This doesn't relate to the (continuing) recent buzz about deleting valuable content, because this isn't about worthless questions that just happen to have good answers; this is about useful questions themselves.
At least twice we have had questions pop up in reference to an event on May 8th where migration was temporarily broken and we had a truly heartwarming display of community one-ness as 16 individuals united to try and migrate one question (with two double-votes). The first was, naturally, when the event was first happening, and Geoff Dalgas reported that an issue was found and a fix would be rolled out (and it subsequently did). Some time later, a second question inquired "How many users does it take to migrate one question?", highlighting the spectacle once again. Lots of us who remembered the original incident sought it out, but it turns out that it was deleted. A 10k managed to find the ID of the question, while others resorted to simply summarizing their memory of the incident. Don't know when it happened, but today I checked and the second question has likewise ceased to be.
Due to time-sensitivity and the fact that the bug should no longer be present, it makes sense for those kinds of questions to be closed as too localized once the bug is fixed - no further contribution is necessary on that part. To me, it would then make sense to keep the question around: it's still useful information. It can be used as a reference if anyone inquires about side-effects and aftermath, as well as a good resource to look back on if something similar happens in the future. So why are we deleting these questions? All that accomplishes is less ability to keep track of what went wrong in the past, which I think is quite harmful to everyone involved in the site.