If an application writes to a path that doesn't exactly match the container path including upper/lower case, or a path that isn't an absolute path, then that path is inside the docker image. This is how you can fill docker image.
If you map a host path that isn't an actual disk or user share, then that host path is in RAM just like the rest of the Unraid OS. This is how you can fill RAM with docker.
You certainly have a lot of dockers. Do you really need all those, or are you just experimenting? Are you really using emby and also plex, for example? Many of those I have no experience with or even know what they are for. Probably we are not going to get into the details of many of them so it would be best if you can figure this out on your own.