With the peak traffic season starting up – summer travels, back-to-school shopping, the highly anticipated Black Friday and holiday season – we know you’re asking yourself, “Can my application handle it?”
In our latest webinar, “Taking the Stress out of Stress Testing,” we discussed how to get started with performance testing, simulate realistic load tests by recording on real web browsers, and get browser-based metrics. There were many questions asked during the session regarding performance testing, so hopefully this blog post answers some of the ones you may have!
Is peak performance determined as the optimal number of users that can be loaded without response and performance being affected? What test contributes to this? Capacity? Spike?
The performance metrics we observe ultimately depends on the test being run. In the event that we wish to monitor the maximum number of users that a web application can support without Service Level Agreements becoming invalid, then a capacity test should help gauge in this detail.
Read more about the different types of performance testing, such as spike, load, and soak testing.
Should a well-defined test case always end with a final state of n+1? Why?
A well-defined test case should be one that can be replayed after recording. The principle of induction was an analogy to explain why we need to play back before we use a web test to run a load test for multiple, concurrent virtual users. The principle is useful when we observe a pattern and want to prove it.
Can we keep LoadNinja scripts locally for organization audit purposes? Or maybe use a source control tool like BitBucket?
Currently, LoadNinja’s artifacts are stored on an external cloud.
How will the load testing be done if you only recorded one case?
We record a transaction once, and then replay it within the recorder to ensure that it can be used in a load test (for multiple concurrent virtual users). After the load test scenario is configured, a load test can be run at scale.
Read more about our record and playback feature.
Does it allow to add check points (text checks) and transaction names? What kind of enhancement does LoadNinja support?
Yes, LoadNinja supports text validations with documentation detailing those validations. You can find document around this support on our LoadNinja Documentation Page.
When a page is being tested and both success and failure are the same page, how would LoadNinja know success from failure?
LoadNinja allows you to debug virtual users in real time and isolate issues away from the successful actions on the page. The Debugger shows where the error occurred and what the related response code is.
Check out this documentation, which further specifies how the Virtual User Debugger assists in this regard.
If we have multiple scripts for different scenarios, can we do bulk updates to those since the scripts are different for each scenario?
If you update the script used for a scenario, LoadNinja should be able to run the test with the updated web test when you next execute the load test scenario.
Can we add Database assertions in LoadNinja test, if Database is in Organization Network?
LoadNinja currently does not integrate with databases at this time.
Check out how you can integrate LoadNinja into your work process by using various technologies and tools.
Can I extract the routes recorded to a well-structured CSV, code, or any kind of file?
You may data-drive the initial URL that is captured from the data uploaded through a CSV file. This is useful when taking a web test recorded in one testing environment (i.e. dev, QA, test) and ensuring that it can be used in another environment (i.e. production).
Can we run iteration-based test for specified duration? For example, 100 transactions in an hour.
We can only configure that a test runs for either a designated amount of time or for a specified number of iterations.
If you’d like to learn more about LoadNinja, check out our features page! You can also start Load Testing now with our free 14-day trial!