Last modified: 2014-09-23 22:58:35 UTC
When the creation of new Labs instances fails, it would be good to get some more output if possible. Now I just see "Failed to create instance".
Yes, we should pass-through error messages from the API. If you are having failures, please find me on irc. It's likely a quota issue.
I spoke with Andrew Bogott about this issue this week, and he said that this is technically a user interface problem, but "getting detailed error messages from openstack is pretty complicated since the failures are [asynchronous]... So not so much a design problem as a 'not worth the months it would take to do right' problem :)" I'm leaving this note here for anyone in the future who wants to work on solving this problem.
Some of the failures are asynchronous, but a lot of the error messages are returned immediately on the API call; we're just not passing them through to the end-user. Additionally, any of the errors that are asynchronous can also be pulled with appropriate API calls, we're just not exposing that. Passing through the immediate failures is likely a pretty easy fix.