Prometheus常用查询

集群指标

 # CPU核心总数
sum(machine_cpu_cores{cluster="$cluster",job="kubelet"}) by(cluster)
 
# CPU请求总数
sum(kube_pod_container_resource_requests{cluster="$cluster",job="kube-state-metrics",unit="core"}) by (cluster)
 
# CPU限制总数
sum (kube_pod_container_resource_limits{cluster="$cluster",job="kube-state-metrics",unit="core"}) by (cluster)
 
 
# 内存总量
sum(machine_memory_bytes{cluster="$cluster",job="kubelet"}) by (cluster)
 
# 内存请求总量
sum(kube_pod_container_resource_requests{cluster="$cluster",job="kube-state-metrics",unit="byte"})
 
# 内存限制总量
sum(kube_pod_container_resource_limits{cluster="$cluster",job="kube-state-metrics",unit="byte"})
 
 
# 磁盘分区已用空间
sum(node_filesystem_avail_bytes{cluster="$cluster",mountpoint=~"/var/log/pods|/data|/"} and on (instance) (label_replace(up{cluster="$cluster",job="kubelet",metrics_path="/metrics/probes"},"instance","$1:9100","node","([0-9.]+)"))) by (mountpoint,cluster)
 
# 磁盘分区总空间
sum(node_filesystem_size_bytes{cluster="$cluster",mountpoint=~"/var/log/pods|/data|/"} and on (instance) (label_replace(up{cluster="$cluster",job="kubelet",metrics_path="/metrics/probes"},"instance","$1:9100","node","([0-9.]+)"))) by (mountpoint,cluster)

Pod 指标

 # CPU 使用率
sum by (namespace,pod) (irate(container_cpu_usage_seconds_total{cluster="$cluster",namespace=~"$namespace",pod=~"$pod",image!=""}[5m]))
/
sum by (namespace,pod) (kube_pod_container_resource_limits{cluster="$cluster",namespace=~"$namespace",pod=~"$pod",unit="core"}) * 100
 
# CPU 资源请求
sum by (namespace,pod) (kube_pod_container_resource_requests{cluster="$cluster",namespace=~"$namespace",pod=~"$pod",unit="core"})
 
# CPU 资源限制
sum by (namespace,pod) (kube_pod_container_resource_limits{cluster="$cluster",namespace=~"$namespace",pod=~"$pod",unit="core"})
 
 
# Memory 使用率
sum(container_memory_working_set_bytes{cluster="$cluster",image!="",namespace=~"$namespace",pod=~"$pod",image!=""}) by(pod, namespace)
/
sum(kube_pod_container_resource_limits{cluster="$cluster",namespace=~"$namespace",pod=~"$pod"}) by(pod, namespace) * 100 != +inf
 
# Memory 资源请求
sum by (namespace,pod) (kube_pod_container_resource_requests{cluster="$cluster",namespace=~"$namespace",pod=~"$pod",unit="byte"})
 
# Memory 资源限制
sum by (namespace,pod) (kube_pod_container_resource_limits{cluster="$cluster",namespace=~"$namespace",pod=~"$pod",unit="byte"})
 
 
# Pod 1小时内重启次数
delta(kube_pod_container_status_restarts_total{cluster="$cluster",job="kube-state-metrics",namespace=~"$namespace",pod=~"$pod"}[1h]) > 0

posted @ 2023-07-17 09:27 liy36 阅读(586) 评论(0) 编辑收藏举报

刷新页面返回顶部

登录后才能查看或发表评论，立即登录或者逛逛博客园首页

相关博文：

· Kubernetes Pod 日志中报 pthread_create failed: Resource temporarily unavailable 错误

· 在pod中配置kubectl使用serviceaccount token 访问集群

· 常用Prometheus查询

· prometheus安装及使用

· Prometheus常用监控指标

阅读排行：
· 全程不用写代码，我用AI程序员写了一个飞机大战
· DeepSeek 开源周回顾「GitHub 热点速览」
· 记一次.NET内存居高不下排查解决与启示
· 物流快递公司核心技术能力-地址解析分单基础技术分享
· .NET 10首个预览版发布：重大改进与新特性概览！

