性能测试框架
Posted FunTester
tags:
篇首语:本文由小常识网(cha138.com)小编为大家整理,主要介绍了性能测试框架相关的知识,希望对你有一定的参考价值。
之前写过一个性能测试框架,只是针对单一的HTTP接口的测试,对于业务接口和非HTTP接口还无非适配,刚好前端时间工作中用到了,就更新了自己的测试框架,这次不再以请求为基础,而是以方法为基础,这样就可以避免了单一性,有一个base类,然后其他的各种单一性请求在单独写一个适配类就好了,如果只是临时用,直接重新实现base即可。下面分享:
package com.fun.frame.thead;
import com.fun.frame.SourceCode;
import com.fun.frame.excute.Concurrent;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;
import java.util.ArrayList;
import java.util.List;
import java.util.concurrent.CountDownLatch;
import static com.fun.utils.Time.getTimeStamp;
/**
* 多线程任务基类,可单独使用
*/
public abstract class ThreadBase<T> extends SourceCode implements Runnable {
private static final Logger logger = LoggerFactory.getLogger(ThreadBase.class);
/**
* 任务请求执行次数
*/
public int times;
/**
* 计数锁
* <p>
* 会在concurrent类里面根据线程数自动设定
* </p>
*/
CountDownLatch countDownLatch;
/**
* 用于设置访问资源
*/
public T t;
public ThreadBase(T t) {
this();
this.t = t;
}
public ThreadBase() {
super();
}
/**
* groovy无法直接访问t,所以写了这个方法
*
* @return
*/
public String getT() {
return t.toString();
}
public void run() {
try {
before();
List<Long> t = new ArrayList<>();
long ss = getTimeStamp();
for (int i = 0; i < times; i++) {
long s = getTimeStamp();
doing();
long e = getTimeStamp();
t.add(e - s);
}
long ee = getTimeStamp();
logger.info("执行次数:{},总耗时:{}", times, ee - ss);
Concurrent.allTimes.addAll(t);
} catch (Exception e) {
logger.warn("执行任务失败!", e);
} finally {
after();
if (countDownLatch != null)
countDownLatch.countDown();
}
}
/**
* 运行待测方法的之前的准备
*/
protected abstract void before();
/**
* 待测方法
*
* @throws Exception
*/
protected abstract void doing() throws Exception;
/**
* 运行待测方法后的处理
*/
protected abstract void after();
public void setCountDownLatch(CountDownLatch countDownLatch) {
this.countDownLatch = countDownLatch;
}
public void setTimes(int times) {
this.times = times;
}
}
下面是几个实现过的基础类:
package com.fun.frame.thead;
import com.fun.httpclient.ClientManage;
import com.fun.httpclient.FanLibrary;
import com.fun.httpclient.GCThread;
import org.apache.http.HttpStatus;
import org.apache.http.client.methods.CloseableHttpResponse;
import org.apache.http.client.methods.HttpRequestBase;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;
import java.io.IOException;
/**
* http请求多线程类
*/
public class RequestThread extends ThreadBase {
static Logger logger = LoggerFactory.getLogger(RequestThread.class);
/**
* 请求
*/
public HttpRequestBase request;
/**
* 单请求多线程多次任务构造方法
*
* @param request 被执行的请求
* @param times 每个线程运行的次数
*/
public RequestThread(HttpRequestBase request, int times) {
this.request = request;
this.times = times;
}
public void before() {
request.setConfig(FanLibrary.requestConfig);
GCThread.starts();
}
protected void doing() throws Exception {
getResponse(request);
}
protected void after() {
GCThread.stop();
}
/**
* 多次执行某个请求,但是不记录日志,记录方法用 loglong
* <p>此方法只适应与单个请求的重复请求,对于有业务联系的请求暂时不能适配</p>
*
* @param request 请求
* @throws IOException
*/
void getResponse(HttpRequestBase request) throws IOException {
CloseableHttpResponse response = ClientManage.httpsClient.execute(request);
String content = FanLibrary.getContent(response);
if (response.getStatusLine().getStatusCode() != HttpStatus.SC_OK)
logger.warn("响应状态码:{},响应内容:{}", content, response.getStatusLine());
if (response != null) response.close();
}
}
下面是数据库的:
package com.fun.frame.thead;
import com.fun.interfaces.ImysqlBasic;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;
import java.sql.SQLException;
/**
* 数据库多线程类
*/
public class QuerySqlThread extends ThreadBase {
