Recent Heroku_ext changes break pg:copy and pg:restore for existing DBs

I don't see how they could have possibly tested this change with their existing automation before pushing it out. Don't see how the "plan" could have passed peer review without questions coming up.

Then, surely they would have seen a spike in pg:copy failures and reacted quickly to communicate the issue and their progress in resolving..

I know stuff happens but this isn't some scrappy startup. It's Heroku owned by Salesforce selling Enterprise plans..