what's the use case? the snapshot parameters are txids, so you'd choose a snapshot txid after the created txid of a relation you're interested in for $1 and the max txid (indicating that the relation is still present) for $3

The snapshot params are `pg_snapshot`s, not xid, I don't know how to get a realistic pg_snapshot! The use-case is simply to be able to investigate a high CPU usage on my database coming from this query. There's indexes in place but I'd like to think about whether they can be improved somehow

oh interesting 🤔 yeah i haven't dug that far in. i'd be curious how the query caller can even be aware of a snapshot unless the value itself is also computed?

the "simplest" approach is to just call the HeadRevision in SpiceDB and grab the revision object to get its snapshot parameter