Add runbook for push image failures

[robrap]

These image push failures are currently being sent to many teams, but it is unclear what needs to be done and who needs to fix issues. See:

• https://github.com/edx/public-dockerfiles/blob/e7b6f3daa91a946dd882a262c1a30b17ec2f9c60/.github/workflows/push-xqueue-image.yaml#L55C11-L58C154

If we create and point to a common runbook in Confluence, then:

1. We can comment on how to deal with common errors (if any arise), and
2. Based on our findings of what errors are common, we can determine who mainly would be fixing these issues. If the errors are typically not unique to the service, then maybe it will make sense for Arbi-BOM or SRE to own these notifications?

The runbook can provide a place for us to gather data, and for the team notified to better understand what they should do.

Note: When I got the email about the following error, https://github.com/edx/public-dockerfiles/actions/runs/12461004340/job/34779989143, it had already resolved elsewhere, so I ignore this issue. Maybe it was failing on:

56.83 Error: retrieving gpg key timed out.

I do not know if re-running would have resolved (I didn't need to, so I never re-ran them), and I don't know if the job could have a retry to keep this failure from happening as much.