Handle ratelimiting

[strugee]

We don't seem to do this properly. Snippet from my local logs against my homeserver:

2023-11-16T20:43:26.875-08:00 🗄️ Get state { room: 'SeaGL2023-Volunteering', type: 'org.seagl.patch.redirect', key: undefined }
2023-11-16T20:43:26.876-08:00 🏘️ Add to space { space: 'SeaGL2023-Information', child: 'SeaGL2023-InfoBooth' }
Rate limited: {"ms":4855}
2023-11-16T20:43:31.943-08:00 🏘️ Add to space { space: 'SeaGL2023-Information', child: 'SeaGL2023-BotHelp' }
Rate limited: {"ms":4768}
2023-11-16T20:43:37.024-08:00 🏘️ Add to space { space: 'SeaGL2023-Information', child: 'SeaGL2023-SpeakerHelp' }
Rate limited: {"ms":4697}
2023-11-16T20:43:42.041-08:00 🏘️ Add to space { space: 'SeaGL2023-Information', child: 'SeaGL2023-SponsorHelp' }
Rate limited: {"ms":4682}
2023-11-16T20:43:47.062-08:00 🏘️ Add to space { space: 'SeaGL2023-Information', child: 'SeaGL2023-Volunteering' }
Rate limited: {"ms":4667}
2023-11-16T20:43:51.952-08:00 🔖 Resolve tag { tag: 'seagl2023-restricted' }

[altsalt]

I think this is being handled and those messages are just notices, but I could be wrong.

[strugee]

I think we're both wrong, actually.

Last night I misread the preceding line and thought that it was continually trying things against the space. But the rooms are different in each message. So I think you're right that it is being handled. However, notice how the counter is decreasing for each attempt - that indicates to me that there's one rate limit being applied to all of these requests, but we don't back off new requests until we've received a rate limit response back. Instead once we get one rate limit response, that should cause all subsequent requests to be deferred until the ratelimit window has passed.