my Nostr indexer runs queries to look up profiles, contact lists and relay lists it doesn't have. normally, a contact list also contains a relay list, but this is non-standard, so that's 3 objects in total.

to speed things up and avoid hitting relay rate limits, i collect pubkeys to query in batches of 70. but that third object - the relay list - is very rare, so i end up running a lot of useless queries. the volume of queries would be a lot smaller if i just ignored it.

it doesn't break anything but it feels a bit wasteful of resources.

i suppose what i could do is keep a list - per connected relay - of objects i have already requested in that session. after all, if the object does appear later on, i should expect the relay to announce that, so it doesn't need further querying.