We were having some issues where visiting the site would occasionally result in a "404 rate limit exceeded error.” This is due to our anti-abuse measures accidentally being sent the IP of the actual docker machine, as opposed to that of the user.

To be 100% transparent, IP logs are stored in a 50 MB ram disk. If any new data is written to a full 50MB disk, old data will be purged to make room. Since it’s in ram, it will also not persist after, say, a shutdown.

These 50MB ram disks are on both the VPS we use to route traffic to the actual dedicated server and the dedicated server itself. Both stored in ram, both 50MB in size. This is so the main server knows what IPs are causing mischief, otherwise it’d just get the IP of the VPS, which wouldn’t be of much use for limiting abuse.

This is only to prevent abuse by automatically banning things like bots or potential DDoS attacks. We will never share/sell or otherwise give your data to anyone. Not that we have much of anything to give, anyway.

Anyway, the hiccups should be resolved now and we should be back to running normally.

  • @BurgerMA
    link
    61 year ago

    Feel free to ask us any questions below if you have any concerns, etc. I suck at explaining the technical concepts in an easily digestable format, but I can try. @Disabled@burggit.moe is way better at this, tbh. It’s why I’m the backend boi.