So recently my unraid webui has been running slowly, sometimes not showing the list of drives for a minute or two, or not at all. This finally culminated in an http 500 internal server error when I tried to load the main page. This isn’t good, and I was still able to access the server over ssh and via smb/nfs, so it wasn’t completely dead.
Time to debug over ssh. I first went to top and saw that shfs was at 100%+ CPU usage. This felt like an odd program to be using more than one CPU core of performance. I then went to the syslogs and saw lots of failed login attempts from a specific IP address. This wasn’t one I recognized, but I thought it might be a container I’m running (unraid api specifically).
I then went and ran docker ps to get the list of containers, and more importantly, the name for my unraid API container. I then ran docker inspect on that container and found it had the IP in question. Finally I went and ran docker kill on the container and the unraid webui started responding again. I guess I had misconfigured the container so it was falling to login, and retrying aggressively enough to stall out the webui.
Note to self, apps that scrape the webui should be configured correctly or else they may have some unintended consequences.