if your relay has a garbage collector on it, it's because it has a bigger store elsewhere that it fetches events from, even if the event was purged from the relay's cache as soon as it's requested again it's fetched and stored after sending it back to the user

you aren't going to use the GC without this being the case because the purpose is distributing the load of data access and concentrating the archive in a small cluster and users read from the caches

Reply to this note

Please Login to reply.

Discussion

to be clear, this is precisely for the purpose of optimizing resource utilization among a number of participating relay operators

maybe if your data was just an endless stream of twitter style notes then you might use a GC standalone and avoid having to monitor disk utilization of your relay, something that would probably save for example Will a lot of time with constantly nukening the database