fix: flaky WebSocket test, usage request queue, and test improvements

- Fix flaky WebSocket passthrough test: allow StatusNormalClosure after
  client close instead of requiring NoError (race condition fix)
- Fix ratelimit 401 test: use PlatformOpenAI instead of PlatformGemini
  for OAuth token cache invalidation scenario (more accurate)
- Add usageLoadQueue: Anthropic OAuth/setup-token accounts sharing the
  same proxy exit are serialized with 1-2s jitter to prevent upstream 429
- AccountUsageCell: add module-level usage cache (5min TTL), unmounted
  safety guard, and integrate enqueueUsageRequest for throttled fetching

backend/internal/service/ratelimit_service_401_test.go

@@ -102,6 +102,8 @@ func TestRateLimitService_HandleUpstreamError_OAuth401SetsTempUnschedulable(t *t
 	})
 }
 
+// TestRateLimitService_HandleUpstreamError_OAuth401InvalidatorError
+// OpenAI OAuth 401 缓存失效出错时仍走 temp_unschedulable
 func TestRateLimitService_HandleUpstreamError_OAuth401InvalidatorError(t *testing.T) {
 	repo := &rateLimitAccountRepoStub{}
 	invalidator := &tokenCacheInvalidatorRecorder{err: errors.New("boom")}
@@ -109,7 +111,7 @@ func TestRateLimitService_HandleUpstreamError_OAuth401InvalidatorError(t *testin
 	service.SetTokenCacheInvalidator(invalidator)
 	account := &Account{
 		ID:       101,
-		Platform: PlatformGemini,
+		Platform: PlatformOpenAI,
 		Type:     AccountTypeOAuth,
 	}