The important and possibly non-obvious pieces of the puzzle are
* Factories are running normally near-100% because they are expensive investments,
* Increasing production capability is a massive investment which takes a lot of money and time, once you decide it, it will take many years before it's in action and then you need to have customers to buy the product for the next 10 years.
* If people can't fully work at said factories due to health restrictions, it will impact the output
* Even normally there are small-scale shortages because supply and demand is so carefully finetuned together. It doesn't require much of extra demand, or much of extra problems to tip that balance off and cause a long backlog in short time, which takes long to rectify (see the first point of near-100% operation to see why; you can just go to 100%, not 110%).
A really simplified calculation would show that if supply is normally working at 95% to satisfy the normal need, if output decreases by 10% to 85.5%, and demand increases from normal (95%) by 10% to 104.5%, during half a year a backlog of (104.5-85.5%)*6months = 114 percentage-months is created, and even if demand goes back to normal (95%) and operation is pushed to 100%, it would take 114 %-months / 6 months = 19 months to satisfy the backlog!
Oversizing all the investments by, say, 25% would pretty much guarantee steady supply in all situations but that would basically be a 25% price increase in all products for no visible benefit under normal conditions.