We all know that containers and microservices should make our applications simple, they should do only 1 thing. With containers and microservices we get rid of some of the complexity within components, however we also add complexity between components. This complexity is often higher than the complexity in our original monolithic architecture, and the real pain comes when you combine state with composition in a distributed architecture. All the things we love about microservices such as restarts and load balancing to scale individual components introduces new potential disasters if not managed carefully!

In this talk I will tell you some rules to help you build and evaluate your fault tolerant architecture with containers and small components. The rules are grouped within 3 sections. Decomposition rules are about how you handle your containers and services responsibilities within the system. Orchestration and synchronization will discuss the more cross cutting concerns of our system. In the end we will talk a bit about the elephant in the room, managing stateful apps.