Testing DMN Performance by turning off Repository and History

Hi,
I’m trying to test the performance of Dmn inside the Camunda 7.6. I have turned off the history as referred inside dmn-benchmark and now I want to turn off the repository to see how the Dmn performs when both history and repository are turned off. The issue is I have no idea how to turn off the repository. I have gone through the documents related to the Repository services but none of them have any steps on how to turn off the Repository Services.

Can someone guide me on how to “turn off” the Repository?

Hi @steja_d,

please have a look the related topic.

Best regards,
Philipp

One reason we asked this is that the blog on DMN performance makes it sound like a trivial thing to do. You should have said that running DMN under those circumstances would be very unusual.

Hi @mppfor_manu,

in the blog post, we described how we measured and provided the code for the benchmark. So all information are there.

However, the Camunda platform is a development-friendly open source engine. If your use case is currently not covered then you can adjust the engine for your needs or provide a pull request.

Best regards,
Philipp

I asked this same question. The answer is, for the entire Camunda BPMN engine it is very difficult to do. See responses below.

We have decided to run a separate Camunda instance in a separate WildFly container in which we have turned off all history. The reason we don’t want to turn off the repository service is that you would lose the easy deployment and management functions the service provides.

If you want the fastest performance, you need to build a standalone DMN engine instance (i.e. without the Camunda BPMN engine). There’s code on Github for such a standalone engine. However, you will have to build your own DMN table deployment and respository functions.

For us, at least for now, we took the middle road and just used a separate instance.