paramIdx) {return true;}//2.判断参数值是不是空，如果是空直接放行//Get parameter value.Object value = args[paramIdx];//Assign value with the result of paramFlowKey methodif (value instanceof ParamFlowArgument) {value = ((ParamFlowArgument) value).paramFlowKey();}//If value is null, then passif (value == null) {return true;}//3.集群限流if (rule.isClusterMode() && rule.getGrade() == RuleConstant.FLOW_GRADE_QPS) {return passClusterCheck(resourceWrapper, rule, count, value);}//4.单机限流return passLocalCheck(resourceWrapper, rule, count, value);}private static boolean passLocalCheck(ResourceWrapper resourceWrapper, ParamFlowRule rule, int count, Object value) {try {if (Collection.class.isAssignableFrom(value.getClass())) {//基本类型for (Object param : ((Collection)value)) {if (!passSingleValueCheck(resourceWrapper, rule, count, param)) {return false;}}} else if (value.getClass().isArray()) {//数组类型int length = Array.getLength(value);for (int i = 0; i < length; i++) {Object param = Array.get(value, i);if (!passSingleValueCheck(resourceWrapper, rule, count, param)) {return false;}}} else {//其他类型，也就是引用类型return passSingleValueCheck(resourceWrapper, rule, count, value);}} catch (Throwable e) {RecordLog.warn("[ParamFlowChecker] Unexpected error", e);}return true;}static boolean passSingleValueCheck(ResourceWrapper resourceWrapper, ParamFlowRule rule, int acquireCount, Object value) {if (rule.getGrade() == RuleConstant.FLOW_GRADE_QPS) {//类型是QPSif (rule.getControlBehavior() == RuleConstant.CONTROL_BEHAVIOR_RATE_LIMITER) {//流控效果为排队等待return passThrottleLocalCheck(resourceWrapper, rule, acquireCount, value);} else {//流控效果为直接拒绝return passDefaultLocalCheck(resourceWrapper, rule, acquireCount, value);}} else if (rule.getGrade() == RuleConstant.FLOW_GRADE_THREAD) {//类型是ThreadSet<Object> exclusionItems = rule.getParsedHotItems().keySet();long threadCount = getParameterMetric(resourceWrapper).getThreadCount(rule.getParamIdx(), value);if (exclusionItems.contains(value)) {int itemThreshold = rule.getParsedHotItems().get(value);return ++threadCount <= itemThreshold;}long threshold = (long)rule.getCount();return ++threadCount <= threshold;}return true;}...
}

三.流控效果为排队等待和直接拒绝的实现

当设置了QPS类型的流控效果为排队等待时，会调用ParamFlowChecker的passThrottleLocalCheck()方法。该方法实现排队等待效果的原理和流控规则FlowSlot通过RateLimiterController实现排队等待效果的原理是一样的。

//Rule checker for parameter flow control.
public final class ParamFlowChecker {...static boolean passThrottleLocalCheck(ResourceWrapper resourceWrapper, ParamFlowRule rule, int acquireCount, Object value) {ParameterMetric metric = getParameterMetric(resourceWrapper);CacheMap<Object, AtomicLong> timeRecorderMap = metric == null ? null : metric.getRuleTimeCounter(rule);if (timeRecorderMap == null) {return true;}//Calculate max token count (threshold)Set<Object> exclusionItems = rule.getParsedHotItems().keySet();long tokenCount = (long)rule.getCount();if (exclusionItems.contains(value)) {tokenCount = rule.getParsedHotItems().get(value);}if (tokenCount == 0) {return false;}long costTime = Math.round(1.0 * 1000 * acquireCount * rule.getDurationInSec() / tokenCount);while (true) {long currentTime = TimeUtil.currentTimeMillis();AtomicLong timeRecorder = timeRecorderMap.putIfAbsent(value, new AtomicLong(currentTime));if (timeRecorder == null) {return true;}//AtomicLong timeRecorder = timeRecorderMap.get(value);long lastPassTime = timeRecorder.get();long expectedTime = lastPassTime + costTime;if (expectedTime <= currentTime || expectedTime - currentTime < rule.getMaxQueueingTimeMs()) {AtomicLong lastPastTimeRef = timeRecorderMap.get(value);if (lastPastTimeRef.compareAndSet(lastPassTime, currentTime)) {long waitTime = expectedTime - currentTime;if (waitTime > 0) {lastPastTimeRef.set(expectedTime);try {TimeUnit.MILLISECONDS.sleep(waitTime);} catch (InterruptedException e) {RecordLog.warn("passThrottleLocalCheck: wait interrupted", e);}}return true;} else {Thread.yield();}} else {return false;}}}private static ParameterMetric getParameterMetric(ResourceWrapper resourceWrapper) {//Should not be null.return ParameterMetricStorage.getParamMetric(resourceWrapper);}
}

