如何在prometheus.yml中配置多个抓取URL？配置问题求助

问题分析 #

你当前的配置存在关键错误：

1. node_exporter端口写错：你提供的node_exporter指标地址是localhost:9182/metrics，但配置里写的是localhost:9100，这会导致无法采集node_exporter的指标
2. 多组件混在同一job里虽可行，但拆分job更便于后续监控维护（可选优化）

推荐配置（拆分job，清晰易维护） #

scrape_configs:
  # 采集Prometheus自身指标
  - job_name: "prometheus"
    metrics_path: "/metrics"
    static_configs:
      - targets: ["localhost:9090"]
        labels:
          group: 'system'
  # 采集node_exporter指标
  - job_name: "node_exporter"
    metrics_path: "/metrics"
    static_configs:
      - targets: ["localhost:9182"]
        labels:
          group: 'system'
  # 采集Wildfly自身指标
  - job_name: "wildfly"
    metrics_path: "/metrics"
    static_configs:
      - targets: ["localhost:9990"]
        labels:
          group: 'system'
  # 采集Wildfly内应用的指标
  - job_name: "application"
    metrics_path: "/application/rest/metrics"
    static_configs:
      - targets: ["localhost:8080"]
        labels:
          group: 'application'

简化修正版（保留原job结构，仅修复端口错误） #

scrape_configs:
  - job_name: "prometheus"
    metrics_path: "/metrics"
    static_configs:
      - targets: ["localhost:9182", "localhost:9990", "localhost:9090"]
        labels:
            group: 'system'
  - job_name: "application"
    metrics_path: "/application/rest/metrics"
    static_configs:
      - targets: ["localhost:8080"]
        labels:
            group: 'application'

验证步骤 #

1. 修改配置后重启Prometheus服务
2. 访问Prometheus UI（默认http://localhost:9090），进入Status -> Targets页面，确认所有target状态为UP
3. 若某个target显示DOWN，直接点击对应的Endpoint链接，确认指标接口是否能正常返回数据

内容的提问来源于stack exchange，提问作者Christian13467