Fix for stuck ratelimit buckets

[mikeshardmind]

Summary

This might not be the best way to handle this, and this still needs testing,

A more intrusive fix here would be to swap this to a pair of semaphore and timed lockout, which I'm trying to avoid having to redesign the ratelimit mechanism.

see thread: https://discord.com/channels/336642139381301249/1355766330996949034

Checklist

• If code changes were made then they have been tested.
  • I have updated the documentation to reflect the changes.
• This PR fixes an issue.
• This PR adds something new (e.g. new method or parameters).
• This PR is a breaking change (e.g. methods or parameters removed/renamed)
• This PR is not a code change (e.g. documentation, README, ...)

[ju]

For record keeping: the fix solved my particular edge case issue.

[mikeshardmind]

Tested, working as intended without causing an issue of waking too soon and not respecting ratelimits, including on longer ratelimits like channel edit.

root issue this fixes is that without this, it's possible for a request to get stuck with nothing to trigger waking it.