Storing CUBE URLs in mongoDB

[Sandip117]

Advantages of saving CUBE URLs in mongoDB:

1. We already store pfdcm endpoints in the DB. Doing the same thing for CUBE would be consistent with pflink's design
2. We can continue to keep the application stateless
3. Less effort to change and re-configure
4. Easily store and remove new API endpoints
5. Clients can decide which CUBE to use at runtime

Advantages of saving as env variables:

1. Technically, only "data" should be stored in a DB
2. Unless a client needs to know application setting, it's better to store config as env variables
3. Easy to read from a file than a DB
4. Clients need not worry about the existence of CUBE

Some key questions:

1. Should "clients" be adding new CUBE end points in pflink ?
2. Do clients really choose from available pfdcm & CUBE options ?
3. Should a single instance of pflink communicate with multiple pfdcm instances in the first place ?
4. Is pflink only a client of pfdcm & CUBE ?
5. Is pflink a separate service/application that "uses" CUBE & pfdcm ?