Being in the index is not a permanent status you earn once. It is a standing judgment Google keeps re-deciding, and a page can hold its place for weeks and then lose it, because each re-crawl is a fresh evaluation of whether the page still deserves a slot. The drop weeks later is not a delayed reaction to something at publishing. It is simply when the next re-crawl and re-score happened to land.
Think of it as a lease rather than a deed. The page occupies index space on terms that get reviewed, and a few things commonly turn the review negative. A core update can raise or shift the quality bar, so a page that cleared the old bar no longer clears the new one. The page’s own value may have decayed, or it may never have drawn the clicks and engagement that justify keeping it, and Google periodically clears pages that earn nothing to keep the index lean. Freshness can lapse on a page that needed to stay current. Or the site’s overall standing can slip, dragging borderline pages out with it.
The timing falls weeks out because re-crawls are spaced and updates roll on their own schedule, not because anything broke. Filing the disappearance under “glitch” is the wrong frame: it was a decision, made on re-evaluation, against signals that changed in the interim. So when a page drops, the useful move is to look back at the weeks before it went, for a core update in that window, content that went stale, traffic that quietly died, or a sitewide change.
One check first, before you treat it as deindexing at all. Confirm in the URL Inspection tool that the page is actually not indexed, rather than still indexed but ranking lower and pulling fewer impressions, which a dashboard can make look like a vanishing. A true drop and a ranking decline are different problems, and only the real drop is the standing-judgment story above.