Observability meets a Flaky Test

“Observability”. A word that is quite popular in tech these days, but not always easy to understand what it means in the real world, right? “How is it not monitoring?” I mean, we’re told that it’s about “unknown unknowns” and that you should be able to answer new questions about a system without having to ship any code. We’re also told that there are three pillars for observability, “logs, metrics and traces”, and that it’s much more than just monitoring. But how does that translate to real-life scenarios from a software tester’s perspective? Well, one day my team was struggling with a flaky test in our CI/CD pipeline. The way we were able to unravel the mystery surrounding that test illustrates a few key observability concepts on the availability of data and the friction in accessing it. In this talk I will tell you the story about that painful flaky test and how it showed us how (the lack of) observability could already be present in our daily lives, without us even realizing it! By the end of this tale, even if you know nothing about observability beforehand, you will understand a bit more. You’ll know some questions you should start asking about your own tests, and have some ideas that will let you find the answers quickly!