Azure Load Testing Managed Service First Thoughts

In a recent announcement, Azure Load Testing has now became GA and in this blog, I will be giving a brief overview of using it along with first thoughts!

Having used a number of Load testing suites throughout the years, I was very curious when Azure Load Testing was announced as a fully-managed Azure service!

Sounding like another great Managed Service offering by Microsoft Azure, I was super keen to try it out!

With cloud consumption ever increasing in organisations, Application performance and resilience is key – we are continuously wanting to reduce downtime while constantly increasing performance and having environments running as efficient as possible, all within a budget.

Why is load testing important?

You may be wondering why is it actually important, is it really needed?

Load testing allows you to simulate real-user situations and scenarios – You can simulate situations with Load testing in terms of user journeys and even user count. Find any issues early with a load testing suite, rather than the end user. End users can be very unforgiving!
Applications can perform very different when under a load – Having a load on your application, you may come across a memory leak or performance issue that doesn’t appear until the application or environment has been greatly utilised.
Creating a code change to your application make result in unexpected behaviour – You could be applying various code changes to your application than may cause issues to the performance or end users journey, successful load testing results will mitigate this.
Forecast scalability of the end solution – You can forecast the costing of the end solution with load testing – seeing how the environment may scale and even approximate costings of when the scaling does happen
Happier end users – Nearly 40% of users leave the site if it takes more than 3 seconds in load. Imagine if out of 1,000 users – 400 don’t return because of poor performance that would have been caught with a successful load test?

Azure Load Testing

With Azure Load Testing being a fully managed load-testing service, you have no need to worry about scaling the testing suite when required – a great start already!

Like all load testing suites – this resource allows you to simulate traffic to your application and environments, monitoring user journeys and tweaking settings to increase end user performance. All this ability within one resource!

How can Azure Load testing be used?

Quite a few options available including:

Regression testing is possible within your CI/CD pipelines, something I will blog at a later stage.
You can create a load test by reusing currently created JMeter scripts
URL based load tests also available
Load test results can be reflected in various metrics including response time & requests/sec. With JMeter scripting, even deeper metrics are available

Running your first URL based load test

Lets run through at creating your first URL based load test, create a new Azure Load Testing resource as below

Once the resource has been created, select tests -> create -> Create a quick test

My test url is https://thomasthornton.cloud/ with default settings for now (as this is an example)

Click run tests & allow the test to run, once completed successfully you will be able to view the results – by selecting Tests & then select your test, in this example Get_thomasthornton.cloud

Example of output from results below, you will be able to begin drilling down into your application, to understand if there is currently any issues, performance related or errors appearing during this simple load test.

Comparing test results

Running more than one of the same test, you can look at comparing results, potentially notice trends and trying to spot any abnormal behaviour.

Select the tests you want to compare like below and then select compare

Comparison below of the two selected tests, as mentioned – you can then begin reviewing different performance metrics between multiple tests. When using JMeter tests, you can have been deeper metrics to review, allowing you to find trends, issues and potential to increase stability and system performance in relation to this.

What next & final thoughts?

The list can be somewhat endless, I have showed very brief usage to what is capable with this awesome resource! It is going to be a great tool in many company tool boxes going forward! Fully managed is a huge bonus, allowing you to load test from the get-go, resulting in faster and more efficient deployments and applications – resulting in happier end users.

What next? I really recommend you check out the following: