alertmanager mesh communication between pods refusing

[Nabapadma-sarker]

Events:
  Type     Reason     Age                  From               Message
  ----     ------     ----                 ----               -------
  Normal   Scheduled  25m                  default-scheduler  Successfully assigned monitoring/alertmanager-main-0 to servertest2-hp-elitedesk-800-g5-sff
  Normal   Pulled     25m                  kubelet            Container image "quay.io/prometheus-operator/prometheus-config-reloader:v0.66.0" already present on machine
  Normal   Created    25m                  kubelet            Created container init-config-reloader
  Normal   Started    25m                  kubelet            Started container init-config-reloader
  Normal   Started    25m                  kubelet            Started container alertmanager
  Normal   Created    25m                  kubelet            Created container alertmanager
  Normal   Pulled     25m                  kubelet            Container image "quay.io/prometheus/alertmanager:v0.25.0" already present on machine
  Normal   Pulled     25m                  kubelet            Container image "quay.io/prometheus-operator/prometheus-config-reloader:v0.66.0" already present on machine
  Normal   Created    25m                  kubelet            Created container config-reloader
  Normal   Started    25m                  kubelet            Started container config-reloader
  Warning  Unhealthy  24m (x5 over 24m)    kubelet            Liveness probe failed: Get "http://10.244.1.95:9093/-/healthy": dial tcp 10.244.1.95:9093: connect: connection refused
  Warning  Unhealthy  10m (x161 over 25m)  kubelet            Readiness probe failed: Get "http://10.244.1.95:9093/-/ready": dial tcp 10.244.1.95:9093: connect: connection refused
  Warning  BackOff    6s (x43 over 13m)    kubelet            Back-off restarting failed container alertmanager in pod alertmanager-main-0_monitoring(8aee3dff-42ef-4f94-8d46-0e7f10853380)

can anyone help which network policy could be the cause for this connection refuse

[kekekekekeshi]

我也遇到这个问题了。进入容器中发现能访问3000端口，并且能请求到数据，但是通过svc就不行。以及健康检查也不行。 我的k8s版本是1.28.2

[kekekekekeshi]

Events:
Type Reason Age From Message

---

Normal Scheduled 57s default-scheduler Successfully assigned monitoring/grafana-5d55b655cf-6pwt5 to n1
Normal Pulling 56s kubelet Pulling image "grafana/grafana:10.2.0"
Normal Pulled 21s kubelet Successfully pulled image "grafana/grafana:10.2.0" in 35.259s (35.259s including waiting)
Normal Created 21s kubelet Created container grafana
Normal Started 21s kubelet Started container grafana
Warning Unhealthy 17s (x4 over 20s) kubelet Readiness probe failed: Get "http://10.244.40.154:3000/api/health": dial tcp 10.244.40.154:3000: connect: connection refused

[kekekekekeshi]

问题已经解决了，主要是网络策略这只允许来着prometheus的流量。这里需要调整一下为允许所有的就行了

[kekekekekeshi]

把ingress的from去掉就可以