feat: Improved rate limit handling#945
Merged
rogerfar merged 3 commits intorogerfar:mainfrom Mar 14, 2026
Merged
Conversation
Contributor
Review Summary by QodoImplement per-host rate limit coordination with separate client groups
WalkthroughsDescription• Separates rate limit tracking into two HTTP client groups for independent pacing • Introduces RateLimitCoordinator to track per-host cooldowns across requests • Adds TORBOX_CLIENT_SLOW for slow operations with dedicated rate limit handling • Improves 429 response handling without requiring timeout overrides • Adds comprehensive test coverage for rate limit coordination and handler logic Diagramflowchart LR
A["HTTP Requests"] --> B["RateLimitHandler"]
B --> C["RateLimitCoordinator"]
C --> D["Per-Host Cooldowns"]
E["Fast Operations"] --> F["TORBOX_CLIENT"]
G["Slow Operations"] --> H["TORBOX_CLIENT_SLOW"]
F --> B
H --> B
B --> I["429 Response"]
I --> D
D --> J["Pause Dequeuing"]
File Changes1. server/RdtClient.Service.Test/Helpers/RateLimitCoordinatorTest.cs
|
Contributor
Code Review by Qodo
1. TorBox user call unhandled
|
![qodo-code-review[bot]](https://avatars.githubusercontent.com/in/484649?s=60&v=4){kind=small}
…and improve test consistency across methods. Updated logging and error handling for rate-limiting scenarios.
omgbeez
force-pushed
the
upstream/even-better-ratelimitng
branch
from
fdc8c55 to
67afe18
Compare
…ption Needed to move the retry handler before the TimeOut handler, and drive the throw from the TimeOut configured by the user for the provider.
rogerfar
approved these changes
Mar 14, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Separates calls into two groups, which allows pacing headers to be tracked separately greatly increasing the reliability of the rate limiting.
Other improvements to the handling of 429 responses so that it's no longer necessary to override request timeouts to mean this on slow queries.
The custom handler is not needed on the standard HTTP client resiliency pipeline now, since the rate limiting on this pipeline will only be a few seconds, and cannot exceed the timeout typically, where on the slow pipeline (used for creating new downloads on TorBox) the limitations are exceptionally long so the handler is needed to determine when a timeout would happen from the throttle, and properly report it to the RateLimitCoordinator.
Other providers should consider testing with the TorBox standard pipeline and adopting it as any provider using CloudFlare or implementing standard headers for pacing and rate limits would benefit without need for changes. If they use really low rate limits on some API groups like TorBox, then use the slow pipeline.
For future consideration, if other providers have more rate limits groups to monitor, or need notably different settings it might be better to use a factory implemented by the provider classes to generate the needed pipelines independently.