real-time data processing, kafka versus pulsar

hey every1, im lookin 2 set up a real-time data pipeline 4 processin user activity. ive been diggin into message queues and stream processin platforms. kafka is the usual suspect, but apache pulsar seems 2 b gainin traction w/ its unified queuin and streamn model. has any1 here worked w/ both, or have strong opinions on which is better 4 high-throughput, low-latency scenarios? what r ur experiences w/ operational complexity and feature sets