private static Logger logger = LoggerFactory.getLogger(QuerySqlThread.class);
String sql;
IMySqlBasic base;
public QuerySqlThread(IMySqlBasic base, String sql, int times) {
this.times = times;
this.sql = sql;
this.base = base;
}
public void before() {
base.getConnection();
}
protected void doing() throws SQLException {
base.excuteQuerySql(sql);
}
protected void after() {
base.mySqlOver();
}
}
下面是concurrent类:
package com.fun.frame.excute;
import com.fun.bean.PerformanceResultBean;
import com.fun.frame.Save;
import com.fun.frame.SourceCode;
import com.fun.frame.thead.ThreadBase;
import com.fun.profile.Constant;
import com.fun.utils.Time;
import com.fun.utils.WriteRead;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;
import java.util.List;
import java.util.Vector;
import java.util.concurrent.CountDownLatch;
import java.util.concurrent.ExecutorService;
import java.util.concurrent.Executors;
public class Concurrent {
private static Logger logger = LoggerFactory.getLogger(Concurrent.class);
/**
* 线程任务
*/
public ThreadBase thread;
public List<ThreadBase> threads;
public int num;
public static Vector<Long> allTimes = new Vector<>();
ExecutorService executorService;
CountDownLatch countDownLatch;
/**
* @param thread 线程任务
* @param num 线程数
*/
public Concurrent(ThreadBase thread, int num) {
this(num);
this.thread = thread;
}
/**
* @param threads 线程组
*/
public Concurrent(List<ThreadBase> threads) {
this(threads.size());
this.threads = threads;
}
public Concurrent(int num) {
this.num = num;
executorService = Executors.newFixedThreadPool(num);
countDownLatch = new CountDownLatch(num);
}
/**
* 执行多线程任务
*/
public PerformanceResultBean start() {
long start = Time.getTimeStamp();
for (int i = 0; i < num; i++) {
ThreadBase thread = getThread(i);
thread.setCountDownLatch(countDownLatch);
executorService.execute(thread);
}
shutdownService(executorService, countDownLatch);
long end = Time.getTimeStamp();
logger.info("总计" + num + "个线程,共用时:" + Time.getTimeDiffer(start, end) + "秒!");
return over();
}
private static void shutdownService(ExecutorService executorService, CountDownLatch countDownLatch) {
try {
countDownLatch.await();
executorService.shutdown();
} catch (InterruptedException e) {
logger.warn("线程池关闭失败!", e);
}
}
private PerformanceResultBean over() {
Save.saveLongList(allTimes, num);
return countQPS(num);
}
ThreadBase getThread(int i) {
if (threads == null) return thread;
return threads.get(i);
}
/**
* 计算结果
* <p>此结果仅供参考</p>
*
* @param name 线程数
*/
public static PerformanceResultBean countQPS(int name) {
List<String> strings = WriteRead.readTxtFileByLine(Constant.LONG_Path + name + Constant.FILE_TYPE_LOG);
int size = strings.size();
int sum = 0;
for (int i = 0; i < size; i++) {
int time = SourceCode.changeStringToInt(strings.get(i));
sum += time;
}
double v = 1000.0 * size * name / sum;
PerformanceResultBean performanceResultBean = new PerformanceResultBean(name, size, sum / size, v);
performanceResultBean.print();
return performanceResultBean;
}
}
redis实现类缺失,因为没有遇到需要单独实现的需求。
关于用代码还是用工具实现并发,我个人看法所有所长,单究其根本,必然是代码胜于工具,原因如下:门槛高,适应性强;贴近开发,利于调优。性能测试,并发只是开始,只有一个好的开始才能进行性能数据分析,性能参数调优。所以不必拘泥于到底使用哪个工具那种语言,据我经验来说:基本的测试需求都是能满足的,只是实现的代价不同。
groovy是一种基于JVM的动态语言,我觉得最大的优势有两点,第一:于java兼容性非常好,大部分时候吧groovy的文件后缀改成java直接可以用,反之亦然。java的绝大部分库,groovy都是可以直接拿来就用的。这还带来了另外一个有点,学习成本低,非常低,直接上手没问题,可以慢慢学习groovy不同于Java的语法;第二:编译器支持变得更好,现在用的intellij的ide,总体来说已经比较好的支持groovy语言了,写起代码来也是比较顺滑了,各种基于groovy的框架工具也比较溜,特别是Gradle构建工具,比Maven爽很多。----此段文字为了撑字数强加的,与内容无关。
点击阅读原文,有兴趣的童鞋可以一起交流
以上是关于性能测试框架的主要内容,如果未能解决你的问题,请参考以下文章
Unity HTFramework框架(四十)Debug的性能监控