I recently started using LoadUI and I have couple (actually many) questions on its functionality.My basic objective is to test the Performance of couple of Web Services. I would be preparing the scripts using SoapUI Pro and would be runnning Performance testing scenarios using LoadUI.PFB my Questions:1) Are SoapUI PRO / Load UI and Load UI agents' supported in Solaris OS (compatibility)2) What is the PORT used for communication between LoadUI controller and Load UI agent.3) How do we set up a Load Test scenario by importing multiple test cases from SoapUI Pro5) How do we distribute the user load (virtual users) among these test cases 6) How do we set up Ramp up / down plans for performance test execution7) How do we pin (capture) the server counters (AUT) in LoadUI How do we Add User Defined Transactions in scripts9) How do we access different graphs (for analsis of performance test Results) like: i. Concurrent users ii. Transaction summary iii. Hits per Second vi. Throughput v. Server Counters data capturing It would be great even if I am provided with a document or a URL answering these questions.

Hi!Your use case sounds like a perfect fit for loadUI+soapUI (honestly!).ValueLabs wrote:1) Are SoapUI PRO / Load UI and Load UI agents' supported in Solaris OS (compatibility)All of these should technically work in Solaris, but we don't officially support Solaris.ValueLabs wrote:2) What is the PORT used for communication between LoadUI controller and Load UI agent.Port 8443 is standard, but it can easily be changed.ValueLabs wrote:3) How do we set up a Load Test scenario by importing multiple test cases from SoapUI ProBasically, the only thing you need to do is to use a soapUI Runner in loadUI and point it to your TestCase. Read more here.LoadUI 2.0 will have larger focus on soapUI integration and features a new soapUI Runner that integrates even better (visually and technically). It will for example make it easier to transfer sessions between soapUI Runners (e.g. Login, Delay, Search, Loop, Split based on probability, Logout or Purchase etc.). The first beta of 2.0 will be released in about two weeks.ValueLabs wrote:5) How do we distribute the user load (virtual users) among these test casesYou do this by using Generator components and route wires to the soapUI Runners. All this is covered in the documentation. In time for the first beta of loadUI 2.0, the documentation will be extended and contain much more video -- but the current documentation should definitely cover your question.ValueLabs wrote:6) How do we set up Ramp up / down plans for performance test executionThere's a Ramp component and we're currently considering adding an additional component just for this.ValueLabs wrote:7) How do we pin (capture) the server counters (AUT) in LoadUIIn loadUI 2.0 Pro, server monitoring is built in. Statistics is automatically saved in local databases, can be compared and mixed with Virtual User statistics.ValueLabs wrote:8) How do we Add User Defined Transactions in scriptsCould you please clarify this? You can load test any soapUI TestCase in loadUI, and soapUI is very flexible and features a lot of scripting capabilities. You can also build your own components and server monitors in loadUI using the same scripting language.ValueLabs wrote:9) How do we access different graphs (for analsis of performance test Results) like: i. Concurrent users ii. Transaction summary iii. Hits per Second vi. Throughput v. Server Counters data capturingThese statistics are all accessed from the Statistics Workbench in loadUI. This is all covered in the documentation here.Hope this answers all your questions!HenrikSmartBear Software

Thank You for prompt reply. Yes, most of my questions are answered.Clarification on My Question 8 (How do we Add User Defined Transactions in scripts): Basically what I mean is, are there any functions like "Start_Transaction {Admin-Login}" and "End_Transaction {Admin-Login}" which would have a set of requests (from LoadUI to AUT Server) in it. So in the final report I would hope to see like "Admin-Login = 0.96 secs" (Avg. response time) Apart from this, can you please let me know how could we change the default port settings to change 8443 (PORT used for communication between LoadUI controller and Load UI agent) to a desired value (we still dont know what all PORTS are open on Environment Under Test)Now, I am eagerly waiting for LoadUI 2.0's release . Sounds like this would make Performance testing even more interesting and exciting.I am attaching a sheet comparing features of Loadrunner, WAPT, LoadUI/SoapUI (Just for reference, might help you guys getting new ideas ). To my knowledge, most of the information provided in it is correct (Sorry!! if you find any discrepancies)

ValueLabs wrote:Clarification on My Question 8 (How do we Add User Defined Transactions in scripts): Basically what I mean is, are there any functions like "Start_Transaction {Admin-Login}" and "End_Transaction {Admin-Login}" which would have a set of requests (from LoadUI to AUT Server) in it. So in the final report I would hope to see like "Admin-Login = 0.96 secs" (Avg. response time)If you design the transaction in soapUI, and use the soapUI Runner in loadUI, statistics are recorded both for the whole transaction and (optionally) the individual steps.ValueLabs wrote:Apart from this, can you please let me know how could we change the default port settings to change 8443 (PORT used for communication between LoadUI controller and Load UI agent) to a desired value (we still dont know what all PORTS are open on Environment Under Test)It's explained here, in the Agents and Distribution page of the documentation.ValueLabs wrote:I am attaching a sheet comparing features of Loadrunner, WAPT, LoadUI/SoapUI (Just for reference, might help you guys getting new ideas ). To my knowledge, most of the information provided in it is correct (Sorry!! if you find any discrepancies)It seems mostly correct to me A few comments/discrepancies though:Web recording is available in soapUI (and therefore loadUI), although it's limited.Regarding Transactions, see my comment above.Rendezvous points can be made as a custom component; but if that's not an option, loadUI doesn't support it yet.Ramp up/down plan can be achieved in loadUI by combining schedulers and generators (ramp and fixed rate). It's not as easy as it could be in the case of rampUp-steady-rampDown (implementing a component for solely this purpose is actually in our backlog, but I won't promise that it makes it into loadUI 2.0).I would definitely say that loadUI supports adding periodic load to the AUT, this can be done with either scheduler components, variance generators, or any combination of these.Adding HTTP headers can be done in soapUI (and therefore loadUI with the soapUI Runner).Regarding the comment Load UI [...] have very less no. of analysis graphs available -- have you had a look at the Statistic Workbench in loadUI (each runner component automatically records 28 statistics)? If so, what do you think is missing?RegardsHenrikSmartBear SoftwareHenrikSmartBear Software

