Containers kinda designed to run one “EXE”, so unifi-cam-proxy follows that model. So you can download source, modifying it, and re-build the container locally using Docker Build. There are techniques like using http://supervisord.org in Dockertfile/enterpoint.sh to have it launch multiple processes…
But I’m thinking it’a the video transcoding that’s using the CPU/memory, than the containers themselves. The FFMPEG library it uses is resource intensive. So I’m not sure it’s going to be dramatically different CPU with two-proxys-in-one container. Memory likely be lower if combined since it doesn’t have the container 2nd image in RAM.