1. **Metric Scraping (Pull Model)**: Instead of applications pushing metrics, the Prometheus server actively scrapes metrics from configured HTTP endpoints (e.g., `http://my-service:8080/metrics`) at configured intervals (e.g., every 15 seconds). 2. **Service Discovery**: Prometheus queries the cloud provider (AWS/GCP) or Kubernetes API to dynamically discover target endpoints as containers scale out. 3. **Data Storage**: Metrics are stored locally in a highly optimized custom Time Series Dat

Q27

# How it Works:

💬Answer

Metric Scraping (Pull Model): Instead of applications pushing metrics, the Prometheus server actively scrapes metrics from configured HTTP endpoints (e.g., http://my-service:8080/metrics) at configured intervals (e.g., every 15 seconds).
Service Discovery: Prometheus queries the cloud provider (AWS/GCP) or Kubernetes API to dynamically discover target endpoints as containers scale out.
Data Storage: Metrics are stored locally in a highly optimized custom Time Series Database (TSDB).
Data Querying (PromQL): Users query the metrics database using Prometheus Query Language (PromQL) to generate dashboards in Grafana.
Alerting: Prometheus evaluates alert rules against metrics; if a rule matches (e.g., up == 0), it pushes an alert to Alertmanager for routing and notifications.

Related Monitoring Questions

View All MonitoringQuestions →

How do you explain the difference between Monitoring and Observability? Are they the same thing?

What are the Four Golden Signals of Monitoring, and why are they critical?

How does the RED methodology help monitor request-driven microservices?

What is the USE methodology, and in what scenarios is it preferred over RED?

Created by

Apurv Gujjar

DevOps & Cloud Engineer

Specialized in:DevOpsAWSGCPKubernetesTerraformDocker

View Portfolio