[Feature] Expose `prometheus.receive_http` received metrics

[lekmitiralf]

Component(s)

prometheus.receive_http

What's wrong?

Problem:
Metrics received by prometheus.receive_http (ex: prometheus_receive_http_receiver_classic_metrics_received_samples_total) are NOT exposed on Alloy's standard /metrics endpoint, blocking external monitoring.
Impact:
Impossible monitoring of proxy/gateway metrics without local TSDB (Mimir/Prometheus).

v0 API endpoints are internal debug only, not scrapable by external Prometheus.

Use Case:
Get recieved metrics from http through the UI
Or, get recieved metrics through a prometheus.echo that output in the /metrics or /api/v0/component/prometheus.echo./metrics
Or get metrics from a scraper

Steps to reproduce

use Alloy as a proxy.
Create a http receiver and send metrics from alloy agent to alloy proxy gateway
❌/api/v0/component/prometheus.receive_http.receiver_classic_metrics/metrics → 404
❌ /api/v0/component/prometheus.echo.received_metrics_exposure/metrics → 404
❌ /api/v0/component/prometheus.scrape.echo_scraper/metrics → 404
❌ no received metrics in the /metrics of alloy UI

System information

Linux 3.10.0-1160.145.1.el7.x86_64 #1 SMP x86_64 x86_64 x86_64 GNU/Linux

Software version

Grafana Alloy 1.13.2

Configuration

// activer le mode debug pour afficher toutes les métriques passantes par le proxy
livedebugging {
  enabled = true
}

// reception des métriques de traces 
otelcol.receiver.otlp "receiver_traces_metrics" {
        grpc {
	  endpoint = "0.0.0.0:4319"
	}
	http {
	  endpoint = "0.0.0.0:4320"
	}
	output {
	  metrics = [otelcol.exporter.otlphttp.sender_to_mimir_otlp.input] 
	}
}

otelcol.exporter.otlphttp "sender_to_mimir_otlp" {
   client {
     endpoint = "http://mimir-instance:3102/otlp"
     tls { insecure = true }
}
}

prometheus.receive_http "receiver_classic_metrics" {
  http {
    listen_address = "0.0.0.0"
    listen_port    = 3102
  }
  forward_to = [prometheus.remote_write.sender_to_mimir.receiver,
  prometheus.echo.received_metrics_exposure.receiver]  // ajouter pour exposer les métriques reçues // à supprimer une fois migration prom vers Thanos
}
// echo permet d'écrire les métriques reçues sur le receive http sur /metrics
prometheus.echo "received_metrics_exposure" { 
  format = "openmetrics"  
}
prometheus.scrape "echo_scraper" {
  targets = [
    { __address__ = "received.internal", job = "receiver_debug"
       },
    { __address__ = "echo.internal", job = "echo_debug" },
  ]
  scrape_interval = "15s"
  forward_to = [prometheus.remote_write.black_hole.receiver]
}

// remote write pour pourvoir scrapper les métriques reçues seulement // à supprimer après migration
prometheus.remote_write "black_hole" {
        endpoint {
                url = "localhost:0"                                                                      
        }
          
        wal {
        truncate_frequency = "1h" // nettoyage des wal toutes les x h
        min_keepalive_time = "5m"  // garde les fichiers au moins x min
        max_keepalive_time = "1h"  // garde au maximum x temps de backlog
        }
}

// envoi vers mimir/thanos à commenter si pas d'envoi
prometheus.remote_write "sender_to_mimir" {
        endpoint {
                url = "http://mimir-instance:3102/api/v1/push"
  
        }

        wal {
        truncate_frequency = "5h" // nettoyage des wal toutes les x h
        min_keepalive_time = "5m"  // garde les fichiers au moins x min
        max_keepalive_time = "24h"  // garde au maximum x temps de backlog
        }
}

Logs

Tip

_{React with 👍 if this issue is important to you.}