It ate the bottle, or the bottle was too big
On this page11
The resin was wrong
A solvent in PET softens, crazes or splits it. "Chemical resistance is the entire decision, check the resin chart first" is the archetype's own words and it is the first failure.
The liner was never specified
A cap with the wrong liner lets a solvent out slowly and an acid corrode inward. The bottle gets blamed and the liner did it.
The container was over the limit
A gallon of a Category 1 flammable in plastic is above the pint the table allows for glass or approved plastic. Being physically possible is not the test.
A metal bottle substituted for a safety can
The metal row of Table H-12 is not a licence to use any metal container. A safety can is a defined object with an approval, and this shelf has neither.
The glass jar on a concrete floor
Glass is the right chemistry and the wrong material near a vice and a hammer. Where it has to be glass, it has to be stored where it will not be knocked.
The cyanoacrylate that cured in storage
A washed bottle that was not dry, or too much headspace. Moisture is the cure trigger and the container is where it comes from.
The thread locker that cured too
Anaerobic, so an airtight container is the wrong choice. This is the one product in the group where the original packaging is usually the right answer.
The finish that no longer matched
Carrier lost through the closure over two years, so the touch up went on thicker and darker. Fill it full, use a liner, keep it dark.
Test the way it will sit
Filled, capped, upright for a month and on its side for a week, warm, in the shop rather than on a desk. Every failure here appears in that test.
And weigh it
Weigh a filled unit, leave it, weigh it again. A measurable loss is a product that will not be usable when somebody reaches for it.