SRE Thailand Bootcamp

Week 1: Monitoring Fundamentals — Prometheus

Theme: Understand Prometheus architecture, instrument applications, write PromQL queries.

Learning Objectives

1. Explain Prometheus architecture (pull model, TSDB, service discovery, federation)
2. Install and configure Prometheus using helm
3. Instrument a Go HTTP server with Prometheus client library
4. Write PromQL queries (rate, irate, histogram_quantile, aggregations)

Reading Materials

• Prometheus Overview
• Prometheus Architecture
• Prometheus Data Model
• PromQL Basics
• PromQL Operators
• Prometheus Best Practices
• Go Client Library Guide
• Google SRE Book - Chapter 6: Monitoring

Finger Exercises (choose 2)

1. Install Prometheus via helm install on a local kind/minikube cluster. Verify targets are up.
2. Write a simple Go HTTP server that exposes custom metrics (http_requests_total, http_request_duration_seconds histogram). Scrape with Prometheus.
3. Practice 10 PromQL queries on the Prometheus demo instance at demo.promlabs.com. Start with rate(promhttp_metric_handler_requests_total[5m]) and work up to histogram_quantile(0.95, ...).

Project

Build a Go HTTP server with /api/users, /api/orders, /healthz endpoints. Expose:

• http_requests_total{method, endpoint, status}
• http_request_duration_seconds (histogram)
• Custom business_orders_created_total
• up / health metric

Deploy to kind, configure Prometheus scrape, verify in Prometheus UI.

Time Budget