This would depend entirely on how much churn your system is doing on containers/volumes/images. Once a day sounds really often for most situations.

"Regularly" = when you're running out of space because of a bunch of built up old stuff.

The operation itself is pretty cheap... but the loss of cache and constantly pulling/writing to disk makes this counter-intuitive, IMO. Agree: prune when there's pressure.

Also, be sure anything you rely on/want to keep is bound to a running container. Otherwise pruning [without filtering] might be eager. Notably, the 'system' context includes networks.