A spike in “crawled, currently not indexed” pages usually signals a quality or duplication problem at scale, not a technical glitch. The status means Google crawled those pages and then judged them not worth indexing, so a sudden batch of them is a value verdict on a group of pages, delivered all at once. The pattern almost always points to thin, duplicate, or low-value content, frequently the byproduct of bulk or templated publishing, where many similar pages went live and Google declined most of them. This is observed behavior across sites rather than a rule Google publishes, so confirm it against what those specific pages actually contain.
The read to take is batch-quality, not batch-error. The pages were crawled successfully, which means Google had no technical trouble reaching them; it simply decided they did not earn a place in the index. So the spike is feedback on the content, not on the crawl. When the affected URLs share a template, come from a programmatic or scaled publishing run, or repeat the same structure with minor variation, the duplication or thinness is the explanation, and the volume of the spike usually matches the volume of the batch that triggered it.
The mistake is treating it as a technical issue to resubmit away. Asking Google to recrawl or reindex pages it already crawled and rejected changes nothing about why it rejected them, so the pages come back as not indexed. Resubmission addresses discovery problems, but this is not a discovery problem; Google found and assessed these pages and returned a verdict. The spike will not clear until the underlying value or duplication issue does.
To act on it, pull the affected URLs from the coverage report and audit the batch for thinness and duplication rather than hitting resubmit. Look for templated content with little unique value, near-duplicate pages that should be consolidated, and pages that exist for volume rather than for readers. Improve, merge, or remove the weak pages, and let the genuinely useful ones earn indexing on their merits. Treat the spike as a quality audit prompt, not a button to press.