Lead Dev 2016 London: Journey from Monolith to Microservices

see http://www.appcontinuum.io/

Which comes first ? The monolith of the microservices ? -> get started with domain bounded context (SOA was without domain bounded context).

1. write (API level) tests

  • don’t break client behaviour
  • high level tests that reflect reality
  • use Pact (https://github.com/pact-foundation/pact-ruby) to write contract in json and generate clients ( Enables consumer driven contract testing, providing a mock service and DSL for the consumer project, and interaction playback and verification for the service provider project)

2. arrange appliction so you can SEE your domain

(Uncle Bob, drill down)

  • less costly to evolve and experiment boundaries
  • delay architecture decisions, always

3. Break out components

  • give me more space in code
  • solidify dependancies ans boundaries
  • DB: migration just touch ONE table, in order to be able to separate data later on

Cost must be > to benefits before moving on

4. promote your first microservice

  • scale a bounded context
  • reuse
  • many other reasons

But it comes with additional costs:

  • service discovery: new application with 0 business value. Increases flexibility
  • circuit breaker: new application with 0 business value, too. The “maybe monad” in a distributed system. Increases resilience and the visibility of systems.

5. Next steps

  • breakup more microservices
  • change communication interface (with a queue system)
  • don’t do anything more: it is up to you

Réactions