SoapUI LoadTest strategy "Thread" behaves strangely + TPS
Hello.
I am using SoapUI 3.6-beta1.
After I found out that it was quite easy to transfer XML from previous response to the next request, I have attempted to use SoapUI for load-testing our application. I have some questions about that.
First of all. If I try to use strategy "Thread", with start threads 10 and end threads 100 (or more), I don't get any results at all. All fields are 0 after test is complete (and also five minuters after, I waited to make sure that no threads were still executing). Do you have to buy "Pro" to be able to work with many threads?
Here are some other examples for this strategy which behaves weird.
Start: 1, End: 15 – Works
Start: 1, End: 20 – Does not work
Start: 10, End: 15 – Does not work
Start: 5, End: 15 – Does not work
Start: 1, End: 20 (pressing stop after 2 threads) – Works
Start: 1, End: 20 (pressing stop after 4 threads) – Does not work
What's the general idea here?
The next question is based on the times it actually works to get statistics:
Is there a way to find out the average TPS for all requests being sent, that is the actual amount of transactions the server can handle? What I do now is copy the entire table (as I cannot select column only), paste in Excel, remove all columns except TPS, create a "SUM"-field (adding all TPS's), and divide by the number of requests I have.
I don't even know if this gives a proper value? But I feel that all request TPS's added together, divided by the number of requests, should give a decent average value.
And I also have to remove all "property-transfer-tasks", since they have a very high TPS (of course), but have nothing to do with what the server can handle.
Regards
/Anders
I am using SoapUI 3.6-beta1.
After I found out that it was quite easy to transfer XML from previous response to the next request, I have attempted to use SoapUI for load-testing our application. I have some questions about that.
First of all. If I try to use strategy "Thread", with start threads 10 and end threads 100 (or more), I don't get any results at all. All fields are 0 after test is complete (and also five minuters after, I waited to make sure that no threads were still executing). Do you have to buy "Pro" to be able to work with many threads?
Here are some other examples for this strategy which behaves weird.
Start: 1, End: 15 – Works
Start: 1, End: 20 – Does not work
Start: 10, End: 15 – Does not work
Start: 5, End: 15 – Does not work
Start: 1, End: 20 (pressing stop after 2 threads) – Works
Start: 1, End: 20 (pressing stop after 4 threads) – Does not work
What's the general idea here?
The next question is based on the times it actually works to get statistics:
Is there a way to find out the average TPS for all requests being sent, that is the actual amount of transactions the server can handle? What I do now is copy the entire table (as I cannot select column only), paste in Excel, remove all columns except TPS, create a "SUM"-field (adding all TPS's), and divide by the number of requests I have.
I don't even know if this gives a proper value? But I feel that all request TPS's added together, divided by the number of requests, should give a decent average value.
And I also have to remove all "property-transfer-tasks", since they have a very high TPS (of course), but have nothing to do with what the server can handle.
Regards
/Anders