Experiment id uniqueness issue with using timestamp as the id #77
Description
The current implementation of using experiment start timestamp as its identifier can cause problems as the uniqueness of the ids cannot be guaranteed. For instance, when two experiments start "simultaneously", both of them are assigned the same timestamp as an identifier. Consequently, all telemetry associated with this identifier gets mixed up, which can lead to confusion and incorrect data analysis.