prometheus

prometheus

prometheus sever

1. Time series databases

2. Data retrieval worker
   targets and units (cpu, memory)

3. Http server

Metrics

两类属性 TYPE 和 HELP
Help是注释
types有 1. counter (How many times X happened) 2. Gauge (What is the current value) 3. Histogram ()

Data retrieval worker

Pull over http endpoints
如果没有/metrics，自己实现的话。那么需要exporter.
exporter会 1. fetches metrics 2. convert formats 3. expose /metrics

service <- exporter <- retrieval worker

Monitoring linux server: node exporter 下载执行会执行node的metrics, 然后configure prometheus去scrape这些metrics.

push的好处

1. 多个instance可以来获取metrics
2. 更好检测是否up

short-lived job

可以发到pushgateway 暂存metrics, 然后prometheus 之后来获取这些metrics.

一些其他的默认参数：

metrics_path: "/metrics"
scheme: "http"