for each swift cluster there are two alarms to monitor container and object availability, in swift_new for codfw-prod those alarms get expanded in icinga without the cluster name and thus result in UNKNOWN.
triaging as normal but once swift gets merged into swift_new this needs to be fixed
define service {
# --PUPPET_NAME-- graphite1002 swift__dispersion_object
active_checks_enabled 1
check_command check_graphite_threshold!http://graphite.wikimedia.org!10!swift..dispersion.object.pct_found.value!95!90!1hours!1!--under
check_freshness 0
check_period 24x7
contact_groups admins
host_name graphite1002
is_volatile 0
max_check_attempts 3
normal_check_interval 1
notification_interval 0
notification_options c,r,f
notification_period 24x7
passive_checks_enabled 1
retry_check_interval 1
servicegroups misc_eqiad
service_description swift object availability
}