Prometheus 學習筆記-1

Devops > Prometheus Prometheus

發布時間 : 2023-11-07 15:31

字數:1.2k 觀看數 :

評論:

前言
介紹
作業環境
Metrics 類型
設定說明
操作流程
PromQL 基礎查詢
數值判斷
範圍查詢
聚合(Aggregation) 查詢
參考文件

前言

本篇記錄如何應用 Prometheus

紀錄學習 HiSKIO 課程 Prometheus + Grafana 監控和警報系統從入門到進階

介紹

Prometheus 官網

Prometheus 是一套系統監控框架，可用來蒐集各項 Metric
因 Prometheus UI 較陽春，故可將 Prometheus 當作資料蒐集器，提供資料給 Garfana 呈現
Prometheus 相關功能可參考 Prometheus - 官方網站

作業環境

Windows 10 Professional (22H2)
Docker Desktop
Docker Compose

課程範例原始碼

Metrics 類型

Prometheus 有四種 Metrics 類型

Counter = 計數器，數值只增不減
Gauge = 與 Counter 不同，Gauge 隨著時間變化可以遞增也可以遞減
Histogram = 柱狀圖，顯示數值區間數量，用以顯示分布情況
Summary = 彙總，顯示數值中位數

設定說明

課程範例原始碼資料夾樹狀結構如下

.
├── docker-compose.yml
├── telegraf/
│   ├── telegraf.d
│   │   ├── amazon.conf
│   │   └── github.conf
│   └── telegraf.conf
└── prometheus/
    └── prometheus.yml

docker-compose.yml

version: "3.0"

services:
  telegraf:
    image: telegraf:1.16.0
    restart: always
    container_name: telegraf
    hostname: telegraf
    ports:
      - 9273:9273
    volumes:
      - ./telegraf/telegraf.conf:/etc/telegraf/telegraf.conf
      - ./telegraf/telegraf.d/:/etc/telegraf/telegraf.d/
    command: telegraf --config /etc/telegraf/telegraf.conf --config-directory /etc/telegraf/telegraf.d

  prometheus:
    image: prom/prometheus:latest
    restart: always
    container_name: prometheus
    hostname: prometheus
    ports:
      - 9090:9090
    volumes:
      - ./prometheus/prometheus.yml:/etc/prometheus/prometheus.yml
    command: --config.file=/etc/prometheus/prometheus.yml

telegraf 部分參考 w4560000 - Telegraf 學習筆記-1

prometheus.yml

global:
  scrape_interval: 5s

scrape_configs:
  - job_name: "telegraf"
    static_configs:
      - targets: ["telegraf:9273"]

scrape_interval 每五秒撈取 Metrics 一次
scrape_configs 欲拉取 Metrics 的服務設定

操作流程

docker-compose up -d

# 確認一下服務
docker-compose ps

# 輸出
NAME                IMAGE                    COMMAND                  SERVICE             CREATED             STATUS              PORTS
prometheus          prom/prometheus:latest   "/bin/prometheus --c…"   prometheus          37 minutes ago      Up 37 minutes       0.0.0.0:9090->9090/tcp
telegraf            telegraf:1.16.0          "/entrypoint.sh tele…"   telegraf            37 minutes ago      Up 37 minutes       8092/udp, 8125/udp, 8094/tcp, 0.0.0.0:9273->9273/tcp

查看 telegraf metrics 開啟 http://localhost:9273/metrics
查看 prometheus 開啟 http://localhost:9090/

PromQL 基礎查詢

目前資料
搜尋 ping_average_response_ms 顯示 ping_average_response_ms Metrics 數據


ping_average_response_ms{environment=”LeoTest”, host=”telegraf”, instance=”telegraf:9273”, job=”telegraf”, pingVersion=”1”, service_name=”github”, url=”github.com”}	38.205
ping_average_response_ms{environment=”LeoTest”, host=”telegraf”, instance=”telegraf:9273”, job=”telegraf”, pingVersion=”2”, service_name=”amazon”, url=”amazon.cn”}	312.184
ping_average_response_ms{environment=”LeoTest”, host=”telegraf”, instance=”telegraf:9273”, job=”telegraf”, pingVersion=”2”, service_name=”amazon”, url=”amazon.com”}	221.355
ping_average_response_ms{environment=”LeoTest”, host=”telegraf”, instance=”telegraf:9273”, job=”telegraf”, pingVersion=”2”, service_name=”amazon”, url=”amazon.de”}	267.097
ping_average_response_ms{environment=”LeoTest”, host=”telegraf”, instance=”telegraf:9273”, job=”telegraf”, pingVersion=”2”, service_name=”amazon”, url=”amazon.jp”}	5.177

操作符號:

= equal
搜尋 ping_average_response_ms{service_name="github"} 顯示 service_name 為 github 的 Metrics 數據


ping_average_response_ms{environment=”LeoTest”, host=”telegraf”, instance=”telegraf:9273”, job=”telegraf”, pingVersion=”1”, service_name=”github”, url=”github.com”}	38.476

!= not equal
搜尋 ping_average_response_ms{service_name!="github"} 顯示 service_name 不為 github 的 Metrics 數據


ping_average_response_ms{environment=”LeoTest”, host=”telegraf”, instance=”telegraf:9273”, job=”telegraf”, pingVersion=”2”, service_name=”amazon”, url=”amazon.cn”}	311.721
ping_average_response_ms{environment=”LeoTest”, host=”telegraf”, instance=”telegraf:9273”, job=”telegraf”, pingVersion=”2”, service_name=”amazon”, url=”amazon.com”}	222.662
ping_average_response_ms{environment=”LeoTest”, host=”telegraf”, instance=”telegraf:9273”, job=”telegraf”, pingVersion=”2”, service_name=”amazon”, url=”amazon.de”}	265.805
ping_average_response_ms{environment=”LeoTest”, host=”telegraf”, instance=”telegraf:9273”, job=”telegraf”, pingVersion=”2”, service_name=”amazon”, url=”amazon.jp”}	5.479

