Permit

Howdy <@U04KNLJQ9Q9>! :slightly_smiling_face:
Yep you’re pretty much spot on :dart:  What you described should work very well.

You can also (have them) expose an API service that queries the data source instead of fully replicating it  (remember that fetching and loading happens in async in the background and not part of the critical path of an OPA query.

You can also have the team produce a <https://docs.opal.ac/tutorials/write_your_own_fetch_provider|custom data-fetch provider> - so they can fully own interaction with the data source they manage.

These and others are all valid paths

Hope this helps- feel free to follow up with additional questions

It does. I think a data fetch provider would be too big an ask for this team? But I’ll review the docs and make a review.

Do you have a recommendation on what database we should use for the data? We can go in any direction with it.

The docs point to using Postgres a lot. Our data is JSON, is it smart to put it into a SQL database and have OPAL convert it back to JSON? Or… :shrug:

Depending on the data source a fetch provider can be as short as 8 lines of Python.

1. You can store JSON in Postgres (We [<http://Permit.io|Permit.io>] do )
2. But you can choose whichever datastore you prefer - Mongo would do as well  

A small thing to add that Postgres supports JSON type if you want to give it a shot :slightly_smiling_face:
<https://www.postgresql.org/docs/current/datatype-json.html>

Thanks <@U02E697H22J> :slightly_smiling_face:
But <@U04KNLJQ9Q9> - you do you. Postgres, Redis, Mogo, Snowflake, … - all good

I appreciate the information! My DB skills are poor, so I’m a little nervous. But I think I can take this and run with it for a while.