grafana agent metrics 集成victoriametrics 试用

以下是grafana agent metrics 集成victoriametrics 的一个简单试用，主要是metrics 的，对于log 以及trace 的后边会介绍

环境准备

• docker-compose 文件

version:  "3"

services: 

  postgres:

    image: postgres:12

    volumes:

    - ./db_data:/var/lib/postgresql/data

    ports:

      - 5432:5432

    environment:

      POSTGRES_PASSWORD: postgrespassword

  agent:

    image: grafana/agent:v0.28.1

    ports:

      - 12345:12345

    entrypoint:

      - /bin/agent

      - -config.file=/etc/agent-config/agent.yaml

      - -metrics.wal-directory=/tmp/agent/wal

      - -enable-features=integrations-next

      - -config.expand-env

      - -config.enable-read-api

    volumes:

      - ./agent.yaml:/etc/agent-config/agent.yaml

      - ./agent-data:/etc/agent/data

  prometheus:

    image: prom/prometheus

    ports:

      - 9090:9090

    volumes:

      - ./promdata:/prometheus

      - ./prometheus.yml:/etc/prometheus/prometheus.yml

    command:

      - '--config.file=/etc/prometheus/prometheus.yml'

      - '--storage.tsdb.path=/prometheus'

  victoriametrics:

    image: victoriametrics/victoria-metrics

    volumes:

      - ./victoriametrics:/victoriametrics

    ports:

      - 8428:8428

    command: 

      - '-storageDataPath=/victoriametrics'

      - '-retentionPeriod=1'

  grafana:

    image: grafana/grafana

    ports:

      - 3000:3000

• 配置
  prometheus.yml

 

global:

  scrape_interval:     1s

  evaluation_interval: 1s

 

remote_write:

  - url: "http://victoriametrics:8428/api/v1/write"

 

scrape_configs:

  - job_name: 'prometheus'

    static_configs:

      - targets: ['prometheus:9090']

 

  - job_name: 'victoriametrics'

    static_configs:

      - targets: ['victoriametrics:8428']

agent.yaml
开启了node exporter 以及process exporter，同时将metrics 写到victoriametrics中

 

# This file serves as an example agent configuration to interact with the

# docker compose environment.

#

# You should pass the following command line flags when running the agent

# locally and using this file:

#

#   -enable-features=integrations-next -config.expand-env -config.enable-read-api

#

# -enable-features=integrations-next is required as the file is configured for

# the integrations revamp.

#

# -config.expand-env is required to expand environment variables. Environment

# variables are used when running the agent inside of docker-compose to connect

# to the other services. When running the agent externally, the expressions will

# default to the appropriate values of the exposed ports.

#

# -config.enable-read-api is optional, but allows you to invoke the /-/config

# endpoint to examine the generated config.

 

server:

  log_level: debug

 

metrics:

  global:

    scrape_interval: 10s

    remote_write:

    - url: http://victoriametrics:8428/api/v1/write

  configs:

  - name: default

    scrape_configs:

    - job_name: avalanche

      static_configs:

      - targets: ['${AVALANCHE_HOST:-localhost:9001}']

#

# Integrations

#

# Uncomment individual integrations below to enable them. Some integrations are

# enabled by default.

#

 

integrations:

  node_exporter:  {}

  ## process

  process:

    process_names:

      - name: "{{.Comm}}"

        cmdline:

        - '.+'

  ## mysql (requires docker-compose mysql profile)

  # mysql_configs:

  # - data_source_name: root@(${MYSQL_HOST:-localhost:3306})/

 

  ## postgres (requires docker-compose postgres profile)

  postgres_configs:

  - data_source_names:

    - postgresql://postgres:postgrespassword@postgres:5432/postgres?sslmode=disable

 

  ## redis (requires docker-compose redis profile)

  # redis_configs:

  # - redis_addr: ${REDIS_HOST:-localhost:6379}

 

  ## dnsmasq (requires docker-compose dnsmasq profile)

  # dnsmasq_configs:

  # - dnsmasq_address: ${DNSMASQ_HOST:-localhost:30053}

  #   leases_path: /tmp/dnsmasq-leases/dnsmasq.leases

 

  ## memcached (requires docker-compose memcached profile)

  # memcached_configs:

  # - memcached_address: ${MEMCACHED_HOST:-localhost:11211}

  #   timeout: 10s

 

  ## statsd

  # statsd: {}

 

  ## consul (requires docker-compose consul profile)

  # consul_configs:

  # - server: http://${CONSUL_HOST:-localhost:8500}

 

  ## elasticsearch (requires docker-compose elasticsearch profile)

  # elasticsearch_configs:

  # - address: http://${ELASTICSEARCH_HOST:-localhost:9200}

 

  ## kafka (requires docker-compose kafka profile)

  # kafka_configs:

  # - kafka_uris: [${KAFKA_HOST:-localhost:9093}]

 

  ## github (requires docker-compose github profile)

  # github_configs:

  # - repositories:

  #   - grafana/agent

 

  ## mongodb (requires docker-compose mongodb profile)

  # mongodb_configs:

  # - mongodb_uri: mongodb://${MONGODB_HOST:-mongodb:27017}

  #   relabel_configs:

  #   - source_labels: [__address__]

  #     target_label: service_name

  #     replacement: 'mongodb'

  #   - source_labels: [__address__]

  #     target_label: mongodb_cluster

  #     replacement: 'mongodb-cluster'

 

  ## cadvisor

  # cadvisor:

  #   disabled_metrics:

  #   - disk

  #   enabled_metrics:

  #   - percpu

启动&效果

• 启动

docker-compose up -d

• 效果

prometheus

victoriametrics 信息

grafana 效果（需要添加victoriametrics prometheus 数据源以及配置dashboard，dashboard 配置了node以及process exporter 的）

说明

利用grafana agent all-in-one 的设计模式，我们只需要通过配置就可以很好的获取主机的metrics 可以简化不少exporter的安装，同时利用victoriametrics 多租户的能力可以很好的支持数据隔离提供不同环境的metrics 安全管理，实际上我以前我写过关于vmagent 与grafana agent 的对比，可以参考

参考资料

https://grafana.com/docs/agent/latest/configuration/metrics-config/
https://grafana.com/docs/agent/latest/configuration/logs-config/
https://grafana.com/docs/agent/latest/configuration/traces-config/
https://grafana.com/docs/agent/latest/configuration/dynamic-config/
https://grafana.com/docs/agent/latest/configuration/integrations/
https://grafana.com/docs/agent/latest/configuration/create-config-file/
https://grafana.com/docs/agent/latest/configuration/
https://docs.gomplate.ca/datasources/
https://www.cnblogs.com/rongfengliang/p/14248068.html