fix(datasources): MySQL browser race + slow row counts

When switching source engine, the sample effect fired with the previous
engine selected object (e.g. public.sales_orders against MySQL), surfacing
"Table hr.sales_orders doesn't exist". Track which engine the selected
object belongs to and only sample when it matches the active engine, plus
guard against stale responses overwriting newer ones.

Row-count for the browser used an exact count(*) which full-scanned huge
tables (~30s on 24M rows). Use planner/statistics estimates and only run a
time-bounded exact count for small (<=50k) tables.

api/sql_console.py

@@ -693,8 +693,20 @@ def _table_row_count(engine: str, object_name: str) -> int | None:
             )
             try:
                 cur = conn.cursor()
-                cur.execute(f'SELECT count(*) FROM "{schema}"."{table}"')
-                return int(cur.fetchone()[0])
+                # Planner estimate first; only do a (bounded) exact count for small
+                # tables so huge tables don't trigger a slow full scan.
+                cur.execute("SELECT reltuples::bigint FROM pg_class WHERE oid = %s::regclass", (f'"{schema}"."{table}"',))
+                row = cur.fetchone()
+                est = int(row[0]) if row and row[0] is not None and row[0] >= 0 else 0
+                if est <= 50000:
+                    try:
+                        cur.execute("SET LOCAL statement_timeout = 4000")
+                        cur.execute(f'SELECT count(*) FROM "{schema}"."{table}"')
+                        return int(cur.fetchone()[0])
+                    except Exception:
+                        conn.rollback()
+                        return est
+                return est
             finally:
                 conn.close()
         if engine == "mysql":
@@ -705,8 +717,22 @@ def _table_row_count(engine: str, object_name: str) -> int | None:
             )
             try:
                 cur = conn.cursor()
-                cur.execute(f"SELECT count(*) FROM `{table}`")
-                return int(cur.fetchone()[0])
+                # Fast estimate from statistics; only do a (bounded) exact count for
+                # small tables to avoid full scans on huge ones (e.g. 24M rows).
+                cur.execute(
+                    "SELECT table_rows FROM information_schema.tables WHERE table_schema=%s AND table_name=%s",
+                    (schema, table),
+                )
+                row = cur.fetchone()
+                est = int(row[0]) if row and row[0] is not None else 0
+                if est <= 50000:
+                    try:
+                        cur.execute("SET SESSION MAX_EXECUTION_TIME=4000")
+                        cur.execute(f"SELECT count(*) FROM `{table}`")
+                        return int(cur.fetchone()[0])
+                    except Exception:
+                        return est
+                return est
             finally:
                 conn.close()
         if engine == "mongodb":