Improve haproxy and ceph dashboard integration
Passive ceph mgr are configured to emit an error with status code 404 for ceph dashboard. Haproxy is configured to treat status 404 as server state "NOLB" instead of marking servers as "DOWN". The clear separation between a passive server with state "NOLB" and a malfunctioning server with state "DOWN" will allow more differentiated alerting on haproxy server states.
Part of https://github.com/osism/issues/issues/1013
We should also add this to the osism/cfg-cookiecutter repository as default.
We should also add this to the osism/cfg-cookiecutter repository as default.
Done in https://github.com/osism/cfg-cookiecutter/pull/610
But it still needs testing