My docker image information sources are (1) hub.docker.com (2) github (3) linuxserver and others. Start here. I recently installed 'linuxserver/jackett' and 'linuxserver/radarr' from scratch, and it's not difficult when configuring the service from the web page itself.
So here's what I found for myself (with a query at the end).
Using PuTTY to connect to the Pi4 over SSH I was able to look at docker and know it was there just by typing docker at the command prompt. Then using command docker search sabnzbd I could reveal a whole load of places I could download a SABnzbd docker image from. Knowing that the linuxserver.io fleet source was a good one (at least according to Servarr for Radarr/Sonarr/etc) I used the command docker pull linuxserver/sabnzbd. Once I'd got that I ran it using docker run linuxserver/sabnzbd.
My understanding is that this is an instance that runs only for the duration of the command issued (so if I inspected a file it would create the container, show me the file and then kill itself off. But the container continues to exist. Is this correct? Is this why I need to prune redundant containers before they get out of control?
While SABnzbd was running in PuTTY I could see it was opening a port (amongst other things), but I couldn't open a webpage to the IP:port. Is this to do with the port being created for the container, but the container port then needing to be mapped to the Pi ports? Do I now need to run the container with startup options, or something, to make the port mapping happen and reveal it over the network?