fix: restored automatic quota failover - #193
Merged
Merged
Conversation
- Retried portable requests after quota rejection without manual resubmission. - Preserved bound quota probes and provider exhaustion responses. - Replaced account-specific continuity evidence at admission. - Added regression coverage for recovery and exhausted pools.
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
After a successful Codex request, quota exhaustion could leak a 429 and leave later requests failing with
bound_unresolvedorturn_state_presence_mismatch. Portable requests now record the rejection and retry an eligible account within the same HTTP request, with a fixed retry budget.Bound continuations retain their account and can probe recovered quota. Cross-account admission replaces old turn-state and response-anchor evidence, so stale account-specific state cannot reach the replacement. Exhausted pools retain the provider 429.
Validation: regression tests reproduce the failures before the fixes, including successful follow-up requests, soft limits, required continuity, all-account exhaustion, and absent replacement state. Full build, vet and race-test results recorded before publication.