Index an archive or tag page only when it carries real search demand and unique value, and noindex the thin, duplicative, auto-generated ones, which is most of them. The decision splits on a single condition: does this page represent something people actually search for, organized in a way they cannot get from the posts themselves? A tag that is a genuine topic hub (a subject readers look for and that pulls together a real body of work) earns indexing. The default for everything else is noindex, because most tag and archive pages are machine-generated lists that add bloat and duplication without adding a destination worth ranking.
The instinct to index all of them for more coverage gets the mechanism backwards. Each auto-generated archive and tag page is usually a thin slice of content that already lives on the individual posts, often near-identical to other archive pages and to the category itself. Indexing them does not extend your coverage, it multiplies low-value URLs that compete with each other and with your real pages for crawl attention and relevance signals. Coverage is about distinct, demand-backed pages, not about how many auto-pages you let into the index.
The exception is genuine and worth protecting. If a tag has become a real topic hub (it has search demand, a curated set of strong posts, and ideally a bit of unique framing at the top), it functions as a useful landing page and should stay indexed. The line is demand plus unique value. A handful of your tags may clear that bar. The dozens generated automatically as you publish almost never will, and leaving them indexed by default is how a small blog ends up with thousands of thin URLs diluting its signal.
For your next audit, pull a list of every archive and tag page and apply the condition one by one. Keep indexed only the ones with real search demand and a reason to exist as their own page, and set the rest to noindex. When in doubt, default to noindex, because a thin archive costs you more indexed than it ever earns.