How hard would be to implement a data source for Druid in growthbook?

Probably not too hard. Does Druid support standard SQL?

https://druid.apache.org/docs/latest/querying/sql.html it does, but not natively

Are you connecting any other data tools to Druid currently? If so, do you mind sharing a blank connection screen from those tools? It helps us figure out which fields are important to have besides a simple host/port.

I made a ticket for it: https://github.com/growthbook/growthbook/issues/220

We will setup a druid to ingest kinesis/s3 data - ideally pump that into Growthbook for testing

As long as there aren't too many quirks with DruidSQL that should be fine. The queries we run are fairly complex, but they are standard enough to work on PrestoDB, BigQuery, Redshift, ClickHouse, and more with almost no changes.

where is it in the codebase?

https://github.com/growthbook/growthbook/tree/main/packages/back-end/src/integrations Probably going to be most similar to PrestoDB