While evaluating Discourse, we do it in a classic way by tweaking configs in Web UI. Multiple contributors are involved including @olblak @halkeye @timja and me. It is perfectly fine for evaluation, but it is also a recipe for disaster in a production system. It would be awesome to move all the configuration to a repository inside jenkins-infra and setup proper configuration management. And maybe even a staging environment for testing…
We don’t have disk access, it’s a sponsored hosted solution. Same with a staging environment. We would have to ask for another instance especially since Discourse isn’t super k8s friendly as part of the build process manages the db. I’ve setup helm charts and tried before. It works, its just take a lot of effort.
I’m of the opinion to not spend limited resources solving problems that are not yet happening incase they might someday happen.
IMHO it is orthogonal. I could imagine the Discourae team interested in offering CasC for the managed service.
Anyway, it is not a top priority. If we screw up in the config, we have old and faithful autobackup