Skip to main content

Architecture

Grafana-Architecture

Architecture

System Overview​

Grafana 12.0.2 is a comprehensive observability platform that provides visualization, alerting, and analytics capabilities for time-series data. The architecture follows a modular design with the following core components:

Core Components

  • Grafana Server: Main application server handling web interface and API

  • Database Layer: PostgreSQL/MySQL for metadata and configuration storage

  • Alerting Engine: Unified alerting system with rule evaluation and notification

  • Data Sources: Connectors to various data systems (Prometheus, InfluxDB, etc.)

  • Plugin System: Extensible plugin architecture for panels and data sources

  • Authentication Layer: Multi-provider authentication and authorization

Architecture Diagram​

┌─────────────────┐    ┌─────────────────┐    ┌─────────────────┐
│   Load Balancer │────│   Grafana Web   │────│   Grafana API   │
│     (Nginx)     │    │     Server      │    │     Server      │
└─────────────────┘    └─────────────────┘    └─────────────────┘
                                │                       │
                       ┌─────────────────┐    ┌─────────────────┐
                       │   Database      │    │   Data Sources  │
                       │ (PostgreSQL)    │    │ (Prometheus,    │
                       └─────────────────┘    │  InfluxDB, etc) │
                                              └─────────────────┘
                                                      │
                       ┌─────────────────┐    ┌─────────────────┐
                       │ Alerting Engine │    │ Plugin System   │
                       │                 │    │                 │
                       └─────────────────┘    └─────────────────┘

Technology Stack

  • Runtime: Go 1.21+ (backend), Node.js 18+ (frontend build)

  • Database: PostgreSQL 12+ / MySQL 8.0+ / SQLite 3.x

  • Frontend: Angular/React components with TypeScript

  • Container: Docker 24.x

  • Orchestration: Kubernetes 1.26+ (optional)

  • Caching: Redis 7.x (optional)

Key Features

  • Unified Alerting: Advanced alerting with multiple notification channels

  • Correlations: Link metrics across different data sources

  • Public Dashboards: Share dashboards without authentication

  • Canvas Panels: Advanced custom visualizations

  • AI-Powered Insights: Machine learning-based anomaly detection

Metrics For metrics we use NodeExporter > Prometheus > Grafana approach

  • We install Node Exporter agent in the target servers.

  • We use prometheus to scrape metrics at managed frequency and store in the path that was declared.

  • Then we integrate Grafana with prometheus by adding it as data source. We build dashboards to visualize the metrics in Grafana.

Logs For Logs we use promtail > Loki > Grafana

  • We need to install promtail agent in the target servers.

  • We use Loki to store the data that was pushed by promtail by attaching time-series metadata in the mentioned storage space.

  • Integrate Grafana with Loki by adding it as a data source. We use PromQL language to query logs depending on the need.