I wanted to make sure that I was going to do it right. I am developing a multitask application. I plan to create one database for each tenant on one raven instance, but would also like to have one instance of my code base (i.e. One deployed mvc webapi instance).
So, I would call some service locator to get a singleton instance of the document store, and then pass it the tenant ID so that I can work against the corresponding session (one session per db tenant).
Provided that each db tenant is relatively small (in hundreds of thousands of documents), the volume of transactions is quite low (hundreds of users per tenant). Is it realistic to expect to run more than multiple tenants on the same server?
I know this is an open question about oranges to apples, but the answer I'm looking for is either a) yes, this is a standard approach to tiered use of a raven and is limited by hardware, or b) you are mistaken, and this may end after several tenants regardless of memory / processing power.
early
source share