I am running jmeter tests with 100 to 1k users successfully,
But I suspect if response time is higher than expected.
My test plan include: get login and post login, get test page , post test page, get search page, post search page....
in total 5 get and 5 post forms.
For same number of users if I reduce number of pages the overall response time decreases
so should I run test for pages separately?
Should I reduce number of users for realistic numbers (5 pages*20 users=100 concurrent users)--for performance test?
Or should I be using distributed system?
What is the best practice?
Current Setup: i5, 8GB Ram, One windows machine with Jmeter 3.1
You load testing should be realistic, otherwise it does not make sense. You need to simulate anticipated users behaviour as closely as possible, only this way you will get the real life picture.
So you need to implement your test scenario to match the way, your site will be used in the real world. If it is alive already you can use Access Log Sampler to replay the traffic. If it is not, you can think about "user groups", like:
50% of not authenticated users will be browsing the site
20% of authenticated users will be browsing the site
5% of users will be in login process
3% of users will search for something
etc.
You can use different Thread Groups to represent different groups of users and Throughput Controller to control the frequency of samplers execution inside the Thread Group. See Running JMeter Samplers with Defined Percentage Probability article for more information on how to distribute the load in a realistic way.
There are N number of factors that will affect the overall response times. number of users / pages etc are also few of them. Do you have any specific requirement for performance test? If you are not sure of non functional requirements, then I think all your questions will be clarified in these posts -
http://www.testautomationguru.com/jmeter-performance-testing-application-of-littles-law-to-workload-models/
http://www.testautomationguru.com/jmeter-tips-tricks-for-beginners/
I think you could be running into jmeter issues. What is the JVM heap and GC algorithm that you are using? How are you measuring the performance of each page? If you are using "Transaction Controller" to measure the performance, any slowness in jmeter will affect your transaction controller values.
Have you correlated with the application logs to really understand that the application is indeed slowing down when you decrease the # of pages? If your application response time is constant, I think jmeter is the culprit here. Move all your measurements to the sampler level instead of transaction controllers and update the results
Related
I have a few question to clarify on my understanding of how JMeter works.
a. Thread Group determine the number of users but it does not determine how many HTML requests are generated per sec ? By default, I notice that every user will send a HTML request at a rate of 2 RPS.
b. If I want to change the RPS per user, then I need to use the Through Put Timer. But the Timer can only lower the request rate from 2 RPS to a lower number. It does not increase the RPS.
c. In order to increase the RPS, I need to add more Threads.
d. Does this mean we are limited to 2 RPS per user ? I see some website have links to many other websites so a webpage refresh would make many requests.
Is this the way JMeter works ?
I have a load test which has 8 transaction (eg CRUD,...). I intend to create a overall Test Plan and I want to use INCLUDE to add all the 8 txn. Do I just record the website and INCLUDE ? What should I include, only the HTML requests ?
I'm also thinking of adding Think Time and Add Variables in the 8 scripts before I INCLUDE.
Do I add the Config Element (eg CSV Dataset Config) in the 8 scripts or the overall Test Plan ?
Thanks.
By default each JMeter thread (virtual user) executes requests as fast as it can. If you want to slow JMeter down to mimic a real user which doesn't hammer the server non-stop and needs some time to "think" between operations - use Timers. More information: How do I Correlate the Number of (Concurrent) Users with Hits Per Second
If you want more RPS - add more threads (assuming that the system under test can give you more RPS)
You should INCLUDE everything which is related to your website (images, scripts, styles, fonts, sounds, etc.) but in the same manner as your browser does, i.e. don't record these requests and instead configure JMeter to download embedded resources and use HTTP Cache Manager so JMeter would request these resources just like browser does. Any requests to "external" websites should be excluded (unless they're also developed and supported and in scope for testing)
That's a good approach, if you use a value more than once it makes sense to declare it via User Defined Variables so you would be able to amend the value only in one place
You add it according to your scenarios, be informed about JMeter Scoping Rules
I have a requirement to test that a Public Website can serve a defined peak number of 400 page loads per second.
From what I read online, when testing web pages performance, virtual users (threads) should be configured to pause and "think" on each page they visit, in order to simulate the behavior of a real live user before sending a new page load request.
I must use some remote load generator machines to generate this necessary load, and I have a limit on how many virtual users I can use per each load generator. This means that if I make each virtual user pause and "think" for x seconds on each page, that user will not generate a lot of load compared to how much it would if it was executing as fast as it could with no configured think time - and this would cause me to need more users and implicitly need more load generator machines to achieve my desired "page loads per second" and this would be more costly in the end.
