A deploy-only 404 needs route, rewrite, and cache evidence before a fix

A debugging note about checking direct URL access, production route output, rewrite rules, and CDN cache before changing app code.

When a 404 only appears after deploy, I try not to label it a routing bug too early. The exact URL, direct browser request, build output, rewrite rule, CDN layer, and stale HTML/asset timing all matter. A cache purge may make the page work, but it can also hide whether old HTML pointed to a route that was not present at the edge. The first useful note is usually which layer served the 404.