App下載

實(shí)例代碼展示Java是如何用線程工廠來(lái)監(jiān)控線程池

猿友 2021-07-23 09:50:38 瀏覽數(shù) (2256)
反饋

Java 中線程池是為了更好地管理、維護(hù)和調(diào)用線程,那么用什么來(lái)管理或者監(jiān)控線程池中的操作呢?下面我將和大家分享關(guān)于 Java 是怎么用線程工廠來(lái)監(jiān)控線程池的內(nèi)容。

ThreadFactory

線程池中的線程從哪里來(lái)呢?就是ThreadFoctory

public interface ThreadFactory {
    Thread newThread(Runnable r);
}

Threadfactory里面有個(gè)接口,當(dāng)線程池中需要?jiǎng)?chuàng)建線程就會(huì)調(diào)用該方法,也可以自定義線程工廠

public class ThreadfactoryText {
    public static void main(String[] args) {
        Runnable runnable=new Runnable() {
            @Override
            public void run() {
                int num=new Random().nextInt(10);
                System.out.println(Thread.currentThread().getId()+"--"+System.currentTimeMillis()+"--睡眠"+num);
                try {
                    TimeUnit.SECONDS.sleep(num);
                } catch (InterruptedException e) {
                    e.printStackTrace();
                }
            }
        };
        //創(chuàng)建線程池 使用自定義線程工廠 采用默認(rèn)的拒絕策略
        ExecutorService executorService=new ThreadPoolExecutor(5, 5, 0, TimeUnit.SECONDS, new SynchronousQueue<>(), new ThreadFactory() {
            @Override
            public Thread newThread(Runnable r) {
                Thread t=new Thread(r);
                t.setDaemon(true);//設(shè)置為守護(hù)線程,當(dāng)主線程運(yùn)行結(jié)束,線程池中線程也會(huì)被釋放
                System.out.println("創(chuàng)建了線程"+t);
                return t;
            }
        });
        //提交五個(gè)任務(wù)
        for (int i = 0; i < 5; i++) {
            executorService.submit(runnable);
        }
    }
}

當(dāng)線程提交超過(guò)五個(gè)任務(wù)時(shí),線程池會(huì)默認(rèn)拋出異常

監(jiān)控線程池

ThreadPoolExcutor提供了一組方法用于監(jiān)控線程池

int getActiveCount()//獲得線程池只當(dāng)前的獲得線程數(shù)量
long getCompletedTaskCount()//返回線程池完成任務(wù)數(shù)量
int getCorePoolSize()//線程池中核心任務(wù)數(shù)量
int getLargestPoolSize() //返回線程池中曾經(jīng)達(dá)到線程的最大數(shù)
int getMaximumPoolSize()//返回線程池的最大容量
int getPoolSize()//返回線程大小
BlockingQueue<Runnable> getQueue()//返回阻塞隊(duì)列
long getTaskCount()//返回線程池收到任務(wù)總數(shù)
public class Text {
    public static void main(String[] args) throws InterruptedException {
        Runnable runnable = new Runnable() {
            @Override
            public void run() {
                System.out.println(Thread.currentThread().getId() + "線程開(kāi)始執(zhí)行--" + System.currentTimeMillis());
                try {
                    Thread.sleep(10000);
                } catch (InterruptedException e) {
                    e.printStackTrace();
                }
            }
        };
        //創(chuàng)建線程池 使用默認(rèn)線程工廠 有界隊(duì)列  采用DiscardPolicy策略
        ThreadPoolExecutor executorService = new ThreadPoolExecutor(2, 5, 0, TimeUnit.SECONDS, new ArrayBlockingQueue<>(5),Executors.defaultThreadFactory(),new ThreadPoolExecutor.DiscardPolicy());
        //提交五個(gè)任務(wù)
        for (int i = 0; i < 30; i++) {
            executorService.submit(runnable);
            System.out.println("當(dāng)前線程核心線程數(shù)"+executorService.getCorePoolSize()+",最大線程數(shù):"+executorService.getMaximumPoolSize()+",當(dāng)前線程池大小:"+executorService.getPoolSize()+"活動(dòng)線程數(shù):"+executorService.getActiveCount()+",收到任務(wù):"+executorService.getTaskCount()+"完成任務(wù)數(shù):"+executorService.getCompletedTaskCount()+"等待任務(wù)數(shù):"+executorService.getQueue().size());
            TimeUnit.MILLISECONDS.sleep(500);
        }
        System.out.println("-------------------");
        while (executorService.getActiveCount()>=0)//繼續(xù)對(duì)線程池進(jìn)行檢測(cè)
        {
          System.out.println("當(dāng)前線程核心線程數(shù)"+executorService.getCorePoolSize()+",最大線程數(shù):"+executorService.getMaximumPoolSize()+",當(dāng)前線程池大?。?+executorService.getPoolSize()+"活動(dòng)線程數(shù):"+executorService.getActiveCount()+",收到任務(wù):"+executorService.getTaskCount()+"完成任務(wù)數(shù):"+executorService.getCompletedTaskCount()+"等待任務(wù)數(shù):"+executorService.getQueue().size());
            Thread.sleep(1000);//每1秒檢測(cè)一次
        }

