Currently, Analytics Builder simulation mode is only capable of consuming data which is stored in MongoDB. This means that long retention periods must be set for data that is expected to be used for simulation or regression testing of Analytics Builder models. Setting long retention periods can be costly (high data storage charges) and has performacne implications in the MongoDB cluster.
We would like to see an option for Analytics Builder to collect historical data through the DataHub API when running in simulation mode. This would allow us to use data that is months or years old in our regression testing, at reasonable cost.
A good suggestion.