prometheus常用exporter

一、node-exporter

node_exporter：用于监控Linux系统的指标采集器。

未在k8s集群内的linux机器监控

GitHub - prometheus/node_exporter: Exporter for machine metrics

常用指标：

•CPU
• 内存
• 硬盘
• 网络流量
• 文件描述符
• 系统负载
• 系统服务
数据接口：http://IP:9100/metrics

第一种方式：linux主机安装

#x86_64主机下载此客户端
wget --no-check-certificate http://foreman.chinamcloud.com:8080/source/node_exporter-1.6.1.linux-amd64.tar.gz
#安装
useradd prometheus -s /sbin/nologin
#x86_64主机
tar zxvf node_exporter-1.6.1.linux-amd64.tar.gz -C /tmp/
#x86_64主机
mv /tmp/node_exporter-1.6.1.linux-amd64 /usr/local/node_exporter
#目录授权
chown prometheus:prometheus -R /usr/local/node_exporter
#封装service
tee /etc/systemd/system/node-exporter.service <<-'EOF'
[Unit]
Description=Prometheus Node Exporter
After=network.target
[Service]
ExecStart=/usr/local/node_exporter/node_exporter
User=prometheus
[Install]
WantedBy=multi-user.target
EOF
systemctl daemon-reload
systemctl enable node-exporter
systemctl start node-exporter

验证监控，访问http://ip:9100/metrics

第二种方式：docker容器启动

docker run -d --name node-exporter --restart=always -p 9100:9100 -v "/proc:/host/proc:ro" -v "/sys:/host/sys:ro" -v "/:/rootfs:ro" prom/node-exporter

配置Prometheus，接入windows的metrics

配置映射里Prometheus.yaml新加一个job

- job_name: linux-exporter
  static_configs:
  - targets:
    - 192.168.1.1:9100
    - 192.168.1.2:9100
    - 192.168.1.3:9100
    - x.x.x.x:9100

将linux的ip加入到配置文件里

常用指标

主机内存使用率 > 90%

(1 - (node_memory_MemAvailable_bytes{environment=~"项目标志"} / (node_memory_MemTotal_bytes{environment=~"项目标志"})))* 100  >90

主机 CPU 使用率> 90%

(1 - avg(rate(node_cpu_seconds_total{environment=~"项目标志",mode="idle"})) by (instance))*100 >90

主机NTP时间差[5m] > 5s

node_timex_tai_offset_seconds{environment=~"项目标志"}>3

主机磁盘使用率>80%

(node_filesystem_size_bytes{environment=~"项目标志",fstype=~"ext.*|xfs",mountpoint !~".*pod.*"}-node_filesystem_free_bytes{environment=~"项目标志",fstype=~"ext.*|xfs",mountpoint !~".*pod.*"}) *100/(node_filesystem_avail_bytes {environment=~"项目标志",fstype=~"ext.*|xfs",mountpoint !~".*pod.*"}+(node_filesystem_size_bytes{environment=~"项目标志",fstype=~"ext.*|xfs",mountpoint !~".*pod.*"}-node_filesystem_free_bytes{environment=~"项目标志",fstype=~"ext.*|xfs",mountpoint !~".*pod.*"})) >80

主机Inode 使用率[5m] > 80%

(1-node_filesystem_files_free{environment=~"项目标志",fstype=~"ext.?|xfs"} / node_filesystem_files{environment=~"项目标志",fstype=~"ext.?|xfs"})*100 >80

二、windows-exporter

GitHub - prometheus-community/windows_exporter: Prometheus exporter for Windows machineswindows-exporter安装部署

1、安装包下载

windows_exporter-0.22.0-amd64.msi

2、直接双击运行该msi程序即可正常安装

安装完成后，打开任务管理器，会看到里面有个windows-exporter的服务

3、使用127.0.0.1:9182端口，可看到对应的metrics

配置Prometheus，接入windows的metrics

配置映射里Prometheus.yaml新加一个job

- job_name: windows-exporter
  static_configs:
  - targets:
    - 192.168.1.1:9182
    - 192.168.1.2:9182
    - 192.168.1.3:9182
    - x.x.x.x:9182

