api: ceph: report whether a health check blocks a rolling restart

The web interface has to show which health checks stop a rolling
restart. On a cluster upgraded to Ceph 19.2.6 or 20.2.4 that is a
smaller set than the checks that are currently firing.

Report what the shared health classification decided for each check,
not whether the check is on the allowlist. The two differ for
OSDMAP_FLAGS, which is never allowlisted but judged by the flags that
are set. On a cluster with only noout set, that check would show up as
blocking while the restart then runs anyway. The answer is per daemon
type, because OSD map flags only ever stop an OSD restart.

Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>