perf: eliminate per-request SQLite queries and optimize proxy latency

- Add in-memory ModelConfigCache (30s refresh, explicit invalidation)
  replacing 2 SQLite queries per request (model lookup + cost override)
- Configure all 5 provider HTTP clients with proper timeouts (300s),
  connection pooling (4 idle/host, 90s idle timeout), and TCP keepalive
- Move client_usage update to tokio::spawn in non-streaming path
- Use fast chars/4 heuristic for token estimation on large inputs (>1KB)
- Generate single UUID/timestamp per SSE stream instead of per chunk
- Add shared LazyLock<Client> for image fetching in multimodal module
- Add proxy overhead timing instrumentation for both request paths
- Fix test helper to include new model_config_cache field

src/dashboard/models.rs

@@ -159,7 +159,11 @@ pub(super) async fn handle_update_model(
     .await;
 
     match result {
-        Ok(_) => Json(ApiResponse::success(serde_json::json!({ "message": "Model updated" }))),
+        Ok(_) => {
+            // Invalidate the in-memory cache so the proxy picks up the change immediately
+            state.app_state.model_config_cache.invalidate().await;
+            Json(ApiResponse::success(serde_json::json!({ "message": "Model updated" })))
+        }
         Err(e) => Json(ApiResponse::error(format!("Failed to update model: {}", e))),
     }
 }