My team is thinking about using redshift for our DW. Is it smart to use RA3 nodes? I like the cross database query option that comes with it. Does it make sense to create different databases as like raw and analytics since you can query across them now? How is the performance with cross database querying?

This depends on how you are planning on constructing your final fact tables and your star schema

1. RA3 yes, if you have such scale. 2. Cross database would be your preference. Won't advise for it as most tools like DBT and Looker aren't very friendly with it. Better to use Schema to differentiate tables.

I am looking at it as querying from one database and load into another one. And maybe a view from a different database into another.

I would probably end up using RA3 if I were using redshift. The choice of single database with schemas vs multiple database comes down to your use case scenario. I personally don't see any need to separate databases (in MPP architecture) anymore especially with workload management. Previously, separating database and hardware was the only way to isolate compute, now you can manage it fine grained through WLM. Going from schema.object to database.schema.object doesn't really add any management efficiency. Neither does putting your raw data in a separate database allow you to take advantage of RA3 cold storage any better as things will likely refresh regularly.