将windows的ip加入到配置文件里

grafana界面配置：

左侧菜单栏，找到DataSource，进行添加数据源，选择Prometheus,然后添加prometheus服务的url地址

导入模版10467

GitHub - prometheus-community/windows_exporter: Prometheus exporter for Windows machines

注意

默认收集只收集了cpu，cs，logical_disk，net，os,service,system,textfile相关的指标，若要开启其他的，需配置—collectors.enabled

1.新增C:\Program Files\windows_exporter\config.yml文件

collectors:
  enabled: "[defaults],process,cpu_info,memory,remote_fx,tcp"
collector:
  service:
    services-where: "Name='windows_exporter'"
log:
  level: warn

2.修改windows启动参数

sc config   windows_exporter  binPath= "\"C:\Program Files\windows_exporter\windows_exporter.exe\" --config.file=\"C:\Program Files\windows_exporter\config.yml\" "

3.重启windows_exporter服务

sc stop windows_exporter
sc start windows_exporter

二、powershell脚本一键安装windows_exporter，`windows_exporter.ps1`

Set-Location -Path $Env:TEMP
(New-Object Net.WebClient).DownloadFile('http://foreman.chinamcloud.com:8080/source/windows_exporter-0.23.1-amd64.msi', $ExecutionContext.SessionState.Path.GetUnresolvedProviderPathFromPSPath('./windows_exporter-0.23.1-amd64.msi'))
Start-Process  ./windows_exporter-0.23.1-amd64.msi
Add-Content "C:\Program Files\windows_exporter\config.yml" "collectors:"
Add-Content "C:\Program Files\windows_exporter\config.yml" "  enabled: ""[defaults],process,cpu_info,memory,remote_fx,tcp"" "
Add-Content "C:\Program Files\windows_exporter\config.yml" "collector:"
Add-Content "C:\Program Files\windows_exporter\config.yml" "  service:"
Add-Content "C:\Program Files\windows_exporter\config.yml" "    services-where: ""Name='windows_exporter'"""
Add-Content "C:\Program Files\windows_exporter\config.yml" "log:"
Add-Content "C:\Program Files\windows_exporter\config.yml" "  level: warn"
sleep 30
sc.exe config "windows_exporter" binpath= """""""C:\Program Files\windows_exporter\windows_exporter.exe"""""" --config.file=""""""C:\Program Files\windows_exporter\config.yml"""""""
sc.exe config "windows_exporter" binpath= """""""C:\Program Files\windows_exporter\windows_exporter.exe"""""" --config.file=""""""C:\Program Files\windows_exporter\config.yml"""""""
sc.exe stop windows_exporter
sc.exe start  windows_exporter
sleep 10

常用指标

主机CPU使用率
100 - avg(irate(windows_cpu_time_total{job=~"$job",instance=~"$instance",mode="idle"}[5m]))*100

主机内存使用率
100 - 100 * (windows_os_physical_memory_free_bytes{job=~"$job"} / windows_cs_physical_memory_bytes{job=~"$job"})

主机磁盘使用率
100-(windows_logical_disk_free_bytes/windows_logical_disk_size_bytes)*100

主机mpc相关进程句柄数
windows_process_handles{environment=~"$environment",job="windows-exporter",process=~'MPC.*|cloudia.*|APPBaseInTT.*|AppBaseInTT.*'}

三、blackbox-exporter

GitHub - prometheus/blackbox_exporter: Blackbox prober exporter

Blackbox Exporter是Prometheus社区提供的官方黑盒监控解决方案，其允许用户通过：HTTP、HTTPS、DNS、TCP以及ICMP的方式对网络进行探测。

应用场景

HTTP 测试
定义 Request Header 信息判断 Http status Http Respones Header Http Body 内容
TCP 测试
业务组件端口状态监听应用层协议定义与监听
ICMP 测试
主机探活机制
POST 测试
接口联通性
SSL 证书过期时间
运行Blackbox Exporter时，需要用户提供探针的配置信息，这些配置信息可能是一些自定义的HTTP头信息，也可能是探测时需要的一些TSL配置，也可能是探针本身的验证行为。在Blackbox Exporter每一个探针配置称为一个module，并且以YAML配置文件的形式提供给Blackbox Exporter。每一个module主要包含以下配置内容，包括探针类型（prober）、验证访问超时时间（timeout）、以及当前探针的具体配置项：
探针类型：http、 tcp、 dns、 icmp.
prober: <prober_string>
超时时间
[ timeout: <duration> ]
探针的详细配置，最多只能配置其中的一个
[ http: <http_probe> ]
[ tcp: <tcp_probe> ]
[ dns: <dns_probe> ]
[ icmp: <icmp_probe> ]

