Gravar-mail: Multi-objective Optimization of Long-run Average and Total Rewards