The concept you need for generating multiple outputs from a single input with ffmpeg is demonstrated very clearly (at least for me it was) at https://trac.ffmpeg.org/wiki/Creating%20multiple%20outputs
basically, ffmpeg -i someinput [output_options] someoutput1 [output2_options] someoutput2 is the pattern. It looks like you already know how to use the rtsp as an input and you know how to generate a single output with a given bitrate. You wan too repeat the entire output block (everything after your rtsp input string), and each time change your resolution with -s and your bitrate (you're using -crf so you'd want to decrease it for higher res/higher quality and increase it for lower-res, lower quality).
Personally, I would use crf for HLS renditions and with libx265, the rule I use is frame_width * frame_height * frame_rate * 0.1 to get the "ideal" bitrate for a given frame size. you'll have to flip the equation if you want to select bitrates to target, then figure out the framesize from that. The ".1" value can go higher, but you won't see much benefit in general (more complex content will better use a higher value) and I rarely go below 0.07.