Hi Frank,

Thanks for your kind words.

Regarding your question about scaling a trame application it really depend how "resource heavy" is your trame application and what kind of deployment setup you have access to. That will influence the kind of setup you might be looking for. Like 1 process for all your apps, 1 process per app/user or one machine per user (then you can go the HPC route with many machines for 1 user).

But let me take a stab at your questions:

1. This depend on the type of application you've created. In general, I prefer 1 process per user using our Docker setup. But if your trame app is light weight, it might be ok to run all your apps/users within the same process like we allow you to do here.

2. I think in your context, using the iframe communicator might be easier. But both apps will need to be served within the same domain. Otherwise, you can create another plain JS client from React and just talk to the trame server, but you will have to connect to the same session which would require you to start the session from react and then connect to it from your iframe.

3. Definitely, that is what our docker image do. It setup an Apache for serving everything from the same entrypoint but handle network forwarding on the per-session basis. You could even use that docker and have the apache in it serve your react app. Or you can use compose to get all the pieces available to you in a simple setup.

Otherwise we are working on an enterprise setup using Django as a landing service/portal which use Kubernetes to run and scale any trame application that you aim to deploy via that service.

Hope that help.

Seb