User these days go online for almost everything from getting in touch with friends to finding a job to buying stuff. And with more and more people accessing the web services, performance really becomes a good check point. Nobody likes to click a link and then wait for like ages. In-fact most people stop visiting a site if it takes longer than about 10 seconds to load a page.
So how do we check the performance of the web service(website/web apps)? Well there are a lot of tools available out there for it. One such tool that is also open source is Jmeter.
Now there are already so many tutorials on how to download and set up Jmeter, so you can just go ahead and Google it. So now lets talk about how to use Jmeter the way I do.
Jmeter also has a built in recorder, but there's a difference in how it works and the way tests recorded with Blazemeter work. You see jemter simply takes all the http requests and hit them onto the server all at once. So if you have recorded a test with built in recorder, what it will do if get every individual http request and treat it independently. So when you're testing a web page for a load of 100 users, what it will do is get all the images links and referenced files and make 100 requests for each of them independently. Whereas, Blazemeter groups all the links in a page so that it generates more realistic load since in this case its like 100 users accessing a web page simultaneously rather then accessing individual image or link from that page.
So, that is how I like to do performance testing. I record all my test scripts using Blazemeter's Chrome Extension and then import those scripts in Jmeter and generate the required load.