当设置了QPS类型的流控效果为直接拒绝时，会调用ParamFlowChecker的passDefaultLocalCheck()方法。该方法采取令牌桶的方式来实现：控制每个时间窗口只生产一次token令牌，且将令牌放入桶中，每个请求都从桶中取令牌，当可以获取到令牌时，则正常放行，反之直接拒绝。

//Rule checker for parameter flow control.
public final class ParamFlowChecker {...static boolean passDefaultLocalCheck(ResourceWrapper resourceWrapper, ParamFlowRule rule, int acquireCount, Object value) {ParameterMetric metric = getParameterMetric(resourceWrapper);CacheMap<Object, AtomicLong> tokenCounters = metric == null ? null : metric.getRuleTokenCounter(rule);CacheMap<Object, AtomicLong> timeCounters = metric == null ? null : metric.getRuleTimeCounter(rule);if (tokenCounters == null || timeCounters == null) {return true;}//Calculate max token count (threshold)Set<Object> exclusionItems = rule.getParsedHotItems().keySet();long tokenCount = (long)rule.getCount();if (exclusionItems.contains(value)) {tokenCount = rule.getParsedHotItems().get(value);}if (tokenCount == 0) {return false;}long maxCount = tokenCount + rule.getBurstCount();if (acquireCount > maxCount) {return false;}while (true) {long currentTime = TimeUtil.currentTimeMillis();AtomicLong lastAddTokenTime = timeCounters.putIfAbsent(value, new AtomicLong(currentTime));if (lastAddTokenTime == null) {//Token never added, just replenish the tokens and consume {@code acquireCount} immediately.tokenCounters.putIfAbsent(value, new AtomicLong(maxCount - acquireCount));return true;}//Calculate the time duration since last token was added.long passTime = currentTime - lastAddTokenTime.get();//A simplified token bucket algorithm that will replenish the tokens only when statistic window has passed.if (passTime > rule.getDurationInSec() * 1000) {AtomicLong oldQps = tokenCounters.putIfAbsent(value, new AtomicLong(maxCount - acquireCount));if (oldQps == null) {//Might not be accurate here.lastAddTokenTime.set(currentTime);return true;} else {long restQps = oldQps.get();long toAddCount = (passTime * tokenCount) / (rule.getDurationInSec() * 1000);long newQps = toAddCount + restQps > maxCount ? (maxCount - acquireCount): (restQps + toAddCount - acquireCount);if (newQps < 0) {return false;}if (oldQps.compareAndSet(restQps, newQps)) {lastAddTokenTime.set(currentTime);return true;}Thread.yield();}} else {AtomicLong oldQps = tokenCounters.get(value);if (oldQps != null) {long oldQpsValue = oldQps.get();if (oldQpsValue - acquireCount >= 0) {if (oldQps.compareAndSet(oldQpsValue, oldQpsValue - acquireCount)) {return true;}} else {return false;}}Thread.yield();}}}
}

四.参数限流是如何进行数据统计

由于参数限流的数据统计需要细化到参数值的维度，所以使用参数限流时需要注意OOM问题。比如根据用户ID进行限流，且用户数量有几千万，那么CacheMap将会包含几千万个不会被移除的键值对，而且会随着进程运行时间的增长而不断增加，最后可能会导致OOM。

