作者:尹正杰
版权声明:原创作品,谢绝转载!否则将追究法律责任。
一.ceph集群的监控
1.查看集群的架构
[root@ceph141 ~]# ceph orch ls
NAME PORTS RUNNING REFRESHED AGE PLACEMENT
alertmanager ?:9093,9094 1/1 10m ago 47h count:1
ceph-exporter 3/3 10m ago 47h *
crash 3/3 10m ago 47h *
grafana ?:3000 1/1 10m ago 47h count:1
mds.oldboyedu-cephfs 2/2 10m ago 20h count:2
mgr 2/2 10m ago 47h count:2
mon 3/5 10m ago 47h count:5
node-exporter ?:9100 3/3 10m ago 47h *
osd 7 10m ago - <unmanaged>
prometheus ?:9095 1/1 10m ago 47h count:1
rgw.yinzhengjie ?:80 1/1 10m ago 104m ceph142
[root@ceph141 ~]#
温馨提示:
不难发现,有grafana,alertmanager,ceph-exporter,Prometheus等组件默认都是安装好的,说白了,无需手动安装。
所以,基于cephadm方式部署的环境,可以直接使用Prometheus监控。若使用的ceph-deploy方式部署,则需要手动配置各组件。
2.查看Prometheus的WEbUI
http://10.0.0.141:9095/targets?search=
3.查看grafana的WebUI
https://10.0.0.141:3000/
4.查看node-exporter
http://10.0.0.141:9100/metrics
5.查看alertmanger
http://10.0.0.141:9093/#/status
二.自实现Prometheus监控参考链接
推荐阅读:
https://github.com/digitalocean/ceph_exporter
https://github.com/blemmenes/radosgw_usage_exporter