I have seen in Load Runner that there is a graph which talks about transaction success and failure. I am assuming that transactions here means certain business functionalities like: login, saveX, addY, deleteZ, etc. Is this understanding correct ?
I was trying to see if something similar is provided by OpenSTA, but I see that in OpenSTA we only talk about Transaction Timers, which represents the amount of time taken to complete a transaction. It doesn't talk about success or failure of the transaction. So how we can identify a transaction was successful or failed because of some reason.
I take a simple example of login transaction: When a user logs in, the transaction is considered successful when the user can see the home page of the application and the transaction is failed if the user sees the same login page with a message like 'Username/password incorrect'. Does OpenSTA provides any means to provide that this login transaction was successful or it failed ?
There is another concept that I see in OpenSTA and thats related to "HTTP transactions", which seems to refer to the HTTP response code received from the server. So if I receive a response code from the server which represents 'page not found' or any error code which means that there was some problem in getting the response from the server then it means that the "HTTP transaction" failed. This information is anyways provided by the HTTP Errors Vs HTTP requests and HTTP errors Vs Elapsed time graphs. But I rarely see these two graphs as helpful because if I do the recording correctly then chances are slim that I will end up with HTTP errors ( unless the server has to bear an unbearable user load ). Any comments on this observation ?
Re: Transaction success/failure
Read about verifying results (pointer in sticky post for this forum) and test cases. Response codes alone are not sufficient to determine success or failure. You will need to load the response into a variable, search for a string which indicates success or failure, and then take some action. Its a manual process, and one you have to reinvent unless you use a script pre-processor to add this sort of code to your scripts.