安装部署blackbox-exporter

1.rancher内导入下面的blackbox-exporter.yml文件到Prometheus

apiVersion: apps/v1
kind: Deployment
metadata:
  generation: 1
  labels:
    cattle.io/creator: norman
    workload.user.cattle.io/workloadselector: deployment-prometheus-blackbox-exporter
  name: blackbox-exporter
  namespace: prometheus
spec:
  progressDeadlineSeconds: 600
  replicas: 1
  revisionHistoryLimit: 10
  selector:
    matchLabels:
      workload.user.cattle.io/workloadselector: deployment-prometheus-blackbox-exporter
  strategy:
    rollingUpdate:
      maxSurge: 25%
      maxUnavailable: 25%
    type: RollingUpdate
  template:
    metadata:
      annotations:
        field.cattle.io/ports: '[[{"containerPort":9115,"dnsName":"blackbox-exporter","hostPort":0,"kind":"ClusterIP","name":"blackbox-port","protocol":"TCP","sourcePort":0}]]'
      creationTimestamp: null
      labels:
        app.kubernetes.io/name: blackbox
        workload.user.cattle.io/workloadselector: deployment-prometheus-blackbox-exporter
    spec:
      containers:
      - args:
        - --config.file=/etc/blackbox_exporter/blackbox.yml
        - --log.level=info
        - --web.listen-address=:9115
        image: prom/blackbox-exporter:v0.16.0
        imagePullPolicy: Always
        livenessProbe:
          failureThreshold: 3
          initialDelaySeconds: 10
          periodSeconds: 2
          successThreshold: 1
          tcpSocket:
            port: 9115
          timeoutSeconds: 2
        name: blackbox-exporter
        ports:
        - containerPort: 9115
          name: blackbox-port
          protocol: TCP
        readinessProbe:
          failureThreshold: 3
          initialDelaySeconds: 10
          periodSeconds: 2
          successThreshold: 2
          tcpSocket:
            port: 9115
          timeoutSeconds: 2
        resources: {}
        securityContext:
          allowPrivilegeEscalation: false
          capabilities: {}
          privileged: false
          readOnlyRootFilesystem: false
          runAsNonRoot: false
        stdin: true
        terminationMessagePath: /dev/termination-log
        terminationMessagePolicy: File
        tty: true
        volumeMounts:
        - mountPath: /etc/blackbox_exporter
          name: vol1
      dnsPolicy: ClusterFirst
      restartPolicy: Always
      schedulerName: default-scheduler
      securityContext: {}
      terminationGracePeriodSeconds: 30
      volumes:
      - configMap:
          defaultMode: 420
          name: blackbox-exporter
          optional: false
        name: vol1

2.rancher内导入下面yml，新增服务发现blackbox-exporter

apiVersion: v1
kind: Service
metadata:
  annotations:
  labels:
    app.kubernetes.io/name: blackbox
  name: blackbox-exporter
  namespace: prometheus
  selfLink: /api/v1/namespaces/prometheus/services/blackbox-exporter
spec:
  ports:
  - name: balckbox
    port: 9115
    protocol: TCP
    targetPort: 9115
  selector:
    app.kubernetes.io/name: blackbox
  sessionAffinity: None
  type: ClusterIP

Prometheus下配置映射新增blackbox-exporter

blackbox.yml

modules:
  http_2xx:
    prober: http
    timeout: 5s
    http:
      valid_http_versions: ["HTTP/1.1", "HTTP/2"]
      valid_status_codes: [200,301,302,403,405]
      method: GET
      preferred_ip_protocol: "ipv4"
  https_2xx:
    prober: https
    timeout: 5s
    http:
      valid_http_versions: ["HTTP/1.1", "HTTP/2"]
      valid_status_codes: [200,301,302,403,405]
      method: POST
      preferred_ip_protocol: "ipv4"
  tcp_connect:
    prober: tcp
    timeout: 2s
  http_403:
    prober: http
    timeout: 2s
    http:
      valid_http_versions: ["HTTP/1.1", "HTTP/2"]
      valid_status_codes: [403,405]
      method: GET
      preferred_ip_protocol: "ipv4"
  https_403:
    prober: https
    timeout: 2s
    http:
      valid_http_versions: ["HTTP/1.1", "HTTP/2"]
      valid_status_codes: [403,405]
      method: GET
      preferred_ip_protocol: "ipv4"
  icmp:
    prober: icmp/etc/blackbox_exporter

