误区：因为prometheus是一个pod不能重启，所以此方案适用于单独非容器安装的prometheus。
1.进入pod
kubectl exec -n istio-system -it grafana-694477c588-8rbvc /bin/bash
2.下载安装Prometheus AlertManager
grafana-cli plugins install camptocamp-prometheus-alertmanager-datasource
3.restart grafana service
service grafana-server restart
4.查看错误原因
grafana-server -homepath /usr/share/grafana
Right. Something is already using port 3000.
Stop the grafana services
systemctl stop grafana.service -l
do a ps aux |grep grafana and kill if any process exits
Then start grafana using
systemctl stop grafana.service -l
Tailf for grafana logs
tailf /var/log/grafana/grafana.log
5.查看本地安装的Plugin
grafana-cli plugins ls
service grafana-server restart
