I'm currently unhappy with the glass-break sensors that are available commercially. I won't rehash my complaints here, you've probably all heard them before. I was brainstorming with a friend and not coming up with anything better idea-wise though, and he suggested the conductive paint used for rear window defrosters on automobiles.
They do sell the paint/epoxy for repair kits (hadn't heard of this til he mentioned it).
Now, I have no idea how you could retrofit this, and given the cost for new windows (even cheap stuff off the shelf at Home Depot runs about $400 each for standard sizes, custom stuff commonly runs upward of thousands each) it probably isn't cost-effective to have someone build them for you.
I think you'd have four traces which could be close to the edge of the panes or potentially even hidden by the sash and rails, maybe even having an appearance of being decorative.
There are a few assumptions here. First off, if the window cracks but doesn't break completely, would that interrupt the trace enough to be detectable? Second, that you could even have a powered device in the moving part of the window without massively over-engineering it. And finally, how long would this even last? I don't think the rear defroster in my car lasted 6 months before there were cold spots in it that didn't melt. Oh, and don't forget the double/triple pane stuff (would this go on the exterior pane, the interior, or both?).
Also it's quite a few gpio pins to dedicate to even a single window.