漫谈 prometheus

以下应用是基于 Springboot 项目的监控,如果是 Springboot 2 版本,可以安装 mircrometer 组件来提供对 prometheus 的支持。

prometheus 官方地址: https://prometheus.io/

Can use 4 kinds of metrics

Prometheus 中提供了四种指标类型(参考:Prometheus 的指标类型),Counter、Timer、 以及直方图(Histogram)和摘要(Summary),后两者是最复杂和难以理解的.

例如Histogram 会在一段时间范围内对数据进行采样(通常是请求持续时间或响应大小等),并将其计入可配置的存储桶(bucket)。

Prometheus Commands Exampls

以下命令更适合与 Grafana 搭配使用。

Sub-Interface QPS (rate 10s, status=200)

rate(http_server_requests_seconds_count{application="PROJECT", uri!="/actuator/prometheus" ,status="200"}[1m]) 

Sub-Interface Latency P90 (rate 1min , seconds) ‘’’ histogram_quantile(0.9, sum(rate(http_server_requests_seconds_bucket{application=”PROJECT_NAME”, uri!=”/actuator/prometheus”}[1m])) by (uri, le)) ‘’’

Sub-Interface Latency P50 (rate 1min , seconds)

 histogram_quantile(0.5, sum(rate(http_server_requests_seconds_bucket{application="PROJECT_NAME", uri!="/actuator/prometheus"}[1m])) by ( uri, le)) 

TOP URL Visits

 topk(10, sum( http_server_requests_seconds_count{application="PROJECT_NAME",uri!="/actuator/prometheus"}) by (uri))

HTTP methods Count

  sum(http_server_requests_seconds_count {application="PROJECT_NAME", uri!="/actuator/prometheus"}) by (status)

Log Count By Level

topk(5,  sum(logback_events_total{application="PROJECT_NAME"}) by (level))
您如果觉得有用,觉得写得还不错,可以一元赞赏作者。
   
微信捐赠图片
   
支付宝捐赠图片