Host system metrics

[dmarcus-wire]

I need a script that captures system level information to understand transcription resource consumption. It should append data to CSV file that contains the following columns

date: The current date in YYYY-MM-DD format.
timestamp: A timestamp in HHMMSS format.
pod name: The name of the Pod (or 'No Pod' if no pod is running).
processor/gpu name: The GPU name, or CPU model if GPU information is unavailable.
core/gpu count: The number of CPU cores or GPUs.
max usage (%): The maximum GPU utilization in percentage.
max gpu temperature (C): The maximum GPU temperature in Celsius.
max pwr:usage/cap (%): The power usage as a percentage of the maximum power limit.
max vram usage (%): The percentage of VRAM used.

[dmarcus-wire]

This script host_metrics.py runs in the background and logs the above metrics every seconds. The data would need to be joined with the csv generated from the evaluation.py to correlate resource consumption.