Monitoring¶

Telegraf¶

将 Telegraf 暴露为 Prometheus 数据格式¶

telegraf.conf

[[inputs.cpu]]
  percpu = true
  totalcpu = true
  collect_cpu_time = false
  report_active = false
  core_tags = false

[[outputs.file]]
  files = ["stdout"]

[[outputs.prometheus_client]]
  listen = ":9273"

telegraf --config telegraf.conf
curl http://localhost:9273/metrics

Prometheus 远端存储到 InfluxDB¶

telegraf.conf

[[inputs.http_listener_v2]]
  service_address = ":1234"
  paths = ["/receive"]
  data_format = "prometheusremotewrite"

[[outputs.influxdb_v2]]
  urls = ["http://192.192.192.6:8086"]
  token = "cL3fFaCMjyEuzU7Bs0uXDApAovt934Y5oE22keyVFHQS1_ELriSLQ6Jneg-bJTVPpHXPddmGENbS6YwwuV9lrw=="
  organization = "admin"
  bucket = "prometheus"

prometheus.yml

global:
  scrape_timeout: 10s
  scrape_interval: 10s
  evaluation_interval: 10s

scrape_configs:
  - job_name: 'prometheus'
    static_configs:
      - targets: ['192.192.192.6:9090']
  - job_name: 'pushgateway'
    static_configs:
      - targets: ['192.192.192.6:9091']
  - job_name: 'node-exporter'
    static_configs:
      - targets: ['192.192.192.6:9100']
  - job_name: 'cadvisor'
    static_configs:
      - targets: ['192.192.192.6:8080']

remote_write:
  - url: 'http://192.192.192.6:1234/receive'