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
We have had to do a couple of reverts (#2560 and #2551) where an image that was successfully built with repo2docker-action does not actually start on a JupyterHub. This should be automatically tested so that it is caught earlier, and handled by the folks changing the images and not 2i2c engineers.
I'd argue we should zoom out on this a little and have a larger effort focused on what should image management/workflow look like.
From a community perspective there are probably lots of clear use cases we can build. So the build/deploy/test cycle can be completely managed by communities and they get sufficient feedback to address issues.
From a personal perspective I think that a product focused approach to image management should remove the need for SRE /support requests around image as a larger product goal.
So the build/deploy/test cycle can be completely managed by communities and they get sufficient feedback to address issues.
+1000
From a personal perspective I think that a product focused approach to image management should remove the need for SRE /support requests around image as a larger product goal.
From a personal perspective I think that a product focused approach to image management should remove the need for SRE /support requests around image as a larger product goal.
We have had to do a couple of reverts (#2560 and #2551) where an image that was successfully built with repo2docker-action does not actually start on a JupyterHub. This should be automatically tested so that it is caught earlier, and handled by the folks changing the images and not 2i2c engineers.
jupyterhub/repo2docker-action#103 is upstream issue I just filed related to this.
The text was updated successfully, but these errors were encountered: