Wow - that are a lot of questions Probably it would be easier to get in touch with us directly and hop on a call to discuss the use case & scenario?
I try to answer here as well - but some questions might need some more discussion:
- How much Zeebe is mature ? Are there some production use cases?
There are production use cases, yes. But we are still in a 0.x state - so not the same guarantees as with Camunda BPM yet.
- Which are performance stats?
This depends on so many factors that I can’t give a general answer. It would start with the question what you mean by performance (throughput? latency?).
- Are some official and/or production ready diagnostic tools available?
What do you exactly search for, what do you want to diagnose?
- Could you provide a comparison or explain differences between Zeebe and Netflix Conductor?
We don’t have a prepared piece on this and it would qualify for an forum or blog post on its own. Maybe this already helps: https://zeebe.io/blog/2019/03/zeebe-comparison-evaluation-criteria/. My spontaneous main points are the use of BPMN in Zeebe (ISO-standard, mature language, supports all important workflow patterns, allows graphical modeling, has a visual that different target groups can understand) and that Zeebe is driven by an open source vendor vs. Conductor is just a project from Netflix put OSS with no guarantees.
- Can Zeebe Enterprise be deployed within Google Cloud?
Yes. You can also get this as managed service with https://camunda.com/products/cloud/
- Giving that Camunda and Zeebe are different products, what can be best solution to use Camunda + Kafka instead of Zeebe (in case Zeebe is not fully ready)?
Sorry - I don’t understand the question fully. Of course you can use Kafka + Camunda, we have multiple customers doing so.
- Does Zeebe Enterprise prodution ready to use Kafka? (because kafka is mandatory for our project)
Sorry, don’t get the exact question.
- What are the main critical issues?
Of what exactly? Zeebe? I can’t give a list - but I would say: We can still learn things that require some refactoring, so we might need to break API’s (that’s why 0.x). Of course we try to avoid that.
And we learn a lot from early adapter on high load, which helps us stabilizing the product. But distributed systems are complex and I am sure there are still bugs in Zeebe (show me one software without ;-)). But I am confident as we have a world-class team that is really good at resolving critical issues quickly. So what I am saying is simply: It is of course still early stage
Happy to discuss the concrete use case and how we can tackle that best!