Jon, The number of cores required is generally determined by some sort of capacity analysis based on the type of workflows you are performing. The minimum would be 2 cores and you may have to go to 4 cores if the workflows were heavier with 50 users (e.g., Insights usage). But, with 50 users performing general hosted feature service queries with a reasonable "think time" between requests (e.g., 6 seconds average), then 2 cores would likely be adequate.
