Thanks for the link to the PDF. Despite some repeated sentences in the document, it nevertheless provides a good starting point to the issue of reliability and burn-in.

"These are design problems and should NEVER have to be found during "burn in". This show reliance on so called "testing" which is pure BS."

Maybe I worded it wrong. One of the problems that we face here is due to poor vendor support on accessories. Heat sink pastes are nothing but chalk powder, transformer cores are just MildSteel sheets - the list is endless. Thus even with the best of design and cross checks we have had surprise failures due to above. Thus such kind of functional tests for an extended period bring out such black sheeps - that are eventually blacklisted.

Raghu