Something that came up in one of Spinecracker's threads may help explain part of the reason beyond those already listed: Inspection stamp impressions located in the stressed area! The localized deformation is a perfect nucleus for "fatigue failure" in an area that wouldn't otherwise be expected to see large cyclical loads. Little things can make a huge difference!
Link to Spinecracker's thread below, ref. Posts #53 through #58- possibly more.:
https://www.milsurps.com/showthread.php?t=21764&page=6
"New '41 Long Branch..."
Railroad bogeys (wheels, specifically) were one of the first well investigated items where tiny inspection stamps in the web led to catastrophic failures (and some derailments!) long into their service lives. I know that Glock Inc. moved the location of their slide proof marks away from the ejection port for the same reason. (I may have even been the one that first recognized the problem, but I can't verify that suspicion- Austria wasn't very forthcoming at the time when dealing with new dramas.)
Other "crack starters" that spring to mind that don't lead to immediate failures: Machining marks, heat treatment micro cracks, Chrome plating, random scratches, corrosion pits...Aside from the Chrome, many of these features can be very small!