The solvent attacked the bottle
Updated
On this page8
Softening and swelling
A wall that gives under finger pressure, a cap that has gone tacky, or a liner swollen enough to jam in the neck. All of them develop over weeks and all of them pass a bench check on the filling day.
Paneling from permeation
A solvent passing out through the wall reduces the internal volume and the bottle draws in. That is diagnostic: a dented bottle with no impact damage is usually telling you about permeation.
The product changes, not the container
Sometimes the bottle looks perfect and the paint has thickened because the thinner left. Weighing a filled, sealed sample at intervals is the cheapest test for this, and almost nobody runs it.
The liner is a separate part with a separate answer
HDPE may be fine for your binder while the foam or the induction-seal facing is not. Of the 236 closures here the liner is published on each one, and the pairing is what you test rather than the shell.
Glass changes the question, not the whole problem
The 42 glass containers on this shelf remove the wall from the argument entirely. The cap, the liner and any fitment are still polymer, so the test list gets shorter rather than empty.
Ask the supplier in writing, per binder
Not "is this solvent resistant" but "has this closure and liner been tested against this solvent at this concentration". The first question gets an optimistic answer and the second gets a useful one.
Keep retained samples and actually look at them
Filled, sealed, dated, one upright and one on its side, checked at one, three and six months. Every failure above is visible by then and none of them is visible now.