I am looking for a solution to deploy Flowable ecosystem in a way, that each module is deployed as a separate microservice. I found in documentation, that it was possible before to build and start Task, Modeler, Admin and IDM separately, but now it is not so straightforward. The only option I find is all-in-one image.
Is it intentional not to go with this option?
Are there any ways to still build individual images and run them? (and connect them to cooperate)
Such separate images still exist in DockerHub Docker Hub, but I would like to be able to build them from source code.
I would also be interested in the answer to your question.
But meanwhile, I wanted to ask you about the strategy you intend to follow concerning the database. Are you accepting the fact that you’ll need to share the database for e.g. tasks and processes or do you already have a strategy on how to separate them?
To be honest I didn’t think about it, but in fact that doesn’t sound like a problem to me. If I understand Flowable architecture correctly, tasks and processes are within same logical domain, thus I don’t see a reason to split it. But maybe you have a usecase where it would make sense.
The reason why I would like to split it into microservices (for lack of better word) is to achieve good scalability - I can imagine, that Task would receive huge traffic (new processes/tasks, interactions) and need to scale, while Modeler (modeling new processes) probably would be less busy and might have less active instances.
I’d suggest you create your own microservices with Spring Boot by embedding the engines you need or use the flowable-rest container image. The apps were joined in a single flowable-ui package to simplify the distribution and use of Flowable, specially for people new to the technology.