Hello,
For resources efficient consumption and simplified operations it would be desirable to have an ExApp runtime environment that can be used for more than one Nextcloud instances. Mainly thinking here about AI ExApps where the CPU, GPU and memory needs can be quite high.
So I did several attempts to make it work, and of course there are many aspects to discuss / take into account but the question is whether from a software point of view if this is supported? This is what happens also for example Talk HPB and Recording Server.
So I'd like to have instance1.cloud.example and instance2.cloud.example using the same, say, translate2 docker container. With some port tweaking and assigning each HaRP a different bind address, I'm able to test the HaRP connection successfully in both instances, but inference fails on the second instance.
I can provide more details if needed. Thanks!