prometheus配置映射内新增job

如配置证书的监控，修改Prometheus内配置映射的prometheus.yml

- job_name: "blackbox_ssl"
  metrics_path: /probe
  params:
    module: [http_2xx]
  static_configs:
  - targets:
    - https://login.demo.chinamcloud.cn
  relabel_configs:
  - source_labels: [__address__]
    target_label: instance
  - source_labels: [__address__]
    target_label: __param_target
  - target_label: __address__
    replacement: blackbox-exporter:9115

如配置TCP端口检测，修改Prometheus内配置映射的prometheus.yml

- job_name: "tcp-port-check"
  scrape_interval: 30s
  metrics_path: /probe
  params:
    module: [tcp_connect]
  static_configs:
  - targets:
    - mpc.server:8088
    - mysql.server:3306
    - redis.server:6379
    - kafka.kafka:9092
    labels: 
      server: 'tcp-port-check'
  relabel_configs:
  - source_labels: [__address__]
    target_label: instance
  - source_labels: [__address__]
    target_label: __param_target
  - target_label: __address__
    replacement: blackbox-exporter:9115

POST 测试，监听业务接口地址，用来判断接口是否在线

- job_name: 'blackbox_http_2xx_post'
  scrape_interval: 10s
  metrics_path: /probe
  params:
    module: [http_post_2xx_query]
  static_configs:
  - targets:
    - https://xx.xxx.com/api/xx/xx/xx/query.action
    labels:
      group: 'Interface monitoring'
  relabel_configs:
  - source_labels: [__address__]
    target_label: __param_target
  - source_labels: [__param_target]
    target_label: instance
  - target_label: __address__
    replacement: blackbox-exporter:9115

常用指标

probe_success == 0 ##联通性异常
probe_success == 1 ##联通性正常

证书过期时间低于30天
(probe_ssl_earliest_cert_expiry-time()) / 86400 <30

http状态码
probe_http_status_code

请求耗时 
probe_duration_seconds{job=~"blackbox_ssl"}
http请求耗时
probe_http_duration_seconds

四、elasticsearch_exporter

GitHub - prometheus-community/elasticsearch_exporter: Elasticsearch stats exporter for Prometheus

elasticsearch_exporter安装部署

1.在k8s内导入下面这个yml文件，创建elasticsearch-exporter

http://elasticsearch.server:9200换成项目上的es地址

若es配置了账号密码，http://账号:密码@elasticsearch.server:9200

apiVersion: apps/v1
kind: Deployment
metadata:
  name: elasticsearch-exporter
  namespace: prometheus
spec:
  selector:
    matchLabels:
      workload.user.cattle.io/workloadselector: deployment-prometheus-elasticsearch-exporter
  strategy:
    rollingUpdate:
      maxSurge: 1
      maxUnavailable: 0
    type: RollingUpdate
  template:
    metadata:
      labels:
        app.kubernetes.io/name: elasticsearch-exporter
        workload.user.cattle.io/workloadselector: deployment-prometheus-elasticsearch-exporter
    spec:
      containers:
      - args:
        - --es.uri=http://elasticsearch.server:9200
        - --es.all
        - --es.indices
        - --es.indices_settings
        - --es.shards
        - --es.snapshots
        - --collector.clustersettings
        image: quay.io/prometheuscommunity/elasticsearch-exporter:v1.6.0
        imagePullPolicy: Always
        name: elasticsearch-exporter
        ports:
        - containerPort: 9114
          name: tcp
          protocol: TCP

2.导入下面yml文件，创建elasticsearch-exporter的服务发现

apiVersion: v1
kind: Service
metadata:
  name: elasticsearch-exporter
  namespace: prometheus
spec:
  clusterIP: None
  ports:
  - name: tcp
    port: 9114
    protocol: TCP
    targetPort: 9114
  selector:
    app.kubernetes.io/name: elasticsearch-exporter

