1、需要下载一个jar包,放入到lib /ext文件中  JMeterPlugins-Standard.jar

2、在服务器上安装 perfmon server agent

这里监控内存我们使用的是:jp@gc - PerfMon Metrics Collectot

在使用之前,我们需要运行/serverAgent/startAgent.bat这个文件,我们需要将serverAgent目录及下面的文件复制到我们测试的服务器上,然后点击打开

3、 创建监控器时选择如图:

PerfMon Server Agent

Installation

You do not need any root/admin privilege. You can just unzip the the ServerAgent-X.X.X.zip somewhere on the server. Then launch the agent using startAgent.sh script on Unix, or startAgent.bat script on Windows.

The agent is written in Java, so you will need JRE 1.4+ to run it. Note you can create yourself the agent package which includes its own JRE so you don't have to install java on the server (We tested it on windows platform). To do this, just create a JRE folder in the agent folder and copy one installed JRE inside. Change the java command inside the .bat file to the path to the java executable and it will work.

Once the agent is running, you can use the PerfMon Metrics Collector Listener to connect to the agents. You can add multiple servers to monitor, one graph can display several kinds of metrics (cpu, memory, etc...), auto-zooming rows for best view.

Usage

To start the agent, simply run startAgent bat/sh file. It will open UDP/TCP server ports where JMeter will connect and query the metrics. Some information will be printed to standard output, informing you on problems gathering metrics (logging verbosity regulated with --loglevel parameter).

You can specify the listening ports as arguments (0 disables listening), default is 4444:

$ ./startAgent.sh --udp-port 0 --tcp-port 3450
INFO 2011-11-25 19:02:14.315 [kg.apc.p] (): Binding TCP to 3450

You can use the --auto-shutdown option when starting the agent to automatically stop it once the test is finished. It is recommended to use this feature only with TCP connections:

$ undera@undera-HP:/tmp/serverAgent$ ./startAgent.sh --udp-port 0 --auto-shutdown
INFO 2011-11-25 19:48:59.321 [kg.apc.p] (): Agent will shutdown when all clients disconnected
INFO 2011-11-25 19:48:59.424 [kg.apc.p] (): Binding TCP to 4444

You can use the --sysinfo option to view available system objects:

$ ./startAgent.sh --sysinfo
INFO 2011-11-25 19:51:25.517 [kg.apc.p] (): *** Logging available processes ***
INFO 2011-11-25 19:51:25.542 [kg.apc.p] (): Process: pid=24244 name=bash args=-bash
INFO 2011-11-25 19:51:25.543 [kg.apc.p] (): Process: pid=25086 name=dash args=/bin/sh ./startAgent.sh --sysinfo
INFO 2011-11-25 19:51:25.543 [kg.apc.p] (): Process: pid=25088 name=java args=java -jar ./CMDRunner.jar --tool PerfMonAgent --sysinfo
INFO 2011-11-25 19:51:25.549 [kg.apc.p] (): *** Logging available filesystems ***
INFO 2011-11-25 19:51:25.551 [kg.apc.p] (): Filesystem: fs=/dev type=devtmpfs
INFO 2011-11-25 19:51:25.551 [kg.apc.p] (): Filesystem: fs=/ type=ext4
INFO 2011-11-25 19:51:25.551 [kg.apc.p] (): Filesystem: fs=/var/run type=tmpfs
INFO 2011-11-25 19:51:25.551 [kg.apc.p] (): Filesystem: fs=/home type=ext4
INFO 2011-11-25 19:51:25.552 [kg.apc.p] (): Filesystem: fs=/var/lock type=tmpfs
INFO 2011-11-25 19:51:25.552 [kg.apc.p] (): Filesystem: fs=/proc type=proc
INFO 2011-11-25 19:51:25.553 [kg.apc.p] (): *** Logging available network interfaces ***
INFO 2011-11-25 19:51:25.554 [kg.apc.p] (): Network interface: iface=lo addr=127.0.0.1 type=Local Loopback
INFO 2011-11-25 19:51:25.554 [kg.apc.p] (): Network interface: iface=eth0 addr=192.168.0.1 type=Ethernet
INFO 2011-11-25 19:51:25.555 [kg.apc.p] (): *** Done logging sysinfo ***
INFO 2011-11-25 19:51:25.555 [kg.apc.p] (): Binding UDP to 4444
INFO 2011-11-25 19:51:26.560 [kg.apc.p] (): Binding TCP to 4444

The --interval <seconds> argument can be used to change metrics collection frequency.

