Everything is now operational! The major factor in this incident was a Docker container which hung with 100% CPU usage, and started filling up the disk and memory. A couple reconfigurations and restarts later, everything is nominal.
We will be setting up better monitoring systems so we're aware of this earlier next time.