First before we get into it you might be asking — what is Traffic Driven Testing?
Traffic Driven Testing utilizes real user sessions and production traffic to automatically generate realistic test scenarios, mocks, and environments to validate Kubernetes applications and microservices through accelerated performance, resilience, and functional regression testing.
INDEX
This testing method is becoming more popular due to the rise of cloud computing and containerization.
Why?
Kubernetes Testing
More and more companies are building on top of Kubernetes and many folks I speak are not be familiar with testing these types of environments.
On top of this testing complex Kubernetes applications presents many challenges.
Like figuring out proper test data to configuring realistic test environments.
Also, constant changes to microservices also make it hard to keep test scripts up to date. This leads to slow test creation, false test failures, and assumptions that don't match reality.
What approach should you take in this case?
A traffic-driven testing approach helps solve these problems.
How?
By using real production traffic to generate better test scenarios.
Sound cool – right? But how do you start?
I spoke with Nate Lee, Co-Founder and Chief Product Officer at Speedscale, to learn how their tool helps teams test Kubernetes apps with more speed and fidelity.
Realistic Testing Scenarios from Live Usage Data
First, you might be wondering – how does this work?
Speedscale uses a “sidecar” proxy to listen to traffic hitting a Kubernetes application and capture backend calls. This production traffic then drives the testing process – from mock services to test environments to test scenarios. Teams can then replay real user sessions at will and multiply traffic to simulate load.
Pretty slick!
This approach leads to testing that closely matches production reality, without a lot of manual script creation or upkeep when services change.
This approach also helps with debugging.
Accelerate Debugging with Isolated Traffic Replays
When issues arise in complex microservices architectures, it’s tricky to pinpoint the root cause.
By replaying traffic and swapping in mocked services, teams can quickly isolate components to narrow down problems. Speedscale co-founder Nate Lee calls this “extracting any plane on the aircraft and sticking it in a wind tunnel.”
You can also get performance data quickly as well with traffic driven testing.
Rapid Performance Testing of Code Changes
Making a code change and wanting quick performance testing feedback?
By capturing production traffic, teams can replay the same scenarios each time and compare metrics like requests per second and latency before and after changes. No need to manually script new performance tests every time.
Test Environment Self-Service
One big killer of automation efforts that I’ve seen is not having proper testing environments.
And as I point out in my post The Four Destroyers of any Automation Testing Script is lack of test environments and data are two of the biggest contributors to failure.
But with Speedscale, the sidecar captures backend traffic so engineers can build the exact test environments needed to exercise specific application flows.
These disposable test environments spin up quickly on Kubernetes without cutting tickets.
You’ve probably have heard a lot about shift-left testing and this approach can help here as well.
Shift Testing Left with Confidence
With the ability to generate test scenarios that match production and configurable test environments on tap, teams can shift testing left in the development cycle. Even short master builds can run key performance and functional tests derived straight from customer usage via production traffic capture. Developers get rapid feedback, while QA has confidence tests reflect reality.
While traffic-driven testing is not a cure-all, this paradigm flip shows a lot of promise for overcoming longtime testing challenges. By tapping into actual production usage, test automation can better keep up with the rapid pace of Kubernetes application development.
Give Traffic Driven Testing
Sound good – right?
Why not try it for yourself?
Get started with Speedscale and see for yourself the power of traffic driven testing for more resilient and performant Kubernetes applications.
