Rapid firing of replaceable events puts an immense burden on relays.
Discussion
depends on the implementation of the relay, I guess
UPDATES are expensive, but relays could be INSERT only ("log style")
this is exactly what happens when you onboard an rss feed type of source. is operating your own relay the only reasonable way to get it done? maybe paying for higher rate limits?
What do you mean?
I mean that there is a legitimate reason to want to emit a rapid-fire stream of kind 30023 events, that being when a new RSS feed is onboarded to nostr.
so if that is a burden, it would be nice to sort out a way to allow the behavior and mitigate the burden.
one idea of how to do that would be to compensate the relay operator for the capacity to railgun a bunch of articles into their relay from time to time.
That's not the issue. 30023s are fine. The same article doesnt get updated very often. The problem is some other kinds that get updated many times per minute. I know I've seen them, but now I don't know exactly what they are or where they're coming from.
Unpopular opinion: replaceable notes is a mistake, relays should be insert-only machines. Replacements should be managed by the clients.
It seems most relays throttle though right? Paid relays + throttling seems to completely prevent the problem. Although, as someone who is writing more clients, I feel that it’s encouraging me to write cursed timeouts. Even worse when I need the user to sign multiple events.
Lol nice Pong simulation. But also, Drivechains are not good.