public final class ParameterMetricStorage {private static final Map<String, ParameterMetric> metricsMap = new ConcurrentHashMap<>();//Lock for a specific resource.private static final Object LOCK = new Object();//Init the parameter metric and index map for given resource.//该方法在ParamFlowSlot的checkFlow()方法中被调用public static void initParamMetricsFor(ResourceWrapper resourceWrapper, /*@Valid*/ ParamFlowRule rule) {if (resourceWrapper == null || resourceWrapper.getName() == null) {return;}String resourceName = resourceWrapper.getName();ParameterMetric metric;//Assume that the resource is valid.if ((metric = metricsMap.get(resourceName)) == null) {synchronized (LOCK) {if ((metric = metricsMap.get(resourceName)) == null) {metric = new ParameterMetric();metricsMap.put(resourceWrapper.getName(), metric);RecordLog.info("[ParameterMetricStorage] Creating parameter metric for: {}", resourceWrapper.getName());}}}metric.initialize(rule);}//该方法在ParamFlowChecker的passThrottleLocalCheck()和passDefaultLocalCheck()方法执行getParameterMetric()方法时被调用public static ParameterMetric getParamMetric(ResourceWrapper resourceWrapper) {if (resourceWrapper == null || resourceWrapper.getName() == null) {return null;}return metricsMap.get(resourceWrapper.getName());}...
}//Metrics for frequent ("hot spot") parameters.
public class ParameterMetric {private static final int THREAD_COUNT_MAX_CAPACITY = 4000;private static final int BASE_PARAM_MAX_CAPACITY = 4000;private static final int TOTAL_MAX_CAPACITY = 20_0000;private final Object lock = new Object();//Format: (rule, (value, timeRecorder))private final Map<ParamFlowRule, CacheMap<Object, AtomicLong>> ruleTimeCounters = new HashMap<>();//Format: (rule, (value, tokenCounter))private final Map<ParamFlowRule, CacheMap<Object, AtomicLong>> ruleTokenCounter = new HashMap<>();private final Map<Integer, CacheMap<Object, AtomicInteger>> threadCountMap = new HashMap<>();public void initialize(ParamFlowRule rule) {if (!ruleTimeCounters.containsKey(rule)) {synchronized (lock) {if (ruleTimeCounters.get(rule) == null) {long size = Math.min(BASE_PARAM_MAX_CAPACITY * rule.getDurationInSec(), TOTAL_MAX_CAPACITY);ruleTimeCounters.put(rule, new ConcurrentLinkedHashMapWrapper<Object, AtomicLong>(size));}}}if (!ruleTokenCounter.containsKey(rule)) {synchronized (lock) {if (ruleTokenCounter.get(rule) == null) {long size = Math.min(BASE_PARAM_MAX_CAPACITY * rule.getDurationInSec(), TOTAL_MAX_CAPACITY);ruleTokenCounter.put(rule, new ConcurrentLinkedHashMapWrapper<Object, AtomicLong>(size));}}}if (!threadCountMap.containsKey(rule.getParamIdx())) {synchronized (lock) {if (threadCountMap.get(rule.getParamIdx()) == null) {threadCountMap.put(rule.getParamIdx(), new ConcurrentLinkedHashMapWrapper<Object, AtomicInteger>(THREAD_COUNT_MAX_CAPACITY));}}}}//Get the token counter for given parameter rule.//@param rule valid parameter rule//@return the associated token counterpublic CacheMap<Object, AtomicLong> getRuleTokenCounter(ParamFlowRule rule) {return ruleTokenCounter.get(rule);}//Get the time record counter for given parameter rule.//@param rule valid parameter rule//@return the associated time counterpublic CacheMap<Object, AtomicLong> getRuleTimeCounter(ParamFlowRule rule) {return ruleTimeCounters.get(rule);}public long getThreadCount(int index, Object value) {CacheMap<Object, AtomicInteger> cacheMap = threadCountMap.get(index);if (cacheMap == null) {return 0;}AtomicInteger count = cacheMap.get(value);return count == null ? 0L : count.get();}...
}

五.参数限流验证请求的流程图总结

本文来自互联网用户投稿，该文观点仅代表作者本人，不代表本站立场。本站仅提供信息存储空间服务，不拥有所有权，不承担相关法律责任。如若转载，请注明出处：http://www.mzph.cn/news/902287.shtml

如若内容造成侵权/违法违规/事实不符，请联系多彩编程网进行投诉反馈email:809451989@qq.com，一经查实，立即删除！

相关文章

多数据源配置（MyBatis-Plus vs AbstractRoutingDataSource）

多数据源配置（MyBatis-Plus vs AbstractRoutingDataSource）

MyBatis-Plus vs AbstractRoutingDataSource MyBatis-Plus多数据源配 1.添加依赖 <dependency><groupId>com.baomidou</groupId><artifactId>dynamic-datasource-spring-boot-starter</artifactId><version>3.5.0</version> <…
阅读更多...
聊透多线程编程-线程互斥与同步-13. C# Mutex类实现线程互斥

