How to Set Up Monitoring with Grafana and Prometheus: A Complete Self-Hosting Guide

To set up monitoring with Grafana and Prometheus, deploy both services using Docker Compose, configure Prometheus to scrape metrics from HTTP /metrics endpoints, and connect Grafana to Prometheus as a data source to visualize time-series data and configure alerts.

The mikeroyal/Self-Hosting-Guide repository, specifically in its README.md monitoring section, provides a production-ready approach to building a scalable monitoring stack using these open-source tools. According to the guide, Prometheus operates as a time-series database that pulls metric data from any HTTP-exposed endpoint, while Grafana connects to Prometheus via the PromQL API to provide visualization and alerting capabilities. This article implements the exact configuration patterns found in the repository to deploy a fully functional monitoring solution.

Architecture Overview

The monitoring stack consists of two core components working together. Prometheus runs a time-series database (TSDB) that scrapes metrics from configured targets using the pull model. It collects raw samples, performs aggregation, and exposes a flexible query language called PromQL via HTTP API.

Grafana functions as the visualization layer that queries Prometheus to render charts, dashboards, and alerts. It reads metrics through the Prometheus HTTP API and can combine data from multiple sources, including Loki logs and VictoriaMetrics.

Typical data flow follows this pattern: exporters (such as node_exporter) expose metrics at /metrics endpoints → Prometheus scrapes and stores the data → Grafana queries Prometheus to display dashboards.

Step-by-Step Setup Guide

Step 1: Install Prometheus via Docker

Launch Prometheus using the official Docker image. The container requires a mounted configuration file and specific command-line flags to define the configuration path and storage location.

Create a docker-compose.yml file in your project directory:

version: "3.8"

services:
  prometheus:
    image: prom/prometheus:latest
    container_name: prometheus
    volumes:
      - ./prometheus.yml:/etc/prometheus/prometheus.yml:ro
    command:
      - "--config.file=/etc/prometheus/prometheus.yml"
      - "--storage.tsdb.path=/prometheus"
    ports:
      - "9090:9090"
    networks:
      - monitoring

networks:
  monitoring:
    driver: bridge

The --config.file flag points to /etc/prometheus/prometheus.yml inside the container, while --storage.tsdb.path defines where Prometheus stores its time-series data.

Step 2: Configure Scrape Targets

Create a prometheus.yml file to define your scrape configurations. This file controls which targets Prometheus monitors and how often it collects data.

global:
  scrape_interval: 15s
  evaluation_interval: 15s

scrape_configs:
  - job_name: "prometheus"
    static_configs:
      - targets: ["localhost:9090"]

  - job_name: "node"
    static_configs:
      - targets: ["host.docker.internal:9100"]

The scrape_interval of 15s tells Prometheus to pull metrics from each target every 15 seconds. The evaluation_interval controls how often Prometheus evaluates alerting rules.

Step 3: Install and Configure Grafana

Add the Grafana service to your docker-compose.yml file. Grafana must connect to the same Docker network as Prometheus to enable service discovery.

  grafana:
    image: grafana/grafana:latest
    container_name: grafana
    depends_on:
      - prometheus
    ports:
      - "3000:3000"
    environment:
      - GF_SECURITY_ADMIN_PASSWORD=admin
    volumes:
      - ./grafana/provisioning:/etc/grafana/provisioning
    networks:
      - monitoring

The depends_on directive ensures Prometheus starts before Grafana. Port 3000 exposes the Grafana web interface, while the environment variable sets the admin password.

Step 4: Provision the Prometheus Data Source

Automate Grafana configuration by creating provisioning files. Create grafana/provisioning/datasources/prometheus.yaml to register Prometheus as a data source automatically:

apiVersion: 1

datasources:
  - name: Prometheus
    type: prometheus
    access: proxy
    url: http://prometheus:9090
    isDefault: true
    editable: true

This configuration uses access: proxy to route queries through the Grafana backend and points to the Prometheus service URL at http://prometheus:9090.

Step 5: Configure Dashboard Provisioning

Create grafana/provisioning/dashboards/dashboard.yaml to enable automatic dashboard loading:

apiVersion: 1

providers:
  - name: "default"
    folder: ""
    type: file
    options:
      path: /etc/grafana/provisioning/dashboards

Place any JSON dashboard files in the ./grafana/provisioning/dashboards/ directory on your host, and Grafana will load them automatically on startup.

Step 6: Launch the Stack

Execute the following command to start both services:

docker compose up -d

Access Prometheus at http://localhost:9090 to verify targets are being scraped. Access Grafana at http://localhost:3000 using the credentials admin/admin.

Import the Node Exporter Full dashboard (ID 1860) or copy pre-bundled JSON files into your provisioning directory to visualize system metrics immediately.

Summary

• Prometheus pulls metrics from HTTP /metrics endpoints using the scrape configuration defined in prometheus.yml and stores them in a time-series database.
• Grafana connects to Prometheus at http://prometheus:9090 to query metrics using PromQL and render visualization dashboards.
• The docker-compose.yml file orchestrates both services on a shared Docker network named monitoring for seamless communication.
• Provisioning files in grafana/provisioning/datasources/prometheus.yaml and grafana/provisioning/dashboards/dashboard.yaml automate the initial setup and configuration.
• Node Exporter provides host-level metrics such as CPU, memory, and disk usage when configured as a scrape target in prometheus.yml.

Frequently Asked Questions

What is the difference between Prometheus and Grafana?

Prometheus is a time-series database that collects and stores metrics using a pull-based model, while Grafana is a visualization platform that queries data sources like Prometheus to create dashboards and alerts. Prometheus handles data collection and storage, whereas Grafana focuses on the presentation layer and user interface.

How do I secure my Grafana installation?

Change the default administrator password by modifying the GF_SECURITY_ADMIN_PASSWORD environment variable in your docker-compose.yml file. For production environments, configure HTTPS using a reverse proxy like Nginx or Traefik, and enable Grafana's built-in authentication providers or LDAP integration.

Can Prometheus monitor services running outside of Docker?

Yes, Prometheus can monitor any service that exposes a metrics endpoint via HTTP, regardless of whether it runs inside Docker, on a virtual machine, or on bare metal. Update the targets array in your prometheus.yml scrape configurations to include the IP addresses and ports of external services or exporters.

How do I set up alerting in this monitoring stack?

Define alerting rules in a separate alerting_rules.yml file and reference it in your prometheus.yml configuration under the rule_files section. Configure Alertmanager to handle notifications, or use Grafana's alerting feature by creating alert rules directly in the Grafana UI that evaluate PromQL queries against your Prometheus data source.

{
  "mcpServers": {
    "instagit": {
      "command": "npx",
      "args": ["-y", "instagit@latest"]
    }
  }
}