    }
}

當(dāng)線程池大小達(dá)到了核心線程數(shù),線程會(huì)被放在等待隊(duì)列。當(dāng)線程池等待隊(duì)列已滿會(huì)開(kāi)啟新的線程。當(dāng)當(dāng)前線程大小達(dá)到最大線程數(shù),等待隊(duì)列也滿了,再提交的話會(huì)執(zhí)行DiscardPolicy策略,直接丟棄這個(gè)無(wú)法處理的任務(wù),最后30個(gè)任務(wù)只剩下15個(gè)了。

原理如圖:

擴(kuò)展線程池

有時(shí)候需要對(duì)線程池進(jìn)行擴(kuò)展,如在監(jiān)控每個(gè)任務(wù)開(kāi)始和結(jié)束時(shí)間,或者自定義其他增強(qiáng)功能。

ThreadPoolExecutor線程池提供了兩個(gè)方法:

protected void beforeExecute(Thread t, Runnable r) { }
protected void afterExecute(Runnable r, Throwable t) { }

線程池執(zhí)行某個(gè)任務(wù)前會(huì)執(zhí)行beforeExecute()方法,執(zhí)行后會(huì)調(diào)用afterExecute()方法

查看ThreadPoolExecutor源碼,在該類中定義了一個(gè)內(nèi)部類Worker,ThreadPoolExecutor線程池的工作線程就是Worker類的實(shí)例,Worker實(shí)例在執(zhí)行時(shí)會(huì)調(diào)用beforeExecute與afterExecute方法。

public void run() {
            runWorker(this);
}
final void runWorker(Worker w) {
                try {
                    beforeExecute(wt, task);
                    try {
                        task.run();
                        afterExecute(task, null);
                    } catch (Throwable ex) {
                        afterExecute(task, ex);
                        throw ex;
                    }
                } finally {
                    task = null;
                    w.completedTasks++;
                    w.unlock();
                }
            }
    }      

部分代碼已省略,線程執(zhí)行前會(huì)調(diào)用beforeExecute,執(zhí)行后會(huì)調(diào)用afterExecute方法。

擴(kuò)展線程池示例

package com;

import java.util.concurrent.ExecutorService;
import java.util.concurrent.LinkedBlockingDeque;
import java.util.concurrent.ThreadPoolExecutor;
import java.util.concurrent.TimeUnit;

public class Text07 {
    public static void main(String[] args) {

        //定義擴(kuò)展線程池 定義線程池類繼承ThreadPoolExecutor,然后重寫(xiě)其他方法
        ExecutorService threadPoolExecutor=
 new ThreadPoolExecutor(5,5,0, TimeUnit.SECONDS,new LinkedBlockingDeque<>()){
     //在內(nèi)部類重寫(xiě)開(kāi)始方法
     @Override
     protected void beforeExecute(Thread t, Runnable r) {
         System.out.println(t.getId()+"線程準(zhǔn)備執(zhí)行任務(wù)"+((Mytask)r).name);
     }
     //在內(nèi)部類重寫(xiě)結(jié)束方法
     @Override
     protected void afterExecute(Runnable r, Throwable t) {
         System.out.println(((Mytask)r).name+"執(zhí)行完成");
     }
     //線程池退出
     @Override
     protected void terminated() {
         System.out.println("線程池退出");
     }
 };
        for (int i = 0; i < 5; i++) {
            Mytask mytask=new Mytask("Thread"+i);
            threadPoolExecutor.execute(mytask);
        }
    }
    private  static  class  Mytask implements Runnable
    {
        private  String name;

        public  Mytask(String name)
        {
            this.name=name;
        }
        @Override
        public void run() {
            System.out.println(name+"正在被執(zhí)行"+Thread.currentThread().getId());
            try {
                Thread.sleep(1000);//模擬任務(wù)時(shí)長(zhǎng)
            } catch (InterruptedException e) {
                e.printStackTrace();
            }
        }
    }
}

優(yōu)化線程池大小

線程池大小對(duì)系統(tǒng)性能有一定影響,過(guò)大或者過(guò)小都無(wú)法方法發(fā)揮系統(tǒng)最佳性能,不需要非常精確,只要避免極大或者極小就可以了,一般來(lái)說(shuō)線程池大小大姚考慮CPU數(shù)量

線程池大小=CPU數(shù)量 * 目標(biāo)CPU使用率*(1+等待時(shí)間與計(jì)算時(shí)間的比)

