Retain databank mapping when a databank is updated
When there are multiple databanks configured and one of those databanks is replaced, all the data mapping gets wiped off from the script. This causes rework. It would be great if we have to update mapping for only the file which was updated and not redo mapping of already configured databanks all over again.Support For Mobile Testing
Since the majority of users hitting a web page are mobile users, adding support for mobile testing is necessary. Of course, there are various levels of mobile testing support - and actually different types of support. I would think that just having the ability to change the browser string so that the load appears to be coming from a mobile device, and maybe having a few bandwidth selections to simulate mobile traffic might be eough initially.Implement compaund Web Test
Dear colleagues, I am pretty sure that compaund Web Test is in your plans for the next LoadNinja update. The purpose of compaund Web Test is similair to the Call Scenario operation from LoadComplete and its purpose to make it possible to create reusable parts of Web Test. For example, assume that I need to load test a scenario when 80% of VUs just browse some goods catalogue and only 20% of all VUs do a purchase. Current implementation of LoadNinja requires that the browse sequence must be recorded (and supported) twice - for browse test and for browse and purchase one. Compaund Web Test should eliminate this duplication.Provide Improved Databank Handling configurations
DISTRO: Option 1: Databank(s) is automatically split into the same number of files as the number of load generators. Each load generator gets its own fraction of the original file. This is the most common use case. A row of data is UNIQUE across all load generators. Option 2: The Databank(s) is sent in it's entirety to each load generator. A row of data is only UNIQUE for a load generator - not across all load generators. END OF FILE BEHAVOR: Option 1: Upon reaching the end of the databank, start again from the beginning of the file. Option 2: Upon reaching the end of the databank, stop the test. DATABANK ACCESS: Option 1: Sequential Option 2: RandomFilter Out Third-Party Hosts
Provide the ability: to see all hosts captured as part of the recording to select which 3rd-party hosts are used (uncheck host to filter it out, check to re-activate) Add: List of Active hosts to the SUMMARY tab in the realtime report List of Inactive hosts to the SUMMARY tab in the realtime report List of Active hosts to the actual report List of Inactive hosts to the actual reportColor Coding *Keyboard Input* items within a script
Actually, I think various types of items should be color coded to stand out. Keyboard Input items should stand out because they will often be edited in some way. For example: Today, I recorded a script and while recording, took a couple of missteps. I had a typo in the password field and had to backspace and retype (twice!). It was easy to go in and edit the initial password to be perfect and then delete the other two password-related items. What took the longest, was LOCATING the keyboard items ;) They are not hidden, but do not stand out.SLA's
Enable validations around the concept of SLA's. Example: In a scenario, you select that Steps can not average greater than 3 seconds response time and no step can have a response time > than 5 seconds. If exceeded, an SLA violation error is thrown and the script continues. Reports should have an SLA section that highlights compliance with the selected SLA.Allow testers to select Cloud Regions to run load generators from
The ability to select, and combine cloud regions is essential. Example: US-based Retailer wants to run a test representing their customer base. 50 percent could be generated in US-EAST and 50 percent US-WEST. Example: Australian company wants to run tests. The load should be generated from AUS as anything else will add unrealistic latency.Test Script Run all vusers sequentially.
Could you give us the ability to run all vusers in test script . When I create a test script i have 100 vUsers which login with different user name and password from a databank. I would like to be able to verify that all 100 vUsers can execute the test script test without errors, thus to verify the correct execution of the steps. Then I can run my scenario using the 100 vUsers knowing that I won't have any failure due to incorrect action on a step. Currently, this verification is performedmanually, which takes a lot of time. In addition, we are having a specific amount of hours per month or year and it does not sound very fair to utilize this time simply to verify that the 100 or more vUsers will pass the test script steps. I hope this will be beneficial to the rest of the LoadNinja users.Ability to call Services external to the site under test
It would be nice to have a step type or something similar that can be added/inserted with the purpose of calling an external service for its response. An example of this would be calling a hosted test service that provides a set of customer details which you can then use in the test to avoid hardcoding values and allowing the test to be more dynamic and thus more maintainable. Another example that comes from my usage in calling out to a hosted test service that has helper functions for the tests, things like GenerateRandomString, GenerateRandomAddress, etc. these can be done via a JS validation but with no centralised storage of JS code we have them hosted external to LoadNinja and call out to it during a test.Global Parameters/Databank
It would be nice to have a global set of parameters that can be referenced across multiple tests, I currently would want to use this to be able to replicate an ability I had in Visual Studio to set the source url at the top most level and it affect every test in the whole load test run so that I can easy swap between environments to be tested. Currently this would require me going into each individual test and editing them
Allow Custom Think Times
There are moments in scripts where I need extra think time to wait for elements to appear. Setting a 'Fixed' time would only add unnecessary time to the entire script, and setting it to the 'Recorded' times is too inflexible. I'm not even sure why you are even able to edit and save the values, as the script ignores them completely. There should be a Custom setting which has the script follow the edited times. These times can be pre-filled with the 'Recorded' or 'Fixed' times it came from.Script / Step Filter for Load Generator Performance Graph
The Load Generator Performance graph has been very helpful in letting us know when the load gets too high, but it would be even better if it could help us figure out 'where'. If it's possible to keep track of which script and/or step is running when bytes are sent and received, it would be very informative to be able to filter this graph (in a stacked manner like other charts) by scripts and steps.
Create a sanbox of 20 users where the tests can be run prior to being committed to the real test.
I have run the load tests twice only to find out the csv had errors in it. There isn't a way to run the script to test the script until it is put into a load test scenario which takes away from allotted time for testing. The 3rd time it ran smoothly. Would be great if there was a way to test the script before committing it to Load test. The playback did not go through all the csv data entries to have them verified.Add Ability to Handle Video Streams
Two Ways to Choose From: Accept the stream and load the same during playback Only accept a certain (selectable) number of bytes during testing. Solution 1 would be able to test the strain that video streams put on the infrastructure and bandwidth. Solution 2 would take into consideration that most of the time a video is not watched all the way through and the user moves on to something else.
LoadNinja Feature Requests
36 Posts