聊透多线程编程-线程互斥与同步-13. C# Mutex类实现线程互斥

目录 一、什么是临界区&#xff1f; 二、Mutex类简介 三、Mutex的基本用法 解释&#xff1a; 四、Mutex的工作原理 五、使用示例1-保护共享资源 解释&#xff1a; 六、使用示例2-跨进程同步 示例场景 1. 进程A - 主进程 2. 进程B - 第二个进程 输出结果 ProcessA …
阅读更多...
stm32week12

stm32week12

stm32学习 九.stm32与HAL库 2.HAL库框架 总架构&#xff1a; 文件介绍&#xff1a; ppp是某一外设&#xff0c;ex是拓展功能 HAL库API函数和变量命名规则&#xff1a; HAL库对寄存器位操作的相关宏定义&#xff1a; HAL库的回调函数&#xff1a; 3.STM32启动过程 MDK编译过…
阅读更多...
opencv HSV的具体描述

opencv HSV的具体描述

色调H&#xff1a; 使用角度度量&#xff0c;取值范围为0\~360&#xff0c;从红色开始按逆时针方向计算&#xff0c;红色为0&#xff0c;绿色为120&#xff0c;蓝色为240。它们的补色是&#xff1a;黄色为60&#xff0c;青色为180&#xff0c;紫色为300。通过改变H的值&#x…
阅读更多...
Java Lambda表达式指南

Java Lambda表达式指南

一、Lambda表达式基础 1. 什么是Lambda表达式&#xff1f; 匿名函数&#xff1a;没有名称的函数函数式编程&#xff1a;可作为参数传递的代码块简洁语法&#xff1a;替代匿名内部类的更紧凑写法 2. 基本语法 (parameters) -> expression 或 (parameters) -> { statem…
阅读更多...
面向对象设计中的类的分类：实体类、控制类和边界类

面向对象设计中的类的分类：实体类、控制类和边界类

目录 前言1. 实体类&#xff08;Entity Class&#xff09;1.1 定义和作用1.2 实体类的特点1.3 实体类的示例 2. 控制类&#xff08;Control Class&#xff09;2.1 定义和作用2.2 控制类的特点2.3 控制类的示例 3. 边界类&#xff08;Boundary Class&#xff09;3.1 定义和作用3…
阅读更多...
C# 封装教程

C# 封装教程

原文&#xff1a;C# 封装_w3cschool &#xff08;注&#xff1a;本文为教程文章&#xff0c;请勿标记为付费文章&#xff01;特此声明&#xff09; 封装 被定义为"把一个或多个项目封闭在一个物理的或者逻辑的包中"。在面向对象程序设计方法论中&#xff0c;封装是…
阅读更多...
量化交易 - RSRS（阻力支撑相对强度）- 正确用法 - 年均收益18%

量化交易 - RSRS（阻力支撑相对强度）- 正确用法 - 年均收益18%

经过研究&#xff0c;发现RSRS的正确用法其实是需要用到两个数据&#xff0c;分别是 n: 一阶拟合样本数&#xff0c;m:求均值方差样本数&#xff0c;其中n比较小 如18&#xff0c;m比较大 如1100 经过调优后&#xff0c;收益率显著上升&#xff01; 如下图&#xff1a; &…
阅读更多...
Oracle expdp的 EXCLUDE 参数详解

Oracle expdp的 EXCLUDE 参数详解

Oracle expdp的 EXCLUDE 参数详解 EXCLUDE 是 Oracle Data Pump Export (expdp) 工具中的一个关键参数&#xff0c;用于指定在导出过程中要排除的对象或对象类型。 一、基本语法 expdp username/password DUMPFILEexport.dmp DIRECTORYdpump_dir EXCLUDEobject_type[:name_c…
阅读更多...
如何使用3DMAX插件PFSpliner将3D对象转化为艺术样条线？

如何使用3DMAX插件PFSpliner将3D对象转化为艺术样条线？

什么是粒子流源（Particle Flow）是3DMAX的一个功能极其强大的粒子系统。它采用事件驱动模型，使用一个名为“粒子视图”的特殊对话框。在“粒子视图”中，您可以将描述粒子属性（如形状、速度、方向和一段时间内的旋转）的单个运算符组合成称为事件的组。每个操作符都提供一组…
阅读更多...
【python】 循环语句（while）

【python】 循环语句（while）

1、循环语句 语法&#xff1a; while 条件:......... #只有条件为真时&#xff0c;才会执行while中的内容。 1.1循环语句基本使用 示例1&#xff1a; print("开始") while 1>2:print("人生得意须尽欢") print("结束") #输出结果&#…
阅读更多...
OOA-CNN-LSTM-Attention、CNN-LSTM-Attention、OOA-CNN-LSTM、CNN-LSTM四模型多变量时序预测一键对比