3.prometheus的配置文件prometheus.yml内新增elasticsearch-exporter的job

- job_name: elasticsearch-exporter
  metrics_path: '/metrics'
  static_configs:
  - targets:
    - elasticsearch-exporter.prometheus:9114

4.验证该job状态是否正常，访问Prometheus

5.导入grafana内模板

常用告警指标说明

# ES集群状态YELLOW
elasticsearch_cluster_health_status{environment=~"csp.*",color="yellow"} =1
# ES集群状态RED
elasticsearch_cluster_health_status{environment=~"csp.*",color="red"} =1
# ES集群状态正常
elasticsearch_cluster_health_status{environment=~"csp.*",color="greeen"} =1
#ES总节点数
sum(elasticsearch_cluster_health_number_of_nodes{cluster=~"haihe3"})
#ES数据节点数
sum(elasticsearch_cluster_health_number_of_data_nodes{cluster=~"haihe3"})

五、kong接入prometheus

1.kong的dashboard内新加插件Prometheus

2. 修改Prometheus.yaml文件配置，新加kong的job监控

静态注册

- job_name: cmc-kong-monitoring
  metrics_path: "/metrics"
  static_configs:
  - targets:
    - kong.cmc:8001

动态注册：使用kubernetes_sd_configs自动发现kong的endpoints【如CSP环境】

#框架cmc下kong
- job_name: csp-cmc-kong    
  kubernetes_sd_configs:
  - role: endpoints
  scheme: http
  relabel_configs:
  - source_labels: [__meta_kubernetes_namespace, __meta_kubernetes_service_name,__meta_kubernetes_endpoint_port_name]
    regex: cmc;kong;http8001
    action: keep
#kong下的kong
- job_name: csp-kong-kong    
  kubernetes_sd_configs:
  - role: endpoints
  scheme: http
  relabel_configs:
  - source_labels: [__meta_kubernetes_namespace, __meta_kubernetes_service_name,__meta_kubernetes_endpoint_port_name]
    regex: kong;kong;8001to8001-kong
    action: keep
#云教kong kong下kongedu
- job_name: csp-kong-kongedu    
  kubernetes_sd_configs:
  - role: endpoints
  scheme: http
  relabel_configs:
  - source_labels: [__meta_kubernetes_namespace, __meta_kubernetes_service_name,__meta_kubernetes_endpoint_port_name]
    regex: kong;kongedu;8001tcp80012
    action: keep

grafana上导入kong的大屏

常用指标

名称                                  类型         单位           说明
kong_bandwidth                       counter      字节数bytes      kong中每个服务消耗的总带宽（字节）
kong_datastore_reachable             gauge          是否           kong对数据库是否可以访问（1：正常 2：不正常）
kong_http_status                     counter      服务数          kong中每个服务的状态码
kong_latency_count                     counter      调用次数         kong中每个服务的调用次数
kong_latency_sum                     counter      调用耗时ms     kong中每个服务的调用耗时
kong_latency_bucket                     histogram      服务数          kong中的服务在各个耗时区间的分布（类型设有：kong内部处理耗时 request请求耗时 ，upstream代理上游服务器耗时）
kong_nginx_http_current_connections     gauge          连接数          kong当前的连接数（状态设有：accepted、active、handled、reading、total、waiting、writing）
    total      客户端请求总数
    accepted    接受的客户端连接的总数
    handled       处理连接的总数。一般来说，除非达到一定的资源限制，否则参数值与接受值相同
    active      包括等待连接的活动客户端连接的当前数量
    reading      当前Kong正在读取请求头的连接数
    writing      将响应写入客户端的连接的当前数量
    waiting      等待请求的空闲客户端连接的当前数量

kong_nginx_metric_errors_total         counter      个数           kong prometheus插件错误的指标数

kubernetes_sd_configs内新增kong=csp灰度框架kong标签

- job_name: kong-cmc-endpoints    
  kubernetes_sd_configs:
  - role: endpoints
  scheme: http
  relabel_configs:
  - source_labels: [__meta_kubernetes_namespace, __meta_kubernetes_service_name,__meta_kubernetes_endpoint_port_name]
    regex: cmc;kong;http8001
    action: keep
  - action: replace
    replacement: csp灰度框架kong
    target_label: kong