We are going for a revamp of our Looker project structure.
We are connecting Looker reports to our main application via SSO embedding.
We have many customers and each customer has the same table structure defined in a different database.
Each model has a list of reports which are available to be viewed. It also includes the base.lookml containing the definition for all the views which are joined as detailed in the discourse document below:
We have to replicate models for each customer to have separate connections pointing to the right database. We are expecting many new customers to come on board soon. Hence, adding a new model each time a new customer joins will not be feasible.
We understand that the connection can be parametrized but with that PDT’s will not work. This is a concern for us since, we would like to use PDT’s for caching and performance purposes.
We are stuck, any direction on how to proceed with the project structure is welcomed.