OOA-CNN-LSTM-Attention、CNN-LSTM-Attention、OOA-CNN-LSTM、CNN-LSTM四模型多变量时序预测一键对比

OOA-CNN-LSTM-Attention、CNN-LSTM-Attention、OOA-CNN-LSTM、CNN-LSTM四模型多变量时序预测一键对比 目录 OOA-CNN-LSTM-Attention、CNN-LSTM-Attention、OOA-CNN-LSTM、CNN-LSTM四模型多变量时序预测一键对比预测效果基本介绍程序设计参考资料 预测效果 基本介绍 基于OOA-CN…
阅读更多...
20250421在荣品的PRO-RK3566开发板的Android13下频繁重启RKNPU fde40000.npu: Adding to iommu gr

20250421在荣品的PRO-RK3566开发板的Android13下频繁重启RKNPU fde40000.npu: Adding to iommu gr

20250421在荣品的PRO-RK3566开发板的Android13下频繁重启RKNPU fde40000.npu: Adding to iommu gr 2025/4/21 14:50 缘起&#xff1a;电池没电了&#xff0c;导致荣品的PRO-RK3566的核心板频繁重启。 内核时间4s就重启。100%复现。 PRO-RK3566 Android13启动到这里 复位&#…
阅读更多...
动态监控进程

动态监控进程

1.介绍: top和ps命令很相似,它们都是用来显示正在执行的进程,top和ps最大的不同之处,在于top在执行中可以更新正在执行的进程. 2.基本语法&#xff1a; top [选项] 选项说明 ⭐️僵死进程&#xff1a;内存没有释放,但是进程已经停止工作了,需要及时清理 交互操作说明 应用案…
阅读更多...
657SJBH西藏藏药特产销售管理系统

657SJBH西藏藏药特产销售管理系统

毕业论文&#xff08;设计&#xff09;文献综述 西藏藏药特产销售管理系统的设计与实现 近年来&#xff0c;随着网络技术特别是Internet技术的普及和发展&#xff0c;电子商务的开发和应用成为一个热门领域&#xff0c;在线藏药特产销售系统就是这其中的一员。 藏药产业在西藏…
阅读更多...
栈和队列--数据结构初阶(2)(C/C++)

栈和队列--数据结构初阶(2)(C/C++)

文章目录 前言理论部分栈的模拟实现STL中的栈容器队列的模拟实现STL中的队列容器 作业部分 前言 这期的话会给大家讲解栈和队列的模拟实现和在STL中栈和队列怎么用的一些知识和习题部分(这部分侧重于理论知识&#xff0c;习题倒还是不难) 理论部分 栈的模拟实现 typedef int…
阅读更多...
RNN的理解

RNN的理解

对于RNN的理解 import torch import torch.nn as nn import torch.nn.functional as F# 手动实现一个简单的RNN class RNN(nn.Module):def __init__(self, input_size, hidden_size, output_size):super(RNN, self).__init__()# 定义权重矩阵和偏置项self.hidden_size hidden…
阅读更多...
二叉查找树和B树

二叉查找树和B树

二叉查找树&#xff08;Binary Search Tree, BST&#xff09;和 B 树&#xff08;B-tree&#xff09;都是用于组织和管理数据的数据结构&#xff0c;但它们在结构、应用场景和性能方面有显著区别。 二叉查找树&#xff08;Binary Search Tree, BST&#xff09; 特点&#xff1…
阅读更多...
一段式端到端自动驾驶：VAD：Vectorized Scene Representation for Efficient Autonomous Driving

一段式端到端自动驾驶：VAD：Vectorized Scene Representation for Efficient Autonomous Driving

论文地址&#xff1a;https://github.com/hustvl/VAD 代码地址&#xff1a;https://arxiv.org/pdf/2303.12077 1. 摘要 自动驾驶需要对周围环境进行全面理解&#xff0c;以实现可靠的轨迹规划。以往的方法依赖于密集的栅格化场景表示&#xff08;如&#xff1a;占据图、语义…
阅读更多...
OpenCV训练题

OpenCV训练题

一、创建一个 PyQt 应用程序&#xff0c;该应用程序能够&#xff1a; 使用 OpenCV 加载一张图像。在 PyQt 的窗口中显示这张图像。提供四个按钮&#xff08;QPushButton&#xff09;&#xff1a; 一个用于将图像转换为灰度图一个用于将图像恢复为原始彩色图一个用于将图像进行…
阅读更多...
最新文章

Copyright @ 2022~2023