Fix timeseries query: use epoch-based bucketing for DateTime64 compatibility

Replace toStartOfInterval with intDiv on epoch seconds, and cast
avg/quantile results to Int64 to avoid Float64 JDBC mapping issues.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

cameleer3-server-app/src/main/java/com/cameleer3/server/app/search/ClickHouseSearchEngine.java

@@ -106,12 +106,13 @@ public class ClickHouseSearchEngine implements SearchEngine {
         long intervalSeconds = Duration.between(from, to).getSeconds() / bucketCount;
         if (intervalSeconds < 1) intervalSeconds = 1;
 
+        // Use epoch-based bucketing for DateTime64 compatibility
         String sql = "SELECT " +
-                "toStartOfInterval(start_time, INTERVAL " + intervalSeconds + " SECOND) AS bucket, " +
+                "toDateTime(intDiv(toUInt32(toDateTime(start_time)), " + intervalSeconds + ") * " + intervalSeconds + ") AS bucket, " +
                 "count() AS total_count, " +
                 "countIf(status = 'FAILED') AS failed_count, " +
-                "avg(duration_ms) AS avg_duration_ms, " +
-                "quantile(0.99)(duration_ms) AS p99_duration_ms, " +
+                "toInt64(avg(duration_ms)) AS avg_duration_ms, " +
+                "toInt64(quantile(0.99)(duration_ms)) AS p99_duration_ms, " +
                 "countIf(status = 'RUNNING') AS active_count " +
                 "FROM route_executions " +
                 "WHERE start_time >= ? AND start_time <= ? " +