feat(antigravity): quota-aware routing with reset-aware fallback

On a 409/429 from Antigravity, fetch live quota to learn the exact
per-model resetAt instead of guessing a backoff, then skip only the
exhausted account/model pair until that time.

- antigravityQuota.js: in-memory quota cache, coalesced concurrent
  refreshes, 30s throttle per connection (applied to failures too),
  keeps known cache when upstream returns 401/403 error payloads
- auth.js: pre-filter exhausted account/model pairs; report the
  earliest quota reset when every account is blocked; skip the
  30-minute cooldown cap so the upstream resetAt is not truncated
- chat.js: antigravity 409/429 falls back on the RAM cache only, no
  persistent modelLock_* for this path
- Logs identify accounts by id prefix, never email or name

Closes #3561

src/sse/handlers/chat.js

@@ -7,6 +7,7 @@ import {
   extractApiKey,
   isValidApiKey,
 } from "../services/auth.js";
+import { handleAntigravityQuotaError } from "../services/antigravityQuota.js";
 import { getSettings } from "@/lib/localDb";
 import { getModelInfo, getComboModels } from "../services/model.js";
 import { handleChatCore } from "open-sse/handlers/chatCore.js";
@@ -300,8 +301,22 @@ async function handleSingleModelChat(body, modelStr, clientRawRequest = null, re
 
     if (result.success) return result.response;
 
-    // Mark account unavailable (auto-calculates cooldown with exponential backoff, or precise resetsAtMs)
-    const { shouldFallback } = await markAccountUnavailable(credentials.connectionId, result.status, result.error, provider, model, result.resetsAtMs);
+    // Antigravity 409/429: refresh live quota to get exact resetAt before locking
+    let quotaResetMs = null;
+    let resetsAtMs = result.resetsAtMs;
+    if (provider === "antigravity" && (result.status === 409 || result.status === 429)) {
+      quotaResetMs = await handleAntigravityQuotaError(
+        credentials.connectionId, result.status, model,
+        refreshedCredentials.accessToken, credentials.providerSpecificData
+      );
+      if (quotaResetMs) resetsAtMs = quotaResetMs;
+    }
+
+    // Exhausted Antigravity model is blocked only in RAM cache until upstream resetAt.
+    // Do not persist a modelLock_* for this path.
+    const shouldFallback = provider === "antigravity" && quotaResetMs
+      ? true
+      : (await markAccountUnavailable(credentials.connectionId, result.status, result.error, provider, model, resetsAtMs)).shouldFallback;
 
     if (shouldFallback) {
       log.warn("FALLBACK", `⇄ ACC:${credentials.connectionName} UNAVAILABLE (${result.status}) → NEXT ACCOUNT`);