I think most people are running into the issue, that they think "microservices" are something new and hip and they don't have a clue on how to break down problems a meaningful way.
PS: Another VERY interesting read, and also maybe a bit unrelated to container hosting is "Testing Strategies in a Microservice Architecture": http://martinfowler.com/articles/microservice-testing/. I'd love to hear some more thoughts about that and how services like giantswarm could help with that...
I don't think this is OT at all. I know those articles and if even Martin Fowler says starting with a monolithic app is not too bad for most problems one should think twice before splitting up everything in tiny pieces.
At the end one needs to re-learn a lot. And due to this I think it is especially challenging doing microservices with technologies that power the average web app (Rails, Django, etc). Maybe it will help to choose a technology stack that emphasizes the values of microservices more (thinking of Clojure or Erlang).
At the end one needs to re-learn a lot. And due to this I think it is especially challenging doing microservices with technologies that power the average web app (Rails, Django, etc). Maybe it will help to choose a technology stack that emphasizes the values of microservices more (thinking of Clojure or Erlang).