ceph-mon quincy/stable 193
Even after going through the following issue,
https://bugs.launchpad.net/charm-ceph-mon/+bug/2043758
COS integration about the alert rules will complain about the setup.
Not sure how critical the substitution though.
unit-ceph-mon-0: 03:41:33 DEBUG unit.ceph-mon/0.juju-log Could not locate cos-tool at: "cos-tool-amd64"
unit-ceph-mon-0: 03:41:33 DEBUG unit.ceph-mon/0.juju-log Skipping injection of juju topology as label matchers
unit-ceph-mon-0: 03:41:33 DEBUG unit.ceph-mon/0.juju-log `cos-tool` unavailable. Leaving expression unchanged: ceph_health_status == 2
unit-ceph-mon-0: 03:41:33 DEBUG unit.ceph-mon/0.juju-log `cos-tool` unavailable. Leaving expression unchanged: ceph_health_status == 1
unit-ceph-mon-0: 03:41:33 DEBUG unit.ceph-mon/0.juju-log `cos-tool` unavailable. Leaving expression unchanged: (
(ceph_health_detail{name="MON_DOWN"} == 1) * on() (
count(ceph_mon_quorum_status == 1) == bool (floor(count(ceph_mon_metadata) / 2) + 1)
)
) == 1
unit-ceph-mon-0: 03:41:33 DEBUG unit.ceph-mon/0.juju-log `cos-tool` unavailable. Leaving expression unchanged: count(ceph_mon_quorum_status == 0) <= (count(ceph_mon_metadata) - floor(count(ceph_mo>
unit-ceph-mon-0: 03:41:33 DEBUG unit.ceph-mon/0.juju-log `cos-tool` unavailable. Leaving expression unchanged: ceph_health_detail{name="MON_DISK_CRIT"} == 1
unit-ceph-mon-0: 03:41:33 DEBUG unit.ceph-mon/0.juju-log `cos-tool` unavailable. Leaving expression unchanged: ceph_health_detail{name="MON_DISK_LOW"} == 1
unit-ceph-mon-0: 03:41:33 DEBUG unit.ceph-mon/0.juju-log `cos-tool` unavailable. Leaving expression unchanged: ceph_health_detail{name="MON_CLOCK_SKEW"} == 1
unit-ceph-mon-0: 03:41:33 DEBUG unit.ceph-mon/0.juju-log `cos-tool` unavailable. Leaving expression unchanged: count(ceph_osd_up == 0) / count(ceph_osd_up) * 100 >= 10
unit-ceph-mon-0: 03:41:33 DEBUG unit.ceph-mon/0.juju-log `cos-tool` unavailable. Leaving expression unchanged: ceph_health_detail{name="OSD_HOST_DOWN"} == 1
unit-ceph-mon-0: 03:41:33 DEBUG unit.ceph-mon/0.juju-log `cos-tool` unavailable. Leaving expression unchanged: ceph_health_detail{name="OSD_DOWN"} == 1
unit-ceph-mon-0: 03:41:33 DEBUG unit.ceph-mon/0.juju-log `cos-tool` unavailable. Leaving expression unchanged: ceph_health_detail{name="OSD_NEARFULL"} == 1
unit-ceph-mon-0: 03:41:33 DEBUG unit.ceph-mon/0.juju-log `cos-tool` unavailable. Leaving expression unchanged: ceph_health_detail{name="OSD_FULL"} > 0
unit-ceph-mon-0: 03:41:33 DEBUG unit.ceph-mon/0.juju-log `cos-tool` unavailable. Leaving expression unchanged: ceph_health_detail{name="OSD_BACKFILLFULL"} > 0
unit-ceph-mon-0: 03:41:33 DEBUG unit.ceph-mon/0.juju-log `cos-tool` unavailable. Leaving expression unchanged: ceph_health_detail{name="OSD_TOO_MANY_REPAIRS"} == 1
...
I think this can be Invalid or Low in the end. It seems that the transformation of the alert rules are done on the other side of the relation. Marking this as Incomplete for the time being and let it expire.