Error Codes Wiki

Prometheus Scrape Target Down — Metrics Endpoint Unreachable or Timeout

Warningsystem

Overview

Fix Prometheus scrape target showing DOWN status when the metrics endpoint is unreachable, returns errors, or times out during metric collection.

Key Details

  • Prometheus scrapes metrics from targets by sending HTTP GET requests to /metrics endpoints
  • Target status DOWN means Prometheus cannot successfully scrape the endpoint
  • Scrape failures can be due to network issues, target crashes, authentication, or timeout
  • The Targets page in Prometheus UI shows the status of all configured scrape targets
  • Service discovery (Kubernetes, Consul, DNS) can add targets that are not yet ready

Common Causes

  • Target application crashed or not exposing the /metrics endpoint
  • Network connectivity or DNS resolution failure between Prometheus and the target
  • Scrape timeout exceeded because the target takes too long to generate metrics response
  • Authentication required (bearer token, basic auth) but not configured in scrape config

Steps

  1. 1Check Prometheus UI Targets page: Status > Targets to see which targets are down and error messages
  2. 2Verify the endpoint is accessible: curl http://target:port/metrics from the Prometheus server
  3. 3Check the target application logs for crashes or /metrics endpoint errors
  4. 4Increase scrape_timeout in prometheus.yml if the target is slow to respond (default 10s)
  5. 5Verify service discovery configuration: ensure labels, namespaces, and selectors are correct

Tags

prometheusscrapemonitoringmetricstarget-down

More in System

Frequently Asked Questions

15-30 seconds is common. Shorter intervals (5s) give better resolution but increase storage and load. Longer intervals (60s) save resources but may miss short-lived anomalies. Match the interval to your alerting needs.