The image below shows the database architecture:
Runtime: The Nubitalk gathering information database, where raw data is stored in real time. This database is not accessible to clients.
Replicated database: The raw data can be replicated on azure for clients exportation. Clients must store the exported data on a place of their convenience, collecting the real time data for as long as they need.
Client database: The place where clients store the exported data (raw or processed). This is where database requests must be made (not to the Azure servers).
The operations are limited to 20 DTUs and data volume is limited to 250GB. Keep in mind that queries have to comply with this restrictions, which means that complex queries may impact on the performance.
Since this is an alternative solution to the Nubitalk reports, where the most common and useful data is already made available, you must contact us to have it implemented.