David shared a diagram/proposed architecture which we collaborated and updated based on the discussion in the meeting
We need to provide guidance, maybe examples, on building appropriate containers with the code needed that can be deployed to the cloud.
This is already a question today, even with the existing server chassis approach
Nigel is looking at how to make this more flexible (with existing chassis initially)
Ljupcho took us through the typical ci/cd process & how source code/open source artifacts are managed
We agreed we are targetting a stateless environment - at least in terms of egeria. State will be managed via connectors to repositories (ie postgres, kafka etc). There is of course 'state' in kubernetes resource definitions also but container data is purely transient
Multiple teams are already using connectors for configuration, audit log, potentially cohort registry, that do not depend on local files. We should expect the same here
An education session on 'Microprofile' will be arranged for a future meeting (possibly next week). We agreed that spring vs microprofile was interesting, but it doesn't address fundamental architectural questions so is a parallel track, as is use of tools like Red Hat quarkus
