Using JMeter for performance testing
Вставка
- Опубліковано 21 бер 2020
- We talk about JMeter, a tool to test your site's throughput and load metering. It involves setting up a test plan and running it against your site with a login.
Unlock unlimited opportunities with 50% off your first month of Coursera Plus
imp.i384100.net/c/3402234/209...
Get Proficient knowledge in Azure services and infrastructures, managing cloud resources, cloud security with our Cloud Support Associate Certificate !
imp.i384100.net/c/3402234/203...
Join the Google Cloud-led certificate program on Coursera and dive into security principles in cloud computing. Master cloud risk management and attack mitigation techniques. Equip yourself with cutting-edge cybersecurity skills.
imp.i384100.net/c/3402234/209...
NordPass - 50% off the Premium plan
go.nordpass.io/SH88h
Join the channel to get access to more perks:
/ @danielpersson
Or visit my blog at:
danielpersson.dev
Outro music: Sanaas Scylla
#jmeter #testplan #performance
Super helpful, thanks a lot for the video, and can't wait for the next one!
Didn't understand the concept of Jmeter. Your video put everything in its place, thanks brother!
great no-nonsense video, helped me a lot.
Past 6 months i was stuck on login, Thank you very much for this video, i am able to login now :) :) You got one subscriber today :)
Thanks! nice tutorial video that walked me through Jmeter effectively !
You just saved my semester, bless you
Summed up with a great example, awesome teaching skills. Very glad I stumbled upon this.
Well explained, thank you so much.
Tack för videon!
Nice video! thanks!
Very simple and to the point. Not beating around the bush 👍
Thanks
I've done some performance testing with HP Load Runner, and I found it's really not something I enjoy doing. I like creating the user scripts, but I don't like any of the other (more important) aspects of performance testing. First, the scenarios (or scripts) that you create need to accurately reflect the traffic to the application. Next, you need to properly balance the scripts so that you have a realistic mix of transactions. And finally, interpreting the results always felt like doing a whole lots of stats.
If you don't get the balance of transactions right, you might end up tuning for the wrong load or over-sizing the system, which results in either outages (because it's tuned for the wrong thing), or a lot of wasted resources.
It's also a common issue that the performance testing environment is not configured the same way as production, which severely limits the value of any performance testing. One really needs to do performance testing in an environment configuration that mirrors what's in production. It's usually suggested that you can scale the test down to match the environment, but in practice that's of limited use, as halving the memory (for example) could have much more of an impact than just halving the performance, depending on how that memory is really being used.
I enjoyed your video, I think it was a nice introduction to JMeter. Please don't take my comments as a criticism :) Performance testing is a really important aspect of software testing and development, and it never gets the attention, or time, or resources that it deserves and needs.
Hi Aaron.
Thank you for your input and experience. You are right that performance testing is a full topic in itself, and some people specialize in it for a reason. I would say I know enough about it to be dangerous, but I can't cover it accurately.
In this case, we had some issues at work, and I did not want to check how much traffic we could take to scale our environment. I was more interested in seeing which calls took the longest and where you could improve the code to handle a larger load. Running the tooling, we found a pain point that we could explore.
But you are right trying to figure out the right amount of equipment or scale values for your cluster is much trickier and requires that you do it in a production-like environment.
Best regards
Daniel
You showed csrf token with the /login request. And how was it handled in jmeter? I don't really get how did it work - you didn't paste any token with /login request
Hi Nokejunky
Thank you for watching my videos.
You are correct, in jmeter you don't have to specify all parameters and cookies you are using. If you have some data that you want to reuse, copy to the next request or connect with another context. Then you only need to describe the relationship or add a handler for that type.
For instance the cookie handler will receive all cookies and return them on the next request. A return value could be reused as a path parameter even if you don't know the value. You only need to mapp it.
I hope this helps
Best regards
Daniel
Gr8
One request,, can you please create some more video on Jmeter which includes - How to read reports, explaination on the same.. Any practial example which shows how to get the issue and what will be the solution.
