fix(dashboard): add COALESCE to SQL aggregations and empty-state handling for charts

Backend: wrap SUM() queries with COALESCE in handle_time_series,
handle_clients_usage, and handle_detailed_usage to prevent NULL-induced
panics when no data exists for a time window.

Frontend: add showEmptyChart() empty-state messages and error feedback
across overview, analytics, costs, and clients pages. Rewrite analytics
loadCharts() to use Promise.allSettled() so each chart renders
independently on partial API failures.

src/dashboard/usage.rs

@@ -104,8 +104,8 @@ pub(super) async fn handle_time_series(State(state): State<DashboardState>) -> J
         SELECT 
             strftime('%H:00', timestamp) as hour,
             COUNT(*) as requests,
-            SUM(total_tokens) as tokens,
-            SUM(cost) as cost
+            COALESCE(SUM(total_tokens), 0) as tokens,
+            COALESCE(SUM(cost), 0.0) as cost
         FROM llm_requests
         WHERE timestamp >= ?
         GROUP BY hour
@@ -155,8 +155,8 @@ pub(super) async fn handle_clients_usage(State(state): State<DashboardState>) ->
         SELECT 
             client_id,
             COUNT(*) as requests,
-            SUM(total_tokens) as tokens,
-            SUM(cost) as cost,
+            COALESCE(SUM(total_tokens), 0) as tokens,
+            COALESCE(SUM(cost), 0.0) as cost,
             MAX(timestamp) as last_request
         FROM llm_requests
         GROUP BY client_id
@@ -255,8 +255,8 @@ pub(super) async fn handle_detailed_usage(State(state): State<DashboardState>) -
             provider,
             model,
             COUNT(*) as requests,
-            SUM(total_tokens) as tokens,
-            SUM(cost) as cost
+            COALESCE(SUM(total_tokens), 0) as tokens,
+            COALESCE(SUM(cost), 0.0) as cost
         FROM llm_requests
         GROUP BY date, client_id, provider, model
         ORDER BY date DESC