Failures come back straight away

When part of a page asked the portal for something it could not serve — a file from a component that was restarting, say — the "no" was sent back by a slower internal route than the original question travelled on. Usually that was fine. Occasionally the answer was simply lost, and whatever was waiting for it waited a full minute, or forever.

Failure answers now come back the same way the request went out. Something that cannot be served says so immediately, so a slow corner of a page stays a slow corner instead of holding the rest up.

Reconnecting…
The connection to the server was interrupted. Trying to restore it…
Trying again…
The connection could not be restored. Reloading the page…
The server was updated. Reloading the page to pick up the latest version.