From my understanding, the event sampling works on a ratio basis, so we can have a 1:10 chance of selecting an event. I'd like to integrate this into some dashboards since it takes so long for them to load.
So my question is, has anyone used this and noticed a big variation in the data returned? Is there anyway to find the margin of error?
I use it, for what its for, which is to speed up trial & error on my queries. If you are any way going with a sampling approach, how accurate does it need to be? It definitely makes searches much faster.