Ratekeeper should not limit based on remote storage servers

After a remote region has failed and the comes back online, the remote storage servers will attempt to catch up as fast as possible. This causes Ratekeeper to limit traffic to the primary region.

No file, test, or entry point is named. Start by locating Ratekeeper and the handling of remote storage servers after a remote region returns, then trace how their catch-up traffic affects primary-region limits. Done means remote recovery traffic no longer causes Ratekeeper to limit traffic to the primary region, with regression coverage for that scenario.

Written by the indexing model from the issue text.