feat(quota): add opt-in Codex auto-ping

Generalize Claude 5h auto-ping into a provider-generic scheduler and add
opt-in Codex auto-ping that warms the next 5h window via a tiny gpt-5.5
request when session.resetAt slides. Default off, per-connection toggle,
failure cooldown, blocking-quota skip, drains stream before success.

Closes #2107

Co-authored-by: Cursor <cursoragent@cursor.com>

src/shared/constants/config.js

@@ -62,16 +62,34 @@ export const CONSOLE_LOG_CONFIG = {
 // Client-side store TTL: how long fetched data stays fresh before re-fetching
 export const CLIENT_STORE_TTL_MS = 60000;
 
-// Claude auto-ping: keep 5h window warm by sending a tiny request right after reset
-export const CLAUDE_AUTOPING_CONFIG = {
-  settingsKey: "claudeAutoPing",        // settings table field
+// Quota auto-ping: keep 5h windows warm by sending a tiny request right after reset.
+export const QUOTA_AUTOPING_CONFIG = {
   tickIntervalMs: 60000,                // scheduler tick
   pingLeadMs: 5000,                     // fire once reset passes (within tolerance)
-  pingModel: "claude-haiku-4-5-20251001", // cheapest model
-  pingText: "hi",
-  pingMaxTokens: 1,
   refreshAheadMs: 300000,               // refetch usage when within 5min of reset
-  fiveHourKey: "session (5h)",          // quota key returned by usage handler
+  failureCooldownMs: 900000,            // avoid failed ping spam while upstream/auth is unhealthy
+  providers: {
+    claude: {
+      settingsKey: "claudeAutoPing",    // preserve existing settings contract
+      quotaKey: "session (5h)",         // quota key returned by usage handler
+      pingModel: "claude-haiku-4-5-20251001",
+      pingText: "hi",
+      pingMaxTokens: 1,
+    },
+    codex: {
+      settingsKey: "codexAutoPing",
+      quotaKey: "session",
+      pingWhenResetAtSlides: true,
+      resetAtDriftMs: 30000,
+      minPingIntervalMs: 600000,
+      skipWhenBlockingQuotaExhausted: true,
+      // Free and Plus Codex accounts both expose gpt-5.5; avoid fallback probes that waste requests.
+      pingModel: "gpt-5.5",
+      pingText: "hi",
+      pingInstructions: "Reply with OK.",
+      pingReasoningEffort: "none",
+    },
+  },
 };
 
 // Re-export from providers.js for backward compatibility