You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
This is something we've been thinking about for a while, and it seems like getting a basic PoC for this should be easier than it sounds.
Roughly this would involve:
Setup a single Kubernetes cluster with auto-scaling rules, which will run the osm-seed instances
Behind some basic authentication, setup a very simple Web UI with a form to fill in values to customize for the osm-seed instance.
In the backend - this can be a simple standalone python app or so, we just run a helm install of osm-seed using the Values sent in by the user via the web form and create a new release / osm-seed instance on the cluster.
Have some basic UI for users to see their active releases, and delete them.
Once we have automated the subdomain creation and SSL generation process in #216, this setup of Web UI -> Values.yaml file -> Helm Install as an initial PoC should not be too crazy.
There will be some practical considerations like #25 and one would need to do some estimation of infrastructure costs, ideally as part of the basic Web UI.
It seems like throwing up a very basic PoC to prove that it "works" should be rather straightforward, and then we could take it from there to see what might be useful to build out, and perhaps get a better sense of what might not be so straightforward / what pain points might be.
This is something we've been thinking about for a while, and it seems like getting a basic PoC for this should be easier than it sounds.
Roughly this would involve:
helm install
of osm-seed using the Values sent in by the user via the web form and create a new release / osm-seed instance on the cluster.Once we have automated the subdomain creation and SSL generation process in #216, this setup of Web UI -> Values.yaml file -> Helm Install as an initial PoC should not be too crazy.
There will be some practical considerations like #25 and one would need to do some estimation of infrastructure costs, ideally as part of the basic Web UI.
It seems like throwing up a very basic PoC to prove that it "works" should be rather straightforward, and then we could take it from there to see what might be useful to build out, and perhaps get a better sense of what might not be so straightforward / what pain points might be.
cc @geohacker @Rub21
The text was updated successfully, but these errors were encountered: