"Purge Everything" reporting success while curl still sees HIT is the classic control-plane vs node gap. On a multi-node CDN the same failure mode shows up when purge is acknowledged centrally but one edge never applied it. We have been building an open-source CDN control plane (CDNCA, AGPL-3.0) so one plane can see what each self-hosted node actually purged, plus a plugin marketplace; marketplace features roll out release by release. Was the miss a wrong cache path, a second cache layer, or the helper talking to a different Nginx than the one serving traffic?