Using Server Agent With Other Applications

Server Agent uses simple plain-text protocol, anyone can use agent's capabilities implementing client, based on kg.apc.perfmon.client.Transport interface. If anyone's interested, start the topic on the support forums and I'll describe how to connect third-party client app to agent.

ServerAgent has simple text protocol and can work on UDP and TCP transports. Most of cases will use TCP.

To have your first talk with the agent, start it locally. Then use telnet utility to connect to it:

user@ubuntu:~$ telnet localhost 4444
Trying 127.0.0.1...
Connected to localhost.
Escape character is '^]'.

If connection has succeeded, you should see "Accepting new TCP connection" message in ServerAgent console log. Type "test" and press Enter in telnet chat, server should answer with short "Yep":

test
Yep

Type "exit":

exit
Connection closed by foreign host.

That's it. You sending a command line, server answering. Command line consists of command, sometimes with parameters. Parameters are separated from command with a colon sign.

Possible commands are:

  • exit - terminates current client session and closes connection to agent, no parameters
  • test - test if server is alive, no parameters
  • shutdown - terminate all client connections and shutdown agent process, no parameters
  • interval - change metrics reporting interval used in 'metrics' command, single parameter is integer value in seconds. Interval can be changed in the middle of metrics reporting. Example: interval:5
  • metrics - starts automatic metrics collection, parameters are metrics list to collect, described below. Example: metrics:cpu
  • metrics-single - calls single metric collection iteration. Example: metrics-single:memory

Specifying Metrics

Metrics list consists of metric specifications, separated by TAB character. Metric collection output consists of float values, TAB separated. Example:

metrics-single:cpumemory
22.0573689416419457.52359562205553

Each metric specification consists of several fields, colon-separated. Short example:

metrics-single:cpu:idle    memory:free
80.02238388360381 57.52359562205553

Fields number is metric-type specific. Possible metric types are:

  • cpu
  • memory
  • swap
  • disks
  • network
  • tcp
  • tail
  • exec
  • jmx

Fields corresponding to each metric type are described at metrics page. Last example (Yep, ServerAgent can be shell exec vulnerability. If you have issue with this, ask me and I'll introduce 'secure' mode, disabling insecure metric types):

metrics-single:exec:/bin/sh:-c:free | grep Mem | awk '{print $7}'
1152488

jmeter 监控服务器的内存,cpu等内容的更多相关文章

  1. Jmeter监控服务器的CPU

    1.Jmeter监控服务器的软件:链接: https://pan.baidu.com/s/1Ag75IXqR2XMe8RpZA4nvHg 密码: yki2 2.(1)Window版本 <1> ...

  2. 『动善时』JMeter基础 — 61、使用JMeter监控服务器

    目录 1.监控插件安装 2.启动监控服务 3.使用JMeter监控服务器 (1)测试计划内包含的元件 (2)HTTP请求界面内容 (3)配置jp@gc-PerfMon Metrics Collecto ...

  3. Jmeter监控服务器笔记

    Jmeter监控服务器-CPU,Memory,Disk,Network性能指标 本文主要说一下如何通过JMeter插件来监控服务器CPU.内存.磁盘.网络等相关资源. 一.下载 首先进入网址https ...

  4. 使用jmeter监控服务器性能指标

    先下载jmeter-ServerAgent Windows下载和Linux下载 https://jmeter-plugins.org/wiki/PerfMon/  找到ServerAgent的下载链接 ...

