We're trying to put the people from the business to do it. We are using APIs, and we have open APIs to define our APIs and the request-response that each call requires and sends. So, to base the mapping on that, there was nothing to help. I know that with some tools, such as Oracle tools, you can see the input and expected output. With drag and drop, you can take one property from the left and drag it to the right, and it does all the mapping itself, but that's not the case with Camunda. So, for me, this is something that can be improved. If you stick to the basics, it can be pretty easy. If you need to extend, as we did with custom plugins, this is rather hard because the documentation is not that great. Everything is there. It is an open-source tool, and we had access to classes and documentation, but there was no great explanation on how to use them. There were also not many examples. It was pretty hard to go this way, but if you stick with the basic functionality, it is a great and easy tool. Another thing that we struggled with has already been fixed. They now offer a SaaS version of Camunda. We struggled a little bit with how to deploy Camunda in our infrastructure. The way it was designed was that the database can easily become a performance bottleneck. It is something that they could improve. They can provide more examples of how to implement a scalable solution using Camunda.
