Backend: Handle errors and notifications for background processing

Since the bulk operations could be applied to a large number of projects at once, we will need to do these operations async.

When applying the attributes to a large number of projects, if we can succeed for most projects and surface errors of any failed cases, that will help the user resolve the failures.

How should we handle errors that happen during these async operations?

A failure for applying attributes to one project should not stop attributes from getting applied to other projects, correct?