Intro & motivation

When I started this project in August 2019, very few documentation about 'Testing as a Service' I found, but no insights or implementation ideas whatsoever.

I did not go on this road having the classical definition of TaaS as 'an outsourcing model, in which testing activities are outsourced to a third party' , but rather create the infrastructure that allowed me to respond to these questions:

• what if everyone can have a complete testing environment with one press of the button, in few seconds / minutes ?
• what if everyone can design better tests, always having clean environments?
• what if the test regression does not last days, but rather hours?

That's power, right?

During this road I started with docker, having in mind 'dream big but take small steps'.

I strongly think that the final destination of this road should be Kubernetes. The estuary stack is compatible with K8s.

I cared about the maintenance aspect over time. It helped me to design stateless, layerless and pure services.

And to conclude, I do not pretend the stack is flexible and powerful, but rather I say: convince yourself.

Why estuary stack

• Language agnostic

Run Python, Java, Nodejs ... it's up to you. The Estuary stack supports any programming language.

• Docker / Kubernetes / Cloud ready
• Granularity and flexibility

Imagine you run a test, a test class, a suite, multiple suites.

• Fast result gathering

Spawning multiple deployments will get you the results in no-time.

• Debug accelerator & Standardization
• Simple test execution with only a REST client: Java/Python … REST client, CI/CD, Web/Desktop/Mobile app

Prepare

Deploy start

Deploy status

Test start

Get results

Deploy end

• Horizontal scalability

Run many docker/kubernetes environments in the same time. Sky is the limit.

• Powerful logging with FluentD. Analytics over time.

Estuary Stack maintenance paradigms

• Client maintenance first

Estuary stack uses the principle: client-first in software maintenance over time.

The stack is built on the principle: do everything from client, do not bother with artifacts/layers that exists on server side

• Env vars first

Configure the stack from env vars. This principle works very well in container world: docker/k8s/cloud.