Datagroups and caching with dynamic schema value

[ryanmjennings]

Hi everyone! I have a somewhat open ended question that I can’t seem to find any information on… We have a Looker instance at my organization that connects to a Redshift DW with multiple schemas. The schema that is the connected to is determined by a user attribute (this article describes how this is implemented: https://cloud.google.com/looker/docs/best-practices/how-to-create-a-dynamic-schema-with-user-attribu...). My question is how do I build datagroups and handle caching in this type of architecture? Since groups of users have difference schemas, and each schema is updated on its own cadence, there’s not really a “central” table for a sql_trigger to check for updates. In addition to this, even if I were to somehow trigger a cache refresh, which schema’s cache would be refreshed? If the answer is “all of them” then when one team update’s their schema, would it cause the instance-wide cache to be reloaded? Has anyone ever implemented caching with the above design? If so, how did you do it? Thanks!