If my only request is to prove that a server can serve 400 page loads per second, I would like to know what difference does it really make if I add think times (and therefore use more virtual users) or not.
Why is generally "think time" considered as something which should be added when testing web pages performance ?
Virtual user which is "idle" (doing nothing) has minimal resources footprint (mainly thread stack size) so I don't think you will need to have more machines
Well-behaved load test must represent real life usage of the application with 100% accuracy, if you're testing a website each JMeter thread (virtual user) must mimic a real user using a real browser with all related features like
handling embedded resources (image, scripts, styles, fonts, sounds, etc.)
using caching properly
getting and sending back cookies
sending appropriate headers
processing AJAX requests like browser does
the most straightforward example of the difference between 400 users without think times and 4000 users with think times will be that 4000 users will open 4000 connections and keep them open and 400 users will open only 400 connections.
How to create proper workload model in JMeter if I have only number of concurrent users and response time as input/requirement? Do I need any additional information in order to load test app?
If these are only info I can get, how to approach load testing in context of using JMeter as load testing tool, any ideas - suggestions - advices?
A "proper workload" would be simulating real life usage of the system under test by "the number of concurrent users".
Just replicate real user behaviour at HTTP protocol level paying attention to pretty much everything i.e.
JMeter's network footprint must be exact replica of the browser's network footprint when it comes to HTTP requests so cross-check the number and the nature of the requests from browser developer tools and JMeter's View Results Tree listener
Make sure to send the same/proper HTTP Headers
As a subset of point 2 pay attention to Cookies
Make sure to download embedded resources like browser does
In addition to point 4 make sure to properly handle cache
Use Timers wisely to simulate think times
If your users are doing different actions configure your JMeter test to distribute them like real users are distributed
Once done just run your test with the anticipated amount of users for the desired duration and compare the real response time against the anticipated
The Avg RT and other KPIs (such as Throughput) are the result of generating or executing the workload with a determined # of VUsers. IMO, the best approach is to generate the load by varying the VUsers as shown in the graph:
This graph also shows a key concept: perfomance (as measure by KPIs) is not linear (although it might appear linear at a small # of VUsers).
I'm using Jmeter to load test an application.
No: of threads is 100.
First HTTP request is the home page.
Second HTTP request is login page where the actual login is made
.
.
n APIs are there.
and what I want to achieve is:
First 100 threads hit the home page concurrently.
Then 100 threads hit the login page concurrently.
.
.
Then 100 threads hit the nth API concurrently.
Add Synchronizing Timer as a child of each request which you need to execute "concurrently"
Set Number of Simultaneous Users to Group by to be equal to the number of concurrent users, in your case 100
Reference: Using the JMeter Synchronizing Timer
Although above approach answers your question better scenario would be starting with 1 virtual user and gradually increase the load, in this case you will be able to correlate increasing load with performance metrics like Response Time, Transactions Per Second, resources consumption of the application under test, etc. as the situation when 100 users open login page and then hit "Login" button at exactly the same moment is highly unlikely to happen and well-behaved load test needs to represent real usage of the application by real users as close as possible.
Threads number 100 ,You need to put each request with synchronizing timer under 3 controllers, See tutorial
I have recorded my web application through template & just to confirm that load test result which i am getting is correct? Just by increasing No of users does it give proper results? Is it enough for load testing of web application?
First of all you need to ensure that your test does what it is supposed to be doing. Recorded tests can rarely be successfully replayed, so normally you should be acting as follows:
Add View Results Tree listener and run your test with 1 user. Inspect request and response details to verify your test steps.
Perform correlation and parametrization if required.
Correlation: the process of identifying and handling any dynamic parameters. Most often people use Regular Expression Extractor for it.
Parametrization: the process of making your test data driven. For example, if your application assumes multiple authenticated users you need to store the credentials somewhere. Most commonly used test element for this is CSV Data Set Config
Make your test realistic. Virtual users simulated by JMeter need to represent real users using real browsers as close as possible with all the related stuff: cookies, headers, cache, etc. See How To Make JMeter Behave More Like A Real Browser to learn how to configure JMeter to act closer to real users. Also real users need some time to "think" between operations so make sure you are using Timers to simulate this behaviour as well.
Only after you apply the above points you should add more virtual users. Again, run your test with 2-3 users and iterations to ensure your test funcitons as designed. Once you are happy with it you can increase the load, but don't overkill your server, increase the load gradually and check the impact of the increasing load on your application, i.e. how response time, throughput and number of errors change as you increase the load. The same is applicable for decreasing the load, don't turn it off at once, decrease the number of virtual users gradually.
Building a Web Test Plan
Building an Advanced Web Test Plan