Updates: I’ve been exploring the ServiceMetadataProvider implementation, particularly "_get_object_internal" in "service.py". It looks like metadata queries currently fetch full datasets and apply filtering client-side via "_apply_filter", and traversal (run → step → task) requires multiple sequential API calls without pagination or batching. I was wondering if there are plans to support server-side filtering or pagination at the metadata service level to reduce the number of requests for large workflows?