Hi,
I've just setup a test cluster with cephadm using quincy. Things work
nicely. However, I'm not sure how to "handle" alertmanager and prometheus.
Both services obviously aren't crucial to the working of the storage,
fine. But there seems to be no built-in fall-over concept.
By default, the active mgr accesses the services using
host.containers.local, thus assuming that they run an the same machine
as the active manager. This assumption is true after the initial
installation. Turning off the host with the active manager activates the
stand-by on another machine, but alertmanager and prometheus are gone
(i.e. not "moved along"). So the active manager produces lots of error
messages when logging into it. Turning the tuned-off machine on again
doesn't help, because alertmanager and prometheus are back, but on the
wrong machine.
I couldn't find anything in the documentation. Are alertmanager and
prometheus supposed to run in some HA-VM? Then I could add the HA-VM to
the cluster with (only) these two services running on it and make the
URIs point to this HA-VM (ceph dashboard set-alertmanager-api-host ...,
ceph dashboard set-grafana-api-url ..., ceph dashboard
set-prometheus-api-host...).
How is this supposed to be configured?
- Michael
_______________________________________________
ceph-users mailing list -- ceph-users@xxxxxxx
To unsubscribe send an email to ceph-users-leave@xxxxxxx