搜尋 ping_average_response_ms{service_name!="github", url="amazon.cn"} 顯示 service_name 不為 github 以及 url 為 amazon.cn 的 Metrics 數據


ping_average_response_ms{environment=”LeoTest”, host=”telegraf”, instance=”telegraf:9273”, job=”telegraf”, pingVersion=”2”, service_name=”amazon”, url=”amazon.cn”}	311.875

=~ match regex

搜尋 ping_average_response_ms{url=~"^amazon.*"} 顯示符合 url 開頭為 amazon. 正則表達式的 Metrics 數據


ping_average_response_ms{environment=”LeoTest”, host=”telegraf”, instance=”telegraf:9273”, job=”telegraf”, pingVersion=”2”, service_name=”amazon”, url=”amazon.cn”}	312.087
ping_average_response_ms{environment=”LeoTest”, host=”telegraf”, instance=”telegraf:9273”, job=”telegraf”, pingVersion=”2”, service_name=”amazon”, url=”amazon.com”}	222.339
ping_average_response_ms{environment=”LeoTest”, host=”telegraf”, instance=”telegraf:9273”, job=”telegraf”, pingVersion=”2”, service_name=”amazon”, url=”amazon.de”}	264.802
ping_average_response_ms{environment=”LeoTest”, host=”telegraf”, instance=”telegraf:9273”, job=”telegraf”, pingVersion=”2”, service_name=”amazon”, url=”amazon.jp”}	5.676

!~ doesn’s match regex

搜尋 ping_average_response_ms{url!~"^amazon.*"} 顯示不符合 url 開頭為 amazon. 正則表達式的 Metrics 數據


ping_average_response_ms{environment=”LeoTest”, host=”telegraf”, instance=”telegraf:9273”, job=”telegraf”, pingVersion=”1”, service_name=”github”, url=”github.com”}	38.009

數值判斷

搜尋 ping_average_response_ms{service_name!="github"} > 100 顯示數值超過 100 的 Metrics 數據數據


ping_average_response_ms{environment=”LeoTest”, host=”telegraf”, instance=”telegraf:9273”, job=”telegraf”, pingVersion=”2”, service_name=”amazon”, url=”amazon.cn”}	312.05
ping_average_response_ms{environment=”LeoTest”, host=”telegraf”, instance=”telegraf:9273”, job=”telegraf”, pingVersion=”2”, service_name=”amazon”, url=”amazon.com”}	222.711
ping_average_response_ms{environment=”LeoTest”, host=”telegraf”, instance=”telegraf:9273”, job=”telegraf”, pingVersion=”2”, service_name=”amazon”, url=”amazon.de”}	265.618

範圍查詢

搜尋 ping_average_response_ms{service_name!="github"}[30s] 顯示近 30 秒內的 Metrics 數據


ping_average_response_ms{environment=”LeoTest”, host=”telegraf”, instance=”telegraf:9273”, job=”telegraf”, pingVersion=”2”, service_name=”amazon”, url=”amazon.cn”}	311.996 @1699346946 315.853 @1699346955 311.564 @1699346965
ping_average_response_ms{environment=”LeoTest”, host=”telegraf”, instance=”telegraf:9273”, job=”telegraf”, pingVersion=”2”, service_name=”amazon”, url=”amazon.com”}	223.442 @1699346955 223.2 @1699346965
ping_average_response_ms{environment=”LeoTest”, host=”telegraf”, instance=”telegraf:9273”, job=”telegraf”, pingVersion=”2”, service_name=”amazon”, url=”amazon.de”}	274.674 @1699346955 272.112 @1699346965
ping_average_response_ms{environment=”LeoTest”, host=”telegraf”, instance=”telegraf:9273”, job=”telegraf”, pingVersion=”2”, service_name=”amazon”, url=”amazon.jp”}	8.695 @1699346955 5.744 @1699346965

聚合(Aggregation) 查詢

搜尋 sum(ping_average_response_ms) by (service_name) 顯示以 service_name 為分群加總 ping_average_response_ms 的 Metrics 數據


{service_name=”github”}	37.998
{service_name=”amazon”}	793.574

搜尋 topk(3, ping_average_response_ms) 顯示最大 ping_average_response_ms 的 Metrics 數據的前 3 筆資料


ping_average_response_ms{environment=”LeoTest”, host=”telegraf”, instance=”telegraf:9273”, job=”telegraf”, pingVersion=”2”, service_name=”amazon”, url=”amazon.cn”}	312.218
ping_average_response_ms{environment=”LeoTest”, host=”telegraf”, instance=”telegraf:9273”, job=”telegraf”, pingVersion=”2”, service_name=”amazon”, url=”amazon.de”}	272.3
ping_average_response_ms{environment=”LeoTest”, host=”telegraf”, instance=”telegraf:9273”, job=”telegraf”, pingVersion=”2”, service_name=”amazon”, url=”amazon.com”}	221.275

其餘 Function 參考 Prometheus Functions

參考文件

HiSKIO 課程 Prometheus + Grafana 監控和警報系統從入門到進階
 Prometheus 官網

轉載請註明來源，若有任何錯誤或表達不清楚的地方，歡迎在下方評論區留言，也可以來信至 leozheng0621@gmail.com
如果文章對您有幫助，歡迎斗內(donate)，請我喝杯咖啡