I think that makes sense. Likely something to feel out as well to see how functional that is in a spreadsheet. The idea behind a spreadsheet is that perhaps it could enable someone to do more labeling more easily with this approach.
On Wed, Mar 4, 2026 at 9:08 AM Goutam K <[email protected]> wrote: > Hi all, > > I recently commented on GitHub Issue #213 regarding the proposed Google > Sheets persister for Burr. > > I wanted to confirm the data model direction before starting > implementation. > > My current plan is to implement a GoogleSheetsPersister similar to the > existing Postgres persister using the Google Sheets Python API. > > Initial layout proposal: > > - one spreadsheet per deployment > - a "states" tab keyed by (app_id, partition_key) > - serialized JSON state + timestamps > - optionally an "events" tab if needed for parity with other persisters > > Issue reference: > https://github.com/apache/burr/issues/213 > > Does mirroring the Postgres persister schema sound like the right starting > point, or is there a preferred structure for new persisters? > > Happy to iterate based on feedback. > > > Thanks, > Goutam >
