Hi All,
I have already worked with Camunda in the context of Microservices Architecture. Now that zeebe is released I am migrating to zeebe and also .net core. My solution has many micro services and I am using docker/kubernetes. I have some questions and doubts:
What is the best architecture to use in the context of a microservices ? Is it better to use Kafka ? http worker ? other ?
I have already walked through all the examples in zeebe tutorials bu I really cannot see a real example of Microservices clear architecture. Is there any good recommendation ?
Anyway, I like camunda and zeebe and I know that zeebe will be the defacto workflow engine for microservices.
Apart from that, there are many ways you can structure it. There is no one “best way”.
It depends on your scale (both workload and your team), what expertise and existing infrastructure you can leverage.
At the end of the day, you try some things, commit to one pathway and drive it all the way, then give a conference talk/write a blog about “lessons learned developing our microservices system”; and maybe release some component of it as an open source project.
There are no shortcuts. Except maybe paying for a consultant to sit with you and analyse your situation and make some recommendations. But even then, you can’t evaluate those without sufficient context.
You’ve got to do POCs and discover the edges of each of the scenarios you’ve outlined, because the most important factor is your organizational expertise.
Thank you for the comment. I have many constraints, one of them is the use of .net core. I have already used Nodejs, Java, Spring, python, rust, kotlin, etc… I think that there is no best way as you said, we have to adopt our solution/architecture and respect the constraints. I am thinking about using Kafka (event-driven), the zeebe c# client and even the http worker to compare(I have many REST services using Asp.net Web API). I will be posting in the zeebe forum and blog
Thanks
If you have an existing REST arch, then the http worker makes sense as a first place to start. If you don’t already have Kafka, I’d try it without Kafka, initially. Adding Zeebe to Kafka increases visibility. Adding Kafka to Zeebe increases operational complexity.