If you're referring to the data feeds, then the way to differentiate between instances and persistence of evars is to compare the evar# and post_evar#.
In workspace, when an evar fires, if you use the "instances" metric, it will give you the number of times that the it actually fired. But if you use any other metric (like page views, occurrences, etc.) it will also count the times that the value persisted (instances + persistence for total occurrences).
In the data feeds, when an evar fires, it populates the evar#. If you want to see the persisted value of the evar you would use post_evar#. So, if you have a row where post_evar# is populated and evar# is blank, that means that it's a persisted value and not an actual firing.
If you want to replicate like the instances metric from workspace, you would count the number of times the evar# fired (not the post_evar#).
The persistence is based on the same rules that you set up for your workspace, so the allocation and expiry. If your evar is set to a visit level expiry in workspace, it will have the same expiry in the data feeds.
