The service_status_cache from discovery only has active_state with all detailed metrics set to None. During collection, get_service_status() was returning cached data instead of fetching fresh systemctl show data. Now always fetch fresh data to populate memory_bytes, restart_count, and uptime_seconds properly.