Printable View
Imagine a situation where we are running a scenario using 1000 vusers...how do we know how many vusers that application is able to withstand at a single point of time... how do we define a scenario for this.
Tx inadvance
[b]Question asked by visitor Subhashini[/b]
Hi Subashini,
Testing an application with 1000 vsuers can be achieved in the following manner -
Create a scenario with your script and Load generators in place. Once done, select the Scenario settings as desired. This would include -
Ramp Up time; Number of User to be ramped up
Run uptil;
Ramp Down time; Number of users to be ramped down.
Check the runtime settings to ensure that the iterations for each user is set to 1.
Once the scenario begins execution, we can monitor the various graphs to verify the progress of the vusers.
Incase the scenario completes, under Analysis tool, go to Transactions / User, Throughput, Peak Period graphs. Incase you see transactions with passed status for all 1000 vusers, definitely your tool can withstand 1000 users.
But incase we find transactions failing for vusers continously after a specified number of vuser, export the errors column. Incase you find a AppServer or WebServer error enlisted, it means that the application has crashed.
Cheers
If you want to know the user threshhold in which the server can handle, you may want to setup your scenario a little differently.
Ramp up 10 vusers every minute and the have the vusers run continuously until the end of the scenario.
If you are testing a web server, you will start to see 500 Response Error Codes at a certain point. By looking at this error graph with the Running Vusers Graph, you should be able to determine the size load the server can handle.
Apart from everything that has been mentioned before this post, I think one of the easiest ways to do this is by matching the response time graph from the monitor section and matching it with the business requirement. If the response time exceeds the Business Requirement RT, the test has failed.