Resolve "The AlertmanagerClusterDown alert fires, even though Alertmanager is healthy"
General MR
Summary
The customer wrote the following with a theory.
The AlertmanagerClusterDown
alert fires, even though Alertmanager is healthy. This happens because Prometheus cannot access the Alertmanager reloader-web
port, due to lack of an AuthorizationPolicy
port entry.
Initial thought was all that had to be added was a authorizationpolicy. However, upon further investigation in order to allow for prometheus to be able to access the necessary port 8080 via a GET involved a network policy
, authorizationpolicy
and peerauthentication
.
Relevant logs/screenshots
The following shows the error when one of the 3 policies/kinds is missing
After the fix with all 3 configured
Relates #146 (closed)