引用：[https://prometheus.io/docs/prometheus/latest/configuration/configuration/](https://prometheus.io/docs/prometheus/latest/configuration/configuration/) 资料[https://github.com/Alrights/prometheus](https://github.com/Alrights/prometheus) Prometheus可以在运行时重新加载其配置。如果新配置格式不正确，则不会应用更改。通过向Prometheus进程发送`SIGHUP`或向`/-/reload`端点发送HTTP POST请求（启用`--web.enable-lifecycle`标志时）来触发配置reload。 这也将重新加载任何配置的规则文件。 prometheus使用的configmap修改后，k8s中的pod并不会自动重新读取，需要使用reload。 `curl -X POST "http://10.96.112.113:9090/-/reload"` **全局配置指定在所有其他配置上下文中有效的参数。 它们也作为其他配置部分的默认值。** * **global：全局配置** * **alerting：告警配置** * **rule_files：告警规则** * **scrape_configs：配置数据源**，称为target，每个target用**job_name**命名。又分为**静态配置和服务发现** ``` global: # How frequently to scrape targets by default. # 默认情况下抓取目标的频率. [ scrape_interval: <duration> | default = 1m ] # How long until a scrape request times out. # 抓取超时时间. [ scrape_timeout: <duration> | default = 10s ] # How frequently to evaluate rules. # 评估规则的频率. [ evaluation_interval: <duration> | default = 1m ] # The labels to add to any time series or alerts when communicating with # external systems (federation, remote storage, Alertmanager). # 与外部系统通信时添加到任何时间序列或警报的标签 #（联合，远程存储，Alertmanager）. external_labels: [ <labelname>: <labelvalue> ... ] # Rule files specifies a list of globs. Rules and alerts are read from # all matching files. # 规则文件指定了一个globs列表. # 从所有匹配的文件中读取规则和警报. rule_files: [ - <filepath_glob> ... ] - 'prometheus.rules' # A list of scrape configurations. # 抓取配置列表 scrape_configs: [ - <scrape_config> ... ] # insecure_skip_verify: true 开启不验证证书。 # Alerting specifies settings related to the Alertmanager. # 警报指定与Alertmanager相关的设置. alerting: alert_relabel_configs: [ - <relabel_config> ... ] alertmanagers: [ - <alertmanager_config> ... ] # Settings related to the remote write feature. # 与远程写入功能相关的设置. remote_write: [ - <remote_write> ... ] # Settings related to the remote read feature. # 与远程读取功能相关的设置. remote_read: [ - <remote_read> ... ] ```