Feature #48102
cephadm: configure HA (cluster flags) for Alertmanager
% Done:
0%
Source:
Support
Tags:
HA
Backport:
Reviewed:
Affected Versions:
Pull request ID:
Description
While HA in metrics (Grafana) is less relevant (manual switchover is the basic workaround), alerting should always perform in HA fashion: otherwise there's no way to distinguish whether alerts are not received because there are no alerting conditions or because the alerting stack is down.
Instructions on how to configure the HA for Alertmanager:
https://github.com/prometheus/alertmanager#high-availability
History
#1 Updated by Sebastian Wagner almost 3 years ago
Isn't the altertmanager already HA by itself? I thought that alertmanager already creates a fault-tolerant cluster on its own?
#2 Updated by Sebastian Wagner almost 3 years ago
- Priority changed from High to Normal
#3 Updated by Sebastian Wagner over 2 years ago
- Priority changed from Normal to Low