问题现象：
部署监控平台需要导入 alert-rule.yaml 中的告警规则，通过 kubectl apply -f alert-rule.yaml 可以正常导入进去，但是在prometheus alert界面看不到对应规则
原因：
prometheus 环境部署是通过 prometheus-operator 进行的部署，operator 会通过 rules 的标签进行匹配是否将规则注入到 prometheus 里面。alert-rules.yaml 的标签不符合 operator 要匹配的标签，所以注入不进去
$ kubectl get prometheus prometheus-operator-operat-prometheus -oyaml # 具体名字根据现场环境填写，格式 kubectl get prometheus <PROMETHEUS_NAME> -oyaml
···
spec:
ruleNamespaceSelector: {}
ruleSelector:
matchLabels: ## rules 规则需要匹配以下两个标签
app: prometheus-operator
release: prometheus-operator-operator
···
alert-rule.yaml
---
apiVersion: monitoring.coreos.com/v1
kind: PrometheusRule
metadata:
labels: #与此标签不匹配，所以注入不进去
app: prometheus-operator
release: prometheus-operator
name: alert-node-rules
namespace: default
spec:
groups:
- name: node-rules
rules:
- alert: 主机内存使用率超过80%告警
expr: sum(100- ((node_memory_MemAvailable_bytes * 100 ) / node_memory_MemTotal_bytes )) by (instance,job) > 80
for: 5m
labels:
severity: "warning"
annotations:
summary: "告警主机："
address: "告警地址："
description: 主机 内存使用率超过80%，当前使用率为 %。
---
apiVersion: monitoring.coreos.com/v1
kind: PrometheusRule
metadata:
labels: #与此标签不匹配，所以注入不进去
app: prometheus-operator
release: prometheus-operator
name: alert-k8s-rules
namespace: default
spec:
groups:
- name: k8s-rules
rules:
- alert: KubeApiserver存活监控告警
expr: absent(up{job="apiserver"} == 1)
for: 1m
labels:
severity: "critical"
annotations:
description: KubeApiserver实例 从prometheus监控目标中失联
解决：
修改 alert-rule.yaml 的标签，使之匹配