As jr performance tester what are the different skill required?
Hi Umesh.
Thank you for watching my videos. I'm not sure if I can talk that much about the reporting feature. I seldomly look at that result, I just use jMeter to produce traffic that I could see the impact of on the server side. So for instance I create a set of 10k users doing 200 request against a service does it handle it gracefully and could I improve the code to do a better job.
I hope this helps.
Best regards
Daniel
@@DanielPersson Thanks a ton!
How would I set up my testing if my website is an HTTPS website rather than locally hosted
Should be easy to handle. You can just add the extra S and things will be handled for you. There might be a requirement to add a certification handler. But the functionality is there.
@@DanielPersson I'm kind of new but would u have a video or any idea how I could access any website using a 443 port?
Hi Stello.
Well if you use HTTPS when you visit a site it will automaticly use port 443.
Adding a handler will save and check the certificate. But I don't think you need to do any extra work. I've tested our production site with the method above and it ran under https and the same method applies. So this video should be enough for testing any website.
Do you have any specific issue with testing a secure site that I can look into or help you with?
Best regards
Daniel
hello Daniel ..good explanation .. how can test the performance of local blockchain network ?
Hi Mohammad.
It all depends on what you want to test. If you want to test the transaction latency that could be done via a socket test. But if you want to check how much data could be processed by the network you need a way to create dummy data and that could be done with some scripting but requires some knowledge of the blockchain network protocol.
Thank you for watching my videos. I hope this helps.
Best regards
Daniel
@@DanielPersson thanks for response .... i have local blockchain network it consist of 7 nodes i need to test performance ... how can do that ?
Hi Mohammed.
What kind of metrics do you want? Is it how much data you can send over the network of nodes, or how many blocks can be handled per second? Do you have any other measurements that could be interesting to check the performance of the network? Without more specifics, it's hard to give you any advice.
Best regards
Daniel
hey, how can 20 users use the same username and password? does the website not use the session system?
Hi Abdullah.
Thank you for watching my videos.
I might have been a bit unclear in my video. When I talk about users, I talk about user sessions, and a user could be logged in from 20 different computers at once. Nothing stops you from creating more than one session.
There is a facility to create new users with unique usernames and passwords so you can test the system with entirely different accounts, but that was not the main focus of this video.
I hope this helps.
Best regards
Daniel
I’m curious how did the jmeter know that you are testing that website? I did not see you copying a link for the login but you performed copying something in the address bar when when you click on the issues. Sorry I didnt get it
Hi Camille
Thank you for watching my videos, and thanks for the question.
At 3.40 or so in the video, I create a configuration setting called Request Defaults, where I add things that should be true for all calls, and in that setting, I add the address localhost:3000 which is the site I want to test. I might have talked about this briefly, so it could be easy to miss.
I hope this helps.
Best regards
Daniel
Thank you so much! I hope you keep on making tutorials for Software QA. You are a big help!
Hi, please make another video where you explain the thing you talked about at 2 min 40 seconds.
Hi Art.
Thank you for watching my videos. I've started to look into a video with that topic but I have not concluded with a good example that explains the process so it might take a while.
Best regards
Daniel
Great, Can you do a video for android application load testing. Thanks
Hi Husney
Thank you for watching my videos.
I have no experience with Android load testing. Sounds like an interesting concept, have you looked into it? Do you have any suggestions on where to start?
Best regards
Daniel
@@DanielPersson yes I hv a idea if u wish I can help u
@@husneyhaneefa9075 Can you help me about android app testing? When I connect my phone with proxy, it disconnect my mobile from network so that I cant record test scenario. Please help me
@@rounakaftab8642 same problem here did you find a solution?
Are you from Sweden or Denmark ?
Hi Jamie
Thank you for watching my videos..
I'm from Sweden.
Best regards
Daniel
good,come from china
Hi David.
Thank you for watching my videos. Glad you liked it and I hope you learned something.
Best regards
Daniel
Just show a simple call first. Instead of that specific case is my suggestion