I have all the proxy_set_header directives in nginx, including the new ones just added in yesterday or the day before.

I noticed this in my logs a short while ago: Dec 17 20:07:03 [lemmy] [2020-12-17T20:07:03Z ERROR actix_http::response] Internal Server Error: LemmyError { inner: {"error":"Too many requests. type: Message, IP: , 180 per 60 seconds"} }

There was a bunch of them over a period of a minute or so.

In cross-referencing the time with the actual nginx access logs, it looks erroneous. There were a flurry of systems accessing a post, but never 180 requests from any single source.

I notice the error message there has a blank IP: - and I wonder if everything in the lemmy backend is able to handle IP’s in the IPV6 format - as nginx is logging requests eg: “::ffff:162.243.128.20”

It’s only been a one-off so far, but I may go increase the rate limiting in the short term if there’s some breakage there.