Hi everyone! I'm Khush Chaudhari, a Full-Stack Developer and 2nd year student at Newton School Technology - ADYPU. I have extensive experience in Python and project documentation, and I’ve spent the last few days diving into Metaflow’s internal data structures. You can see my work on https://github.com/khushthecoder I'm targeting the Metadata Service Pagination project. I've read through

metaflow/plugins/metadata_providers/service.py

and traced how

ServiceMetadataProvider._get_object_internal

constructs REST paths like

/flows/{id}/runs

— I can see there's no

limit

or

offset

parameter passed, which means every listing endpoint returns unbounded responses. I've also looked at the metadata service routes in

metaflow-service

to understand the server side. I've installed Metaflow, ran tutorials 1-5 locally, and I'm working on a PR for issue #3019 (

_metaflow

parameter bug in

MetaflowObject.__init__

). @Sakari Ikonen — for the pagination project, are you leaning toward cursor-based pagination (using

created_at

as the cursor) or offset-based (`limit`/`offset`) for the metadata service endpoints? Cursor-based would handle concurrent inserts better but requires all endpoints to have a consistent sort order.