Prometheus Grafana

Prometheus & Grafana Monitoring

Core Concepts

Prometheus is the industry standard for cloud-native metrics monitoring and alerting.

1. The Pull Model

Prometheus actively "scrapes" (pulls) metrics from target services via HTTP endpoints (typically /metrics). This is the opposite of systems like StatsD which expect services to push data to them.

• Services must expose metrics in the Prometheus plaintext format using a client library.
• PromQL: A powerful functional query language used to extract time-series data.

2. Alertmanager & Grafana

• Alertmanager: Handles alerts generated by Prometheus. It deduplicates, groups, and routes them to receivers like Slack, PagerDuty, or Email.
• Grafana: The visualization layer. It queries Prometheus via PromQL and displays the data in rich, customizable dashboards.

Monitoring Architecture Map

%%{init: {"theme": "default", "flowchart": {"useMaxWidth": true}}}%%
flowchart TD
    subgraph Targets ["Application Workloads"]
        A["Microservice 1 (/metrics)"]
        B["Microservice 2 (/metrics)"]
        C["Node Exporter (CPU/RAM)"]
    end
    
    subgraph Prom_Stack ["Prometheus Ecosystem"]
        D["Prometheus Server (TSDB)"]
        E["Alertmanager"]
    end
    
    subgraph Visualization ["Dashboards & Alerts"]
        F["Grafana (Dashboards)"]
        G["Slack / PagerDuty"]
    end
    
    D -.->|"Scrape (Pull HTTP)"| A
    D -.->|"Scrape (Pull HTTP)"| B
    D -.->|"Scrape (Pull HTTP)"| C
    
    F -->|"Query (PromQL)"| D
    D -->|"Fire Alert Rules"| E
    E -->|"Route Notifications"| G