About generate clusters full log file

[henls]

Metrics such as gpu_util,mem_util are needed to predict task size in Lucid and QSSF algorithms, but these metrics are not available in helios publicly available dataset. May I know the way to get this data, or am I understanding it correctly?