The problem is that for this type of thing credible reports of failure carry more weight than reports of success. If you're making a big bet on the reliability of file system you want to mitigate risk, and for all I know either you haven't tickled a bug I would or you've just been lucky so far.
When I looked hard at this 2-3 years ago, it was clear to me that without a tape backup system I couldn't justify purchasing, ZFS just wasn't there yet. I've heard nothing since them to convince me it has been sufficiently polished to get over my threshold or required reliability.
ADDED: I don't think it can be said that ZFS is established to the same degree that the parent's examples are (Java, BEA and Oracle's DB).