]> git.apps.os.sepia.ceph.com Git - ceph-ci.git/commit
mgr/prometheus: track individual healthchecks as metrics
authorPaul Cuzner <pcuzner@redhat.com>
Thu, 16 Sep 2021 23:24:29 +0000 (11:24 +1200)
committerPaul Cuzner <pcuzner@redhat.com>
Fri, 22 Oct 2021 00:32:39 +0000 (13:32 +1300)
commite0dfc02063ef40cf6a1dc6e3080d0a856ceff050
tree0319a077dbcbebea1ff1f7a72d2e1a058f59cd4f
parentb70647e8796c78a4f551a1edd9ebceb06255dced
mgr/prometheus: track individual healthchecks as metrics

This patch creates a health history object maintained in
the modules kvstore.  The history and current health
checks are used to create a metric per healthcheck whilst
also providing a history feature. Two new commands are added:
ceph healthcheck history ls
ceph healthcheck history clear

In addition to the new commands, the additional metrics
have been used to update the prometheus alerts

Fixes: https://tracker.ceph.com/issues/52638
Signed-off-by: Paul Cuzner <pcuzner@redhat.com>
doc/mgr/prometheus.rst
monitoring/prometheus/alerts/ceph_default_alerts.yml
monitoring/prometheus/alerts/test_alerts.yml
src/pybind/mgr/prometheus/module.py