hmmmm disk space enough i think
Never used LV pool and it has been a long time since i used publish/export, but if you look at your script, you are creating a snapshot, then publishing it, which creates an image, and exporting to it zip, which it has to compress the image (I am not sure if LXD creates another copy before it zips it either).
So by the time you zip it you have the snapshot (??) container 32GB, the image 32GB, and the temporary zip file 32GB, plus have all the other containers and images on the pool. If that is not enough you running export which is now creating another copy to your server another 32GB. So your pool is 197GB, I guess that is why it is slowing down when zipping, because of resources are low during this process.
I remember when i was runing into similar problem it was always the same, disk space was getting low and everything crawled to a halt.
if you dont have this problem on smaller containers, then you know why.