历史上的今天：
2021-07-17 部署RabbitMQ集群

公告

昵称： liy36
园龄： 5年5个月
粉丝： 4
关注： 4

+加关注

2025年3月

日

一

二

三

四

五

六

liy36

Prometheus常用查询

Prometheus常用查询

集群指标

Pod 指标

公告

搜索

常用链接

我的标签

随笔分类

随笔档案

阅读排行榜

推荐排行榜

	# CPU核心总数
	sum(machine_cpu_cores{cluster="$cluster",job="kubelet"}) by(cluster)

	# CPU请求总数
	sum(kube_pod_container_resource_requests{cluster="$cluster",job="kube-state-metrics",unit="core"}) by (cluster)

	# CPU限制总数
	sum (kube_pod_container_resource_limits{cluster="$cluster",job="kube-state-metrics",unit="core"}) by (cluster)


	# 内存总量
	sum(machine_memory_bytes{cluster="$cluster",job="kubelet"}) by (cluster)

	# 内存请求总量
	sum(kube_pod_container_resource_requests{cluster="$cluster",job="kube-state-metrics",unit="byte"})

	# 内存限制总量
	sum(kube_pod_container_resource_limits{cluster="$cluster",job="kube-state-metrics",unit="byte"})


	# 磁盘分区已用空间
	sum(node_filesystem_avail_bytes{cluster="$cluster",mountpoint=~"/var/log/pods\|/data\|/"} and on (instance) (label_replace(up{cluster="$cluster",job="kubelet",metrics_path="/metrics/probes"},"instance","$1:9100","node","([0-9.]+)"))) by (mountpoint,cluster)

	# 磁盘分区总空间
	sum(node_filesystem_size_bytes{cluster="$cluster",mountpoint=~"/var/log/pods\|/data\|/"} and on (instance) (label_replace(up{cluster="$cluster",job="kubelet",metrics_path="/metrics/probes"},"instance","$1:9100","node","([0-9.]+)"))) by (mountpoint,cluster)

	# CPU 使用率
	sum by (namespace,pod) (irate(container_cpu_usage_seconds_total{cluster="$cluster",namespace=~"$namespace",pod=~"$pod",image!=""}[5m]))
	/
	sum by (namespace,pod) (kube_pod_container_resource_limits{cluster="$cluster",namespace=~"$namespace",pod=~"$pod",unit="core"}) * 100

	# CPU 资源请求
	sum by (namespace,pod) (kube_pod_container_resource_requests{cluster="$cluster",namespace=~"$namespace",pod=~"$pod",unit="core"})

	# CPU 资源限制
	sum by (namespace,pod) (kube_pod_container_resource_limits{cluster="$cluster",namespace=~"$namespace",pod=~"$pod",unit="core"})


	# Memory 使用率
	sum(container_memory_working_set_bytes{cluster="$cluster",image!="",namespace=~"$namespace",pod=~"$pod",image!=""}) by(pod, namespace)
	/
	sum(kube_pod_container_resource_limits{cluster="$cluster",namespace=~"$namespace",pod=~"$pod"}) by(pod, namespace) * 100 != +inf

	# Memory 资源请求
	sum by (namespace,pod) (kube_pod_container_resource_requests{cluster="$cluster",namespace=~"$namespace",pod=~"$pod",unit="byte"})

	# Memory 资源限制
	sum by (namespace,pod) (kube_pod_container_resource_limits{cluster="$cluster",namespace=~"$namespace",pod=~"$pod",unit="byte"})


	# Pod 1小时内重启次数
	delta(kube_pod_container_status_restarts_total{cluster="$cluster",job="kube-state-metrics",namespace=~"$namespace",pod=~"$pod"}[1h]) > 0