There's an effect, on any website, that when someone even semi-popular mentions something, everyone goes to it at once, it will likely fall over. To run an njump you would probably want to at least cache the previous results on a CDN or something right? I guess, I didn't realize you weren't doing that. It looks like it's just a server with the stuff cached on it's disk.
Or, it sounds like some of the issue is it's getting requested many events it cannot find? IE, they just randomly ask for random IDs with no relay hints or anything and you have to go scrape that for them each time?
I admire the idea, but it all sounds hard and prone to being a victim of it's own success.
