[metrics/system/process] on Darwin cpu total ms seems off

[leehinman]

On Darwin if you run filebeat and top shows ~500% CPU for filebeat we would expect cpu.total.time.ms to be roughly 5 times the uptime.ms. This isn't the case. For example I see:

uptime.ms	cpu.total.time.ms
77881	9649
96485	11956
115261	14307

If I run the same thing on Linux, the numbers are as expected.

These stats come from

elastic-agent-system-metrics/metric/system/process/process.go

Line 187 in f03b1a2

func (procStats *Stats) GetSelf() (ProcState, error) {