[v3.8.50] fix(antigravity): lock full quota per exact model - #8630
Merged
diegosouzapw merged 4 commits intoAug 6, 2026
Merged
diegosouzapw merged 4 commits into
diegosouzapw merged 4 commits into
Conversation
The Antigravity 429 path now uses lockExactModel instead of the previous family-scoped Gemini lockout, so a single-model exhaustion no longer disables sibling models on the same account. Update the existing 'family cooldown starts around 30s' test to assert the new contract: - cooldownMs is bounded short (<= 60s) instead of the inferred 30s+ family lock - the exact failing model is locked (allRateLimited / cooldownScope=model) - sibling models on the same connection stay eligible - clearModelLock is the existing success-path hook that releases the lock Refs diegosouzapw#8630
Owner
|
Re-homed to |
Resolves conflicts against release's own refactor of the Antigravity retry loop (executeOnce -> runAntigravityAttempt/tryResolveRetryFromErrorBody) and against diegosouzapw#8050's already-merged 404/not_found exact-model lockout, by generalizing accountFallback.ts's key-lookup helpers to check all 3 lockout key shapes (quota-family, diegosouzapw#8050's literal-model 404 key, and this PR's own exact-model key) instead of picking one side. Co-authored-by: diegosouzapw <8016841+diegosouzapw@users.noreply.github.com>
Extracts the exact-model (non-family-scoped) lockout key/entry logic introduced for Antigravity into a new leaf module, open-sse/services/accountFallback/exactModelLock.ts, following the same DI pattern already used by lockoutEviction.ts and cooldownCap.ts. This brings accountFallback.ts, antigravity.ts and chatCore.ts back under their frozen file-size baselines without changing behavior. Co-authored-by: costaeder <costaeder@users.noreply.github.com>
This was referenced Aug 6, 2026
muhamadgalihsaputra
pushed a commit
to niyatna/NiyatnaRoute
that referenced
this pull request
Sep 27, 2026
…zapw#8630) Validated in local merge-train T6 (ungrouped batch 1)
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
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.
A full-quota Antigravity 429 currently persists
rate_limited_untilfor the whole connection. That disables unrelated models on the same account.\n\nThis introduces an exact provider/account/model lock that coexists with existing quota-family locks, routes Antigravity quota paths through it, and keeps the connection eligible for sibling models. A successful request for the same model clears the exact lock through the existing success path.\n\nThe lock is intentionally in-memory; the connection-wide persisted cooldown is not reused because it would reintroduce the cross-model outage.\n\nTests: Antigravity quota-family suite (7 passing).