Indeed Alberto.

I would personally add that the decision on which metrics to monitor will typically depend on different things.  First of all, especially for an IP switch as referenced here, it will largely depend on the kind of environment.  Managing an office network or an IP network that carries media flows for example is a completely different task, and that would be reflected also in your choice of metrics.  Secondly it also could depend on your overall policies and strategies.  E.g. I know that people sometimes work with a policy that a severity critical is only used for alarms that are guaranteed to be catastrophic for  the services (i.e. your service is guaranteed to be down), major for events that are highly likely to be service affecting at least, etc.   So your choice of metrics and the severity used might be impacted by such policies.

I note that you mention that you do not have an overview of the available metrics.  But note that if you go to the default Alarm Template, it will show you all the metrics (note that there is a button on the top right that allows you to filter the list to the actual monitored metrics, or to see all of the metrics available - so you may have to switch that if not all metrics are available).  Note that you can also go to the driver of the element, and if you right click on the version, you have an export option there to export all the metrics in CSV format, and this will give you all the metrics available for that type of element.

Also note that if you enable alarming on a specific metric, there could already be some default alarm thresholds that are suggested (that is the case if the driver does include default suggestions for that metric).

Let us know if this helps or feel free to further elaborate on what it is that you wonder about, and we can see if we can provide some more specific input.