For present day needs, it is pretty difficult to scale/maintain a system built in "monolithic" way. Micro-services, is an architectural style, a development style of building enterprise application comprising of small, independent, customer focused, scalable services. Docker & Kubernetes are kind of enablers for microservices architecture. At this juncture, test automation is a very important requirement for a successful, scalable microservices architectured applications.

How do you fasten the process of adding right test cases for microservice architectured applications & automating them? Anwser is, 1. invest less time on repeatable activities (like failed tests debugging & root cause analysis- these should be automated). 2. Get Continuous data which can result in adaptable actions which covers business value. In this talk, we will explore about strategies to test a distributed microservice architectured application by analyzing API flows using trace IDs and build test strategies based on the coverage & unexplored areas. We will effectively make use of both Selenium Logs & microservices logs to develop a coherence between them by time series approach, Trace IDs & span ids. This will identify if there are any automation failures corresponding functional issues via infrastructure/framework which we developed.

We will also talk about challenges faced to test, automate & maintain a high resilient test automation framework. We will look into tools & infrastructure(Zipkin, ELK etc) that are required to quickly test & certify a build and to achieve continuous intelligence in testing.

In contemporary test automation world, we are running 1000's tests every day. Even though most of our test cases are reliable and stable, debugging failures will consume lot of time, if not handled appropriately. This is an effort to reduce the analysis of failures and come up with probable root cause for each and every failure.

In this talk, we will explore all the approaches we have taken to drastically reduce the time to debug the failures & let you concentrate more on adding new tests. We will also talk about the approaches which could not yield results for us initially & why. We will look into building a failures' dashboard, driven by pattern based classification of logs(both server & automation tool) with reporting.