Expose host/vm powerered state metrics - #42

I wondered this or add a label to vm/host_info metrics - my idea behind these were to host all non-metric information about particular vSphere object... but may be cumbersome to use in dashboards if only a label...

Thanks for the feedback.

I considered adding this as labels on vm_info/host_info, but I chose explicit 0/1 metrics because they are easier to use in PromQL, dashboards and alerts.

For example:

• vmware_vm_powered_on == 0
• count by (host) (vmware_vm_powered_on == 0)
• vmware_host_connected == 0

With only labels on *_info metrics, users would need to filter on label values, which is useful for metadata but less convenient for alerting and numeric aggregation.

That said, I agree that power_state/connectivity are also descriptive information. So i let this PR on draft some time (and will rebase it) to see if there are any more feedback. I can still:

1. keep the explicit metrics only,
2. add the state as labels on vm_info/host_info as well,
3. or switch to the *_info label approach only.

Great PR @Zophar78. Thank you!
I favor using explicit metrics to enable simpler and more cost-effective filtering for both graphing and alerting. While adding this data as a label might not be a major problem in this specific case, I believe using a new metric is a better approach. As stated in the official documentation: 'The change of any label's value, including adding or removing labels, will create a new time series.' (https://prometheus.io/docs/concepts/data_model/)

Ok PR is now ready for review/merge.