Prometheus monitoring K8S cluster Pod
Currently cAdvisor kubelet integrated into the assembly, can kubernetes each cluster node kubelet started using metrics cAdvisor interface provides access to relevant data of the performance index for all nodes container.
The default port cAdvisor provide services to *** *** 4194, offers two interfaces:
- Prometheus Interface Format Indicator: nodeIP: 4194 / metrics (or exposed through kubelet cadvisor interfaces nodeIP: 10255 / metrics / cadvisor);
- WebUI interface Interface: nodeIP: 4194 / containers /
The above data are in the format of the output interfaces of prometheus.
- Interface kubelet metrics of nodes cAdvisor provided by the node acquires
- There are performance related data containers.
Exposed interface address:
https://NodeIP:10255/metrics/cadvisor
https://NodeIP:10250/metrics/cadvisor
Grafana deployment and visualization in the K8S
grafana is a visual panel has a very nice graphics and layout display, full-featured metrics dashboards and graphics editor that supports Graphite, zabbix, InfluxDB, Prometheus, OpenTSDB, Elasticsearch etc. as a data source, than Prometheus comes charts show too many powerful, more flexible, has a wealth of plug-ins, more powerful.
- Download: https: //grafana.com/grafana/download
Cluster resource monitoring template number: 3119
1, into a newly created address Prometheus
2. Import Template: Use the template 3119
3, to see the effect
4, adjust the monitoring indicators based on the source server
5, modify the parameters according to the situation
6, to see the effect