Thanks for providing the clarifications.Regarding the Analysis Graphs: I just had a glance at Workbench component and I felt there is no much information about statistics (graphs) on Errors during the execution e.g. Errors per sec, type of errors etc.Actually I have note started working on Analysis part (Workbench Component). Only after digging into it completely I would be able to give proper feedback.

ValueLabs wrote:I just had a glance at Workbench component and I felt there is no much information about statistics (graphs) on Errors during the execution e.g. Errors per sec, type of errors etc..All Runners have Failures statistics (Total and Per Second) [Statistics Reference]./Henrik

Questions on LoadUI | SmartBear Community

16 Replies

phanisrikar
Contributor
14 years ago
If that would be easy to develop custom component for rendezvous point, then it will definitely help us in the below scenario

We are trying to check the GET and UPDATE methods of REST webservices are working properly when more than 200 - 300 users send the request to the server at a time.

and if we run the test in distributed mode, does the test case properties will be logged in table log component ? because it is currently not logging the any properties. Please let us now more info/issues on load testing in distributed mode

Thanks
Phani

SmartBear_Suppo

Alumni

14 years ago

phanisrikar wrote:
If that would be easy to develop custom component for rendezvous point, then it will definitely help us in the below scenario

We are trying to check the GET and UPDATE methods of REST webservices are working properly when more than 200 - 300 users send the request to the server at a time.i

phanisrikar wrote:
If that would be easy to develop custom component for rendezvous point, then it will definitely help us in the below scenario We are trying to check the GET and UPDATE methods of REST webservices are working properly when more than 200 - 300 users send the request to the server at a time.i

Sorry, could you explain more in detail where in this scenario that a Rendevouz component would be handy.

phanisrikar wrote:

and if we run the test in distributed mode, does the test case properties will be logged in table log component ? because it is currently not logging the any properties. Please let us now more info/issues on load testing in distributed mode

phanisrikar wrote:
and if we run the test in distributed mode, does the test case properties will be logged in table log component ? because it is currently not logging the any properties. Please let us now more info/issues on load testing in distributed mode

Shouldn't be any issues. I assume that this is loadUI 2.0 beta1? Any messages in the agent terminal?

Do you see other things in the TableLog, such as TimeTaken? If not, what if you connect a Fixed Rate Generator to a TableLog?

Regards

Henrik
SmartBear Software

ValueLabs
Contributor
14 years ago
This component looks like an ideal suit for scenario that requires Ramp up - Steady state - Stop (hope the "Stop" here is not abrupt but it follows a steady ramp down pattern).

My Question-

Suppose I set:

Ramp Duration: 20
Peak Rate: 60
Peak Duration: 120

What I understand from the above set scenario is, in first 20 secs of the test, the no. of requests are ramped up to 60 / sec and then this rate (60 / sec) is maintained for 120 secs, then the test stops.

Is this right?

Because during the test, I feel the request rate is going well beyond 60 / sec during first 10-15 secs of execution.

Thanks.
SmartBear_Suppo
Alumni
14 years ago
ValueLabs wrote:
Because during the test, I feel the request rate is going well beyond 60 / sec during first 10-15 secs of execution.

Can you elaborate "I feel the request rate"?

What if you add a chart in the Statistics Workbench for the Runner's Sent requests per second? That is: Sent > Per Second. Does the line look like the one in the Ramp Sequence Generator's documentation?

Regards

Henrik
SmartBear Software
ValueLabs
Contributor
14 years ago
I mean the display in ramp sequencer shows the correct request rate (it follows ramp up - steady state - ramp down) but the requests running in the attached runner doesn't match with the ramp sequencer.

E.g: In the attached screen shot the ramp sequencer says 45.1 requests are being sent out to runner but the runner shows 0 requests being processed.

So am I doing anything wrong?
Issue.doc150 KB
mbalajiwale
New Contributor
14 years ago
Hi,

How do we create multiple users in Load UI? Which component is used for that purpose?
Also, under VUGenerators, we are using 'Random' component and setting the value for 'Base Rate' say 10, 20, 100 and so on. So is this Base Rate value number of HTTP requests per second sent to the server or is it number of users per second?

Please confirm/correct us.

Thanks

Forum Discussion

Questions on LoadUI

16 Replies

Recent Discussions

Unable to authenticate license in Ready API Login button unresponsive

Assertions Panel issue to update assertions since update 3.63

ReadyAPI websocket plugin issue

Related Content

TestComplete interview questions

Userforms Groupbox question

Data generator question