初始化测试步伐超过用户负载速度 [英] Pace of Initialization test exceed user load pace

查看:72
本文介绍了初始化测试步伐超过用户负载速度的处理方法,对大家解决问题具有一定的参考价值,需要的朋友们下面随着小编来一起学习吧!

问题描述



我正在使用VSTS 2008对应用程序进行负载测试。我的测试步骤如下:

1。对每个用户执行一次初始化测试2。我为每个用户使用Fiddler运行我测试的Web测试脚本。为每个用户重复相同的测试直到时间到了

我的目的是确定在每个用户登录一次后,所有用户在一段时间内重复执行相同的测试会使服务器过载。

我已经尝试了所有3种测试混合设置,测试时间为30分钟,最多100个用户。我使用的是0初始用户的步进加载模式,步长为10秒,每秒增加2个用户。

Hi,

I am using VSTS 2008 for load testing an application. My steps for testing is as listed below.

1. Perform an initialization test once for every user
2. Run the web test script i recorded using Fiddler for every user
3. Repeat the same test for each user till time's up

My aim is to determine whether after logging in once for each user, performing the same test repeatedly within a period of time by all users will overload the servers.


I have tried all 3 test mix settings with a test time of 30 mins and maximum 100 users. I am using a step load pattern of 0 initial users, a step duration of 10 secs and an increment of 2 users per sec.

当我使用前两个测试混合设置时(总数)测试和虚拟用户的数量),我注意到传递的初始化测试的速度总是超过用户负载的速度 - 甚至在用户加载之前就执行初始化测试。这没有任何意义,因为我需要在执行初始化测试之前要加载的用户。但是,当我使用基于用户速度的测试组合时,传递的初始化测试的速度与用户加载的速度一致 - 初始化测试将仅在用户加载时执行。为什么会这样?

When i use the 1st two test mix settings (Total number of tests and number of virtual users), i noticed that the the pace of the passed intialization tests always exceed the pace of the user load - initialization test is performed even before the user is loaded in. This doesnt make any sense as I need the user to be loaded before performing the intialization test. However, when i use the test mix which is based on user pace, the pace of the passed intialization test is consistent with the pace of the user load - Initialization test will only be perform when the user is loaded in. Why is this so?

我不明白测试混音的不同用法。根据我在线阅读的内容,使用前两个测试混合设置不应该是一个问题,因为我正在运行一个测试。根据用户速度使用测试组合对我的要求不准确,因为我不知道每个用户在现实生活中的节奏。理想的情况是重复运行测试一段时间并检查服务器和带宽的性能。

I dont understand the different usuage of the test mixes. Based on what i read online, using the 1st two test mix settings shouldnt be an issue since i am running a single test. Using a test mix based on user pace is not accurate for my requirements as i wouldnt know the pace of each user in real-life. The ideal scenario would be to run the test repeatedly for a period of time and check the performance of the servers and bandwidth.

请提供建议。谢谢

推荐答案

'使用以记录的思考时间为中心的正态分布'&  选择'基于虚拟用户数量[测试模型] &设置随机思考时间,如[5到20秒]

'Use normal distribution centered on recorded think times' & choose 'based on the number of virtual users [test model]' & set random think time like [ 5 to 20 seconds]

任何混淆然后让我知道

问候,

Himanshu


这篇关于初始化测试步伐超过用户负载速度的文章就介绍到这了,希望我们推荐的答案对大家有所帮助,也希望大家多多支持IT屋!

查看全文
登录 关闭
扫码关注1秒登录
发送“验证码”获取 | 15天全站免登陆