線程池死鎖

如果線程池執(zhí)行中,任務(wù)A在執(zhí)行過(guò)程中提交了任務(wù)B,任務(wù)B添加到線程池中的等待隊(duì)列,如果A的結(jié)束需要B的執(zhí)行結(jié)果,而B(niǎo)線程需要等待A線程執(zhí)行完畢,就可能會(huì)使其他所有工作線程都處于等待狀態(tài),待這些任務(wù)在阻塞隊(duì)列中執(zhí)行。線程池中沒(méi)有可以對(duì)阻塞隊(duì)列進(jìn)行處理的線程,就會(huì)一直等待下去照成死鎖。

適合給線程池提交相互獨(dú)立的任務(wù),而不是彼此依賴的任務(wù),對(duì)于彼此依賴的任務(wù),可以考慮分別提交給不同的線程池來(lái)處理。

線程池異常信息捕獲

import java.util.concurrent.ExecutorService;
import java.util.concurrent.SynchronousQueue;
import java.util.concurrent.ThreadPoolExecutor;
import java.util.concurrent.TimeUnit;

public class Text09 {
    public static void main(String[] args) {
        //創(chuàng)建線程池
        ExecutorService executorService=new ThreadPoolExecutor(5,5,0, TimeUnit.SECONDS,new SynchronousQueue<>());
        //向線程池中添加兩個(gè)數(shù)相處計(jì)算的任務(wù)
        for (int i = 0; i <5 ; i++) {
            executorService.submit(new Text(10,i));
        }

    }
    private  static class  Text implements  Runnable
    {
        private  int x;
        private  int y;
        public  Text(int x,int y)
        {
            this.x=x;
            this.y=y;
        }
        @Override
        public void run() {
            System.out.println(Thread.currentThread().getName()+"線程x/y結(jié)果的為"+x+"/"+y+"="+(x/y));
        }
    }
}

可以看到只有四條結(jié)果,實(shí)際向線程池提交了五個(gè)任務(wù),但是當(dāng)i==0時(shí),產(chǎn)生了算術(shù)異常,線程池把該異常吃掉了,導(dǎo)致我們對(duì)該異常一無(wú)所知

解決辦法:

1.把submit改為execute

2.對(duì)線程池進(jìn)行擴(kuò)展,對(duì)submit進(jìn)行包裝

package com;

import java.util.concurrent.*;

public class Text09 {
    public static void main(String[] args) {
        //創(chuàng)建線程池  使用自定義的線程池
        ExecutorService executorService=new TranceThreadPoorExcuter(5,5,0, TimeUnit.SECONDS,new SynchronousQueue<>());
        //向線程池中添加兩個(gè)數(shù)相處計(jì)算的任務(wù)
        for (int i = 0; i <5 ; i++) {
            executorService.submit(new Text(10,i));
        }

    }
    public  static class  Text implements  Runnable
    {
        public  int x;
        public  int y;
        public  Text(int x,int y)
        {
            this.x=x;
            this.y=y;
        }

        @Override
        public void run() {
            System.out.println(Thread.currentThread().getName()+"線程x/y結(jié)果的為"+x+"/"+y+"="+(x/y));
        }
    }
    //自定義線程池類 對(duì)TranceThreadPoorExcuter進(jìn)行擴(kuò)展
    private  static  class  TranceThreadPoorExcuter extends  ThreadPoolExecutor
    {

        public TranceThreadPoorExcuter(int corePoolSize, int maximumPoolSize, long keepAliveTime, TimeUnit unit, BlockingQueue<Runnable> workQueue) {
            super(corePoolSize, maximumPoolSize, keepAliveTime, unit, workQueue);
        }
        //定義一個(gè)方法用于傳入兩個(gè)參數(shù) 第一個(gè)是要接受的任務(wù) 第二個(gè)是Exception
        public  Runnable warp(Runnable r,Exception e)
        {
            return new Runnable() {
                @Override
                public void run() {

                    try {
                        r.run();
                    }
                    catch (Exception e1)
                    {
                        e.printStackTrace();
                        throw e1;
                    }
                }
            };
        }
        //重寫(xiě)submit方法
        @Override
        public Future<?> submit(Runnable task) {
            return super.submit(warp(task,new Exception("客戶跟蹤異常")));
        }
        //還可以重寫(xiě)excute方法
    }
}

此方法使用了自定義的線程池,重寫(xiě)線程池中的submit方法,在submit方法中,把要傳入的任務(wù)參數(shù)帶一個(gè)捕獲異常信息的功能就可以捕獲線程池異常。

到此本篇關(guān)于用實(shí)例代碼展示 Java 如何使用線程工廠監(jiān)控線程池的文章就介紹到這了,想要了解更多相關(guān) Java 線程的其他內(nèi)容請(qǐng)搜索W3Cschool以前的文章或繼續(xù)瀏覽下面的相關(guān)文章,希望大家以后多多支持我們!


0 人點(diǎn)贊