PromQL 常用语句

一、PromQL 操作符
1.1、查看根分区大小

node_filesystem_free_bytes{mountpoint="/"} /node_filesystem_size_bytes{mountpoint="/"}

1.2、查看根分区已用百分比

100 - node_filesystem_free_bytes{mountpoint="/"} /node_filesystem_size_bytes{mountpoint="/"} * 100

(1-(node_filesystem_free_bytes{mountpoint="/"} /node_filesystem_size_bytes{mountpoint="/"})) * 100

1.3、磁盘利用率大于20% 小于 50%

20 < (node_filesystem_free_bytes{mountpoint="/"} /node_filesystem_size_bytes{mountpoint="/"}) * 100 <= 50

1.4、磁盘利用率大于20% 并且小于等于50%

(node_filesystem_free_bytes{mountpoint="/"} /node_filesystem_size_bytes{mountpoint="/"}) * 100 > 20 and (node_filesystem_free_bytes{mountpoint="/"} /node_filesystem_size_bytes{mountpoint="/"}) * 100 <= 50 

1.5、磁盘利用率大于20% 或者小于等于50%

(node_filesystem_free_bytes{mountpoint="/"} /node_filesystem_size_bytes{mountpoint="/"}) * 100 > 20 or (node_filesystem_free_bytes{mountpoint="/"} /node_filesystem_size_bytes{mountpoint="/"}) * 100 <= 50 

1.6、磁盘利用率小于30% 或者大于等于50%

(node_filesystem_free_bytes{mountpoint="/"} /node_filesystem_size_bytes{mountpoint="/"}) * 100 < 30 or (node_filesystem_free_bytes{mountpoint="/"} /node_filesystem_size_bytes{mountpoint="/"}) * 100 >= 50 

二、PromQL 常用函数
2.1、探测网站访问大于3s

sum(probe_http_duration_seconds)> 3

2.2、集群磁盘总剩余量

sum(node_filesystem_free_bytes{mountpoint="/"})/(1024^3)

2.3、状态码200 请求总数

sum(prometheus_sd_kubernetes_http_request_total) by (status_code)
{status_code="200"}
6332
{status_code="<error>"}

2.4、统计请求头和请求状态码数量

sum(prometheus_http_requests_total) by (code,handler)

2.5、统计访问前三的状态码总数

topk(3,sum(prometheus_http_requests_total) by (code,handler))

2.6、统计访问最后三位的状态码总数

bottomk(3,sum(prometheus_http_requests_total) by (code,handler))

2.7、ceil取整（四舍五入）

ceil(node_filesystem_free_bytes/1024^3)

2.8、取整数（直接去掉整数）

floor(node_filesystem_free_bytes/1024^3)

2.9、从小到大排序

sort(node_filesystem_free_bytes/1024^3)

2.10-、从大到小排序

sort_desc(node_filesystem_free_bytes/1024^3)

2.11、计算数据在一段时间内的增长数

increase(prometheus_http_requests_total{handler="/api/v1/metadata",service="prometheus-k8s",code="200"}[1h])

increase(kubelet_http_requests_total{metrics_path="/metrics"}[10m]) /600
{endpoint="https-metrics", instance="192.168.122.160:10250", job="kubelet", long_running="false", method="GET", metrics_path="/metrics", namespace="kube-system", node="master", path="metrics", server_type="readwrite", service="kubelet"}
0.03333333333333333

rate(kubelet_http_requests_total{metrics_path="/metrics"}[10m])
{endpoint="https-metrics", instance="192.168.122.160:10250", job="kubelet", long_running="false", method="GET", metrics_path="/metrics", namespace="kube-system", node="master", path="metrics", server_type="readwrite", service="kubelet"}
0.03333333333333333
irate(kubelet_http_requests_total{metrics_path="/metrics"}[10m])