The networkpolicy here

https://github.com/prometheus-community/helm-charts/blame/b9b681eec0219b6a5300f75a2df6fd48b7c5551a/charts/kube-prometheus-stack/templates/alertmanager/networkpolicy.yaml#L39

only allows traffic from prometheus to alertmanager. However, if traffic from prometheus to config-reloader web port isn't allowed, the alertmanager cluster down alert will fire

The attached image shows the current networkpolicy and then after changing it to also allow traffic to the configreloader web port