analyze_worker: workers to gather and report dynamic metadata to debusine

#8 (closed) mentions that workers have analyze_worker method that generates dynamic metadata of the worker.

I thought that it might be a good idea to start a discussion on which dynamic metadata is useful and how to gather the metadata. I was thinking of:

• RAM memory in the system (based on /proc/meminfo? via some Python library or implemented)
• Storage available (which mount point?)
• Storage type / speed? (based on hdparm quick benchmark? bonnie++? If it's a cloud based worker it's not as easy as magnetic or SSD)
• CPU power (number of processors as per /proc/cpuinfo? What would be good indicator for the type of tasks? Or a quick CPU benchmark since this can live in a cloud and the numbers might not be representative of the real capacity?)

Has anyone think of other dynamic metadata gathered by the worker itself?

Were you thinking of something very different @hertzog and not the "power" of the worker?

I'm assuming that architectures and distributions are set statically via debusine_worker.conf configuration file (mentioned in #4 (closed))