Dealing With a Microservice-based Architecture… What?!
Microservices are great until you have to deploy them, then they can be the worst
Welcome to the first installment of “Ask a Dev”, where I’m going to try and directly answer questions from people in our industry about what we do and how we do it.
Today’s question is packing a punch, because it deals with microservices and what are the logistics around deploying a whole application based on them.
To be more specific, Ricard Nikens on Twitter asked:
“Something I still haven’t quite figured out is how to build an entire product based on microservices. Like, if I have 3 microservices for the product I want to build, I have 4 GitHub repos, right?
Do I install these microservices as dependencies from GitHub? Do I have to publish them as private npm packages? How is this dealt with?
And if I have an update on one of these microservices, how do I update it in the whole product?
And then, also, how do I get it all up and running? Do I need a bash script to start several node processes on different ports and have the main application know where these microservices are at?”