I’m looking for some advice because I feel like I might be heading in the wrong direction.
We’re planning to store between 500 and 5000 tuples of a specific type, and we’ll need them later for search and authorization. Since we also need to support authorized counts (to search and then check), we decided that our best option is approach#3 — reading all tuples from FGA and using them in every query to our database. We’re trying to avoid projections if possible.
Looking at the OpenAPI specs, I noticed:
listObjects returns up to 1,000 items or runs for a maximum of 3 seconds
There’s an experimental stream endpoint that is unlimited, but it’s not available in the TypeScript SDK
Given our volume, are we making a mistake by trying to implement approach #3? Are there better strategies we should consider, or are there plans for pagination support in the future?
Sorry I tried adding links but for some reasons I can’t
Thank you for the quick reply @omer1, if you are on the paid plan then I would encourage you to open a ticket for the support, they will be able to address your questions answear the fastest.
There’s an experimental stream endpoint that is unlimited, but it’s not available in the TypeScript SDK
You already noted that it’s not in TypeScript (it’s only available in Python at the moment). While it does stream, it does not provide unlimited output. It will timeout based on time. It will return results but halt after 3 seconds.
As for pagination, there are some technical challenges with our database implementation and we do not have pagination support planned at the moment.
