There have been amazing articles on the subjects of migrating from a monolith to a microservice architecture, e.g., this is probably one of the better examples. The benefits and drawbacks of the architecture should be pretty straightforward. I want to talk about something else, though: the strategy. We build monoliths since they are easier to get started with. Microservices usually rise out of necessity when our system is already in production.

However, there are a lot of questions that arise when deciding when to do the migration or not – how do you decide the boundaries of a service? How do you verify the self-healing properties of your microservice architecture?