Skip to content

Lesson 5: Observability (Prometheus & PromQL)

🧠 The Concept (Explain Like I'm 5)

If you drive a car without a dashboard, you won't know you're out of gas until the engine stops. Prometheus is the engine sensor that records speed and gas levels every 5 seconds. Grafana is the beautiful dashboard that shows you the dials.

🏢 The Enterprise Context

  • Pull-based Metrics: Prometheus doesn't wait for apps to send data. It actively reaches out to a /metrics web page on every Pod every 15 seconds and downloads the numbers (Scraping).
  • PromQL: The query language used to ask complex questions, like "Give me the 99th percentile HTTP response time over the last 5 minutes."

🗺️ Visual Architecture: The Prometheus Scraping Flow

flowchart LR
    subgraph Apps ["Kubernetes Pods"]
        App1["Payment Service<br/>(Exposes /metrics)"]
        App2["Auth Service<br/>(Exposes /metrics)"]
    end

    Prom["Prometheus Server"] -->|HTTP GET /metrics<br/>every 15s| Apps
    Prom --> TSDB[("Time Series DB")]

    Grafana["Grafana Dashboard"] -->|PromQL Queries| TSDB
    Prom -->|Threshold Exceeded| Alert["Alertmanager (PagerDuty)"]