13.4 Reading Prometheus Source | Reading & Contributing to Open Source

Repository Structure

prometheus/prometheus:

Directory	What's There
`cmd/prometheus/`	Main binary
`scrape/`	Scrape loop (HTTP polling)
`storage/`	TSDB (time series database)
`promql/`	Query language engine
`web/`	HTTP API and UI

How Scraping Works

scrape/manager.go
  → Creates a scrape pool per target group
  → Each pool runs N scrapers (goroutines)
  → Each scraper:
    1. Sleeps for the scrape interval
    2. HTTP GET to the /metrics endpoint
    3. Parses Prometheus exposition format
    4. Writes samples to TSDB

This is a worker pool pattern (Module 7) with rate limiting (scrape interval).

How Exporters Work

An exporter is an HTTP server (Module 8) that exposes a /metrics endpoint:

import "github.com/prometheus/client_golang/prometheus"

var httpRequests = prometheus.NewCounterVec(
    prometheus.CounterOpts{
        Name: "http_requests_total",
        Help: "Total HTTP requests.",
    },
    []string{"method", "path", "status"},
)

func init() {
    prometheus.MustRegister(httpRequests)
}

// In your handler:
httpRequests.WithLabelValues("GET", "/api/pods", "200").Inc()

Everything you learned in this course comes together: structs (metrics), interfaces (collectors), HTTP servers (exporter), concurrency (scrape pools).