  5. ubuntu查看系统资源占用(内存,cpu和进程)

    ubuntu查看系统资源占用(内存,cpu和进程) 1  top 查看ubuntu的资源占用的命令为$: top    说明:top命令就可以查看内存,cpu和进程了,很方便 top: 主要参数 d: ...

  6. linux查看内存cpu占用

    linux查看内存cpu占用top 命令  按q退出 可以添加额外选项选择按进程或按用户查看如: top -u gitu PID:进程idPR:进程的优先级别,越小越优先被执行NInice:值VIRT ...

  7. vs中调试程序查看变量在内存中的内容的方法

    vs中调试程序 查看变量在内存中的内容的方法 https://blog.csdn.net/guojg1988/article/details/42922149 原文链接:http://www.sows ...

  8. JMeter监控服务器CPU、内存的方法

    jmeter也可以像loadrunner一样监控服务器CPU.内存等性能参数,不过需要安装一些插件 一.首先下载监控服务器的插件   链接:https://pan.baidu.com/s/1o9Zuw ...

  9. jmeter ---监控服务器CPU, 内存,网络数据

    JMeter如何收集获得服务器cpu,内存,磁盘,网络等相关资源使用率的信息 1. JMeter 自带的Monitor Results 监控 JMeter 自带的Monitor Results 在官网 ...

随机推荐

  1. html框架—多对话框(相同id)处理

    一个网站的数据大多数都是异步刷新的,这没什么好说的,然后现在很多前后端框架,大家都知道框架很好用,不用自己写样式,只要利用框架上的语法就能做出漂亮的动态的效果来,而用框架的话大多数的动态效果都是动态生 ...

  2. SVN 常识

    1.相关博客 http://my.oschina.net/u/1780920/blog/425792 2. 文件红色:表示文件没有添加到服务器 绿色:表示没有更新新的修改到服务器 普通黑色:表示和服务 ...

  3. 图解Android触摸事件分发

    Android中触摸事件传递过程中最重要的是dispatchTouchEvent().onInterceptTouchEvent()和onTouchEvent()方法. View和Activity有d ...

  4. Android程序意外Crash后自动重启

    1.自定义UncaughtExceptionHandler public class UnCeHandler implements UncaughtExceptionHandler { private ...

  5. 【读书笔记】iOS网络-测试与操纵网络流量

    一,观测网络流量. 观测网络流量的行为叫做嗅探或数据包分析. 1,嗅探硬件. 从iOS模拟器捕获数据包不需要做特别的硬件或网络配置.如果需要捕获这些数据包,那么可以使用嗅探软件来监听回送设备或是用于连 ...

  6. [读书笔记] Head First 设计模式

    OO基础 抽象 封装 多态 继承 OO原则 原则 描述 封装变化 找出应用中可能需要变化之处,把它们独立出来,不要和那些不需要变化的代码混合在一起. 把会变化的部分取出来并“封装”起来,好让其他部分不 ...

  7. iOS中数据传值的几种方式

    值传递:基本数据类型的变量之间的数据传递 //值传递不会改变变量的值 void func(int a) { a = ; } int main(int argc, const char * argv[] ...

  8. iOS PushMeBaby日志提示SSLwrite():-36 94

    在测试远程推送时用PushMeBaby来模拟服务器发推送消息 但是点击push按钮后手机没有收到推送消息,多点击几下后程序就崩了,强行运行后日志输出了 SSLwrite():-36 94这句话 出现这 ...

  9. js获取url

    location.href 返回完整的url location.origin  返回带协议的主机域名  如http://www.test.com location.pathname  返回url中路径 ...

  10. 入CTF坑必不可少的地方-保持更新

    0x00 前言 没有交易,没有买卖,没有排名,纯属分享:p 0x01 CTF介绍 CTF领域指南CTF介绍大全CTF赛事预告 0x02 CTF练习 BIN:reversingpwnableexploi ...