3 Ways to fail your product launch on Kubernetes

A storm in autum of 2021 overloaded the production environment with requests. All services affected ran in the Kubernetes cluster on Azure and should have scaled up to cover the load. Instead the response times went through the roof.

This case study shows the journey taken to transform this non-stable setup into a well-performing one. Along the way, you will learn what configurations in Kubernetes you should watch out for and what the impacts of changes in these configurations are. In addition you will learn how important profiling is ahead of major launches but also in general.

The journey gives you a realistic depiction of a cloud journey with legacy software and inside an existing environment. You will leave the talk equipped with the tools necessary to prevent the same mistakes that led to the overloaded production environment of this case study.