monitoring: alert for pool fill up broken

Fixes: https://tracker.ceph.com/issues/44991
Signed-off-by: Volker Theile <vtheile@suse.com>
(cherry picked from commit e197e4d7f481395e2cef3e7acf61254906d5c32b)

diff --git a/monitoring/prometheus/alerts/ceph_default_alerts.yml b/monitoring/prometheus/alerts/ceph_default_alerts.yml
index aebd816156066196da0b4eec3374ceab36d0fe89..a0074f0779f0c1c691b5b0b6e723ea3bebcaf338 100644 (file)
--- a/monitoring/prometheus/alerts/ceph_default_alerts.yml
+++ b/monitoring/prometheus/alerts/ceph_default_alerts.yml
@@ -232,7 +232,7 @@ groups:
           (
             predict_linear(ceph_pool_stored[2d], 3600 * 24 * 5) >=
             ceph_pool_max_avail
-          ) * on(pool_id) group_right(name) ceph_pool_metadata
+          ) * on(pool_id) group_left(name) ceph_pool_metadata
         labels:
           severity: warning
           type: ceph_default