When creating a new Microservice you typically need to add a lot of boilerplate to the code, such as logging, metrics, authentication, SSL, secrets/credentials, etc... All this ends up overshadowing the actual logic of the service itself and results in a lot of dependencies and code. Because of this, we at Soluto created a template for the boilerplate, which worked great...
Until we had to upgrade one of the dependencies which resulted in huge effort of code upgrades and deployment across each and every service that uses the template. And that’s besides the fact we needed to make a separate template for each language we used (and over the years, the number of programming languages we use at production increased dramatically).
We needed a way to simplify all this. We wanted to get rid of the boilerplate while maintaining the functions listed above. So we decided to delegate the entire responsibility to the Kubernetes infrastructure. In this lecture, I will do a live coding session and show you how to remove the boilerplate from your code and move it to the infrastructure, and discuss the benefits and limitations of this approach.
15:30-16:00 | Track 2
Not My Problem: Delegating responsibilities to the infrastructure
Designed and developed software for the last decade, passionate about product design, software architecture and developing open source software.
In the last few years I gave various talks on UI, functional/reactive programming and multi-container architecture at internal meetups and public events